Many keystrokes have been registered, and much ink spilt, in attempts to produce a universal, etic, one-size-fits-all definition of household. None exists. More realistically, anthropologists have recognized the particularities of the situations they study, and directed their efforts to analysing them. Most agree with Hammel (in Netting et al. 1984: 40-1) that households simply are ‘the next bigger thing on the social map after an individual’, and proceed then to answer the ‘who’, ‘what’ and ‘where’ questions that concern householdly things.
Household ideology and composition
Aside from single-person domestic units, households consist of groups of persons. What ties these groups together? Often, but not always, it is kinship and marriage, and household members can trace such links among themselves. Some households consist of families — groups constructed of spousal, parent-child, and sibling bonds — but families may also be distributed in two or more households, or one household may contain relatives beyond these primary kin (as well as non-kin members). Families are emic and ideological groupings that knit persons together on grids of marital, parental and filial relations. Households have their own cultural dimensions, and these need to be examined separately from those of families, even when household and family personnel largely overlap (Yanagisako 1979; Netting et al. 1984; Gray and Mearns 1989).
In Japan they do not, and a series of cultural concepts exists that sorts people into groupings different from the Western folk notions of family and household. Shotai is the set of persons who share a common budget, regardless of kinship connections; kazoku is a group that has a sense of belonging together, whether or not they co-reside (and might include an adult child not living with a parent); ie means both the house itself and its transgenerational line of occupants (ancestral and living), of which there can be only one married pair per generation; dozoku is a larger group that includes an ie and branch houses founded by sons other than the one in any generation who continues the ie line (but by no means all sons form such branch households); shinseki includes houses linked to an ie through out-marriage of its daughters (Kitaoji 1971; Yanagisako in Netting et al. 1984).
Which is family, which is household? Or is it better to approach Japanese social organization through Japanese categories? If we focus on ie, the salient unit in rural Japan, we learn that its table of organization and terms of address have their own logic, which does not coincide with that of biological families and kinship. Males (in English translation) may be ‘old man, retired’, ‘house-head’, ‘successor’, ‘younger son’ (who must leave to marry), or ‘grandchild’; women may be, ‘old woman, retired’, ‘housewife’, ‘bride/young wife’, ‘daughter’ (who also must leave to marry), or ‘grandchild’. Men may be recruited as ‘successor’ by either birth to their ‘house-head’ father, marriage to his daughter, or adoption; a family bloodline may die out, and the ie continue.
Such positional household systems may be culturally elaborated as in Japan, or less so, but still behaviourally significant (Carter in Netting et al. 1984). In Accra in West Africa, adult household residents may occupy the mutually exclusive roles of child in their household of rearing, employee, independent co-member, solitary adult, conjugal partner, single parent, or grandparent; gender-patterned, life-cycle household role sequences may also be identified. For many Ghanaian adults their household of rearing may not be a parent’s household; a large proportion are fostered in other households as children, just as many of the children resident at any point in time live with adults other than their own parents (Sanjek 1982; 1983). More broadly, adoption and fosterage occur widely beyond Japan and Ghana, and are important in understanding both the ideology and composition of households where these practices are of organizational significance.
One cross-cultural scheme that is sensitive and adaptable to such ethnographic variation categorizes five major household types (Hammel and Laslett 1974; Sanjek 1982). Solitaries are single-person households; subtypes consist of single, divorced, widowed or duolocally married persons. No family households have no spousal pair or parent—child members, but may be comprised of other relatives (siblings, cousins, grandparents and grandchildren), or only of non-related room-mates. Simple family households include both spousal couples with or without children, and male and female single-parent households; an important subtype in many societies are mother— child households in which the father resides elsewhere, sometimes with another adult woman. Extended family households are simple family cores that add other kin, but not other spousal couples or parent—child units; they may be extended laterally (with siblings of simple family core adults), or lineally, both up (to include perhaps a parent of a married pair) and down (adding a co-resident grandchild). Multiple family households contain two or more discrete simple families (e.g. a couple and two married sons, two divorced sisters or widowed cowives and their children, or a four-generation Japanese ie), and may be extended with other kin as well. All five types may also include live-in household workers, less satisfactorily labelled ‘servants’, a term with unneeded connotations; boarders, who pay to eat and often to sleep in a household, and lodgers, who pay only to sleep, may also be counted as members.
This framework permits an overview by ethnographic region of ideologically dominant, and often numerically prominent, household forms, without resorting to such local usages as ‘the Hindu joint family’ or ‘the zadruga\ In composition, these Indian and Balkan lineal—lateral multiple family households, each with much symbolic elaboration, resemble Chinese and patrilineal African counterparts. Lineal multiple family households include the Japanese ie and Western European and Irish stem families (Goody 1972; Netting et al. 1984; Gray and Mearns 1989). Simple family households, with the husband-wife or ‘nuclear family’ subtype normative, are characteristic of England, the USA, Latin America, the Caribbean and many hunter-gathering societies. In these cases, extended families often form to meet physical or economic distress; and among elites (which excludes the hunter-gatherers) multiple family households are not uncommon, especially at family seats and retreats.
In all household regimes, the phenomenon of "developmental cycles occurs, meaning that a household may assume different forms through time as members arrive and depart, and new households begin (Goody 1972); the normative form may occur only rarely, although strategies to achieve it or adjust to its passing may preoccupy many members of the society otherwise. Berreman (1975), for instance, shows neatly how the North Indian Pahari "polyandrous household ideal (one wife, multiple husbands) actually produces other classifiable forms as members age and die. Too many developmental cycle analyses, however, take only the household as the focal unit; more careful study exposes how domestic cycles may differ by gender (especially where conjugal separation is frequent), and that household management and socialization strategies probably anticipate this (Sanjek 1983).
For some anthropologists, to consider household ideology and composition before ‘domestic functions’ is to put the normative cart before the behavioural horse. They ask first what a person must do to be considered a member of a household. Ethnographers generally answer this question in three different ways: a person must sleep there, or eat there, or make some economic contribution, whether or not they sleep or eat there.
As to other activities there is less contention. Households are agreed to be crucibles of identity (Gray and Mearns 1989), primary locations in which life-cycle phases are enacted (childhood, family establishment, elderhood), rites of passage celebrated or planned, inheritance decided and ancestors (who may symbolically and even physically underlie the household) be venerated. But must the persons who perform these and other activities in householdly places also be householdly groups?
No. Goody (1972) distinguishes dwelling (sleeping) units, reproductive (eating) units, and economic units, pointing out that each may be organized independently, with personnel of one including individuals from two or more of the others. Sanjek (1982) documents how, in an Accra neighbourhood, the universal activities of production (gainful work), social reproduction (the day-to-day reprovisioning of labour power emphasized in feminist anthropology), consumption, sexual union, and socialization of children are organized within, across, and outside of household units (compare Vatuk in Gray and Mearns 1989 with Netting et al. 1984). Social relations among local residents and consociates elsewhere involved in these activities constitute a platform on which class alignments and political ideologies are constructed. Households in urban Ghana are not the building blocks of larger social forms (if they may be elsewhere; see Gray and Mearns 1989); social practice is.
Sanjek defines these Accra households, which social practice variously joins and cuts against, on the basis of where persons sleep and store possessions, a decision following a West African cultural logic also identified by other regional ethnographers. Vatuk (in Gray and Mearns 1989) presents similar data concerning social practice for a South Indian Muslim group, but defines households according to where people eat, a decision in accord with South Asian cultural logic. There are also anthropologists who define households neither by sleeping nor eating groups but by pooled economic contributions -remittances sent one way by urban migrants for instance, or foodstuffs sent the other way by rural members. Again, this decision follows local cultural logic (‘two locations, one household’). Thus, Goody’s three units may each appropriately be used for answering the ‘what’ question about households.
Structures, locations and activities again
Hunter-gatherers, shifting cultivators and tent-dwelling pastoralists may move their household groups as their economies require and their impermanent or portable dwellings permit. In more permanent built environments, like West African savannah settlements or Southwestern American pueblos, house owners may reassign, divide or add new rooms as household composition changes. But when space is a commodity that households rent, or laws regulate building use or redesign, household groups (that may not wish to) must sometimes divide into separate locations, even though they may continue to organize themselves as one economic unit or cooking group.
Other people’s households are also workplaces for millions of persons worldwide (in contrast to artisans, business proprietors, professionals or job contractors whose own home is also their workplace). Women and men who are hired by households to cook, clean, care for children or elders, and other tasks are found throughout the world (Sanjek and Colen 1990). Some live in, frequently with an ambiguous relationship to household members (‘She’s just one of the family’; ‘I hate not living in my own home’). Others live elsewhere in their own households, and see the household of their employment simply as a workplace.
Trajectories of change
Although some anthropologists argue convincingly that enlarging households is adaptive in circumstances of extreme poverty, the global trend is to smaller household size. Often this is accompanied by ideological strictures concerning household composition (nuclear ‘family values’) and canons of domesticity (Lofgren in Netting et al. 1984; Williams 1994). Beyond this, with increasing world population the result is more households — more consumption units, more renters and mortgage payers, more telephone and cable subscribers, more furniture buyers and credit-card holders. Moreover, the fastest growing household type in the USA and elsewhere is solitaries.
Anthropological research on household ideology, composition, activities and location also confronts new territory. The study of lesbian and gay households raises easily settled typological questions, but also more challenging organizational ones (Lewin 1993). Major social policy debates centre on households, including issues regarding growing numbers of elderly persons and criminalized young males who reside in institutional and congregate settings. The number of teenage mothers in the USA is actually declining, but the ‘problem’ is magnified both by their greater current likelihood of independent household formation, and by later and fewer births among the female majority (Williams 1994). Finally, anthropologists who document the lives of homeless persons confront dilemmas of how households survive without homes.