Académique Documents
Professionnel Documents
Culture Documents
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=uncpress.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
University of North Carolina Press is collaborating with JSTOR to digitize, preserve and extend access to
Social Forces.
http://www.jstor.org
Social Networks in the Modern City
HENRY W. IRVING, University of Hull
ABSTRACT
Social network, an interesting theoretical cconcept, has suffered through difficulties in
developingfrom it any operationaldevices suitablefor use in ordinarysocial surveyresearch.
Here, one such device is presented, and its utility is examined in the contrasting urban
contexts of Hull and Los Angeles. Existing theories about the correlatesof networkdensity
are partiallyconfirmed and partially contradictedby the findings. In particularthere is an
attempt to reevaluate the part played by kinship in the social networks of modern urban
society.
"And these three things abideth-class, role and network-and the greatest of these
is network" (Mitchell). When Mitchell wrote these words, network was certainly
looming large on the sociological scene. The seminal work of Elizabeth Bott and
her followers, in their studies of conjugal role segregation, had placed the idea of
network firmly in the language of social inquiry, by suggesting that it provided a far
more useful contextual framework for studying informal social relations than the
more traditional "group." Frankenburg had attempted to give the concept of
network a more general importance by placing various communities on a spectrum
of "social redundancy." Mitchell's own former colleagues in Rhodesia had given
an impressive array of research teeth to the idea by their participant observer
studies of the spread of rumour and the extension of influence in African urban
situations. J. A. Barnes, the doyen of the network group, had been attempting, for
some fifteen years, to wrap the network concept in a theoretical garb derived from
graph theory. Mitchell's enthusiasm was certainly well-founded. Yet most social
scientists will agree that network has not attained the conceptual stature of class or
role. I would like to suggest that this stunted growth stems not from the lack of
theoretical investment, for there has been plenty of that, but from the reluctance
of social scientists to operationalize the idea. Several workers have specifically
attempted to test the Bott hypothesis in a variety of social contexts (Aldous and
Strauss; Nelson; Udry and Hall), but their efforts merely attracted the approbrium of
"narrow operationalism" (Turner). Such hostility to genuine attempts to build a
body of working empirical devices must surely be inimical to the development of
the idea of a social network. Without spawning a set of measuring devices that have
widespread and relatively straightforward usage, the construct must inevitably
remain stillborn.
Other reasons, apart from the perennial hostility of self-styled theorists,
underlay the slow operational development of the network idea. In the first place lay
the heavy hand of the Bott hypothesis. So influential were Bott's ideas in the early
years of network that it came to be regarded, not as an interesting social fact to be
investigated in its own right, but as an intervening variable in the explanation of
conjugal role segregation. Yet the idea of network density, as a way of characterizing
867
868 / Social Forces / vol. 55.4, june 1977
3. Mesa del Mar A tract development with houses of a price and quality
to appeal to lower middle and skilled working classes.
indications in the literaturesuggest that, in the upper reaches of the managerialand pro-
fessional class, the careerof the husbandmay be so time consuming and involving that the
maritalrelationshipbecomes segregated(264).
Whether such segregation is associated with a tightening of the network,
Bott does not say, but her remarksserve to illustratethe complexity of the relation-
ship between class and network density. Her work, based on twenty intensively
studiedfamilies in the East End of London, is a telling example of the explanatory
potency of the anthropologicalapproach. To gain wider validity, however, her
network hypotheses must be tested among a larger sample of respondents, in a
wider variety of contexts and unfettered by a preoccupationwith conjugal role
segregation.To this end, the Hull and Californiandata were analyzed.
In terms of urbancharacteristicsit would be difficult to find more contrast
than exists between Hull and that partof GreaterLos Angeles which lies in Orange
County,surroundingthe old ruralservice centerof Santa Ana. Hull is essentially an
industrialcity, spawned, admittedly,in the Middle Ages, but assuming its present
urbanform in Victorian times with the arrivalof the railway and the consequent
expansion of its fishing, docks, and raw materialprocessing industries.Residential
expansion in the twentieth century has, of course, been considerable, but it has
simply fleshed out, with council and private housing, the bare bones of the
Victoriancity. OrangeCountyis quintessentiallythe freeway city of the 1960s-the
sort of urban place which Webber would consider a "non-place urban realm"
(a, b). Homes are new, people are residentially mobile, households typically
compriseyoung families or single people. It is astonishing,therefore,that the index
of networkdensity shows such similarityof frequencydistributionbetween the two
populations. (Table2). One would have expected, following Bott and Webber,that
Hull would show a markedlycloser-knit distribution.In fact, the reverse is true:
more people thanexpected in OrangeCountyhave tight networks;more people than
expected in Hull have loose networks. Social activity is obviously not so orderedby
the hardfacts of mobility as Webberwould suggest.
Bott's caution about the complexity of the relationshipbetween class and
networkdensity is heavily underlinedby both sets of data. Threeindicatorsof social
class-occupation, education and car ownership-failed, with one exception, to
reveal any significantassociationwith the index of networkdensity.The exceptional
case was the education variable for the California sample. The categorization
system was differentfrom thatused in Hull, for almost 50 percentof the Californian
respondents reported education up to or beyond the bachelor's level, and this
difference may underlie the positive finding. Graduates,evidently, exhibit a ten-
dency towards closeness of network-a neat illustration, perhaps, of the Bott
postulationthatprofessionaland managerialpeople undergowhat might be called a
social convergence. Certainly,it seems as though it is the close networks of the
graduateswhich underlie the anomalous finding reportedearlier, that Californian
networks were more closed than Hull ones. Car ownership in Californiais a very
poor indicatorof social class, for 240 out of 300 respondentfamilies had two or
more cars, but there is sufficientoverall evidence to show that, whateverthe causes
872 / Social Forces / vol. 55:4, june 1977
Table 2. NETWORK
DENSITY
INHULLANDORANGECOUNTY
Loose Close
knit knit
Hull 168 186 354 = 1275
df = 1
Orange County 101 199 300
I <0 001
~~~~p
269 385 654
Occupation
Loose Close Loose Close
Non 65 8No 8 10
manual 65 81 manual 68 14
Manual 123 105 Manual 3 3 59
Q = 0 17 Q = 0 07
Education
Loose Close Loose Close
minimum
school 110 10 9 Non graduates 58 98
leaving age
mPnost -m 58 77 Graduates 43 143
Q = 0 14 Q = 0 32
Car ownership
Loose Close Loose Close
No car 92 101 One cOr 25 33
Ca r 76 85 Two or 75 165
Car 76 85 cars
~~~more 7 6
Q = 0?00 Q = 0 25
(To be significant at the 5% confidence level with this sample size, Q must have a value of at least 0.26)
company or with the occasional friend, report looser knit networks. This is an
interesting finding and one which points tellingly towards the value of using
personalitymore often as a variablein standardsocial inquiry,though some caution
mustbe added. The personalityinventorywas designed to categorizethe personality
of the actualrespondent;the networkdensity index attemptsto categorizethe social
ambience of the couple in the case of married respondents. The association is
significant enough, however, to report some confirmationfor Bott's assertionthat
individualpersonalityis an importantvariableunderlyingnetworkdensitypatterns.
The strengthof the relationshipsbetween age and network density on the
one hand, and extraversionand networkdensity on the other, pointedtowardsmore
874 I Social Forces I vol. 55:4, june 1977
Household
category
Loose Close Loose Close
Family 69 85 Family 65 131
Non Non
famfly 99 101 family 36 68
Q =009 Q =003
Length of
residence
Loose Close Loose Close
under 73 95 Under 66 101
Ten years five years
Ten years 95 91 Five years 34 98
and over and over
Q = 0 15 Q =-0 30
Age of
respondent
Loose Close Loose Close
45 and 100 73 45 and 18 69
over over
Under 68 113 Under 83 130
45 45
Q= -039 Q= 042
(To be significant at the 5% confidence level with this sample size, Q must have a value of at least 0.26)
Network density
Fairly Very
Loose close Close close
Normal 53 38 17 7 115
St Ieb 28 10 4 0 42 2
Introvert (3 26) (2425) x = 31 68
Personality Neurotic 26 16 2 2 46 df = 12
category Introvert (3 9)
Stable 33 29 14 1 77 p <0001
Extravert
Neurotic 28 17 20 9 74
Extravert (1 4) (1 5) (5 5) (6 3)
168 110 57 19 354
O Age (old)
QXY = 0 38
QXY: Tied T = 0 36
QXY: Diff T = 0 40 No effect
QTY = 0 31
QTY: Tied X = 0-31
QTY: Diff X = 0 31 No effect
Fairly Very
Loose close Close close
Hessle Road 19 20 11 10 60
Anlaby Park 21 16 11 5 53
x 2 = 44-2
New Iand 24 29 7 0 60
df =15
Garden Village 39 13 8 0 60
p <0.0001
North Hull Estate 35 14 9 2 60
Avenues 30 18 11 2 61
168 110 57 19 354
Fairly Very
Loose close Close close
Turtle Rock 22 34 0 4 60
X2 = 30 16
Mesa Verde 18 23 15 4 60
df = 12
Mesa del Mar 12 31 14 3 60
P <0 002
Apartments 29 17 9 5 60
Ghetto 20 23 13 4 60
101 128 51 20 300
Localization
Loose Close Loose Close
Localized 75 42 Localized 51 68
Non- Non -
Localized 93 144 1Localized,
50 131
Q 0-46 Q = -0Q33
Intensity
Loose Close Loose Close
Not Intense 149 103 Not Intense 83 103
Intense 19 83 Intense 18 96
Q -0 72 Q = -062
Spread
Loose Close Loose Close
Concentrated 113 62 Concentrated 82 88
Q - -0 60 Q = -068
Kin - orientation
Loose Close Loose Close
Q= +0 60 Q=+0 .27
Localized Localized
/ \ /\~~~~peimn
Intense Close knit Intense specime
.Close ns knit
A consequent model An antecedent model
CONCLUSIONS
This investigation has provided some confirmation for a number of the ideas
formulated by Elizabeth Bott. In one importantinstance, however, the findings
contradicther. It was earliersuggestedthatthe position of kin in the social network,
a position muchemphasizedby Bott and Gluckman,shouldbe carefullyscrutinized.
Methodologically, the inclusion of kin makes nonsense of certain techniques of
network inquiry, and for this reason was left aside in the construction of this
particularnetworkdensity index. Theoretically,however, if only for its traditional
emphasis, it merits some examination. Bott and Gluckman suggest that kinsfolk
occupy key positions in the social network.Litwakand Szelenyi, on the otherhand,
claim that in modern mobile society primary groups have become functionally
differentiated, with kin, neighborhood, and friends serving quite separate social
functions, with relatively little overlap (Litwak and Szelenyi). It was not possible,
for reasons alreadyoutlined, to examine precisely the position of kin in the social
network of each respondent. It is, however, interesting that the network density
index shows a significantassociation with the index of kin orientation-a measure
of the amount of interaction given to kinsfolk as opposed to non-kin. Those
respondentswho are kin-orientedin their social interactionwere those who reported
loose-knit networks among their non-kin interactors,and vice versa. This seems to
suggest, contraryto the assertionsof Bott and Gluckman,that tightnessof network
connectednessamong kin and non-kin are two ratherseparatephenomena,and that
kin and neighbors are not cemented together in an integrated huddle of social
familiarity.Even down the Hessle Road, as tight-knitand as localized a community
as is today easily found, the Bott stereotypedid not exist. Perhapsthis stereotype
belongs essentially to a social era now largely past, when ruralantecedentsstill had
theirdeep-rootedresidualsin the cities of Englandand America. One need not go as
far as Litwak in claiming a structuraland functional differentiationof all primary
groups, but it does seem as though kin are becoming increasinglyregarded as a
separaterealm of social supportand responsibility.Webber'smobile urbansociety
has taken its toll in this respect. In others, however, it has had less impact. Social
networks remain close-knit in a surprisingvariety of urban situations, and they
continue, even in this mobile age, to remain substantiallyrooted to the residential
locality.
REFERENCES
Aldous, J., andM. A. Strauss. 1966. "Social Networksand ConjugalRoles: A Testof Bott's
Hypothesis." Social Forces 44(June):576-80.
Barnes, J. A. 1969. "Graph Theory and Social Network: A TechnicalComment on Con-
nectedness and Connectivity."Sociology 3(May):215-32.
Bossard, J. H. S. 1945. "The Law of Family Interaction."AmericanJournal of Sociology
50(January):292-95.
Bott, Elizabeth. 1957. Family and Social Network. 2d ed. London:Tavistock, 1971.
Frankenburg,Ronald. 1966. Communitiesin Britain. London:Penguin.
Gluckman,M. 1971. "Preface." In E. Bott, Family and Social Network. London:Tavistock.
880 / Social Forces / vol. 55:4, june 1977