The Strength of Weak Ties Revisited

[Sociological Theory, Volume 1 (1983), 201-233.]

In this chapter I review empirical studies directly testing the
hypotheses of m y 1973 paper " T h e Strength of Weak Ties"
(hereafter " S W T " ) and work that elaborates those hy-
potheses theoretically or uses them to suggest new empirical
research not discussed in my original formulation. Along
the way, I will reconsider various aspects of the theoretical
argument, attempt to plug some holes, and broaden its
base.

T H E S T R E N G T H O F WEAK TIES:
A NETWORK THEORY REVISITED
Mark Granovetter
S T A T E UNIVERSITY O F N E W Y O R K ,
STONY BROOK

The Argument Recapitulated

T h e argument asserts that our acquaintances (weak ties) are less
likely to be socially involved with one another than are our close
friends (strong ties). T h u s the set of people made u p of any individual

I a m indebted to Everett Rogers w h o first suggested this study, inviting
it for a special session of the International Communications Association meet-
ings o n the weak-ties hypothesis. T h e first version was delivered at this session
in Acapulco o n May 21, 1980. A version closer to the present one was delivered
a t the conference o n Structural Analysis, April 4, 1981, SUNY-Albany. I a m
indebted to participants in these two conferences for their generous
comments- especially Fernando Morett, Scott Feld, Nan Lin, a n d Ronald
Rice. T h i s chapter has also drawn heavily from the comments and literature
review of Ellen Granovetter.

202 Sociological Theory

and his or her acquaintances comprises a low-density network (one in
which many of the possible relational lines are absent) whereas the set
consisting of the same individual and his or her close friends will be
densely knit (many of the possible lines are present).
T h e overall social structural picture suggested by this argument
can be seen by considering the situation of some arbitrarily selected
individual- call him Ego. Ego will have a collection of close friends,
most of whom are in touch with one another- a densely knit c l u m p of
social structure. Moreover, Ego will have a collection of acquaintances,
few of whom know one another. Each of these acquaintances, however,
is likely to have close friends in his own right and therefore to be
enmeshed in a closely knit clump of social structure, but one different
from Ego's. T h e weak tie between Ego and his acquaintance, therefore,
becomes not merely a trivial acquaintance tie but rather a crucial
bridge between the two densely knit clumps of close friends. T o the
extent that the assertion of the previous paragraph is correct, these
clumps would not, in fact, be connected to one another at all were it
not for the existence of weak ties (SWT, p. 1363).
It follows, then, that individuals with few weak ties will be
deprived of information from distant parts of the social system and will
be confined to the provincial news and views of their close friends. This
deprivation will not only insulate them from the latest ideas and
fashions but may put them in a disadvantaged position in the labor
market, where advancement can depend, as I have documented else-
where (1974), on knowing about appropriate job openings at just the
right time. Furthermore, such individuals may be difficult to organize
or integrate into political movements of any kind, since membership in
movements or goal-oriented organizations typically results from being
recruited by friends. While members of one or two cliques may be
efficiently recruited, the problem is that, without weak ties, any
momentum generated in this way does not spread beyond the clique.
As a result, most of the population will be untouched.
T h e macroscopic side of this communications argument is that
social systems lacking in weak ties will be fragmented and incoherent.
New ideas will spread slowly, scientific endeavors will be handicapped,
and subgroups separated by race, ethnicity, geography, or other
characteristics will have difficulty reaching a modus vivendi. These

Strength of Weak Ties 203

themes are all taken u p in greater detail, with supporting evidence, in
SWT.
I now wish to review the past eight years' literature on weak ties.
First, I will review work focusing on the impact of weak ties on indi-
viduals, then work relating to the flow of ideas and the sociology of
science, and, finally, work evaluating the role of weak ties in affecting
cohesion in complex social systems.

The Impact of Weak Ties on Individuals

An early draft of S W T was entitled "Alienation Reconsidered:
T h e Strength of Weak Ties." In this draft I argued that weak ties, far
from creating alienation, as one might conclude from the Chicago
school of urban sociology-especially from Louis Wirth-are actually
vital for a n individual's integration into modern society. Upon further
reflection it is clear that this argument is closely related to certain
classic themes in sociology. In the evolution of social systems, perhaps
the most important source of weak ties is the division of labor, since
increasing specialization and interdependence result in a wide variety
of specialized role relationships in which one knows only a small seg-
ment of the other's personality. (See the perceptive comments of Sim-
mel, 1950, pp. 317-329.) In contrast to the emphasis of Wirth, and also
Toennies, that role segmentation results in alienation, is the Durkheim-
ian view that the exposure to a wide variety of different viewpoints
and activities is the essential prerequisite for the social construction of
individualism.
In a provocative article, Rose Coser (1975) takes u p some of these
themes. She describes the complexity of role set-to use Robert Mer-
ton's expression for the plurality of others with whom one has role
relations-as a "seedbed of individual autonomy." In Simmel's view,
she recalls, "the fact that an individual can live u p to expectations of
several others in different places and at different times makes it possible
to preserve an inner core, to withhold inner attitudes while conforming
to various expectations" ( p . 241). Furthermore, persons "deeply
enmeshed in a Gemeinschaft may never become aware of the fact that
their lives d o not actually depend on what happens within the group
but on forces far beyond their perception and hence beyond their con-
trol. T h e Gemeinschaft may prevent individuals from articulating


their roles in relation to the complexities of the outside world. Indeed,
there may be a distinct weakness i n strong ties" ( p . 242).
Coser then elaborates the cognitive ramifications of this conun-
drum: "In a Gemeinschaft everyone knows fairly well why people
behave in a certain way. Little effort has to be made to gauge the
intention of the other person. . . . If this reasoning is correct . . . the
manner of communication will tend to be different in a Gesellschaft.
Hence, the type of speech people use should differ in these two types of
structures" ( p . 254). She relates this difference to Basil Bernstein's dis-
tinction between restricted and elaborated codes of communication.
Restricted codes are simpler- more meanings are implicit and taken
for granted as the speakers are so familiar with one another. Elaborated
codes are complex and universal-more reflection is needed in organiz-
ing one's communication "when there is more difference between those
to whom the speech is addressed" ( p . 256). While some weak ties may
connect individuals who are quite similar, of course, there is, as I
pointed out in SWT, "empirical evidence that the stronger the tie
connecting two individuals, the more similar they are, in various
ways" ( p . 1362). T h u s Coser's argument applies directly to the distribu-
tion of weak and strong ties. She concludes that in "elaborated speech
there is a relatively high level of individualism, for it results from the
ability to put oneself in imagination in the position of each role
partner in relation to all others, including oneself" ( p . 257). She goes
on to argue that the social structure faced by children of lower socio-
economic backgrounds does not encourage the complex role set that
would, in turn, facilitate the development of "intellectual flexibility
and self-direction" (p. 258).
T h i s discussion casts a different light on some of the arguments
of SWT. There I argued that while West Enders, for example, did have
some weak ties, they were embedded within each individual's existing
set of strong ties rather than bridging to other groups. I interpreted this
lack of bridging as inhibiting organization because it led to overall
fragmentation and distrust of leaders. Coser's argument suggests
further that bridging weak ties, since they d o link different groups, are
far more likely than other weak ties to connect individuals who are
significantly different from one another. T h u s , in addition to the over-
all macrostructural effect of bridging weak ties, I could also have
argued that they are exactly the sort of ties that lead to complex role sets


a n d the need for cognitive flexibility. T h e absence of flexibility may
have inhibited organization against urban renewal, since the ability to
function in complex voluntary organizations may depend on a habit of
mind that permits one to assess the needs, motives, and actions of a
great variety of different people simultaneously.
There is no special reason why such an argument should apply
only to lower socioeconomic groups; it should be equally persuasive
for any set of people whose outlook is unusually provincial as the
result of homogeneous contacts. In American society there is thus some
reason for suggesting that upper-class individuals as well as lower-class
people may suffer a lack of cognitive flexibility. Baltzell (1958) and
others have described in detail the cloistered features of upper-class
interaction; Halberstam (1972) has suggested that such a social struc-
ture creates inflexibility in the form of arrogance and a sense of infalli-
bility and had much to d o with American involvement in the Vietnam
War.
At a more mundane level, I argued (SWT, pp. 1369-1373; 1974,
pp. 51-62) that weak ties have a special role in a person's opportunity
for mobility- that there is a "structural tendency for those to whom
one is only weakly tied to have better access to job information one does
not already have. Acquaintances, as compared to close friends, are more
prone to move in different circles than oneself. Those to whom one is
closest are likely to have the greatest overlap in contact with those one
already knows, so that the information to which they are privy is likely
to be much the same as that which one already has" (1974, pp. 52-53).
In my empirical study of recent job changers (1974), I found, in fact,
that if weak ties are defined by infrequent contact around the time
when information about a new job was obtained, then professional,
technical, and managerial workers were more likely to hear about new
jobs through weak ties (27.8 percent) than through strong ones (16.7
percent), with a majority in between (55.6 percent).
Three pieces of empirical research offer partial confirmation of
this argument. Langlois (1977) studied a large sample of men and
women in a branch of the Quebec provincial government. Langlois
notes that even though this branch had "attempted to formalize the
recruitment of its members as much as possible" ( p . 217), 42.7 percent
of the 2,553 individuals in the sample found their jobs through per-
sonal contacts. Using frequency of recent contact as the definition of tie


strength (but with slightly different cutting points from mine), he also
found that weak ties were indeed often the ones that resulted in a new
job, but the pattern varied strongly by occupation. Administrative o r
managerial employees had a pattern very much like the one I reported:
35.5 percent using weak ties, 15.8 percent strong ones, and 48.7 percent
intermediate. Professionals and office workers also were heavy users of
weak ties (30.8 percent and 25.8 percent but, unlike managers, used
strong ties even more frequently (51.0 and 44.4 percent). Semiprofes-
sionals found only 13.1 percent of jobs through weak ties and blue-
collar workers 19.1 percent; the former found 44.9 percent of jobs
through strong ties, the latter only 19.1 percent.
Ericksen and Yancey (1980) studied a probability sample of 1,780
adults aged sixty-five and under living in the Philadelphia area in
1975. Respondents who had significant help from another person in
finding their current job were classified as having used ties (1980,
p. 14). If the person providing the help was identified as a relative or
friend of the respondent, the tie was considered strong. If the person
was classified as a n acquaintance, the tie was considered weak.
Ericksen and Yancey note that most acquaintances were work-related
and about two thirds of the strong ties were relatives. A majority of
respondents used some form of personal connection to land the job. Of
those who were not self-employed, 41.1 percent used strong ties, 15.6
percent weak ties, and 43.3 percent formal means or direct application
(p. 15). It is hard to compare this classification of ties to my
trichotomy- the operational definitions are different, there are two
categories instead of three, and the population at risk here is of broader
socioeconomic background.
One set of results is of special interest, however. Ericksen a n d
Yancey found that less-well-educated respondents were those most
likely to use strong ties for jobs: " The rate drops among respondents
who attended college and is balanced by a correspondingly large
increase in the likelihood of using weak ties and a slight increase in the
use of bureaucratic procedures" ( p . 24). In fact, 31 percent of managers
used weak ties in finding jobs, a figure close to that found by Langlois,
though 30 percent used strong ties, a larger figure than in the Canadian
sample. Regression analysis was then implemented to detemine
whether the strength of ties used had any impact on income (net of other
variables). Results indicated that the use of strong ties had n o consis-

Strength of Weak Ties

tent impact; for weak ties, the overall effect on income is substantial and
negative-opposite to the predictions of the weak-ties hypothesis. But
there is a significant weak-ties/education interaction ( p p . 24-25):
"Weak ties actually lead to a reduction in income among the poorly
educated, but . . . this reduction grows smaller with increasing levels
of education such that there is a small increase among high school
graduates . . . and this increase grows larger with further increases in
education. T h u s , for that group of well-educated respondents where
weak ties are most likely to be used we see that the effects of using the
weak ties are most positive."
L i n , Ensel, and Vaughn (1981) use similar definitions of weak
and strong ties to probe the relation between tie strength and occupa-
tional status attainment for a representative sample of men aged twenty
to sixty-four in an urban area of upstate New York. Those ties identi-
fied by respondents as acquaintances or friends of friends were classi-
fied as weak whereas friends, relatives, or neighbors were considered
strong ties. Their method was essentially similar to that used by
researchers such as Blau, Duncan, and Featherman-the construction of
structural equation models, or path analyses, to measure the relative
contribution of different independent variables to some dependent vari-
able, in this case occupational status (as measured by the Duncan Socio-
economic Index). Their central finding was this: T h e use of weak ties
in finding jobs has a strong association with higher occupational
achievement only insofar as the weak ties connect the respondent to a n
individual who is well placed in the occupational structure. This con-
clusion is illustrated in the path diagram from their article (Figure 1).
For the first job, the direct combination of tie strength is negligi-
ble; for the current one it is larger but still less than the indirect effect.
T h e indirect effect is large because the great majority of weak ties used
in finding jobs connected respondents to high-status individuals: 76.2
percent of weak ties (compared to 28.9 percent of strong ones) for the
first job and 70.7 percent (compared to 42.9 percent of strong ones) for
the current job were to informants of high occupational status (defined
as a score of 61 to 96 on the Duncan scale). T h e most likely interpreta-
tion of these findings is that weak ties are more efficient at reaching
high-status individuals, so that if such ties are available they are pre-
ferred. But since only 34 percent of jobs in this sample were found
through weak ties (among those whose job was found through social


Figure 1. Effects of Tie Strength and Source Status on Attained
Occupational Status.

OTY
-0.292 0.645

STY Y
-0.117

R 2 = 0.473
Y

Current Job
STY= strength of tie

OTY = source status

Y = attained status

ties) it appears that many individuals had no choice but to fall back on
strong ties.
These studies clarify the circumstances under which weak ties
provide unusual advantage. T h e argument of SWT implies that only
bridging weak ties are of special value to individuals; the significance
of weak ties is that they are far more likely to be bridges than are strong
ties. It should follow, then, that the occupational groups making the
greatest use of weak ties are those whose weak ties do connect to social
circles different from one's own. In Langlois's Canadian study, the
most frequent users are managers and professionals-just the catego-
ries, to use Robert Merton's terms, most likely in an organization to be
cosmopolitans rather than locals and most likely to deal with acquain-
tances in other organizations or other branches of the same organiza-
tion. Homans has argued that high-status individuals are more likely
in all groups to have contacts outside the group (1950, pp. 185-1 86).
Ericksen and Yancey too find managers to be the group with the
highest frequency of jobs found through weak ties. How should we
interpret the interaction effect, in their data, between weak ties and
education in determining income? I suggest that in lower socioeco-
nomic groups, weak ties are often not bridges but rather represent
friends' or relatives' acquaintances; the information they provide
would then not constitute a real broadening of opportunity-reflected


in the fact that the net effect of using such ties on income is actually
negative. In higher groups, by contrast, weak ties do bridge social
distance; thus if there are n o lucrative job openings known to one's own
social circle at a given moment, one may still take advantage of those
known in other circles. Here the net effect of weak ties on income is
strongly positive.
Consistent with this interpretation is the finding of Lin and col-
leagues (1981) that weak ties have positive effects on occupational sta-
tus only when they connect one to high-status individuals. For those of
lower status, weak ties to those of similar low status were not especially
useful, whereas those to high-status contacts were. In the latter case the
status difference alone strongly suggests that the ties bridged substan-
tial social distance. When high-status respondents use weak ties of
similar status, there is no status difference to seize on for evidence that
such ties bridge; here we must speculate that the hypothesized tendency
of high-status individuals to have more bridges among their weak ties
is in effect.
These interpretations, though consistent with the data, could be
better supported by detailed field reports of the exact circumstances
under which respondents used weak ties. Some findings, such as Lang-
lois's, of great strong-tie use by professionals and little weak-tie use by
semiprofessionals are simply not explained by the arguments of S W T
and thus await further speculation.

Excursus on the Strength of Strong Ties

Lest readers of S W T and the present study ditch all their close
friends and set out to construct large networks of acquaintances, I had
better say that strong ties can also have value. Weak ties provide people
with access to information and resources beyond those available in
their own social circle; but strong ties have greater motivation to be of
assistance and are typically more easily available. I believe that these
two facts d o much to explain when strong ties play their unique role.
A general formulation is suggested by Pool (1980), who argues
that whether one uses weak or strong ties for various purposes depends
not only on the number of ties one has at various levels of tie strength
but also on the utility of ties of different strength. T h u s people for
whom weak ties are much more useful than strong ties may still be


constrained to use the latter if weak ties make u p an extremely small
portion of their contacts; conversely, one for whom strong ties are more
useful may be socially isolated and forced to fall back on weak ones.
T h u s the analytic task is to identify factors affecting these variations.
Pool argues, for example, that the number of weak ties is increased by
the development of the communications system, by bureaucratization,
population density, and the spread of market mechanisms. Further, he
suggests that average family size affects the number of weak ties, since
where "primary families are large, more of the total contacts of an
individual are likely to be absorbed in them" (p. 5 ) .
Peter Blau has suggested that since the class structure of modern
societies is pyramidal, and since we may expect individuals at all levels
to be inclined toward homophily- the tendency to choose as friends
those similar to oneself-it follows that the lower one's class stratum,
the greater the relative frequency of strong ties. This happens because
homophilous ties are more likely to be strong and low-status individu-
als are so numerous that it is easier for them to pick and choose as
friends others similar to themselves.' A literal interpretation of this
comment would lead us to expect upper-status individuals to have
large numbers of weak ties, since there are so few others of high status;
it would further follow that many of these weak ties would then be to
others of lower status, since the latter would be so numerous. T h i s
conclusion does not accord with ethnographic accounts of upper-class
life that stress the importance of strong ties to other members of the
upper class. But it does suggest why the upper class must invest so
much in institutions such as private clubs, special schools, and social
registers; the effort to maintain a network of homophilous strong ties is
more difficult here than for lower strata. (See, for example, Baltzell,
1958.)
Important as it may be to know what an individual's total net-
work looks like before we can assess the meaning of that person's use of
a strong or weak tie, there exist few data that allow us to take this factor
into account. It may be that recent work on the sampling of large social
networks will allow us to make progress in this area. (See, for example,
Granovetter, 1976, and Frank, 1981).
More can be said about the value to individuals of ties of differ-
ent strength. Here Pool observes that "the utility of weak links is a
function of the security of the individual, and therefore of his wealth. A


highly insecure individual, for example a peasant who might starve if
his crop fails, is under strong pressure to become dependent upon one
or a few strongly protective individuals. A person with resources o n
which he can fall back can resist becoming dependent on any given
other individual and can explore more freely alternative options"
(1980, p. 5).
T h i s hypothesis can be strongly supported empirically. In my
study of job finding, for example, I found that those whose job was
found through strong ties were far more likely to have had a period of
unemployment between jobs than those using weak ties (1974, p. 54). I
suggested, moreover, that those in urgent need of a job turned to strong
ties because they were more easily called on and willing to help, how-
ever limited the information they could provide. Murray, Rankin, and
Magill (1981) studied social and physical scientists at one Canadian
and one American university: Most found jobs through strong rather
than weak ties. They interpret this finding as contradicting my
hypothesis that salient job information derives from weak ties ( p . 119).
But more than 80 percent of their data concern first academic jobs—
situations of considerable insecurity for new Ph.D.s who have few use-
ful contacts in their discipline as yet and typically rely on mentors and
dissertation advisers who know them and their work well. (This is the
definition of strong tie used by Murray and colleagues.) They d o find
that the proportion using strong ties for jobs subsequent to the first is
still high-47 percent versus 58 percent for first jobs-but the data for
the 47 percent consist of about fifty individuals, in one university,
where response rate barely exceeded 50 percent. Even if the figure were
representative, it would need to be disaggregated by career stage; thus
the present hypothesis suggests that as professors move further away
from their first academic job, their reliance on strong ties should
decline. T h e question o f whether respondents face unemployment or
not also would be relevant here; when individuals are denied tenure,
for example, one would expect a greater reliance on strong ties, other
things being equal, than if it were not strictly necessary to find a new
job.
A purely theoretical model from economics bears directly on this
question: Boorman (1975) used economic theory and network ideas to
suggest when rational economic actors might choose to allocate their
time a n d energy to weak ties as compared to strong ones. H e assumes


that strong ties require more time to maintain than weak ones and that
if one hears of a job, one offers the information to strong ties (if any are
unemployed) and otherwise to weak ties. These simple assumptions
lead to a complex mathematical model. T h e results, however, can be
summed u p simply: If the probability of unemployment in the system
is low, rational individuals will invest all their time in weak ties and
such a situation will be a Pareto-optimal equilibrium; for a high prob-
ability of unemployment, on the other hand, the only stable equilib-
rium is one in which only strong ties are maintained, though such an
equilibrium is not Pareto-optimal. (That is, the situation of some
actors could be improved without that of any others being worsened.)
T h i s model, based on assumptions and ideas entirely different from
those reported in my theoretical and empirical work, ends u p with
remarkably similar conclusions-which suggests that more attention
ought to be paid in such studies to the level of employment security
enjoyed by different participants in the labor market. (Doorman's
model is elaborated and extended in important ways by Delaney 1980.)
Employment difficulty is not the only occasion that prompts the
use of strong ties. All sorts of everyday problems have this result. Sum-
ming u p studies of helping networks in a Toronto suburb, Wellman
indicates that the "closer (stronger) the intimate relationship (as meas-
ured by the respondents' ordinal ranking of the intimates), the more
the perceived availability of help becomes a salient defining compo-
nent of that tie. Closeness is apparently the single most important
defining characteristic of helpful intimate relationships; . . . 56 per-
cent of the first closest ranked intimates are relied on in emergencies
. . . while only 16 percent of the sixth closest intimates are" (Wellman,
1979, pp. 1222-1223).
A number of studies indicate that poor people rely more on
strong ties than do others. Ericksen and Yancey, in a study of Philadel-
phia, conclude that the "structure of modern society is such that some
people typically find it advantageous to maintain strong networks and
we have shown that these people are more likely to be young, less well
educated, and black" (1977, p . 23). In their words: "Strong networks
seem to be linked both to economic insecurity and a lack of social
services. As long as the unemployment rate is high the threat of living
in poverty is real, and as long as large segments of the population find
access to medical services, day care, and social welfare services problem-


atic, we can expect to find reliance on strong networks to continue
among them" ( p . 28)
T w o ethnographic studies demonstrate the same point: Stack
(1974) studied a black, urban American, midwestern ghetto, Lomnitz
(1977) a shantytown on the fringes of Mexico City. Without apparent
knowledge of one another's work, and despite the enormous cultural
differences between these two populations, the investigators came to
nearly identical conclusions. Stack: "Black families living in the Flats
need a steady source of cooperative support to survive. They share with
one another because of the urgency of their needs. . . . They trade food
stamps, rent money, a T V , hats, dice, a car, a nickel here, a cigarette
there, food, milk, grits, and children. . . . Kin and close friends who
fall into similar economic crises know that they may share the food,
dwelling, and even the few scarce luxuries of those individuals in their
kin network. . . . Non-kin who live u p to one another's expectations
express elaborate vows of friendship and conduct their social relations
within the idiom of kinship" (1974, p p . 32-33, 40). Lomnitz: "Since
marginals are barred from full membership in the urban industrial
economy they have had to build their own economic system. T h e basic
social economic structure of the shantytown is the reciprocity net-
work. . . . It is a social field defined by an intense flow of reciprocal
exchange between neighbors. T h e main purpose . . . is to provide a
m i n i m u m level of economic security to its members" (1977, p. 209).
T h e similarity extends also to the use of fictive kinship as both effect
and cause of further reciprocity.
T h i s pervasive use of strong ties by the poor and insecure is a
response to economic pressures; they believe themselves to be without
alternatives, and the adapative nature of these reciprocity networks is
the main theme of the analysts. At the same time, I would suggest that
the heavy concentration of social energy in strong ties has the impact of
fragmenting communities of the poor into encapsulated networks with
poor connections between these units; individuals so encapsulated may
then lose some of the advantages associated with the outreach of weak
ties. T h i s may be one more reason why poverty is self-perpetuating.
Certainly programs meant to provide social services to the poor h a v e
frequently had trouble in their outreach efforts. From the network
arguments advanced here, one can see that the trouble is to be expected.

Sociological Theory

Weak Ties in the Spread of Ideas

In S W T I suggested application of the argument on weak ties to
the study of innovation diffusion (1973, pp. 1365-1369). T h i s sugges-
tion has been taken u p by Everett Rogers in his analysis of family
planning over a ten-year period in rural Korean villages; in combina-
tion with other network variables, this argument helps explain the
level of adoption in these villages (Rogers, 1979, pp. 155-157; Rogers
and Kincaid, 1981, p p . 247-249).
T h i s argument applies not only to the diffusion of innovations
but to the diffusion of any ideas or information. It has been taken u p
especially with regard to the spread of cultural and scientific ideas.
Fine and Kleinman, for example, in an article entitled " Rethinking
Subculture: An Interactionist Analysis" (1979),assert that the sociolog-
ical concept of subculture is deficient because it is stripped of its inter-
actional origins and regarded mainly as a set of disembodied symbols.
They confront, in particular, the paradox that large numbers of indi-
viduals, most of whom have never been in contact with one another,
nevertheless manage to sustain common understanding and mean-
ings-as in the example of youth culture. They reject the view that
such a common culture can be mostly explained by the pervasive influ-
ence of mass media: "While media diffusion can result in widespread
knowledge, one must not equate the extent of information spread with
method of transmission. Much that is communicated by the mass
media is not transmitted or used by audiences" ( p . 9). Furthermore,
many cultural items never transmitted by the media are known
throughout an extensive network: "Youth cultures offer excellent
examples of subcultures which provide a set of communication chan-
nels external to the media. Much material which is common knowl-
edge a m o n g young people-dirty jokes, sexual lore, aggressive
humor . . . -is not communicated by the adult-controlled media"
( p . 9). They go on to suggest that the

speed at which children's lore is spread across great dis-
tances . . . suggests the role of weak ties. In addition
to the school peer group, children who have been geo-
graphically mobile may maintain friendships over many
miles. T h e childhood pastime of having pen pals is an


example of this phenomenon. Likewise, the distant
. , , cousins who populate American extended families
provide children with others with whom to compare
their life situations and cultures. Since children's culture
has both regional and local variations, these kin ties can
provide a mechanism by which cultural traditions breach
geographical chasms. . . . T h e spread of culture from an
individual in one local social network to a n acquaintance
in another local social network seems to be a critical
element for the communication of cultural elements
within a subculture, [pp. 10-11]

O n e suggestive empirical study consistent with this argument
was carried out by L i n , Dayton, and Greenwald (1978).Volunteers in a
tri-city area of the eastern United States agreed to forward a booklet to
designated but previously unknown target persons through a chain of
personal acquaintances (see Milgram, 1967). Lin and colleagues inves-
tigated this question systematically by defining strength of tie in two
different ways: by recency of contact and also by the type of relation-
ship named by respondents sending the booklet along to the next per-
son. Data based o n both measures showed that successfully completed
chains made much more use of weak ties. T h e authors sum u p by
saying that "participants in the successful chains tended to utilize
fewer strong ties in their forwarding effort. T h e successful terminals
[those who reached the target] dramatically showed that they had weak
ties with the targets" ( p . 163). While this experiment is artificial in the
sense that n o information or ideas were actually being transmitted with
these booklets, the efficacy of weak ties in reaching socially distant and
unknown targets suggests that the process cited by Fine and Kleinman
to explain the diffusion of cultural ideas and symbols across wide seg-
ments of a society-via weak ties-may indeed operate as hypothesized.
What makes cultural diffusion possible, then, is the fact that
small cohesive groups who are liable to share a culture are not so
cohesive that they are entirely closed; rather, ideas may penetrate from
other such groups via the connecting medium of weak ties. It is a
seeming paradox that the effect of weak ties, in this case, is homogeni-
zation, since my emphasis has been the ability of weak ties to reach out
to groups with ideas and information different from one's own. T h e
paradox dissolves, however, when the process is understood to occur


over a period of time. T h e ideas that initially flow from another setting
are, given regional and other variations, probably new. Homogeneous
subcultures d o not happen instantly but are the endpoint of diffusion
processes. What cannot be entirely explained from arguments about
diffusion is why groups in California and New York, with initially
different orientations, adopt enough of one anothers' cultures to end
u p looking very similar. Weak ties may provide the possibility for this
homogenization, but the adoption of ideas cannot be explained purely
by structural considerations. Content and the motives for adopting one
rather than another idea must enter as a crucial part of the analysis.
T h e active role of individuals in a culture cannot be neglected lest the
explanation become too mechanistic. Fine and Kleinman note that
"culture usage consists of chosen behaviors. . . . Culture can be
employed strategically and should not be conceptualized as a condi-
tioned response. Usage of culture requires motivation and, in particular,
identification with those who use the cultural items. T h u s , values,
norms, behaviors, and artifacts constitute a subculture only insofar as
individuals see themselves as part of a collectivity whose members
attribute particular meanings to these 'objects'" (1979, pp. 12-13).
T h i s point can be clarified by contrasting the diffusion of
subcultural items to that of scientific information. T h e scientific case is
different in that the adoption of innovations is supposed not to be
arbitrary, as in subcultures, but to be governed by accepted tests and
standards. T h a t the supposed difference is only one of degree is sug-
gested by Chubin's model of scientific specialization (1976), which is
similar to Fine and Kleinman's analysis of cultural groups. He points
out that despite considerable division of labor in science, few problem
areas are likely to be sociometrically closed-any scientific field has a
center and a periphery, and the periphery may be defined by its
members' weak ties with the center and to other scientific groups.
T h e importance of this notion is clear. If "the innovativeness of
central units is shackled by vested intellectual interests (or perspectives)
then new ideas must emanate from the margins of the network"
( p . 460). Furthermore, as I suggested in S W T for the case of high-risk
innovations ( p . 1367), Chubin points out that marginals, in science,
can better afford to innovate; the innovations, if useful, are seized o n by
the center. T h i s sequence of events may go unnoticed because the
"adoption is sure to affect the innovator's position in the specialty as


well. Weak ties are transformed; the former marginal may become the
nucleus of a cluster" ( p . 464). A similar position is taken by Collins
(1974), who reports an empirical investigation of eleven laboratories in
Britain, the United States, and Canada involved in the development
and production of a certain type of laser. Arguing from his data and
from theoretical considerations, Collins contends that the idea of an
"invisible college" is misleading because it suggests too coherent an
internal structure. For Collins the likely importance of weak ties in
scientific innovations throws "serious doubt on the validity of the
questionnaire response as a direct indicator of the flow of real scientific
innovatory influence" ( p . 169).
T h e most comprehensive attempt, in a scientific setting, to test
empirically the validity of my arguments on weak ties is that of Fried-
kin (1980). He sent questionnaires to all faculty members in seven
biological science departments of a large American university and
received ninety-seven responses (71.3 percent of the relevant popula-
tion). T w o alternative definitions of weak tie led to similar outcomes.
T h e results reported rest on the following definition: T w o scientists
were said to have a weak tie if one reported having talked with the other
about current work, but the other made n o such report. Where both
made this statement about the other the tie was defined as strong. (See
S W T , p. 1364 n., for a discussion of the definition of mutual choices as
strong ties.)
Friedkin tests a number of my propositions systematically. O n e
test concerns what I called local bridges-ties between two persons that
are the shortest (and often the only plausible) route by which informa-
tion might travel from those connected to one to those connected to the
other (SWT, p p . 1364-1365). I argued that while not all weak ties
should be local bridges, all such bridges should be weak ties-an
argument central to the assertion that weak ties serve crucial functions
in linking otherwise unconnected segments of a network. Friedkin
found that there were eleven local bridges in the network; all were weak
ties (1980, p. 414). Moreover, this result is much stronger than might
have been expected by chance: 69 percent of ties among respondents
were weak and 31 percent were strong. By a binomial test of signifi-
cance, therefore, the chance of such a result, if ties were randomly
chosen to be local bridges, would be only 0.017.


Other findings predicted in S W T materialize strongly in these
data. I argued, for example, that the stronger the tie between two peo-
ple, the greater the extent of overlap in their friendship circles. I con-
tended too that people with strong ties to third parties are more likely
to be acquainted than are those with weak ties to those parties, who in
turn are more likely than if they had no mutual friends. All these
propositions are verified in substantial detail (Friedkin, 1980, p p . 415-
417). Friedkin concludes that this "evidence suggests that local bridges
tend to be weak ties because strong ties encourage triadic closure,
which eliminates local bridges. Other things being equal, weak local
bridges will tend to be maintained over time, while strong local bridges
will tend to be eliminated" ( p . 417). Finally, the hypothesis that inter-
g r o u p ties (as opposed to intragroup ties) consist disproportionately of
weak ties is assessed: 77 percent of interdepartmental ties, compared to
65 percent of intradepartmental ties, are weak ties ( p = 0.002).
T h e assertions about bridging can also be cast in terms of
transitivity- the tendency of one's friends' friends to be one's friends as
well. In S W T I asserted that transitivity could be expected of strong
ties, but not especially of weak ones, since the rationale for
transitivity-if A chooses B and B chooses C, it is inconsistent for A not
to choose C-is irrelevant for weak ties: A may not even know C, and if
he does he will not find it inconsistent not to be interacting with his
acquaintance's acquaintance (SWT, p. 1377). In a study of a n Israeli
kibbutz with 280 members, Weimann (1980) measured the strength of
ties by tenure, importance, and frequency. Using a program written by
Samuel Leinhardt (SOCPAC II) that compares the frequency of transi-
tive triads to that expected by chance, Weimann found that "networks
of strong ties are significantly tending to transitivity, while networks of
weak ties lack this tendency, and in some cases even tend to intransitiv-
ity. . . . Weak ties, relatively free from the tendency to transitivity, are
less structured, thus enabling them the role of bridging separate cliques
or subgroups, carrying information to all the network's segments"
(1980, p p . 16-17).
Weimann finds also, however, that strong ties are not irrelevant
in information flow-the speed of flow, credibility, and especially
influence are all greater through strong ties and, in fact, "most of the
influence is carried through strong ties" (1980, p. 12). He suggests a


division of labor between weak and strong ties: Weak ties provide the
bridges over which innovations cross the boundaries of social groups;
the decision making, however, is influenced mainly by the strong-ties
network in each group ( p . 21).
Weimann also points out that weak ties play an important cohe-
sive role in the kibbutz-a social unit formerly regarded as tightly
organized. "Encouraged by growing heterogeneity, the process of seg-
mentation . . . limited the power of traditional social forces and threat-
ened some of the basic principles of the kibbutz, namely direct
democracy, equality, and participation. . . . Conversation networks in
a kibbutz play . . . the role of social control mechanism: Gossip
becomes one of the social forces suppressing deviants and keeping the
obedience to the common norm. . . . By the transmission of gossip
items (mainly in weak ties, as shown in this research), the kibbutz
social system can keep solidarity, sanctions, and obedience in a hetero-
geneous, segmented social group" (1980, p p . 19-20).
Friedkin points to the need for greater precision about the regu-
larity and type of information transmitted through different kinds of
ties:

Granovetter's theory, to the extent that it is a pow-
erful theory, rests on the assumption that local bridges
and weak ties not only represent opportunities for the
occurrence of cohesive phenomena . . . but that they
actually do promote the occurrence of these phenomena.
A major empirical effort in the field of social network
analysis will be required to support this aspect of Grano-
vetter's theoretical approach. . . . It is one thing to argue
that when information travels by means of these ties it is
usually novel, and, perhaps, important information to
the groups concerned. It is another thing to argue that
local bridges and weak ties promote the regular flow of
novel and important information in differentiated struc-
tures. One may agree with the former and disagree with
the latter. If we accept the proposition that regular flows
of information depend on the presence of multiple short
paths between persons, than a local bridge does not
represent a likely path of information flow, though it
represents a possible path of such flow. . . . O n e might
argue that such information as does flow by means of
local bridges is crucial to the social integration of differ-


entiated populations, that is, that regular flows of infor-
mation between differentiated groups are not crucial to
their systemic integration. If so, one is asserting that there
are different bases of macro and micro integration; for
example, that macrointegration can be based on weak ties
which permit episodic transmissions of information
a m o n g groups, while microintegration is based on a
cohesive set of strong ties which permit regular transmis-
sions within groups. [1980, pp. 421-4221

Weak Ties and Social Organization

Friedkin's emphasis on weak ties as the basis of macrolevel
rather than microlevel integration is similar to Peter Blau's argument
that since "intimate relations tend to be confined to small and closed
social circles . . . they fragment society into small groups. T h e integra-
tion of these groups in the society depends on people's weak ties, not
their strong ones, because weak social ties extend beyond intimate cir-
cles (Granovetter, 1973) and establish the intergroup connections on
which macrosocial integration rests" (Blau, 1974, p. 623).
In this section, therefore, I consider a number of studies that
address the role of weak ties in organizing groups larger than the
primary groups of microsociology. T w o such studies consider the
problem of integration of diverse groups within formal organizations.
Karweit, Hansell, and Ricks (1979) do not present new data but rather a
stimulating review of the literature on how features of peer groups
within schools affect the educational aspirations and achievements of
their members. After noting that numerous investigations have sought
to document the socializing power of the peer group, they remark:
' T h e dyadic view implies that peer socialization to different values
occurs only through close friendship ties. However, other peer
relationships- such as admiration for someone quite different from
oneself- may be more important socialization sources than close
friendships" ( p . 19). (Recall the evidence cited in S W T that ties to those
very different from oneself are much more likely to be weak than
strong, p. 1362.) Further, they question whether it is proper policy to
create a situation where the values of one group in school (such as
high-achievers) can be efficiently assimilated by others. A more desira-
ble peer structure, they suggest, would be one in which "diverse cul-


tures can exchange information and support without necessarily
becoming more similar" ( p . 19).
They then consider how my argument about weak ties may be
applied to biracial school settings in the United States, suggesting that
"racial integration in the classroom can be achieved by arranging class-
room structures to produce enough weak contacts to connect black and
white cliques, rather than by encouraging strong biracial friendships
[the usual strategy]. T h i s result would be immediately applicable in
schools because weak ties are easier to stimulate through realistic
organizational innovations. . . . If the idea that racial integration has
to occur in strong contacts at the dyadic level can be relaxed, many
possibilities for planned intervention in schools to foster racial
accommodation become feasible" ( p . 20). Finally, Karweit and col-
leagues suggest that a "good deal of students' alienation from school
may be associated with their lack of indirect contacts with student
leaders and their consequent inability to contribute to student decision-
making processes" ( p . 26). T h i s point is reminiscent of my argument in
S W T that West Enders in Boston may have been loath to join the fight
against urban renewal because a lack of bridging weak ties left most of
them without even indirect access to leaders of such an organization. If
the same argument applies to school settings, the strategy of encourag-
ing bridging weak ties, suggested by Karweit and colleagues, could
have the effect not only of linking culturally different groups but of
reducing student alienation and increasing social solidarity.
Judith Blau presents a case study of successful integration in a
children's psychiatric hospital in New York City (1980)and argues that
this integration can only be understood by considering the role of an
extensive network of weak ties. T h i s public hospital has a staff of two
hundred and serves severely impaired children. Treatment is difficult
and outcomes uncertain. Although comparable institutions elsewhere
are marked by high staff turnover and low morale, Blau notes that this
is decidedly not the case at the Childrens' Center. She attributes the
high morale at the center to the surprising predominance of weak ties
among staff members-so many that "all two hundred staff members
are on a first-name basis" (p. 6). Interaction is so evenly distributed
that there is an absence of cliques, though she did discover "a highly
differentiated system of specialized staff relations" forming stable sub-
networks ( p . 8).


These subnetworks have many different foci of organization
reflecting the complex arrangement of the hospital into departments
(Psychology, Nursing), committees, programs (art, recreation), residen-
tial units, and clinical teams focused on specific sets of patients. If the
ties in these stable subnetworks were strong, by my argument this
would tend to close them off from one another, so that they would
develop into cliques; the overwhelming predominance of weak ties,
even if structured, produces and maintains, instead, a situation in
which each subnetwork overlaps extensively with many others, and a
large number of the weak ties serve bridging functions. Blau found, for
example, that neither "homogeneous work groups nor strong friend-
ship relations could be identified. . . . T h e institution's intolerance of
close dyadic ties is expressed by the ritualized avoidance patterns
a m o n g those who have a sexual or family alliance outside of the insti-
tution. . . . T h i s suggests . . . that in a complex structure . . . exten-
sive weak networks can remain viable only when close ties are
prohibited. . . . For when dimensions of structure intersect and staff
are integrated in subnets of multiple crosscutting role relations, close
bonds with some will threaten working relations with others" ( p p . 20-
21). Further, "an individual's access to opportunities and resources
can only be fully exploited if he or she is linked with others in diverse
positions furnishing different information, but strong ties tend to
involve closed circles that limit [such] access. . . . Since information is
so widely diffused throughout the hospital structure, it is imperative
for staff to sustain bridging intergroup connections, further weakening
bonds of ingroup solidarity" ( p . 21).
Relating her findings to organization theory, Blau notes that the
problem of integrating large numbers of diverse specialists in a formal
setting is not simple; in the present case, there are psychiatrists, social
workers, doctors, dentists, teachers, art and music therapists, and learn-
i n g disability specialists, as well as the usual aides. T h e standard solu-
tion to this problem is strong mechanisms of control built into a
formal hierarchy. As an alternative to this formal solution some clinics
have tried to create "familylike and egalitarian relations." Blau points
out, however, that both solutions exhibit forms of strong ties ( p . 19). In
the bureaucratic solution, the ties are hierarchical; in the democratic
clinics, many of which have reacted against the formal model, "tena-
cious ties provide a matrix of close primary group relations unifying


the entire structure. These strong ties strikingly resemble patterns
observed in small communities, summer camps, and Jesuit monastic
orders" ( p . 20). T h u s the weak-tie solution of the Bronx psychiatric
hospital seems significantly different from either of these patterns.
Blau implies that it is better by associating the hierarchical pattern
with ineffectual coordination of health care delivery; one might
assume, moreover, that the democratic solution, which depends on a
network of strong ties to bind an institution together, would be severely
limited regarding the size of the system so bound. Attempts to extend
the size of institutions would confront the constraint that individuals
could not sustain the requisite number of strong ties, leading to frag-
mentation of the institution into cliques with a corresponding loss of
morale and integration.
If the weak-tie mode of organizational integration is in fact
efficient and leads to high morale and good services for the general
theoretical reasons Blau suggests, can the model be exported to similar
settings? T o answer this question we must understand how such a
pattern came about originally. Blau suggests that there was a con-
scious attempt to develop a new kind of structure, but it is unclear
whether the founders understood the structural implications of their
early decisions.
In a larger setting, that of entire communities, Breiger and Patti-
son (1978) use the methods of blockmodeling (see White, Boorman, and
Breiger, 1976) to argue that weak ties play the bridging roles I have
suggested in integrating communities and that, moreover, it would be
possible to infer the weak versus strong quality of certain ties entirely
from algebraic manipulation and reduction of the raw sociometric data
even without other information. They take sociometric data collected
by Laumann and numerous associates (Laumann and Pappi, 1976;
L a u m a n n , Marsden, and Galaskiewicz, 1977) in a German city, Alt-
neustadt, and an American one called Towertown (both pseudonyms).
T h o u g h the patterns are different in the two cities, Breiger and Patti-
son show that they share certain structural features suggesting
the importance of weak ties. In technical terms, the joint homomor-
phic reduction of the two blockmodel semigroup multiplication tables
generates a common structure in which certain algebraic relations
are satisfied- relations that would be predicted by the arguments
of SWT.

224 Sociological Ties

In their analysis Breiger and Pattison studied three types of ties
in the two communities- social, community affairs, and business-
professional- and found that social ties function as strong ties, that
business-professional ties are weak, and that community-affairs ties are
strong in relation to business ties but weak in relation to social ones
(1978, p p . 222-224). T h i s characterization is consistent with the ethno-
graphic accounts of the communities, even though it can be developed
without knowledge of those accounts. While it is not possible here to
give an adequate treatment of the mathematical complexities of the
Breiger-Pattison analysis, it should be stressed that the algebraic role
structure they predict o n the basis of the weak-ties argument is not one
that can be found empirically by any variety of curve fitting; the
hypothesis is entirely falsifiable, but it is, in the present case, confirmed
for the two communities.
In S W T I suggested that individuals with few weak ties were
unlikely to mobilize effectively for collective action within their com-
munities, arguing that the West Enders described by Gans were for this
reason ineffective in fighting urban renewal. (See Gans, 1961, and my
exchange with Gans on this question in American Journal of Sociol-
ogy, September 1974.) Steinberg (1980) puts this suggestion into a gen-
eral context, noting that there are "two dominant schools of thought
o n the relationship between community attachment and participation
in organized protest." O n e is the mass society argument: Protest results
from the "sudden activation of previously 'unattached' individuals or
uprooted collectivities." T h e second school, largely in response to the
first, argues that "attached individuals or organized collectivities are
most likely to engage in sustained protest" (1980, p. 2). T o argue the
importance of weak ties in organization is a position halfway between
these two; the second school has had little to say about preexisting
social ties, and Steinberg notes the surprising "paucity of empirical
research which systematically examines the social ties of the members
of protest groups. . . . We need systematic microlevel inquiries that
examine the social ties of initiators and initial recruits before and after
the formation of conflict groups in different contexts as well as the
effects of these ties" ( p . 3).
Steinberg's own work is a longitudinal study that analyzed "the
politically relevant social ties of the initiators and initial recruits in five


conflict groups mobilized around [different] educational issues in a
suburban community" (p. 3). In all five cases, local school authorities
resisted the dissidents' goals; as a result, they were forced to initiate new
groups ( p . 4). For all groups, Steinberg asked about preexisting ties
between initiators and initial recruits, those among the recruits, and
those between group members and nonmembers relevant to educa-
tional affairs. She found that of the seven individuals responsible for
initiating the five groups, none was seriously integrated into the com-
munity; all were women who occupied a "marginal position in rela-
tion to the network of groups and individuals active in community
affairs" ( p . 17). Furthermore, of twenty initial members recruited
directly by initiators, nineteen stemmed from preexisting ties and fif-
teen of these were weak; the other four were concentrated in one group.
(Strong ties were those described by respondents as representing good
friends.)
T h e group recruited on the basis of strong ties "was linked to
the fewest organizations and individual memberships were concen-
trated in the same organizations which formed a dense network. Later
recruits tended to join the same groups as the founders. Groups formed
on the basis of weak ties, on the other hand, were linked to more
organizations that were loosely knit and individual memberships
tended to be scattered throughout these organizations" (p. 19). T h e
strong-tie group was ultimately unsuccessful, whereas three of the
other four groups were able to implement many of their aims. Stein-
berg concludes that although the initiators of successful groups were
marginal individuals in the community, they were able to recruit peo-
ple who had occupied leadership positions and were linked to a dense
network of school activists. "The evidence suggests, tentatively, that
where innovations are controversial, a mobilization strategy based on
the activation of weak ties is more likely to facilitate adoption of the
goal and integration into the school decision-making structure"
(Steinberg, 1980, p. 25).
Here we see the intricate interplay between weak and strong ties
in structuring outcomes and mediating the competing claims of var-
ious community groups. T h e final study to be reported also points to
such an interplay and contains elements of conflict and cohesion. T h e
study itself reports only the cohesion achieved by the business com-
munity; left implicit is the extent to which this cohesion may imple-


ment the goals of that community in conflict with those of competing
groups such as labor or consumers. Bearden and colleagues studied
interlocking directorates among American corporations-a tie between
two corporations was said to exist when at least one individual sat o n
the boards of directors of both. It has long been noted that such inter-
locks permit interfirm control or collusion.
Ever since the investigation of interlocking directorates began
early in the twentieth century, one of the persistent issues has been
whether the corporate network consists of cliques (or interest groups),
which might then be seen as competing with one another, or of one
large densely connected web, reflecting the overarching influence of
unity among capitalist leaders. A number of studies (reviewed i n
Bearden and others, 1975, pp. 1-16) have produced inconclusive results
on this question.
T h e study to be reviewed here, Bearden and others (1975), is
described in a widely circulated (though unpublished) paper; the most
extensive study of interlocking directorates ever carried out in the Unit-
ed States, it covered the 1,131 largest American corporations during
the period 1962-1973. From various sources Bearden and colleagues
collected data on all 13,574 directors of these companies in 1962. In
resolving the issue of interest groups versus one large connected net-
work, Bearden and coworkers comment that if "all interlocks are
treated as having equal strength, the whole network is so highly
connected that the identification of cliques is very difficult." They
point out, however, that when a n officer of firm A sits on the boards of
firms B and C, two types of interlocks are created. T h e A-B and A-C
interlocks are functional or strong interlock ties since there is a direct
connection between an officer of one corporation and the board of the
other. T h i s is the kind of interlock that might occur, for example,
"when one company places someone o n the board of another company
because of stock ownership, buyer-seller relationship, or other func-
tional relationship" (Bearden and others, 1975, p. 27). O n the other
hand, the individual from firm A also creates an interlock between the
boards of firms B a n d C, which may have n o direct business ties with
one another. This interlock can then be seen as incidental, or weak.
Another weak tie is the interlock created between two companies when
a n individual from outside the corporate world entirely-from a uni-
versity or law firm, for example-sits on both boards.


Bearden and colleagues argue that in order to uncover cliques or
interest groups it is necessary to compare the corporate network as a
whole to that formed only by strong interlock ties. They find that when
the entire network is analyzed, there is "no sign of interest groups
. . . [but] lack of clustering reflects the prevalence of weak ties in the
interlock network" ( p . 60). When only strong ties are analyzed, definite
cliques appear, "with banks playing the central roles in all the clus-
ters" (p. 68). They comment:

Recent interlock literature has produced contra-
dictory indications about the survival of the "interest"
groups uncovered by pre-World War II research. . . . It is
our contention that both sides of the dispute are correct:
T h e integration of New York, Boston, Chicago, Phila-
delphia, and California centers of business into a
national and even international network of corporations
has occurred simultaneously with the maintenance and
further development of interest groups. . . . T h e inten-
sive clusters are created by strong ties. . . : interfirm
stockholding, . . . indebtedness, enduring economic
interdependence. . . . Like the small groups studied by
Granovetter, these intense ties tend to produce cliques
which are in some ways competitive and exclusive of each
other. T h e extensive national network is formed from
weak ties. . . . They do not imply specific interfirm eco-
nomic connections; instead they reflect an overall com-
mon orientation and interest, the need for common
action across cliques, and a growing sense of national
and international interdependence among large corpora-
tions. T h u s . . . the network . . . contains both the coor-
dinative leverage of weak ties and the cliquishness of
strong ties; national cooperation among most firms as
well as competitive antagonism between clusters; unity
and conflict simultaneously. [pp. 51-52]

Their argument is oriented, then, to the force of weak intercor-
porate ties in creating cooperation and coordination in the national
network of firms. It seems likely also that such ties play a role analo-
gous to that highlighted by Fine and Kleinman in helping to maintain
a common cultural consciousness among the managers of large
corporations- in transmitting information and influences on stylistic
matters, political judgments, and broad social trends. It could be


argued that such cultural homogeneity is a necessary condition for any
large-scale corporate coordination.

The Question of Verification

In this account of empirical research bearing on the argument of
SWT, I have shown that the argument has in fact been useful in clarify-
ing a variety of phenomena ranging from effects of social relations on
individuals, to the diffusion of ideas and innovations, to the organiza-
tion of large-scale social systems. But still one may usefully ask: Do
these studies show that the argument is empirically verified? Many of
the studies cited did not set out systematically to test the argument of
SWT. In some cases that argument came in handy to explain empirical
findings that would have otherwise been anomalous. This is the case
for Rogers, J. Blau, and Bearden and colleagues. While these studies
certainly lend credence to the argument, the procedure for finding
them-either personal contact with the authors or else their citation of
SWT-is not an unbiased procedure for testing the argument. There is
n o way to know, for example, about empirical studies in which the
S W T argument was considered, rejected, and not mentioned because it
did not fit the facts; scholars can hardly be expected to cite every argu-
ment that does not help explain their anomalies.
Furthermore, a number of the studies cited are mainly theoreti-
cal efforts- proposing that, in their subject area, weak ties can be seen
as serving important functions, but not actually bringing a substan-
tial body of empirical data to bear on this assertion. T h e work of Coser,
Boorman, Fine and Kleinman, Chubin, and Karweit and colleagues fits
this category. These studies are more valuable in providing stimulating
leads for future research than in providing direct confirmation of the
argument.
Other studies have set out, as part of their work, to put the
argument to a systematic test: Langlois; Lin, Ensel, and Vaughn;
Ericksen and Yancey; Murray, Rankin, and Magill; Rogers; Lin, Day-
ton, and Greenwald; Breiger and Pattison; Weimann; Steinberg; and
Friedkin have done so. T h e results of these studies are very encourag-
ing, but not conclusive. As Friedkin points out, one needs to show not
only that ties bridging network segments are disproportionately weak
but also that something flows through these bridges and that whatever


it is that flows actually plays a n important role in the social life of
individuals, groups, and societies. While some of these studies d o make
such a showing, the case remains incomplete.
T h i s review has also highlighted a crucial feature of the original
argument that has important bearing o n its verification: I have not
argued that all weak ties serve the functions described in SWT- only
those acting as bridges between network segments. Weak ties are
asserted to be important because their likelihood of being bridges is
greater than (and that of strong ties less than) would be expected from
their numbers alone. T h i s does not preclude the possibility that most
weak ties have no such function. It follows that an important part of
further specifying the argument would be more systematic investiga-
tion of the origin and development of those ties which bridge as com-
pared to those which do not. In S W T I suggested that for a community
to have many weak ties which bridge, there must be several distinct
ways or contexts in which people may form them. I went o n to point
out that Charlestown (in Boston), which successfully fought urban
renewal, as compared to the West End, which did not, had a rich
organizational life and its male residents worked within the commu-
nity. T h e implication was that weak ties formed in these contexts were
more likely to bridge than weak ties that result by meeting friends of
friends-in which case the new tie is clearly not a bridge. None of the
work reviewed here has taken u p this point; a recent paper by Feld
( 1981 ), however, develops a new theoretical perspective based o n the
issue of what social foci organize the formation of social ties. Work
proceeding from this perspective may shed new light o n the issues
raised here. T h e most pressing need for further development of net-
work ideas is a move away from static analyses that observe a system at
one point in time and to pursue instead systematic accounts of how
such systems develop and change. Only by careful attention to this
dynamic problem can social network analysis fulfill its promise as a
powerful instrument in the analysis of social life.

Note

1. Discussant's comments o n the material in this chapter, which
was presented as a paper at the Conference o n Contributions of Net-
work Analysis to Structural Sociology, Albany, N.Y., April 4, 1981.


References

BALTZELL, E. D.
1958 Philadelphia Gentlemen: T h e M a k i n g of a National
U p p e r Class. Glencoe, 111.: Free Press.
B E A R D E N , JAMES, A N D OTHERS
1975 " The Nature and Extent of Bank Centrality in Corporate
Networks." Paper presented at the 1975 meeting of the
American Sociological Association, San Francisco.
BLAU, JUDITH
1980 "When Weak Ties Are Structured." Unpublished manu-
script, Department of Sociology, State University of New
York, Albany.
B L A U , PETER
1974 "Parameters of Social Structure." American Sociological
Review 39(5):615-635.
B O O R M A N , SCOTT
1975 "A Combinatorial Optimization Model for Transmission
of J o b Information T h r o u g h Contact Networks." Bell
Journal of Economics 6(1):216-249.
B R E I G E R , R O N A L D , A N D PATTISON, PHILIPPA
1978 " The Joint Role Structure of T w o Communities' Elites."
Sociological Methods and Research 7(2):213-226.
C H U B I N , DARYL
1976 " The Conceptualization of Scientific Specialties.? Socio-
logical Quarterly 17(4):448-476.
COLLINS, H A R R Y
1974 " The TEA Set: Tacit Knowledge and Scientific Net-
works." Science Studies 4:165-186.
COSER, ROSE
1975 " The Complexity of Roles as Seedbed of Individual
Autonomy." In L. Coser (Ed.), T h e Idea of Social Struc-
ture: Essays i n H o n o r of Robert Merton. New York: Har-
court Brace Jovanovich.
D E L A N Y , JOHN
1980 "Aspects of Donative Resource Allocation and the Effi-
ciency of Social Networks: Simulation Models of Job


Vacancy Information Transfers Through Personal Con-
tacts." Ph.D. dissertation, Yale University.
ERICKSEN, E . , A N D Y A N C E Y , V.
1977 " The Locus of Strong Ties." Unpublished manuscript,
Department of Sociology, Temple University.
1980 "Class, Sector and Income Determination." Unpublished
manuscript, Temple University.
FELD, SCOTT
1981 " The Focused Organization of Social Ties." American
Journal of Sociology 86(5): 5- 1035.
101
FINE, GARY, A N D KLEINMAN, SHERRYL
1979 "Rethinking Subculture: An Interactionist Analysis."
American Journal of Sociology 85(1):
1-20.
FRANK, OVE
1981 "A Survey of Statistical Methods for Graph Analysis."
Sociological Methodology 12:110-155.
FRIEDKIN, N O A H
1980 "A Test of the Structural Features of Granovetter's
'Strength of Weak Ties' Theory." Social Networks
2:4 1 1-422.
GANS, HERBERT
1961 T h e Urban Villagers. New York: Free Press.
GRANOVETTER, M A R K S.
1973 " The Strength of Weak Ties." American Journal of Soci-
ology 78(6):1360-1380.
1974 Getting a Job: A Study of Contacts and Careers. Cam-
bridge, Mass.: Harvard University Press.
1976 "Network Sampling: Some First Steps." American Jour-
nal of Sociology 81:1287-1303.
HALBERSTAM, DAVID
1972 T h e Best and the Brightest. New York: Random House.
HOMANS, GEORGE
1950 T h e H u m a n G r o u p . New York: Harcourt Brace
Jovanovich.
K A R W E I T , N A N C Y , H A N S E L L , S T E P H E N . A N D RICKS, M A R G A R E T
1979 " The Conditions for Peer Associations in Schools."
Report no. 282, Center for Social Organization of Schools,
Johns Hopkins University.


LANGLOIS, S I M O N
1977 "Les Reseaux Personnels et la Diffusion des Informations
sur les Emplois." Recherches Sociographiques 2:213-245.
L A U M A N N , E., A N D PAPPI, F.
1976 Networks of Collective Action. New York: Academic
Press.
L A U M A N N , E . , M A R S D E N , P . , A N D GALASKIEWICZ, J .
1977 "Community-Elite Influence Structures." American Jour-
nal of Sociology 83:594-63 1.
L I N , N A N , D A Y T O N , P A U L , A N D G R E E N W A L D , PETER
1978 "Analyzing the Instrumental Use of Relations in the
Context of Social Structure." Sociological Methods and
Research 7(2):
149-166.
L I N , N A N , ENSEL, W., A N D V A U G H N , J .
1981 "Social Resources, Strength of Ties and Occupational Sta-
tus Attainment." American Sociological Review 46(4):
393-405.
L O M N I T Z , LARISSA
1977 Networks and Marginality. New York: Academic Press.
MILGRAM, STANLEY
1967 "The Small-World Problem." Psychology Today l(May):
62-67.
M U R R A Y , S . , R A N K I N , J . , A N D M A G I L L , D.
1981 "Strong Ties and Job Information." Sociology of Work
and Occupations 8(1):1 19- 136.
POOL, ITHIEL
1980 "Comment o n Mark Granovetter's 'The Strength of Weak
Ties: A Network Theory Revisited.'" Read at the 1980
meeting of the International Communications Associa-
tion, Acapulco, Mexico.
ROGERS, EVERETT
1979 "Network Analysis of the Diffusion of Innovations." In
P. Holland and S. Leinhardt (Eds.),Perspectives o n Social
Network Research. New York: Academic Press.
ROGERS, EVERETT, A N D K I N C A I D , D. L A W R E N C E
1981 Communication Networks: Toward a New Paradigm for
Research. New York: Free Press.


SIMMEL, G E O R G
1950 T h e Sociology of Georg Simmel. (Translated and edited
by Kurt Wolff.) New York: Free Press.
STACK, C A R O L
1974 All Our Kin. New York: Harper & Row.
STEINBERG, LOIS
1980 "Preexisting Social Ties and Conflict Group Formation."
Paper presented at the 1980 meeting of the American Soci-
ological Association, New York.
WEIMANN, G A B R I E L
1980 "Conversation Networks as Communication Networks."
Abstract of Ph.D. dissertation, University of Haifa, Israel.
WELLMAN, BARRY
1979 "The Community Question: T h e Intimate Networks of
East Yorkers." American Journal of Sociology 84(5):
1201-1231.
W H I T E , H A R R I S O N , BOORMAN, SCOTT, A N D B R E I G E R , R O N A L D
1976 "Social Structure from Multiple Networks: I. Block-
models of Roles and Positions." American Journal of
Sociology 81:730-780.

The Strength of Weak Ties Revisited

More Related Content

Similar to The Strength of Weak Ties Revisited (20)

More from augustodefranco . (20)

Recently uploaded (20)

The Strength of Weak Ties Revisited