1. HISTORICAL DISCOVERIES AND DEBATED HYPOTHESES

As French historians André Burguière and François Lebrun have observed, in the 1960s ‘any historian dealing with the family since the sixteenth century would have spoken of the gradual rise of the modern nuclear family in Europe under the influence of the economic transformations and the “modernizing” effect of the state’.Footnote ⁷ Common to most sociological theories, rather passively accepted by historians, ‘was a belief in the passage of the traditional family, featuring large and complex domestic groups, into the present-day restricted family, with the residential group limited to the conjugal unit’. It was assumed that in the past the size and composition of domestic groups had been very much the same all over Europe and that the differences that could be observed in the twentieth century were to be explained by economic or cultural lags caused by differential rates of modernization. However, a time would soon come, it was predicted, when national and regional differences would dissolve and ‘uniformity [would] reign with European families responding to a common policy’.Footnote ⁸ These beliefs began to be challenged in the 1970s, as part of a general retreat from the rigid evolutionary framework that had come to dominate the social sciences. In the field of family studies, however, further and specific reasons to dispute the received wisdom were provided by the surprising results of a spate of historical investigations, which indicated that new maps of family forms and marriage patterns in pre-industrial Europe had to be drawn and raised questions of relevance not only to history but to sociology and other social sciences as well. According to some scholars, as we shall see, these questions remain crucial in order to understand the role of the immediate family and of the larger kinship network as sources of social security in present-day Europe. They emphasize that, contrary to sociological predictions, large variations can still be detected, and suspect that such variations are not due to different rates of modernization, but rather to different paths to modernity heavily dependent on, and intertwined with, historical and cultural differences of long standing.

One of the first scholars to try to take stock of the new evidence unearthed by these studies was Alan Macfarlane in an article published in 1980, the primary purpose of which was to inform social anthropologists and historical geographers about some recent achievements of historical research on marriage and the family in Europe.Footnote ⁹ One major achievement was John Hajnal's discovery that for at least four hundred years, through the sixteenth to the nineteenth centuries, there had been, in Macfarlane's words, a ‘huge “fault” line, running down from north to south’,Footnote ¹⁰ roughly from St Petersburg to Trieste: to the east of this line a pattern of early and universal marriage had dominated; the countries to the west had, by contrast, been characterized by a pattern of delayed marriage for both men and women, with high proportions of both sexes never marrying. A few years later, spurred by Peter Laslett and the Cambridge Group, other studies had located in western Europe ‘a second demographic fault line’,Footnote ¹¹ running across from west to east: to the north of this line, households in the past were small in size and ‘simple’ (nuclear) in structure, consisting of parents, some unmarried young children and possibly servants; to the south, they were larger and often ‘extended’ in structure. ‘We thus have’, Macfarlane concluded, ‘three demographic regions, the eastern, the western (north) and the western (south)’.Footnote ¹²

This article inaugurated a series of attempts by historians and historical demographers to map the European family. It will suffice to mention three well-known propositions put forward in rapid succession in the early 1980s: Richard Smith's suggestion that a ‘Mediterranean’ marriage and family pattern was detectable well before the sixteenth century;Footnote ¹³ Hajnal's distinction between a north-west European ‘simple household’ system of household formation and a ‘joint household’ system to be found in southern and (with some modifications) eastern Europe;Footnote ¹⁴ and, finally, Laslett's partition of Europe into four macro-regions: ‘Western’, ‘West-central’, ‘Mediterranean’ and ‘Eastern’.Footnote ¹⁵ It should be noticed that all these attempts shared an emphasis on long-term continuity and, ultimately, on the role of culture. Indeed, Macfarlane had already argued that the demographic structures uncovered by historians were coterminous with broad cultural regions: it was a ‘pleasing irony’, he had commented, ‘that one of the “hardest” of the social sciences [i.e. demography] should be nudging us [anthropologists] towards that realization of the importance of “culture” which, so some have argued, is the special contribution of anthropology’.Footnote ¹⁶ Another point to be noticed, in view of its relevance to the study of the relationship between kinship and social security, is that one of the less evident but actually most important aims of the research agenda set by Laslett was to shed light on how families or domestic groups had taken care of vulnerable categories such as orphaned children, widows, the elderly and other needy people. The discovery that in pre-industrial times England and more generally north-western Europe displayed a marked prevalence of neo-local, simple family households implied that many individuals were inevitably left without familial support, in a condition of actual or potential hardship. The small and structurally simple households of England and north-western Europe were therefore vitally dependent on external support, but where did this support come from? The existence in England of a deep-rooted and highly developed state system of poor relief led Laslett to suggest tentatively that external support had come essentially from what he liked to call ‘the collectivity’, a term by which he meant collective institutions such as parishes, municipalities, charitable bodies or indeed the state, which transcended the level of both the family as a bounded co-residential unit and the kinship networks made up of relatives living outside the household.Footnote ¹⁷ The clearest formulation of this claim was provided by Laslett himself some ten years later in the form of his ‘nuclear-hardship hypothesis’, which maintains that in England and north-western Europe, where simple-family households were dominant, transfers from the collectivity were of the highest consequence, whereas transfers from kin were of little significance.Footnote ¹⁸

Once again, it is relevant to observe that the grounds on which these and other similar claims were based are partly cultural (in England and north-western Europe, moral obligations to reside with one's kin in order to provide support when needed were either absent or weak), which explains why the nuclear-hardship hypothesis entailed a set of opposite predictions for those European regions where the role of kinship was believed to have been of much greater importance for primarily cultural reasons, namely southern Europe and possibly eastern Europe. The important point for us is that this hypothesis proposed that the macro-regional family and marriage patterns tended to correspond – geographically as well as typologically – to contrasting systems of welfare provision: the role of both kinship and family, it was contended, had been far greater in the southern and eastern parts of Europe than in the north-western countries, where the long-term prevalence of intrinsically vulnerable nuclear families had been made possible by support provided by the state, or had perhaps urged the creation of a system of state or otherwise public welfare. This argument was encouraged by the geographical contrast (possibly related to divergences between Catholic and Protestant poor relief policies) habitually drawn by historians between a ‘northern’ welfare system – best exemplified by England, and characterized by comprehensive, parish-based outdoor relief – and a ‘southern’ or ‘continental’ system, ‘supposedly haphazard and limited in its finances and scope, and based on the “indoor” assistance provided in large hospitals’.Footnote ¹⁹

A favourite way of testing this hypothesis was through the investigation of the living arrangements of the elderly, and the evidence coming from research on southern Europe appeared to provide convincing corroboration. It showed that in the areas characterized by joint family systems the great majority of old people lived in large complex households. In a sharecropping community of central Italy, to cite one of the best studies, David Kertzer ascertained that over 70 per cent of those aged over 65 in 1880 resided in extended or multiple households. No one had been left alone, and virtually no widows or widowers lived in households without at least one of their children.Footnote ²⁰ Similar patterns and proportions were found in the geographical belt extending from northern Spain and the Pyrenees through southern France to the western Alps, where stem family arrangements were widespread.Footnote ²¹

As is well known, in the 1980s these theses became subjected to a severe critique. The attack was mainly launched either by anthropologists or by historians influenced by anthropology, who denied the legitimacy of macro-regional approaches for several reasons. It was contended, for instance, that the alleged cultural regions were hardly homogeneous, and that supposedly distinctive traits were detectable on both sides of boundaries that appeared to be far more blurred than originally envisaged by the proponents of the macro-regional theses. These arguments were used especially to criticize the so-called ‘Mediterranean family model’,Footnote ²² but more generally, to quote Kertzer, ‘the whole enterprise of branding major areas of Europe as having a particular type of household system’ was judged by many to be ultimately misleading. Instead of looking for broad regional uniformities, the task of family historians was to develop a theoretically more comprehensive and sophisticated approach where economic, demographic, ecological and cultural factors would all be taken into account in order to properly address the unexpected degree of variability.Footnote ²³ This was tantamount to proclaiming the superiority, indeed the triumph, of micro-approaches over macro-approaches. By the middle of the 1990s the ‘Hajnal/Laslett model’ had been seriously questioned, and a further blow was inflicted by historical studies of assistance and poor relief that strongly suggested that the contrast between a ‘northern’ and a ‘southern’ model of welfare provision either did not exist or its importance had been greatly exaggerated.Footnote ²⁴

It therefore came as a surprise when David Reher, a former critic, published an article in which he contended that contrasts between macro-regions in Europe – and particularly between the northern and southern halves of western Europe – not only had existed in the past but persisted in the present, as demonstrated by the continuing much greater strength of family ties in the Mediterranean countries.Footnote ²⁵ The care of the elderly again provides a critical test. Reher has no doubt that in Mediterranean Europe the family was far more essential for the well-being of its more vulnerable members than in northern latitudes and that the care of the elderly especially fell almost exclusively on the family – as shown still today by the much higher proportion of old people living in institutions in northern Europe than in Italy, Spain or Portugal. There is a significant qualification, though: Reher believes that joint-family households prevailed only in very few southern European areas, and that southern Europe was actually characterized by simple, neo-local households. This is questionable.Footnote ²⁶ What really matters for us is, however, that Reher stresses that co-residence was not the only means of taking care of the elderly. Alternative means were, he says, ‘the circulation of the elderly among the households of their offspring, or the spatial proximity between the homes of the elderly and those of their children’.Footnote ²⁷ In view of Reher's proclivity to maintain that in southern Europe too nuclear arrangements have historically been much more frequent than complex households, it is plain that he regards these two latter ways of providing welfare to have been definitely more important than co-residence. A visitor from northern Europe would have been surprised not so much by differences in household structure as by the divergent ethics almost palpable in northern and southern Europe.

Initially spurned or utterly rejected, especially by historians and anthropologists, as crude, simplistic and prone to reducing nuanced differences to black versus white contrasts, Reher's thesis and his reliance on the contrast between ‘weak’ and ‘strong’ ties have eventually enjoyed a huge success. Although they are certainly open to debate, it seems significant that some distinguished historians have taken sides with Reher,Footnote ²⁸ and that even those who remain critical are conceding that they cannot avoid having to come to grips with his theses.Footnote ²⁹ Even more significant are the numerous references to his work in the sociological and demographic literature on present-day welfare systems, which suggest that he has been more effective than Laslett himself in convincing social scientists of the relevance of historical work to the issues they are investigating. Whatever one may think of his arguments, there can be little doubt that, along with the partly similar ones put forward by other leading scholars,Footnote ³⁰ they compel us to face squarely a number of questions which are both big and thorny: should we emphasize the commonalities found by studies on family and kinship in various parts of historic and contemporary Europe or rather the divergences they have detected? If divergences do exist, can they be traced to structural or cultural factors? Should we give more weight to similarities in trends or to persisting differences in levels? And, finally, behind the new residential arrangements and the incipiently novel ways of reckoning kinship that are becoming manifest all over Europe, can we discern genuine and radical change or, rather, the signs of deep-lying continuity?

2. RECENT CROSS-NATIONAL COMPARATIVE RESEARCH: FINDINGS AND QUESTIONS

The last decade has been marked by intense macro-regional comparative research bearing, directly or indirectly, on the relationships between kinship, social security and the welfare state. Several major cross-national studies followed in the footsteps of such efforts as the pioneering survey on ‘Social Networks and Support Systems’ launched in 1986 by the International Social Survey Programme (ISSP), or Frank Castles' attempt to cluster European countries into ‘families of nations’ displaying different patterns of kinship recognition, family obligations, formal legislation and policy-making.Footnote ³¹ They include, among others, the new 2001 ISSP survey on social relations and support systems, the SOCCARE comparative project on social care in western Europe undertaken in 2000–2003Footnote ³² and the 2002 Eurostat study on the family arrangements of women and men in Europe in the late twentieth century,Footnote ³³ not to mention the fourth and fifth ‘waves’ of the ‘World Values Survey’ and, of course, SHAREFootnote ³⁴ and KASS.Footnote ³⁵ In addition, a number of comparative studies adopting a variety of macro-regional approaches have been published, ranging from works specifically focused on the relationships between family and welfare in EuropeFootnote ³⁶ to Göran Therborn's world-wide history of the family in the twentieth century.Footnote ³⁷

Although space inhibits any detailed analysis here, it seems fair to say that by and large these studies agree that by the beginning of the twenty-first century there is little evidence that the process of convergence foretold by modernization theorists has been completed or, indeed, that it has proceeded very far. Within Europe, in particular, pronounced differences between countries and ‘families of nations’ continue to be spotted. The results of SHARE, a large research project covering eleven western European countries, are especially telling. Household structure, residential proximity to kin and contact with them were among the first variables to be explored, and the verdict has been that ‘there are important differences among the “strong family countries” in the South and the “weak family countries” in the North’.Footnote ³⁸ It is worth remembering that the first cross-national surveys conducted in the 1980s had signalled that, although the nuclear family consisting of parents and unmarried children proved to be the dominant type everywhere in western Europe, in the south the proportion of parents living with adult children was consistently higher that in north-western Europe. Perhaps even more significantly, they had also shown that in Mediterranean Europe the growing nuclearization of households had been countered by a much stronger tendency for married children to settle close to their parents.Footnote ³⁹ More than a quarter of a century later, SHARE still finds that the spatial pattern of proximity between elderly parents and their adult children ‘exhibits a very clear North–South divide’,Footnote ⁴⁰ and a clear north–south gradient is equally noticeable with respect to rates of co-residence and frequency of contacts among adult family generations. If it is true that living arrangements shape patterns of care-giving, then these differences should be mirrored in the crucial field of family support: predictably, SHARE data reveal, in this respect too, ‘a strong North/South European divide’.Footnote ⁴¹

Interestingly, SHARE has also paid considerable attention to financial transfers between generations, and again significant differences have emerged between northern and southern countries: although intergenerational transfers tend to flow downwards all over western Europe, younger respondents receive more from parents in the north, older respondents from children in the south.Footnote ⁴² This emphasis on a divide between northern and southern Europe is due to the fact that all the countries in the SHARE sample belong to western Europe, but similar conclusions concerning the differing strength of family ties across nations are reached by studies covering both western and eastern Europe, such as the ‘Immigrazione e Cittadinanza in Europa’ (ICE; ‘Migration and Citizenship Rights in Europe’) survey,Footnote ⁴³ or indeed KASS. There seems, therefore, to exist a broad consensus that cross-national and macro-regional differences are not just a figment of classificatory imagination. This, however, raises the big and thorny issues referred to above, the first one being the perennial question of the relative importance of ‘structural’ and ‘cultural’ factors in accounting for these differences.

3. STRUCTURE AND CULTURE

It has been rightly remarked that comparative cross-national research is almost inevitably confronted by the contrasting lures of structuralism and culturalism: while structuralist theories expect similar macro-characteristics to have equifinal effects in different national contexts, the alternative approach assumes that cultural factors can impinge on social and economic structure to produce different outcomes from similar structural conditions. Thus, ‘if contrasts between countries are dominant, they point towards cultural explanations, but if similarities predominate they indicate the force of general processes’.Footnote ⁴⁴

The increasing ‘nuclearization’ of Italian domestic groups in the second half of the twentieth century is a case in point. In the early 1980s census data indicated that the proportion of ‘complex’ households had been halved from 22.4 per cent in 1951 to 11.3 per cent thirty years later, whereas the proportion of nuclear families has steadily grown in the same period from 67 per cent to nearly 80 per cent.Footnote ⁴⁵ This was initially interpreted as part of a wider transition from a traditional to a modern social order entailing Italy's belated convergence with northern-European styles of family life. The story seemed simple enough: spurred by the combined effects of the two parallel processes of industrialization and urbanization which Italian society was experiencing at an accelerated pace, the individualistic and modern nuclear family was eventually ousting the traditional, patriarchal family of old and severing, or at least loosening, the links between newly married couples and their relatives. Of course, by weakening the links with parents and kin, the modern couple was also, to a large extent, relinquishing their help. But in northern Europe this had been made possible at a relatively early stage by the aids which married couples were entitled to receive from the state, and it seemed legitimate to surmise that in Italy, too, this move had been favoured by the policies promoted by the state, especially in the 1970s, when a number of measures specifically intended to support the family had been enacted. In recent years, however, an alternative interpretation has been put forward by several Italian sociologists and demographers, who have contended that the ‘break’ with parents and kin might have been far less drastic than the declining proportion of complex families would make us believe. They have drawn attention to evidence which suggests that the old patterns of co-residence have actually been replaced by new patterns of residential proximity between parents and children and, more generally, between relatives. This evidence points to a twofold process: while it reveals a growing desire on the part of the younger generations to seek privacy and greater autonomy through the creation of separate households at marriage, it also suggests that both the new couples and their parents continue to seek residential proximity. The immediate motivation behind this search for proximity is mainly the parents' wish to ‘help downwards’, for example by buying their children a house located not far from their own house. In the longer term, however, this residential choice comes to comply with a widespread moral value that asks adult children to ‘help upwards’ by taking care of their elderly parents. All in all, the seeming metamorphosis of coresidence into proximity is seen to be symptomatic on the one hand of significant cultural continuity as regards the role and importance of close kin as sources of social security, and on the other of significant cultural difference between Italy and the other southern European countries, where levels of residential proximity are very high, and northern Europe, where a much greater degree of spatial dispersion of relatives has been observed.Footnote ⁴⁶

Theorists of a structuralist persuasion can obviously counter that both this difference in residential patterns and the difference in the direction of transfers reflect differences in the levels of welfare systems, older needy people being more frequently dependent on their children in the southern (or eastern) countries, where welfare regimes are weaker, than in northern Europe. Yet, cultural factors cannot be easily ruled out. As Claudine Attias-Donfut, Jim Ogg and François-Charles Wolff remind us in their analysis of SHARE evidence, two main models have been advanced to explain financial transfers: ‘altruism’ and ‘exchange’. If altruistic motivations prevail, one should expect transfers to flow from the least to the most financially needy generation irrespective of any present or future reciprocating help. By contrast, ‘in the exchange model, financial transfers from one generation reflect the payment of services and visits provided by the other generation. Services and their financial counterpart are embedded in current and future obligations of reciprocity.’Footnote ⁴⁷ SHARE data show that in the northern countries ‘parents are more likely to give money to children who live far away from them than to children who live nearby’, which is in line with the predictions of the ‘altruistic’ model, whereas financial transfers in the southern countries appear to be part of intergenerational exchanges and, ultimately, to be driven by more binding moral obligations.

This indicates that there is a case for culture, but also that such a case is by no means settled in view of the plausibility of alternative ‘structural’ explanations. To tip the balance in favour of cultural explanations, crucial help would seem to come from history, as suggested by Therborn's wide-ranging inspection of family history in the twentieth century, which leads him to distinguish – on the basis of their marriage and fertility patterns as well as of their internal power relations – seven ‘family systems’ in the modern world. Therborn emphasizes that a striking and not altogether expected finding to come out from census data and recent surveys is that over the past hundred years ‘all family systems, throughout their changes, have tended to preserve specific characteristics’.Footnote ⁴⁸ European family systems, in particular, appear to be separated from one another by boundaries which do not coincide with present-day national borders (as hypotheses stressing the role of institutional factors would lead one to expect), but rather with the ‘fault lines’ identified by historical demographers in the 1960s and early 1970s. He observes, for instance, that ‘the classical European family divide, running from Trieste to St Petersburg … is still visible in 2000’,Footnote ⁴⁹ and the map of contemporary European family systems he proposes is not too different from the one drawn by Laslett in 1983.Footnote ⁵⁰ Reher's contention that macro-regional contrasts not only existed in the past but persist in the present would seem to be vindicated, and the case for cultural explanations made much stronger.

Therborn's stand is epitomized by his theoretical decision to view family systems as ‘geocultures’, a term selected to indicate that they take their colouring ‘from the customs and traditions, from the history of a particular area, a cultural wrapping which may remain after structural, institutional change, leaving imprints on the new institutions’.Footnote ⁵¹ Therborn is forced, however, to acknowledge that the ‘cultural terrain’ is problematic ‘given the heterogeneity of the notion of culture’.Footnote ⁵² In a rather different vein, other scholars have pointed out that a serious limitation of cross-national projects is their tendency to explore only variables that lend themselves to measurement but are hardly sufficient to pinpoint the role of culture. As noted by Vern Bengtson in an influential article, a comparatively easy way of ‘measuring’ kinship consists of focusing on tangible assets and ‘exchanges that can be counted’, such as the amount of financial support given to or received from members of other generations.Footnote ⁵³ This is precisely what sociologists and economists have now done for some years, with often surprising results. But is it legitimate to infer, and ‘measure’, what we might call ‘kinship culture’ from behaviours? It is Bengtson himself who warns us that kinship relations ‘reflect not only behaviors (frequency of contact, help given or received) but also emotional-cognitive dimensions (feelings of closeness and bonding, similarity of values and opinions)’.Footnote ⁵⁴ If we take these dimensions into consideration, the whole picture can change considerably, as it is at least conceivable that in a given country kinship may be ‘weak’ in terms of contacts but ‘strong’ from an emotional-cognitive point of view. Clearly, very much depends on the notion of culture that is used. Daniele Gaggio, an economic historian with anthropological leanings, has pointedly argued that sociologists and economists tend to see culture as a set of norms made observable by behaviours, whereas most anthropologists and many historians conceive of culture as ‘a system of signs and symbols (rather than behaviours and institutionalised norms), which are highly unstable, open to encounters and contaminations, and decipherable only through a prolonged hermeneutic process’.Footnote ⁵⁵ This process, we may note, usually involves a prolonged stay in the field. This is one of the reasons why the KASS project has trodden different research paths: on the ‘macro’ side, historical investigations have first been carried out in order to produce, for each of the eight countries in the sample, sizeable accounts of the legal, economic, demographic, political and cultural contexts of the changing relations between families, the state and other welfare agencies, paying special attention to the twentieth century. On the ‘micro’ side, intensive fieldwork has been conducted in nineteen localities and fieldworkers were asked to combine survey and ethnographic research strategies by administering very detailed questionnaires while simultaneously undertaking a prolonged period of ‘participant observation’ to gain close familiarity with local social organization and culture. The final part of this article will tap the historical and ethnographic evidence collected by this project to discuss some broad questions concerning difference and similarity in European kinship patterns and their relations to the various welfare regimes.

4. A NEW EUROPEAN ‘KINSHIP CULTURE’ OR PERSISTING DIFFERENCES?

A general tendency highlighted by the historical accounts to emerge from KASS is a withdrawal of the welfare state, paralleled by efforts to bring the family back to take upon itself support duties which had been previously taken over by the state.Footnote ⁵⁶ Even in Sweden there is now pressure for families to mobilize their resources to take care of the elderly and disabled.Footnote ⁵⁷ A crucial question is whether the European family of the new millennium will be able to fulfil these duties, for it is evident that deep structural transformations in the labour market and no less profound social and demographic changes, such as the decline of fertility and nuptiality and the growing instability of marriages, have made it very different from what it was in the ‘golden age’ of the first decades after the Second World War. Yet, it would be dangerous to accept the argument that family and kinship are largely the same thing and are destined to fade together. It would rather seem that the weakening of the family as a bounded co-residential unit founded on marriage is entailing an increase in the importance of kinship ties, not least because the instability of marriage is making kin an essential resource for support in situations of family crisis.

As we have seen, this has led many scholars to argue that all over Europe we are witnessing a ‘rediscovery’ of kinship. It should not be forgotten, however, that this reinstatement of kinship is taking place in times of a retreat of the welfare state: in Sweden, ‘legislators, administrators and experts attempt to re-invigorate and strengthen kinship ties which during modernization had been relatively ignored’;Footnote ⁵⁸ and in post-1989 Germany the legal definition of kin has been broadened, so as to increase the number of individuals who can be summoned to take care of needy kin.Footnote ⁵⁹ Although legal obligations may not be perfectly mirrored by actual practice, nevertheless the patently greater number of kin categories that are legally expected to support a needy relative in Mediterranean Europe vis-à-vis northern and central Europe has long been taken as an indicator of the strength of kinship in these societies as well as of the tendency to regard the role of the state as subsidiary to that of the family.Footnote ⁶⁰ It would now seem that northern and central European countries are moving towards what Jane Millar and Andrea Warman have called ‘a more “Italian” model’.Footnote ⁶¹ If we add that there is mounting evidence that inheritance and other intergenerational transfers, contrary to Durkheim's prediction that inheritance would lose its salience in modern societies,Footnote ⁶² are acquiring greater significance in a period of declining welfarism,Footnote ⁶³ we are entitled to suspect that the rediscovery of kinship ties in contemporary Europe is mainly obeying instrumental imperatives.

According to some leading scholars, however, this is not the whole story. Summarizing the results of the KASS ethnographic studies, Martine Segalen asserts that, although there is no disputing that instrumental reasons have contributed to the recent rise of kinship, these studies demonstrate that European kinship is undergoing a shift from ‘interest’ to ‘emotion’, from status-based moral obligations to ‘elective’ relations free from material dependence and economic calculation. This is a point made with special force by the ethnographic accounts of kinship and social security in Swedish, German and French localities. ‘A small and self-selected group of close kin’: this is, according to Segalen, ‘the new kinship form in contemporary Europe, with strong commonalities in meaning and function’.Footnote ⁶⁴

This statement proclaiming the advent of a new and single kinship culture in Europe is at odds with the emphasis others have placed on differences between nations and macro-regions – differences, it should be noted, that are by no means absent in the KASS ethnographic record or in the quantitative evidence piled up by KASS fieldworkers. Indeed, the findings of KASS local surveys appear to be entirely consistent with the macro-regional differences revealed by large-scale surveys and actually suggest an intriguing similarity between the geography of present-day European residential arrangements and the map of pre-industrial household patterns proposed by Laslett in 1983. No less importantly, the uniquely detailed information yielded by the ‘kinship network questionnaires’ administered by KASS fieldworkers makes visible a strong association, at the local level, between household size, household structure, the density of the kinship network (measured by the proportion of ego's relatives living nearby) and the frequency of locally endogamous marriages. The values of all these variables vary substantially and tend to be higher precisely in those parts of contemporary Europe, heavily concentrated in the southern and eastern regions of the continent, where obligations to support kin are morally and legally more binding. This corroborates the view that in some European regions individuals are enmeshed in social universes where the weight of kinship relations is heavier than in others and where kinship networks differ considerably from the ‘small and self-selected groups’ described by Segalen. Interestingly, whatever the dimension that is taken into account, the two polar positions among the nineteen fieldwork sites are most often occupied by two localities that epitomize the north–south gradient, namely Vällingby, a Stockholm suburban neighbourhood investigated by Hans Marks and David Gaunt, and Tramonti, a southern Italian village studied by Carlo Capello and Nevill Colclough.Footnote ⁶⁵ This polarity is reminiscent of the similarly extreme contrasts between northern and southern European countries or areas that have been the subject of previous analyses.Footnote ⁶⁶ However, these differences lend themselves to diverging evaluations: while acknowledging that they are real, some scholars believe that they should not be given undue emphasis and prefer to draw attention to commonalities.Footnote ⁶⁷

Opinions are therefore divided between those who believe that a unified ‘kinship culture’ is emerging in Europe and those who maintain that cultural differences of long standing are sturdily persisting in the face of change. The issue is all the more intricate and puzzling as there might be truth to both alternatives. It seems legitimate to wonder whether the question is posed correctly – whether, in particular, we are really facing an alternative, or should rather be careful to avoid the familiar pitfall of confusing trends with levels. If it is incontestable that the current rediscovery of kinship, largely spurred by its increasing weight as a source of social security, is a widespread tendency reducing previous gaps between European nations, it seems equally undeniable that differences still exist: Segalen herself has recently remarked that some major international surveys show that European countries ‘are situated at different points along a scale of intensity with which kinship ties are maintained’, with Italy and other Mediterranean countries ‘highest on such a scale’.Footnote ⁶⁸ The trouble is that anthropologists are especially disinclined, as we have seen, to accept that the cultural aspects of kinship can be measured and scales ranking their strength devised, which entails that such differences are denied any cultural significance. Is this sceptical attitude completely justified? Therborn is surely right that ‘crude rates are crude measures’, but ‘occasionally census numbers glimmer’;Footnote ⁶⁹ census data and other statistical evidence show that after the period of relative convergence of family and marriage patterns experienced by most of Europe around the mid-twentieth century, ‘new’ spatial differences have emerged which closely resemble the ‘old’ differences brought to light by historical demographers. There seems to have been, therefore, a reconnection with old patterns which sits uneasily with the theses of those who hold that the last decades of the twentieth century have seen the establishment of a new and unified kinship culture.

This explains why the question of the role of long-term continuities in shaping the relations between kinship and social security is receiving considerable attention from supporters of the notion of ‘path dependence’,Footnote ⁷⁰ who theorize that the future will be strongly conditioned by the past. Especially the evidence pointing to the existence of macro-regional clusters has fuelled speculation about the influence of culture on welfare systems. However, both empirical tests and theoretical assessments of these hypotheses have stumbled against several difficulties. One such difficulty is that most analyses have rather rigidly grouped countries either into welfare typologies or into macro-regions, thereby generating classification problems of specific countries and obscuring internal differences within groups of countries. Trouble arises also from a tendency to equate history and culture, to conceptualize the latter as a package of constraining norms and to infer from the existence of long-term continuities the idea that these norms are almost immutable. And finally, structural and cultural explanations are only too often seen as dichotomous alternatives. As a result, it becomes very tempting, for instance, to trace the greater pervasiveness of the family as a welfare agency in southern Europe back to deep Mediterranean cultural roots, without bothering too much about the historical specificities of institutional development in this area.

There are different ways of coping with these problems. In order to overcome the difficulties plaguing typological and macro-regional approaches, in a recent article Matthijs Kalmijn and Chiara Saraceno have adopted a ‘continuous approach’ whose main principle consists in treating the underlying concept on which the typology or grouping of countries is based as a continuous variable and to assign countries a score on this continuum. They have also argued that the role of need has been neglected in previous analyses and that instead of comparing countries in relation to the overall level of support parents receive from children, it is more profitable to examine the degree to which children respond to the needs of parents. This has enabled them to use SHARE data to demonstrate that the effect of parents' need on the support received from a child is indeed stronger in southern European countries, which are more ‘familialistic’ both in subjectively perceived intergenerational obligations and in implicit policy assumptions.Footnote ⁷¹ This would seem to confirm, as Kalmijn and Saraceno themselves note, that there are persisting differences in Europe. They are keen to add, however, that the ‘familialistic’ Mediterranean countries and the more individualistic countries of north-western Europe ‘differ for both cultural and institutional reasons, and these two dimensions are correlated’.Footnote ⁷²

The challenge now is to explore this correlation more satisfactorily. This can be done using statistical tools, but a different perspective can be provided by ethnographic studies. As we have noticed, there has been much speculation about the influence exerted by culture on welfare systems. There has been less interest in exploring how, and to what extent, the welfare state has influenced family and kin relations so far and is likely to influence them in the future. This was one of the central questions KASS intended to tackle, and revealing glimpses have indeed come especially from the ethnographic studies. They show, in particular, that in countries with ‘weak’ welfare regimes such as Italy and Poland, in the absence of satisfactory or affordable state provision, the support and the mutual help offered by kin is crucial in both urban and rural sites. The contrast with the ‘perceived feeling of individual security’ generated by the Swedish welfare state is, once again, striking. The ethnographic material from Sweden illustrates very effectively what being born and growing up in a social environment in which ‘family and kin are somewhat freed from the burden of economic support and care’ means in terms of expectations.Footnote ⁷³ In his manifesto of modern ethnographic method, Bronislaw Malinowski acknowledged in 1922 that ‘in certain results of scientific work – especially that which has been called “survey work” – we are given an excellent skeleton, so to speak, of the tribal constitution, but it lacks flesh and blood’.Footnote ⁷⁴ Ethnographic studies aimed at giving ‘flesh and blood’ to the abstract relations between structural and cultural factors can reveal that major life decisions may have not only an economic but also a cultural background, and that this background is not a rigid heritage of remote times but is continuously moulded by ‘a well-established symbiotic interaction between the welfare state and kinship structures’ (as suggested for Sweden)Footnote ⁷⁵ or, alternatively, by the painful insufficiency of state support – a dearth that familialistic ideologies contribute to perpetuate.

5. CONCLUSION

Reading the literature on family and kinship in present-day Europe reviewed in this article suggests that differences ‘in meaning and function’ do exist and predominate over commonalities. Similarity in trends should not obscure significant differentials in levels, all the more so since it is not clear whether the ‘rediscovery of kinship’ reported by scholars in many European countries can be seen as a proper trend or, rather, as a widespread but heterogeneous tendency. The resemblance between contemporary spatial patterns and the macro-regional differences brought to light by historical demographers implies that the future of kinship in Europe can hardly avoid being largely path-dependent. However, context-dependent factors are also at work: indeed, there are signs that, owing to the generalized decline of the welfare state, the gap between the ‘weak’ family ties of Scandinavia and north-western Europe and the ‘strong’ ties of southern and eastern Europe is lessening. There is a paradox, though. In the 1960s sociologists predicted that a time would soon come when national and regional differences would dissolve and uniformity would reign, with European families responding to a common policy aimed at unburdening them of obligations towards kin. The evidence surveyed in this article suggests that, at least in certain respects, European countries are approaching some degree of uniformity, but they are unexpectedly converging towards forms of welfare provision that are closer to the familialistic models of southern Europe than to the ‘modern’ models of Scandinavia and north-western Europe.