Mortality in the past: every second child died

As explained in the following footnote, this data is available from the Human Mortality Database.

Regarding the number of children that people had, the metric that I would ideally need here is the average number of children per woman (or per couple). This is sometimes reported as average family size, but I was not able to find this data. But I could find data on the total marital fertility rate.

The sources for both the number of children and the rate of deaths for the three countries are the following:

Sweden:
A total marital fertility rate of 7.62 children per married woman is reported in Table II (page 40) in M. Anderson (Ed.) (1996) – Population Change in North-Western Europe, 1750–1850. Extramarital children were rare in Sweden at the time. Anderson estimates it at 2%.

The under-15 mortality rate is taken from the Human Mortality Database and corresponds to the average of the annual observations between 1750 to 1780.

Bavaria, Germany:

This data is taken from John Knodel’s research: John Knodel (1970) – ‘Two and a Half Centuries of Demographic History in a Bavarian Village’. Population Studies 24, no. 3 (1 November 1970): 353–76.

According to his study married women had on average 5.6 children and saw on average almost three (2.8) of their children die before they were 15 years old. Knodel also includes data for an earlier period, but cautions against relying on it, writing: “The figure shown for couples married between 1692 and 1749 is undoubtedly spuriously high, resulting from the frequent omission of infant and child deaths from the parish registers during the period.” Knodel suggests that even the data after 1750 (which is shown here) is likely an underestimate of the true mortality.

France:

In France the average married woman had about 8 children in the period 1740 to 1769. This is the total marital fertility rate taken from Table II (page 40) in M. Anderson (Ed.) (1996) – Population Change in North-Western Europe, 1750–1850. 

Youth mortality rates for France are reported in Volk and Atkinson. For the period 1600-1700 the authors report an estimate of 40-50%. They also present an estimate for the period 1816-50 when the mortality rate was 44%.

This research is carried out in the area of Pueblo Viejo, Cahuachi, Estaqueria and Atarco in the Nasca valley. 

The mortality is quoted after Anthony A.Volk Jeremy A. Atkinson (2013) – Infant and child death in the human environment of evolutionary adaptation. In Evolution and Human Behavior. Volume 34, Issue 3, May 2013, Pages 182-192.

The original paper is Drusini, A. G., Carrara, N., Orefici, G., & Bonati, M. R. (2001). Paleodemography of the Nasca valley: Reconstruction of the human ecology of the southern Peruvian coast. Homo, 52, 157–172.

This research is carried out in several sites on the S’Illot des Porros in Mallorca.

The mortality is quoted after Anthony A.Volk Jeremy A. Atkinson (2013) – Infant and child death in the human environment of evolutionary adaptation. In Evolution and Human Behavior. Volume 34, Issue 3, May 2013, Pages 182-192.

The original paper is Alesan, A., A. Malgosa, and C. Simó (1999) – Looking into the Demography of an Iron Age Population in the Western Mediterranean. I. Mortality. In American Journal of Physical Anthropology 110, no. 3 (November 1999): 285–301. The life expectancy at birth was 23 years.

For details see the paper Anthony A.Volk Jeremy A.Atkinson (2013) – Infant and child death in the human environment of evolutionary adaptation. In Evolution and Human Behavior. Volume 34, Issue 3, May 2013, Pages 182-192.

When relying on evidence from modern hunter-gatherers one needs to be cautious of how representative these societies are of those in the past. This is because recent hunter-gatherers might have been in exchange with surrounding societies and “often currently live in marginalized territories”, as the authors say. Both of these could matter for mortality levels, so that the mortality rates are higher or lower than in historical times.

To account for this, Volk and Atkinson have attempted to only include hunter-gatherers that are best representative for the living conditions in the past; they limited their sample “only to those populations that had not been significantly influenced by contact with modern resources that could directly influence mortality rates, such as education, food, medicine, birth control, and/or sanitation.”

All but one of these studied societies in Volk and Atkinson are modern hunter-gatherers. The one study on mortality rates of paleolithic hunter-gatherers finds a higher youth mortality rate: 56% did not survive to puberty.

In the literature on global health and modern health statistics, the most commonly studied age-cutoff is the age of five and the share of children dying before they are five years old is referred to as ‘child mortality’. A low age-cut off makes sense because the mortality in early childhood is typically substantially higher than in late childhood. Nevertheless a focus on the first five years gives only a partial view on the mortality of children. Childhood of course doesn’t end at the age of five, and in this article I’m relying on a more common definition of childhood and am considering all deaths up to the end of puberty or up to the age of 15. Which cutoff is used varies between the different studies on which this account is based on.

A higher cut-off also has the advantage that it makes it possible to connect with historical and archaeological research on the mortality of children. Especially in archeological records it is not possible to determine the precise age at which a child died, but it is possible to differentiate between a child and an adult. 

The researchers Anthony Volk and Jeremy Atkinson, on whose research I am primarily relying here, have brought together most of the historical estimates I am reporting here. Their literature search focused on the  share of children who died before reaching “approximate sexual maturity at age 15”. See: Anthony A.Volk Jeremy A.Atkinson (2013) – Infant and child death in the human environment of evolutionary adaptation. In Evolution and Human Behavior. Volume 34, Issue 3, May 2013, Pages 182-192. https://www.sciencedirect.com/science/article/pii/S1090513812001237#s0015

The mortality up to the end of puberty is less commonly reported in modern health statistics. But it is of course also estimated by health statisticians and at the end of this post you find the estimates from the IGME at: https://childmortality.org

(Note that Volk and Atkinson refer to the mortality up to “approximate sexual maturity” as child mortality, while I am following the established language in global health statistics where child mortality is reserved for mortality up to the age of five.)

The demographers Toshiko Kaneda and Carl Haub estimate the number of humans that were ever born to be about 100 billion. Applying a child mortality rate of 50% this would mean that about 50 billion children died throughout human history. This is of course a rough estimate, but it gives us some idea of just how many children died.

Despite the great pain caused by poor health and early death humanity discovered extremely few medical remedies and practices that were effective over the course of hundreds of generations.

While the rate of discoveries was extremely slow, there were certainly some breakthroughs that were relevant in the context of the time. One such example is that from the bark of the Cinchona tree Quinine was extracted which was effective in treating malaria. For other drugs see Wikipedia’s List of drugs by year of discovery, which includes some other remedies.

We discuss this in more detail here and also in the first footnote of this article. From 10,000 BCE to 1700 the world population grew by only 0.04% annually..

Anthony A.Volk Jeremy A. Atkinson (2013) – Infant and child death in the human environment of evolutionary adaptation. In Evolution and Human Behavior. Volume 34, Issue 3, May 2013, Pages 182-192.

See also M.E. Lewis (2007) – The bioarchaeology of children. Cambridge University Press, NY.

The Indian Knoll site was investigated by Francis Johnston and Charles Snow C.E. Snow. See F.E. Johnston, C.E. Snow (1961) – The reassessment of the age and sex of the Indian Knoll skeletal population: Demographic and methodological aspects. In American Journal of Physical Anthropology, 19, pp. 237-244.

And also Indian Knoll skeletons. The University of Kentucky, Reports in Anthropology, Vol. IV, No. 3, Part 11 University Press of Kentucky, Lexington, KY (1948)

The 28% infant mortality rate is reported in Volk and Atkinson based on Trinkaus (1995).

Erik Trinkaus (1995) – Neanderthal mortality patterns. Journal of Archaeological Science, 22 (1995), pp. 121-142. Online here https://www.sciencedirect.com/science/article/pii/S0305440395801707

Chamberlain (2006) also reports very high mortality rates for subadult Neanderthals (Homo Neanderthalensis). See: Andrew T. Chamberlain (2006) – Demography in Archaeology. Cambridge University Press (Cambridge Manuals in Archaeology).