Poppel, F. van, Bloothooft, G., Gerritzen, D., Verduin, J. (1999). 'Naming for Kin and the Development of Modern Family Structures: An Analysis of a Rural Region in the Netherlands in the Nineteenth and Early Twentieth Centuries', The History of the Family, 4, 261-295.

NAMING FOR KIN AND THE DEVELOPMENT OF MODERN FAMILY STRUCTURES:

AN ANALYSIS OF A RURAL REGION IN THE NETHERLANDS IN THE NINETEENTH AND EARLY TWENTIETH CENTURIES

FRANS VAN POPPEL, GERRIT BLOOTHOOFT, DOREEN GERRITZEN AND JAN VERDUIN

• Abstract
• Introduction
• The meaning of naming for kin
• How did naming for kin develop in the Netherlands: two hypothesis on the degree of naming for kin
• Differences in naming for kin by social class and sex of the child
• The research area
• The sources
• Methodological issues
• Problems in spelling and name variation
• A hierarchy for naming
• Patterns in naming for kin
• The sequence in namesakes of children
• Naming the first child
• Necronymic naming
• Death of father of mother
• The second forename
• Homonymic naming within the same family
• Differences by social class and over time
• Forerunners of change
• Discussion
• References
• The authors

Abstract:

Our conclusion is that the early rise of the conjugal family, living independent of extended family, and the new attitude that recognized the child as an autonomous individual had no impact on the degree of naming for kin. In a more general sense, doubt is raised on the idea that changes in family structures and mentality directly express themselves in naming practices.

INTRODUCTION

During the past two decades sociologists, demographers, anthropologists and historians have started to use first names to study a wide diversity of historical subjects. Examples of work along this line include studies in which first names given to children were used to study the development of social taste, identification and distinction (Besnard 1984; see also Besnard & Desplanques 1986, Schnapper 1984, and Lieberson and Bell 1992 for modern names), and the degree of interaction between persons from different ethnic groups (Watkins and London 1994). Several historical studies have been published in which forenames were used to chart alterations in the structure of the family over time and to identify changing relationships between the lineage, kin and wider community. This topic has been a central part of the work of French and American historians (see in particular the contributions in Dupâquier et al. (1984), Smith (1984, 1985, 1994), Tebbenhoff (1985) and Main (1996)). The basic assumption of much of this work is that the information on names given to children offers a direct behavioral expression of a variety of attitudes regarding the family, the kinship network and the wider social environment (Smith 1985).

Although names given to children make up a valuable source of information for historical and sociological research, some comments have to be made on the direct relationship which sometimes is assumed to exist between the mentality of the population and the naming pattern. It is true that changes in naming for kin may be interpreted as manifestations of changes in the relation between kin. A decrease in the proportion of children named for grandparents and aunts and uncles may for example be understood as an indication that the relation with these kin members has become more detached; yet conversely the custom to name children for these family members cannot be interpreted simply as an expression of a strong attachment between the members of the family network. A custom to name children kin which is strictly adhered to, can become something obvious and in that case the relation between naming for kin and the attitude towards the members of the kin network has to be interpreted cautiously. This is even more the case when one tries to relate naming patterns to changes in the wider social environment and when changes in naming are interpreted as manifestations of these social or attitudinal changes. Research has shown that radical political and religious changes such as the Christianization, the Reformation (at least in the Netherlands), the French and Russian Revolutions and Hitler’s assumption of power did not result in fundamental changes in the naming pattern (see respectively Kohlheim 1977 and 1988; Gerritzen 1994, pp. 217-218; Dupâquier 1981, p. 144; Davis 1968 and Seibicke 1991, p. 122). Nor is it plausible that economic and social changes inevitably lead to changes in naming patterns. Examples of this may be found in the work of Mitteraurer and Alhaug.

Given the historical correspondence between transmission patterns in the naming system and the transmission of property through inheritance- the son sharing the name of the head of the family also had a claim on the material possessions, the presumptive heir was named for the testator - it is tempting to conclude that it was the transition of an economy based on inheritable property to one based on individual wage labor that has caused a decline in the percentage of children named for kin. Mitterauer, however, has argued that this economic transition has been a long process that took place much earlier than the change in naming for kin. The same lack of correspondence between social changes and changes in naming patterns can be observed as far as another aspect of the changing naming pattern in the twentieth century is concerned, to wit the new corpus of names and the more individualized naming pattern. These new patterns started among members of the educated bourgeoisie, although one could have expected that precisely this group, which was more than others blessed with cultural capital, ought to be the guardian of the naming tradition (Mitterauer 1993, pp. 421-422). Another example can be found in the work of Alhaug, and concerned the possible relation between naming practices and changes in the position of women. Towards the end of the last century Norwegian women lost their traditional dominant position in dairy farming such as cheese making as dairy factories took over their tasks. This weakening of the position of women relative to that of men did however not run parallel to the observed increase in the frequency to name women for men. The custom to derive women’s names from men’s names had already developed in the beginning of the nineteenth century, long before the change in the position of women took place (Alhaug 1990, pp. 142-143). The relation between naming for kin and the development of family structures thus is a complex one, and research into this relationship urges for reflection on the relation between naming and mentality.

The meaning of naming for kin

Initially, in the custom of naming for kin, the passing on of a forename of a family member to a newborn child, the idea of bringing an earlier generation to life is to be found. In the course of time the idea that ancestors lived on through their names has faded into the background and naming the child for a member of the older generation became a tradition. In addition to the idea of reincarnation, the practice to name for kin was a way to express that the child was part of a wider family network. Zonabend (1984) has argued that in historical populations the person who is giving the name to a new-born child delivers a message of a familial or social meaning. The child-naming expressed that the child was part of a genealogical chain stretching forward over the generations. There is a historical correspondence between transmission patterns in the naming system and the transmission of property through inheritance. For a long time in history, naming for kin and transmission of the estate (or, in some cases, transmission of a honorary position) were in the eyes of the community synonymous: the son sharing the name of the head of the family also had a claim on the material possessions, the presumptive heir was named for the testator (Sangoi 1987; Tebbenhof 1985; Zonabend 1984).

Infant naming practices in the past (as is the case in many non-Western societies today) thus reveal, as Scheper-Hughes (1992, pp. 413-415) has put it, a 'thinking in roles'. This implied that the necessity to transmit according to the rules the symbolic and real heritage of the family, a heritage of which the first name was an element, had priority over the unique personality of the child (Schnapper 1984). That the desire to respect the naming rules (and the person who was named for) was stronger than the idea that every child had a right to his or her own individual name, becomes evident from two phenomena: the fact that several children in the same family might be given (a variant of) the same name and the fact that a new-born could inherit the name of an older, deceased sibling (De Moel 1975; Smith 1985; Imhof 1985). Both practices can be interpreted as an indication that the child is not looked upon as an autonomous personality with an unalienable right to an 'individualized' self (Scheper-Hughes 1992; Smith 1985).

From historical studies of child-naming practices in the Netherlands we know that naming for kin was the custom in the Netherlands in the late middle ages (Thierry de Bye Dólleman 1974). For the period 1850-1950, two folklorist studies are available, both based on questionnaires send to local correspondents (Sierksma 1946; Blok 1954). They showed that the first son was usually named after his father's father, the first daughter after the father's mother or the mother's mother, differing according to region and town or village. At the birth of the following child it was then the 'other side's' turn: if the previous name came from the father's side, the mother's father or mother, depending on the child's sex, would supply the next name, and vice versa. The third and fourth children would receive names from the grandparents whose names had not yet been passed on, where necessary by changing the sex of the name (Cornelis (male) > Cornelia (female), Jannetje (female) > Jan (male)). Subsequent children were named after aunts and uncles, where as a rule if the father's parents had been first to have their names passed on, the father's brothers and sisters would be given preference over the mother's (Sierksma 1946). If a child died, out of respect for the naming rules the next born child in the family would be given its name (Boekenoogen 1897, p. 122; Blok 1954).

As the authors of the two folkloric studies considered their own surveys 'unorganized and scanty' (Sierksma 1946) and based on 'rather arbitrary answers' (Blok 1954), they are not really reliable sources for the study of child-naming practices and leave many aspects aside.

In this article, we will try to fill the gaps that the above-mentioned studies have left. For that purpose, use will be made of a large dataset, giving information on names of children and names of members of the kin network over a long period of time in a rural region of the Netherlands, called the Alblasserwaard and Vijfheerenlanden. Since the data are quantitative and derive from a large fraction of the population our dataset first of all offers the possibility to calculate the frequency with which naming for kin was applied. The data refer to the nineteenth and early twentieth century, a period in which many changes in family structure took place and it is interesting to see whether these are reflected in the naming pattern. A further advantage is that this data set offers the opportunity to study which family members were named for. More detailed answers can be given to questions in which the family historians are interested such as the role played by the sex of the child, the influence of the death of one of the parents, necronymic naming, and premarital births and pregnancies. By using information on the social position of the parents of the child, the social class differences in name giving practices might be studied. As the data cover a period of 130 years, time trends in naming patterns can easily be depicted.

The organization of this paper is as follows. Section 3 discusses two alternative hypotheses on the level and time trends in naming for kin in the Netherlands. Next these hypotheses are further specified. Section 5 gives a short overview of the research area that is studied and the sources that have been used for the present study. A discussion of the methodological problems which one encounters in studying naming patterns is given in section 6. In section 7 we try to make explicit the set of rules used by parents, in giving names to their children. In section 8 we derive some conclusions regarding the development of the family and kinship in the Netherlands.

HOW DID NAMING FOR KIN DEVELOP IN THE NETHERLANDS: Two hypotheses on the degree of naming for kin

Two competing hypotheses might be formulated on the characteristics of the naming pattern in the Netherlands during the nineteenth and early twentieth centuries.

On the basis of the existing literature dealing with the history of the family and the demographic history of Western Europe, one might state that the conjugal family, the family that lived independent from extended kin, came into existence in the Netherlands relatively early, and that a new attitude towards children, characterized by intimacy and sentiment, developed there at the same time. To this might be added that at least in the Protestant part of the Netherlands the institution of godparentage ceased to play a role, as a result of which parents had a formal opportunity to abandon the old ritual of naming newborn children for kin and in principle were free to choose any name they wished. One might expect that as a result of these changes, in the nineteenth century only a small proportion of children were still named for kin and that during the second half of the nineteenth and the first decades of the twentieth century the custom to name for kin has further decreased its strength.

On the other hand, more restricted information, based on scattered data from various sources gives the impression that the early rise of the nuclear family may well have taken place without bringing along a fundamental change in the importance of the wider kin network. Viewed from this perspective, one might assume that naming for kin was still the dominant practice until very recently and that it did not undergo fundamental changes during the nineteenth and early twentieth centuries

Hypothesis 1. Naming for kin decreased during the nineteenth century

During the period of the Republic (1581-1795), Dutch household structure departed in a fundamental way from that of the neighboring countries (Van der Woude 1972; Damsma 1993, pp. 34-41). Only few nuclear families lived together with relatives and three-generation households and living-in siblings were extremely rare. Dutch households were on average smaller, more tightly organized and more independent of extended family intervention than elsewhere in seventeenth-century Europe (Schama 1987, p. 386), giving the impression of a high degree of family individualization. Legislation and jurisdiction regarding degrees of kinship within which marriage was prohibited, comments of moralistic authors on relations with relatives, published personal documents etc. all showed that the contacts with the circle of relatives did not constitute obligations of a penetrating nature (Haks 1982). Analogous to this rise of the conjugal family, a new attitude towards children developed characterized by intimacy and sentiment. It is assumed that the Dutch were among the first who recognized the child as an autonomous individual, who had a claim to the affections and passionate attachment of the mother. According to Schama (1987, pp. 481-561; see also Gélis 1986, p. 314), the Dutch Republic was first of all a Republic of Children. Numerous iconographic sources reveal the praxis of a new mode of managing children, unparalleled outside the Republic, and characterized by emotional attachment of Dutch parents to children. Dekker (1995, pp. 211-212; see also Dekker and Groenendijk 1991) has convincingly shown on the basis of a large set of personal documents such as diaries, autobiographies etc. that the approach of Dutch parents to their children was characterized by deeply felt parental love. After 1750, the attitude towards death and the reaction which the community expected when parents had their children robbed from them by death, changed: parents more and more often proceeded to keeping the memory of their deceased child alive and the Christian resignation toward personal loss became under dispute. The need to portray the deceased child had grown during the seventeenth and eighteenth centuries, particularly among the Protestants, where it replaced earlier ways to keep the memory of the deceased alive (Dekker 1995, pp. 224-225)..

In much of the writing about the history of the family, the notion is stressed that Protestantism has played a significant role in the establishment of the modern family through its exaltation of the household above the community and through its strengthening of the bond between husband and wife and between parents and children. Protestantism might have played a role in decreasing frequencies of naming for kin also in a more direct way, as the role of godparenthood, a practice which was closely associated with child-naming practices, changed fundamentally during the Reformation. As a consequence of the reduced theological significance of godparenthood and the continuing criticism of the outward show that often took place when godparents appeared at baptism (Olthuis 1908, pp. 191-193, 199-206) - godparenthood gradually dropped out of use and has been absent from Dutch Protestantism since the beginning of the nineteenth century (Oskamp 1988, pp. 75-78, 82-89). This abandonment of the institution of godparentage could in principle undermine the old naming rituals, and free parents to choose any name they wished (Main 1996). Several authors have suggested that in areas where extended kin no longer had to play a role as godparents, parent-centered naming of the firstborn of both sexes as a result became popular than naming children for more distant kin such as the grandparents. That was for example the case in New England, where the father or the father and mother together presented children for baptism (Smith 1985; Main 1996).

One might expect that as a consequence of this more narrow range of kin naming for kin in the Netherlands played a less important role than elsewhere and that its importance further decreased in the nineteenth and twentieth centuries. Due to the stronger emphasis on the individuality of the child one might in the same period expect a decrease of the percentage of children named for a deceased older sibling. By breaking with the tradition to name their children for kin, parents could show that the influence of the family at large had decreased (Besnard 1984; Dupâquier 1984). They could demonstrate their belief that every child had a right to his or her own individual name, and could express that they themselves as name givers were persons completely differing from the child, with a different position and self-image (Mitterauer 1993, pp. 381-391). In particular the decline in the proportion of first sons and first daughters who did share forenames with parents or grandparents could then be considered indicative of the decreasing strength of the bond between relatives (Smith 1985; Simon 1989). The decrease in necronymic naming at the same time might indicate that parents began to perceive their children as unique individuals.

How convincing the information on the early rise of the conjugal family might be, restricted information from several sources gives the impression that the changes in household structure and composition have left essential parts of the kin network unaffected until very recently.

Ishwaran concluded in his study on family life that Dutch culture even in the 1950s could be characterized by a strong emphasis on family unity and on close contacts with the wider kin group (Ishwaran 1954, pp. 96-108). Given the fact that his study dealt exclusively with the urban middle-class in the western part of the Netherlands, a group which is considered to serve as a model of the modern family type, the results of this study are ver revealing. Ishwaran concluded that a considerable interest in relatives existed even when the family units lived at great distance from each other. On family festivals, such as wedding anniversaries, family units gathered from widely separated areas. People possessed an intimate knowledge of grandparents, great-aunts and great-uncles, and easily recognized their cousins up to the first and second degree. The kinship obligations that were still felt in the 1950s were mainly economic in times of need, when the individual or the family as a whole met with hard times, or social ones such as attendance at family occasions. The binding force which often held three generations together and rendered social contacts easy and possible was the grandmother. Other relatives such as uncles, aunts and first cousins had, but to a somewhat lesser extent, the same functions as the grandparents. For two rural municipalities in or near the research area (Arkel and Kedichem), a more restricted study of value orientations and living circumstances was held in the end of the 1950s (Douma 1961). This study showed that families kept rather frequent contacts (had regular visits, correspondence, or called) with own parents, brothers and sisters and that contacts with nieces and nephews, aunts and uncles were much less common. Although independence of the kin network was highly valued, help of family members in case of illness, economic distress etc. was rather frequent. The sovereignty of parents in matters of upbringing of children was strongly stressed and advice and intervention of kin were not appreciated: only the grandparents were allowed to give their opinion on the education of the children.

A second indication that until recently relations with the wider kinship system were very important comes from the discussion on the revision of the Civil Code in the 1950s. The 1912 Poor Law had stipulated that relief would only be given by a public institution in those cases when other individuals, bound by law to help their relatives, were not in a position to help. The Civil Code among others singled out the spouse, (grand)parents and (grand)children as liable for the costs of each others maintenance. Such costs had to be met whether or not these relatives shared a household. The obligation of individuals to support their relatives remained in force almost unaltered until the 1960s. It was only after World War II that a vivid discussion developed on the maintenance obligations between grandparents and grandchildren. A majority of Parliament decided in 1956 that due to the fact that the emotional bond between grandparents and grandchildren had decreased so strongly, the maintenance obligation between them could better be abolished (De Regt 1985).

The 1950-surveys, as well as the discussion on the Poor Law suggest that feelings of solidarity with the members of the extended family were rather strong until recently in different segments of Dutch society. Parents might therefore have continued to name their children for kin thus expressing this solidarity with the wider kin group. Naming for kin will in that case have been the dominant practice during the nineteenth and early twentieth centuries and may not show a fundamental change in the period studied here.

Differences in naming for kin by social class and sex of the child

In the above, two competing hypotheses were formulated relating to the general frequency with which naming for kin was applied and to the time-trends in naming for kin. On the basis of the literature, more specific assumptions may be formulated on the development in the pattern of naming for kin for boys and girls and for different social classes.

Strong indications exist that boys were named for kin more frequently than girls. Seibicke maintains in general terms that the principle to name for kin was applied to boys in particular (Seibicke 1982, p. 117). Andersen makes mention of a survey of Carstensens in Niebüll and Deezbüll in Die Bökingharde (Northern Germany) among 100 families with at least two sons and two daughters in the period 1750-1850. It appeared that 80% of the first two sons were named for their grandfathers, 10% for their great-grandparent and 10% for others; 66% of the girls were named for their grandmothers and 6% for their great grandmothers (Andersen 1977, p. 158). Berger argues that the operation of the patriarchal society is identifiable by the fact that the son as family heir has to carry on the tradition of the family in his forename as well (Berger 1967, p. 308; this survey relates to the twentieth century). Debus too points to the relationship between the higher frequency of naming for kin amongst boys and the idea that they, more than girls, are considered as the persons who carry on the family (Debus 1988, p. 56; this survey too relates to the twentieth century, in particular the post World War II period). The contemporary naming pattern in the Netherlands is also characterized by the fact that boys are named for kin a little bit more frequent than girls (Gerritzen 1996b).

Apart from the assumption that boys are named for kin more frequently than girls we also expect that due to the dominance of the patriarchal principle in society, which ensures greatest power, prestige and influence for males, names from the paternal line will be given more frequently to children than names form the maternal line. Tebbenhof (1985; see also Simon 1989, p. 51 and Burguière 1984) has suggested that a tendency to emphasize one lineage over the other may reflect a conjugal imbalance of power. Symbolic permanence in the family in the Netherlands indeed rested with the male side: in a legal sense an imbalance of conjugal power existed (Braun 1992). On the other hand, Dutch inheritance practices emphasized that property flowed equal through both the male and female lines and actively promoted the equality of all heirs (Verrips 1983, pp. 51-53; De Haan 1994, pp. 55-60; 79-80). Naming in the paternal line can also be observed in the family name, which runs through the father’s line, and in the patronym, the first name of the father which is added to the forename of the child. Although this custom can partly be explained by the need to indicate the descent from the father, it can also be interpreted as an expression of patriarchal ideas. The fact that all children receive the father’s surname may have reinforced the tendency to give priority to names from the paternal line. We therefore expect that the perpetuation of the names from the male kinship line was considered more important than that of the female line. In some social classes, this tendency may be more pronounced than in others. Research has for example shown that among farmer’s families the frequency of contact with the family from the paternal side was higher than that with the family of the wife. Douma (1961, p. 566) has related this to the fact that women usually resided on the farm of the husband or his father, as a consequence of which she was more involved with the family of her husband and her family went out of sight.

Two factors are usually mentioned in explaining social class differences in naming for kin: property and the degree of contact with the outside world.

Several authors have suggested that given the differences in the importance that inheritance practices have for the various social classes, differences in naming by social class could also be expected. Upper class families were able to transmit to their children economic and cultural capital such as property, formal and informal knowledge, social status etc. This disposed the inheritants to a sense of continuity, to the feeling that one has certain obligations towards the family. In Hingham (Mass.) the wealthiest group in the town placed more weight on the generational depth of their families and was more likely to name the first child for a grandparent and eventually to name a child for the maternal grandfather or maternal grandmother (Smith 1985). The situation of the upper class is more or less comparable with that of farmers and artisans. Paternal authority and domestic dependence, families still functioning as units of production, less separation between work place and family home and the family as a place of learning and service were characteristic of this group (Mitterauer and Sieder 1982, pp. 134-138).

In the lower social classes, the bond between the generations was less important from an economic point of view. At a relatively young age, children from this class entered the labor force, and the character and conditions of labor life could easily lead to an alienation from their social background and regional origin. The mental and emotional life in adulthood was determined to a lesser degree by the family at large and depended more on the situation in the nuclear family. One would thus expect a lower degree of naming for kin (in particular naming for the grandparental generation) in the lowest social class, reflecting the family relationships, typical for this class, which in their turn depended on their laboring conditions.

A second factor which must be taken into account in studying social differences in naming practices is the different degree of contact which social classes have with the outside world. Renewal in naming practices often is the consequence of a relationship that a group establishes with people living outside their usual world. In this respect, higher social classes have an advantage as they often have earlier and more frequent contact with new developments taking place elsewhere; lower classes followed at a distance (Leys 1976, pp. 153-156; see also Bozon p. 1987; Boutier and Perouas 1984 and Andersen 1977, p. 204). In the past, the modernization of naming practices in the lower classes therefore almost always were depended on and following those amongst the upper layers of society . The above arguments give rise to conflicting expectation on the development of the naming pattern. We expect that in the lower social classes a smaller proportion of children is named for kin and at the same time we hypothesize that new developments in the naming pattern will first be seen in the higher social classes.

THE RESEARCH AREA

In this article we make use of a dataset, relating to a region called the Alblasserwaard and Vijfheerenlanden (for short: the Alblasserwaard). The research area is situated in the Rhine delta in the heart of the Netherlands and consists of two adjoining polder regions: Alblasserwaard and Vijfheerenlanden (see Figure 1). The area is bounded to the north, west and south by distributaries, and to the east by a high dyke situated between the Lek and Merwede rivers which runs from north to south. The total number of inhabitants in this region in 1830 was 42,700, a number which increased to 80,500 in 1900 and 97,400 in 1930.

Figure 1. Map Showing the Location of the Research Area

Agriculture, in particular cattle rearing, was by far the most important economic sector in this region. Hemp was grown on small, heavily-fertilized fields but its importance declined in the nineteenth century and increasing emphasis came to be placed on cheese-making and, in the twentieth century, on pig farming. Willow coppices had been cultivated along the riverbanks of several villages and these provided the raw materials for making barrels (outside the region) and baskets (within the region). The rivers near to various villages were fished. A typical farm was about 15 hectares in size and in most cases owned by the farmer; roughly 20 percent of the farms were rented. Depending on the size of the farm and their own particular family circumstances, farmers employed non-live-in permanent workers and day laborers and single, live-in farmhands and maids. Farmers were serviced by local craftsmen, shopkeepers and merchants. More sophisticated facilities were obtained from several small towns such as Gorinchem and Leerdam which functioned as regional centers. During the nineteenth century, the agricultural emphasis of several villages on the banks of the rivers declined as a result of the advent of shipyards and their associated engineering works. Several brick factories were established in the easternmost part of the research area.

The Alblasserwaard made up an integral part of the Dutch economy. Agriculture was specialized and market-oriented, the transport system was relatively well-developed (the region was linked by sailing vessels and coaches with nearby cities like Rotterdam and Dordrecht) and industry had developed from the beginning of the nineteenth century on. Almost the whole population of the Alblasserwaard and Vijfheerenlanden belonged to the Dutch Reformed Church, the Hervormden, or to the Gereformeerden (Calvinists), a dissenting group with an even more orthodox tendency than the Dutch Reformed (Verrips 1983, pp. 300-302).

THE SOURCES

The sources for this study consist of a sample of published and unpublished genealogies of families and cover a period of 130 years in the nineteenth and early twentieth century. The data relate to the rural population of the area, in particular to families formed in the period 1820-1940. For the largest part the data were extracted from unpublished genealogies. Data on 51 lineages could be included in the database. For 23 lineages, data related to different branches, covering the whole period whereas for 28 lineages the data were of a more restricted character, concerning for example only one branch covering the whole or a large part of the period.

In principle, the database includes the following variables:

1. For the married couple: forenames and surnames, forenames and surnames of the parents, place and date of birth, place and date of marriage, rank order of the marriage, place of residence, occupation and date of death.

2. For all births: forenames, sex, date of birth and date of death.

Missing data were as far as possible supplemented by consulting the birth, marriage and death registers of the municipalities concerned, and other relevant sources. The date of death was included whenever the child had died before it had reached age one. For those children who died after their first birthday, the date of death was also very often available. Each apparent irregularity in the reproductive process, like a birth interval longer than two years between two consecutive births, or a last birth at a relatively low age of the mother, was checked. In total, several thousands checks, additions and corrections were done. Nonetheless, when studying names shortcomings such as typing

errors in computerizing records from the original, inadvertent spelling errors made by the official originally taking down the information, unreadable handwriting, misunderstanding of the information on the part of the recording official, lack of care by the individuals or the officials in the actual spelling of information on the records, letter substitutions or omissions which the genealogists did not consider worth correcting or which due to fashion become acceptable ways of spelling, will always remain attached to the data.

Our final database included 1,439 men; 1,358 were married only once, 79 had married two times while two were married three times. Thus, there were 1,522 marriages (1,358 + 79x2 + 2x3 = 1,522) for which we had information on the complete reproduction process. The database only included women who were in their first marriage. This last requirement is necessary because otherwise naming may have been influenced by children from earlier marriages of the mother. The database contained information on 9,596 children. The fertility level in the region thus was very high: only 59 out of the 1522 marriages (3.9%) remained childless; more than 37% of all marriages had 10 or more children. Marriages contracted in the period 1800-1839 on average had 7.2 children (stillborn children included), those contracted in the period 1840-59 and 1861-79 6.9 children. It was only after 1870 that fertility started to decrease: to 6.3 in the cohort married in 1880-99, to 5.1 in cohort 1900-19 and to 3.8 in cohort 1920-39. Child mortality was high: 18.9% of all male children and 17.9% of the female children died in their first year of life; rates first increased and declined only after 1880 from a maximum of 25.0 per 100 live-born children to a value of 3.8 per 100.

To study the differences between social classes, we divided our study population into three groups, according to the occupation of the father at the time of marriage or at the time of birth of his children. Three groups were distinguished: laborers (490), farmers (661) and the petty bourgeoisie (371). This last group included self-employed artisans, shopkeepers, skippers and merchants and a few lower level professionals. Agricultural and other laborers were combined into one group.

METHODOLOGICAL ISSUES

The methodology of studies dealing with the relationship between the kinship structure and naming patterns is rather straightforward: one asks whether children, differentiated by sex and birth order, shared the same names as their parents, grandparents, or other kin. If two family members are found sharing the same first name, one infers that the younger was named for the older (Smith 1985; Dupâquier 1981). It is thus only on the basis of correspondence in names between the child and a member of the kin network, and without any further information on the intentions of the name givers that we have to decide whether or not the child has been named by the parents for a relative with the same name.

For a complete study of naming-practices, it is necessary to reconstitute families over at least three generations to identify all members of the kin network, including uncles and aunts of the children, susceptible to being chosen as namesakers. One thus needs complete family reconstitutions of three families: that of the paternal grandparents, that of the maternal grandparents, and that of the parents themselves. This was the case for 126 families only. In seven of these families the man married twice. From these second wives we had brothers and sisters in two cases only. Because of this low number we decided not to analyze the naming related to aunts and uncles from the second marriage. In addition to this set of data, we had the set of 1,439 families which could be analyzed completely as far as naming for grandparents and parents from first and second marriages is concerned.

Research into naming for kin first of all has to determine whether there is correspondence in name between the child and a member of the kin network. If there is a name correspondence with a relative we assume that the child has been named for this relative. Before being able to ascertain whether there is such a correspondence, two major problems have to be solved:

(1) what name correspondence is needed to consider a child to be named for a relative,

(2) how do we count occurrences if there is more than one candidate, for instance, a father, a grandfather and an uncle with the same forename, or even candidates with the opposite sex?

PROBLEMS IN SPELLING AND NAME VARIATION

The most strict requirements we can pose on name correspondence between a child and a relative is full correspondence in spelling plus a correspondence over all names that constitute the forename. For many reasons these requirements would be too rigid. For example, the spelling of forenames may show variation, for example Cornelis becoming Kernels, there could be phonetic variation (Marrigje becomes Marigje), or the ending (Adriana becomes Adriaantje) or the form of the name, i.e. the choice of an abbreviated form of the full name (Jan versus Johannes, Neeltje versus Cornelia) may differ. But also looser correspondences between names exist such as between Gerrit and Gerard. A name may also switch sex: grandmother Gerrigje, grandson Gerrit; uncle Teunis, niece Teuntje (in traditional Dutch naming practice, the diminutive form was mainly used for women, so the diminutive endings, -tje, -ke, -igje, -chien etc. can be seen as sex markers). In such cases one may be tempted to choose for an underlying etymological basis of forenames. Unfortunately, it is not sure that such an argument has been used by the parents when they named the child. We then may run into an incorrect interpretation of a correspondence between names. Ideally we should be able to distinguish between variants and what were considered to be genuinely different names. Yet, we have chosen for etymological base names and we considered forenames with equal base name as variants. To determine the etymological base name, use was made of Van der Schaar (1994). The 843 different forenames of all persons mentioned in the database (318 different male forms and 525 different female forms) could be assigned one of 185 different base forms. A correspondence between two names was considered to exist if the names had at least one base form in common. Some names were ambiguous with respect to the base form in that more than one base form could apply (for example the base form of Marrigje could be Maria and Margaretha). In these cases we assigned two base forms.

A HIERARCHY FOR NAMING

Another methodological problem we come across when studying naming for relatives is that there can be more than one relative with the same base name as the new-born child. On their turn, these relatives may have been named for each other or for the same ancestor. These complexities may contaminate the insight in the naming pattern. For that reason we have preferred to include in the counting of name correspondences not all possible cases of naming for kin but only those considered most relevant and probable. We have used a chain of inferences of higher and lower probability according to which a name correspondence might be interpreted (see figure 2).

Figure 2. Hierarchical Structure Used for Interpreting Name Correspondence

If a child has the same name as both X and Y, we assume that the child was named for the oldest of X and Y. In our scheme, this is the relative that has a higher position in the hierarchy. We furthermore assume that fathers are ranked above the brother of the mother, and that mothers are ranked over sisters of the father. This implies that naming for more narrow kin is considered more plausible than naming for more distant kin. A child could have been named for more than one relative when relatives had the same name and were at the same level in the scheme. This implies that the total number of cases of naming is higher than the total number of children which were named for kin.

The highest in the hierarchy is the naming of a boy for a deceased brother, or the naming of a girl for a deceased sister. This situation should be considered as the effect of child mortality and therefore these occurrences are counted separately. That doesn’t alter the fact that, in particular for children dying at higher ages, the name correspondence expressed the parent’s wish to name the newborn for the deceased child in stead of the desire to restore the disturbed naming pattern.

As can be seen, grandparents have the same rank, as have uncles and aunts. No further detailed ranking among these can be justified. A minor complication concerns the uncles and aunts, because there may be more than one with the same name. For instance, the father may have several brothers with the same name, although the younger all have died and the elder has been named for them. In such a case we only count the name correspondence once.

If we count the naming for relatives according to this scheme, we will miss the special situation when the son is named for the father while the father in his turn is named for his father, or when the daughter is named for the mother and the mother is named for her mother. Since we considered these types of naming for parents of importance as they show a special emphasis on maintaining a certain forename in every generation, we have counted occurrences separately. We also have to realize that correspondence between the name of the child and that of the father may in principle also be understood as naming for a great-grandparent or uncle/aunt.

A further complication may arise when a child gets more than one name (this happens in only 16.5% of all cases). It is very well possible that the names that constitute the full forename where each chosen for an individual reason. Consequently, our main research concerned correspondences for each name of a child separately, although we also looked for full name correspondence. Because also relatives can have more than one forename, basically lots of possibilities exist that could be considered. In practice, however, the second or third forename of a relative often corresponds to the first forename of another relative. If we give priority to a correspondence to a first forename over a correspondence to the second or next forename, the likelihood that we have to consider a correspondence to a second or next forename is that low that we decided to consider the first forenames of relatives only.

In the next paragraphs we will discuss our analyzes of the naming for kin in the Alblasserwaard. The following aspects will be dealt with: (1) the patterns in naming for kin: how many children are named for their grandparent, their parents or their aunts and uncles? (2) the sequences in namesakes of children: what kind of relation exists between the position within the family and naming for relatives? (3) the naming of the first child (4) naming for a deceased sibling (necronymic naming) and the influence of the death of the father or mother on the naming pattern (5) the influence of the death of the father (6) naming for kin in the second forename (7) homonymic naming within the same family and (8) differences between social classes and the developments over time.

PATTERNS IN NAMING FOR KIN

Tables 1 and 2 present the results of the analyzes of naming for kin in the Alblasserwaard for the first forename (83.5% of the children has only one first name). Table 1 shows the result of the smaller set based on family reconstructions that include aunts and uncles, while table 2 shows the results for the full set, not including aunts and uncles.

We will first discuss the summaries of the tables. The children are divided in three groups. The first group (about 19%) has been named for an earlier child and actually shows the influence of child mortality. The second group is named for relatives and detailed results for this group are given in the main table. For the third group of children we could not explain their names. This group comprises about 22% of the names of the children in the full set, whereas it includes only 3.5% for the set including aunts and uncles. The latter figure learns that almost every child was named for kin. This indicates that one really needs a full family reconstruction to arrive at an understanding of naming strategies. Nevertheless, it does not imply that our full set is useless in that respect: as long as the naming concerned grandparents or parents, which are high in our hierarchy of name sakers, precise figures are available. Cross-sex naming is likely to be overestimated for the full set, however, as aunts and uncles are higher in the hierarchy than relatives of the opposite sex.

The main tables show the numbers of kin for which children were named. To ease the interpretation, the results are presented in two ways. (1) Because the number of grandparents and parents each equals the number of families investigated, we computed for these groups of kin the probability that a person was named for in a family by dividing the counts by the number of families. (2) For every relative, or group of kin when it concerned aunts, uncles, older brothers and sisters, we also calculated the percentage of the total number of children that were named for him/her/them. In the latter total, the children named for an older child were excluded (and are therefore put between parentheses) to eliminate the influence of child mortality. If grandparents, aunts or uncles of paternal and maternal side were of the same name, we counted the naming two times. Therefore, the total number of cases in which naming took place is higher than the number of children named for kin. This effect is in the order of 10-15%, as is also shown in the summary. For the father and mother the number of cases is added in which the father's father or the mother's mother had the same name in order to show the fraction by which the father and mother were named for.

Table 1. Patterns in Naming for Kin -- Christian names: subset including aunts and uncles (126 first marriages, 7 second marriages) (please refer to text for further explanations).

Table 2. Patterns in Naming for Kin -- Christian names: full set excluding information on aunts and uncles (1439 first marriages, 81 second marriages, 2 third marriages) (See text for further explanations)

As could be expected, the detailed results show the highest percentage of naming for grandparents, parents and aunts or uncles of the same sex. The grandparents are the most important group that is named for: in the full set, 63.8% of the sons and 66.6% of the daughters is named for one of his/her grandparents. Naming for the parents (including 28% double naming with a grandparent) is not unusual: 15.6% of the sons had the same name as his father and 11.1% of the daughters shares the name with their mother. It has to be stressed that naming for a father/mother could be intended as naming for a great-grandfather/mother. In the more restricted set, aunts and uncles make up a major category in relative terms: 33.6% for sons, 24.2% for daughters. The lack of these data for aunts and uncles translates for the total set into more cross-sex naming. Table 1 shows however that cross-sex naming is fairly infrequent, a total of 4.8% for sons and 12.8% for daughters. The difference between sons and daughters in this respect mainly originates in the naming of daughters for grandfathers (7%). As Gerritzen (1996b) has shown, many female names are derived from male names whereas the number of male names derived from female ones is very small. This factor may explain why daughters are named for grandfathers but sons hardly are named for grandmothers.

Girls and boys are named for almost equally frequent (sons 96.4% and daughters 94.5% if naming for deceased siblings is excluded), but naming in the paternal line is applied more often. When it comes to the grandparents, the parents of the father are significantly more often chosen (p<0.01; based on the full set); the father too scores significantly higher than the mother (p<0.01). Amongst aunts and uncles, a difference between the male and female lines cannot be observed (see table 1).

THE SEQUENCE IN NAMESAKES OF CHILDREN

The above results are all related to accumulated counts, where the order of naming within a family was not considered. In table 3 we present for each of the (grand)parents the probability distribution that the n^th grandchild of the same sex is named for him or her. These probabilities are based on the full set. For the uncles and aunts we can only present the counts for the nth child in the subset of 126 families (also presented in table 3). In all these cases children that were named for an older child are neglected since they only reflect child mortality.

Table 3. The order in which relatives were named after as reflected by the probability of n^th child of the same gender being named after a (grand)parent (data based on full set) and the number of aunts and uncles the n^th child was named after (data based on subset of 133 families)*

It appears that the grandfather from father's side had by far the highest probability to be named for in the first son (it stands to reason that this son is only in approximately half the number of cases also the first child; see the next paragraph). The grandfather from mother's side had its peak for the second son with wide distributions for the father and the uncles. The third son is most often named for an uncle. For the daughters, the grandmothers both had a maximum to be named for in the first daughter, but the distribution for the grandmother from mother's side is wider and lower. For the aunts, too, the third daughter is most often named for one of them.

NAMING THE FIRST CHILD

The choice of a name for the firstborn of either sex usually carries special significance for the parents, relative to subsequent offspring. Although Dutch law did not recognize primogeniture, the mores did to a certain extent. This is for example shown in the authority patterns among the siblings of the nuclear family. The eldest child had a certain share in the responsibility for the care and socialization of younger brothers and sisters (Ishwaran 1959, pp. 145-146).

In particular naming of the first child is characterized by an outspoken preference of naming for kin from the paternal line: the data for the full set indicate that when the first child is a boy, he is named for the father’s father in 79.3% of the cases. In 21.8% of the cases, he is named for the mother’s father. When the first child is a daughter, she is named for het father’s mother in 64.0% of the cases whereas she is named for her mother’s mother in 26.7% of the cases. These results agree more or less with those of the surveys of Sierksma (1946) and Blok (1954) who stated that the Alblasserwaard was one of the regions where the first child was named for the father's father or mother.

The order by which children were named for can also be investigated explicitly. We did this for both sets but will present results for the full set only. From a comparison with the smaller set with aunts and uncles it appeared that in the full set the naming for unknown kin in most cases referred to aunts and uncles. Again we did not consider children named for an older child. Because there are very many possible sequences we limited the search to those sequences that were found in more than 1.5% of the families. Results are presented in table 4, firstly for the first son, secondly for the combination of first and second son, and thirdly for the combination first, second and third son. The same results are shown for daughters. The results for the first son and the first daughter conform with the previous result that the first son is in most cases named for the paternal grandfather. When families in which no son was born are left aside (8.1%), 64.9% of all first sons were named for their grandfather (49.3 plus 6.1 plus 4.2 divided by 100-8.1). Despite this clear preference, in 35.1% of all families another relative is chosen first. Of course, most of these cases relate to families in which the son was not the firstborn child but was preceded by a girl. For the first daughter too, a clear preference for the mother of the father could be found (48.8%), (for families with daughters: (37.5 plus 3.6 plus 2.3 divided by 100-11.0) but in 51.2% of the families another relative is chosen. Again, this usually meant that the daughter was not the firstborn child but was preceded by a boy.

For the more complex sequences with two or three children the only consistent pattern that arises is that the grandparents are named for first and are followed by uncles, aunts and parents. Deviations from this pattern occur quite regular, and also within the general pattern all kinds of variations are found with as the most common pattern <father father / father mother / uncle or other> for sons and <mother father / mother mother / aunt or other> for daughters.

Table 4. The order in which children were named after relatives: individual figures for sons and daughters, for the first child, second children, and first three children combined. Data based on the full set of 1,439 families*

Note: A difference of more than 0.03 is significant for probabilities higher than 0.30; the difference for probabilties of less than 0.10 should exceed 0.02 to be significant; all at p<0.01.

A special group of first born children are the 51 children (27 sons, 24 daughters) that were born before marriage and were legitimated at the time of marriage. Naming for kin was common even in this situation, but it is remarkable that grandparents are named for relatively infrequent and aunts and uncles relatively frequent (see table 3). In this sense, their naming pattern deviates from that of the other firstborn children. On the other hand, the majority of these children, particularly the sons, were named for the parents of the yet unmarried father: 13 sons were named after the father of the father and only 3 after the father of the mother; 10 daughters were named after the mother of the father and 7 after the mother of the mother. Part of the explanation probably lies in the fact that in a number of cases, a marriage was already scheduled before the child was born (children which were not legitimated were not included in our study). In that sense the situation of the couple resembles that of families in which a child was born within a period of seven months after marriage. Among Dutch orthodox Protestants the number of these 'enforced' marriages was much higher than among any other Christian group (Miedema 1989). In the Alblasserwaard too, a large proportion of firstborn children was conceived before marriage. Children born within seven months of marriage made up 42.5% of all firstborn legitimate children; in addition to that, another 8.1% was born within 7-9 months after marriage. Remarkably enough, orthodox Protestants have a very strict moral code as far as premarital sex is concerned and an enforced marriage is seen as the confirmation of a relationship that was sinful from the start. Enforced marriages were therefore not blessed in church unless a public confession of guilt had taken place (Verrips 1983, pp. 130-135). Nevertheless, sexual contact before marriage was a common matter and experimentation was widespread as one disliked 'buying a pig in a poke'. In addition to that, the fact that the first child was on its way was also a cause for celebration, since procreation was considered a sacred duty (Miedema 1989). Only by keeping in mind the paradox between the strict moral code and tolerant conduct, one can explain that most of the children that were legitimated by marriage were named for kin.

NECRONYMIC NAMING

The desire to respect the naming rules could have as its consequence that the next born child inherited the name of the older, deceased sibling (necronymic naming). To study this we took the set of first sons and daughters that had died before they had reached age five as a starting point. It appeared that only about 27% of the deceased children was not named for in a younger child. When those families are studied in which a deceased child was followed by at least one child of the same sex, 86.9% of the deceased boys was named for a in the next boy and 84.7% of the deceased girls was named for in the next girl. However, when a deceased boy was followed on by a girl, 26.1% of them was named for the deceased child and when a deceased girl was followed by a boy, this percentage was only 10.1%. This difference can be explained by the fact that transposing a boy's name into a girl's name (such as Jan becoming Jannigje) is always possible, whereas the other way around is difficult in the case of a female base name (Geertruida, Elisabeth, Margaretha). There is no significant variation of this attitude across social groups or in time. The last observation means that we don’t find any influence of the strong decline of child mortality after 1880 (from a maximum of 25.0 per 100 live-born children to a value of 3.8 per 100).

One might suppose that the particular personality of a child dying soon after birth would be less apparent to its mother and father than that of an older child; it would not have had time to grow into its name and make the name, in a real sense, its own for keeps. Consequently, one would expect that naming practices, in an era of high infant mortality, would display an inverse correlation between age at death and transfer naming, with infants more likely than older children to have necronymic successors. In reality, there is hardly such a difference between both groups of children: of the 1,471 non-last children who died during their first year of life 82% of the boys and 78% of the girls were named for in a next child; for the 495 non-last children who die after their first birthday (the average age at death was four year), these percentages were respectively 75 and 72. The differences are too small to assume that parents made a distinction between infants and older children in transmitting the name of a deceased child.

DEATH OF FATHER OR MOTHER

A special situation arose when the father died during the pregnancy of his wife or when the mother died in childbirth. In particular the death of the father had a strong impact of the choice of the forename: the 21 children of fathers who died during pregnancy all were named for him, even the daughters. As only one mother died in childbirth, the naming pattern in this situation could not be ascertained. What we could do is examine how children were named in those families in which the mother died within one month after the birth of the child; for a part of this group we may assume that their death could be anticipated. Of the 22 daughters, 14 were named after their mother; in view of the low frequency with which children were named after their mother (around 10%), this number is higher than could be expected by chance alone. Of the 12 sons only one had a name which corresponded with that of his mother (Cornelia > Cornelis).

Another unusual family situation arose after the death of one of the spouses followed by a remarriage of the surviving spouse. As remarrying widows were not included in our study, the analysis here is restricted to remarrying widowers. In only 10 out of 60 second male marriages, a correspondence could be observed between the name of one of the children from the second marriage and the name of the deceased former spouse. Although Sogner and Dupâquier (1981; see also Thierry de Bye Dóllemans 1974) have suggested otherwise, it was not the rule that children of a remarriage were given the name of the first spouse.

THE SECOND FORENAME

For the second forename the number of cases of unknown naming is higher than for the first forename (14% versus 3.5% for the set with aunts and uncles). There may be three factors responsible for this. (1) We interpret the second name only in terms of the forename of a relative, but the second name could well find its basis in another second name, especially if a double name is named for as a whole. Further analyzes show that this is the case for 18.2% of the boys and 13.9% of the girls with a double name (full set). (2) The second name may have been used for a more liberal choice of a name by the parents (Dupâquier 1981; Simon 1989, p. 136). In this respect it is interesting to know that Wilhelmina, the name of the Dutch queen during the period 1898-1948, is in the top-10 of second names (and Willemina as well) and not in the top-25 of first names (Willemina neither). (3) The difference can be partly related to reduced accuracy due to the lower number of names involved. The more detailed results for the second name related to naming for kin are in line with the results for the first forename.

HOMONYMIC NAMING WITHIN THE SAME FAMILY

A phenomenon which not directly bears upon naming for kin but which nevertheless might be explained by this custom is the presence, within one family, of living children bearing the same name (homonymy). This certainly was not uncommon in the Alblasserwaard: amongst sons we found 84 of such couples, amongst daughters 40 (taking the base names as the starting-point ). In most cases, homonymic naming concerned a common base name, making therefore some variation possible in the base form such as Arie and Adrianus and Antonia and Teuntje. Notwithstanding the fact that mortality of infants was high and that very frequently a new-born child inherited the name of an older, deceased sibling, we are not just like that allowed to take it for granted that when we find two identical names in one family the older child has died. Homonymic naming could be observed more frequently among twins than among newborns: 14 of the 139 twins received the same name. This higher proportion might perhaps be explained as a preventive measure taken by the parents to guarantee that they would have a same-name successor as they recognized the high death risks of twins (see De Moel 1975, pp. 43-44). From the fact that two children within the same family could be given the same name we might deduce that the rules of naming for kin were very strictly applied. For, this situation was a consequence of naming for different members of the kin network with identical forenames, whose turn it was, according to the rules of the game, to act as namesaker.

DIFFERENCES BY SOCIAL CLASS AND OVER TIME

We expressed the expectation that social class differences in naming practices were characterized by higher frequencies in naming for kin amongst farmers and the petty bourgeoisie than laborers. This hypothesis could not be confirmed: only tiny differences between social classes were found for the full set. Amongst laborers, 92.4% of all firstborn boys and 89.1% of the firstborn girls was named for their grandparents whereas for farmers and the petty bourgeoisie these figures were respectively 92.8% and 95.4% (for boys) and 89.7% and 85.8% (for girls).

More detailed information is given in table 5. In this table, a distinction is made between the sex of the child that is named for and naming for the paternal and maternal line is given separately. Amongst the petty bourgeoisie a slightly stronger preference existed to name the first son for the father’s father (85.1%) than was the case among laborers (75.7%) and farmers (78.7%). (As children might show a name correspondence with more than one member of the kin network, totals in this table do not sum up to the figures give in the above).

Table 5. Percentage of first grandchildren named after a grandparent, by gender and social class (full set)

Social classes did differ in the degree in which they gave multiple names to their children. In the Alblasserwaard there has been an increase in the number of children that were given a double name over the 19th century. Between 1815 and 1845 less than 10% of the children (both sons and daughters) were given a double name, after 1845 this figure rapidly increased to about 20% in 1875 which percentage was maintained over the remaining observation period (until 1930). This result is also exemplified by the number of families that had at least one child with a double name: for the petty bourgeoisie this amounts to 62.5% of the total number of families, whereas this percentage is 52.7 and 45.4 for the laborers and farmers respectively. Earlier studies, also in the Netherlands, indicate that this fashion started earlier among girls than among boys and that the origin lays in the highest social classes and was adopted later on by the lower social classes (Ebeling 1993, p. 36; Sangoi (1987) for a comparable development in France and Simon (1989) for Germany). In Figure 1 the development is given of the percentage of multiple names per social class (the number of triple forenames is that low that we did not consider these in further analyzes). As is clear from the figure, the petty bourgeoisie in the Alblasserwaard started to give more double forenames to their children, the other two classes followed. The second observation is that the new fashion was applied to boys and girls in the same measure. That the petty bourgeoisie was the pioneer in giving multiple names might perhaps be explained by the fact that they came into touch with other regions more frequently than others. Therefore they adopted changes in the naming habits relatively early (Leys 1976, pp. 153-154).

Figure 3. Development of Percentage of Multiple Names by Social Class.

More or less in line with the differences in the tendency to use multiple names were the differences between social classes in the degree of variety in naming. This uniformity in naming was measured by the concentration index, indicating how strong a handful of very popular names dominated. Among farmers' sons, 55.3 per cent received one of the ten most popular names. Among laborers that was only 48.7 per cent and among members of the petty bourgeoisie even only 41.9 per cent. For girls' names the same social class gradient was found: the concentration index was respectively 37.1, 32.6 and 28.7 per cent. One has to realize though that the degree in which naming was concentrated on a restricted number of names depended partly on the degree of homogeneity of the regional origin of our study population and the family relationships that existed between the members of our population.

Changes over time in the degree in which children were named for kin were hardly observable: even in the most recent period almost each and every child was named for kin. Naming for kin thus did not decrease in importance, and that was the case for each of the social classes. Data for the full set indicate that of the firstborn children born in marriages which were entered into in the period before 1850, 93.5% of the boys and 85.7% of the girls were named for grandparents: for marriages in the period 1850-1900 these percentages were respectively 94.5% and 88.8%, and for marriages entered into after 1900 respectively 93.1% and 90.4%. Amongst girls thus even a slight increase could be observed. During the same period we find an increase in the degree in which children were named for the father’s father (75.6% > 78.8% > 83.9%) and the mother’s mother (21.8% > 27.0% > 30.1%), as is shown in table 6.

Table 6. Percentage of first grandchildren named after a grandparent, by gender and period of marriage (full set)

The stability of the custom to name for kin can also be observed in the more restricted set with aunts and uncles for the unknown naming - for that matter all names from the usual naming pool - is not concentrated in the last period. Even these exceptions to the rule therefore cannot be interpreted as a first indication of the abandonment up of the tradition to name for kin.

FORERUNNERS OF CHANGE

Naming for kin has a conserving effect the naming pattern, as it is a system which leaves only little room for change (see Miedema 1982, pp. 176, 196 and 199 and Ebeling 1984, p. 309). Yet in the Alblasserwaard too there are signs visible which point to changes in the naming pattern. These do not concern however the abandonment of the old pattern to name for kin as such but variation in form in transmitting the name: choosing an other spelling of the name, an other ending and an other form of the name. Again these changes stretch over a long period of time as usually decades pass before a name is given again. In looking at these variations in form, a tendency towards a "de-regionalization" of names appears to exist in the Alblasserwaard (as reference we used a sample from the census of 1947). A clear tendency can be observed toward a replacement of names, spelled in a way typical for the province of Zuid-Holland or for the region, by names which have a much wider diffusion. This de-regionalization of names and the growing preference for names which were common in the Netherlands as a whole, for example AdriaantjeÞAdriana 34 versus AdrianaÞAdriaantje 11 times, GerrigjeÞGeertje 26 versus GeertjeÞGerrigje 12 times, was not very outspoken among men but very clear among women (see also Gerritzen 1996a and Gerritzen to be published). Andersen observed in his research the same development: around 1850 he observed a strengthening of the tendency to replace old dialect forms of names by the official form (Andersen 1977, 204).

We postulate that this tendency to deregionalization, this change in the pattern of name variation happened under the influence of the integration of the region in the national community. For a large part of the nineteenth century, people living in rural areas felt on the whole more integrated into their community than was the case for urban dwellers: the members of the community were stronger bound together in terms of common values and norms, had stronger neighborhood ties, and most marriage partners were from the same area (Ishwaran 1954, pp. 214-226). The Dutch countryside lost many of these characteristics as a consequence of an increased degree of participation in the national community (Knippenberg 1988), in particular during the period after 1910 (Verrips 1983, p. 44). On the scale of the locality, the greater exchange of people and ideas with the outside world, the declining use of local dialects, the partial replacement of regional newspapers by national ones, the establishment of a largely uniform curriculum in free compulsory primary education, improved communication via postal services, railways and roads, conscription with the forced temporary migration it implied, the advent of the national administration with its uniform documents, industrialization, and the economic integration, might lead to the abandonment of many time-honored customs, of which naming-patterns were only one aspect (Rice 1979).

DISCUSSION

The main conclusion from our study is that naming for kin was the dominant aspect of the naming pattern in the part of the Netherlands that we studied, the Alblasserwaard. Of all the members of the kin group it were the grandparents that had by far the highest probability to be named for in their children. The majority of families honored three and more frequently four of the grandparents by transmitting their forenames on to their children. The second group consisted of aunts and uncles of the child: it usually only was their turn after the grandparents were named for. Naming for the parents also occurs; they normally are placed third. Only a few percent of all children bears a name from outside the family. These case do not imply a renewal of the naming pool as these name all belonged to the standard stock.

The outspoken preference to express the continuity of the lineage by naming the firstborn son and daughter for grandparents and not for parents shows that this custom was not only a Mediterranean practice, as Smith (1984; 1985) has suggested, but was also the rule in a northwest-European country as the Netherlands (and Germany, see Simon 1989).

On the basis of our results we may conclude that the early emergence in the Netherlands of the conjugal family, living independent of extended family, and the new attitude that recognized the child as an autonomous individual did not have an impact on the frequency of naming for kin. Interesting in this context is the fact that there was hardly any difference in naming for a deceased child between children who died in their first year of life and children who died at higher ages. The presence within one family of living children bearing the same name also indicates that naming was not used to individualize children. The respect for the naming rules was stronger than the idea that a name had to express the identity of a completely unique person.

Our second hypothesis, suggesting that even when the conjugal family unit is dominant, ideas related to lineage and to extended kinship and obligations to the kin group can remain active, therefore has our preference. Even when the married couple and their children are central, parents might have a sense of the multigenerational dimension of the family, and attach importance to the extended kin, an importance which can be exemplified by the transmission of the forenames of extended kin to their children, as well as by other actions such as visiting each other, offering support and advice and feeling obligations towards each other.

One is tempted to conclude that the rule to name for kin was a reaction to the fact that the nuclear family lived apart from the extended family: naming for kin is a way to express that although the nuclear family is spatially separated from the larger family, an emotional bond with the other members of the kin network still exists.

It appeared that naming daughters for kin was as common as naming sons for kin. Yet again, in the Alblasserwaard a preference existed for the paternal line as the source of names, at least for grandparents and parents. Furthermore, the majority of parents chose the name of their eldest child, regardless of its sex, from the paternal line. This can be explained by drawing on historically rooted conceptions of the roles of males and females in society. Another explanation may be found in the need to mark the descent of the child: it is more obvious to do so in the paternal line than in the maternal one. Noticeable is here the fact that all children, the daughters included, of fathers who died during pregnancy were named for their father. Nonetheless, on the whole the naming pattern shows a reasonably balanced distribution of names from the paternal and maternal lines. This indicates that people felt the need to stress the equivalence of the two families of origin and to strengthen the ties between the two sides by alternating the name-linkages.

Given the fact that inheritance played a more important role among farmers and petty bourgeoisie than among laborers, we expected differences in naming for kin between social classes. We did however not find differences in the degree to which social classes adhered to the naming pattern. In this context one might point to the fact that Van der Schaar (1953) stated that in the western part of the Netherlands, which was 'always more or less democratic', social status differences in naming have never been large. Even in the Middle Ages, a strong difference in baptismal names between the more distinguished families and the farmers and commoners could not be observed.

In our research area, the use of multiple forenames started earlier and was more common among the petty bourgeoisie than among laborers and farmers. We suggested that this and the tendency to a decrease in the dominance of a restricted number of names might be explained by the fact that they were more frequent in touch with the urban environment which made them more open to change than the others groups.

Naming for kin remained very frequent in the period that we studied here. A comparison of the name forms which underwent changes during the transmission on the next generation nonetheless showed that the tendency existed to chose for those forms which were less regionally-bound.

Could the high and constant frequency of naming for kin even in the early twentieth century be explained by the fact that almost the whole population in our area belonged to a group, characterized by a very specific religious model of the world and a distinct style of life (Verrips 1983, 300-302) (thus suggesting that our study population is not at all representative for other rural or industrializing Dutch rural areas in the period concerned). Orthodox tendencies predominated among the Dutch Reformed and Calvinists in the area. Both groups were characterized by the fact that they stressed the obedience to the authority of the parents, who were invested with that authority by God, and the emphasis they laid on the fact that children had to pay loyalty, love, and respect to their father and mother. In more general terms too, compliance and submission to authority was valued high. Following differences in predispositions between the sexes, man and woman have different tasks and these differences lead to a different upbringing of boys and girls: the boy is predestined for public life and the girl for motherhood (Sturm 1988, pp. 125-154). The stress on authority, obedience, family and inequality between the sexes, although only differing in degree but not in kind with the situation in most other groups of Dutch society, could have stimulated the conservation of a naming pattern which stressed naming for kin and preferences for names from the paternal line.

Yet it would be incorrect to relate this continued existence until at least the first decades of the twentieth century of the custom to name for kin only to these specific conservative characteristics of the Alblasserwaard. After all, naming for kin has also maintained an important position in contemporary Dutch naming: even nowadays, around 45% of all children is named for kin (Gerritzen 1997b). The use of these namesakes in most cases is restricted to the official names though, for the first name used in day-to-day life is almost always a modern name (Gerritzen 1997a; see also Dupâquier et al. 1987, p. 29; Bosshart 1973, pp. 26 and 43, and Berger 1967, pp. 307-308). The attribution of a second or third forename makes it possible to respect the traditional pattern of naming and at the same time to express the individual taste and creativity of the parents.

As our study did not produce a unambiguous relationship between naming and mentality the question remains how to interpret the absence of change in the pattern of naming for kin: are we allowed to conclude, given that naming for kin continued to determine the choice of the name, that the strength of the family ties remained constant during the period that we studied? Or do the results indicate that the changing family relationships which resulted from a diversity of social and economic changes had no hold on the custom to name for kin? An answer to this aspect of change has to come from research into the role which is played by naming for kin in contemporary naming. By studying differences between parents who refrain from naming for kin and parents who respect this tradition insight might be gained into the relationship between attitude towards the family and the custom to name children for kin.

REFERENCES

Alhaug, G. 1990. "Womens names in -ine - a great name fashion in Norway in the 19th century."Pp. 137-144 in Proceedings of the 17th International Congress of Onomastic Sciences, edited by E.M. Nähri. Vol. 1. Helsinki: University of Helsinki/Finnish Researh Centre for Domestic Languages.

Andersen, Chr. 1977. Studien zur Namengebung in Nordfriesland. Die Bökingharde 1760-1970. Bräist: Bredstedt.

Berger, H. 1967. Volkskundlich-soziologische Aspekte der Namengebung in Frutigen (Berner Oberland). Bern: Verlag Paul Haupt.

Besnard, P. and G. Desplanques. 1986. Un prénom pour toujours. La cote des prénoms hier, aujourd'hui et demain. Paris: Editions Balland.

Besnard, Ph. 1984. "De la sous-exploitation des prénoms dans la recherche sociologique." Pp. 51-60 in Le prénom. Mode et histoire, edited by J. Dupâquier, A. Bideau and M.-E. Ducreux. Paris: Editions de l'école des hautes études en sciences sociales.

Blok, D. P. 1954. "De naamgeving van het eerste kind." Mededelingen van de Centrale Commissie voor onderzoek van het Nederlandse Volkseigen 6: 7-8.

Boekenoogen, G.J. 1897."Persoonsnamen.", Pp. 107-131 in De Zaanse Volkstaal; Bijdrage tot de kennis van de woordenschat in Noord-Holland; met aanvullingen door G.J. Boekenoogen en K. Woudt en een voorwoord door Jo Daan. Vol. I. (Leiden:). (Reprint Zaandijk: J.Heijnis Tsz.1971).

Bosshart, L. 1973. Motive der Vornamengebung im Kanton Schaffhausen von 1960-1970. Diss. Freiburg/Schweiz: Paulusdruckerei.

Boutier, J. and L. Perouas. 1984. "L'évolution des prénoms en Limousin du XIe au XXe siècle." Pp. 67-78 in Le prénom. Mode et histoire, edited by J. Dupâquier, A. Bideau and M.-E. Ducreux. Paris: Editions de l'école des hautes études en sciences sociales.

Bozon, M. 1987. "Histoire et sociologie d'un bien symbolique, le prénom." Population 42: 83-98.

Braun, M. 1992. De prijs van de liefde. De eerste feministische golf, het huwelijksrecht en de vaderlandse geschiedenis. Amsterdam: Het Spinhuis.

Brinkgreve, Christien and Bram van Stolk. 1997. Van huis uit. Een onderzoek naar sociale erfenissen. Amsterdam: Meulenhoff.

Burguière, A. 1984. "Prénoms et parenté." Pp. 29-36 in Le prénom. Mode et histoire, edited by J. Dupâquier, A. Bideau and M.-E. Ducreux. Paris: Editions de l'école des hautes études en sciences sociales.

Damsma, D. 1993. Het Hollandse huisgezin (1500-heden). Utrecht/Antwerpen: Kosmos.

Davis, P.A. 1968. "Soviet Russian Given Names." Names 16: 95-104.

Debus, F. 1988. "Aufgaben, Methoden und Perspektiven der Sozioonomastik." Pp. 41-77 De naamkunde tussen taal en cultuur, edited by Rob Rentenaar en Ellen Palmboom. Amsterdam:Cahiers van het P.J. Meertens-Instituut nr. 1.

Dekker, J.J.H. and L.F. Groenendijk. 1991. "The Republic of God or the Republic of Children? Childhood and Child-rearing after the reformation: an appraisal of Simon Schama's thesis about the uniqueness of the Dutch case." Oxford Review of Education 17: 317-335.

Dekker, Rudolf. 1995. Uit de schaduw in 't grote licht. Kinderen in egodocumenten van de Gouden Eeuw tot de Romantiek. Amsterdam: Wereldbibliotheek.

Douma, W.H. 1961. Het gezin op een verstedelijkend plateland. Wageningen: Bulletin nr. 20. Afdeling sociologie en sociografie van de Landbouwhogeschool.

Dupâquier, Jacques. 1981. "Naming-practices, godparenthood, and kinship in the Vexin, 1540-1900." Journal of Family History 6: 135-155.

Dupâquier, Jacques. 1984. "Introduction." Pp. 5-12 in Le prénom. Mode et histoire, edited by J. Dupâquier, A. Bideau and M.-E. Ducreux. Paris: Editions de l'école des hautes études en sciences sociales.

Dupâquier, Jacques, A. Bideau and M.-E. Ducreux, eds. 1984. Le prénom. Mode et histoire. Paris: Editions de l'école des hautes études en sciences sociales.

Dupâquier, Jacques., J.-P. Pélissier and D. Rebaudo, 1987. Le temps des Jules. Les prénoms en France au xixe siècle. Paris: Editions Christian.

Ebeling, R. A. 1993. Voor-en familienamen in Nederland. Geschiedenis, verspreiding, vorm en gebruik. Den Haag: REGIO-PRojekt Groningen, Centraal Bureau voor Genealogie.

Gélis, Jacques. 1986. "L'individualisation de l'enfant.", Pp. 311-329 in Histoire de la vie privée. Vol. 3., edited by P. Ariès and G. Duby. Paris: Editions du Seuil.

Gerritzen, Doreen. 1994. "Twee revoluties in de persoonsnaamgeving." Naamkunde 26: 207-242.

Gerritzen, D. 1996a. "Voornamen in Friesland bij de volkstelling van 1947". Friezen. Fryske Nammen 10: 25-36.

Gerritzen, Doreen. 1996b. "Seksespecifieke kenmerken van voornaamgeving." Naamkunde 28: 107-136.

Gerritzen, Doreen. 1997a. "Hedendaagse naamgeving in Nederland." Naamkunde 29 (in press).

Gerritzen, Doreen. 1997b. "Motieven bij de keuze van voornamen." Naamkunde,29 (in press).

Gerritzen, Doreen. (to be published). "Unterschiede in der Vornamengebung in der Stadt und auf dem Lande in den Niederlanden im Jahre 1947." in Proceedings of the 18th International Congress of Onomastic Sciences.

Haan, Henk de. 1994. In the shadow of the tree. Kinship, property and inheritance among farm families. Amsterdam: Het Spinhuis.

Haks, Donald. 1982. Huwelijk en gezin in Holland in de 17de en 18de eeuw. Assen: Van Gorcum.

Imhof, Arthur.E. 1985. "From the old mortality pattern to the new: implications of a radical change from the sixteenth to the twentieth century." Bull. Hist. Medicine 59: 1-29.

Ishwaran, K. 1954. Family life in the Netherlands. The Hague: Uitgeverij van Keulen.

Knippenberg, Hans and Ben de Pater. 1988. De eenwording van Nederland. Nijmegen: SUN.

Kohlheim, V. 1977. "Zur Erforschung der Diffusion onomastischer Innovationen." Pp. 1-34 in Beiträge zur Namenforschung Neue Folge. 12.

Kohlheim, V., Zur Verbreitung sprachlicher und onomastischer Neuerungen. In: Beiträge zur Namenforschung Neue Folge 23 (1988), p. 156-176.

Leys, O. 1976. "Sociolinguistische aspecten van de persoonsnaamgeving." Naamkunde 8: 137-158.

Lieberson, S. and E.O. Bell. 1992. "Children's first names: An empirical study of social taste." Am.J. of Sociology 98: 511-554.

Main, G.L. 1996. "Naming children in Early New England." Journal of Interdisciplinary History 27: 1-27.

Miedema, H.T.J. 1982. "Bijdragen tot de voornamengeografie van Friesland. De voornamen van 24 duizend Friezen in 1811-1812." Naamkunde 14:169-207.

Miedema, N. 1989. "The Orthodox Protestants and enforced marriage" Pp. 5-23 in Dutch dilemmas: anthropologists look at the Netherlands, edited by J. Boissevain et al. Assen: Van Gorcum.

Mitterauer, Michael and R. Sieder. 1982. The European Family. Chicago: University of Chicago Press.

Mitterauer, Michael. 1993. Ahnen und Heilige. Namengebung in der europäischen Geschichte. München: Verlag C.H. Beck.

Moel, C.J. De. 1975. "Doopnaamgeving te Edam-Volendam in de 17e en 18e eeuw." Bijdragen en Mededelingen van de Commissie voor Naamkunde en Nederzettingsgeschiedenis van de KNAW 29: 5-18.

Olthuis, H.J. 1908. De doopspraktijk der gereformeerde kerken in Nederland. 1568-1816. Utrecht: G.J.A. Ruys.

Oskamp, P. 1988. Doopborgen. Profiel en profijt. Ph. D. Dissertation University of Amsterdam. Amsterdam.

De Regt, Ali. 1985. "Onderhoudsplicht en verhaal van steun 1912-1965." Sociologisch Tijdschrift 12: 405-445.

Rice, J.G. 1979. "Studying the modernization process at the scale of the locality." Pp. 93-106 in Time, space and man: Essays on microdemography, edited by J. Sundin and E. Söderlund. Stockholm: Almqvist & Wiksell International.

Sangoï, J.-C. 1987. "Transmission des prénoms et réproduction sociale en Bas-Quercy xviiie-xixe siècles." Pp. 263-293 in Annales de Démographie Historique. Paris: Editions de l'E.H.E.S.S.

Schaar, J. van der. 1953. Uit de wordingsgeschiedenis der Hollandse doop- en familienamen. Assen: Van Gorcum & Co.

Schaar, J. van der. 1994. Spectrum Voornamenboek (revised by D. Gerritzen and J. B. Berns). Utrecht: Het Spectrum (1962).

Schama, Simon. 1987. The embarrassment of riches, An interpretation of Dutch culture in the Golden Age. New York: Alfred A. Knopf.

Scheper-Hughes, Nancy. 1992. Death without weeping. The violence of everyday life in Brazil. Berkeley: University of California Press.

Schnapper, D. 1984. "Essai de lecture sociologique." Pp. 13-22 in in Le prénom. Mode et histoire, edited by J. Dupâquier, A. Bideau and M.-E. Ducreux. Paris: Editions de l'école des hautes études en sciences sociales.

Seibicke, Wilfried. 1982.. Die Personennamen im Deutschen. Berlin/New York:De Gruyter.

Seibicke, Wilfried. 1991. Vornamen. Frankfurt am Main: Verlag für Standesamtswesen.

Sierksma, K. 1946. De traditie der voornamen. Naarden: Uitgeversmaatschappy A. Rutgers.

Simon, M. 1989. Vornamen wozu? Taufe, Patenwahl und Namengebung in Westfalen vom 17. bis zum 20. Jahrhundert. Münster: F. Coppenrath Verlag.

Smith, D.S. 1984. "Child-naming practices as cultural and familial indicators." Local population studies 32: 17-27.

Smith, D.S. 1985. "Child-naming practices, kinship ties, and change in family attitudes in Hingham, Massachusetts, 1641 to 1880." Journal of social history 18: 541-566.

Smith, D.S. 1994. "Continuity and discontinuity in Puritan naming: Massachusetts, 1771." The William and Mary Quarterly 51: 67-91.

Sogner, S. and J. Dupâquier. 1981. "Introduction." Pp. 1-11 in Marriage and remarriage in populations of the past, edited by J. Dupâquier et al. London: Academic Press.

Sturm, J.C. 1988. Een goede gereformeerde opvoeding. Kampen: J.H. Kok.

Tebbenhoff, E.H. 1985. "Tacit rules and hidden family structures: naming practices and godparentage in Schenectady, New York 1680-1800." Journal of social history 18: 567-585.

Thierry de Bye Dólleman, M.1974. "Gewoonten en gebruiken met betrekking tot de naamgeving aan kinderen vóór 1600." Holland 6: 289-297.

Verrips, J. 1983. En boven de polder de hemel. Een antropologische studie van een Nederlands dorp 1850-1971. Groningen: Wolters-Nordhoff.

Watkins, S.C. and A.S. London. 1994. "Personal names and cultural change. A study of the naming patterns of Italians and Jews in the United States in 1910." Social Science History 18: 169-209.

Woude, A. M., Van der. 1972. "Variations in the size and structure of the household in the United Provinces of the Netherlands in the seventeenth and eighteenth centuries." Pp. 299-318 in Household and family in past time, edited by P. Laslett (with the assistance of R. Wall). London; Cambridge University Press.

Zonabend, F. 1984. "Prénom et identité." Pp. 23-28 in Le prénom. Mode et histoire, edited by J. Dupâquier, A. Bideau and M.-E. Ducreux. Paris: Editions de l'école des hautes études en sciences sociales.

THE AUTHORS

Frans van Poppel is Senior researcher at the Netherlands Interdisciplinary Demographic Institute (NIDI), P.O. Box 1165, 2502 AR The Hague, The Netherlands. Gerrit Bloothooft is researcher at the Department of Computers and Arts, Utrecht University, Trans 10, 3512 JK, Utrecht, The Netherlands. Doreen Gerritzen is free-lance researcher, Groesbeeksedwarsweg 259, 6521 DJ Nijmegen, The Netherlands. Jan Verduin is a former Lecturer in the Department of Geography, State University of Utrecht, Valklaan 17, 3738 GE Maartensdijk, The Netherlands.