Household and family during urbanization and industrialization: efforts to shed new light on an old debate

1. Early sociological visions

Frédéric Le Play (1855, 1872) was the first scholar who investigated shifts in family configurations in a systematic way, and he was one of the pioneering social scientists who argued that society was changing under the influence of urbanization and industrialization (Ruggles, 2012). In Les ouvrier Européens and Les ouvriers des deux mondes, studies which were based on empirical fieldwork of 132 families in different European, North-American, Asian, North-African and Asian countries, Le Play classified different types of families according to their organization, form and functioning (Wall, 1983). He focused on families as he thought that they were ‘the true social building blocks of society’ (Thornton, 2005a, p. 4). Le Play believed that differences in family forms through time and space were the result of diverging social and economic developments (Mogey, 1955; Thornton, 2005b).

Le Play distinguished between patriarchal families (what we would call today extended families), stem families and unstable families (nowadays referred to as nuclear families). He found patriarchal families to be widespread in non-Western countries. In these families, newly-wed sons moved with their spouses into the household of their parents and stayed under the authority of the head of the household. Family capital stayed undivided, the maintenance of the family line was guaranteed and customs and tradition were transmitted from one generation to the next. Stem families, which according to Le Play were found in many parts of Europe, shared most of these characteristics with patriarchal families, but differed regarding their form. In stem families, parents chose one of the children as their successor, usually the oldest son. This person stayed with his spouse and children in the household of the parents and perpetuated the family line, while siblings moved out upon marriage and formed new families and independent households. The third type of family, the unstable family – highly criticized by Le Play for its moral decay1 – was more or less exclusively found among the working classes in the urban industrial areas of the Western world.

Le Play argued that patriarchal families, which had been common in rural areas of Europe before the French and industrial revolution, were being replaced more or less completely by unstable families; the same was true for stem families, although on a smaller scale. The transition from patriarchal and stem families to unstable families was, in the eyes of Le Play, a consequence of the fact that families had less capital to hand down from one generation to the next, due to the shift to wage labor. Another factor, which according to Le Play drove the shift in family configurations, was the introduction of equal inheritance among offspring, which was established in France in 1793 (Desan, 1997). Declining family capital and equal inheritance in combination with massive rural-to-urban migration led to a decline in authority among heads of households, and stimulated children to leave the household at an early age. Households broke up, family lines were not continued, moral decay occurred and parents were left alone by the time they were in need of care from their children.

Many of the founding fathers of sociology, including Ėmile Durkheim, Ferdinand Tönnies, Georg Simmel and Max Weber, during the next decades stressed that because of urbanization and industrialization, a transition from multigenerational to nuclear families took place in nineteenth- and early twentieth-century Europe, and a diametric distinction was drawn between preindustrial rural and urban industrial societies. Preindustrial rural societies (what Tönnies referred to as Gemeinschaft) were characterized by extended families and close social and economic relations continued to exist among family members throughout the individual's life course. Religion, as well as customs and habits, regulated social life in the village, and solidarity among villagers was essential (Jackson, 1997). Urban industrial societies (what came closest to what Tönnies described as Gesellschaft) were, by contrast, characterized by nuclear families and weakened kinship ties, due to an increase in the physical distance among kin members, as a result of migration, and economic needs took precedence. Cities were perceived as places where life was faster, more organized, more bureaucratic and where, contrary to villages, anonymity, chaos, loneliness and confusion were prevalent. Solidarity as well as social control were weaker in cities than in villages and the formation of social networks was harder in the urban environment. City life was more individualistic and the construction of a personal identity became necessary. Last, but not least, the influence of religion and religious customs and habits declined (Liang, 2008).

In the United States, scholars of the Chicago School of Sociology started to theorize further on the relationship between industrialization, urbanization and family and population change. This led to the formulation of the so-called theory of social breakdown, which is, in essence, twofold (Hareven, 1982). On the one hand, industrialization led to a transition from extended to nuclear families, as families consisting of one conjugal unit with children were believed to have fitted the requirements of industrial life better than families living with three or more generations under one roof (Burgess & Locke, 1945).2 On the other hand, rural-to-urban migration led to the disintegration of the family unit and the erosion of kinship networks. Stripped of ties with family members and village culture and habits, rural-to-urban migrants became uprooted upon arrival in a new city. The lack of a social network caused, according to the adherents of the Chicago School of Sociology, all kind of social evils, including alcoholism, extra-marital births, prostitution and crime (Park, 1928; Park & Burgess, 1925; Thomas, Park, & Miller, 1971 [1921]; Thomas & Znaniecki, 1918; Wirth, 1928).

2. The null hypothesis and the Hajnal line

After a century of theorization, most sociologists agreed that towards the middle of the twentieth century, urbanization and industrialization had weakened family ties and had brought about a large-scale shift from extended to nuclear families in the Western world, as the latter were better fitted for urban life, and because many of the functions of extended families had become obsolete, as a result of ongoing economic specialization (Ruggles, 2012). Moreover, the idea that industrialization and urbanization were the big watersheds in the social history of the modern Western world which had driven most of the social and demographic change in the previous decades, had become deeply embedded. Few people at the time must have imagined that some discoveries in the near future would seriously undermine this picture of the past.

In the early 1960s some sociologists, among them Sidney Greenfield (1961) and William Goode (1963), started to seriously call into question that the prevalence of nuclear families in the Western world was a ‘functional consequence of the urban-industrial revolution’ (Greenfield, 1961, p. 312). However, the big blow to this paradigm would come from the historical sciences, in which, under the influence of the Annales School in France and The American Economic History in the US, a quantitative turn was taking place. As a result, historians started to work increasingly on family and population issues. The seminal work on family reconstitution by Henry (1967) encouraged historians to systematically explore, collect and analyze quantitative historical records on family and demographic issues in the past. In light of this, Peter Laslett started to explore nominal census material. To his own surprise he discovered that in the seventeenth-century English village of Clayworth most inhabitants lived in nuclear families, while he had expected, on the basis of the sociological studies by Le Play and others, to find many extended families that included family members belonging to different generations and various conjugal units within one single household. This unexpected finding signaled the start of large scale data research on family configurations in the past, in which Laslett himself played a key role. In 1969 he organized a conference at the Cambridge Group for the History of Population and Social Structure, in which a large number of scholars presented research on family structure in the past in a wide range of European countries, as well as North America and Japan. The findings were later brought together by Peter Laslett and Richard Wall in the book Household and Family in Past Time. In the introduction to the book Laslett described Le Play as the ‘central architect of the myth that families had been extended in the past and had changed to nuclear as a result of the transition from a rural agricultural to an urban industrial society’ (Thornton, 2005a, p. 2). Against this ‘myth’, Laslett set the ‘continuity’ or ‘null hypothesis’, stating that on the basis of the assembled evidence in the volume, the nuclear family had been the dominant family type in England and Europe before, during and after the industrial revolution, and that the mean household size had stayed more or less unaltered in the previous centuries (Laslett, 1972).

Laslett's hypothesis was reinforced by Hajnal (1965, 1982) finding that Europe, west of the line St. Petersburg–Trieste, had been dominated by a unique marriage pattern, of late marriage and high proportions of life-time singles, in the centuries before WWII. This unique marriage pattern was according to Hajnal primarily based on the principle of neo-locality and facilitated by the system of life cycle service. While elsewhere in the world newly-wed couples moved into the household of the parents of the groom or the bride, Western European couples set up independent households. In order to do so, they had to be financially independent, a moment which most people reached only at an advanced age. Moreover, a considerable proportion of the population never reached financial independence at marriageable ages. Hajnal thereby alleged that families had been nuclear in Western Europe in the previous centuries, long before urbanization and industrialization had begun. Later, some scholars reversed cause and consequence, by stating that the Western European marriage pattern paved the way to industrialization (Macfarlane, 1986, 1987; Hartman, 2004).

3. Refinements, revisions and alternative views

Although Laslett and Hajnal are widely praised for their contributions to historical demography, critics have emerged that offer other insights. In the case of the Western European marriage pattern, some interesting alternative explanations have been suggested regarding, for example, parental authority (Klep, 2005) and the attractiveness of single life (Guinnane, 1991). Contributions by Faragó (1998), Kaser (2001), Mitterauer (2003) and Szołtysek (2007), among others, have shown that Central Europe had much more variation in marriage patterns than has been suggested by Hajnal. Other scholars showed that even within Western Europe there was much more diversity in marriage patterns (Benigno, 1989; Kertzer & Hogan, 1989; Ortmayr, 1995). Moreover, research by Katherine Lynch (1991, p. 83) has shown that urban areas were characterized by ‘an exaggerated version of the Western European marriage pattern’, mainly because of the high turnover of migrants. Last but not least, comparisons with other world regions led to new insights and to the question of how unique the Western European marriage pattern, in fact, is (Engelen & Wolf, 2005; Engelen & Puschmann, 2011).

Regarding the null hypothesis, it became clear that in many regions of Europe, the nuclear family had not been the sole family form in the past, and that in some areas, more complex families had existed; in particular, stem families, which proved that Laslett's ‘crusade’ against Le Play was undeserved. Laslett (1983) himself revised his older vision on the basis of new empirical evidence and divided Europe in four different zones. According to this revised view, only Northwestern Europe had been dominated by nuclear families. Others refined this scheme again or suggested different classifications. Fauve-Chamoux (1995) proved, for example, the existence of the stem family in the Pyrenees, and more recently showed that stem families actually existed and still exist in many Asian societies (Fauve-Chamoux & Ochiai, 2009)3. Kertzer and Hogan (1989) suggested, on the basis of evidence from Casalecchio, that the Mediterranean pattern suggested by Laslett (1983) was in need of revision and that the relation with industrialization and urbanization was more complex than expected. With regard to the US, Tamara Hareven (1982) found that large numbers of French Canadian families, in the textile city of Manchester in the first three decades of the twentieth century, had been extended, and that their move to an industrial city had neither destroyed family solidarity nor their multigenerational form. Emmanuel Todd (1990) came up with a new pre-industrial family typology, which he mapped for Europe. Todd's work was, however, soon heavily criticized, and Richard Wall and Antoinette Fauve Chamoux, being unable to present a better classification, proposed to simply give up the challenge of classifying and mapping family configurations for pre-industrial Europe (Dribe, Manfredini, Oris, & Ritschard, 2010).

Much of the criticism on the work of Laslett focuses on measurement issues (Ruggles, 2012). As families and households evolve over time, longitudinal source material is better suited for the study of family configurations than the cross-sectional records used by Laslett. Richard Wall (2009) showed that Laslett's method was unable to detect a majority of the families, which were identified by Le Play as stem families. The reason for this is that a family might be nuclear at a given moment in time when a census was taken, but might turn into a multigenerational household at a later point, once one of the children marries and moves with the spouse in the parental household. Moreover, with time multigenerational families break up, which means that nuclear families in the census might have been stem or extended families in the past. Next, there are simple demographic explanations for the low prevalence of multigenerational families: because life-expectancy in the past was low and both age at marriage and age at first childbirth were high, the time that more than two generations of family members lived in one household would have been brief. This, however, does not mean that all previous generations in the Western past preferred the nuclear family type, but rather that most of them lived a majority of their lives in families consisting of only two generations of family members, because of the demographic constraints.

What is more problematic is that Laslett successfully dismantled the sociological paradigm that industrialization and urbanization had changed family life, but did not replace it with any new theory, nor did he give guidelines for future research (Janssens, 1993). Some scholars at a session dedicated to the fortieth anniversary of the book Household and Family in Past Time at the 2012 SSHA meeting in Vancouver even stated that the absence of any concluding chapter in the book seems to have stunted historical family demographers to a certain degree. There might be some truth in that. After all, apart from all kinds of interesting case studies, no major large-scale attempts to study the impact of industrialization and urbanization on family configurations, marriage patterns and family formations for larger parts of Europe, have been conducted during the past decades (Ruggles, 2012; Szołtysek, Gruber, Zuber-Goldstein, & Scholz, 2011). This shortcoming might also be related to a shift from the family or the household to the individual as main object of analysis. This is a result of the successful advent of the life-course paradigm and the introduction of event history analysis to the field of historical demography (Kok, 2007).

4. The special section

This special section is an effort to shed new light on the old debate about the impact of urbanization and industrialization on family configurations, marriage patterns and family formation by presenting new empirical results for areas in Europe, which have previously been largely neglected. The contributors to this special section make use of cross-sectional and longitudinal sources, and they apply both old and new methods. While existing historical demographic studies on family configurations, marriage patterns and family formation have largely focused on rural populations and on natives (Lynch, 1991), the contributors in this special issue focus instead mainly on migrant populations in medium and large urban areas. After all, Lucassen and Lucassen (2009) have proved that in the heyday of the industrial revolution, a strong increase in migration, especially rural-to-urban migration, took place in Europe. The driving force behind this special section is the belief that the impact of urbanization and industrialization on family configurations, marriage patterns and family formation, must have been larger than suggested by previous empirical studies; and it is our aim to re-open this old, inspiring debate. We will now present a short overview of the contributions to this special section, and we will conclude this introductory article by suggesting some new potential lines of research.

The first contribution to this special section is written by Rolf Gehrmann and is titled German towns at the eve of industrialization: household formation and the part of the elderly. This study is based on recently discovered census material from the states of the German Customs Union (GCU). Gehrmann used a sample from the 1846 census lists, consisting of 11,131 individuals from 11 different towns. The paper shows that there were clear differences in the age of leaving home and the mean age at first marriage between towns in the western and eastern part of the GCU. In the eastern part, women left the family of orientation early on, married at a relatively young age, and they resided more often in nuclear families. Moreover, in the east, more or less all women married and fertility was high. In towns in the western part of the GCU, by contrast, women left home relatively late, married at an advanced age, and there was a considerable proportion of life-time singles. Stem families prevailed and fertility was restrained. Next, Gehrmann found a contrast between pre-industrial and industrial towns. The latter were characterized by a lower mean age at first marriage, especially among women. Moreover, it seems that industrialization and nuclearization were highly related, as industrial towns showed a higher prevalence of nuclear households, underlining Le Play's traditional point of view that industrialization transformed stem families into nuclear families. Gehrmann also found significant differences between the urban and rural population in Germany, especially for men. In urban areas, men left home earlier and married at a younger age. At the same time the risk of definite celibacy was higher in towns, and the odds of finding elderly kin members in the household was lower. This leads Gehrmann to conclude that at least in Germany, modernization fostered a transition towards nuclear households.

The article with the title ‘The true social molecule’ Industrialization, paternalism and the family. Half a century in Le Creusot (1836–1886) is written by Jérôme Bourdieu and Lionel Kesztenbaum. The authors aim to understand how family formation took place and evolved in Le Creusot, a village in the heart of France, which as a result of rapid industrialization and strong rural-to-urban migration from surrounding areas, turned into a booming town in the course of the nineteenth century. Bourdieu and Kesztenbaum observe how families were formed and dissolved over time, studying intergenerational co-residence from 10 tightly linked censuses with five-year intervals. Le Creusot was dominated by the Schneider company, which was led by a disciple of Le Play. The Schneider company brought a paternalistic family policy into practice – promoting the extended family – by controlling access to the labor market, housing and schooling. However, this family policy could not hinder the fact that industrialization and rural-to-urban migration drove family formation into an opposite direction. Accordingly, people started to marry and have children at older ages. This was especially true for first generation rural-to-urban migrants. Moreover, intergenerational co-residence declined, although the proportion of elderly in town was on the rise.

While the studies by Gehrmann (2014) and Bourdieu & Kesztenbaum (2014) deal with the impact of urbanization and industrialization on family configurations, marriage patterns and family formation in a general way, the contribution to the special section by Paul Puschmann, Per-Olof Grönberg, Reto Schumacher and Koen Matthijs (2014), focuses exclusively on the impact of industrialization and urbanization on marriage and family formation among migrants. From a functionalist point of view, sociologists believed that urbanization, migration and nuclearization went hand in hand, as nuclear families were more geographically mobile than more complex family types. At the same time, scholars of the Chicago School of Sociology believed that the adaptation process of rural-to-urban migrants was highly problematic, and that migrants ended up on the edge of urban society, because they lacked a social network. However, studies by Anderson (1971), Hareven (1982) and Janssens (1993) suggest that extended families persisted in the urban industrial environment, that extended families offered the best shelter for rural-to-urban migrants in industrial cities, and that they facilitated urban integration. At the same time, other studies suggest that integration was not as difficult as has been suggested by the Chicago School of Sociology (Sewell, 1985; Lucassen, 2004). However, it is true that migrants married later and less than the urban native population, which has often been regarded as a sign of bad integration (Lynch, 1991; Moreels & Matthijs, 2011; Oris, 2000)

In light of this, Puschmann, Grönberg, Schumacher and Matthijs (2014) studied access to marriage and reproduction of domestic and international migrants, who moved as singles to Antwerp and Stockholm in the latter half of the nineteenth and the early twentieth century. The authors make use of the Antwerp COR* database (Matthijs & Moreels, 2010), and the Stockholm historical database (Geschwind & Fogelvik, 2000). Both databases consist of highly reliable longitudinal information, which allows researchers to ‘reconstruct’ (parts of) the life courses of migrants. In their analyses the risk of first marriage and first child birth function as indicators of the process of social inclusion. They focus on Antwerp and Stockholm, as these cities experienced major differences in terms of industrialization. In the course of the nineteenth century, Antwerp developed into a port city, without a noteworthy industrial sector. The opposite occurred in Stockholm, as the Swedish capital turned into an industrial hotspot. While the authors expected that industrialization would create niches in the urban labor market, and would create opportunities for migrants to marry and form families, the results of the discrete time event history analyses tell a different story. The chances of social inclusion, in terms of access to marriage and reproduction, were higher for migrants in Antwerp than in Stockholm. With industrialization expanding in Stockholm, the odds of getting married and starting a family became lower for migrants. In Antwerp, where no major industrialization process took place during the period of study, remarkably enough, the odds of social inclusion became higher. The study of Puschmann, Grönberg, Schumacher and Matthijs shows once more that the relation between industrialization and family formation is far from straightforward.

5. Suggestions for future research

While many of the old questions regarding the impact of urbanization and industrialization on family configurations, marriage patterns and family formation have not been answered (conclusively), new questions have arisen. We still aim to get a better understanding of how family types, marriage patterns, and family formation changed through time and space. Ideally this should lead to the creation of maps which include changes over time. Moreover, the role of urbanization and industrialization, as well as other factors, has to be re-assessed. It is necessary to identify and explain differences between countries, regions and cities and between different ethnic, cultural and socio-economic groups. Next, we want to know under which specific circumstances extended and stem families disappeared and marriage and family formation were advanced. Factors such as occupational structures, the availability of wage labor, public education, female participation at the labor market, housing supply, living standards and norms and values, will have to be taken into account. At the same time, it is necessary to understand how it was possible that, in certain industrial cities, extended families persisted or even increased in number (Ruggles, 1987). What was the role of migrants in the persistence of extended families? Did they increase the number of extended families as they moved into households of kin, and why did migrants have less access to marriage and reproduction? Was there a certain social breakdown and did migrants become uprooted? Were certain groups of migrants excluded from the marriage, labor, and housing markets? And if so, why was this precisely the case for these specific groups? Did they bring less human capital with them in terms of education and skills, or did they face discrimination because of cultural or ethnic differences?

While gender has become an important variable in historical demography and family sociology, it is striking that changes in family configurations, marriage patterns and family formation systems have hardly been studied from a gender perspective. Existing studies suggest that females might play a leading role in changes regarding family life. Tine De Moor and Jan Luiten van Zanden (2010) suggested that it was no coincidence that the Western European marriage pattern came into being at a moment when the legal and social position of women in Western European societies was improving. More recently, the Arab countries also witnessed a strong rise in ages at marriage, while the rights and opportunities of women are on the rise (Engelen & Puschmann, 2011). Moreover, Matthijs (2002) demonstrated that women played a major role in the weakening of the Western European marriage pattern in Flanders towards the middle of the nineteenth century. He found that mean ages at first marriage declined earlier among females than among males, which he explained in terms of the relative deprivation of women. Females were increasingly denied labor market opportunities, and, as a consequence, started to focus more on the home and family life.

Apart from providing the reader with some substantive suggestions, we also want to make some methodological recommendations. One of the weaknesses in the existing research, and also in the contributions to this special section, is the way in which industrialization and urbanization are operationalized. Instead of measuring industrial development, towns, cities, and whole areas are classified as being industrial or non-industrial, or as going through a process of industrialization or not. Something similar is true for the distinctions made between urban and rural geographic entities. Instead of measuring the degree of urbanization, places are classified as villages, towns, or cities, without taking the number of inhabitants or the size and amount of urban functions into account. In order to make more far-reaching statements about the relationship between urbanization and industrialization, on the one hand, and family configurations, marriage patterns and family formation, on the other hand, it will be necessary to move away from case-studies. By doing so, researchers can take into account larger numbers of geographic entities, for each of which, measures of urbanization and industrialization are also addressed through use of multilevel analyses. Soon, it will become perfectly feasible to study, for example, family formation on the individual level for different regions in Europe, while taking into account measures of industrialization and urbanization on the municipality level with the help of census information (especially also occupational and industrial census material). Additionally other sources, such as passed down statistical accounts of municipalities, governmental departments, industrial companies, post services, railway companies, etc. could be used. In the case of industrialization we can think about measures like the share of industrial workers in the total working population, the number of industrial plants, steam engines or industrial machines present in a municipality. With respect to the degree of urbanization, we can think about the number of inhabitants, population density, urban size in hectares, the presence of one or more railway stations, the number of post offices or the number of letters sent on a yearly basis, the availability of schools and universities, and so on. Research on the impact of industrialization and urbanization on social mobility has already moved in such a direction, and the first fruitful results have recently become available (Van Leeuwen & Maas, 2010; Zijdeman, 2010; Schulz, 2013).

Another interesting line of research about the study of life course trajectories can be explored with the help of event history analysis and sequence analysis. An application of event history analysis to marriage and family formation is presented in this special section by Puschmann, Grönberg, Schumacher and Matthijs (2014), but many other interesting applications are possible. It can be assessed whether children in places where urbanization and industrialization took root, moved more often and at earlier ages out of the household and whether their risk of becoming lifetime singles was lower. Likewise it can be tested whether men in industrial cities had better chances of becoming a head of the household and whether they reached this stage earlier in life, compared to men living in the countryside or in non-industrial cities. Moreover, it has been suggested that families in urban industrial places were either pioneers in the fertility transition or latecomers (Oris, 1996; Eggerickx, 2001). The question is whether urban migrants brought innovations in terms of reproduction, or that they adapted their fertility behavior to that of the native population. This can be assessed by modeling spacing and stopping behavior of different groups of migrants and natives, by focusing on birth intervals and the age at last birth (Moreels, Vandezande, & Matthijs, 2010; Moreels & Vandezande, 2012). Next, the likelihood that elderly dependent family members moved in urban industrial households can be modelled, as well as the risk of having living-in lodgers or boarders (Ruggles, 1987). Additionally, it can be investigated whether urban families in industrial cities were more mobile than families in rural areas and in non- or pre-industrial towns, by modeling the risk of out-migration and/or residential mobility. An interesting study by Kok, Mandemakers, and Wals (2005) found that the urban poor moved very often within their own district, as this was part of a coping strategy. The question is whether this is typical for industrial cities or that comparable strategies existed also elsewhere among the poor.

With the help of sequence analysis, it is possible to assess whether the sequence and duration of different statuses (e.g. living with parents and siblings in the family of orientation, being a head of a single household, living with a spouse without children, living with a spouse and children, etc.) in the life course changed over time, as a consequence of urbanization and industrialization. Sequence analysis is a relatively new, but promising, technique which has already led to some very interesting results (Dribe et al., 2010; Schumacher, Matthijs, & Moreels, 2013).

Whatever type of approach is chosen to study family configuration, marriage patterns and family formation in times of urbanization and industrialization, researchers will profit from the ever-increasing amount of data which is becoming available through, among others, IPUMS4, the North-Atlantic Population Project5, the Mosaic Project6 and the European Historical Population Samples Network7.

Acknowledgements

We are grateful to the Max Planck Institute for Demographic Research (MPIDR) and the Scientific Research Network Historical Demography (Research Foundation Flanders, FWO) for financing the meetings of the Urban demography network in Rostock (Germany) and Leuven (Belgium), which paved the way for this special section. Moreover, we would like to thank Siegfried Gruber (Max Planck Institute for Demographic Research, Rostock), Jan Kok (Radboud University Nijmegen/KU Leuven), Koen Matthijs (KU Leuven), Andrejs Plakans (Iowa State University) and Mikolaj Szołtysek (Max Planck Institute for Demographic Research, Rostock) for their useful comments and suggestions. Last but not least, a word of thanks goes to Robyn Donrovich (KU Leuven) for the language editing.

Notes

^1. According to Le Play moral decay was a result of the advent of unstable families, as religion and tradition in patriarchal families had been handed down from one generation to the next, from father to son. The dethronement of the head of the household had led to a situation in which religion and tradition were no longer passed down the line as families broke up.

^2. Ideas which were later further elaborated by Parsons (1963).

^3. In the case of Japan, the existence of stem families was already known for a long time.

^4. http://www.ipums.org/

^5. https://www.nappdata.org/napp/

^6. http://195.37.34.109/cgi-bin/index.plx

^7. http://www.ehps-net.eu/