Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
FAMILY A N D H O U S E H O L D : •9631
T H E ANALYSIS O F
DOMESTIC GROUPS
INTRODUCTION
For the essential features of the individual family, as of all other social institutions,
depend upKtn the general structure of a given society and upon the conditions of life
therein A careful and detailed analysis of family life and of different aspects of the family
unit m connection with other social phenomena is therefore necessary Such an analysis
enables us to descnbe the said unit m a complete and exact way (91, p 6)
In the more than half-century that has passed since Malinowski ex-
pounded on the procedure for amving at a "scientific, correct, and useful
definition of the family in Australia" (91), it has become commonplace to
charge him with having fallen far short of these goals Less often evaluated
IS the success that anthropologists in general have attained in lUuminatmg
161
162 YANAGISAKO
"do not even belong to the same universe of discourse," and F.M Keesing
(80, p 271), who views kinship and locality as two distinct pnnciples of
organization. Bender maintains that ' famihes, as kinship units, must
be defined stnctly m terms of kinship relationships and not in terms of
co-residence" (8, p 493) For Bender, the empincal difference denves from
the observation that in numerous societies famihes do not form households,
and in even more instances, households are not composed of famihes
The contrast between kinship and locality as different pnnciples of orga-
mzation also lies behind the more specific distinction between family and
household which prevails in studies of peasant commumties Here the
family as a jurally defined, corporate km group is distinguished from the
household as a collection of kin and sometimes nonkm who share a common
residence The corporate nature of the family denves from the jural nghts
to property, usually land, which its members hold in common (29, 51, 106,
151) Accordingly, for Freedman (51, p 9) a joint family exists m China
whenever two or more men are coparceners of a chia (family) estate,
regardless of whether these men are mamed or whether they and their
respective wives and children live in different residences An obvious conse-
quence of this jural defimtion of the family ts that in societies such as China
and India, men, as lnhentors of property, are placed at the core of the
family, while women are classified as "subordinate or fnnge members" who
have lesser nghts of ownership in the estate (151)
The focus on property rights attaches to the family the specific function
of control over property, including its transmission But the more general
distinction between the family as a kinship unit and the household as a
residential unit, which Bender carnes funher than anyone else, is an at-
tempt to avoid a functionalist definition ot the family Bender asserts that
functional definitions of the family are inadequate because many of the
functions that have been construed as "family functions'* are sometimes
fulfilled by coresidenual groups* that are not based on kinship relationships
and are in other instances earned out by neither families nor households
As an alternative, he proposes that we define the family in purely "struc-
tural" terms, because in all societies people recognize kinship relationships
and use these to form social groups, and because these relationships can be
organized in only a hmited number of logical ways Like the others who
pose this distinction. Bender never makes explicit precisely what he means
by "kinship" relationships or "kinship"* units as distmct from coresidential
relationships or relationships ansmg from propinquity Yet it is evident
from Bender's amplifications on the subject that by "kinship" relationships
he means genealogically defined relationships, that is to say, relationships
that can be traced through one or more parent-child or marnage linkages
164 YANAGISAKO
also some set of activities These activities, moreover, are usually related to
food production and consumption or to sexual reproduction and childrear-
ing, all of which are glossed under the somewhat impenetrable label of
"domestic" activities Yet since all the activities implicitly or explicitly
associated with the term household are sometimes engaged in by sets of
people who do not live together, several anthropologists (8,22,128) contend
that we would do better to employ alternative terms in our ethnographic
descnptions Bulmer (22), for one, questions the adequacy of the term
"household" for descnbmg a situation where overlapping sets of people
participate m meal shanng, gardening, and coresidence He proposes that
in these circumstances the "domestic group," m which people acknowledge
common authonty in domestic matters, is a more salient social unit Seddon
(128) offers the term "budget unit" to refer to a group of individuals having
d common "fund" and exchangmg goods among each other without reckon-
ing, as distinct from individuals who live together within one homestead
In the same vem, Bender (8) pomts out that because the term household
?s always tainted by an implicit association with certam domestic functions
It IS often misused and consequently is not useful for comparative purposes
He suggests that we substitute the term' coresidential group" for household
and that we draw a clear distinction between such groups and any functions
they might fill In the case of a society hke the Mundurucu then. Bender
identifies two kinds of coresidentiai units one composed of adult males
hvmg in men's houses and the other composed of groups of women and
children By labeling these units coresidential groups rather than
households, we avoid making the false assumption that either of these
coresidential units are the most important groups for the performance of
domestic functions in Mundurucu society We would recognize, further-
more, that because domestic functions are for the most part earned out
through reciprocal interactions between adult males as a group and adult
females as a group, "the whole village forms the domestic unit, the sexual
division of labor in domestic activities being at the village level'" (8, p 495)
Bender's separation of families, coresidential groups, and domestic func-
tions IS useful to the extent that it prods the ethnographer to exphcate the
exact nature of the social unit he is labehng a family or coresidentiai unit
and to descnbe precisely the functions it performs rather than assuming
them or leaving the reader to fill m with his own cultural assumptions In
addition, by substituting the term coresidential group for household,
Bender allows for the coexistence of several types of coresidential groups
at different levels within the same society An individual may simulta-
neously belong to two or more nested coresidential groups for example, a
nuclear family hut, in a patnlaterally extended family compound, withm a
patnlmeai descent-based settlement
166 YANAGISAKO
control over household size and composition, many of wbicb are embedded
m "kinship" customs such as marnage and adoption practices and tbe
tuning of family division. For instance, in a remarkable piece of detective
work m bistoncal demography, T C Smitb (140) discloses bow m at least
one village m eigbteentb century Japan, peasants were able to prolong tbe
family's pbase of "maximum farmmg efficiency" by postponing tbe lnbent-
mg son's mamage as long as possible and m tbe meantime detainmg the
departure of bis siblings from tbe bousebold Furtbermore, we cannot
assume, as do Cbayanov and Sablms, that because wage-labor is absent,
households have access only to their own labor As Donham (37, 38)
convincingly argues, there may be other institutional contexts in whicb
labor IS transferred from one household to another Accordingly, tbe mem-
bers of bousebolds among the Malle of soutbwestem Etbiopia work about
tbe same lengtbs of time regardless of tbe bousebold's dependency ratio and
stage in tbe developmental cycle Ratber tban workmg longer hours, mid-
dle-aged men sponsor more work parties tbrough which tbey are able to
gain in tbe net transfer of labor between bousebolds (37)
Fortes's (45) model of tbe developmental cycle of domestic groups also
relies on demograpbic processes to explam tbe vaneties of domestic groups
found in a community Accordmg to Fortes, " wben it is recognized tbat
tbese so-called types are m fact phases in the developmental cycle of a smgle
general form for eacb society, the confusion vamshes" (45, p 3) The
vanous domestic groups observed at one pomt m time can be ordered into
a single developmentfj sequence, as in the case of matnfocat families m
(formerly) Bntish Guiana (136) Hence, the developmental cycle may ex-
plain wby a census sbows only a small percentage of tbe households in a
community conformmg to the ideal type (10, 72) As all domestic groups
pass tbrough different stages of the developmental cycle at different times,
a census will catch only a few in tbe ideal, complete pbase For tbis reason,
an ideal bousebold type such as the Eastem European zadruga is better
regarded " not as a form, but as a transitory state m the development
of the bousehold" (72, p 142)
While tbe developmental cycle must be reckoned witb m any analysis of
bousehold or family form, m most cases it cannot explain all the observed
vanation because factors other than tbose stemmmg from tbe process of
social reproduction may operate to produce a diversity of domestic groups
(29, 59, 141) E N Goody's study (59), for example, reveals tbat tbe
developmental cycles of compounds are charactensucally different among
tbe tbree estates (commoner, Muslim, and rulmg) m tbe divisional capitals
of central Gonja Hence, as Fortes (47, p 18) now concedes, a model of a
umform developmental cycle cannot explam wby the actual bistones of
different families entail different developmental sequences
FAMILY AND HOUSEHOLD 169
wbicb sbapes tbe actions of maximizmg mdividuals Tbe two views are well
illustrated in tbe hterature on Cbmese famibes, wbere most researcbers (1,
28,29,51,158) appear to agree on tbe pivotal role of property, but disagree
over tbe manner by wbich property ngbts are transmitted from one genera-
tion to tbe next and, consequently, over wbat detennmes tbe ttmmg of
property division Accordmg to Freedman (50, 51) and Abem (I), tbe
Cbmese fatber bas jural autbonty over tbe family property until his deatb.
Coben (28, 29), in contrast, mamtams tbat as soon as a son mames be
becomes bis fatber's jural equal and can demand divmon of tbe family
estate For Freedman. tbe greater tbe autbonty of tbe fatber, tbat is, tbe
more his autbonty is buttressed by his pobtical and economic position in
extradomestic spberes, the longer be can delay division of tbe estate by bis
sons For Coben, what detennmes the timmg of division is not tbe extent
of tbe fatber's autbonty over sons wbo wisb to be free of it, but tbe economic
self-mterest of the sons. The pursuit of this disagreement, which may denve
in part from actual regional differences, would hardly seem to be productive
as It would lead only to further rafymg the faulty dichotomy of jural and
economic spheres which muddles discussions of property relationships To
view property eitber as the mere vehicle of jural relationships (46) or as an
economic relationsbip between people and tbmgs (29) is to misconstrue
wbat IS more correctly an "entire system of rules, ngbts, and expectations
of role" (73, p 207) or a "system of social relationsbips" (16, p 204) wbich
must be explamed rather tban adduced as explanation
In tbe same bght, we must eschew narrow conceptions of property rela-
tionsbips as man-land relationsbips determined by "ecological factors " Tbe
limitations of a narrow ecological framework are well documented m G A
Colber's (30) study of family orgamzation and land tenure m six Maya
commumties Although be began witb a cultural ecological framework
empbasizmg tbe local adaptation of people to resources. Collier became
mcreasmgly aware tbat an adequate explanation of tbe relation between
family orgamzation and land tenure required an understandmg of tbe re-
gion-wide system of lnteretbnic and lnterclass relationsbips linkmg famibes
to hamlet, townsbip, state, and nationai processes Goldscbmidt & Kunkel
(55) reacb a similar conclusion in tbeir cross-cultural survey of the relation-
ship between degree of land scarcity, pattem of land lnbentance, and "fam-
ily stmcture" (by wbich they mean bousehold composition) m 46 peasant
commumties Tbey find that the "ecological context" suggested by E R
Wolf (159) cannot explam tbe different pattems of land inhentance in tbese
communities because there is no conclusive pattem of association between
land scarcity and eitber residence or mbentance pattems Consequently,
tbey direct our attention to Wolfs (159) "bierarchical social context" by
suggesting several ways m wbicb the relationship between peasants and the
FAMILY A.ND HOUSEHOLD 171
ebtes who control state organization (and thereby land tenure pattems)
shapes tbe form of mbentance and, consequently, peasant family structure
Goody's (61, 62) grand theory of tbe evolution of domestic organization
also focusses on tbe transmission of property as a key feature of the domes-
tic domain Goody proposes to explam not just the structure of domestic
groups, but a whole cluster of interlinked elements, includmg mamage
transactions, mbentance pattems, descent groups, form of mamage, do-
mestic roles, and kinship terminology Central to his theory is tbe contrast
be draws between African bndewealth societies and European and Asian
dowry societies, in wbicb two ver>' different forms of mamage transactions
generate far-reacbing consequences for domestic organization Bndewealtb,
as a transaction between tbe km of the groom and tbe kin of the bnde, forms
part of a circulatmg societal fund which has a leveling effect on wealth
differences In contrast, dowry, as a type of premortem inhentance of the
bnde, sets up a conjugal fund for a marned couple, thereby reproducing
wealth differences As different mechanisms for redistnbutmg property,
tbese mamage transactions are linked by Goody (61) to other aspects of
social organization Dowry systems are seen as inherently bilateral, because
tbey distnbute ngbts m a manner tbat does not link property to sex
Furtbermore, as a form of "diverging devolution," dowry entails tbe trans-
mission of property outside the unilineal descent group, thus weakenmg the
corporate nature of these descent groups In companson, mhentance witbm
tbe unilmeal descent group in Afncan bndewealtb societies strengtbens
tbeir corporate nature Finally, bndewealth and dowry have different conse-
quences for the "domestic role compendia" (62, p 41) The status of women
as heirs to sigmficant property in a dowry system leads to monogamy
(sometimes accompamed by concubinage), the concept of conjugal love,
and the mdividuabzation of the mother (61, p 37, 62, pp 42, 51) In
contrast, bndewealtb is associated with polygyny, diffuse domwtic relation-
ships, and tbe classificatory role of mother
There can be bttle doubt that Goody's typological scbeme deserves to be
a major focus of discussion and researcb on marnage transactions and
domestic organization for some time to come Most provocative are tbe
questions raised about bow the transmission of property—whether in tbe
form of mamage payments, premortem inhentance, or postmortem mben-
tance—shapes both the internal structure of domestic units and economic
itnd political processes usually construed as extemal to tbe domestic do-
main What remains to be debberated and funher researched is tbe empin-
cal generality and analytic utility of his typology For example, there is
ample evidence in tbe ethnograpbic literature on bndewealtb transactions
in cognatic societies (e g 6) and in the more "loosely structured" umhneal
descent societies of highlands New Gumea fe g 149), wbich does not appear
172 YANAGISAKO
to fit well with bis African bndewealth model Indeed, one wonders whether
we are due for another round of debate over tbe utibty of "Afncan models"
as was our fate m the case of umhneal descent systems
Even more disputable is Goody's characterization of dowry m European
and Asian societies Accounts of dowry transactions in Japan (140), south-
ern Italy (33), and Cbma (93) do not sustain Goody's claim tbat dowry is
best seen as a form of female mbentance of "male property " Even Goody's
coeditor Tambiab (151) cites cntical differences among tbe Soutb Asian
societies of nortbem India, soutbem India, Ceylon, and Burma in jural
conceptions of female ngbts to property, in tbe manner m wbicb property
IS actually transmitted to women, and in tbe consequences of tbese trans-
missions for domestic organization These recalcitrant cases, of wbich I
suspect we will hear more m the future, are ample remmders tbat the task
awaiting us is to "decompose the property of tbe family or tbe conjugal
estate into its exphcitly recognized components and see what ngbts hus-
band and wife enjoy m relation to them at mamage and divorce" (151, p
153)
Although It reqmres refinement. Goody's contrast between bndewealth
and dowry, at least as two ideal types, provides a promismg msigbt into tbe
impact of mamage and its accompanying transactions in tbe sbaping of
domestic relationships. His causal model in which technological factors of
production are tbe pnmary movers in the evolution of domestic organiza-
tion probably will not stand so well agamst the test of time His attempt to
denve bistoncal inferences from cross-sectional data by subjecting correla-
tions between mode of property transmission and societal institutions to
patb analysis does little justice to tbe complexity of histoncal developments
that bave led to tbe kind of dowry system he descnbes [see Stone (146, 147)
for a discussion of the sigmficant cbanges tbat occurred in early modem
England witb regard to tbe control of property by family beads and wom-
en's legal ngbts to property] Tbe problem, of course, is inberent in any
evolutionary scbeme that rests on a crude succession of types For bowever
sophisticated the quantitative hardware, one cannot denve bistoncal pro-
cess from abistoncal, cross-sectional data
There is a final issue tbat I fear will be overlooked in the debate over
Goody's typology and evolutionary scheme, because it is so embedded
withm Goody's discussion that it is likely to go unnoticed In Goody's
evolutionary account of the shift from a bndewealtb to a dowry system,
radical transformations occur in mode of production, division of labor,
stratification, property relations, centralization of political autbonty, mar-
nage forms, km terminology, and domestic role relationsbips Yet one tbing
remams constant the nuclear family is the basic productive unit of society
(62, p 20) Because he sees the nuclear family as everywhere and for all time
FA^MILY AND HOUSEHOLD 173
tbe basic and natural unit of society. Goody never feels compelled to explain
wby, when intensive agriculture created greater production surpluses and
differences in styles of life, people became concemed to reproduce these
life-styles in their own progeny as agatnst the members of tbe lineage or
otber kinship or locality grouping For Goody, there is no need to explain
tbe emergence of the nuclear family as a socially and culturally significant
unit because it was already, and always, present In the final section of this
paper, I will retum to a cntical commentary of this assumption, which is
held by many more antbropologists than just Goody
A second productive resource cited as a determinant of household struc-
ture IS labor Labor requirements are singled out by Pasternak, Ember &
Ember (108) as the most powerful determinant of the formation of extended
family households Usmg a randomly selected sample of 60 societies, they
test their hypothesis that
extended family households are likely to emerge and prevail in a society when (in
the absence of slave or hired labor) work outside the home makes it difficult for a mother
to tend children and/or perform her other regular nme-consummg domestic tasks, or
when the outside activities of a father make u difficult for him to perform his subsistence
work (108 p 121)
"work away from tbe home for more tban balf tbe day for at least (cumula-
tively) 30 days of the year" (108, p 118)
Tbe most vexatious issue tbat Pasternak, Ember & Ember and tbe otbers
(55, 62) wbo attempt statistical cross-soaetal compansons encounter is
wbetber tbe units tbey have selected are appropnate for comparative pur-
poses Tbe evidence tbat societies contam different frequencies of a range
of bousebold types surely cballenges tbe classification of wbole societies mto
two categones of tbose witb "extended family bousebolds" versus those
witb "mdependent family bousebolds " Our goal of understandmg and
explaining domestic orgamzation may be better achieved by investigations
of tbe diversity of domestic umts m societies and tbe articulation of tbese
domestic umts witb one anotber, ratber tban by compansons tbat reduce
wbole societies to a smgle bousebold type, family structure, or mamage
transaction
Just as labor reqmrements are used to explam cross-societal vanations in
domestic groups, so a large number of researcbers bave used it to explain
tbe frequency witb wbicb a particular housebold type appears m a society
or segment of a society In reviewmg ecological studies, Nettmg (102) cites
several studies whicb purportedly sbow that bousebold composition among
horticultunsts and agnculturalists vanes with tbe type and amount of labor
reqmred for effective crop production In his own study, for example, he
attnbutes tbe differences between tbe large, extended famibes of sbifting
cultivators and tbe nuclear family bousebolds of mtensive agnculturalists
m the same Nigenan plateau environment to diffenng labor reqmrements
(100,101). Along tbe same lines. Sabbns (119) argues tbat under conditions
of widely dispersed resources tbe extended family provides a more efficient
productive unit. Nakane (97) attnbutes tbe larger-tban-average mean
bousehold size in two Japanese commumties to tbe labor requirements of
tbeir productive activities (fisbmg m one, silk cultivation m tbe otber), and
Dorjabn (39) cites labor demands of upland nee farmmg as the pnme factor
m bousebold size and type among tbe Mayoso Temne In tbe case of tbe
Cbmese family, botb Pastemak (107) and Coben (29) stress tbe need for tbe
efficient allocation of labor as a primary determinant of family form
In all these studies, the authors assume tbat once baving identified tbe
productive activity of tbe bousehold, tbey can ascertain tbe technologically
determmed labor reqmrements of tbat activity and proceed to sbow that tbe
household efficiently meets these labor requirements They fail to explain
adequately, however, tbat it is not merely tbe tecbnological requirements
of production wbicb necessitate a particular bousebold type, but tbe entire
manner in which production is socially orgamzed In otber words, a produc-
tive activity can be accompbsbed by a vanety of bousehold types depending
on how tbat production is socially organized, includmg tbe division of labor
by sex and age, tbe use of hired labor, cooperation in productive tasks
FAMILY AND HOUSEHOLD 175
production units witb extremely limited resources and are greatly subject
to the forces of nature, the market, and the state, they expenence some
degree of random oscillation m social mobibty (129, p 112) Among the
Russian peasantry in tbe early twentietb century, for example, tbere was
considerable multidirectional and cycbcal mobibty of bousebolds (129) T
C Smitb (140) also cbaractenzes tbe situation in Tokugawa Japan as one
in wbicb farm famihes were not likely to enjoy an economic advantage over
tbeir neigbbors for many years On the other band, it would be wrong to
conclude that because of this mobility observed differences m wealtb are
mere transitory advantages that have no consequence for domestic orga-
nization Despite the mobibty in Smith's village, tbe highest probabibty was
for a family to remam in the same landholdmg category from one tax
register to another—an average interval of 12 years (140, p 119)
Sigmficant differences in wealth among peasants also may be overlooked
because tbey are so often obscured by kinship relationsbips wbicb bind tbe
landed to the landless and the land-ncb to tbe land-poor (95, 129, 130)
Consequential wealth differences may exist between tbe smaller coresiden-
tial units of a single kin-based compound (54) or between brotbers wbo bead
the main and branch bouseholds of a family (140)
Tbe complexity of tbe intemal differentiation of tbe peasantry, or of any
kmd of community, sbould not deter us from investigating its relation to
the vanations we observe in tbe composition and mtemal structure of
domestic groups, as well as in their extemal relationships Indeed, it is this
very complexity—the fluctuations in size and wealtb, tbe social mobility,
and tbe kin ties tbat bind togetber bousebolds in different strata—that
underlines our need to mvestigate the economic and social interdependence
of different sectors of society
kmship change as much as theirs, since many of their notions bave been
borrowed from anthropology and sociology Indeed, the common guiding
frameworks may be more easily discemed m the context of a newly develop-
ing field of inquiry such as the historj^ of the family After I discuss the
methods and analytical strategies employed by family histonans, I move on
to d review of anthropological studies of family and kinship change
^ A review of the Journal of Family History the History of Childhood Quarterly, the Journal
of Interdisciplinary History, the 4inale'i (Trance), and Pasi and Present (England) confirms
ihis concern with the familv
178 YANAGISAKO
directed by Laslett Tbe main tbmst of Laslett's work (85-87) is tbe refuta-
tion of the theory tbat a shift from an extended family system to a nuclear
family system accompamed European mdustnabzation His discovery that
early census matenals sbow tbe small, nuclear family bousehold to bave
predominated m England before the mdustnal revolution is ample evidence
of tbe contnbution tbat bistoncal demograpby can make to the construc-
tion and revision of bypotbeses about tbe bistory of tbe family Having
found similar evidence of tbe predominance of small bouseholds elsewbere
in premdustnal Westem Europe and even m Tokugawa Japan (76), bow-
ever, Laslett boldly extends his thesis to assert tbat "bttle vanation in family
orgamzation can be found m buman bistory" (86, p. lx) Wbile he is aware
of tbe distmction between bousehold size and family organization, at times
Laslett lapses mto the unfortunate practice of confusmg family organization
with residence pattems Because be relies pnmanly on aggregate data on
bousebold size and composition, Laslett's work suffers from a narrowness
as well as a pletbora of metbodological and conceptual problems [see (12)
for an excellent cntique of Laslett's work] A few of tbese problems are
wortb bstmg bere because tbey plague tbe work of otber quantitative bis-
tonans of tbe family equating residence pattems with family structure and
even kinship organization, paymg httle heed to relationsbips between resi-
dential units, failing to consider regional diversity, failmg to take into
consideration the developmental cycle of domestic groups, and makmg
unwarranted assumptions about tbe common cntena for definmg
bouseholds used by census takers in different societies
Tbese conceptual and metbodological muddles are well illustrated by the
hterature on tbe history of the black family m the United States In reassess-
ing the slave expenence tbrougb reconstructmg tbe bistory of black fami-
lies, scbolars bke Gutman (69, 70) assert tbat, contrary to received wisdom,
blacks in tbe Umted States lived predommantly in two-parent bousebolds
botb before and after emancipation Gutman's figures certainly attest to tbe
need for a reassessment of popular notions of tbe black family But be and
otber family bistonans (53) muddle the issue by assuming that if husbands
are present m the bousebold, tbey play a central role m tbe family and,
tberefore, tbat tbese cannot be "matnfocal," "matnarcbal," or female-
beaded bousebolds Yet tbe pomt tbat presence or absence m a bousebold
IS qmte a different tbmg from an mdividual's structural role m tbe family
was made convincingly clear 20 years ago by R T Smitb (136)
Fortunately, not all family bistonans wbo rely on quantitative methods
suffer from tbese faibngs. Some bistonans of tbe black family (77, 131) are
quite cognizant of the distinction between household composition and fam-
ily structure, and tbose antbropologists wbose works appear m bistoncal
antbologies bave tned to dnve home this distinction (60, 71) Tbe most
F-VMILY AND HOUSEHOLD 179
analytically refined works seem to come from bistonans wbo use a broader
range of source matenals in conducting intensive researcb on a specific
locality at a single point in time, tbus avoidmg an unbealtby rehance on
aggregate figures wbicb obscure substantial differences among strata, re-
gions, and etbmc groups In some histoncal studies of the family, the
authors take into account tbe developmental cycle of domestic groups, even
in using aggregate data (10, 13), place the household withm its proper
kinship context (82, 156), attend to locai vanations (65, 142), and pay heed
to the influence of other cultural domains, such as rebgion, on family
ideology (7, 88) Greven (67,68), for example, utilizes botb tbe genealogical
bistones of specific famili^ and aggregate data on tbe entire population in
his case study of population, land, and family in a seventeentb-century New
England town His reconstruction of the household and kinship networks
m the community enables bim to mterpret the organizational stmcture of
individual households within the context of tbeir relationships with other
households Similarly, Kent's study (82) of fifteenth-century anstocratic
families in Florence shows tbe relations among the individual, housebold,
and bneage to be more subtle tban is suggested by overly dicbotomized
conceptual frameworks which portray the household and lineage as mber-
ently antagonistic structures Finally, histoncal researcb on mbentance
systems in different regions of Westem Europe (14. 153) promises to
provide us with useful analyses of the interplay between the inhentance laws
of the state, local customs, and the actual mbentance practices of different
sectors of society
Like anthropologists, social histonans are searching for analytically pro-
ductive ways to identify and descnbe family change Stone's (146, 14"?)
efforts to display tbe cbanges in English family structure in tbe early mod-
em penod lead bim to the construction of a typology of three successive
family types While there is much that can be objected to in his mterpreta-
tion of evidence (154), his attention to pubbc ideologies of tbe family as well
as to family functions contnbutes much ncher matenal for an analysis
of histoncal change than the absorption with surface level similanties of
demographic statistics can offer
The most formidable problem which confronts histonans of tbe family
IS the matching of data sources with the ngbt kinds of questions Wbere
they are least successful m their endeavors is when, like Laslett (86) and
Gutman (69, 70), they attempt to answer questions about family stmcture
for which their quantitative sources are ill suited, or wben, like Shorter (132,
133), they try to "extract emotional motivations from unwillmg statistics"
(C Daniels, unpublished paper) Tbey are most successful wben, bke T C
Smitb (140), they make creative but judicious use of quantitative matenals
to calculate measures of age-specific mortality and fertility, life expectancy.
180 YANAGISAKO
nuptiality, and birth spacing, and supplementmg these with literary and
legal records, amve at conclusions about population growth, infanticide,
and tbe dynamics of tbe bousebold in a specific locality
Altbougb tbe research of histonans has certainly yielded much of sub-
stantive mterest and sigmficance for discussions of the bistory of tbe West-
em family, tbe imphcations of tbese results for our understanding of family
cbange are obscured by tbe absence of a coberent tbeoretical framework
A ratber odd grab-bag of conceptual tools and analytic strategies bas been
hfted from tbe social sciences by bistonans Perhaps most lU-advised is tbe
proclivity of some social bistonans (86, 87, 132, 133) to infer changes in
sentiments and cultural conceptions from alterations in demographic pat-
tems Because the illiterate masses leave behind mainly vital statistics wbicb
do not speak of attitudes and behefs, they are particularly vulnerable to this
kind of interpretation Given tbe available source matenals, tbe adoption
of psycboanalytic models to speculate about tbe histoncal consequences of
childhood expenences in different penods also seems particularly lnappro-
pnate Few histonans go as far as deMause (35) to propose a "psychogenic
theory of bistory" which applies Freudian and neo-Freudian tbeones to the
history of cbildbood, but an emphasis on socialization and an lmphcit, if
undeveloped, theory of the importance of early cbildbood expenence in
sbapmg personabty and culture is present m tbe works of otber family
bistonans (86, 87) It is because of this socialization bias tbat Laslett
cbooses as tbe predominant famibal type in a society tbe form in which the
largest percentage of children have grown up "in their most impressionable
years" (86, p 67) This remvention of personality and culture tbeory leads
bistonans like Laslett to place an inordinate empbasis on a single family
function to tbe detriment of otber, no less important, functions (12, 156)
The most popular import from tbe social sciences, bowever, has been a
somewhat loosely conceptualized, Parsonian eqmbbnum model of society
In general, family structure and function are seen as adjustmg to tbe shifting
extemal conditions of tbe economy and political orgamzation (11, 118, 146,
147) Economic factors may be singled out as having a major causal influ-
ence on the family, as in Greven's study (67) wbere tbe control of land by
tbe town's foundmg fatbers leads to tbe estabbsbment of "extended patnar-
cbal famibes " In any event, alterations m family form and functions are
portrayed as attempts to make domestic relationsbips more congment witb
tbe outside social environment I include withm tbe general category of a
Parsonian equihbnum framework tbe "actor-onented" approacbes (4, 5,
118) wbicb purport to explain particular family structures by reconstruct-
ing tbe social options open to maximizing mdividuals For despite tbe
interest of histonans such as Anderson (4, 5) in decision-making processes
h \MILY AND HOUSEHOLD 181
and the inclusion of formalist premises such as "all actors have a number
of goals, tbe attainment of which would maximize tbeir satisfactions or
psycbic rewards" (4, p 8), in the end the explanation of family relationships
IS a functionalist one Hence to explam the high incidence of coresidence
of mamed children and parents in the early mdustnal English town of
Preston, Anderson (4) cites the economic benefits for both tbe marned
couple and parents Much rarer are the histonans like N Z Davis wbo
recognize tbe conflicts and tensions as well as the consistencies between
family life and political, economic, and religious institutions Davis's discus-
sion (34) of tbe disjuncture between "pnvatistic family values" and tbe
"more corporate values" held by the same families in early modem France
conveys a more refined notion of cultural systems than do the analyses of
family histonans who are inclined toward tightly integrated functionalist
schemes
Several family histonans (e g 63. 133) have been attracted to the simple
dichotomies of traditional vs modem, rural vs urban, mechanical solidanty
vs organic solidanty and instrumental vs affective families While this is
hardly tbe place to review tbe sbortcomings of a modemization framework,
suffice It to say that it is somewhat ironic that histonans would find a static
functionalist perspective so appealing just at a time wben anthropologists
and sociologists are searching for more dynamic models of society
As I implied m the beginning of this section, the answer to the question
of why histonans have not been more successful in integrating their findings
into a unified theory of family change lies, at least in pan, in the shortcom-
ings of our own discipline It is to this body of literature that I now tum
except for family firms and some control ot joint property, decisions m the set of
extra-familtal km in a modern Western urban society are made in nuclear family units,
however influenced they mav be by kin ties outside (44, p 456)
Finh, Huben & Forge are nght that some of the debate is created by
researchers who tend to overstate the significance of km ties But in addition
to Its androcentric bias, tbeir distinction gives undue weight to centrahzed
authonty as the defining feature of the extended family Kin groups can
make decisions and take cooperative action without a central authonty
And even when a dispersed family is not specifically a political or economic
group, this does not mean that it cannot be used for pohtical or economic
purposes (15) It is wrong, therefore, to accept the absence of km groups
resembling those found in lineage societies as evidence of a decline in the
lmponance of tbe extended family
If we are to refine our analysis of family and housebold change, we must
begm to ask new questions of our data Questions as to whether or not the
extended family or kinship in general has declined in industnal-urban soci-
ety or whether the family has endured even under conditions of rapid social
change have impeded our progress toward a more refined analysis of
cbange After all, structural cbange and structural continuity in family and
kinship institutions are not mutually exclusive phenomena (135, 163) One
way to move toward a more refined analysis of change in family and kmship
IS to examine the relationship between change in the ideology of family and
kinship and change in actual institutional arrangements In the case of the
Balkan zadmga, Hammel (71) concludes that recent changes can be at-
tnbuted largely to alterations in demographic rates and extemal constraints
while the underlying kinship pnnciples (e g vinfocality, agnatic bias, pa-
tnfocality, and lineage organization) have remained relatively unmodified
Brown (20) suggests similarly that we cannot assume tbat mdustnal-urbani-
zation in Japan has necessanly resulted m the demise of tbe dozoku (a group
of households related in a network of mam and branch ties) even though
the classic type of dozoku with its economic correlates is no longer viable
The ideology which has yielded dozoku organization can persist despite
alterations m the observable organizational forms
Tbe cntical question which anses when we employ this analytical strat-
egy IS how do we identify the underlying, ideological pnnciples of forms hke
the Balkan zadruga and the Japanese dozoku to discem whether ideological
change has occurred"* For the zadruga, the underlying kinship pnnciples
extracted by Hammel (71) consist of an assortment of things, some of which
appear to be normative ruies or preferences (patnfocality), other of which
are observable behavioral tendencies (agnatic bias), and still otbers of which
(vinfocalitv) are ambiguously either obser^-able pattems or normative rules
184 YANAGISAKO
As some of tbe basic pnnciples are outcomes of other pnnciples wben tbey
are combined with certam demograpbic and social constraints, tbe meamng
of "ideological pnnciples" is equivocal
Tbe assessment of ideological cbange. of course, is difficult where tbe
antbropologist, hke tbe bistonan, does not have direct evidence of former
cultural ideals, values, and meanmgs Most commonly, the anthropologist
IS forced to reconstmct past ideologies from contemporary accounts. If be
tben compares tbis abstracted ideology of the past with tbe more ncbly
contextualized contemporary ideology, his evaluation of cbange may be
limited by tbe vagueness of bis constmcted model of tbe past If we are to
bnng to our analysis tbe knowledge tbat cbange m observable behavior does
not necessanly mean tbat cultural ideologies bave altered, we need a con-
ceptual framework tbat can belp us to systematically differentiate and
display tbe mteraction between tbese aspects of family and kinsbip Tbe
great promise of sucb an analytic strategy lies m its capacity to identify tbe
dynamic tension between ideology and action as a possible source of cbange
(163), thereby transcending the bmitations of an analytic framework that
lnvanably attnbutes cbange to tbe constramts imposed by factors extemal
to family, bousehold, and kmsbip
may be altogether the wrong strategy This is tme of sets of households that
share a corporate identity as well as those that do not, but which may have
other Significant relationships It is too often thought that by labeUing a
domestic umt a "stem-family" we have adequately descnbed both its inter-
nal organizational structure and the relationships that articulate it with
groups and mdividuals outside it But if we compare, for example, stem
family households m eighteenth century Austna (10) with stem family
households m eighteenth century Japan (140), the economic relationships
between households in these two societies are clearly different
One kind of analytic advance we can make is illustrated by Cohen's (29)
dissection of Chmese family organization He observes that while the Chi-
nese family (chia) is a discrete kin group.
It can display a great deal of vanation in residential arrangements and in the economic
ties of its members For purposes of analysis the chia estate, economy, and group can
be considered as three basic components in chia organization (29, p S8)
The chia estate is "that body of holdings to which the process of family
division IS applicable," the chia group is "made up of persons who have
nghts of one sort or another to the chia estate at the time of family division,"
and the chia economy "refers to the exploitation of the chia estate as well
as to other mcome-producing activities lmked to its exploitation through
remittances and a common budgetary arrangement" (29, p 59) Byexamm-
ing vanations in the connections between chia components, Cohen succeeds
in clanfying some of the problems encountered when the Chinese family is
descnbed m terms that wrongly assume an ideal chia m which the estate,
group, and economy are umfied
An analytic framework such as Cohen's, of course, is useful only when
applied where the common ownership of property is the key element bmd-
lng together domestic units Where this is not the case, as among landless
peasants, hunter-gatherers, or wage-laborers, we may usefully identify other
significant components of family organization, including the commensal
group, the production group, and the budget group m which reciprocal
exchange occurs without accounting (cf 90, 128) However, as I have al-
ready remarked, the aggregate of people engaged in any of these activities
may change throughout the production cycle, the exchange cycle, or the
individual's hfe cycle Consequently, it seems more analytically strategic to
begm With an investigation of the activities that are central to the domestic
relationships m each particular society, rather than with its domestic
groups If we start by ldentifj^ng the important productive, ntual, political,
and exchange transactions m a society and only then proceed to ask what
kinds of kmship or locality-based units engage in these activities, and in
FAMILY A.ND HOUSEHOLD 187
ders if this is really the case or whether their unifiwi treatment stems more
from their approach to the domestic domain Their discussion ofthe politics
of mamage among the gentry (54, p 131) makes it clear that there is much
more complexity m mamage relationships among the gentry than among
commoners and consequently greater status distmction among gentry sib-
lmgs If this iS the case, we would expect to find some salient differences
between commoner and gentry sibling relationships and in the relationships
between households within a common houseyard which are linked by a
sibhng tie These matters are never discussed, however, and we are led to
believe that despite significant differences between the two status groups in
the operation of public kmship institutions, kinship in the pnvate domam
IS essentially the same for both
This kmd of oversight occurs, it seems, because Geertz and Geertz begin
their analysis of kinship with a discussion of the domestic domain and then
never retum to it after they have extensively analyzed the public domain
As a result, their analysis of interpersonal relationships m the domestic
domain does not benefit from their exposition ofthe complex forces shaping
the public domain of kmship Ind^d one wonders whether the thinner
of our descnptions of domestic relationship^ is partly an artifact of our
habit of beginning with the domestic and then moving to the pohtico-
lural—a habit that we may have unthinkingly inhented from Malinow-
ski
The distinction between the domwtic and politico-jural domains (or the
pnvate and public domains) calls for stnct scnitmy not just because of its
analytical consequences, but because it is the encompassing framework for
a cluster of notions which pervade anthropological studies on the family and
household Included within this dense conceptu^ network is the conviction
that the core of dom«tic relations is the mother-child bond While there
may be differences in the manner in which the mother-child unit is hnked
to larger orgamzational structures, the bond itself is perceived as essentially
the same everywhere and denved from the biological facts of procreation
and nurturance (45, p 8, 46, pp 251, 255-56. 47, p 21, 49, p 37, 58, p
18) Closely tied to this notion is the idea that the mother-child relationship
IS constrained by affective and moral convictions generated by the expen-
ence of "mothering" necessary for the biological survival of human off-
spnng These affective and moral constraints permeate the entire domain
of domestic relationships, thereby distingmshing them from those relation-
ships which are ordered by political and jural prmciples Finally, as was
seen in the attempts to define family and household, there is the belief that
reproduction—that iS, the provision of properly enculturated personnel to
fill social positions necessary for the perpetuation ofthe social order—is the
pnmary activity of the domestic domain
190 YANAGISAKO
Inquines mto women's relationships with people outside their own do-
mestic group refute the notion that it is lnvanably men who link mother-
child umts to larger institutional structures in society Women's
involvements m exchange transactions (149, 155), in informal women's
commumties (99, 111, 160), and in urban kin networks (164) are now
mterpreted as having significance for extradomestic arrangements rather
than as mere extensions of women's domestic onentation [see for example
(56)] Moreover, domestic relationships are often so lnextncably lnter-
meshed with relationships of political alliance that to separate the domestic
aspects from the political aspects is to misconstrue these relationships M
Strathem's (149) discussion of divorce and the attnbution of blame for
divorce on women's "will" by Mt Hageners is a particularly telling demon-
stration that "domestic" aspects of the conjugal relationship sometimes
cannot be usefully separated from the alliance aspects of the relationship
Taken together, these studies push beyond the recognition that domestic
relationships are influenced by extradomestic, politico-jural considerations
to the realization that domestic relationships are pan and parcel of the
political stmcture of a society Although some initial explorations on sex
roles may have been guided by a domestic versus public distinction (52,113,
121), more recently there appears to be an emerging consensus that this
dichotomy is analytically unproductive and empmcally unfounded (109,
110, 114) Too many studies of women's "domestic" activities have dis-
closed that these have pohtical as well as reproductive consequences for us
to continue to accept the domestic/public dichotomy as a descnption of
social reality It now seems more productive to mterpret the dichotomy as
"a cultural statement masking relations which are highly problematic"
(110)
As research on women has taken a closer look at gender ideologies, it has
also begun to dispute the notion that the biological facts of reproduction
produce an immutable mother-child relationship "Motherhood," it tums
out, IS not everywhere construed in the same manner, nor do all gender
ideologies place equal emphasis on "motherhood"' as an aspect of woman-
hood (115, 116) Revelations about nannies, wet nurses, and other "surro-
gate" mothers disclose that even in our own Euro-Amencan history one
cannot assume that "motherhood" has always entailed the same functional
components or the same components of meamng (18, 40)
Instead one asks what the defimtion of the domain of kinship is for each
culture studied By abstractmg normative mles from concrete, observable
actions (which mcludes verbal statements), the anthropologist denves the
system of symbols and meanmgs pertainmg to kmship relationships (126,
P 38)
Because it directs us to conduct thorough investigations of native concep-
tual categones, symbohc analysis produces ncher and more precise ethno-
graphic accounts than do analyses that fail to mterpret social umts and
actions Within their relevant contexts of meanmg The advantages of sym-
bohc analysis as an analytic tool for the comparative study of family and
kinship are attested to by studies (79, 127, 134, 157, 165) that employ
Schneider's approach For example, Inden & Nicholas (79) demonstrate
that a Bengah kinship umt, which from a purely genealogical perspective
appears to be identical to the Euro-Amencan "nuclear" family, is con-
structed out of very different cultural meanmgs and normative expectations
than are Euro-Amencan families In addition, symbohc analysis enables us
to see that native km categories, including family and household, are often
polysemic, that is, they encompass a range of different meanmgs (26,40,98,
117, 134, 165) R I Rosaldo's analysis (117) reveals that the Ilongot
category name "be.rtan" is used in a number of different senses that cannot
be conveyed adequately by any smgle, reduced anthropological concept,
such as the deme or nonunihneal descent group By mtroducmg a dia-
chronic perspective, he is able to show that these category n a m ^ are best
interpreted as "a means of identifymg bounded groups at different phases
in a single histoncal process" (117, p 18) In my analysis of Japanese-
Amencan kmship (165), I conclude similarly that the category "relative"
has different meamngs and is composed of different types of umts dependmg
upon the cultural context Hence, if we are to imderstand the nature of
kinship categones in a society, we must investigate the diverse meanmgs
attached to them in actual usage And contrary to the claim that this kind
of ethnographic specificity makes comparative studies unfeasible, Needham
(98) and others (117, 124) argue persuasively that it puts us m a better
position to make compansons because it allows us to "see the social facts
m a less distorted way " (98, p 70)
FAMILY AND HOUSEHOLD 193
the existence of slavery produced a social order which assigned to the coloured
people, whether slave or free, the lowest rank m the social hierarchy to perpetuate
this hierarchy it was essential to proscnbe mamage between the dominant and the
dominated groups Yet, partly for demographic reasons, white men had to resort to the
coloured community for women By virtue of the hierarchical nature of the soaety these
unions as a rule took on the form of sporadic or stable concubinage Both the free men
and the slave women aquiesced m this mfenor form of mating because, due to the racial
overtones of the system, it was a most effective means of social advance to them and
particularly to their ofisprmg The pnnciple of hypergeneration applies as well to mtrara-
cial unions (92, p 124)
Hence, it was the hierarchical nature of the social order that produced the
'sexual margmahzation" of colored women, and this m tum produced
concubmage and matnfocahty
As R T Smith (139) perceptively notes, Martinez-Aher's analysis
eclipses other explanations of matnfocahty m the Canbbean because it does
not focus on isolated 'economic factors" like income or occupation, which
cannot adequately account for the diverse types of domestic union observed
By focusing on relations of inequality, includmg conjugal umons, between
members of different strata, she casts light onto the marnage relations and
forms of domestic union within each stratum
A third study that bnngs social inequality into a discussion of domestic
orgamzation is Gough's (66) reanalysis of Nuer society as a society in
Transition rather than in timel^s equilibrium Gough's reexamination of
Evans-Pntchard's (41, 42) data on Nuer domestic relations uncovers the
fact that a large proportion of the population did not conform to the
•^'agnatic pnnciple " By presenting evidence ofthe differences between Nuer
anstocratic lineages and Dinka or Nuer commoner lmeages in forms of
domestic umon, postmantal residence, and the tracing of descent, Gough
IS able to argue persuasively that Evans-Pntchard overlooked a marked
"skewing" in the operation of the agnatic pnnciple among different seg-
ments of the population Having recognized this unevenness in the opera-
tion of the agnatic pnnciple. she offers a compelling histoncal explanation
of this vanability
What IS particularly notable m Gough's analysis is that she draws on two
different sets of matenais, each of which Evans-Pntchard presented m a
separate monograph (41, 42) Evans-Pntchard, of course, was committed
to the idea that " the relations between the sexes and between children
and adults belong rather to an account of domestic relations than to a study
of pohtical institutions" (41, p 178) Consequently, he excluded such rela-
tions from his analysis of Nuer social structure and relegated them instead
to a second volume in which he discussed incest prohibitions, mamage,
types of domestic unions, and interpersonal kinship relationships Gough's
creative contnbution denves from her unwilhngness to accept this separa-
tion of the domestic and political spheres By bnnging together in a single
analysis what Evans-Pntchard separated into two, she elucidates the work-
ings of a social system that is as much produced by relations between the
sexes and between children and adults as it is produced by relations between
men Her reanalysis should lay to its final rest the false notion (cf 49, p 22)
that one can descnbe the political system of a society without taking mto
account the web of interpersonal relationships of kinship
All three of these works show marnage to be central to the creation of
different types of domestic groups Ail three are able to display the lnterrela-
196 YANAGISAKO
19) IS the behef that the vital function of the family is the beanng and
reanng of children Otherwise there would be no reason for choosing moth-
ers and children as the atom of the family Goodenough's definition, there-
fore, lromcally suffers from the same ethnocentnsm that he attnbutes to
Murdock's definition By assummg that wherever we find mothers and
children (a biological given) they form the core of the family, he has himself
taken a functionally significant unit in our society and treated units that
resemble it elsewhere as if they were fundamentally the same things While
he IS undoubtedly right that in every human society mothers and children
can be found, to view their relationshtp as the universal nucleus of the
family is to attnbute to it a social and cultural significance that is lacking
in some cases Merely because one can identify a relationship that bears a
genealogical resemblance to our own mother-child relationship does not
prove that people everywhere attach to this relationship the same cultural
meanmgs and social functions, nor that it forms the structural core of larger
kinship groups We do not, after all, insist that unilineal descent groups are
present everywhere just because in every society we can ferret out of
genealogies a set of umhneal descent relationships Just as a unilineal de-
scent group requires the attachment of concrete functions and a culturally
recogmzed identity, so any unit designated as the nuclear family must be
shown to engage in some socially significant activity and to be imbued with
some consequential meaning Goodenough does not feel compelled to
present us with proof that the mother-child dyad everywhere has a central
functional and meaningful role, because he assumes that nurturance by the
mother is required for the biological survival of human oflfspnng and that,
consequently, all people must attnbute cultural import to this fact But, as
I indicated in the preceding section, closer scrutiny of the functional and
meanmgful entailments of "motherhood" does not sustain these assump-
tions (18. 40, 115, 116) We should no more infer that mothers and their
dependent children are the irreducible core of the family, because every-
where they have some kmd of socidUy recognized tie, than we should
conclude that the sibhng tie is the core o*" the family because it is likewise
mvanably recognized
As the final candidate for a umversal definition of the family. Bender's
(8) charactenzation of the family as a strictly kinship phenomenon poses a
slightly difierent set of problems As people everywhere recogmze "kinship"
relationships. Bender might seem to have hit upon the sole lnvanant aspect
of the family Yet if we define the family in what has been disclosed to be
stnctly genealogical terms, how then do we recognize us boundanes'' If we
completely divorce the family from any functional considerations, there is
but one way to decide who are the members of a family in a panicular case
that IS hy asking the natives to identify the tulturally meaningful ''kinship"
198 YANAGISAKO
the legitumzation of children, nghts m children and exchange of sexual and other
nghts are nearly everywhere associated with unions between men and women It follows
from the umversahty of kinship and the near umversality of mamage that, by definition,
family relationships are nearly umversal (9, p 238)
groups are organized throughout the whole range of human societies" (60,
pp 118, 124) This minimization of dom^tic vanability spnngs forth from
Goody's assumptions about the way m which the physiological and psycho-
logical concomitants of childbeanng, childreanng, and food preparation
structure the activities of domestic units (139) The reluctance to recognize
that in different societies widely varying and shifting assemblages of people
participate in these activities bespeaks of an unstated absorption with the
biological requirements of sexual reproduction
To commence a search for "the family." whether general or panicular,
by seeking out genealogical relationships is to begin with the assumption
that reproduction is the pnmary function of the family For if our investiga-
tion begins with the identification of affinal and filial links and then proceeds
to discover the groupings formed out of these bonds, parenthood and mar-
nage and, inevitably, reproduction must by definition emerge as the irreduc-
ible core of the family As Schneider (126) rightly argues, the use of the
genealogical gnd in kinship studies commits us to the position that the
biological facts of reproduction are what kmship is all about
The conviction that the reproduction of society's members is the essential
function of the family reveals that we have not progressed as far past a
Mahnowskian conception of the family as we would like to claim We may
have recognized the error in Mahnowski's reduction of all kinship institu-
tions to extensions of relationships within the elementary family, but we
continue to accept many of his notions about the nature of the family itself
Our placement of the family withm the domestic domain with its moral and
affective constraints, our fixation on genealogical definitions of the family,
and, underlying all, our emphasis on reproduction as the core of the family's
activities all betray our Mahnowskian hentage
The belief that the facts of procreation and the intense emotional bonds
that grow out of it generate an lnvanant core to the family is what sustams
our search for universals But the units we label as families are undeniably
about more than procreation and socialization They are as much about
production, exchange, power, inequality, and status When we fully ac-
knowledge that the family is as much an mtegral pan of the political and
economic structures of society as it is a reproductive unit we will finally free
ourselves from an unwarranted preoccupation with its procreative functions
and all the consequent notions embodied within such a stance There is
nothing wrong, of course, with a functional analysis of the family, for if the
family is a salient unit m any society it must have attached to it some
functions—whether these are symbolic functions, activity functions, or
both What is wrong is to decide a pnon that the diverse array of social units
we call families fulfill the same set of functions or that their pnmary
function IS always the same If we are to cast aside this premise and instead
200 YANAGISAKO
seek out the functions of the family in each society, we must at the same
time abandon our search for the irreducible core of the family and its
universal defimtion Our usage of the terms "family" and "household" will
then reflect an awareness that they are, like "nuirruige" and "kinship,"
merely "odd-job" words, which are useful m descnptive statements but
unproductive as tools for analysis and companson (98, p 44) The dilemmas
we encounter in cross-cultural compansons of the family and household
stem not from our want of unambiguous, formal definitions of these units,
but from the conviction that we can construct a precise, reduced definition
for what are inherently complex, multifunctional institutions imbued with
a diverse array of cultural pnnciples and meamngs Indeed, the only thmg
that has thus far proved to be unvarying in our search for the umversal
family is our willingness to reduce this diversity to the flatness of a genealog-
ical gnd
ACKNOWLEDOMENTS
Literature Cited
1 Ahem, E M 1973 The Cult of the Cognatic Social Structure, ed G N
Dead m a Chinese Village Stanford A{^>ell, pp 66-86 Washington DC
Stanford Umv Press 280 pp Am Anttm>pol Assoc 160 pp
2 Alexander, J 1977 The role of the male 7 Auwers, L 1978 Fathers, sons, and
in the tmddle-ctass Jamaican family / wealth m colonial Wmdsor, Connecti-
Comp. Fam. Stud 8 369-89 cut / Fam. HisL 3 136-49
3 Alexander, J 1978 The cultural do- 8 Bender, D R 1967 A refinement of the
mam of mamMc Am. EthnoL 5 5-14 concept of household families, co-resi-
4 Anderson,M 1971 Family Structure tn dence, and dcmiestic functions Am. An-
Nmeteenth Century Lancashtre Cam- thropoL 69 493-504
bndge, England Cambndge Umv 9 Bender, D R 1971 De facto families
Press 230 pp and de jure households m Ondo Am.
5 Anderson, M 1972 Household stnic- AnthropoL 73 223-41
ture and the mdustnal revolution mid- 10 Berkner, L K 1973 The stem family
nmeteenth-century Preston m compara- and the developmental cycle of the
tive perspective In Household and peasant household an 18 coitury Aus-
Family tn Past lime, ed P Laslett, R tnan example. In The American Family
Wall, m) 215-36 London Cambndge in Social-Histortcal Perspective, ed M
Umv Press, 623 pp Gordon, pp 34-58 New York St Mar-
6 AppeU,G N 1976 The Rui^jis social tm's 428 pp
structure in a cognatic society and its 11 Berkner, L K 1973 Recent research
ntual symbohzation In 71ie pieties of on the lustory of the fomily in Western
Borneo Explorations tn the Theory of Europe / Marriage Fam 35 395-405
FAMILY AND HOUSEHOLD 201
12 Berkner, L K. 1975 The use and mis- nous Families Evanston Northwestem
use of census data for the histoncal Umv Press 380 pp
analysis of family structure / Interdis- 28 Cohen, M L 1970 Developmental pro-
ctp. His. 5 721--38 cess in the Chinese domestic group In
13 Berkner, L K 1976 Infaentance, land Family and Kinship in Chinese Society,
tenure, and peasant family structure a ed M Freedman, pp 21-36 Stanford
German regional companson In Fam- Stanford Uiuv Press 269 pp
ily and Inheritance. Rural Society in 29 Cohen, M L 1976 House United,
Western Europe, 1200-1800, ed J R House Divided the Chinese Family tn
Goody, J Thtrsk, E P Thompson, pp Taiwan New York Columbia Univ
71-95 Cambndge, England Cam- Press 267 pp
bndge Univ Press 421 pp 30 Colher, G A 1975 Fields of the Tzotztl
14 Berkner, L K , Shaffer. J W 1978 The The Ecological Bases of Tradition in
joint family in the Nivemais J Fam Highland Chiapas Austin Univ Texas
Hist 3 150-62 Press 255 pp
15 Bloch, M 1971 Placing the Dead 3! Colher, J F 1974 Women m politics
Tombs, Ancestral Villages and Kinship In Woman, Culture and Society, ed M
Organization tn Madagascar London Z Rosaldo, L Lamphere, pp 89-96
Seminar 241 pp Stanford Stanford Univ Press 352 pp
16 Bloch, M 1975 Property and the end of 32 Cram, J B 1976 Ngenifan ntual pro-
affinity In Marxist Analyses and Social cess m a Bomean nee harvest In Stud-
Anthropology, ed M Bloch, pp 203 - ies in Borneo Societies, ed G N Appell,
28 New York Wiley 240 pp pp 51-63 Northern Illinois Umv Cen-
17 Bohannan, P 1963 Social An- ter for Southeast Asian Studies Spec
thropology New York Holt. Rinehart Rep 12 DeKalb Center for Southeast
& Winston 421 pp Asian Studies, Northem Illinois Umv
18 Boon, J A 1974 Anthropology and 157 pp
narmies Man 9 13''-4O 33 Davis. J H 1973 Land and Family in
19 Brandes, S 1975 Migration, Kinship Pisticci London Athlone 200 pp
and Community Tradition and Transi- 34 Davis, N Z 1977 Ghosts, km and
tion in a Spanish Village New York progeny some features of family life in
Academic 220 pp early modem France Daedalus 106
20 Brown, K 1968 The content of dozoku 87-i14
relationships in Japan Ethnology 35 deMause, L 1974 The evoluaon of
1 113-38 childhood In The History of Childhood,
21 Buchler. I R , Selby, H A 1968 Km ed L deMause, pp 1-73 New York
ship and Social Organization New Psychohistory Press 450 pp
York Macmillan 366 pp 36 Deiuch, B S 1974 Sex and power in
22 Bulmer, R N H 1960 Leadership and the Balkans See Ref 31, pp 243-62
social structure among the Kyaka people 37 Donham, D L 1977 A field theory of
of the western highlands district of New household production. Presented at
Guinea. PhD thesis Australian Na- A.nn Meet Am Anthropol Assoc,
tional Univ, Canberra, Aust 76th, Houston
23 Burch, T K 1972 Some demographii. 38 Donham, D L 1978 Production in a
determinants of average household size Malle community, Southwestem Ethi-
an analytic approach See Ref 5 pp opia. 1974-75 PhD thesis Stanford
91-102 Umv, Stanford, Calif 309 pp
24 Carlos, M L , SeUers, L 1972 FamUy, 39 Doqahn, V R 1977 Temne household
kinship structure, and modernization in size and composition rural changes
Latin Amenca Lai Am Res. Rev over time and rural-urban differences
1 95-124 Ethnology 16 105-28
25 Chayanov, A V 1966 The Theory of 40 Dnimmond, L 1978 The transatlantic
Peasant Economy ed D Thomer. B nanny notes on a comparative sermot-
Kerblay, R E F Smith Homewood lcs of the family m Enghsh-speaking so-
111 Irwm 317 pp cieties Am Ethnol 5 30-43
26 Chock, P P 1974 Time, nature and 41 Evans-Pntchard, E E 1940 The Nuer
spint a symbolic analysis of Greek- Oxford Clarendon 271 pp
Amencan spintual kmship 4m Eth- 42 Evans-Pntchard, E E 1951 Kinship
nol 133-46 and Marriage among the Nuer Oxford
27 Ciignel, R 1970 Many fVtves, Many Clarendon 183 pp
Powers 4uthority and Power m Po-'vg}-- i3 Fallers, L A , Levy, M J Jr 1959 The
202 YANAGISAKO
146 Stone, L 1975 The nse of the nuclear 156 Wheaton. R 1975 Farmly and kinship
family in early modem England See m Westem Europe the problem of the
Ref 118, pp 13-58 joint family household / Interdiscip
147 Stone, L 1977 The Family Sex, and Hist 5 601-28
Mamage in England 1500-1800 Lon- 157 Witherspoon, G 1975 Navaho Kinship
don Weidenfeld & Nicolson 800 pp and Mamage Chicago Univ Chicago
148 Strathem, A 1973 Kinship, descent Press 137 pp
and locahty some New Guinea exam- 158 Wolf, A P 1976 Chinese domestic or-
ples See Ref 138, pp 21-34 ganization the view from Taiwan. Pre-
149 Strathem, M 1972 W^omen in Between sented at Conf Anthrop in Taiwan,
Female Roles in a Male World Mount Portsmouth. New Hampshire
Hagen. New Guinea London Seminar 159 Wolf, E R 1966 Peasants. Englewood
372 pp Cliffs Prentice-Hall 116 pp
150 Talmon-Garber, Y 1970 Social change 160 Wolf, M 1972 Women and the Family
and kinship ties In Families in East and m Rural Taiwan. Stanford Stanford
West Socialization Process and Kinship Univ Press 229 pp
Ties ed R Hill, R Konig, pp 504-24
The Hague Mouton 630 pp 161 Wngley, E A 1966 Family reconstitu-
15! Tambiah, S J 1973 Dowr\'and bnde tion In An Introduction to English His-
wealth and the property nghts of torical Demography, ed E A Wngley,
women in South Asia See Ref 61. pp pp 96-^159 London Weidenfeld &
59-169 Nicoison 283 pp
152 Terray, E 1972 Marxism and Primi- 162 Wngley, E A 1969 Population and
tive" Societies. Transl M Ktopper History New York McGraw-Hill
New York Monthly Rev Press 186pp 256 pp
t53 Thirsk, J 1976 The European debate 163 Yanagisako, S J 3975 Two processes
on customs of inhentance. 1500-1700 nf change m Japanese-Amencan kin-
See Ref 13, pp 177-91 ship J Anthropol Res. 31 196-224
154 Thomas, K 1977 The changing famih 164 Yanagisako, S J 1977 Women-cen-
Times Lit SuppL 3943 1226-27 tered kin networks in urban bilateral
155 Weiner A B 1976 Women of Value kinship Am Ethnol 4 207-26
Men of Renown New Perspectives in 165 Yanagisako, S J 1978 Vanance in
Trobriand Exchange A.ustin l^niv Amencan kinship implications for cul-
Texas Press 299 pp tural analvsis 4m. Ethnol 5 15-29