GLOBALIZATION, MIGRATION AND FAMILY DIVERSITY
Susan C. Ziehl
COBISS 1.01
1.	INTRODUCTION
Globalization has been the buzz-word in the social sciences since the 1990’s. As in the case of most academic (and political) issues, it has been the subject of debate and controversy. There are differences of interpretation as to what this concept means, and even when different scholars agree on the meaning, they disagree as to whether it represents something ‘really new’. But whereas globalization has been the issue of debate within the social sciences and sociology in particular, it has had very little impact on the field of family studies. There are a number of possible reasons for this. Firstly, there has been a tendency for globalization to be conceptualized primarily in economic and political terms. It has been seen as synonymous (or at least principally about) the economic and political integration of societies across the globe and one of the debates has been about whether or not this has actually occurred. The family is seen as something different from the polity and economy - a different institution which raises different empirical and theoretical concerns. It is not unimportant that the family has traditionally been seen as representing the ‘private realm’ while the economy and polity have been seen as constituting ‘the public realms’. This is, of course, an ideological distinction which (family) sociologists have, for decades now, been questioning. But this has not changed the fact that there has been very little dialogue between scholars who have participated in globalization debates and those interested in family studies.
It would be an oversimplification to suggest that globalization debates have been solely around economic and political issues. Notions of cultural imperialism, universal citizenship and the link between cultures and identity have also been part of the debate. But those who have taken this route have seldom made connections between that aspect of the globalization debate and family life. Or at least where that connection has been made, it has been mentioned rather than analysed in any degree of depth.
Another possible reason for the lack of dialogue between these areas of academic debate and enquiry, concerns the ‘type’ of scholars involved. The majority of those interested in broad global trends in economic and political affairs are male, while family sociology is still predominantly a female preserve (numerically that is). This
Dve domovini • Two Homelands 19 • 2004, 47-68
is not a hard and fast distinction. There are a number of male family sociologists and female economists. But on the whole, the gender divide between the various academic disciplines and sub-fields of Sociology remain. For instance, at the last ISA meeting in Montreal, about 70% of the contributors to the Family Research Committee were female, almost exactly the same percentage as the proportion of male contributors to the Economy and Society Research Committee. In the case of the Ad Hoc Committee on Globalization, the proportion of male contributors was close to 90%. It is therefore possible that this gender divide in academic specializations is another reason for the lack of dialogue between globalization and family scholars.
Again, there are of course exceptions such as Smith & Wallerstein’s (1992) work on households and world systems theory. But to date their work has not been integrated into the field of family sociology. (See also Therborn, 2001.)
Although I have identified some reasons for the lack of dialogue - this is somewhat surprising since the issues which have been raised in globalization debates, overlap to a large extent with debates which have been taking place in Family Sociology. The communality is a concern with the question of change (extent thereof or lack thereof) and the direction of change: towards uniformity or diversity.
2.	GLOBALIZATION DEBATES
Drawing on Held (1999) Giddens (2001) describes the globalization debate in terms of three positions: Hypergloblizers; Sceptics and Transformationalists. Hyperglobaliz-ers claim that we have indeed entered a new era (the ‘global age’) - driven by a global economic system (capitalism) into which all societies are integrated. Also associated with this view is the idea of global governance and a reduction in social inequality. Hyperglobalizers further claim that we are approximating a ‘global civilization’, that is, a common set of values and norms that govern the behaviour of individuals all over the globe (Giddens, 2001: 58-61).
On the other hand, the skeptics argue that the world economy is far less integrated today than was the case in the 19th century and that social inequality has been on the increase (at least when one compares the various regions of the world). It also affords the state greater power to decide how and if it wishes to be integrated into the world economy. In contrast to the hyperglobalisers, these theorists do not believe that we are moving towards greater cultural unifonnity but rather a ‘clash of civilizations’ (Giddens, 2001: 58-61).
The third position (transfonnationalists) draws on elements to these two. Like the hyperglobalizers, they believe that we have entered a new era marked by an unprecedented rise in the degree of global interconnectedness. They see modernism as the force behind this transformation and are more open-ended\ambivalent about the consequences of this trend. They see both greater integration and fragmentation as features of the globe’s future. This is the position adopted by Giddens who argues that the ‘hypers’
place too much emphasis on economic factors while the skeptics underestimate the extent to which the world order has changed. For him globalization is not a one-way process but is more complex and open ended (Giddens, 2001: 61).
Giddens is one of the few sociologists to draw a connection between globalization and the family. He writes:
“Globalization is fundamentally changing the nature of our everyday experiences /.../forcing a redefinition of intimate and personal aspects of our lives, such as the family, gender roles, sexuality, personal identity /.../ We are faced with a move towards a new individualism (before which) the weight of tradition and established values is retreating” and that “traditional frameworks of identity are dissolving and new patterns of identity emerging.” (Giddens, 2001: 61)
At first glance these passages seem to convey the hyperglobizers’ position. The choice of words (like forcing; retreating; dissolving) give the impression of globalization as a wave which is pushing us (all of us?) in a particular direction. But this is just an impression. Giddens differentiates his position from the ‘hypers’ by claiming that globalization is not a one-way process towards a particular end. What globalization is forcing us to do, claims Giddens, is to reconsider our identities in the light of greater knowledge of various cultures and the end product is a global citizen whose identity is formed by numerous cultural sources. To illustrate this idea he uses the example of “a black urban South African” who “today might continue to be strongly influenced by the traditions and cultural outlooks of his tribal roots at the same time as he adopts cosmopolitan styles and tastes - in dress, leisure pursuits, hobbies and so forth - that have been shaped by globalizing forces” (2001: 64). What emerges from this passage is an image of someone being presented with a series of cultural repertoires and then choosing from among them to form a type of hybrid identity.
This does seem plausible even if a little patronizing to those who (still) live in traditional society. ‘They’ are seen as slaves to their culture, whereas ‘we’ modem people actively choose who we want to be and what our identity is. We can make and remake our identities and beliefs at will. But let us accept Giddens’ portrayal of the ‘global citizen’ as one engaged in “the ongoing process of creating and re-creating our self-identities” (2001: 62). The question that then arises is what those choices are. In other words, what is the outcome of those choices? Does any particular pattern emerge? Are we all making the same choices (MacDonaldization) or different ones (greater cultural diversity)? Is it only a matter of time before the “black urban South African” also sheds his\her belief in the ancestors and thereby “the cultural outlook of his (her) tribal roots”? But more pertinent to this article is the question of how globalization is affecting family life.
3.	FAMILY CHANGE AND DIVERSITY
A concern with changing family patterns is by no means new. It goes back to writers such as Bachofen (1861); Maine (1861); Morgan (1877); McLennan (1886); Le Play (1871) and Engels (1902) - all of whom identified various stages which they believed the family has gone through historically. The idea of an evolution in family patterns was echoed in Talcott Parsons’ work - his central theoretical argument being that there is a fit between the nuclear family and a modem industrialized economy. This idea was taken further by William Goode (1963 & 1964) who argued in favour of a world-wide trend towards the nuclear family pattern as more and more societies industrialize (convergence thesis). However, in the 1960’s and 1970’s these ideas became increasingly unpopular as Laslett and the Cambridge group provided historical data in support of the view that in England at least, ‘the family’ has always been nuclear. So, if the modem family is nuclear and the pre-modern one was also nuclear, there can be no evolution or change of family patterns. But it was not long before other scholars took issue with Laslett and his colleagues claiming that ‘the family’ has indeed changed historically (Shorter, 1975; Stone, 1977, for example). According to these scholars, the modem family is significantly different from its pre-modem counterpart.
More recently, the notion of family diversity has emerged very strongly as a theme of family sociology. And once again this has been couched in terms of change: the most popular position being that there is an increasing trend away from the (conventional) nuclear family and towards ‘family diversity’ (divergence thesis). The question of whether or not and how family patterns have changed historically has therefore been a central concern of family sociology.
Against this background, I once again pose the question of why Family Sociologists have not engaged directly with the globalization debate. After all, that debate is also about change and the direction of change. Is the contemporary era sufficiently different from a previous one so that one can call the former ‘a global village’ and the latter by some other name (non-globalized)? Also, part of the globalization debate has been about the direction of change. Is the world becoming more standardized (Mac-Donaldization thesis) or more diversified (cultural diversity)?
Since roughly the 1990’s a ‘hegemonic’ position has emerged regarding the ‘diversity debate’. It is one which says that in the course of the twentieth century there has been a steady move away from the conventional nuclear family and towards other family forms. This process is seen as having progressed so far that there is today no standard model of the family or a majority family form. Divorce, increasing acceptance of ‘alternative lifestyles’ and growing ethnic diversity are seen as the major factors behind this development. This position is closely associated with the move towards redefining the family (away from the conventional nuclear family) and the substitution of terms like ‘families’ and ‘family diversity’ for ‘the family’. This is a process which I myself have participated in - by defining the family as a social institution which encompasses a variety of cultural ideas (beliefs) about family life and a variety of family
patterns (Ziehl, 1997). It is also a process from which I have learned a great deal - my main conclusion being that the notion of family diversity as an empirical reality has been greatly exaggerated in certain instances. Below, I illustrate this idea using data from Great Britain. I then expand the discussion to include the European Union and later South Africa. My central argument will be that the notion of family diversity has more applicability on a continental\European and global level, than a national one.
4.	THE ‘DEATH’ OF THE CONVENTIONAL NUCLEAR FAMILY?
The quote below aptly illustrates what has become the orthodoxy in family studies.
“Many Sociologists believe that we cannot speak about ‘the family’, as if there is one model of family life that is more or less universal. The dominance of the nuclear family was steadily eroded over the second half of the twentieth century /.../ Less than a quarter of households in Britain conform to the model of the traditional family. There are also pronounced differences in family patterns across ethnic minority groups /.../ For example, Asian households often contain more than one family with children, while black communities are characterized by a large number of lone-parent families. For these reasons it seems more appropriate to speak of ‘families’. Referring to ‘families’ emphasizes the diversity of family forms. While as a short-hand term we may often speak of‘the family’, it is vital to remember what a variety it covers.” (Gidddens, 2001: 174)
Giddens goes on to reject functionalist theories of the family because they “neglect variations in family forms that do not correspond to the model of the nuclear family. Families that did not conform to the white, suburban, middle-class ‘ideal’ were seen as deviant” (2001: 175).
Evidence of this claim is usually sought in empirical data which show a move away from nuclear family households such as that presented below:
Table 1:
Distribution of Family and Household Types in Britain: 1961-1998
Household Type	1961	1971	1981	1991	1998
One Person Household	11	18	22	27	28
Couples (no children)	26	27	26	28	28
Couple with dependent children	38	35	31	25	23
Couple with non-dependent children	10	8	8	8	7
Household Type	1961	1971	1981	1991	1998
(Nuclear Family Pattern)	(85%)	(88%)	(87%)	(88%)	(86%)
Lone Parents Dependent children	2	3	5	6	7
Lone Parents Non-dependent children	4	4	4	4	3
Multi-family households	3	1	1	1	1
Two or more unrelated adults	5	4	5	3	3
Dependent children include all children up to the age of 15 plus all those persons aged 16-24 who are economically inactive (mainly in education) and who are living with at least one of their parents (European Commission, 2001: 115).
Source: Guardian, 27 March 2000, p. 3, in: Giddens, 2001: 176.
It is from this table that Giddens derives the statement that today less than a quarter of households in Britain are nuclear families. But is this evidence of a decline in the nuclear family and if such evidence did exist, is the decline due to an increase in ‘alternative lifestyle’ choices or indeed greater ethnic diversity?
4.1	Evidence of the decline?
Focusing on nuclear family households alone is not, in itself, evidence of the decline in the nuclear family (model/system/pattem) over time. Even the claim that nuclear family households represents a minority of all households is not evidence that it does not ‘predominate’ (if that is meant in the statistical sense). As one can notice from the table above, even in the 1960’s only a minority of households were nuclear family households (38% if one only includes those with dependent children and 48% if non-dependent children are included as well).
What one needs to focus on, in determining any change in the (statistical) predominance of the nuclear family pattern, are the various other household structures that make up the ‘normal’ nuclear family domestic life cycle. These are couple households (before the birth of children and after they have left home) and the single person household (when one spouse has died and sometimes before marriage). When one adds together these three phases of the nuclear family domestic life cycle, one notices that 85% of households were in one of those phases in 1960 compared with 86% in 1998. This means that the proportion of households, which fall into one of the phases of the ‘normal’ nuclear family domestic life cycle, has either remained stable or indeed increased.
Looked at from this perspective then, there is hardly any evidence of a massive decline in the popularity of the conventional nuclear family model. This is not to say that there has been no change over the roughly 40 year period depicted in the table. If
one focuses on single parent families with dependent children one notices that they have increased from 2% of households in 1961 to 7% of households in 1998. One can, of course, present these statistics more dramatically by saying that single parent families are three times more common today than was the case in 1961 or have experienced a 300% increase etc. However, they still represent a small proportion of all households despite the dramatic rise in the divorce rate.
Returning to the argument that nuclear family households (couples with dependent children) have declined: What are the reasons for this? The unprecedented rise in the divorce rate is undoubtedly one of the reasons. It manifests itself in both the increase in single parent households and single person households. But the other reasons are predominantly demographic. Couples are marrying later today than was the case in the 1950’s. But the majority of the population still marries at some stage in their lives. This means that the number of couple households (before the birth of children) increases. The other major demographic trend, which has impacted on household patterns, is the increase in longevity. On average, people are living much longer today than was the case in the past. This increases the proportion of couple households (after children have left home) and single person households (when one spouse dies). Indeed, probably the most dramatic change in household composition patterns we have seen in the course of this century, has been the increase in single person households. And this is mainly because people are living longer and to a lesser extent because they are waiting longer before marrying.
4.2	How has increased ethnic diversity in Britain impacted on family patterns?
It is in this area that I believe the closest connection between globalization and family patterns can be made. As is probably well known, since the Second World War there has been massive immigration of people from the former colonies into Britain and Western Europe as a whole. In the case of Britain these migrants came mainly from Asia; the Caribbean and Africa. Elliot writes the following in this connection:
“Over the past forty years, ethnic divisions have assumed heightened salience in national and international politics /.../ These divisions involve cultural differences and allegiances, are in general associated with marked inequalities of power and wealth and tend to be at their sharpest wherever they coincide with racial and/or religious distinctions. At their center lie differences in gender and family structures. Gender and family patterns may reflect longstanding cultural traditions, are frequently governed by deeply held religious beliefs and are integral to a people’s identity. However, they are inevitably challenged, and become the source of intense anxieties, wherever ethnic groups share a common territory and must negotiate a shared way of life.” (1996: 40) (emphasis added)
Elliot discusses the impact of ethnic diversity on family patterns in Britian with respect to the two main immigrant groups: Britains of Afro-Caribbean descent and those of Asian descent. In both cases she draws on research which compares the ‘ideal typical’ working class family pattern in the Caribbean and rural South Asia in the 1950’s and 1960’s with the family patterns which characterize the working class section of these communities in Britain today.
While Elliot’s depiction of working class family life in the Caribbean in the 1950’s and 60’s is somewhat confusing, it seems to involve the following: relatively high rates of extended families, female-headed families and non-marriage. Migration to Britain appears to have involved an initial increase in the popularity of marriage followed by a decline. It also seems to have involved the ‘virtual disappearance’ of both the ‘grand-mother family’ and extended families more generally (1996:45). This assertion is based on research which shows that only 1% of children (from Afro-Caribbean descent) are raised by grand mothers today (Griffiths, 1983 in Elliot, 1996: 45). Data on marital status by ethnic group further shows that in the period 1986-8 the Afro-Caribbean population was less likely to be married than ‘White’ Britains (33% vs 51%) and more likely to be divorced (7% vs 4%) (Elliot, 1996: 45). Solo parenthood is also shown to be more common among Afro-Caribbeans than ‘native Britains’. According to Haskey, in 1987-9 about half of all families with dependent children in the Afro-Caribbean community were headed by a lone parent compared with only 15% in the case of ‘White’ Britains. In Elliot’s view, what these data show is both change and continuity in Afro-Caribbean family life as a result of migration. She summarises her views as follows:
“(The data) show that in Afro-Caribbean communities, extended family relationships are weaker than they were in the Caribbean, that the grandmother family is virtually non-existent and that there may be tendencies towards higher rates of marriage /..../ Nevertheless, Afro-Caribbean family patterns remain distinctive in terms of the institutional weakness of marriage, the presence of woman-headed families (and) the marginality of men ... “ (Elliot, 1996: 57)
She finds similar degrees of change and continuity in Asian family patterns. But in this case the nature of rural South Asian families of the fifties is more clearly spelt out. As far as household structure is concerned “the ideal typical rural South Asian family was characterised by the formation of three-generation patrilocal households (consisting of a man, his sons and grandsons, their wives and unmarried daughters) ...” (Ballard, 1982, in Elliot, 1996: 49).
“Ballard shows that the prototypical rural Asian family is constructed around an ideology of patrilineal cooperation. This ‘traditional Asian family’ is frequently cited as epitomizing the classical extended family believed to
be characteristic of subsistence societies. It stands in sharp contrast with the conventional Western conjugal family and in even sharper contrast with ‘liberated’ Western sexual and family values.” (1996: 49)
How has migration impacted on Asian family life? Elliot claims that all the available evidence indicates that migration to Britain resulted in the severe disruption of kinship ties but that these have subsequently been reconstituted:
“.. the salience of kinship bonds in Asian culture represented an important resource in the reworking of Asian life in Britain /.../ though preferences for nuclear-family households are emerging (Stopes-Roe and Cochrane, 1990) Asian households remain larger and are more likely to contain extended family members than either ‘white’ British or British Afro-Caribbean households (Haskey, 1989b; Ballard and Kalra,1994).”(Elliot, 1996: 51)
Elliot finds further evidence of the “continuing integrity of Asian family values” in the “low rates of unmarried cohabitation, divorce and solo parenthood among Asian peoples /.../ together with their low rate of inter-marriage with people of other ethnicities” (Elliot, 1996: 57).
It would seem then that British citizens of South Asian descent are adding more conservatism to British family life while at the same time creating more diversity by having a greater propensity to extended family living, than other Britains. On the other hand, Britains of Afro-Caribbean descent are adding to the array of more ‘liberal’ family structures by showing a greater propensity to ‘lone parent families’.
“In sum, it appears that the gender and family structures of Afro-Caribbean and Asian groups have changed in the context of the cultural, economic and political opportunities and constraints of life in Britain but in ways that are shaped by their own cultural logic. They remain distinctively different from Anglo-Saxon gender and family structures. The gender and family structures of other ethnic minority groups (including European minority groups) can also be expected to display their own cultural specificities. Afro-Caribbean and Asian patterns are thus indicative of the ethnic diversification of British society in the latter part of the twentieth century.” (Elliot, 1996: 58)
Elliot is undoubtedly correct when she asserts that British society has undergone a process of ethnic diversification. However, the question that arises is the extent to which this has impacted on family patterns in British society as a whole. If one looks back at Table 1, one notices that since the 1960’s the proportion of all households that are sole parent households has increased (from 2 to 7% or 6 to 10% if independent children are included) but that ‘multi-family households’ have declined to almost nothing (3% to 1%). The short answer to the question of the extent to which ethnic diversification
has translated into greater family diversity is therefore: ‘not much’ and the explanation lies a phrase contained in the quote from Elliot above ‘minority groups’. Immigrants from outside Europe and their descendants, today constitute only about 5% of the British population - a figure that is expected to stabilize at around 9% (Elliot, 1996). Their ‘distinctive’ family patterns are therefore unlikely to impact greatly on the general pattern of family life in that society. Moreover, it is not just the case that the vast majority of households in Britain fall within one of the phases of the ‘normal’ nuclear family domestic life cycle (Table 1) but also that the majority of the population find themselves in that situation as well. This is illustrated in Table 2.
Table 2:
Households and People in Great Britain 1981 and 1992
Type of Household	% of households	% of people	% of people
	1981	1981	1992
One Person	22	8	11
Married Couple	26	20	23
Nuclear Family (Dep. Children)	32	49	40
Nuclear Family (Indep. Children)	8	10	11
Nuclear Family Pattern	88%	87%	85%
Lone parent dependent children	4	5	10
Other	9	8	5
Source: Adapted from Social Trends 13 & 24, in: Haralambos and Holbom (1995: 355).
The argument being raised above is not that no change has occurred in the field of family patterns over the last 50 years. Rather, it is that the extent of the change is not as dramatic as family sociologists (and the media) have led us to believe. Despite a significant increase in the divorce rate; despite increasing acceptance of cohabitation as a prelude or alternative to marriage; despite increasing acceptance of non-marital childbearing and despite increasing ethnic diversity, the vast majority of households and people in Britain fall within the nuclear family pattern.
4.3	Making the facts fit the ideology: statistical vs moral norms
How have we got to this point where, what sociologists and the media have been telling us about family life, is at variance with the empirical data? I submit that it is because family sociologists, in particular, have failed to adequately distinguish between moral and statistical norms. But more than that, it is because family sociologists have
been trying to seek support for their view that non-conventional family structures ought to be accepted as legitimate or ‘normal’ (a moral stance), in empirical data showing how common or uncommon different household structures are (statistical frequencies). Note Giddens’ rejection of functionalist theories of the family because they imply that middle class families are ‘ideal’ and others are deviant (p. 5). Note also the phrase “the dominance of the nuclear family (has been) steadily eroded” without specification of what kind of dominance is at issue (Giddens, 2001: 175).
Put bluntly, the reasoning has been as follows: If the conventional nuclear family is no longer the statistical norm, then it should no longer be regarded as the moral norm and other family situations should enjoy legitimacy as well. The problem with this reasoning is that it is an attempt to deduce a moral position from an empirical claim about how frequently something occurs. Even if it were possible to show that the majority of a population does not follow the conventional nuclear family pattern, that is not evidence for the claim that it should lose its position as the moral norm or cease to be regarded as legitimate. Conversely, the fact that non-conventional family structures represent a small proportion of all households is not, in itself, proof for the claim that they should not be regarded as legitimate or accepted by the public at large. One cannot deduce a ‘should’ claim from an ‘is’ claim. The legitimacy or not of any particular family or household structure is a moral question that can only be answered on that level. It may well be the case that within family sociology circles at least (and in the media), the nuclear family no longer enjoys the position of moral norm and other family situations are regarded as legitimate. But the empirical evidence shows that it is still the statistical norm.
This raises the question of what we mean by (family) diversity. Is it sufficient for 50%; 20% or even 10% of the population to be ‘different’ before we talk of diversity? It would appear that if the 1990’s was the decade of deconstructing the concept of‘the family’ it is now time for us to start deconstructing the notion of ‘family diversity’. The argument raised above is that if 50% is the cut-off point (half conventional and half non-conventional) or even 20%, then there is not much family diversity in Britain today. The same applies to the United States. The argument that will be raised below is that the notion of family diversity has greater applicability on a regional scale (when we look at Europe as a whole) and globally.
5.	FAMILY DIVERSITY IN THE EUROPEAN CONTEXT
When Household and Family in Past Times was first published (Laslett & Wall, 1972), it was heralded as providing conclusive proof for the claim that in pre-industrial Europe the nuclear family pattern already predominated. This was however, a simplification and distortion of the research, since as Kertzer (1991) points out, the book does contain chapters on Serbia and France which show significant regional variation in family patterns in pre-industrial times - the main difference being the proportion of
‘extra-nuclear kin’ in households. Kertzer (1991) furthermore cites research conducted on communities in Southern Europe (Italy) and Eastern Europe (Hungary) which also show higher levels of extended family households in those areas than in the case of North-Western Europe and England, in particular. Laslett himself has responded to the ‘undue haste’ with which people have generalized from data pertaining to England to the rest of Europe by drawing a distinction between four “tendencies in domestic group organization in traditional Europe” (in Wall, Robin & Laslett, 1983: 256). He labels these North-West, Central, Mediterranean and East. He presents these as points on a continuum with the North-West areas having the lowest proportion of ‘resident kin’ and multigenerational households and Eastern Europe having the highest. This classification is very similar to Hanjal’s distinction between two pre-industrial marriage systems in Europe: The North-Western pattern or simple household system and the South-East or ‘joint family household system’ - the latter being marked by a younger age at marriage; patrilocal post-marriage residence and therefore higher incidence of complex family households than in the case of the North-West pattern (in Kertzer, 1991: 158). As indicated below, there is some evidence to suggest that these differences still persist today.
The European Commission has provided data on the distribution of the population between households in the European Union in 1988 and 1999. The data for 1999 are presented in Tables 3 & 4. Note that the unit of analysis in this instance is the individual rather than the household. The first column of Table 3 therefore shows that 12% of the EU population was living alone; 24% with a partner; 36% with a partner and dependent children. Table 3 further shows that 11% of the European population was living in households consisting of 3 or more adults with dependent children. I am taking these to be extended family households. This distribution of household structures is not the same for each country within the Union. Rather, distinct patterns emerge which can be linked to the distinction between the nuclear family pattern and the extended family pattern.
Elsewhere, I have argued that the barometer or test of the prevalence of one or other of these family patterns is not the prevalence of nuclear family households or the extended family households themselves (Ziehl, 2001). Rather, when looking for evidence of one of these family systems, one needs to focus on what is rare or unusual in that family system. In the case of the nuclear family system it is extended family households that are rare. In this system, it is unusual for a married couple to live with one or other of their parents or for a child to be raised in a household which also contains a grandparent. In the case of the extended family system, it is single person households and couple households that are unusual. In the extended family system it is rare for someone to live alone either before marriage or once a spouse has died or to live with a spouse only. This is because extended family systems are governed by the rule of patrilocality (rather than neolocality) - a man brings his wife into his parental home. The extended family system also includes nuclear family households since not all sons can be accommodated in this way.
In Table 3 only those countries with below average proportions of their population living in single person households are included. They have further been arranged from lowest to highest (proportions of the population living alone). The countries included in Table 3 are also those with above average levels of extended family living.
Table 3:
Countries with Above Average Levels of Extended Family Households
Type of Household	EU *	Spain	Portugal	Ireland	Greece	Italy	Luxem.	Total
Single Person	12	5	5	7	8	9	10	7
Couple Household	24	16	16	14	21	18	20	18
Nuclear Family Household	36	34	39	43	38	37	43	39
Single Parent Household	4	2	2	4	2	2	4	3
Extended Family Household	11	21	20	18	13	13	12	16
Three or more adults	14	22	18	14	18	21	12	18
Total	100%	100	100	100	100	100	100	100
* No data for Sweden. Source: European Commission, 2001: 115.
Table 3 shows that in these societies (which with only two exceptions, are located in Southern Europe), it is unusual for people to live alone. Living with a spouse only is also less common in these countries than in the European Union as a whole. Finally, in all these countries, the proportion of the population living in extended family households is well above the average for Europe. Indeed, in Spain, Portugal and Ireland they are almost twice as common than is the case in Europe as a whole. These differences are even starker if we compare these (mainly) Southern countries with those further north. Table 4 shows those countries which have above average levels of single person households and below average levels of extended family households.
As can be noted from Table 4, all of these countries are in the northern or central parts of Europe. Comparing these two regional blocks gives clear evidence of family diversity within Europe itself (Table 5).
Susan C. Ziehl Table 4:
Countries With above Average Levels of Single Person Households
Type of Household	EU*	Den.	Ger.	Fin	Neth.	U.K.	France	Belgium	Aust.	Total
Single Person	12	17	16	16	14	13	13	12	12	14
Couple Household	24	28	29	26	29	27	25	23	22	26
Nuclear Family Household	36	36	34	41	35	33	43	42	33	37
Single Parent Household	4	3	4	6	3	8	5	5	3	5
Extended Family Household	11	7	7	4	9	8	7	8	14	8
Three or more adults	14	8	10	6	9	12	8	11	15	10
Total	100	100	100	100	100	100	100	100	100	100
Source: European Commission, 2001: 115. * Data for Sweden not available.
Table 5:
Regional Variation in Family Patterns in European Union 1999
Type of Household	European	“Southern”	Northern &
	Union	Europe	Central Europe
Single Person	12	7	14
Couple Household	24	18	26
Nuclear Family Household	36	39	37
Single Parent Household	4	3	5
Extended Family Household	11	16	8
Three or more adults	14	18	10
Total	100	100	100
Nuclear Family Pattern	(72)	(64)	(77)
Extended Family Pattern	(47)	(55)	(45)
Source: European Commission, 2001: 115.
The classical extended family pattern is associated with only two household types: extended and nuclear family households. An extended family household usually consists
of: Ego, her spouse, children, son’s wife and grandchildren. This becomes a nuclear family when ego and her spouse die and stays such until a grandchild starts having children. When one adds together these two phases of the extended family pattern one notices that a minority of people in Northern and Central Europe (45%) but a majority of those in ‘Southern’ Europe (55%), fall into one of those phases. It is true that the majority of people in Southern Europe fall within the nuclear family domestic life cycle (64%) as well. But this hides the fact that living in extended family households is twice as common in the ‘South’ as compared to the North.
Spain and Denmark are at opposite poles of the nuclear vs extended family pattern divide. Extended family living is three times more common and single person living is three times less common in Spain than Denmark. Closer examination reveals that in Spain there are no people aged 30 or below who are living alone and therefore no single person households in that age category. This compares with 24% of the Danish population and 18% of the EU as a whole (Table 6). Furthermore, when one focuses on the 65+ age category one notices that 18% of the Spanish population in that age category are living in a household arrangement other than alone or with a spouse, compared to only 2% in the case of Denmark (Table 7). It is my understanding that these ‘other households’ are extended families. Among the elderly, living in an institution is also more than twice as common in Denmark than Spain (5% vs 2%) (Table 7).
Table 6:
Single Person Households by Age
	E.U.		Denmark		Spain	
	% People	% SPHH	% People	% SPHH	%People	%SPHH
All	12*	100*	17*	100*	5*	100*
-30	2	17	4	24	0	0
30-64	5	42	7	41	2**	40
65+	5	42	7	41	3	60
SPHH: Single Person Households
*May not add up to the total due to rounding.
** I have corrected this up from 1, as the data show that 1% of population is living alone and female and 1% living alone and male.
Source: European Commission, 2001: 115.
Susan C. Ziehl Table 7:
Elderly Population by Household Situation and Age Group, 2001
Population 65+ living:	E.U.	Denmark	Spain
Alone	32%	42%	22%
With Partner	54%	52%	58%
Other household	9%	2%	18%
Institution	4%	5%	2%
Total	100%	100%	100%
Source: European Commission, 2001: 116.
The argument raised previously is that globalization has had very little impact on family patterns in Great Britain. Longitudinal data for the European Union suggests that this is the case for Europe as a whole as well. Indeed, between 1988 and 1999 the proportion of the European population living in one of the phases of the nuclear family pattern has increased (from 69 to 72%) (Table 8). But my second argument is that while one finds little family diversity when focusing on one country or one continent, this is not true when we compare different countries and regions. The data show that while living alone and living with a spouse only, is not completely unheard-of in ‘Southern’ European countries, there is evidence to suggest that a different family pattern or domestic life cycle is being followed in the ‘South’ as compared to the northern parts of Europe. In this article it is not possible to delve into the reasons for those differences. The argument that will be raised below is that the notion of family diversity has even more applicability on a global scale.
Table 8:
Population in Households in European Union, 1988 & 1999
Household Type	1988	1999
Single Person	10	12
Couple Household	21	24
Nuclear Family Household Dependent Children	38	36
Single Parent Household	3	4
Extended Family Household	14	11
Three or more adults	14	14
Total	100	100
Nuclear Family Pattern	(69)	(72)
Extended Family Pattern	(52)	(47)
Source: European Commission, 2001: 115.
6.	TOWARDS A GLOBAL PERSPECTIVE ON FAMILY DIVERSITY
In the 1960’s, William Goode put forward the idea (with qualification) of a global movement towards uniformity in family patterns and claimed that the convergence was towards the nuclear family pattern (1963; 1964).
“In all parts of the world and for the first time in world history, all social systems are moving fast or slowly toward some form of the conjugal family system and also toward industrialization /.../ with industrialization the traditional family systems - usually extended or joint family systems /.../ are breaking down.” (Goode, 1964: 176)
I am not aware of any research that has been done on a global scale to test this hypothesis and therefore to determine whether Goode’s prediction has in fact come true. In the South African context numerous studies have been undertaken to determine whether Black South Africans are increasingly living in nuclear family households (see Ziehl, 2001). This research is based on the fact that African family systems are different from Western family systems - the first being based on the extended family and the second on the nuclear family model. The question that has been addressed is therefore whether or not Black South Africans are increasingly adopting the family patterns characteristic of Western societies. Some have argued that it is politically incorrect to ask this question (Russell, 1994). Indeed, in elaborating her views on this matter, Russell specifically refers to the process of globalization claiming that despite this process, the kinship systems of Black South Africans are likely to remain strong:
“(There is) compelling evidence /.../ of the persistence among black metropolitan households of distinctive patrilineal householding principles, which prove resilient and compatible with industrial urbanisation. It alerts us to the probability that as Africa becomes more firmly entwined in the world economy, distinctively African social patterns are likely to emerge and persist.
In South Africa they are likely to predominate and prevail.” (1994: 66)
Elsewhere, I have argued that the problem with this debate is that we have not had the data on which to base claims about changing family patterns in South African society as a whole (Ziehl, 2001). The studies referred to earlier have all been either small-scale or lacked a longitudinal dimension or both. In the 1990’s Steyn (1995) conducted a survey of household structures in South Africa using a representative sample. However, she only included the urban areas and her study was not longitudinal. I have analysed data from the 1996 South African census to determine the distribution of household types in South Africa as a whole. At the time of writing this process is incomplete since only the data for Black and White South Africans have been analysed. But since Black South Africans represent the vast majority of all South Africans (77%), I will be us-
ing data from that section of the population, as a proxy for South African society as a whole. Below, I address two questions: (1) Are South Africans increasingly following the nuclear family pattern and (2) Are there distinct differences between the family patterns of South Africans and those of Western societies such as Britain?
6.1	Convergence or Divergence
It is not possible to compare the 1996 Census with the previous one (1991) for two reasons. Although, the 1991 census questionnaire included a question on household structure, the responses were not coded and the data not analysed. Secondly, the 1991 census excluded the former (independent) homeland areas whereas the 1996 one did not. This means that the geographic base of the two censuses differs. Longitudinal data, to measure change in family patterns in South African society as a whole, is therefore not available. The next best option is to compare the distribution of household structures in urban and non-urban areas. This is provided in Table 9. Given that Black South Africans have traditionally followed the extended family pattern, one is immediately struck by the high levels of single person households in both urban and non-urban areas. This is accounted for by the fact that in 1996, domestic workers who live on the property of their employers were classified as heading their own households. More solid evidence of changing family patterns can be found in the ‘couple household’ category. In the past and still today, it is very unusual for Black South Africans to live with a spouse only. Only 7% of all Black households were couple households in 1996. However, the urban figure is significantly higher than the non-urban one. Urban dwellers are twice as likely to live with a spouse only, than rural dwellers. Here, then, there is some evidence of the breaking up of the extended family pattern and the adoption of the nuclear family pattern associated with urbanization. I am relying here on only one data-source, though, and South African census data (like that of other societies) are fraught with problems. So I offer this observation as a tentative null hypothesis while we await better research on family life in South Africa.
Table 9:
Household Structure
South African Census 1996
	Black		
Type of Household	Urban	Non-Urban	Total
Single Person	21%	13%	17%
Couple Household	10%	5%	7%
Nuclear Family Household	22%	18%	20%
	Black		
Single Parent Household	11%	18%	15%
Extended Family Households	23%	30%	27%
Three or more unrelated Adults	3%	2%	2%
Unspecified	11%	16%	13%
Missing value	0.07%	0.02%	
Total	100.00%	100.00%	100%
6.2	South Africa and Britain
The idea that South Africans may increasingly be adopting the nuclear family model in no way contradicts the view that on the whole South Africans do not follow the nuclear family pattern. As can be noted from the table below, only a minority of households (44%) fall within the nuclear family pattern. Moreover, a comparison with Britain shows that whereas in the latter case, the nuclear family pattern is clearly the statistical norm (accounting for 86% of all households), this is not the case in South Africa. Finally, whereas only 1% of households in Britain are extended, this applies to 27% of all South African households. In my view, then, the table below aptly illustrates the notion of family diversity on a regional or global level.
Table 10:
Distribution of Household Types: Comparison Britain and South Africa
Type of Household	Great Britain 1998	South Africa 1996
Single Person	28	17
Couple Household	28	7
Nuclear Family Household	30	20
Nuclear Family Pattern	(86%)	(44%)
Single Parent Household	10	15
Extended Family Household	1	27
Three or more adults\unrelated individuals	3	2
Not Specified	0	13
TOTAL	100	100
Source: Giddens, 2001: 176. Own calculations from data provided by Statistics South Africa.
7.	CONCLUSION
The purpose of this article has been to draw links between debates around globalization and those that have been taking place within the field of family studies. In both cases there has been dispute about whether the contemporary era is sufficiently different from the previous ones, or not. Among those who believe that we have in fact entered a new and different era, there is disagreement about whether that means that identities and behavioural patterns (including family patterns) have become more uniform (and/or Westernized) or more diversified.
The reigning orthodoxy in family studies is one which says that in England and Western Europe, the nuclear family pattern did once predominate but that it has lost that position due to increasing rates of divorce and increasing ethnic diversification. There are connections here with the globalization phenomenon since the post second world war migration from parts of Asia, Africa and the Caribbean is identified as one of the main reasons for the increased ethnic diversity in countries like Britain. Without disputing that some change has occurred (in the field of family patterns), I have critiqued the notion that it has been so great that the nuclear family has been knocked off its statistical perch - at least in the case of individual societies within North Western Europe. Seeing the nuclear family household as a phase in the domestic life cycle, which makes up the nuclear family pattern, makes this clear.
With regard to the divergence\convergence debate I have argued that on the societal level and in the case of Western European societies such as Britain, there is very little evidence of family diversity despite the migration flows of the post second world war era. I have further argued that on the continental level the division between North and South, which was identified by family historians for the pre-industrial era, seems to be persisting. In the ‘South’ family life is far more likely to involve a period of extended family living than is the case in the ‘North’. The notion of family diversity is therefore more applicable in this context. Finally, I have argued that when one compares a Western society such as Britain with an African society such as South Africa, one notices significant differences in family patterns. Family diversity therefore also applies in this context. Whether globalization will eventually eradicate those differences and African culture succumbs to its pressures, remains to be seen.
BIBLIOGRAPHY
Bachofen, J. 1861. Mutterrecht (Maternal Law). In: Engels, F. Origin of the Family,
Private Property and the State. 1902. Chicago: Charles H. Kerr & Company. Elliot, F. 1986. The Family: Change or Continuity? London: MacMillan.
Elliot, F. 1996. Gender, Family and Society. London: MacMillan.
Engels, F. 1902. Origin of the Family, Private Property and the State. Chicago: Charles
H. Kerr & Company.
European Commission. 2001. The Social Situation in the European Union. Eurostat.
Giddens, A. 2001. Sociology (Fourth Edition). Oxford: Polity Press.
Goode, W. 1963. The Family. Englewood Cliffs: Prentice Hall.
Goode, W. 1964. World Revolution and Family Patterns. New York: The Free Press.
Hajnal, J. 1983. Two Kinds ofPre-industrial Household Formation System. In: Wall, R. et al. (eds.). Family Forms in Historic Europe. Cambridge: Cambridge University Press, pp. 65-104.
Haralambos, M. & Holbom, M. 1995. Sociology - Themes and Perspectives. London: Unwin Hyman.
Held, D. et al. 1999. Global Transformations. Oxford: Polity Press. In: Giddens, 2001. Sociology. Oxford: Polity Press.
Hochschild, A. 1997. The Time Bind. In: Giddens, 2001. Sociology. Oxford: Polity Press.
Kanter, R. 1977. Men and Women of the Corporation. New York: Basic Books.
Kertzer, D. 1991. Household History and Sociological Theory. Annual Review of Sociology, vol. 17,pp.155-179.
Kertzer, D. & Laslett, P. (eds.). 1995. Aging in the Past - Demography, Society and Old Age. Berkeley: University of California Press.
Laslett, P. & Wall, R. (eds). 1972. Household and Family in Past Time. Cambridge: Cambridge University Press.
Le Play, F. 1871. L 'Organisation de la Famille. Paris: Tequi.
Macfarlane, A. 1978. The Origins of English Individualism. Oxford: Basil Black-well.
Maine, H. 1861. Ancient Law. London: Murray.
McLennan, D. 1886. Primitive Marriage (Studies in Ancient History). In: Engels, F. 1902. Origin of the Family, Private Property and the State. Chicago: Charles H. Kerr & Company.
Morgan, L. H. 1877. Ancient Society (Researches in the Lines of Human Progress from Savagery, through Barbarism, to Civilization). New York: Henry Holt & Company.
Parsons, T. 1956. The American Family: Its Relations to Personality and the Social Structure. In: Parsons, T. & Bales, R. (eds). Family, Socialization and Interaction. New York: The Free Press.
Russell, M. 1994. Do Blacks live in Nuclear Families? An Appraisal of Steyn’s work on Urban Family Structure in South Africa. South African Sociological Review, 6/2, pp. 56-67.
Shorter, E. 1975. The Making of the Modern Family. Glasgow: Fontana\Collins.
Smith, J. & Wallerstein, I. 1992. Creating and Transforming Households. Cambridge: Cambridge University Press.
Steyn, A. (1995). Urban household structures in the Republic of South Africa. In: C. Yi, (ed.), Family formation and dissolution: Perspectives from East and West. Taipei: Academia Sinica, pp. 169-204.
Stone, L 1977. The Family, Sex and Marriage in England 1500-1800. London: Wieden-feld & Nicholson.
Therbom, G. 2001. African Families in a Global Context. Paper delivered at the Worshop on Global Porcesses. Cape Town, November 29 - December 2, 2001.
Wall, R; Robin, J & Laslett, P. (eds). 1983. Family Forms in Historic Europe. Cambridge: Cambridge University Press.
Ziehl, S .C. 1997. Family diversity -A South African perspective focusing on Whites in Grahamstown. Unpublished doctoral thesis, Rhodes University, Grahamstown. Ziehl, S. C. 2001. Documenting Family Patterns in South Africa: Are Census Data of any Value? Paper delivered at the Workshop on Global Porcesses. Cape Town, November 29 - December 2, 2001.
POVZETEK
GLOBALIZACIJA, MIGRACIJE IN RAZNOLIKOST DRUŽINSKIH VZORCEV Susan C. Ziehl
Do nedavnega je bilo le malo dialoga med raziskovalci in teoretiki globalizacije na eni strani ter znanstveniki, ki se ukvarjajo z družinskimi študijami, na drugi strani. Pogosto beremo, da globalizacija vpliva na družinske vzorce; kakšni natanko so ti učinki in kakšna je njihova narava, pa pravzaprav še ni bilo dovolj pojasnjeno. Namen tega prispevka je torej poskus vzpostavitve povezav med diskusijama o globalizaciji in o raznolikosti družinskih vzorcev.
V prvem delu avtorica obravnava nekatere vzroke za pomanjkanje dialoga med glo-balizacijskimi in družinskimi študijami. V drugem delu analizira učinek globalizacije in še posebej migracij na družinske vzorce v Evropi. Končno primerja južnoafriške družinske vzorce s tistimi, ki so značilni za eno od evropskih dežel, in sicer za Veliko Britanijo. Raziskava jo privede do sklepa, da je imela globalizacija doslej minimalen vpliv na družinske vzorce, ki jih najdemo v posameznih evropskih družbah, in da razlike med regijami, kar zadeva tradicionalno prevladujoče družinske vzorce, ostajajo znotraj evropskega konteksta skoraj še ravno tako opazne kot nekoč. Prav zato je seveda tem manj verjetno, da bi lahko bile te razlike kakorkoli povezane z globalizacijo. Druga ugotovitev pa je, da se prava raznolikost prevladujočih družinskih vzorcev kaže šele na globalni ravni, kar postane očitno ob primerjavi razreza družinskih vzorcev v eni od afriških in eni od evropskih dežel. Ali bodo globalizacijskiprocesi sčasoma izničili tradicionalne razlike v prevladujočih družinskih vzorcih na globalni ravni, pa je v tem trenutku nemogoče napovedati.
Susan Ziehl je doktorica znanosti in izredna profesorica na Department of Sociology and Industrial Sociology, Rhodes University, Grahamstown, Južna Afrika. Na Univerzi Stellen-bosh je diplomirala in magistrirala iz ekonomije, doktorirala pa je na Univerzi Grahamstown. Predmet njenih raziskav in objav so strukture družin in gospodinjstev, feminizem in sodobna reprodukcijska tehnologija, zakon o družini in multikulturalizem, enostarševke družine in afirmativno dejanje.