I can’t recall where I read that 40% number, unfortunately. And boiling such a wide range of social experiences down to a single number is eventually going to get me in a lot of trouble with anthropologists.

Here’s a decent breakdown of the evidence for various state and non-state ancient societies[0]. Obviously there is some significant overlap between state and non-state societies, but in general the violent death rate in ancient societies is really high compared to even the 1800s, an era we consider violent by modern standards.

0: https://ourworldindata.org/ethnographic-and-archaeological-e...