Author:

Abstract:

In the recent literature on young age mortality a considerable attention has been devoted to the phenomenon of death clustering: the unequal distribution of (infant) mortality between families. Infant deaths seem to occur in a limited number of families. So far death clustering has been observed many populations, both historical and contemporary. The causes of this death clustering are still very unclear. This is partly due to the fact that there are no good tools available to unravel the characteristics, causes, significance and consequences of mortality clustering. However there is a consensus that this perspective on mortality might yield new and relevant information, not only in describing and explaining the historical mortality transition, but also for our understanding of the fertility transition.This study examines death clustering in the district of Antwerp in the second half of the 19th century. The study has a double set-up. Firstly, attention is paid to the methodological aspects of clustering: the development and application of methods and techniques to measure and visualize clustering. Therefore infant mortality is first approached from the perspective of the loss of children within the family, then from the mortality of the individual child, and finally from the unequal distribution of mortality between populations. For the latter, the Antwerp analyses are compared with data from northern Sweden. This allows to introduce a new dimension adding to the debate on the causes of death clustering, i.e. the effect of infant mortality from previous generation to the survival of children of the next generation.One of the main results is that mortality rates of infants in successive generations are strongly correlated, both along the maternal and the paternal line. Those who lose their infant brothers and sisters, later on have a significantly greater risk to lose their own infants. This correlation is caused by a complex interplay of factors, such as socialization of educational practices and breastfeeding traditions, intergenerational inheritance of social status, regional influences, but also genetic defects. In practice it proves very difficult to detect and separate the combined contributions of these factors.A genealogical history of infant mortality is one of the reasons for the unequal distribution of infant deaths on families: the mortality of infants is clustered in a small number of families, in part because the parents of those families descended from families with high infant mortality.These findings have several implications, both methodological and theoretical. For the empirical reconstruction and the theoretical interpretation of mortality, death clustering, and the decline of infant and child mortality it is statistically necessary and theoretically useful to consider family characteristics. In turn this is crucial for explaining the demographic transition, the transition from high to low fertility and mortality.