I was reading this article, and it's honestly better cited than a lot of things I read, but one of the claims it includes contradicts my expectations. Specifically, this claim:

In the best-selling book, Sapiens: A Brief History of Humankind (audiobook), world renown historian, Yuval Noah Harari, describes the early periods of the agricultural revolution:

“With the move to permanent villages and the increase in food supply, the population began to grow. Giving up the nomadic lifestyle enabled women to have a child every year. Babies were weaned at an earlier age – they could be fed on porridge and gruel.” [1]

But, Noah Harari notes that this pursuit of an easier life led to unintended consequences:

“The extra hands were sorely needed in the fields. But the extra mouths quickly wiped out the food surpluses, so even more fields had to be planted. As people began living in disease-ridden settlements, as children fed more on cereals and less on mother’s milk, and as each child competed for his or her porridge with more and more siblings, child mortality soared.”

Based on the context in the article, the agricultural revolution being described is the Neolithic Revolution around 10,000 BC. The source of the claim is listed as the book "Sapiens: A Brief History of Humankind" by Yuval Noah Harari, but I don't have access to that book in order to check that.

The specific thing I'm questioning is that during the agricultural revolution "child mortality soared". That seems extremely counter-intuitive to me, since I would expect increased consistency and availability of food and shelter to have a positive effect on mortality rates. Additionally part of the reasoning listed seems to be a cycle of "need more food -> need more people to farm the food -> need more food for those people" but I had thought the entire premise of the Neolithic Revolution is explicitly that people would be able to farm more food than they could consume, meaning that the "more people" step would increase food production more than enough to cover those people rather than ending up back at needing more food.

However, all of that is just "my logic" and vague memories from history classes I didn't really pay close attention to. I can't be confident with just that, so did infant mortality rates increase substantially during the Neolithic Revolution?

  • So I'm assuming they're probably talking about the agricultural revolution that took place in pre-history (circa 10,000 BC), as opposed to any of the others referred to at en.wikipedia.org/wiki/Agricultural_revolution?
    – DenisS
    Commented Nov 7, 2018 at 15:21
  • @DenisS Yes, that seems to be the one; the preceding section indicates that it is discussing a period around 9500-8500 BC known as the Agricultural Revolution, and there isn't any indication that it would be talking about different Agricultural Revolutions in each section. I'll add that to the question. Commented Nov 7, 2018 at 15:26
  • 1
    I'd also refer to it as the "Neolithic Revolution" because that seems to be another common name that doesn't get it confused with more modern "Agricultural" Revolutions.
    – DenisS
    Commented Nov 7, 2018 at 15:29
  • 3
    By "rates" we're talking as a percentage of the total infant population, and not just raw numbers of deaths, right? Commented Nov 7, 2018 at 15:43
  • @PoloHoleSet The phrase used in the claim is "child mortality", and definitions of that phrase I've seen typically are proportional not absolute, for example from Wikipedia: "Child mortality refers to number of child deaths under the age of 5 per 1000 live births". If the claim intended it as "absolute number of child deaths" then I think evidence of that would be most of the way to a good answer. Commented Nov 7, 2018 at 16:05

1 Answer 1


Probably someone can give a more detailed answer, but skeletal data seems to support this increase; form Bocquet-Appel, Science, 2011:

During the economic transition from foraging to farming, the signal of a major demographic shift can be observed in cemetery data of world archaeological sequences. This signal is characterized by an abrupt increase in the proportion of juvenile skeletons and is interpreted as the signature of a major demographic shift in human history, known as the Neolithic Demographic Transition (NDT).

There's also a graph in the paper on this:

enter image description here

[The photo mentioned in the caption is just a bunch of bones, so I've omitted it here.]

Keep in mind that it went hand in hand with an increase in fertility though. Actually the Bocquet-Appel paper discusses this:

When compared with the Contemporary Demographic Transition (CDT) as described for Western industrialized societies, the NDT was its mirror image. In the CDT, the decline in mortality was followed by a decline in fertility, but in the NDT, increased fertility was followed by increased mortality. The CDT is slowing the growth of the world population, but the NDT was its springboard. In both cases, however, the time lag between the two stages produced an interval in which fertility exceeded mortality and resulted in a rapid increase in the population.

A more reserved paper by Gage & DeWitte (2009) notes that

About the only issue that is reasonably certain is that fertility exceeded mortality for some period of time during the transition so that populations grew. The idea that mortality and fertility increased (the consensus view) is based on indirect evidence and assumption.

I'll also note the title of your question (incorrectly) asks about infant mortality, but the quote you gave (and seemingly still doubt) is about child mortality, which is the topic of the Bocquet-Appel (and other) papers on the NDT. Infant mortality is defined as the death of children under the age of one year; child mortality typically refers to deaths under 14 years; in the Bocquet-Appel paper this is extended to 19 years.

  • I wasn't aware of the difference between "child mortality" and "infant mortality", though that wouldn't have changed my skepticism. Based on the unusual choice of "up to 19" in this paper and references to "under 20" in the article I read this seems likely to be at least part of the data used for the claim in the first place. Commented Nov 10, 2018 at 16:37
  • 1
    I can not read the original Bocquet-Appel paper. As it is the major source of your answer, I like to add that the graph is not too convincing as scientific evidence. It is quite variable and there are more data points after the "Neolithic Revolution" too. Has there been any hypothesis testing done in the paper in addition to just regression? (and was it controlled for unequal variance?) Commented Nov 13, 2018 at 22:10
  • @TimMMarsouin: no there isn't any formal hypothesis testing in that paper. Commented Nov 13, 2018 at 22:23
  • Just presenting this graph without any measurement of variance like confidence intervals makes it not really valuable and meaningful. Also the regression model seems overly complicated. We should be sceptical. Commented Nov 15, 2018 at 5:53

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .