![Sample our Social Sciences journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110822/TOC-Subject-Sample-2021-Social-Sciences.jpg"})
Abstract
This paper examines the partner selection of the lower classes during an urban crisis period in early industrial Belgian cities. It was found that in this period characterized by an economic transition, overpopulation, migration and a low standard of living, social heterogamy was high, whereas social homogamy increased, or was ‘restored’, in the subsequent period. The urban crisis effect on partner selection contradicts the claims of modernization theory that there was a gradual increase in societal openness and that societal openness was typically modern, but it fits the idea of the informalization of marriage, a process marked by an increase in unmarried cohabitation and illegitimacy.
Notes
1 Ascribed positions, like gender, age, geographical and social origin, are assigned to an individual beyond his or her effort, and are the opposite of achieved positions. This is a central distinction in the sociology of modernization (CitationRitzer, 1992).
2 The standard modernization thesis might of course also be wrong (CitationBeck, 1992). Another factor is the lack of a common scheme for the classification of occupational titles (Van de Putte & Miles, Citation2005; Maas & van Leeuwen, Citation2005).
3 That is, in case the correlation between social origin and social position is not too strong.
4 The influence of social structure is context dependent. We will control for its influence in the empirical analysis. The role of meeting opportunities is difficult to assess, we do not make claims about it, but see the methodology section.
5 In many 18th- and 19th-century cities (e.g. Grenoble, Geneva), a distinction is made between the rooted persons, who are seen as the ‘core’ of the city, those who own and transmit the local town culture and the city pride, and the mobile persons, who are just going through. Migrants meant that marrying a native bride was seen as a good strategy of stabilization and integration in the city.
6 It is important to distinguish this period of early political class formation from the subsequent period (1891–1913). A new political climate, in which political class formation and the ‘social question’ were central issues, swept over Belgium. The socio-political atmosphere was troubled since the strikes of the mid-1880s, sometimes framed as a ‘violent workers' revolt’ (CitationLamberts, 1999, p. 332). The advent of universal (plural) manhood suffrage (1893) and the entry of the Socialist party into parliament served to exacerbate this atmosphere. In this climate of massive working class discontent, the Labour movement became a mass movement. In our view, at that time, the process of class formation led to new group belonging criteria (weaker boundaries within the lower classes) and this led (again) to heterogamy, at least in Flanders, which was indeed observed in Ghent (CitationVan de Putte, 2005).
7 For approximately 50% of the spouses' fathers were not alive at the wedding. This lack of information on this group implies that the results are not necessarily applicable to the whole population, but only to the population of spouses with living fathers. Simulations that take a (possibly) higher level of intergenerational mobility and a (possibly) lower level of social control for those with dead fathers into account, while also controlling for the changing amount of spouses whose father was dead, shows that at least for Ghent the trend of heterogamy is valid for the whole population under these assumptions (see CitationVan de Putte, 2005). Of course, it cannot be excluded that spouses with dead fathers react completely different to the living conditions during the urban crisis.
8 Information on the period before 1800 is difficult to collect as it was only after 1796 that the government created marriage certificates. In parish certificates, only a few occupational titles are recorded.
9 As a preparatory step we applied the Historical International Standard Classification of Occupations (HISCO) scheme (Citationvan Leeuwen, Maas & Miles, 2002). HISCO is a functional classification distinguishing between occupations on the basis of the tasks associated with them. Each occupational group gets a five-digit code (e.g. 75,400 for ‘weaver’). Other information, for example on employment status, is stored in separate codes.
10 Yet, there were more ‘labourers’ and ‘factory workers’ in Ghent, and they probably worked in textile factories. This reflects the different origin of the workers. While Ghent recruited unskilled workers among its Lumpenproletariat, Verviers attracted proto-industrial textile workers coming from the neighbouring countryside and the suburbs.
11 Yet, this ‘pole position’ of Verviers does not seem to be the obvious answer to population growth, as – although really important – it has never reached the levels and the rhythm observed in other Walloon industrial areas, such as Liège or Seraing. Of course, the other industrial centres were also overpopulated, but the overcrowding was by no means equal to that observed in Verviers. The specificity of the wool city may be explained by the bad railway connections. The railway only reached the city in 1843, and good connections with the neighbouring villages were only present from the 1870s. In these circumstances, workers had no other choice than living near their factory.
12 These are, of course, no migration figures. First, unmarried cohabitation is not taken into account, although probably strong for migrants and in the first half of the 19th century (CitationVan de Putte, 2005). Second, natives and migrants may differ in marriage intensity. Third, there may be a time lag between arrival and marriage.
13 We cannot distinguish between those with and without a father present, as the latter category is too small.
14 To compare father and son we used the HISCO codes in the two-digit form (e.g. 95=construction worker).
15 The small number of sons with the same occupation but with a different SP level are coded in the last category. This situation may occur where there is information on employment or hierarchical status, as in this case other SP levels are assigned (e.g. master carpenter and carpenter are coded in the same occupational group, but in different SP levels).
16 The mixed group is probably more integrated than the least rooted group. For example, a migrant marrying a native bride in Verviers is someone who wanted to move to the rooted group, and his marriage was already a success, as well as a promise for further integration in the group of real Verviétois.
17 For each category of social origin (e.g. SP level 2), the percentage of fathers of the bride that belongs to this category is calculated. For each groom with this specific social origin (here: SP level 2), this percentage is the chance to marry homogamously. The results show whether there are differences in the chance of marrying homogamously between locations and groups, controlling for the group size. For information on this procedure, see CitationVan de Putte (2005).
18 And if this changed during the period under observation, this probably only led to an increased difficulty to observe the expected effect on heterogamy in the first half of the 19th century. The slums in Ghent (composed of very small dwellings built around a square or in a small street, without individual sanitation) restricted meeting opportunities between members of different classes, as there were only poor inhabitants. As these slums were already numerous in the first half of the 19th century, this may result in higher levels of homogamy rather than heterogamy.
19 For the use of log-linear analysis in the study of (historical) social mobility, see CitationGrusky and Fukumoto (1989) and Van Leeuwen and Maas (Citation1996, Citation2002).
20 In a so-called saturated model, an individual parameter is calculated for all cells. In topological models, groups of cells are constrained to have the same parameter, assuming that the differences between the cells are negligible.
21 Homogamy parameters are similar to inheritance parameters in research on intergenerational mobility, that is, parameters that refer to the diagonal cells of mobility tables.
22 The models can be described in this way: Model 1={l, p, g, b, pg, pb, lp, lg, lb, lpg, lpb, D1(gb, 1, p), D2(gb, 1)}, where l=location, p=period, g=SP level father groom, b=SP level father bride, pg=the association between period and SP level father groom,…, lpg=third-order effect of location, period and SP level father groom,…, D1=design matrix for the association between SP level father groom and bride (SP level 1–3) with one parameter (gb) calculated for each period; D2=design matrix for the association between SP level father groom and bride (SP level 4–5) without period as the grouping variable. Model 2={l, p, g, b, pg, pb, lp, lg, lb, lpg, lpb, D1(gbp, 3, l), D2(gb, 1)} where D1=design matrix for the association between SP level father groom and bride (SP level 1–3) and period, per location, with one parameter (gb) for each period, calculated per location.
23 The parameter estimates of all the effects in the model allow the calculation of the model expected frequencies. If there are other effects, not included in the model, that strongly determine the observed frequencies, the model expected frequencies will be rather different from the observed (true) frequencies.
24 Otherwise the ‘saturated model’ (containing all effects) would always be the best model, as observed and model expected frequencies are by definition the same in this model. Saturated models are not theory driven, and hence have the disadvantage that they are very difficult to interpret.
25 If many cells of the partner selection table have zero or a few observations the estimation of the parameters is biased.
26 If we also include the middle class and the elite, we need to add interaction parameters to the model (e.g. between the presence of parents and social origin). This unnecessarily complicates the analysis. The claims about the effect of the urban crisis are made for the lower classes, not for the middle class and the elite. Taking this perspective of the groom does not limit the procedure for controlling for the effect of group sizes (CitationVan de Putte, 2005). If we take the other perspective (selection of brides with lower class social origin) the results are not different.
27 We cannot simply use this variable in model 2A as there is an overlap with the variables measuring the presence of the father. The category ‘native*native’ (geographical homogamy) is a combination of the category ‘native’ of the variables that measure the presence of father groom and the presence of father bride.
28 A short note on the interpretation of the parameter estimates of the logistic regression analysis. The parameters of a logistic regression analysis show how many times more (if higher than 1) or less (lower than 1) the chance is of marrying heterogamously if one belongs to a specific category of a variable compared with the reference category of that variable. In a model containing interaction effects, the main effects of the variables in the interaction effect (e.g. between period and family witnesses) show the effects within the reference group of the other variable in the interaction effect. For example, the parameters for family witness show the difference in the chance of marrying heterogamously in the reference period. The period effect shows the effects for the reference category of ‘family witness’. The interaction effects of ‘family witness’ and period show how this period effect differs in the other categories of ‘family witness’.
29 The negative Bic values for both models signify that the models are better in terms of fit and parsimony than the saturated model. This means that omitting more specific effects (e.g. different homogamy indices per SP level) does not lead to a significant loss of information.
30 Mind that the models are very parsimonious. Here we are not interested in what precisely happens in the off-diagonal cells. We do not have any hypotheses on whether people marry upwardly or downwardly.
31 Model information:
32 To this must be added that in the period 1874–1890, the effect of working in the same sector as one's father disappeared (parameters obtained by the same model, but with 1873–1890 as the reference period: 1.077; p=0.815), which confirms our interpretation.
33 This view on the limited impact of one's social origin during the urban crisis may seem to contradict the widely held vision that in times of crisis the importance of family support increased (Kocka, Citation1984; van Poppel & Nelissen, Citation1999). The latter vision is, of course, plausible, in the sense that in these conditions presumably all potential channels of support were used. But this does not necessarily imply that social origin was more rigidly used as a criterion of partner selection. The conditions under which social origin is a criterion in partner selection are that one's social origin is, first, known by the parties involved, and, second, estimated, at the time of marriage, as more useful compared with the social positions and characteristics of the spouses themselves. In our opinion, in times of urban crisis these conditions were not fulfilled.
34 Also in Ghent there was a high level of illegitimacy (Art, Citation1972; Cloet, Citation1991, pp. 438–439) and probably unmarried cohabitation (CitationVan de Putte, 2005). According to CitationAlter (1988, p. 112), Verviers “had a moderately high illegitimacy ratio (about 9% of all births were illegitimate) and a very high rate of bridal pregnancy (30–40%)”. Mind that unmarried cohabitation was probably not strongly related to patterns in social homogamy. An analysis of street homogamous marriages in Ghent, among which were the marriages between spouses who were cohabitating before marriage, shows that these marriages did not differ from other marriages in terms of their homogamy pattern (CitationVan de Putte, 2005).