Our exploration of demographic behavior in the sample villages begins with an examination of mortality, one of the two core elements of the demographic transition. The determination of adult mortality is difficult with family reconstitution data because of problems associated with the determination of the period at risk for persons whose history of vital events is known only incompletely. At best, the level of adult mortality can be estimated only within some plausible range. Precise estimates of infant and child mortality are far more readily obtained. In this study, analysis of mortality experience is limited primarily to the risks of dying in infancy or childhood and to the special case of maternal mortality; assessment of general adult mortality itself is not attempted, although indications of overall life expectancy are derived indirectly from the infant and child mortality estimates.

Following a discussion of data quality issues and problems of measurement as they relate to estimating mortality, the current chapter presents findings on levels and trends in infant and child mortality, including an examination of neonatal and post-neonatal mortality risks. Estimates of overall life expectancy are derived from the level and age patterns of mortality under age 10. Finally, the seasonal patterns of infant mortality are examined. Chapter 4 then investigates socio-economic and demographic differences in infant mortality, including a detailed examination of the association of infant and child mortality with maternal age, birth order, sibship size, and birth intervals.

Most historical investigations of nuptiality in western Europe have focused on the relatively late entry into first marriage for both sexes and the substantial proportions never married that were typical in the past. Far less attention has been paid to the patterns of marital dissolution and remarriage that were characteristic of the prevailing socio-economic and demographic systems, despite their potential significance for the individuals and societies involved.

The studies of historical patterns and trends of remarriage in Europe that have been done are usually based on data on the distribution of marriages according to the marriage order of the spouses. Results have typically revealed a substantial decline in the proportionate share of marriages involving people remarrying, and a concomitant growth in the proportion of primary marriages (those between never-married men and women). In their analysis of English trends, for example, Schofield and Wrigley argue that while 25–30 percent of those marrying in the sixteenth century were remarrying, this proportion had declined to only 10 percent by the nineteenth century. Data from local family reconstitution studies in France, including those by Cabourdin and Bideau, confirm the existence of a trend, from the sixteenth to the nineteenth century, toward a decreasing proportion of remarriages.

There has been considerably less work done on the actual probabilities and rates of remarriage. In their history of the European family, Mitterauer and Sieder assume that the probability of remarriage underwent a ‘continuous’ decline over the last two centuries, although they provide no supporting evidence.

The genealogies for all the villages in the sample contain at least some introductory material about the village, although in the cases of Öschelbronn, Kreuth, Vasbeck, and especially the two East Frisian villages, the information is quite minimal. Based primarily on the descriptive material contained in the introductory sections, some of the more salient features of the villages' history and then social and economic characteristics can be summarized.

The villages in Baden

Given the close proximity of the four villages in Baden, it is useful to start their description with some information on the general history of the local area. The strategic position between the Rhine and the Black Forest meant that the villages are located in a corner of Germany that was regularly overrun and held by competing armies during the many wars of the past centuries. At various times, soldiers of Sweden, France, Austria, and Prussia were quartered in the area, creating considerable hardship through looting and demands for provisions. There was some respite from war during the middle of the eighteenth century but the area again became a battleground following the outbreak of the French Revolution and in the Wars of the First Coalition. Not until the Napoleonic period was there any extended period of peace again. The degree of seigniorial control varied across villages and over time.

The relationship between mortality and fertility has been a matter of interest since the early days of population research. Concern with the nature and extent of the relationship has heightened since the formulation of demographic transition theory, which attributes central importance to the timing and interdependence of the secular declines in birth and death rates in Western demographic experience. Fertility can influence mortality in a variety of ways and in Chapter 4 the impact of the birth interval on infant mortality was examined. Most interest, however, has focused on the effect of mortality, and particularly infant and child mortality, on reproductive behavior. Recent conceptualization of this relationship has identified several different potential effects. These include an insurance (or hoarding) effect whereby couples have extra children in anticipation of child mortality, a replacement effect whereby couples have replacement births in response to their own actual (as opposed to anticipated) experience with child loss, and an involuntary physiological effect on the interval between births attributable to the interruption of breastfeeding and the consequent shortening of the postpartum non-susceptible period. At a different level, a societal effect is sometimes identified which operates indirectly (and presumably unconsciously) through social customs to adjust the community fertility level to the community mortality level.

The number of births born out of wedlock that were eventually followed by the marriage of the parents, and hence should be classified as legitimate according to the definition used in the present study, is probably biased downward due to two problems. First, some mothers of illegitimate births might marry the father outside the village and hence not have the marriage recorded in the parish registers on which the village genealogy is based. Second, during periods when the name of the father of the illegitimate child is not given in the parish register, the genealogist might not know that the father was the eventual husband of the child's mother, especially if the child died before the marriage or no further mention of the child is found in the parish registers subsequent to the parents' marriage.

The possibility that a number of subsequent marriages of unwed parents could have taken place outside the village is suggested by the fact that a substantial proportion of unwed mothers whose child appears as non-legitimized in the genealogy did not remain permanently in the village where the birth occurred and thus did not remain under observation as far as the local records were concerned. While the departure of someone from the village is only occasionally noted explicitly in the village genealogies, it can be inferred from the lack of a death date or from a mention of the death occurring elsewhere.

Although family reconstitution represents a new technique for organizing demographic data, the approach has been central to the work of genealogists for centuries. Genealogical studies are often based on the same parish records used by historical demographers, but for a number of reasons these studies are of limited value as data for historical demography. Perhaps the overriding limitation is that the genealogist usually chooses the families to be included in his study on the basis of common ancestry or descent; they thus constitute a selective sample which is not likely to be representative of the general population, or even of any clearly defined substrata. Moreover, since most genealogies are based on scattered sources, they are unlikely to be a complete compilation of vital events even for the family units included.

Sources of data: village genealogies

One unusual but relatively unknown exception is the Ortssippenbuch (literally ‘book of local kinsmen’), or village genealogy, which serves as the major source of data for this study. Genealogies of this type are unique to Germany. Unlike most genealogies, in which the births, deaths, and marriages of a particular family line are traced regardless of where various branches of the family may have moved, the village genealogy encompasses the vital events of all families that ever resided in a particular village insofar as these events are recorded in the local records.

Much of the premarital sexual activity in German villages in the past occurred among couples who subsequently married. This is apparent not only from the statistics on prenuptial births already presented in the previous chapter but even more so from the statistics on births that, although born after a couple married, were clearly conceived prior to the wedding date. This chapter focuses on the outcome of such premarital sexual activity, examining more closely the phenomenon of prenuptial births and exploring in detail the phenomenon of bridal pregnancy. Before proceeding to the analysis of these phenomena, however, a brief discussion about several issues relevant to the measurement of bridal pregnancy is in order.

Measuring bridal pregnancy

Fewer problems are involved in the measurement of bridal pregnancy than is the case for illegitimacy and its division into legitimized and non-legitimized births, as discussed in the previous chapter. In the present study, a woman is considered to have been pregnant at marriage if she gave birth by the start of the eighth elapsed month following the wedding date. While normal variance in the biological period of gestation will result in some women being misclassified with respect to bridal pregnancy, the margin of error should be minor. The net result is probably a slight underestimation of the proportion pregnant at marriage since use of an eight-month interval errs on the conservative side, excluding more births that were actually prenuptially conceived than falsely including births postnuptially conceived.

In any demographic study of the past, the validity of the findings depends on the accuracy and completeness of the source material and, if applicable, on the appropriateness of the techniques employed to overcome known shortcomings in the data. In this study, which is based on analysis of family reconstitution data from German village genealogies, the data are generally treated as if they are accurate although, in the case of data on infant and child mortality, only for specified periods of time for some villages. In this appendix, the quality of the data is assessed and the basis for determining the periods during which mortality data are assumed to be accurate is provided.

Accuracy of transcription and reconstitution

While the data for this study come directly from the village genealogies, those genealogies are in turn based on the parish registers and, in some cases, also the civil registers. Thus the accuracy of the data being used depends both on the completeness with which vital events were reported in the parish and civil registers and on the accuracy with which the genealogist transcribed these events and compiled them into reconstituted family histories. Details of a study have been reported elsewhere in which a sample of entries in the registers of four villages was compared with events recorded in the corresponding village genealogies to see if the events were included and accurately stated.

Infant-feeding practices, particularly the prevalence and duration of breastfeeding, are an important determinant of infant mortality in situations where modern hygienic conditions are lacking, and an important determinant of fertility where deliberate birth control is not practiced. Fortunately some information is available on breastfeeding for a number of areas in Germany, at least for the end of the period covered by the present study. There can be no question that sharp regional differences in breastfeeding existed around 1900. Moreover, there is some limited evidence that many of the regional patterns had existed for a considerable period in the past. While the quantity and quality of information varies for the areas in which the sample villages are located, at least some indication of the pattern prevailing around 1900 is available for all. In addition, indirect evidence can be derived from the reproductive histories of couples in the village genealogies. In brief, the evidence indicates that the three Bavarian villages are located in areas where it was not common to breastfeed at all or to breastfeed for only very short durations. In all the areas where the other sample villages are located, breastfeeding appears to have been the general rule, although differences with respect to duration were probably substantial. Fairly prolonged breastfeeding appears to have been most common in Waldeck and East Friesland, while more moderate breastfeeding characterized the areas of Baden and Württemberg where the sample villages are located.

The decline in fertility associated with the demographic transition has been characterized as a shift from a system in which reproduction was largely controlled through social institutions and customs to a system where the private choice of individual couples plays the major role. This shift of control from the societal to the family level involves the emergence and spread of a fundamentally new pattern of reproductive behavior that has been described in the previous chapter as family limitation. In brief, the fertility transition can be viewed as representing the transformation of a population's reproductive pattern from one characterized by natural fertility to one in which family limitation predominates. In this chapter, the nature and extent of marital fertility control in the sample village populations are explored, in the course of which several relevant methodological and substantive issues are addressed.

As described in the previous chapter, at the core of family limitation are attempts to stop childbearing prior to the end of the biologically defined reproductive span. This has implications for both the age at which women terminate childbearing and the age patterns of fertility. The exploration of the shift of natural fertility to family limitation begins by examining indices that utilize these implications for detecting deliberate marital fertility control. In particular, the Coale–Trussell index of fertility control (m) and the age of mothers at last birth are examined in some detail.

In the previous chapter, inter-village differences in mortality were examined and found to be substantial, especially along regional lines. Differences in mortality risks by age, both within the first year of life and within the childhood ages, and shifts in these differences over time were also analyzed. This chapter continues the examination of infant and child mortality by focusing on the differential risks associated with socio-economic status and demographic characteristics. Owing to the limited amount of information contained in the village genealogies, the examination of socio-economic differentials relates only to the occupation and village leadership status of the child's father. However, a far more extensive analysis of demographic differentials is possible because of the considerable information on demographic characteristics that results from the record linkage involved in the family reconstitution process. Analyses of demographic differentials in this chapter focus on infant and child mortality risks associated with the sex of the child, maternal age, birth order, sibship size, and birth interval.

A particularly interesting consideration in the study of infant and child mortality in the past is the possibility that child neglect or abusive child-care practices served as a way of limiting family size prior to the widespread use of contraception or abortion. While infanticide immediately following birth was not unknown in eighteenth- and nineteenth-century Europe, it was probably limited largely to desperate unwed mothers and was rare among married couples.

In Germany, as indeed throughout most of Europe, the vast majority of reproductive behavior during the eighteenth and nineteenth centuries took place within the confines of marital unions as formally defined by the Church and the legal system. The discouragement of extramarital fertility was critical for the western European preindustrial demographic system, in which nuptiality was the principal mechanism through which a precarious balance between population and local resources was achieved. Nevertheless, virtually nowhere was the suppression of births out of wedlock complete, and in some areas and during some periods non-marital fertility made more than a trivial contribution to overall fertility levels. Moreover, while marriage often marked the beginning of a couple's reproductive career, this was by no means always the case. Prenuptial births to couples who subsequently married were not unusual, and even when a couple's first birth was postnuptial, a prenuptial conception often preceded, and in some cases may have precipitated, the marriage itself.

Births born or conceived out of wedlock are of concern not only for their demographic significance but also for what they imply about social life in the past. Their impact on the lives of the parents and children directly involved, as well as on the community at large, is a matter of historical interest. Moreover, they can serve as an imperfect indicator of premarital and extramarital sexual activity.

While the reconstituted family histories contained in the village genealogies are rich in information documenting demographic behavior, only limited information is typically available to indicate the social or economic standing of the couple in the community. Such information is usually limited to the mention of the husband's occupation. For some families even this information is lacking. Occasionally there is mention of the husband's legal status in the community, some honorary position, or position in the village leadership, but such information is far less common than occupational designations. No similar information is provided for the wife. In some cases the occupation, legal status, or honorary position of the husband's and wife's fathers are available, although less frequently than for the husband himself. In this study, only information on the husband's occupation or related designation and, to a much lesser extent, status as a village leader is used to differentiate socio-economic groups within the villages.

The purpose of this appendix is to describe in some detail the construction of the occupational and village leader classification employed. Note should be taken that the occupational classification used in the present study is a revised version of the scheme used in previous analyses of the sample villages and thus the occupational differentials in demographic behavior differ from those published in earlier articles.

Occupational classification scheme

Constructing meaningful occupational groupings for past populations from the limited information derived from parish registers is an extremely difficult task.