Jump to ContentJump to Main Navigation
Similarity in DifferenceMarriage in Europe and Asia, 1700-1900$

Christer Lundh and Satomi Kurosu

Print publication date: 2014

Print ISBN-13: 9780262027946

Published to MIT Press Scholarship Online: May 2015

DOI: 10.7551/mitpress/9780262027946.001.0001

Show Summary Details
Page of

PRINTED FROM MIT PRESS SCHOLARSHIP ONLINE (www.mitpress.universitypressscholarship.com). (c) Copyright The MIT Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in MITSO for personal use (for details see www.mitpress.universitypressscholarship.com/page/privacy-policy). Subscriber: null; date: 16 January 2019

Prudence as Obstinate Resistance to Pressure

Prudence as Obstinate Resistance to Pressure

Marriage in Nineteenth-Century Rural Eastern Belgium

Chapter:
(p.261) 8 Prudence as Obstinate Resistance to Pressure
Source:
Similarity in Difference
Author(s):

Michel Oris

George Alter

Paul Servais

Publisher:
The MIT Press
DOI:10.7551/mitpress/9780262027946.003.0008

Abstract and Keywords

The nineteenth century challenged the adaptive capacities of rural population in eastern Belgium. Families succeeded in maintaining local cultures and ways of life by imposing strict controls on marriage, a traditional response that intensified during the century. This chapter considers family control and escape from that control among the unmarried aged 18–44 in two rural communities. We focus on the effects of household composition and local contexts on their uneasy transition to adulthood, considering also partner choice and premarital sexuality in a social and demographic system under pressure. Even more than others, farming families avoided marrying during periods forbidden by the Catholic Church, their intervals between marriage and first birth were longer, illegitimate births were rarer and their expectations about marriage were higher.

Keywords:   Marriage, rural families, partner choice, sexuality, Belgium, migration

The nineteenth century challenged the adaptive capacities of rural populations in eastern Belgium. Both economic and demographic changes upset the balance between population and economic resources, and families and communities actively used marriage and migration to counter Malthusian pressure. The growth of factory production caused the collapse of proto-industry, which was widespread in eastern Belgium, and the last pockets of subsistence farming were converted to commercial agriculture. At the same time the demographic trends pointed toward faster population growth. Mortality was decreasing, and fertility rose slightly after 1850 and did not begin to decrease until the last quarter of the nineteenth century (Alter, Neven, and Oris 2004; Alter, Neven, and Oris 2010). Rural families succeeded in maintaining their local cultures and ways of life by imposing strict controls on marriage, a traditional response that intensified during the nineteenth century. The Industrial Revolution also offered a more modern demographic response to population pressure in the form of migration to rapidly growing industrial cities. This chapter considers family control and escape from that control among the unmarried adults aged 18 to 44 in two rural communities. We focus on the effects of household composition and local contexts on their uneasy transition to adulthood.

In the first section, we describe the transformation of the eastern Belgian countryside in the nineteenth century, and in the next section, we show how access to marriage was a key element in the management of tensions between growing populations and decreasing resources. Thereafter, the data and samples used to analyze the determinants of marriage and outmigration are discussed. The following three sections place the Eurasian hypotheses in the east Belgian context, present the results, and examine the additional constraint of partner choice. Finally, (p.262) we explore the interrelations between control of sexuality and control of marriage and the use of premarital pregnancy in parent—child negotiations over the timing of marriage.

Rural Worlds in Transition

Our research focuses on two areas located in eastern Belgium: the municipality of Sart in the Ardennes region and three villages (Charneux, Clermont, and Neufch â teau) situated in the heart of the Pays de Herve. Despite the small distance between them, the Ardennes and the Pays de Herve have different geographic characteristics, as well as different agrarian, demographic, social, and economic histories. Moreover all these localities were less than 20 kilometers from the rapidly growing textile factories in Verviers and less than 40 kilometers from the mines and mills around Li è ge. Rural and urban were interdependent in nineteenth-century eastern Belgium, and the populations of the Ardennes and the Pays de Herve could not have developed as they did without opportunities for migration to nearby urban and industrial places (Neven 2003; Oris and Alter 2001).

Prudence as Obstinate Resistance to PressureMarriage in Nineteenth-Century Rural Eastern Belgium

Map 8.1 Locations of Sart and Pays de Herve

(p.263) The divergent experiences of our two case studies are due to very different ecologies. The municipality of Sart consisted of half a dozen hamlets located on the northern slopes of a high plateau of peat bogs and forests called the “Hautes Fagnes.” In the eighteenth century, iron forges used charcoal from these forests, but they disappeared after coke-based forges developed in the Li è ge basin. Most of the inhabitants were smallholders or renters of middle-sized farms. They depended heavily on the area’s abundant communally owned forests for wood and forage for livestock, as well as slash-and-burn agriculture, an archaic practice that continued into the nineteenth century. An 1847 Belgian law encouraged the sale of common land and the formation of large estates, but large farms were not economically viable in the Ardennes (Hoyois 1981: 154–76). In the second half of the nineteenth century, more progressive agricultural practices emphasizing livestock were introduced, and timber was sold to the coal mines around Li è ge (Alter and Oris 1999b).

The Pays de Herve, a plateau of rolling meadows and orchards between the Vesdre valley and the Dutch border, had a quite different history. During the sixteenth century, merchant-clothiers from Verviers recruited farmers to spin and weave wool for sale abroad. Involvement in the emerging modern economy also stimulated an early transition from subsistence farming to commercial agriculture based on cattle breeding and the production of cheese and fruits. Hervian farmers prospered under this proto-industrial system. Landownership in creased, and common lands disappeared (Servais 1982). When spinning machines reached Verviers in the early nineteenth century, this system was suddenly destroyed. Rural spinning had disappeared by 1830, and handloom weavers experienced the same fate around 1860. By 1860 the decline in peasant property ownership had reached its nadir, and most farmers were renting land from noble or bourgeois owners. The modernized agriculture of the Pays de Herve remained profitable and dynamic during the second half of the nineteenth century because the demand from the urban markets rose continuously (Neven 2003).

The divergent economic conditions in the two study areas were reflected in their population histories. The Pays de Herve as a whole responded to the demise of proto-industry with a 10 percent population loss between 1800 and 1830. Between 1805 and 1831 the populations of Charneux and Clermont decreased by 16 and 11 percent, respectively. During the next twenty years, the decline continued at a slower pace. In 1850, 2,126 people had their legal residences in (p.264) Charneux (as opposed to 2,732 in 1805) and 2,095 in Clermont (as opposed to 2,424 in 1805). In the middle of the nineteenth century, under the influence of economic reorganization, the demographic curves of the two localities stabilized until 1890 (Neven 2003: 112). In Sart the traditional economy strained to support an increase from 1,815 inhabitants in 1812 to 2,541 in 1851. Between 1851 and 1900 the population of Sart fell back to its starting level. These population movements were not determined by the “natural” difference between fertility and mortality, but by extensive migration in a dynamic regional economy.

The stressful process of ruralization of the countryside had beneficial effects on mortality, especially in comparison with the “epidemiological depression” occurring in industrial towns. After 1850 an increase in the standard of living severed the Malthusian link between prices and mortality in both the Hervian localities and Sart (Alter, Neven, and Oris 2004: 181–86, 205). As cottage industry was marginalized, the contact between the countryside and the local urban centers diminished, and more or less balanced exchanges transformed into unidirectional movements from countryside to town. Communities in the Ardennes and Pays de Herve escaped from the Malthusian “positive check” (mortality), which appears when the number of people to be nourished exceeds the available resources. Indeed these localities achieved more than eliminating crisis mortality. The life expectancies in the Sart and Hervian municipalities were among the highest in Belgium. Between 1812 and 1900 the expectation of life at birth in Sart was around 42 years, with an increase of one year between the periods 1812 to 1846 and 1847 to 1866. By 1880 the life expectancy had advanced to almost 50 years, and remained at this level during the 1880s (Alter, Neven, and Oris 2004: 182). Better off than the Ardennes, the three Hervian localities attained expectations of life at birth of 48.4 and 49.7 years for men and women, respectively, between 1847 and 1900 (Neven 2003: 640). This favorable trend contrasted with an epidemiological depression affecting industrial areas in the third quarter of the nineteenth century. The explosive growth of cities near the rural sample locations caused deteriorating sanitary conditions (Oris and Alter 2001). For example, the life expectancy in the coal-mining city of Tilleur, a suburb of Li è ge, was only 38 years during the period 1847 to 1880.

The cruel paradox is that the favorable conditions in the countryside depended on the outmigration of surplus population to those dangerous cities. In the second half of the nineteenth century, outmigration gradually reduced the population of Sart to 2,091 inhabitants in 1900. (p.265) This net loss of 450 persons hides a much larger outflow because births outnumbered deaths by around 1 percent of the population each year. A similar flow of outmigrants occurred in the Pays de Herve.

The Heart of Rural Resistance: Access to Marriage

Until 1890, surplus labor left the countryside for the industrial towns. The increase in the life expectancy during this period confirms that outmigration was an important safety valve, but it did not relieve the constant pressure on those who stayed. Residents of Sart and the Pays de Herve did not marry earlier in the second half of the nineteenth century, and the proportion of definitive celibacy remained exceedingly high. Access to marriage remained extremely restricted in the second half of the nineteenth century, contrary to what happened in industrial areas, where the Malthusian “preventive check” was weak after the middle of the century (Oris 2000).

We measure the extent of Malthusian pressure with the Coale index of the proportion of married women among women aged 15 to 49 (Im). Figure 8.1 presents annual estimates of Im, showing that the conditions became increasingly less favorable for marriage for women living in Sart. Im declined from 0.491, which was already a relatively low level, to 0.391 between 1816 and 1831. Then the preventive check was relaxed for about twenty years until a population decline began in the 1850s. The next period was characterized by stability, and Im stayed at around 0.400. Finally, there was a brief accentuation of control during the late 1880s and early 1890s before nuptiality began to recover at the end of the nineteenth century (Alter and Oris 1999a: 140). The Coale index of the proportion married was even lower in the Pays de Herve, and the trend was clearly declining. The proportion married was highest at the beginning of the series in the middle of the nineteenth century (0.360), but its level was already very low at this time. Im fell until 1858 to 1860 (0.310) before a period of recovery in the 1860s. Compared with nineteenth-century data from provinces across Europe reconstructed by the European Fertility Project, the Hervian values of Im are among the lowest ever observed. Only two Scottish counties and the poorest cantons of Central Switzerland show lower levels of marriage. Very low values of Im (0.320 to 0.340) can also be found in a few Finnish, Irish, and Flemish provinces (Coale and Treadway 1986; Watkins 1986), which were among the poorest rural areas of Europe at that time. The Pays de Herve, however, was a relatively wealthy agricultural area, as (p.266)

Prudence as Obstinate Resistance to PressureMarriage in Nineteenth-Century Rural Eastern Belgium

Figure 8.1 Index of proportion married (Im) in Sart and Pays de Herve, 1812–1900

Sources: State Archives in Liege, Municipalities (Population Registers) and Civil Registration

we explained in the preceding section, but its inhabitants maintained extremely strict control over the access to marriage.

We can conclude that in both Ardennes and the Pays de Herve, families and collectivities desperately resisted the economic and demographic threats to their way of life. Indeed, while industrial towns offered an escape, those who stayed were subject to extremely strict control over the formation of new households through marriage. Not surprisingly, the ages at first marriage were very high.

In Sart the average age at marriage was about 27 for women and 29 for men during the studied period (see figure 8.2). The marriage system in the Pays de Herve also appears to have been extremely restricted. Among males, the average age at first marriage was 29 years around 1850, and then increased gradually to fluctuate between 30 and 31 between 1855 and 1885. In the last fifteen years of the nineteenth century, it decreased again to below 30 years of age. The average marriage age for Hervian females evolved in the opposite way. It was higher at the beginning and the end of the period. In 1864, when the ages at marriage were the lowest, the average woman waited until the age of 27.5 for her first marriage. Late marriage was combined with nonmarriage: 21.4 percent of the women in the Pays de Herve and 15.4 percent in the Ardennes never married.

The importance of late marriages and a high level of celibacy for the control of population growth is demonstrated in table 8.1. The total (p.267)

Prudence as Obstinate Resistance to PressureMarriage in Nineteenth-Century Rural Eastern Belgium

Figure 8.2 Mean age at first marriage by sex in Sart and Pays de Herve, 1814–1898.

Note: Five years moving average.

Sources: State Archives in Liege, Municipalities (Population Registers) and Civil Registration

Table 8.1 Total fertility rate and total marital fertility rate above age 20 in Sart, Pays de Herve, and Tilleur, 1812—1899

Period

Sart

Pays de Herve

Tilleur

TFR

TMFR 20+

TFR

TMFR 20+

TFR

TMFR 20+

1812–1819

4.6

7.9

1820–1829

4.4

7.7

1830–1839

4.2

7.8

1840–1849a

4.8

8.5

3.9

8.4

4.8

6.8

1850–1859

4.9

8.9

3.6

8.2

5.5

8.0

1860–1869

4.7

8.8

3.9

8.9

5.9

7.9

1870–1879

4.7

8.9

4.0

9.2

5.5

7.4

1880–1889

4.4

8.5

3.8

9.0

1890–1899

3.3

6.7

3.1

8.1

1812–1899

4.4

8.2

3.7

8.7

5.6

7.7

Sources: State Archives in Liege, Municipalities (Population Registers) and Civil Registration.

Note: TFR = total fertility rate; TMFR 20+ = total marital fertility rate above age 20.

a. 1846–1849 for Tilleur and the Pays de Herve.

(p.268) fertility rate (TFR) shows the average number of births per woman.1 The average woman had more than four births in nineteenth-century Sart (4.4), somewhat fewer than four in the Pays de Herve (3.7), and more than five in the industrial town of Tilleur (5.6), which we include for comparison. The differences among these communities were primarily due to the timing of marriage, which was particularly late in the Pays de Herve and relatively early in a growing industrial place like Tilleur. When we look at fertility within marriage as measured by the total marital fertility rate above age 20 (TMFR 20+), the differences among these communities are much smaller, ranging from 8.7 in the Pays de Herve to 7.7 in Tilleur. In effect, the low level of marriage in the Pays de Herve reduced the overall fertility by about 5 children (3.7 versus 8.7), compared with a reduction of only 2 children per woman in Tilleur (5.6 versus 7.7).

This overview shows a demographic system in rural eastern Belgium under extreme stress. A balance between economic resources and population was successfully maintained, and even improved, despite the strong pressure on individuals, families, and collectivities resulting from high fertility and falling mortality. In a previous publication, we describe this as an illustration of Davis’s “multiphasic response” model. Birth control did not play a role until the end of the nineteenth century, not because communities were socially conservative but because they relied upon alternative responses to population growth (Alter, Oris, and Neven 2010). Successful adaptation was costly, because it was based on two undesirable alternatives: postponing marriage or leaving the location. The Malthusian preventive check (late marriage and definitive celibacy) played a key role in rural eastern Belgium by drastically reducing the average number of children per woman. Young couples had to overcome major obstacles before they were able to establish a new family. The alternative was to abandon the countryside for a new way of life in the cities.

Determinants of Marriage and Migrations: About the Data and Samples

In their introductory chapter to this volume, Lundh and Kurosu observe that “quite little could be said about which factors influenced the timing and incidence of marriage or about the way they interacted” (Lundh and Kurosu: chapter 1 of this book). The collective enterprise of the Eurasian Project is to highlight “the complexity of the (p.269) mechanisms behind individual and household decisions on marriage matters” (Lundh and Kurosu: chapter 1 of this book). Here we discuss the EAP hypotheses in the context of nineteenth-century eastern Belgium.

Nuptiality acted together with outmigration to control the population size at an aggregate level, but they also acted together at the household level, at which their respective roles may not be easily distinguished. In Sart, the Pays de Herve, and elsewhere in Wallonia, the nuclear family system was dominant, and the rule of neolocalism was followed. Marriage implied departure from the parental family: if not immediately, at least after a short time (see chapter 4 in this book, as well as Capron and Oris 2000: 237–43, 254, or Leboutte 1988: 184–85). In practice, some outmigration was directly linked to marriage. In cases in which individuals did migrate in order to marry and settle down elsewhere, marriage was not explicitly mentioned in the sources. In the population registers on which we rely, only the departure was mentioned, since the marriage took place in another locality and after the outmigration. This implies that marriage should not be studied as an isolated event but rather as a complement of the outmigration of unmarried people.

Our analysis includes marriage and migration as competing risks, which we have drawn from a combination of official documents. The Belgian population register has already been presented in the previous EAP volumes. Here we recall that registration was organized by household with (in principle) one page per domestic unit and one line per individual. Descriptive variables (name, age or date of birth, place of birth, matrimonial status, occupation, etc.) were listed in columns, and vital events were recorded during the period between censuses. Among those events are outmigration and marriage.

The most important imperfections in our sources affect marriages celebrated in other locations and external and internal moves. Under-registration of mobility has already been discussed in the EAP mortality volume (Alter, Neven, and Oris 2004; see also Alter, Devos, and Kvetko 2009). The recording of marriages is crucial for this chapter. Under the Belgian Civil Code all marriages were recorded in civil registers by a local official. In most cases those records were transferred to the population register, and we have linked the marriage registers to the population registers to capture any marriage dates not found in the population registers. This linkage, however, only applies to marriages that were celebrated and recorded in our study areas.

(p.270) Following the local custom, the wedding was usually celebrated where the bride resided. Several situations were then possible. First, if the groom was from the same place as the bride and they established a household in the municipality, the information is complete. Second, if the groom was from a different municipality and the bride from a study parish, we have some information about him and his parents from the marriage register in the sample area, but we know nothing about the groom’s household before the marriage or the marriage market in his village. If the new couple settled in Sart or the studied Hervian villages, we can follow their fate after the marriage. However, if the bride moved to her husband’s municipality after the wedding or they outmigrated together to a third location (e.g., an industrial town), they disappear from our sample. Consequently we are not able to observe the couple’s first birth to determine whether it was conceived prenuptially. The information about males from our sample villages who married women from other places may also be ambiguous. If the new couple settled in one of our study areas, their new marital status was recorded in the population register, and the date of the marriage is often given too. If they established a new household in a different location, however, we know only that the man left one of the studied areas (i.e., outmigration was recorded in the population registers), but we do not know whether he remained single in a new location or married in his fianc é e’s location.

Some outmigration seems to have been motivated by a desire to marry. Migrants from rural Wallonia arriving in Tilleur had an Im index of 0.670 between 1867 and 1880—double the value observed in our rural samples! The Im of these rural migrants is a little overestimated, because several newcomers declared themselves to be married on their arrival but recorded a marriage in Tilleur a short time later. Nevertheless, it means that some couples of rural origin married in the industrial towns, where they could escape the Malthusian tension between population and resources that acted as a brake on marriage (Oris 1995, 2000).

These limitations in the sources underscore the importance of studying marriage and outmigration at the same time. However, they also affect the subsamples available for specific kinds of analysis, which vary throughout the chapter. First, we will study all the marriages recorded among the single adults aged 18 to 44 whose life courses can be reconstructed from the population registers. For the reasons mentioned above, we observe more marriages for women than for men, respectively 1,110 and 889 in Sart and 1,096 and 790 in the three Hervian (p.271) municipalities. Second, when we look at endogenous versus exogenous marriages, we use unions celebrated in our study areas that are recorded in the civil marriage registers. Finally, when we look at prenuptial conceptions, we examine couples who stayed in our sample localities for at least eight months after the wedding. These restrictions are imposed by the sources, and it is important to keep in mind the selection processes behind them when interpreting the results.

Determinants of Marriage: Eurasian Hypotheses and the Eastern Belgian Context

In the tradition of the Eurasian Project, models are located at the crossroads of potential determinants at the individual, family, and macro levels (Bengtsson 2004a). At the individual level, age and sex are the fundamental demographic categories. For both Sart and the Pays de Herve, we study unmarried people aged 18 to 44, and we estimate separate models for men and women. In addition every person is located in a dynamic family and macro-level context. Without denying the importance of personal ambitions and individual initiative, we recognize that the family and the local community were responsible for imposing Malthusian constraints on individual behavior. Although difficult to measure, the importance of the family and of the interactions between household economic status and behavior are generally recognized. Nevertheless, Devolder (1999) also insists that we must take into account the role of the rural local community. He argues that the diffusion of the Western European marriage pattern was directly linked to historically specific agricultural conditions: “It is the rural community only that could prevent young people to marry and live on the peasant lands. The rural group forced young people to wait until a farm was available …” (Devolder 1999: 24). We will examine the potential pressure exerted by the local community on unmarried people in a mature Malthusian system with indicators of macro level demographic and economic conditions. We argue that the effects of those contextual variables on individuals were mediated by household composition. table 8.2 displays the means of the variables used in the regressions below.

First and foremost, the family environment depended upon the presence or absence of parents. Following the 1804 Napoleonic Civil Code, sons younger than age 25 and daughters under 21 needed the formal consent of their parents to contract a marriage (Servais 2005: 7–8). (p.272) Secular customs reinforced the parental role in marriage, while also taking into account the individual consent of young people, especially young women. Ethnographic surveys by the Mus é e de la Vie Wallonne in Li è ge have extensively documented nineteenth-century rural practices. When a young man wanted to start courtship, he went to the girl’s home, usually on Sunday evening, with a friend who carried a bottle of gin. After a few words, the friend asked the head of the family for permission to make a toast and demanded a glass. The father’ s acceptance was a clear sign that he was favorable. All those present toasted, drinking from the same glass, starting with the head of family and ending with the young men. The next time, it was the suitor who proposed a toast “to the health” of the young woman whom he was courting. If she was agreeable, she moved her chair beside that of the gallant, who presented the glass to her. Then she wetted her lips, and the suitor drank the rest of the glass. However, if the young woman politely refused, saying “thanks,” the significance was clear (Enqu ê tes … 1927–1930: 95–96, 103; Pinon 1978).

Although their power was not absolute, parents—especially the father—had a legal right to delay a marriage and a customary right to intervene in both the timing and the choice of spouse. Although the life expectancy was high in our rural samples, marriage was so late that only 40 percent of the unmarried people under the age of 25 lived with both a father and a mother who were still alive (Neven 2003: 245). The absence of one or both parents was likely to modify behaviors because gender balance was very important in the economy of rural households. For example, a mother living alone could prefer to keep an unmarried son in her household to replace the lost labor of her husband in the family economy. We have further refined the measures of the presence or absence of parents to account for the short-term effects of the death of a parent. The “present/absent” distinction is important, but it does not tell us whether an individual was orphaned at the age of 5 or at the age of 25. For this reason the first 3 years after the death of a parent are categorized separately to identify the immediate effect of a parent’s death on the young adult’s life course (Alter and Oris 1999a).

As with parental presence in the household, we are interested in the influence of siblings on individual marriage chances. Although family priorities varied across societies and might differ for marriage and outmigration, the impact of an individual’s position within the sibling group has been attested in historical populations (see Bras and Neven (p.273) 2007; S é galen and Ravis-Giordano 1994; Van Poppel et al. 2003). Transfers from parents to children trying to establish themselves in agriculture would have been strongly affected by the number of siblings who shared the same ambition. Moreover, at a time when settlement possibilities were limited and waiting times were long, we should expect that siblings differed in their expectations and experiences. Older and younger siblings faced different economic conditions within the family and had different obligations to their parents and to each other. For these reasons we use covariates taking into account sex and birth rank (i.e., older/younger) to detect differing life-course strategies among siblings.

As the economic context changed and opportunities to establish new households diminished, we expect that marriage had a strong socioeconomic dimension. Local custom in the study areas was based on the principle of neolocality, popularly expressed as“mariage demande m é nage” (a marriage requires a household). A new couple was expected to buy or rent at least a small house, to have the necessary household equipment (as part of the bride’s trousseau), and to hold either enough land for subsistence or an occupation to supplement it (Servais 2005). Given the sources, the best indication of a household’s socioeconomic status is the occupation of the household head. Occupations were rarely recorded for young women, and even the activities of young men are underrecorded. In Sart almost 20 percent of the households including an unmarried person aged 18 to 44 were headed by someone without a clear socioeconomic status (usually a widow), and this was the case for 15 to 17 percent of the domestic units in the Pays de Herve. These were the most vulnerable households, as were the 9 to 10 percent of day laborers and the 10 percent of textile workers still surviving in the Pays de Herve after the collapse of the proto-industry. In the middle of the socioeconomic hierarchy were the farmers, who were more dominant in Sart (56 percent) than in the Pays de Herve (40/45 percent), where various craftsmen constituted 12 to 15 percent of the households. The local elite consisted of the 9 to 10 percent of heads of household belonging to the petty bourgeoisie (mainly traders, white-collar workers, and a few property owners), who were still people of modest means. (See table 8.2 for the details.) We expect more restrictive marriage patterns among the farmers, who were most affected by population pressure in rural societies, compared with the poor and less rooted on the one side and the better-off households on the other side. (p.274)

Table 8.2 Means and frequencies of variables used in Sart and Pays de Herve, 1812–1899

Covariates

Sart 1812–1899

Pays de Herve 1846–1899

Men

Women

Men

Women

Occupation of head of household

Day laborer

8.7

9.1

10.5

9.8

Textile worker

0.9

1.6

9.8

10.0

Farmer

55.8

56.3

38.4

44.9

Craftsman

5.0

5.7

14.7

11.9

Petty bourgeoisie

10.2

7.8

9.5

8.6

Other

19.4

19.4

17.2

14.8

Presence of parents (mother/father)

Present/present

41.9

43.0

46.2

45.6

Present/absent

21.3

20.8

18.5

16.2

Absent/present

14.8

15.5

12.3

13.8

Absent/absent

22.0

20.8

23.0

24.5

Presence of sibling

Older brother

37.4

39.9

38.4

39.5

Older sister

36.0

36.6

41.5

38.8

Younger brother

51.1

55.2

49.1

48.7

Younger sister

48.8

48.6

50.1

48.1

Period

1812–1845

32.4

37.7

1846–1874

37.1

32.0

53.3

53.8

1875–1890

17.8

17.2

29.5

30.1

1891–1899

12.8

13.1

17.2

16.1

Fluctuation in rye prices

0.128

0.149

Fluctuation in cost of living

0.226

0.192

Sex ratio

134

134

100

99

Ratio young/old

2.19

2.19

1.97

1.97

Person-years at risk

23,955

18,297

29,127

29,834

Sources: See table 8.1.

Notes: Sex ratio = male to female ratio of unmarried 20 to 35 (per 100). Ratio young/old = persons 20 to 29/persons 55 to 64. Fluctuation in rye prices = deviation from the Hodrick–Prescott trend. Fluctuation in cost of living = deviation from the Hodrick-Prescott trend.

(p.275) Household characteristics, especially socioeconomic status, operate within a local community context. Macro variables characterize aspects of social life and the regional economy as they changed across time. Bengtsson (1993) introduces regional prices into individual-level event-history analysis to estimate the influence of short-term trends in economic conditions on life-course transitions, including differences in the vulnerability of socioeconomic groups within a village. In our model the regional/communal context is represented by four covariates, capturing both the economic and the demographic situation of the localities.

First, deviations from the trend of a price index indicate short-term economic fluctuations. For Sart, we use the annual price of rye, which was a vital cereal in the village economy, especially for the well-being of poor people (Alter, Neven, and Oris 2004). For the Pays de Herve, we use a regional cost of living index, because Hervian agriculture was highly commercialized and local farmers were obliged to buy most of their food (Oris, Alter, and Neven 2005). In each case, we use the difference between the log of the observed price and the log of the trend in prices for that year, which was computed using the Hodrick–Prescott filter (see Alter, Neven, and Oris 2004).

Second, we include indicator variables for time periods, which capture medium-term changes in economic and social conditions. The first half of the nineteenth century (1812–1846) was a period of population growth in Sart, which put increasing strain on a largely subsistence economy. During the next thirty years, the precarious economic balance in Sart was transformed by the sale of common lands and the arrival of American wheat. In the last quarter of the nineteenth century, we observe a restructured economy due to the conversion of the agrarian system from subsistence farming to a commercial economy based on livestock. In the Pays de Herve, the first half of the century saw the destruction of the prosperous proto-industrial system and the collapse of peasant ownership of land, but a new equilibrium was reached from 1850. The region pioneered the transition to commercial agriculture for the urban food markets, so the depression of 1873 to 1890 had little effect in the Pays de Herve (Oris, Alter, and Neven 2005).

Third, the ratio between the number of people aged 20 to 29 and those aged 55 to 64 is an indicator of the impact of population growth on marriage prospects. We hypothesize that a relatively large cohort of young people will have more limited economic opportunities: “large birth cohorts lead to increased competition for scarce resources such as (p.276) land, jobs and housing, which in turn leads to postponed marriages and higher marriage ages. In this way cohort sizes express the extent of general crowding and competition for scarce resources” (Lundh 1999a: 229). In a Malthusian society this covariate should be inversely related to the likelihood of marriage, because the younger generation cannot establish itself when the previous generation occupies the limited number of economic niches. Finally, another demographic variable is the sex ratio (male to female) of the unmarried people aged 15 to 39, which describes the gender balance in the marriage market. When unmarried women are in short supply, we expect the ages at marriage for women to fall and for men to rise.

The Determinants of Marriage and Outmigration: First Results

In the preceding section we saw that migration was sometimes an alternative and sometimes a pathway to marriage. Those who were unable to marry in Sart or the Pays de Herve could migrate to the expanding industrial agglomeration around Verviers. Urban life offered not only a large number of potential mates, but also a way to escape the constraints that hindered marriage in the rural communities. table 8.3 offers a comparative perspective on marriage and the outmigration of unmarried people in both Sart and the Pays de Herve.

Family characteristics played a great role in the decisions of the young adults in rural eastern Belgium, for instance, the socioeconomic status of the household head. The results confirm the hypothesis that the children of farmers had a distinct matrimonial pattern characterized by more controlled access to marriage. Thus the offspring of farmers were more likely to marry and less likely to migrate than others. As expected, farmers were the most rooted in their communities and had the best chances to marry, although marriage was a slow and uneasy transition even for them. In Sart unmarried people living in the household of a day laborer, the poorest families, were much less likely to marry and twice as likely to leave the location. Outmigration represented an escape from a precarious and limited situation for both men and women.

In modern and proletarianized Pays de Herve too, children of day laborers were more likely to outmigrate than farmers ’ children, but the context was different from that of Sart. Population turnover due to migration was already well established on the Hervian plateau, and farmers constituted a lower proportion of the population. Moreover (p.277) the relative risks of marriage for the sons and daughters of day laborers were higher than in the families of farmers in the Pays de Herve: a result that becomes statistically significant when we run a model in which the socioeconomic status is controlled for by region of birth (not shown here). Hervian day laborers acted like the urban proletariat, whose children enjoyed greater freedom. Parents did not have strong arguments for retaining their children at home and were not worried about the risks of a misalliance. In the late eighteenth and throughout the nineteenth century, the local elite observed and criticized this absence of prudence among the lower classes in both the countryside and the cities, and they contrasted this attitude with the patience of farmers ’ children, who waited for their father’ s death or retirement to take over the farm (Alter 1988: 145–48; Haesenne-Peremans 1981: 324). Such new behaviors had spread in the nineteenth-century Pays de Herve but not in the poorer Ardennes.

Among the unmarried aged 18 to 44, those who were living with both parents (42 percent in Sart and 46 percent in the Pays de Herve) were subject to the strongest parental control. The risks of both marrying and outmigrating were generally lowest when both the father and the mother were present in the household (see table 8.3). Those living with no parents, some of whom were servants and farm laborers, might be expected to be more mobile, but widows or widowers were also less able to retain their children at home than married couples. Previously we argued that the death of either parent set in motion negotiations and adjustments increasing the likelihood of marriage and migration, and the results in table 8.3 support this perspective. In both Sart and the Pays de Herve the risks of marriage tended to be higher in the first three years after a parent’s death than after three years, while the migration risks were lower in the immediate aftermath of a parent’s death in the Pays de Herve and following a father’ s death for women in Sart. This suggests that some couples who had been discouraged from marrying by their parents were free to marry after a parent died, and, at least in the Pays de Herve, these marriages replaced outmigration.

Additionally the results in table 8.3 show that co-residence with brothers or sisters substantially reduced both the chance to marry and the chance to outmigrate in both Sart and the Pays de Herve regardless of sex or birth order. We interpret this as an indication of unobserved differences among families regarding preferences and the demand for labor. From this viewpoint, the number of siblings at home was a (p.278)

Table 8.3 Cox regression estimates (relative risks) of marriage and outmigration: unmarried 18 to 44 in Sart and Pays de Herve, 1812–1899

Covariates

Men

Women

Marriage

Outmigration

Marriage

Outmigration

Sart 1812–1899

Pays de Herve 1846–1899

Sart 1812–1899

Pays de Herve 1846–1899

Sart 1812–1899

Pays de Herve 1846–1899

Sart 1812–1899

Pays de Herve 1846–1899

Relative risk

p-value

Relative risk

p-value

Relative risk

p-value

Relative risk

p-value

Relative risk

p-value

Relative risk

p-value

Relative risk

p-value

Relative risk

p-value

Occupation of head of household

Day laborer

0.65

0.00

1.11

0.37

2.00

0.00

1.14

0.08

0.76

0.02

1.12

0.26

2.39

0.00

1.29

0.01

Textile worker

0.98

0.96

0.80

0.09

1.85

0.02

1.01

0.89

0.50

0.02

0.90

0.33

1.48

0.17

0.90

0.23

Farmer

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

Craftsman

0.94

0.67

1.01

0.92

1.33

0.04

0.90

0.14

0.98

0.87

0.93

0.48

1.46

0.01

0.96

0.65

Petty bourgeois

0.65

0.03

0.86

0.27

2.54

0.00

1.21

0.02

1.00

0.98

0.81

0.08

2.12

0.00

1.04

0.63

Other

0.85

0.25

1.06

0.17

1.65

0.00

1.24

0.02

0.87

0.27

0.96

0.10

1.61

0.00

1.43

0.04

Presence of parents (mother/father)

Present/present

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

Present/absent

1.06

0.67

1.96

0.00

0.87

0.37

1.90

0.00

1.16

0.21

1.21

0.21

0.89

0.51

2.26

0.00

Absent/present

0.99

0.95

2.32

0.00

1.25

0.12

2.31

0.00

1.12

0.35

1.22

0.26

1.07

0.73

1.65

0.00

Absent/absent

0.95

0.78

0.89

0.76

2.07

0.00

2.83

0.00

1.24

0.27

1.35

0.24

2.23

0.00

2.24

0.00

Present/dead < 3 years

0.93

0.75

1.21

0.26

0.90

0.65

0.61

0.00

1.66

0.00

1.64

0.00

0.50

0.08

0.53

0.00

Absent/dead < 3 years

1.23

0.46

3.65

0.01

1.22

0.57

1.24

0.41

0.95

0.90

1.61

0.24

0.37

0.33

0.65

0.28

Dead < 3 years/present

1.35

0.18

1.14

0.55

0.88

0.69

0.51

0.00

1.42

0.08

1.66

0.01

0.91

0.78

0.63

0.00

Dead < 3 years/absent

0.89

0.71

4.25

0.00

2.13

0.02

1.24

0.42

1.60

0.10

2.46

0.01

2.87

0.00

1.04

0.91

Both dead < 3 years

3.03

0.01

3.70

0.00

2.50

0.08

0.74

0.16

1.73

0.24

1.29

0.36

2.99

0.07

0.84

0.48

Presence of sibling

Older brother

0.69

0.00

0.74

0.00

0.75

0.00

0.75

0.00

0.76

0.00

0.92

0.23

0.71

0.00

0.72

0.00

Older sister

0.65

0.00

0.70

0.00

0.70

0.00

0.74

0.00

0.74

0.00

0.73

0.00

0.87

0.09

0.80

0.00

Younger brother

0.85

0.03

0.82

0.01

0.82

0.00

0.83

0.00

0.76

0.00

0.94

0.35

0.78

0.00

0.86

0.01

Younger sister

0.69

0.00

0.82

0.02

0.80

0.00

0.91

0.07

0.91

0.14

0.96

0.56

1.11

0.19

1.00

0.97

Period

1812–1845

1.48

0.00

0.46

0.00

1.08

0.43

0.42

0.00

1846–1874

1.26

0.04

0.99

0.90

0.88

0.08

0.83

0.00

1.18

0.07

1.01

0.86

0.89

0.20

0.81

0.00

1875–1890

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1891–1900

1.53

0.00

0.97

0.87

0.64

0.00

1.08

0.42

1.26

0.05

1.22

0.18

0.86

0.20

1.29

0.03

Sex ratioa

0.96

0.19

1.01

0.40

1.22

0.00

1.00

0.62

0.99

0.65

0.99

0.17

1.13

0.06

0.99

0.08

Ratio young/olda

0.99

0.70

0.96

0.03

0.96

0.14

0.98

0.10

0.96

0.01

1.00

0.82

0.97

0.40

0.99

0.57

Fluctuation in rye prices

0.83

0.28

1.12

0.43

0.89

0.43

1.33

0.07

Fluctuation in cost of living

1.23

0.22

0.99

0.93

0.90

0.50

0.96

0.82

Events

889

790

1,141

1,908

1,110

1,096

806

1,767

Person-years at risk

23,955

29,127

23,955

29,127

18,297

29,834

18,297

29,834

Overall p -value

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.00

Sources: See table 8.1.

Notes: Sex ratio = male to female ratio of unmarried 20 to 35. Ratio young/old = persons 20 to 29/persons 55 to 64. Fluctuation in rye prices = deviation from the Hodrick-Prescott trend. Fluctuation in cost of living = deviation from the Hodrick-Prescott trend.

(a.) Rescaled to represent a 10 percent change.

(p.279) (p.280) consequence rather than a cause of marriage and migration behavior. Families with more adult children were more able to put them to work, and the sibling group was large because children in those families were less likely to leave home by marrying or migrating.

The macro variables in table 8.3 had much less impact than the individual- and household-level variables. The sex ratio among unmarried people aged 20 to 35 was very well balanced in the Pays de Herve (100 men for 100 women) but largely skewed in Sart (134 men per 100 women). Young women in Sart were more attracted to jobs in Verviers than men, while a balanced combination of male and female skills remained vital to the family economy in the Pays de Herve (Alter and Oris 1999a: 138–39; Neven 2003: 123). Although the unbalanced sex ratio in Sart could be expected to have increased the marriage chances for women and decreased them for men, we do not see that effect. Instead, both women and men were more likely to outmigrate during years in which the male-to-female ratio was higher. This suggests that the marriage market was not limited to those living in Sart.

The ratio of persons aged 20 to 29 to those aged 55 to 64 was included in the models to examine the effects of population structure on marriage opportunities. A high ratio of young to old adults could be hypothesized to have implied fewer opportunities for the young to establish new households and a greater tendency for them to migrate elsewhere. We find some, albeit quite weak, support for the hypothesis. The results indicate a negative effect of the young-to-old ratio on the marriages of women in Sart and men in the Pay de Herve but no effect on outmigration. Similarly marriage and migration were not responsive to short-term fluctuations in the price of food, which are a proxy for the variations in the standard of living and economic stress (Bengtsson 2004). Except for an increase in the outmigration of single women in Sart, no consistent pattern emerges from the analyses. Although the situation of unmarried adults in eastern Belgium was very difficult, the impediments to marriage primarily operated on individuals within families, and variations in macro-level economic and demographic conditions played a marginal role.

If some aspects of marriage and outmigration were similar in Sart and the Pays de Herve, differences across time periods highlight contrasts in the timing of economic development. In Sart, 1875 to 1890 were the worst years for marriage because the arrival of American cereals forced the transition from subsistence farming to a form of commercial agriculture requiring less labor. During those years outmigration was (p.281) at the highest level observed during the nineteenth century. By the 1890s the population of Sart had decisively adapted to a new rural economy, and couples were adopting birth control to relax the constraints on marriage in the future. In contrast, farmers in the Pays de Herve, who already operated commercial agriculture based primarily on livestock, suffered less during the late-nineteenth-century agricultural depression and entered the twentieth century resistant and self-controlled.

An Additional Constraint: Partner Selection

Marriage in the countryside was a complicated life transition not only because young adults had to wait, often a long time, for an opportunity to form a new household but also because the right partner had to be found. As in most of preindustrial Europe, the customs in rural eastern Belgium privileged endogamous marriages, especially among farmers, who needed both male and female skills to operate a farm (Segalen 1980). This is confirmed empirically in an analysis of marriages in a dozen Belgian communities by Van de Putte et al. (2005), showing that the levels of endogamy were higher in the villages than in the cities. Endogamy was mostly explained by the proportion of farmers in the population, which also drove the differences between the rural areas. Indeed “the higher level of absolute heterogamy measured in the Pays de Herve was the result only of the relatively small numbers of farmers, meaning that at the level of the daily experience as reflected in the level of absolute heterogamy we see societal openness caused by the presence of a large number of skilled and lower-skilled workers, such as those in the textile industry. Nonetheless, preference is strong for a partner from the same socioeconomic origin, as is shown by controlling for the effect of group sizes” (Van de Putte et al. 2005: 211). The analyses of Van de Putte et al. (2005) were run exclusively on the married population in the sample municipalities. Since we are using the population registers, we are able to include in the analysis of marriage risks the additional constraint on marriage of the preference for a partner of the same social origin.

Tables 8.4 (Pays de Herve) and 8.5 (Sart) display the results from a new set of models. As in table 8.3, the design is competing risk between marriage and migration, but here we distinguish among endogamous, exogamous, and “indeterminate” marriages. As in the previous regressions, we try to estimate the influence of determinants at the (p.282)

Table 8.4 Cox regression estimates (relative risks) of endogamous, exogamous, and indeterminate marriage, and outmigration in Pays de Herve, 1846–1890

Covariates

Endogamous marriage

Exogamous marriage

Indeterminate marriage

Outmigration

Relative risk

p -value

Relative risk

p -value

Relative risk

p -value

Relative risk

p -value

Sex

Male

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

Female

1.79

0.00

1.70

0.00

1.08

0.24

0.85

0.00

Place of birth

Pays de Herve

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

Rural area

0.76

0.24

0.80

0.28

1.03

0.82

1.62

0.00

Urban/industrial aria

1.10

0.80

1.32

0.33

0.53

0.01

1.52

0.00

Other

0.27

0.03

0.64

0.21

0.48

0.00

1.91

0.00

Occupation of head of household

Farmer

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

No activity

0.63

0.08

2.17

0.00

1.15

0.21

1.14

0.06

Day laborer

0.39

0.00

1.84

0.00

1.24

0.06

1.34

0.00

Textile worker

0.66

0.04

1.14

0.52

0.93

0.57

1.02

0.73

Craftsman

0.53

0.00

2.07

0.00

0.99

0.89

0.94

0.33

Petty bourgeoisie

0.24

0.00

1.61

0.01

0.90

0.38

1.14

0.04

Presence of parents (mother/father)

Present/present

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

Present/absent

0.71

0.10

1.06

0.72

2.31

0.00

1.19

0.00

Absent/present

1.25

0.22

1.33

0.07

1.49

0.00

1.16

0.02

Absent/absent

0.45

0.00

0.57

0.00

1.70

0.00

1.97

0.00

Presence of sibling

Older brother

0.75

0.00

0.84

0.03

0.79

0.00

0.92

0.00

Older sister

0.62

0.00

0.60

0.00

0.76

0.00

0.92

0.00

Younger brother

0.84

0.01

0.85

0.00

0.92

0.02

1.03

0.11

Younger sister

0.88

0.03

0.90

0.05

0.95

0.12

1.07

0.00

Period

1846–1872

1.07

0.62

1.37

0.01

0.76

0.00

0.88

0.00

1873–1890

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

Fluctuation in cost of living

1.00

0.97

0.99

0.42

1.01

0.08

1.00

0.84

Events

393

952

541

3,675

Overall p -value

0.00

0.00

0.00

0.00

Sources: See table 8.1.

Note: Fluctuation in cost of living = deviation from the Hodrick-Prescott trend.

(p.283) (p.284)

Table 8.5 Cox regression estimates (relative risks) of endogamous, exogamous, and indeterminate marriage, and outmigration in Sart, 1812–1890

Endogamous marriage

Exogamous marriage

Indeterminate marriage

Outmigration

Covariates

Relative risk

p-value

Relative risk

p-value

Relative risk

p-value

Relative risk

p-value

Sex

Male

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

Female

1.36

0.00

1.87

0.00

1.73

0.00

0.99

0.90

Place of birth

Sart

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

Outside Sart

0.62

0.01

0.83

0.31

0.89

0.32

2.20

0.00

Occupation of head of household

Farmer

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

No activity

0.59

0.03

0.60

0.09

0.73

0.02

1.47

0.00

Day laborer

0.50

0.00

1.26

0.20

0.75

0.05

1.99

0.00

Craftsman

0.44

0.00

1.43

0.06

0.76

0.10

1.43

0.00

Petty bourgeoisie

0.92

0.74

1.05

0.86

0.74

0.14

1.94

0.00

Presence of parents (mother/father)

Present/present

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

Present/absent

0.80

0.101

0.68

0.026

2.11

0.000

1.11

0.31

Absent/present

1.09

0.483

1.28

0.083

1.09

0.464

1.16

0.15

Absent/absent

0.57

0.000

0.73

0.058

1.65

0.000

2.30

0.00

Presence of sibling

Older brother

0.83

0.01

0.75

0.00

0.85

0.00

0.84

0.00

Older sister

0.59

0.00

0.74

0.00

0.77

0.00

0.81

0.00

Younger brother

0.83

0.00

0.89

0.03

1.02

0.59

0.98

0.60

Younger sister

0.95

0.31

0.84

0.00

0.83

0.00

1.06

0.07

Period

1812–1846

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

1847–1874

0.90

0.33

0.92

0.52

1.20

0.04

6.04

0.00

1875–1890

0.57

0.00

0.87

0.37

1.26

0.02

6.14

0.00

Fluctuation in rye prices

0.99

0.33

0.98

0.15

0.99

0.14

1.01

0.05

Events

816

629

554

1,947

Overall p-value

0.00

0.00

0.00

0.00

Sources: See table 8.1.

Note: Fluctuation in rye prices=deviation from the Hodrick-Prescott trend.

(p.285) (p.286) individual, household, and macro levels. Since weddings were usually celebrated in the bride’s parish even when the couple settled some-where else, we frequently know that a marriage happened, but we do not observe the spouse and his household of origin (in about 28 percent of marriages). We are therefore unable to determine whether those unions were endogamous or exogamous, or to disentangle homogamy by socioeconomic status and by place of residence. The population at risk remains the same as in table 8.3, namely unmarried persons aged 18 to 44 living in Sart and the Pays de Herve. Joint estimates for men and women are presented in tables 8.4 and 8.5, the higher relative risk associated with sex reflecting the earlier age at marriage among women.

When comparing endogamous versus exogamous marriage, the two rural areas appear similar from many perspectives. Obviously the socioeconomic status of the household head was the main determinant (see tables 8.4 and 8.5). In the Pays de Herve, for instance, unmarried adults living in farmer-headed households were about twice as likely to enter an endogamous union as all the other socioeconomic groups. All the groups, except textile workers, were more likely than farmers to find a spouse in a different occupational group. As we expected, farmers had a distinctive marriage pattern! In Sart, the differences between farmers and the other socioeconomic groups in the risk of an endogamous marriage were similar, but the contrast in exogamous marriage seems to be affected by a higher proportion of indeterminate marriages among farmers.

Other family variables also played a role. Both endogamous and exogamous marriages were less likely for those living apart from their parents or whose parents were deceased. These independent young people, some of whom were servants, were more geographically mobile and more likely to outmigrate or to form “indeterminate” marriages with a spouse from a different location. As we saw above, people who lived with both parents were the least likely to marry outside their location of origin. Likewise living with siblings of any age or sex decreased the likelihood of any kind of marriage or outmigration.

Although short-run price fluctuations had no influence on any type of marriage in these rural samples, we do see structural changes across time periods. In Sart, endogamous marriage tended to decrease during the nineteenth century. Compared with the first half of the nineteenth century (1812–1846), the relative risk of an endogamous marriage was only 0.57 after 1875, and the risk of an “indeterminate” marriage increased with time as mobility grew. In the Pays de Herve, where the (p.287) core of local society resisted change, the risk of endogamous marriages did not change after 1846, while a decrease in exogamous marriages was compensated for by a rise in marriages with partners from other villages.

One final finding is that natives of Sart and the Pays de Herve were more likely than residents born elsewhere to form unions that were endogamous with respect to socioeconomic status. We suggest that this reflects the higher density of social networks among those who had lived in the same location since birth.

The choice between waiting for a local marriage and leaving home in search of better prospects was not the same for every young adult in Sart or the Pays de Herve. In both places young people born into a core of farm families waited longer to marry a spouse from a local family very similar to their own. In these families the respect for norms relating to when and whom to marry was strongest among young adults living with both parents. Across generations socially and geographically endogamous unions reinforced the homogeneity of families at the heart of rural society. The constraints imposed on partner choice and the timing of marriage in these families implied a high level of tension between parents and unmarried adults.

Controlling Marriage, Controlling Sexuality: A Few Paradoxes

During the nineteenth century eastern Belgium had a low to moderate level of illegitimacy: fewer than 3 percent of births in the Pays de Herve and around 6 percent in Sart. In view of the large number of unmarried adults in these populations, we might think that social control was largely effective in limiting sexual activity to marriage (Kok 2005). Respect for institutions is also reflected in the compliance with prohibitions on marriage during Advent and Lent by the Catholic Church (Neven and Oris 2003). Yet women were frequently pregnant on their wedding day. The prevalence of premarital pregnancies points to conflicts within the marriage system.

Premarital pregnancy is not easy to interpret, since it can occur in very different situations. Laslett (1980: 8) suggests that most pregnant brides resulted from “accidents” during long periods of engagement.2 Flandrin (1976: 180) evokes mariages de r ép aration: couples who had no common project for the future but were forced to marry to repair their mistake. In some places, the fertility of the bride was tested in a “trial marriage” and pregnancy was a positive result leading to marriage.3 (p.288) Pregnancy has also been seen as a trap set by the bride to oblige her lover to marry her. Alter (1988: 115–20) argues that lovers used pregnancy to overcome the resistance of parents, who objected to either the timing of a marriage or the choice of partner. When families derive economic benefits from the labor of their children, parents have an incentive to postpone the marriages of their children, and as we saw above, the Belgian Civil Code made it very difficult to marry without parental consent. When a young woman was pregnant, however, her parents were compelled to agree to a wedding. Otherwise, they faced condemnation from the local elite, the priest, and, last but not least, an age-graded community in which the local “youth society” supported their fellows (Enquêtes … 1963–67: 237–40).

We examine this issue by estimating event-history models separately for marriages in which the bride was or was not pregnant. We treat these two kinds of marriage as different events that were “competing risks.” An unmarried person is considered to be at risk of either type of event, and each type of marriage is treated as a censoring event for the other type. We must, however, keep in mind that we exclude marriages that were followed by outmigration in the next eight months, since we must determine whether the bride was pregnant when she married. Consequently we include fewer marriages than in the preceding analyses. The sample is restricted to couples who settled in the sample villages after their weddings, namely those who were more geographically stable than their peers. Thus the focus is on the tensions within families in the stable core of these rural societies.4

Table 8.6shows event-history models for marriages without and with premarital conceptions in Sart and the Pays de Herve. Overall, there are few noticeable differences, although the estimates are more likely to be statistically significant for marriages without premarital conception, which were more numerous. Pregnant brides were more common among the daughters of poor families, especially day laborers. Contemporary observers expected children of the landless to marry freely, since no patrimony was endangered and the risk of misalliances was absent. We find here evidence of a different moral culture among rural day laborers and the urban proletariat (see chapter 4 in this book and Neven and Oris 2003: 15–17).

The differences between the two types of marriage stand out when we look at the effects of parental presence. In most cases the death or absence of a parent increased the likelihood of a marriage without a pregnancy and reduced the likelihood of a premarital pregnancy. For (p.289) example, in Sart, women living without any parents were 53 percent more likely to marry without a premarital conception and 47 percent less likely to marry after a premarital conception. This pattern is also observed in the Pays de Herve. Although these results are not always statistically significant, this finding fits the model of parent–child negotiation. Children who had already decided on a spouse could use a pregnancy to put social pressure on their parents if they had previously been refused permission for marriage (Alter 1988: 120; Leboutte 1988: 401; Oris 2000). If parent–child conflicts over marriage were common in rural eastern Belgium, it could explain the different behavior of people whose parents were absent. The death of one or both parents changed the negotiation. Young people were more able to decide the timing of marriage themselves and less likely to postpone marrying until pregnancy left little choice. Waiting for parents to transfer property was no longer necessary, and they could marry without resorting to social pressure to extract parental consent.

Two other interesting results emerge from table 8.6. First, marriages with premarital conceptions declined over time in both Sart and the Pays de Herve. Using pregnancy to obtain parental consent was less necessary when outmigration had become an acceptable option. The rural population that stayed behind was becoming more socially coherent, and strict social controls were being maintained, if not strengthened (Neven and Oris 2003). Second, marriage was insensitive to short-term economic fluctuations in the models presented above, and we have argued that tensions at the individual, family, and macro levels were so high that short-term considerations were not important. table 8.6 indicates that marriages without premarital conception became less frequent in Sart when the price of rye went up. In bad times prudent couples waited, and those who could not wait provoked an emergency to obtain permission to wed.

Conclusions

Both the Pays de Herve and Sart provide good illustrations of Davis’s model of the demographic transition as a multiphasic process (Alter, Oris, and Neven 2010). While these communities faced a double bind from contracting economic opportunities and rising rates of natural increase, they responded with a combination of old (late marriages and high proportions of never marrying among the stayers) and new (outmigration to industrial towns) behaviors. Outmigration increased, but (p.290)

Table 8.6 Cox regression estimates (relative risks) of marriage with or without prenuptial conception: unmarried women aged 18 to 44 in Sart and Pays de Herve, 1812–1899

Sart 1812–1899

Pays de Herve 1846–1899

Without prenuptial conception

With prenuptial conception

Without prenuptial conception

With prenuptial conception

Covariates

Relative risk

p-value

Relative risk

p-value

Relative risk

p-value

Relative risk

p-value

Occupation of head of household

Farmer

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

Day laborer

0.71

0.07

1.09

0.79

0.85

0.19

1.57

0.09

Petty bourgeoise

1.21

0.23

1.17

0.69

0.92

0.58

1.33

0.42

Textile worker

0.76

0.04

0.63

0.22

Craftsman

0.67

0.00

1.16

0.58

Other/no activity

1.13

0.29

1.18

0.47

1.12

0.43

1.71

0.11

Presence of parents(mother/father)

Present/present

1.00

ref.

1.00

ref.

1.00

ref.

1.00

ref.

Present/absent

1.20

0.16

1.01

0.97

1.26

0.17

1.09

0.84

Absent/present

1.14

0.32

1.06

0.81

1.29

0.19

0.79

0.68

Absent/absent

1.53

0.05

0.53

0.17

1.54

0.11

0.37

0.32

Present/dead < 3 years

1.72

0.01

1.44

0.32

1.62

0.01

1.19

0.71

Absent/dead < 3 years

1.02

0.97

0.64

0.56

1.52

0.35

3.70

0.28

Dead < 3 years/present

1.40

0.15

1.40

0.39

1.67

0.01

1.57

0.45

Dead < 3 years/absent

1.92

0.04

0.81

0.74

2.56

0.01

3.47

0.30

Both dead < 3 years

1.90

0.22

1.08

0.94

1.27

0.43

1.66

0.63

Presence of sibling

Older brother

0.84

0.11

0.93

0.73

0.65

0.00

0.64

0.00

Older sister

0.77

0.02

0.75

0.21

0.58

0.00

0.45

0.00

Younger brother

0.77

0.01

0.81

0.32

0.74

0.00

0.59

0.00

Younger sister

1.03

0.75

1.12

0.58

0.84

0.00

0.68

0.00

Period

1812–1845

1.00

ref.

1.00

ref.

1846–1872/74a

1.40

0.07

0.58

0.08

0.93

0.44

1.19

0.46

1873–1890

1.00

ref.

1.00

ref.

1875–1899

1.35

0.03

0.46

0.00

1891–1900

1.04

0.76

0.46

0.06

Sex ratiob

0.98

0.60

1.00

0.97

0.93

0.87

2.35

0.46

Ratioyoung/oldb

0.96

0.04

0.95

0.12

1.23

0.44

0.85

0.80

Fluctuation in rye prices

0.97

0.01

0.84

0.66

Fluctuation in cost of living

1.00

0.69

0.99

0.47

Events

604

148

768

123

Overall p-value

0.00

0.01

0.00

0.00

Sources: See table 8.1.

Notes: Sex ratio = male to female ratio of unmarried 20 to 35. Ratio young/old = persons 20 to 29/persons 55 to 64. Fluctuation in rye prices=deviation from the Hodrick-Prescott trend. Fluctuation in cost of living = deviation from the Hodrick-Prescott trend.

(a.) Sart: 1846–1874, and Pays de Herve: 1846–1872.

(b.) Rescaled to represent a 10 percent change.

(p.291) (p.292) strict control over access to marriage was maintained throughout the nineteenth century within a core community of farming families. The young adults in these families paid a price to stay on the land by waiting longer to marry than their contemporaries who moved to the cities.

The responses of late marriage and outmigration meant different things for different socioeconomic groups. Children born into the poorest families, the landless day laborers, were more like urban workers because they were more mobile and less respectful of norms limiting sexual behavior. In poor households, parents were often dependent on the earnings of adult children and reluctant to forgo their earnings, which set the stage for conflict and negotiation over the timing of marriage. Moreover marriage had costs, even when parents were not expected to provide land or other property. Couples had to obtain official documents (birth certificates, parental consent, etc.), wait for formalities (publication of bans), and satisfy social obligations within the village community. The local youth group expected to be entertained, and it was more expensive if the groom was from another village (Doppagne 1978).

The marriage strategies of farm families differed from those of landless laborers. They were the most endogamous and also the most stable, from both occupational and spatial points of view. Young people of farming origin were more controlled than those from other socioeconomic groups, resulting in stronger respect for social norms. Farming families avoided marrying during periods forbidden by the Catholic Church, their intervals between marriage and first birth were longer, and illegitimate births were rarer. They also had higher expectations about marriage. The groom was expected to acquire skills and capital, the bride would have assembled a trousseau, and both should bring land to the new household. Their wedding was also expected to be more generous to the local community and youth society. In meeting these requirements, farming families relied on social networks for privileged access to the “invisible” market in which transactions often involved kin (Servais 2009).

In addition we have seen a tendency of families to resist change and discourage the transition of property and household headship to the younger generation. Young adults living with a widowed parent or on their own were more likely to marry than those living with both parents. They were also less likely to marry if they lived with a sibling of any age or sex. These results suggest a strong binding force within families (p.293) that resisted change. Parents used their influence to delay or discourage marriages, and young adults had two responses. Many waited until one parent died, which broke the spell by redistributing both property and power. Some chose to force the issue by engaging in sexual relations, knowing that a pregnancy would trump parental authority in the court of public opinion. Over time the second strategy became less important. For poor families, higher incomes in the cities created greater opportunities and spread bourgeois models of respectability. Farm families consolidated both their economic position and their adherence to strict social norms.

Despite very different agricultural endowments and economic conditions, we found many similarities between Sart and the Pays de Herve. This may be due partly to the fact that from 1860 many efforts were made in the Ardennes to move in the direction of the Hervian model (Servais, 2013). Whatever the reason, marriage behavior in both areas reflected the importance of the principle of neolocality in a period when the opportunities to form new households were contracting. Under these circumstances, human agency, “in the form of collective culture, institutions and social norms and in the form of actors” (Lundh and Kurosu, chapter 1 in this book), resulted in extremely tight control of access to marriage. Late marriage and outmigration were temporary responses until the integration into the modern economy and the adoption of family limitation removed the threat of population pressure in these rural communities.

Notes

(1.) Here the TFR is estimated from cross-sectional data on the age-specific fertility for each decade.

(2.) In Montzen, a village located on the border of the Pays de Herve, M. Lennarts observes from notarial archives the strong relation in the Modern Period between the promise of marriage and sexual intercourse (Lennarts 1984: 153).

(3.) Among the Li è ge miners, this behavior was attested by qualitative sources (Oris 1993: 59). Lesthaeghe notices that in Flanders, following the Napoleon era, out-of-wedlock pregnancies entailed more out-of-wedlock births than premarital pregnancies. He thus concludes that the ecclesiastical authorities were increasingly less able to oblige the “sinner,” the “offender,” to marry the young girl (Lesthaeghe 1991: 268).

(4.) This is also a different sample from the one studied in chapter 4 of this volume. (p.294)

Notes:

(1.) Here the TFR is estimated from cross-sectional data on the age-specific fertility for each decade.

(2.) In Montzen, a village located on the border of the Pays de Herve, M. Lennarts observes from notarial archives the strong relation in the Modern Period between the promise of marriage and sexual intercourse (Lennarts 1984: 153).

(3.) Among the Li è ge miners, this behavior was attested by qualitative sources (Oris 1993: 59). Lesthaeghe notices that in Flanders, following the Napoleon era, out-of-wedlock pregnancies entailed more out-of-wedlock births than premarital pregnancies. He thus concludes that the ecclesiastical authorities were increasingly less able to oblige the “sinner,” the “offender,” to marry the young girl (Lesthaeghe 1991: 268).

(4.) This is also a different sample from the one studied in chapter 4 of this volume. (p.294)