2. Shifts in marriage seasonality in Europe, across the nineteenth and twentieth centuries

There is a rich tradition in Europe of studying shifts in marriage seasonality, and the propensity to marry during Lent in particular. Researchers such as Houdaille, Sardon, Lesthaeghe, and López-Gay have explored changes in the dates when weddings were held, and noted the impact of historical events, such as the French Revolution, as well as religious differences between Catholics and Protestants in Belgium.Footnote ¹⁶ Recent studies on West Flanders (Belgium) by Hideko Matsuo and on the Netherlands by Frans Van Poppel emphasise that both industrialisation and evolving ideologies can influence marriage seasonality.Footnote ¹⁷ Theo Engelen's work on seven Dutch provinces also highlights the lasting significance of religion in shaping marriage calendars right up until the twentieth century.Footnote ¹⁸

Examinations of Italy's modernisation, by researchers such as Gabriele Ruiu and Marco Breschi, suggest that couples were persistently discouraged from celebrating their marriage during Lent.Footnote ¹⁹ Similar conclusions have been drawn by Anna Coppa, Emmanuel Sanna, and Maria Enrica Danubio for various Italian regions.Footnote ²⁰ Enzo Lucchetti et al. explored patterns of marriage seasonality in France, Italy, and Spain over the past two centuries, and found that these were influenced by changes in religion, work, and climate.Footnote ²¹

In contrast, studies trying to understand processes of modernisation in England, Sweden, and Germany place more emphasis on economic factors, as exemplified by Ann Kussmaul's work on preindustrial England; Christer Lundh, Martin Dribe and Bart Van de Putte's research on Scania, Sweden; and John E. Knodel's work on a number of German villages.Footnote ²² Notably, in England, an increase in the rate of marriage in December was more associated with industrialisation and the adoption of a new working schedule than with Advent, highlighting how regional differences in the interplay of religious and economic factors can affect marriage patterns, especially in areas where Catholic and Protestant populations co-exist, as in Germany.Footnote ²³

In the case of Spain, it is worth mentioning Àngels Torrents’ long-term analysis of the timing of marriage for an industrial town in the province of Barcelona, which showed that during the seventeenth and eighteenth centuries both agricultural and religious cycles placed restrictions on when marriages could take place, but these restrictions disappeared as the nineteenth century progressed.Footnote ²⁴ Similarly, José Miguel Martínez-Carrión, in his study on the South East of Spain during the nineteenth and twentieth centuries, concluded that, despite the constraints imposed by the agricultural calendar, it was church restrictions on marriage during Lent that were most significant.Footnote ²⁵ Antonio González-Martín has also shown that the number of marriages during Lent in Andorra increased over time between 1606 and 1960.Footnote ²⁶ Barcelona, South East Spain and Andorra each underwent different modernisation processes, and hence show varying levels of secularisation.

The above review shows that when interpreting variation in marital seasonality, an economic and social view tends to be used to explain the changes observed over time. Some of the studies do, however, mention the relaxation of ecclesiastical norms or changes in ideology as possible causes of shifts in the marriage calendar, although these factors are not included in the statistical analyses reported in the papers; they are only mentioned in the interpretative part of the narratives. In this article, both of these themes are taken into account through innovative analysis which considers the political period in which marriages took place, while also recognising that ‘when to hold a wedding’ was a cultural element transmitted from one generation of a family to the next.

3. The political, socio-economic, demographic and family context of the Barcelona Area, in the nineteenth and twentieth centuries

Throughout the nineteenth century European states adopted measures to separate political and religious power. Durkheim (1912) saw the state as the quintessential agent of secularisation,Footnote ²⁷ and it is important to recognise that the relationship between the state and the church was largely influenced by the politics and political ideology of different periods. In the case of Spain, politics across the eighteenth and nineteenth centuries alternated between two regimes: absolutism and liberalism, the second of which could take one of two forms; being either ‘progressive’ or ‘moderate’. Such political swings could have significantly influenced the secularisation process within the country.Footnote ²⁸ There have been three waves of secularisation in Spain to date, with the first running between the beginning of the nineteenth century and the Civil War of 1936–1939, and it is the relationship between the political swings within this period and marriages taking place in Lent that is the focus of our analysis.Footnote ²⁹

Spain's first liberal phase can be described as ‘progressive’ and encompassed the Cortes of Cádiz (1810),Footnote ³⁰ the War of Independence (1808–1814), which is considered to mark the beginning of the liberal revolution in Spain; the Trienio Liberal (Three Liberal Years or Liberal Triennium) (1820–1823); and the years of liberal government during the minority of Queen Isabella II (1833–1843). These years saw the suppression of the Inquisition (although this was briefly restored, along with the monarchy, during the Absolutist Sexennium, between 1814 and 1820); the disentailment or confiscation (desamortización) of church properties; and a decree abolishing the payment of tithes to the church. During the summer of 1835 there were also significant episodes of anticlericalism in various towns in Catalonia such as Reus, Barcelona and Sabadell. These involved assaults on – and the burning of – a number of convents and the murder of over 70 individuals connected to the church.Footnote ³¹ These riot-like events, during the first Carlist war,Footnote ³² along with the killing of friars in Madrid the previous year, acted as the catalyst for the forced expulsion of the regular clergy – viewed as the Carlists’ main allies by the supporters of the Regent Maria Cristina, who were their opponents in the war; the suppression of religious orders, and the confiscation of church property.Footnote ³³ This whole period, from the first to the fifth decade of the nineteenth century, which saw the church lose its economic power and fall under the supervision of the state, is therefore considered to constitute the initial manifestation of the first wave of secularisation in Spain.Footnote ³⁴

The second political phase, which started in the years of Queen Isabella II's majority, 1843–1868, can be categorised as one of moderate liberalism. Although, when the Constitution was written in 1845, during the Moderate Decade which ran from 1844 to 1854, it stated that the Spanish nation was Catholic and relations with the Holy See had been restored. Furthermore, in 1851 a Concordat between the Spanish monarchy and the Vatican state was signed. The Concordat implied that the state declared itself confessional (officially recognising Catholicism as the state religion) and that the church would offer a public service overseen by the state. However, the Concordat also marked the end of the clerical reforms that had begun with the disentailment of 1836, as the aim was now to build a national church with a rational and professional structure. It remained in effect until 1979, apart from 1873–1874 and during the 1930s when it was temporarily repealed.

Spain returned to progressive liberalism for two years during the Progressive Biennium in 1854–1856. During this period, disentailment recommenced, and efforts were made to establish religious freedom. There was then a return to moderate liberalism before progressive liberalism reappeared with the Democratic Sexennium of 1868–1874 (Sexenio Democrático, also known as Six Democratic Years), following the Glorious Revolution of 1868. It is of note that when revolutionary juntas were being formed in Catalonia, the junta in the town of Reus, in the district of Tarragona, argued for equality among citizens, eliminated church privileges, and established civil marriage. The first civil marriage in Spain thus took place in Catalonia in 1868,Footnote ³⁵ although civil marriage was not implemented nationwide until 1870, when the Provisional Law of Civil Marriage made state-sanctioned marriage obligatory for the first time. The Provisional Law was considered one of the triumphs of the 1868 Revolution, although it was moderate in character, allowing a religious ceremony to take place prior to the civil one, unlike in France. In addition, civil marriage remained subject to the same impediments and indissolubility as marriage celebrated in church. The Provisional Law only remained in place until February 1875. Its short duration means that the church was the only institution performing marriages for most of our study period.Footnote ³⁶

In Alicante, one of the few cities in Spain where the implementation of civil marriage in 1870–1875 has been studied, it is observed that 37 per cent of marriages during this period were celebrated in church before a civil ceremony took place, while 28 per cent only had a civil ceremony. How the remaining marriages were conducted is unknown. It appears, however, that while civil marriage was implemented in Alicante, the law's detractors spoke of its limited impact across Spain.Footnote ³⁷ Throughout the progressive Democratic Sexennium, the freedom of worship, assembly, and education, as well as the secularisation of cemeteries, were implemented. The suppression of the Society of Jesus, the Jesuits, was also decreed, and the process of desamortización was resumed, running alongside the reforms laid out in the 1851 Concordat. Nevertheless, the state continued to provide economic support for the church. It is important to highlight that in 1868 the First International was established in Spain (the International Workingmen's Association) with the formation of the country's first anarchist group, following Mikhail Bakunin, in Catalonia. It is also worth mentioning that the First Spanish Workers’ Congress, during which the Regional Spanish Federation of the International Workers’ Association (IWA) was established, was held in Barcelona in 1870. This made a decisive contribution to the spread of the anarchist movement in both Catalonia and Spain which signalled growing scepticism towards both political and religious organisations.Footnote ³⁸ At the same time, federal republicans, who had made political headway in Catalonia during the Constituent Cortes elections of 1869, were promoting the practice of atheism.

After the restoration of the Bourbon monarchy, there was a return to a period of moderate liberalism. This ran from 1874 to 1931, although our study only covers the period until 1880. It is worth noting that in 1876 the Constitution was amended to establish religious tolerance, although still affirming that Catholicism was the state religion. In fact, the Bourbon Restoration also brought a reinstatement of the church's influence within society, as previously suppressed religious orders were reinstated and new congregations founded.Footnote ³⁹

From an economic point of view, the Barcelona Area experienced significant proto-industrialisation in the eighteenth century. This was based on rural industry and on calico factories mainly located in the city of Barcelona. The region's agricultural areas also saw an expansion of cultivated land, irrigation, and crop rotation.Footnote ⁴⁰ This led to further specialisation in the production of wines and spirits, particularly in Catalonia's coastal counties; some of which are included in our study area. Catalonia also specialised in wine productionFootnote ⁴¹ and had close connections with the colonial market. At the end of the century, early industrialisation arrived with the introduction of the mechanical spinning jenny, but the War of Independence, which ran from 1807 to 1814, slowed the pace of industrial development, as did the loss of Spain's American colonies, although it retained Cuba, the Philippines and Puerto Rico. It was not until the 1830s that an unprecedented boom in production began, as the cotton textile industry drove an important surge in industrialisation. This led to a significant decrease in the primary sector and the integration of the factory system in the secondary sector, transforming the region's occupational structure.

The city of Barcelona became one of the first industrial centres in Southern Europe, benefitting from the proximity of its port.Footnote ⁴² The lack of coal in Catalonia encouraged the use of water-powered machinery, and this led to industrial development along the region's rivers such as the Llobregat (which also lay in the Barcelona Area), Besòs, and Ter. It is also worth mentioning the industrialisation of the towns of Sabadell and Terrassa, both centres of the wool industry, as these dominated the regional and national markets. They lay close to Barcelona and had railway connections to the port.Footnote ⁴³

In demographic terms, the Barcelona Area went from having just over 60,000 inhabitants in 1717 to over 370,000 in 1860. This can be explained by the significant growth of its urban nuclei such as Sabadell, Terrassa and the city of Barcelona. Barcelona's population alone increased from nearly 36,000 inhabitants in the early decades of the eighteenth century to almost 250,000 in the second half of the nineteenth century. Over the same period, Sabadell's population multiplied tenfold and Terrassa's sixfold until they both had more than 10,000 inhabitants. This growth has traditionally been attributed to in-migration from rural areas, but it is worth highlighting that migration from other industrial basins, such as those of the Ter and Llobregat, also played an important role, especially in the case of Sabadell,Footnote ⁴⁴ as a notable portion of these migrants were already skilled.Footnote ⁴⁵

The Barcelona Area was also one of the earliest regions within southern Europe to experience fertility decline. This decline began in the first half of the nineteenth century,Footnote ⁴⁶ and has also been considered to be an indication of secularisation.Footnote ⁴⁷ Both the fertility decline and the economic transformation occurred in a society where the stem family, characterised by impartible inheritance, was the predominant family type, present in almost all social groups.Footnote ⁴⁸ In the literature, such families are generally characterised as more traditional and less receptive to innovation.Footnote ⁴⁹

4. The Barcelona Historical Marriage Database and the genealogical reconstitutionpr

In our analysis we used the Barcelona Historical Marriage Database (BHMD), which contains over 600,000 marriage licenses recorded in the Llibres d'Esposalles (Marriage License Books) from 1451 to 1905 for the entire Diocese of Barcelona, of which the Barcelona Area is a principal component (see Figure 1). The Diocese comprised approximately 250 parishes in 1900 and the Marriage License Books recorded the payment of taxes imposed on all marriages before they could be conducted within the Diocese. The fee charged in each case depended on the occupation or social position of the groom. Pope Benedict XII (1328–1423), an Antipope during the Avignon Papacy also known as Pope Luna, initially granted permission for this tax to be collected, with the primary objective of financing the construction and maintenance of Barcelona Cathedral. In this article we focus on the 284,122 marriages recorded between 1715 and 1880 (our full data sample).

Figure 1.

The Barcelona Area in the Spanish and Catalan contexts.

Source: Barcelona Historical Marriage Database (BHMD). Authors’ elaboration.

Each entry in the Licence Books records: the date of payment, which almost always coincides with the date of the marriage celebration; the spouses’ forenames, surnames and previous marital status; the names of their parents; the date and parish of the celebration; and the groom's occupation was also generally noted. Throughout the period of the study, those paying taxes were documented and classified into seven socioeconomic categories, ranging from nobles at the top to artisans and peasants at the bottom. Those applying to marry, but unable to pay the fees, were placed in a further category known as Amore Dei, exempt from paying as they were considered to be too poor. It should be noted, however, that the grooms in such cases typically declared that they had a trade.

The study focuses on the Barcelona Area’ or the Oficialitat de Barcelona, the most populated deanery among the four that comprised the Diocese of Barcelona. The deanery included towns such as Barcelona, Mataró, Sabadell and Terrassa which had experienced proto-industrialisation in the mid-eighteenth century and early industrialisation in the nineteenth before becoming the most industrialised region in Spain (Figure 1).Footnote ⁵⁰ The information on marriages available from the other three deaneries within the diocese contains significant gaps, discouraging its use.

Data harmonisation and coding have been applied to the BHMD. The information on grooms’ occupations were coded using the Historical International Classification of Occupations (HISCO).Footnote ⁵¹ Afterwards, the occupations were converted into social groups using the Historical International Social Class Scheme (HISCLASS).Footnote ⁵² The twelve categories presented in HISCLASS have been regrouped into six groups reflecting the socioeconomic characteristics of the Barcelona Area (See Table 1). We labelled the topmost layer in the social stratification the ‘Elite’. This encompassed all members of the nobility (mainly the lower nobility) as well as those in high positions of institutional power, such as government officials, who are usually placed among Higher Managers in HISCLASS. Below this group lie the liberal professionals, followed by traders (including those in high and medium-levels of commerce), those artisans who belonged to guilds, farmers of all kinds and, finally, day labourers.

Table 1.

Codes and classifications used for the main occupations and social status groups found among grooms in the Barcelona Area (1715–1880)

Source: See Figure 1.

All geographical locations (parishes) mentioned in the Marriage Licence Books have been coded to the municipal boundaries they belonged to in 1860, and all the parishes within the corresponding boundaries have been geo-referenced. This approach allowed us to identify all the rural parishes and major urban areas at that time. Additionally, as outlined in the previous section, we delineated three political phases within our study period so that we could examine the potential impact of different political regimes on rates of marriage during Lent. The first, absolutist, includes the years classified as part of the Old Regime (in the eighteenth century), and contrasts with the two nineteenth century phases of progressive and moderate liberalism.

The nominal information contained in each marriage licence allows genealogical reconstruction, and was used to link together family members in order to examine the existence of intergenerational transmission of marriage patterns. A subset of families (Genealogical Sample) was reconstructed by linking parents’ marriages to those of their children (see Figure 2). The assumption was made that an individual would appear as a bride or groom before being observed as a parent on their child's marriage licence.

Figure 2.

Grooms’ social group structure in all cases (Full S.) and in the genealogical sample (Gen. S.) in the Barcelona Area (1720–1880).

Source: See Figure 1.

Note: In the period 1720–1759, socio-occupational information is derived from parents' marriages in the genealogical sample, as children's marriages are only available from 1750 onwards. Starting from 1760, the information is based on children's marriages.

To carry out the genealogical reconstitution, we used the BuscaDescendències (Looking for Offspring) software developed for the ‘Five Centuries of Marriages’ ERC project.Footnote ⁵³ This software employs an algorithm which calculates three similarity measures between strings, considering the length of each string and the position of each letter. The string comparison is based on the ‘Levenshtein Distance’ and a combination of the ‘Bag Distance’ and ‘Longest Common Substring distances’. Furthermore, the algorithm is adapted to Catalan phonetics. Constraints were also introduced into the linkage process to ensure plausible links between marriages: a minimum difference of 15 years and a maximum difference of 50 years between the marriage of the parents and their child were set. It is important to note that the record linkage can result in multiple links, due to the prevalence of common Catalan and Spanish forenames and surnames, so that connections can be made between unrelated individuals who share the same first and last names.Footnote ⁵⁴ The total number of links generated automatically surpassed 400,000 for the period 1715–1880.

We employed a conservative approach to validate links. First, we considered only those with a similarity score of 85 per cent or higher. Second, we eliminated any duplicated links between parents, permitting only one to one linkage. Third, we only considered licenses which had complete and reliable data on groom's occupation and payment of the licence fee. As a result, the final tally of genealogical connections amounted to approximately 123,000 links, composed of children who married from 1750 onwards.Footnote ⁵⁵ In terms of its socioeconomic composition, the genealogical sample was quite similar to the full sample, as shown in Figure 2.

We should, however, note a potential issue with our record linkage regarding the proportion of linked sons and daughters. In the BHMD dataset, complete nominal information on parents was available only for grooms during the eighteenth and nineteenth centuries, while brides typically did not have their fathers’ name and surname along with their mothers’ first name available. After removing all duplicate entries (children with two potential sets of parents), we observed a discrepancy in the total number of linked sons and daughters, with 32 per cent of daughters and 68 per cent of men in the genealogical sample between 1750 and 1880 (see Table 2).

Table 2.

Descriptive table of the main variables used in the descriptive and regression analyses

Source: See Figure 1.

The study has further limitations to be considered, such as sample selection issues, particularly those concerning migrants. The urban centres of Barcelona, Mataró, Sabadell, and Terrassa served as significant attractors of migratory flows during the studied period, but we are unable to observe the transmission of marriage behaviour in Lent if individuals migrated into or out of the Barcelona Area. Our results are therefore restricted to those families where the parents married within the study area, and children remained within its boundaries. The geographic scope of the study is nevertheless extensive, as the study area encompassed around 90 municipalities by the end of the nineteenth century.

Another important issue is the absence of information regarding the ages of the bride and groom in the Marriage Licence Books. This poses a drawback for demographic analysis, as it hampers the assessment of how age-related factors may have affected marriage trends. This inconvenience is offset, however, by the large number of marriages contained in the Marriage Licence Books and the breadth of their temporal and geographical coverage. Moreover, since we have information on the marital status of grooms and brides at the time of marriage, we are able to use this variable to distinguish between individuals marrying for the first time and those remarrying; as the latter tend to be older than the former, this variable can act as a proxy for age.

5. Methods used to measure and study Lenten Marriages

Our analytical strategy was based on two main steps. The first step was to take accurate measurements of marriage seasonality in order to assess the extent to which marriages took place in Lent in the Barcelona Area during the period 1715–1880. We used two different indeces of marriage seasonality. The first of these was the so-called Seasonality Index (SI). This captures the degree of seasonality within marriages by comparing the number of marriages expected and observed each month across the year. It has commonly been used in the literature when considering socioeconomic change.Footnote ⁵⁶ The second index was the Lent Seasonality Index (LSI). Lent, as is well known, is a 40 day liturgical period based on the lunar calendar, which means that it's timing varies from year to year; it can start anywhere between the 4th of February and the 10th of March. The LSI requires a more accurate calculation of the measurement of seasonality around a specific period in the religious calendar. Moreover, it also offers a novel way to look at the impact of particular political moments on the desacralisation of private lives and the separation of church and state – key element of European secularisation – that took place during economic and social modernisation. It also allows us to observe how families influenced the timing of marriage among their younger generations. The LSI takes the exact days in which Lent took place into consideration, and thus avoids having to use the month of March as a proxy for Lent. While the latter is frequently done, it does not always include all the Lenten days as some can fall in February or April.Footnote ⁵⁷ The SI and LSI can be formulated as follows:

$$SI = \displaystyle{{M_x/D_x} \over {M_y/D_y}}\quad LSI = \displaystyle{{M_q/D_q} \over {M_y/D_y}}$$

M_x is the number of marriages per month, M_y is the number of marriages per year, D_x is the number of days of a given month, D_y is the number of days per year, M_q is the number of marriages in Lent and D_q is the number of days in Lent.

In the second step of our analytical strategy, we tested different determinants that might influence the probability of marrying in Lent in the Barcelona Area. We took the LSI as our dependent variable in order to assess the impact of socioeconomic factors and political regime (absolutism and liberalism)/political tendency of the goverments within liberalism (moderates vs progressives) amongst the cases in the full sample and that of the demographic and familial characteristics in the genealogical sample across the 1750–1880 period (see Table 2). We used linear probability models (LPMs) as this type of model makes the interpretation of regression coefficients for binary outcomes easier than if logit regression is used. Additionally, LPMs enable comparison across samples, which is not feasible with odds ratios.Footnote ⁵⁸ LPMs also provide protection from omitted variables which are not associated with our independent variable, unlike in logistic regression.

We created a categorical independent variable defining the political period in which a couple married, based on the political phases experienced in Spain during the nineteenth century (see Table 2). The Old (absolutist) Regime was assumed to last from 1715 until 1807, while the moderate liberal goverments (in which, for statistical robustness, we included the 1814–1819 and the 1824–1833 periods of absolutism) and the progressive liberal goverments alternated throughout the nineteenth century. Additionally, we incorporated different interactions between the political periods, the individual's social background, and their demographic and familial characteristics in our models, to determine if different combinations might amplify or modify the likelihood that a couple would observe the restrictions around celebrating marriage during Lent. Furthermore, because we expected a linear increase in the number of Lenten marriages over time, all our models included a year trend indicator to control for such temporal trends.

6. Measuring Lenten marriages in the Barcelona Area across the eighteenth and nineteenth centuries

Since the end of the eighteenth century, the Barcelona Area witnessed the emergence and gradual establishment of industrialisation. Concurrently, there was a progressive shift in the distribution of marriages throughout the year. In the period, 1720–1759 – which does not correspond to the political periods mentioned above, as we only seek here to describe the phenomenon – marriages were concentrated in February, while March was avoided. However, by our last period, 1840–1879 – also not corresponding to the political periods –, marriages occurred much more evenly across the year, apart from a concentration in December (see Figure 3). This shift towards a peak of marriages in December has also been observed in other countries, such as Sweden.Footnote ⁵⁹

Figure 3.

Index of marriage seasonality (SI) by month; in Barcelona Area, by a 30 year period, during 1720–1880.

Source: See Figure 1.

Note: The vertical axis displays the Index of marriage seasonality, while the horizontal axis shows months. Each line represents the period in years marriages took place.

Over the long term, marriages during Lent steadily increased in frequency from the early eighteenth to the late nineteenth century (see Figure 4). Growth in the Lent Seasonality Index (LSI) became notably steeper from the 1780s onwards, coinciding with the changes in occupational structure brought about by the relatively early industrialisation of Catalonia and the Barcelona Area.Footnote ⁶⁰

Figure 4.

Index of relative frequency of Lent marriages (LSI) in the Barcelona Area, during 1715–1880, showing political periods.

Source: See Figure 1.

Note: The vertical axis displays the index of relative frequency of Lenten marriages (LSI) while the horizontal axis shows years when marriages took place. The legend on the figure represents the following periods: (1) Old Regime, (2) Liberal Revolution, (3) Absolutist Sexennium, (4) Liberal Triennium, (5) Ominous Decade, (6) Maria Cristina & Espartero Regency, (7) Moderate Decade, (8) Progressive Biennium, (9) Moderate Biennium, (10) Liberal Union, (11) Moderatism Crisis, (12) Democratic Sexennium, (13) Restoration of the Monarchy.

Several marked peaks and troughs in the LSI can also be observed during the first half of the nineteenth century. For instance, the highest value occurs around 1808, the first year of the Peninsular War (1808–1814) during which Napoleon's troops invaded Spain. This tallies with increased non-observance of religious holidays observed in Catalonia during the same period mentioned by Moliner Prada, and the increase in divorce cases demonstrated by Costa.Footnote ⁶¹ Marriages during Lent increased during Progressive periods such as 1814–1819 and 1823–1833, while during the periods when the absolute monarchy was restored in 1814–1819 and 1823–1833 they decreased (see Figure 4). Progressive periods also saw decreases in Easter observance in cities such as Barcelona.Footnote ⁶² Moliner Prada asserts that this was a consequence of a partial secularisation of the people, which was not as evident in other Spanish cities.

The period 1840–1880 was characterised by smaller fluctuations in the occurrence of marriages during Lent. There is a linear increase in the LSI towards zero over time, indicating that Lenten marriages were becoming as normally distributed as other marriages; i.e., the population were no longer observing religious guidelines regarding Easter. This change occurred in parallel with an increase in the number of couples obtaining their marriage license free of charge, as socioeconomic inequality increased in the Barcelona Area.Footnote ⁶³

When social groups are considered, the Elite never showed a seasonal attitude towards Lent, while Farmers were the group most likely to avoid Lenten marriages across the study period. However, both groups were experiencing a decline in relative numbers (see Figure 5). Artisans and, in particular, Traders exhibited a clear evolution in their marriage seasonality, shifting from avoidance of Lenten marriages to more flexible behaviour in the nineteenth century. Throughout the eighteenth and nineteenth centuries, the commercial sector played a prominent role in the formation of the bourgeoisie of Barcelona, making a decisive contribution to the economic modernisation of the region.Footnote ⁶⁴ Traders were involved in both the import of manufactured goods and grain and the export of viticultural products. There was also significant trade with Spain's colonies. By the nineteenth century merchants in the Barcelona Area were highly dynamic with multiple investments and closely tied to industrial development as well as the execution of public works. It is important to note, however, that particularly during the second half of the nineteenth century, merchants, industrialists, and others remained important pillars of the Church in the Barcelona Area, a trend that continued into the twentieth century.Footnote ⁶⁵

Figure 5.

Index of relative frequency of Lenten marriages (LSI) in the Barcelona Area, during 1715–1880, for selected social groups.

Source: See Figure 1.

Note: The vertical axis displays the index of relative frequency of Lenten marriages (LSI) while the horizontal axis show years were marriages took place.

Although each social group presented a different level of observance of the proscription on Lenten marriages, a general increase in the LSI is observed, indicating that there was a common move away from the observance of Lent. There was a significant increase in the number of marriages during Lent in all social groups from the 1780s onwards, even among the farmers in rural areas, which suggests that the change was due not just to economics but that a cultural/religious shift in attitudes to marriage was also taking place. According to evidence provided in episcopal reports from different Catalan dioceses from the early nineteenth century onwards, these new attitudes can be understood as a growth in the loss of Christian values amongst the population as they increasingly distanced themselves from religious practices. Later in the century such attitudinal changes would be associated with the advancement of liberalism, socialism, freemasonry, and Protestantism.Footnote ⁶⁶ Similar attitudinal changes have also been noted in rural Catholic societies in the Netherlands and France where, in the latter case, they would be accompanied by an early decline in fertility (see Figure 5).Footnote ⁶⁷

7. Political and societal factors influencing the likelihood of getting married during Lent

One of the main issues when disentangling the determinants behind the increasing incidence of Lenten marriages is discerning between time trends and the influence of political, cultural and societal factors. We started by using the full sample dataset of marriages to test the dependent variable of propensity to marry in Lent by social group, controlling for the annual trend between 1715–1880. Our first model (Model 1, Table 3) indicates that the social group to which a groom belonged determined the probability that he would get married during Lent. Our reference group in the model was ‘Farmers’ who were the group least likely to marry during Lent, while better-off groups such as the elite, liberal professionals and traders were most likely to do so; they had a 4 to 3 percentage point higher probability of marrying in Lent than the farmers.

Table 3.

Linear probability model regression estimates on the probability of marrying in Lent by social groups and political periods

Source: See Figure 1.

Note: Only those coefficients with statistically robust standard errors (SE) are shown. *p < 0.05, **p < 0.01, ***p < 0.001.

In Model 2 (Table 3), we included the different political periods seen in Spain during the eighteenth and nineteenth centuries. It is noticeable that the earlier Progressive periods (P), such as the Liberal Revolution during the Napoleonic Wars and the Liberal Triennium, showed higher incidences of Lenten marriage. To put the Liberal Triennium in context, this period saw a widespread decline in traditional celebrations and the emergence of politically- and civically-oriented festivities.Footnote ⁶⁸ This trend continued throughout the nineteenth century, albeit at a slightly slower pace, and can be seen in Barcelona during the Moderate Decade (1844–1854), when radical progressivism prevailed and gave wings to subversive popular culture.Footnote ⁶⁹ It is worth recalling that it was during this decade that the Concordat of 1851 was signed, and the Church began to regain some of its social standing before the advent of more moderate liberalism.Footnote ⁷⁰ In any case, all political periods in the nineteenth century consistently show probabilities of Lenten marriages some 1.5 to 4 percentage points higher than those prevailing during the Old Regime in the previous century.

Andrés Gallego (1994), amongst others, has documented the increasing disregard for the traditional marriage calendar in Spain, accompanied by a decline in the observance of religious practice, since the late eighteenth century.Footnote ⁷¹ In the particular case of Catalonia, it is worth highlighting the reports of Catalan bishops Antoni Palau i Termens (1862) and Pantaleón Montserrat Navarro (1868), which depict Barcelona as ‘a spiritually backward city where corruption of customs was fostered by the tolerance of prostitution, gambling, blasphemy, and other vices’.Footnote ⁷²

Overall, when predicting the probability that marriages will take place in Lent at different times, while controlling for time trends and incorporating the interaction with political periods, our models confirm that all political periods in the nineteenth century exhibited a higher probability of Lenten marriages than that of the Old Regime, as shown in Figure 6. While the probability of getting married in Lent was 0.07 in the Old Regime, during the early progressive periods, this probability rose to 0.1 during the Liberal Revolution (1808–1813) and to 0.09 during the Liberal Triennium (1820–1823), and thereafter the probability stabilised at around 0.08.

Figure 6.

Probability of marrying in Lent by political period, 1715–1880, controlling for social group and year trend, in the Barcelona Area.

Source: See Figure 1.

Note: The probabilities were calculated taking the interaction between political periods and social groups into account, and adjusting for the year trend using Model 2 (Table 3).

Next, as shown in Table 4, we successively introduced various political, socioeconomic, and demographic variables related to marriage to examine their potential association with the incidence of Lenten marriages. In Model 3, in addition to the different social groups, we include absolutist (reference category), moderate, and progressive political regimes. Once time trends were accounted for, only the terms of progressive government showed a significant difference from the reference category, displaying an average probability of Lenten marriages that was 1.2 percentage points higher than that of the absolutist regime, as shown in Figure 7.

Table 4.

LPM regression estimates of the probability of marrying in Lent, interactions between social groups, political regimes and sociodemographic characteristics

Source: See Figure 1.

Note: Coefficients and robust standard errors (SE) displayed. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 7.

Probability of marrying in Lent by political regime and social group, in the Barcelona Area 1715–1880.

Source: See Figure 1.

Note: Predicted probabilities based on model 7 (Table 5) using interaction terms (table not shown) in our regressions between the political phase and the social groups. Coefficients and robust standard errors (SE) displayed. *p < 0.05, **p < 0.01, ***p < 0.001.

Table 5.

LPM regression estimates on the probability of marrying in Lent intergenerationally, the Barcelona Area (1750–1880)

Source: See Figure 1.

Note: Coefficients and robust standard errors (SE) displayed. *p < 0.05, **p < 0.01, ***p < 0.001.

In Model 4, we introduce additional socioeconomic and demographic variables, specifically: the proportion of individuals exempt from paying the marriage license fee because of poverty (Amore Dei), and the percentage of couples where at least one of the spouses had been married previously (remarriage). Exemption from the marriage license fee showed a slight positive association with an increased likelihood of being married during Lent. However, the proportion of marriages involving the remarriage of a widow or widower had a more substantial effect as they had an approximately 2 percentage-point higher probability of marrying in Lent than couples where both bride and groom were marrying for the first time. The Amore Dei and remarriage variables suggest that, in some instances, individuals who were facing more vulnerable situations, such as poverty or widow(er)hood, may have been less selective when choosing a marriage date, or they may have preferred to marry during periods when celebrations were more muted, such as Lent which was seen as a time when lavish behaviours were to be avoided. Weddings during Lent did not include a nuptial Mass, which reduced the social significance of the ceremony,Footnote ⁷³ which must have been a benefit for more disadvantaged couples.

In Models 5 and 6, we included interaction terms between the political regimes, remarriage and Amore Dei variables. The interaction coefficients show that there was no additive effect of combining a particular type of political regime with being remarried or marriage tax-exempt, suggesting that all three factors contributed to the occurrence of Lenten marriages separately. This may be because the prevailing political regime is more of a contextual determinant, while remarriage and poverty were more structural factors in determining whether a marriage took place in Lent.

Our final step, shown in Figure 7, was to predict the probability of Lenten marriages under the different types of political regime for each social group (based on Model 7, see Table 5). Amongst the elite, liberal professionals, and traders, there was no difference in the probability of marrying in Lent across the different political regimes. However, amongst Artisans and farmers, in particular, we observe a clear gradient with a lower probability of marrying in Lent under an absolutist regime, followed by successively higher probabilities under moderate and then progressive governments.

8. Was the change in the marriage calendar also a family matter?

In this section, we investigate whether factors such as the sex of an individual or their marriage order within their sibling group, which was taken as a proxy for birth order,Footnote ⁷⁴ affected their willingness to break with the traditional marriage calendar and marrying during Lent. We also consider whether the timing of weddings was indeed transmitted from parents to their children, affecting their attitude towards the liturgical and religious calendar, and generating a sense of belonging to a particular social group. In societies such as Catalonia, where there were strong family ties and the stem family was the predominant family system, such vertical transmission of culture was particularly important, as it helped to sustain the social system. Under the region's inheritance system, the sole heir to an estate received three quarters of its total value as their inheritance the moment they signed their marriage contract; the remaining one quarter of the estate was, by law, distributed among the heir's siblings.

Although the stem family in Catalonia has frequently been associated with rural areas it was the predominant family form across almost all social groups in the region,Footnote ⁷⁵ and adapted completely to proto-industrialisation and industrialisation.Footnote ⁷⁶ Nuclear families also followed a system of impartible inheritance by instituting a single heir. This is because only a stem family could produce another stem family, as the parents would only remain with the heir. The rest of the children would end up creating a nuclear family unless they married an heir/heiress and moved in with their in-laws. Also, in the long run the stem family would necessarily become a nuclear family as the parents would die before their heir child. Thus, many families appointed an heir to whom they could pass on their assets, even if they owned very little. This logic reinforced vertical family structures, with the likely result that attitudes and behaviours towards religion and its practice were passed down from one generation to the next.

We therefore modelled the likelihood of marrying in Lent using the genealogical database (see section 4) which incorporates links between parents and children. In Models 7 and 8 (Table 5), we included different familial and demographic characteristics of both the parents (Generation 1; G1) and children (Generation 2; G2). In Model 7, we observe that the likelihood of the younger generation marrying in Lent was higher if their parents did too: the probability of children marrying in Lent was 1.3 percentage points higher if their parents also married in Lent than if their parents married at another time of year. This strongly suggests that the family is a conveyor of religious attitudes, even although these attitudes are not necessarily traditional ones. In other words, beliefs, norms, values, attitudes, and behaviours are not static, but adapt to changing economic and political times. Although it might have been thought that being female would be associated with marked differences compared to men, this was not the case. Marrying in a different parish from the one in which one's parents married – which was used as a proxy for migrant status – seemed to have a slight effect on the chances of marrying in Lent. Here, the distance migrated could encourage individuals to break with the traditional religious attitudes which they had absorbed during childhood and adolescence. Overall, our findings relating to parents’ behaviours showed that the latter did, to some extent, influence their children's decisions on whether or not to marry during Lent, except when children migrated before they married. Thus, the observed transmission of a preferred time to marry between generations of the same family indicates that the family was indeed a key institution in the transmission of beliefs, norms, values, attitudes, and behaviours.Footnote ⁷⁷

While we observe a positive association between parents’ and children's Lenten marriages, the same does not hold for the transmission of the month of marriage, which suggests that what was being transmitted was not the timing of marriage per se, but a specific inclination or disinclination to marry during Lent.

In Model 8 (Table 5), we included the social group to which the younger generation in the genealogical dataset belonged. In essence, this displays a pattern similar to the one observed in the full dataset. However, it is interesting to note that even when controlling for the children's social group, the coefficient for the intergenerational transmission of non-observance of Lent was still statistically significant; suggesting once again, that parental influence had an independent effect, regardless of the children's social group. The family is thus seen to have been more important than social group or class in influencing marriage behaviour.Footnote ⁷⁸

In the last set of models, shown in Table 6, we included all the sociodemographic and socioeconomic variables we had for both parents (G1) and children (G2), together with the information on the political regimes under which G1 and G2 married. In Model 9, we see that the political regime in which G2 married was correlated with their probability of marrying during Lent, but the latter was not correlated with the political regime in which their parents married. However, the coefficients in Model 9 are always read in comparison to the reference group: absolutism, and during periods of absolute monarchy, as we have seen, the proportions of Lenten marriages was much lower, particularly during the eighteenth century. In Models 10–13, we therefore included dummies for the Moderate periods against all other political regimes (Model 10 and 11) and for progressive periods against all other political regimes (models 12 and 13). When each case is taken separately, we can see that when children marry in either moderate or progressive periods there is a positive correlation with Lenten marriages, and a matching positive effect on the intergenerational transmission of a propensity to marry during Lent (Model 10 considers moderate periods and Model 12 Progressive ones). However, the interaction between political regime and the intergenerational transmission of marriage outside Lent shows no statistically significant or substantial additive effect of having parents who married in Lent and the individual being analysed being exposed to a particular liberal period. In other words, the results show that both structural family influences and contextual political influences coexisted independently.

Table 6.

LPM regressions estimates in the probability of marrying in Lent in the genealogical sample, the Barcelona Area (1750–1880)

Source: See Figure 1.

Note: Coefficients and robust standard errors (SE) displayed. *p < 0.05, **p < 0.01, ***p < 0.001.