With the birth of modern sociology in Western Europe in the latter half of the 19th century, social change was seen as evolutionary. Extending the theories of physical evolution, scientists saw social systems (including family systems) evolving towards higher forms with the stronger surviving the weaker (Spencer 1972). They saw Western Europe as the pinnacle of social development, they were surprised at the variety of social systems that existed in other cultures, and they concluded that the low level of economic development of these societies was related to their low level of social development.

PERCEPTIONS OF FAMILY SYSTEMS

It also seemed to 19th century scientists that the more primitive the society, the more extended were its family systems; or the more developed it was, the more the family system followed the nuclear pattern. The broad conclusion was drawn and held that the extended family had been a victim of the industrial revolution (Ogburn and Tibbits 1963).

These arguments were refined by the structural-functional sociologists of the 1950s. These writers referred to the process of differentiation (societal units becoming more specialised) in modernising societies - that is, as a society modernises, its units become increasingly more specialised. They saw the modern, nuclear family as better structured to accommodate this process of differentiation. The family they were referring to consisted of father, mother and their children with the mother not being in paid employment. Being smaller, this family form was seen as better equipped than other family forms to operate in the modern economy which revolves around individual achievement and social and geographic mobility (Parsons and Bales 1955). It was predicted that as all world cultures moved towards industrialisation, their family systems would approach some variant of the nuclear type (Goode 1963).

Thus, up to the 1960s, progression from extended to nuclear families via the industrial revolution was the prevailing perception of the history of Western family systems. At this time, however, careful reconstructions of past populations in England and other countries in western and northern Europe revealed the predominance of the nuclear family household well before the onset of the industrial revolution (Laslett and Wall 1972). This has been confirmed frequently in subsequent studies, and, indeed, it is now postulated by this school of thought that the nuclear family household in England has its origins as far back as the Christian revolution in the 4th century (Goldthorpe 1987).

It was from about that time that the Church introduced restrictions upon marriage and inheritance which had the effect of increasing its own wealth while weakening links in the wider kin network. It promoted a change in land tenure from 'folkland' or land held in customary tenure by a kin group to 'bookland' or land held under a written title deed and subject to the written will of the land holder. Strategies of kin groups to retain their lands and hence their power were severely limited by the restrictions imposed by the Church (Goody 1983). A similar battle for supremacy is being played out today in many developing countries between the emerging nation state and the powerful kin groups or lineages in the society. In many of these instances, however, it is the kin grouping which prevails by gaining control over the functioning of the state.

If industrialisation had any impact on family structure it was to increase the extent of extended family households among working people in towns and cities where there was a housing shortage (Anderson 1971). By the late 19th century, however, with the extension of credit to working class families and the growth of worker's cottages, these poor, urban extended families declined in significance (Anderson 1980). The evidence shows clearly that before and after the industrial revolution young couples in England were expected to set up their own separate households and, indeed, that the aged were also expected very largely to live separately from their children (Laslett 1989).

With this historical background and as the predominant cultural forms in Australia have their origins in English systems, we could expect to find that Anglo-Australian family households have also always been nuclear. Again, Australian studies tend to confirm that 19th century households were nuclear (McDonald and Quiggin 1985). Finally, the almost total absence of extended family households among Australians born in Australia is confirmed in recent studies (McDonald 1991).

THE FAMILY BEYOND THE HOUSEHOLD

The historical research discussed above very carefully refers to families as households or domestic groups - that is, people who live together and share the same kitchen. The emphasis in this definition of family is upon the demography of domestic groups. Censuses and reconstructions of past populations conventionally use the household as the basic unit. Australian censuses define the household as a group of persons living and eating together as a domestic unit. Censuses allow for multiple family households and show that they are very uncommon in the Australian context; however, they do not allow us to examine the incidence of multiple household families.

The concept of the family beyond the household is outside the scope of censuses. Hence the household becomes the dominant group unit in statistical reports. As the domestic group or household is the major purchasing unit in the society, the image of family generally conveyed through marketing is also that of the household family. Thus, based on what is easily observed and measured, there is a strong tendency to equate family with household and to ignore the family beyond the household.

IDEALISED FAMILY MORALITY

Rather than considering the demography of domestic groups as the criterion for determination of the prevailing family type in the society, we could instead focus attention upon the family type which is pre- eminent in the idealised morality of the society.

With this leap, there seems to be a great deal of evidence that the nuclear family has become pre-eminent in our idealised morality in the past 300-400 years. That is, in the world of ideas, the nuclear family is indeed a creation of the age of industrialisation, urbanisation and modernisation. But the creation probably emanates more from the changing history of ideas than from structural changes in the society. In the past 300-400 years, there has been increasing emphasis on the autonomy of the individual, on romantic love, on the privacy of the couple, on affectionate relationships between parents and their children, on the protection by the state of the privacy of the individual or small group, on legal changes which reduce the power of extended family over the nuclear family (Stone 1977). The emergence to pre-eminence of the nuclear family as the idealised family morality can be seen as part of this general progression towards the small and the private. The idealising of the nuclear family probably reached its zenith in the 1950s when current psychology tended to label anyone who did not aspire to this ideal as deviant (Ehrenreich 1983).

From the late 1960s onwards, however, in this progression of ideas, the idea of the nuclear family has itself come under pressure as the rights of individual family members have come to take precedence over the rights of individual families (McDonald 1988). In this era, we have seen a 'war over the family' waged on one side by those who seek to protect the idealised 'breadwinner' model of the nuclear family against the trend to uphold the rights of individual family members, and on the other side by those who would prefer to see the concept of the nuclear family eliminated altogether.

As time passes, this 'war over the family' can increasingly be seen as an ideological conflict conducted almost in isolation from any reality in the society itself. People continue to want and value family life, and 80 per cent of all children live with both their natural parents. But the 'breadwinner' model is disappearing and the shell of privacy of the nuclear family is cracking open so that the rights of individuals within families can be protected.

The argument so far

• The history of the demography of domestic groups indicates clearly that it is a myth that the nuclear family replaced the extended family as the common form of household sometime in the past 200-300 years. The northern and western European nuclear family household has been shown to extend as far back as we can reliably measure.
• On the other hand, there are strong arguments that in the realm of idealised morality, the emergence to preeminence of the nuclear family in the past 200-300 years is a reality.
• Families, however, can be defined in ways other than households and idealised morality. When we examine these other dimensions, the nuclear family as described in the appendix is called into question.

THE FAMILY AS A UNIT OF COOPERATION AND EXCHANGE

Beyond demography and ideology, family can be considered as the central core in the support networks of individuals. Using this definition of family, evidence continues to emerge that the extent of contact, cooperation and exchange between family members living in different households is very great indeed. The lack of recognition for these flows of exchange is the basis of the conventional wisdom that the extended family is insignificant in the Australian context.

When we extend our definition of family to family relationships across household boundaries, the definition of 'family' becomes more complex. In societies where roles and obligations between kin are rigidly prescribed, it is possible to provide a structural definition of the extended family. In the Middle East, for example, extended families are patrilineal, patriarchal and endogamous (marriage occurs within the extended family). Upon marriage a woman becomes a member of her husband's extended family, but because of endogamy, is still connected to her own family of origin. In South Asia, the arrangements are similar except that marriages are exogamous (outside the extended family) and so the woman upon marriage loses connection with her family of origin and owes all obligation to her husband's family.

In Western societies like Australia, roles and obligations are not rigidly prescribed in respect of kin living in different households. The reality, therefore, is that people define their own families. They do this according to four main dimensions: the relationship to each person; the purpose or activity concerned; particular circumstances applying in their case; and their perceptions of the nature of obligations.

THE RELATIONSHIP

Relationship refers to the particular paired relationship between two persons, such as mother-daughter, brother-sister, cousin-cousin. The expectation would be that the closer the family relationship, the more likely that the family member would be included in the definition of family. Subject to the purpose, this means that, in most circumstances, the defining out of a parent or a child is a deliberate act of exclusion. As the relationship distance increases, defining in or out becomes more a decision about inclusion as distinct from exclusion. For example, whether or not a cousin is part of one's family would tend to be an 'including' decision rather than an 'excluding' decision.

PURPOSE OR ACTIVITY

The family definition will vary according to the purpose or activity for which 'family' needs to be defined. For example, the kin who attend ceremonial events such as marriages and funerals often will be more widely extended than the kin that are asked to provide personal care during a major illness or regular child care. Yet another grouping will gather together at Christmas time. This will differ from those family members from whom we would ask financial assistance in a crisis. Yet another group would be acceptable if we needed emotional support. Some others we keep in close contact to exchange news. We construct different families for different purposes.

CIRCUMSTANCES APPLYING

The family definition will also vary according to the particular circumstances applying in each case. Most obviously, deceased relatives cannot occupy the place that they might have occupied. Other circumstances which may affect the inclusion or exclusion of a relative within the definition of 'family' include geographic distance, availability of other kin, the financial circumstances of both the relative and the person defining family, alternative sources of particular needs particularly provision of support by the state or community organisations, and so on. For example, the husband's mother may provide assistance with child care because the wife's mother is working or lives a long way away. Or the wife's parents may be asked for financial assistance rather than the husband's parents because they are in a better financial position.

PERCEIVED OBLIGATIONS

Finally, the self-definition of family will depend on perceived obligations of one person to another. These obligations are a product of somewhat vaguely defined societal expectations, the history of the relationship between the individuals concerned, and the perceived capacity of each to receive or provide the particular 'service' for which family is being defined. Sometimes, these obligations are specified in the law, such as the responsibility of parents to provide support to their non-adult children or the responsibility of one spouse to support the other. In some countries (such as China and Singapore) adult children are required to support their parents.

OBLIGATIONS DETERMINED BY THE STATE

The Family Law Act requires all parents to support their children aged less than 18 years, and all Australian states have child protection legislation which provides for children to be taken into state care if these obligations are not being met. The obligation of support continues beyond a marriage breakdown and to parents who are not co-resident with the child. Recent developments in the area of child support indicate an increased willingness on the part of the state to enforce these legal obligations.

Obligations to spouses, although stated with equal force in the Family Law Act, are not enforced or investigated to the same extent as obligations to children. While it is possible in theory to take your co-resident spouse to the Family Court in order that he or she provides you with financial support, and it has been done, it is difficult to imagine this approach being used very often in a continuing marriage.

In other contexts, particularly social security, state definitions of obligations are inconsistent and ambivalent. For Job Search Allowance, there is a presumption in policy that a child is dependent on parents until age 21. For eligibility for independent Austudy the presumption is that the parents have financial responsibilty for the child to age 25, irrespective of whether the child is or is not living with the parents. Age 25 is also used to define family coverage for private medical insurance. At the other end, eligibility for the Sole Parent Pension cuts out when the youngest child reaches age 16. Eligibilty for the Sole Parent Tax Rebate, however, continues until the child reaches age 18 or to the 25th birthday where the child is a full-time student. The Carer's Pension can be obtained by a homosexual person caring for his or her partner, but in the payments of pensions and benefits, homosexual couples are treated as separate individuals. There is also the anomalous situation of unsupported 15-year-olds who, having been disowned by their parents, fall between the crack of federal and state responsibility. It would not be too cynical to say that this hodge-podge of arrangements has much more to do with government fiscal considerations than it does with perceived family support obligations.

EXPECTATION, ALTRUISM OR RECIPROCITY

Most obligations or responsibilities in the family sphere are not determined by the state but by vaguely defined societal or personal expectations, notions of reciprocity, the history of the relationship between the two individuals and, debateably, the level of altruism of the provider (d'Abbs 1991).

In an innovative approach to the investigation of perceived obligations in the United Kingdom, Finch (1989) has made use of 'vignettes', short stories of the past relationship between two people. Having heard the story, the survey respondent is then asked whether one of the two people should or should not be expected to provide a particular form of help to the other. The stories are varied to include instances where the person now seeking help had either provided or refused help to the potential service provider in the past. By this means, the dimensions of reciprocity, altruism and morality can be explored.

Finch concludes: 'Membership of the kin group into which you were born is automatic and irrevocable: you do not join and there is a sense in which you cannot leave. The lifelong nature of relationships with one's family of origin mark them as distinctive and different from all others ... [however] ... Commitments to assist one's kin are not automatically 'there', ready-made for the fully-fledged adult to take on board. They develop and change over time; they get reaffirmed through reciprocal assistance. [The sense of obligation] derives from commitments built up between real people over many years, not from an abstract set of moral values.' (p.241)

Another approach to evaluation of perceptions is that used in the Institute's Australian Living Standards Study. In this study, the respondent is asked how likely he or she would be to ask specified people (mainly relatives) for support if they needed it. The questions cover financial, practical and emotional support. Responses to these questions reflect the combined effects of expectations and specific circumstances affecting the likelihood that the specified person could or would provide such support. Results of this study for the outer Melbourne area of Berwick are reported in this issue of Family Matters (Millward 1992). Data relating to child care in Berwick also show that, when parents are working and require child care from outside the household, the principal sources of such care are the grandparents of the child followed up by other relatives of the family.

Like this example of child care, most Australian research on the functioning of extended families focuses on actual behaviour rather than upon how people might be expected to behave in hypothetical situations. The major studies in Australia are the Families Surveys conducted in 1982 and 1992 by the Australian Bureau of Statistics (described by Kilmartin in this issue of Family Matters), the Institute's study (d'Abbs 1991) of support networks in different Australian local contexts (reviewed by Wolcott in this issue of Family Matters), and the Institute's Australian Living Standards Study (some results of which are reported in this issue by Millward). McCaughey's (1987) A Bit of A Struggle provides case study material for Geelong. With regard to the aged, extensive material is provided in the Australian National University's Ageing and Families project (Kendig 1986).

All of these studies indicate that the extended family is the central core of the support networks of most Australians. Indeed, the importance of the extended family may well have increased in recent times because the need for support from outside the household has increased at the same time as governments have been curtailing support of families.

For example, the surge of mothers with young children into paid employment has led to an equally rapid rise in demand for child care. The principal providers of child care in Australia are grandparents and other relatives. In the previous generation, this support was not required to nearly the same extent because the proportion of mothers in the labour force was much smaller. Regular child care by grandparents or other relatives also implies that grandparents are living relatively close to their children (this is confirmed in the article by Millward elsewhere in this issue). It is often presumed that 'past' society was sedentary and that family members lived close to each other within the same town or village. However, the growth of large cities and the ubiquitous motor vehicle may have led in fact to families these days being closer together than they were in the past.

It has also been shown that immediately upon divorce, many people who leave the matrimonial home return to their parents' home for a period (McDonald 1986). The rapid increase in divorce in recent years has made this more common. There is also some evidence that young people in their twenties, now facing more economic difficulty than those of the immediate past, are very likely to have received major financial assistance from parents in the financing of housing (Glezer 1991).

In an assessment of family relationships in England over time Finch (1989) concludes: 'The idea of the golden age in which family relationships were stronger than they are today is clearly a myth, without foundation in historical evidence' (p.85).

CONCLUDING REMARKS

This issue of Family Matters is directed towards a re- evaluation of the importance in our society of family relationships beyond the boundary of the household. The idealised family morality which we carry with us is that of the isolated nuclear family. This idealised morality together with demographic definitions of families has led us to believe that we are a society which has spurned the extended family, that once the umbilical cord of co-residence is cut, relatives become just like anybody else in the community - people we try to avoid helping.

This image of ourselves is supported by the obvious difficulty of many persons in our community who are isolated from extended family, such as homeless young people and some aged persons. An emphasis on support networks as a defining characteristic of families, however, indicates that most of us have very close bonds with family members with whom we do not live, that the extended family is alive and well. Indeed, if it were not so, the number of unsupported people on our streets would be far greater than it is.

REFERENCES

• Anderson, M. (1971), Family Structure in Nineteenth century Lancashire, Cambridge University Press, Cambridge.
• Anderson, M. (1980), Approaches to the History of the Western Family, 1541-1871, Macmillan, London.
• d'Abbs, P. (1991), Who Helps: Support Networks and Social Policy in Australia, Australian Institute of Family Studies, Melbourne.
• Ehrenreich, B. (1983), The Hearts of Men: American Dreams and the Flight from Commitment, Anchor Press/Doubleday, New York.
• Finch, J. (1989), Family Obligations and Social Change, Polity Press, Cambridge.
• Glezer, H. (1991), 'Cycles of care: support and care between generations', Family Matters, No.30, December, pp.44-46.
• Goldthorpe, J. (1987), Family Life in Western Societies: A Historical Sociology of Family Relationships in Britain and North America, Cambridge University Press, Cambridge.
• Goode, W. (1963), World Revolution and Family Patterns, Free Press, New York.
• Goody, J. (1983), The Development of the Family and Marriage in Europe, Cambridge University Press, Cambridge.
• Kendig, H., ed. (1986), Ageing and Families: A Social Networks Perspective, Allen and Unwin, Sydney.
• Laslett P. and Wall, R. (eds) (1972), Household and Family in Past Time, Cambridge University Press, Cambridge.
• Laslett, P. (1989), A Fresh Map of Life: The Emergence of the Third Age, Weidenfeld and Nicholson, London.
• McCaughey, J. (1987), A Bit of a Struggle: Coping with Family Life in Australia, McPhee Gribble, Melbourne.
• McDonald, P. and Quiggin P. (1985), 'Lifecourse transitions in Victoria in the 1880s' in Grimshaw, P., McConville, C. and McEwen, E., Families in Colonial Australia, Allen and Unwin, Sydney.
• McDonald, P., ed. (1986), Settling Up: Property and Income Distribution on Divorce in Australia, Australian Institute of Family Studies, Prentice- Hall, Sydney.
• McDonald, P. (1988), 'Families in the future: the pursuit of personal autonomy', Family Matters, No. 22, December, pp.42-47.
• McDonald, P. (1991), 'Migrant family structure', in Funder, K. (ed), Images of Australian Families, Longman Cheshire, Melbourne.
• Millward, C. (1992), 'Keeping in Touch: Extended Family Networks', Family Matters, No. 32, August.
• Ogburn, W. and Tibbits, C. (1963), 'The family and its functions', in Freedman, R., Principles of Sociology, Holt, Rinehart and Winston, New York.
• Parsons, T. and Bales, R. (1955), The Family: Socialization and Interaction Process, Free Press, New York.
• Spencer, H. (1972), On Social Evolution: Selected Writings, University of Chicago Press, Chicago.
• Stone, L. (1977), The Family, Sex and Marriage in England 1500-1800, Wiedenfeld and Nicholson, London.

Appendix

THE FAMILY CONVENTIONAL WISDOM

In Australia, the nuclear form of the family has predominated throughout this century. As the dictionary of sociology tells us, the nuclear family is a small group composed of husband, wife and immature children which constitutes a unit apart from the rest of the community.

There may by variants of this family form when parents split up or someone dies, but certainly extended families are almost unknown in the predominant Australian culture.

Extended families, where relatives other than parents and children cooperate in various forms of exchange, are common among Australian Aboriginal people and Australians who come from Mediterranean cultures or from Asia, but are rarely found in Anglo-Australian culture.

As many of the old photos show, we used to have extended families in the past, but industrialisation, urbanisation and modernisation brought an end to all that.