Cold Takes

Unraveling the evidence about violence among very early humans

Holden Karnofsky -

Click lower right to download or find on Apple Podcasts, Spotify, Stitcher, etc.

As I've asked Has Life Gotten Better?, I've run into some intense debates about how violent early humans were.

Trying to unravel and understand the points of disagreement has been confusing, but necessary to get a decent picture of trends in quality of life over very long time periods. The rate of violent deaths is one of the few systematic, meaningful-seeming metrics for assessing quality of life before a couple hundred years ago, and strong claims are made on both sides.

As of now, I believe a different story than either "violent death rates have consistently gone down over time" or "early societies were remarkably peaceful." Here's my current impression:

Hypothesis Advanced by Time period My take
(A) There was a rise in violent death rates sometime around (or before) 10,000 BCE, as people moved from a nomadic lifestyle (moving from place to place) to a sedentary one (staying in one place indefinitely). Some anthropologists, including critics of Better Angels of our Nature Varies by region, but first transition would be sometime around (or perhaps before) 10,000 BCE I'd guess this is true, with low confidence.
(B) There was a fall in violent death rates during the initial transition from non-state to state societies. Better Angels of our Nature Varies by region, but first development of major states was around 5,000 BCE. I am unconvinced / don't find this particularly likely.
(C) Today's society (especially the developed world) has lower violent death rates than prehistoric societies, both nomadic and sedentary. Better Angels of our Nature 10,000 BCE and earlier vs. today I'd guess this is true, with moderate confidence.
(C) There was a fall in violent death rates in Europe starting around 1300, and the US around 1600. Better Angels of our Nature 1300 CE and later Seems true based on homicide data, though large-scale atrocities could complicate the picture (future post).
(D) It's unclear what happened to violent death rates in between the initial transition to sedentary societies and 1300 CE. (Additional observation I'm making) Period between 10,000 BCE and 1300-1600 CE Seems true

Simplified visual timeline version:

Below, I will:

  • Lay out the best data I know of (from Our World in Data) for violent death rates in pre-state societies. I'll briefly discuss the interpretation ("B" above) that early societies were very violent, and became less so as powerful states arose.
  • Examine the main criticism I’ve seen of this data. The criticism is that one needs to distinguish between “sedentary” pre-state societies (which stay indefinitely in one area, and tend to be relatively hierarchical) and “nomadic” societies (which move from place to place, tend to be more egalitarian, and are hypothesized to be more representative of the distant past).
  • Try to pull apart and compare (1) “sedentary” societies; (2) “nomadic” societies; (3) today’s world. I think that "sedentary" societies look more violent than "nomadic" ones, implying that violence rose at some point in the past (lending support to (A) in the chart above), while both look more violent than today's world (lending support to part of (C) in the chart above).
  • Discuss the state of the evidence on what happened to violent death rates in between prehistory and today (which will cover (D) and part of (C) in the chart above).

Data on violent death rates from pre-state societies

Chapter 2 of Better Angels of Our Nature makes comparisons between violent death statistics in non-state societies vs. state societies. Since (it seems) there were no "states" from the beginning of our species until the emergence of the first states around ~5,000 BCE,1 the presumption is that these statistics are a decent way to get at the prevalence of violence in the distant pre-state past, compared to the prevalence of violence in later periods with states.

For a summary, I'll lay out these graphics from Our World in Data, which are essentially the same as those from Better Angels, with some updates and corrections.

One immediate observation: I don't think this data provides much evidence about the argument that the early transitions from non-state to state societies led to a decline in violence ("B" in the chart above).2 This is because:

  • The state societies presented are all from 1600 CE and later, other than "Central Mexico, 1419-1519 CE" (whose violent death rate fits right in with the non-state rates).
  • 1600 CE is a very long time after the appearance of the first states. So this is a comparison between non-state societies and recent or modern state societies - not between non-state societies and the state societies that closely followed them.3

What this data may (and, as I'll argue below, does) show is that today's societies appear to have low violent death rates compared to the distant past.

Nomadic vs. sedentary societies

I've looked for criticisms of the arguments about early violence given in Better Angels. One of the major ones is that the earliest societies weren't just non-state societies - they were specifically nomadic foraging societies, which were more egalitarian than other types of non-state societies. And the high violent death rates above mostly don't come from nomadic foraging societies.

For an example of this criticism, see quote in footnote from Hunter-Gatherer and Human Evolution by Richard B. Lee.4 (For some other criticisms and my responses, see Appendix 1.)

In a sense, this critique isn't directly contradicting Better Angels, which has explicitly chosen to take a "state vs. non-state societies" lens rather than an "earlier humans vs. later humans" lens.5 But if nomadic foragers were less violent than other societies and most early societies were nomadic foragers, this would mean that there was at least one historical case of violent death rates going up (even if they later went back down).

From what I can tell, the distinction between nomadic foraging societies and sedentary societies is widely recognized as real and important, though much about it is disputable and disputed. I have seen a lot of casual references to this distinction; the most thorough discussion I've found is in Chapter 9 of The Lifeways of Hunter-Gatherers (abbreviated for the rest of this piece as "Lifeways"; more here on why I consider this source especially helpful/credible). Here's a quote giving the basic idea:

If I asked the average anthropology student to imagine a group of hunter-gatherers, it is most likely that the Ju/’hoansi would come to mind: small, peaceful, nomadic bands composed of men and women with few possessions and who are equal in wealth, opportunity, and status. Yet ... the average student is also aware of cases that easily overturn that image: large, sedentary, warring, possession-laden Northwest Coast societies, where men boasted of their exploits, status, and power ...

[the latter] are nonegalitarian societies, whose elites possess slaves, fight wars, and overtly seek prestige. Although anthropologists have long considered [these societies] to be exceptions, products of resource-rich environments, archaeologists continue to discover evidence of nonegalitarian foraging societies in many environments ...

It's important to note that when Lifeways uses the term "egalitarian," it's only in a relative sense - for example, it specifically states that even "egalitarian" societies have inegalitarian gender relations.6 "Egalitarian" in this context seems to refer to (a) general norms against explicit hierarchy and boasting of status;7 (b) perhaps a lack of permanent power hierarchy.8

The remainder of that chapter implies that nonegalitarian societies are more likely to be sedentary (occupying a single area indefinitely) as opposed to nomadic (needing to periodically move in order to get to an area with more resources). One rough intuition I've seen is that it's hard to accumulate wealth in a society that periodically relocates entirely (and that lacks currency, banking, etc.) At the same time, this chapter emphasizes that this association (and the idea of "egalitarian" societies) can be oversimplified (quotes in footnote).9

The reason this might matter is the claim that - to quote Lee again - "nomadic foragers ... reflect most closely the characteristics of ancient foragers." The idea being that for most of human history (hundreds of thousands, or even millions of years), most humans were mobile and "egalitarian," and it was relatively recent (perhaps somewhat, but not long, before the Neolithic Revolution of ~10,000 BCE) that sedentary societies started popping up.

This is a view that many seem to hold, including Lee as quoted above, as well as Robert Kelly (the author of Lifeways)10 - though I have also seen it disputed by those who believe that early humans could have been largely nonegalitarian. (Better Angels is one of the sources that argues the latter - quotes in footnote).11

To recap this section:

  • There is a common association between societies' being nomadic and egalitarian, though I haven't seen a systematic examination of the association, and believe that this idea is at least somewhat oversimplified.
  • There is a common view that nomadic, egalitarian societies are most representative of the distant past, though this too is disputed.

With this in mind, I think it makes sense to try to distinguish between nomadic/egalitarian and sedentary/nonegalitarian societies in the data provided above. I'll do that next.

Violent death rates in nomadic societies, sedentary societies, and today's societies

Unfortunately, it is not at all straightforward to determine whether a given society is "nomadic/egalitarian" or "sedentary/inegalitarian."

  • These distinctions are only sometimes flagged in the papers that discuss violence stats.
  • They don't appear to be clear-cut or undisputed.
  • I've often been unable to get confident about what category a given society belongs in via searching, sometimes because societies gradually transition from nomadic to sedentary over time (so the timing of when the violence stats were assessed becomes crucial).
  • Both Better Angels and Our World in Data rely primarily on three "aggregator" sources that pull many estimates of violent death rates from previous papers. But when I examined these, I felt that none gave reliable classifications.
    • Keeley 1996 provides the largest number of estimates. Unfortunately, it says little about its methods for deciding what to include, and doesn't distinguish between nomadic and sedentary societies. Additionally, commentary from a later paper (Bowles 2009, noted below) made me concerned about its reliability.12
    • Bowles 2009 distinguishes explicitly between sedentary and nomadic societies in one of its tables.13 However, this table seems to contain 4 meaningful errors in its 8 rows (see appendix). (BTW, I think this topic is cursed. Lifeways contains a table that should be a collection of homicide rates among foraging societies, but is an erroneous reprint of the previous table; and one of Our World in Data's key tables is also an erroneous reprint of the previous table, which they haven't fixed even though I emailed them about it, though in that case I was able to find the right table by guessing the URL. At least this blog post is perfect.)
    • Gat 2006 mostly doesn't distinguish between sedentary and nomadic.14

To try to get at the distinction, I tried to identify "nomadic forager" societies using Lifeways, Youngberg and Hanson 2010, and the Bowles 2009 classifications, in the set of societies that Our World in Data lists violent death rates for.15

I was only able to identify three:

Society Violent deaths per 100,000 people per year Why I think they count as nomadic foragers
Murngin 330 Bowles 2009 classifies them this way
Tiwi 160 Youngberg and Hanson 2010 names them as one of five societies that best meet the criteria for "societies that most resemble the social environment where most human psychology seems to have evolved: small bands of nomadic foragers." Lifeways also labels the Tiwi "egalitarian" in Table 7-9.
!Kung (aka Ju/'hoansi) 42 Very commonly cited as an example of a nomadic, egalitarian society, e.g. they are the example given when Lifeways explains the distinction

Next, I'll discuss what this tells us about (a) violent death rates in nomadic vs. sedentary societies; (b) violent death rates in early societies vs. today's.

Violent death rates in nomadic vs. sedentary societies

The violent death rates above seem significantly lower than what you'd get if you weren't separating out nomadic foragers.

  • In context of the full Our World in Data set, these societies rank as the 17th, 25th and 28th most violent out of 30.
  • The median violent death rate (per 100k people per year) for that set is 420, compared to 160 for the median of this set.
  • I believe that most of the other, generally more violent societies are probably sedentary, based on their locations16) and a few confirmed cases.
  • A couple other data points show a similar pattern of sedentary populations having notably (but not astronomically - something like 2x) higher violent death rates.17

This implies - if we accept the (non-consensus, but widely held) point that the earliest humans were mostly nomadic - that violent death rates likely rose as people transitioned from nomadic to sedentary societies.

Violent death rates in nomadic societies vs. today's world

It's not totally straightforward to compare the above figures to figures from today's societies, because we need clarity on what exactly "deaths from violence" are - do they include homicides? Battle deaths? Suicides? Poisonings? Deaths from violent accidents and encounters with animals?

  • My current impression is that the violent death figures cited above are inclusive of battle deaths and homicides (which I'd guess can be a fuzzy line for small bands of people),18 while it's ambiguous whether they include other sorts of violent deaths.19
  • I haven't found a ready-made analogous statistic for today's societies (something combining homicides, battle deaths, and potentially more), so I tabulated the figures myself here (based on this Our World in Data source, as well as this population data), creating both a "high" and "low" violent death rate estimate.
    • The "low" estimate is just deaths from terrorism, interpersonal violence, and "conflict and terrorism".
    • The high estimate adds in executions, poisonings, self-harm deaths, and road injuries (a big one).

Here's my summary:

Society Violent deaths per 100,000 people per year
Murngin 330
Tiwi 160
!Kung (aka Ju/'hoansi) 42
World (today) - high estimate 35.4
USA (today) - high estimate 35.2
World (today) - low estimate 7.4
Western Europe (today) - high estimate 6.6
USA (today) - low estimate 6.2
Western Europe (today) - low estimate 0.3

The upshot - today's world looks nowhere near as violent as the few foraging societies we have figures for.

A couple of additional notes on these comparisons in a footnote.20

Note that there are significant caveats to the data on foraging societies, and I wouldn't be shocked if it turned out upon further investigation that - when accounting for large-scale wars and atrocities - today's world looks at least as violent as nomadic foraging societies. (I wouldn't be shocked, but it's not what I currently expect.)

Comparison to societies in between then and now

As discussed above, it looks like the societies most likely to be representative of our most distant past have high violent death rates compared to today. It also looks like violent death rates went up at some point in the distant past (around or prior to 10,000 BCE).21

What about in between ~10,000 BCE and today?

There isn't a huge amount we can say about this period. Interestingly, all of the data we have on violent death rates seems to be either (interpreted as) being about the very distant past (many thousands of years ago) or about the relatively recent past (1300 CE and later).

My rough understanding of why this is:

  • To make guesses about the distant past, we can look at currently-existing societies that seem to have "never modernized" in the sense of adopting agriculture or becoming part of large-scale states.22
  • Starting in 1300 CE, we start to have (in a small number of European countries) systematic written records of things like court cases that make it possible to estimate homicide rates.
  • But for the in-between period, we have neither.23

Still, we can look at the data we do have on 1300 CE and later, and draw some clues from that. My current overall guess:

  • Violent death rates fell at some point in between ~10,000 BCE and the kind of modernization that came with court records and made homicide rate estimates possible (around 1300 CE in Europe, later elsewhere).
  • After that sort of modernization, they fell fairly steadily.

I'll explain these points next.

The earliest systematic data I've seen from actual past homicide reports (as opposed to inference from ethnography or archaeology) is from this chart (there is a similar one in Better Angels):

The starting homicide rate is 25-55 per 100k people per year, which looks low compared to the figures above (especially the figures that don't distinguish between nomadic and sedentary societies!) This implies that there was some decline in violence even before 1300 (in addition to the big decline afterward, shown).

However:

The chart is only of homicides, and other sources of violent deaths could complicate the picture. For example, eyeballing this chart suggests that we might want to add something like 50 to the figure for the UK just to account for battle deaths; and we don't know how much more we'd have to add if we wanted to include suicides, executions and violent accidents.

Western Europe may not be representative. Some of the only other homicide data I've seen from around this period is in the following charts from Better Angels:

The starting rates here don't look clearly lower than the figures for nomadic forager societies.24

My current overall guess: after violent death rates ~doubled with the transition from nomadic to sedentary societies (discussed above), they then fell back to the original rates (but not necessarily much below) at some point in between then and the kind of modernization that came with court records and made homicide rate estimates possible. After that sort of modernization, they fell fairly steadily, as shown in the above charts.

I've now covered each of (A)-(D) discussed in the intro. Here's the graphical summary again:

Whew! I hope you found that at least somewhat less confusing to read than I did to write. Having gotten something resembling clarity on violent death rates over the long run, I can use this for the broader question of whether life has gotten better.

The story for "overall quality of life" looks broadly similar to the story for violent death rates: things look like they got worse around 10,000 BCE (when agriculture was developed), then we have a big mystery between then and the early 2nd millennium, and since then there was some improvement (which accelerated in the Industrial Revolution a couple hundred years ago). This is a different picture from either "Life has gotten steadily better throughout our history" or "Life was best in the state of nature," both of which I think are more common memes.

 

Footnotes

  1. I haven't tracked down a satisfying source for this, though I also haven't seen much controversy over it (not that that gives too much comfort). The Wikipedia page on states makes this claim, as does Better Angels ("It took around five thousand years after the origin of agriculture for true states to appear on the scene," from chapter 2, citing four sources). ↩

  2. Better Angels also makes an additional argument that the transition to state societies brought a reduction in violence: ""The major cleft in the graph, then, separates the anarchical bands and tribes from the governed states. But we have been comparing a motley collection of archaeological digs, ethnographic tallies, and modern estimates, some of them calculated on the proverbial back of an envelope. Is there some way to juxtapose two datasets directly, one from hunter-gatherers, the other from settled civilizations, matching the people, era, and methods as closely as possible? The economists Richard Steckel and John Wallis recently looked at data on nine hundred skeletons of Native Americans, distributed from southern Canada to South America, all of whom died before the arrival of Columbus.59 They divided the skeletons into hunter-gatherers and city dwellers, the latter from the civilizations in the Andes and Mesoamerica such as the Incas, Aztecs, and Mayans. The proportion of hunter-gatherers that showed signs of violent trauma was 13.4 percent, which is close to the average for the hunter-gatherers in figure 2–2. The proportion of city dwellers that showed signs of violent trauma was 2.7 percent, which is close to the figures for state societies before the present century. So holding many factors constant, we find that living in a civilization reduces one’s chances of being a victim of violence fivefold." I haven't dug into this study but I have a number of reservations, including generic skepticism of any one study, and the observation that <10% 1800 of people worldwide lived in cities at ~every point before (this is based on a calculation i did off HYDE data). ↩

  3. While some of the non-state data is also recent, I believe it is intended as evidence about the very distant past, when the vast majority of people lived in non-state societies (as opposed to today, when almost none do). ↩

  4. "Historically nomadic foragers (HNFs), small in scale, mobile, and egalitarian, reflect most closely the characteristics of ancient foragers, a point emphasized by Fry (2006, 2013). But the bellicose school loads their sampling procedures with groups that depart sharply from this pattern.

    "Mounted foragers of the American Great Plains (De Maillie 2000) and sedentary nonegalitarian foragers of California (Heizer 1978) and the north west coast of North America (Suttles 1990; Flannery & Marcus 2012, pp. 66-87; Daly 2014) all demonstrated significant levels of war-like behaviors. Yet, horse transport on the plains and stockaded settled villages on the west coast are completely absent from the archaeological record of pre-Neolithic foragers. But at least these are examples of hunter-gatherers ...

    "In his 2011 book, Pinker does address the differences between foragers and farmers, but he still loads his sample with cases that are not representative of HNFs [Historically Nomadic Foragers]. For example, in his table 'Rate of Death in Warfare in Nonstate and State Societies' (Pinker 2011, figures 2– 3, p. 53), the 27 nonstate cases are heavily loaded with New Guinean and nearby farming societies (12 of 27) and Californian and Plains Indians (5 of 27); only 5 of 27 of the cases remotely qualify as HNFs.

    (Lee doesn't say which 5.) ↩

  5. "For all these reasons, it makes no sense to test for historical changes in violence by plotting deaths against a time line from the calendar. If we discover that violence has declined in a given people, it is because their mode of social organization has changed, not because the historical clock has struck a certain hour, and that change can happen at different times, if it happens at all. Nor should we expect a smooth reduction in violence along the continuum from simple, nomadic hunter-gatherers to complex, sedentary hunter-gatherers to farming tribes and chiefdoms to petty states to large states. The major transition we should expect is at the appearance of the first form of social organization that shows signs of design for reducing violence within its borders. That would be the centralized state, the Leviathan ... What happens, then, when we use the emergence of states as the dividing line and put hunter-gatherers, hunter-horticulturalists, and other tribal peoples (from any era) on one side, and settled states (also from any era) on the other?" From Better Angels of our Nature, Chapter 2.  ↩

  6. "Even the most egalitarian of foraging societies are not truly egalitarian because men, without the need to bear and breastfeed children, are in a better position than women to give away highly desired food and hence acquire prestige. The potential for status inequalities between men and women in foraging societies (see Chapter 9) is rooted in the division of labor (vCollier and Rosaldo 1981: 282; Bliege Bird, Codding, and Bird 2009)." ↩

  7. Lifeways: "There are people in every society who will try to lord it over others, but egalitarian cultures contain ways to level individuals, to 'cool their hearts' as the Ju/’hoansi say. Humor is used to belittle the successful but boastful Ju/’hoan hunter; if that fails, he will be shamed with the label !xka ≠xan, 'far-hearted,' meaning mean or stingy (Lee 1988: 267). The Martu berate such people with warnings that they are 'like rocks,' with no compassion (Bird and Bliege Bird 2009: 44). Wives use sexual humor to keep a husband in line; and gambling, accusations of stinginess, or demand-sharing maintain a constant circulation of goods and prevent hoarding." ↩

  8. "Group leaders are common in egalitarian societies – Shoshone 'rabbit bosses' for example – but they are temporary and have their position only because they have demonstrated skill at a particular task. Their leadership does not carry over into other realms of life, nor is it permanent." ↩

  9. "Rather than attributing nonegalitarian foragers simply to 'resource abundance,' as many have done in the past, we will see that sedentism, the resource base, geographic circumscription, storage, population pressure, group formation, and enculturative processes all play a role ... The term 'egalitarian' does not mean that all members have the same of everything – goods, food, prestige, or authority. Not everyone is equal in egalitarian societies, but everyone has (or is alleged to have) equal access to food, to the technology needed to acquire resources, and to the paths leading to status and prestige ... Even in this regard, the inheritance of material wealth (especially productive land) and relational wealth (political connections) give some individuals a head start in life ... Egalitarianism can mask hierarchy. Australian Aboriginal men acquire authority and power in religious affairs by disengaging from property, by giving away meat, for example. But one can only disengage from property if, at some level, one claims a right to it (see Bird and Bliege Bird 2009). Appeals to autonomy and equality by informants in egalitarian societies often contradict an ethnographic reality in which some members have higher status and greater access to resources than others. We have already seen that people are well aware of, give greater prestige to, and may lose some of their autonomy to men who are good hunters. Differences in autonomy are perhaps especially pronounced between men and women." ↩

  10. "On the strength of archaeological data, it is reasonable to assume that nonegalitarian society developmentally followed egalitarian society. On the northwest coast, for example, slavery appears about 1500 BC, warfare by AD 1000, and nonegalitarian societies by at least AD 200 (Donald 1997; Ames and Maschner 1999; Ames, 2001, 2008; Grier 2006; see also reviews of global prehistory by R. C. Kelly [2000] and Keeley [1996], and Kennett [2005] on California’s Channel Islands). Egalitarian behaviors and an egalitarian ethos were adaptive for quite a long time in human history before the selective balance tipped in favor of nonegalitarian behaviors and a nonegalitarian ethos (Cohen 1985)." ↩

  11. From Better Angels of our Nature, chapter 2: "The nonstate peoples we are most familiar with are the hunters and gatherers living in small bands ... But these people have survived as hunter-gatherers only because they inhabit remote parts of the globe that no one else wants. As such they are not a representative sample of our anarchic ancestors, who may have enjoyed flusher environments. Until recently other foragers parked themselves in valleys and rivers that were teeming with fish and game and that supported a more affluent, complex, and sedentary lifestyle. The Indians of the Pacific Northwest, known for their totem poles and potlatches, are a familiar example."

    Bowles 2009, one of the main sources of violence data, seems to take a similar view: "Because hunter-gatherer populations occupying resource-rich areas in the Late Pleistocene [starting over 100,000 years ago] and early Holocene were probably sedentary (at least seasonally), I have included wars involving settled as well as purely mobile populations." ↩

  12. "I studied all available archaeological and ethnographic sources that present (or are cited as presenting) relevant data. Of these 34 sources, 14 were found to present data that were unrepresentative (for example, when warfare was primarily with modern agricultural populations), unreliable, or inadequate ... prehistoric warfare was frequent and lethal, but somewhat less so than estimates based on data in the standard source for these estimates (6)." (6) is a reference to Keeley 1996. I also note that Bowles 2009 is later, examined Keeley 1996, and contains far fewer data points than Keeley 1996, implying that it chose to discard a lot of what Keeley 1996 lists. ↩

  13. The one using ethnographic data, i.e., studies of today's foraging societies; it does not do this in the table using archaeological data ↩

  14. It does specify that some of its figures come from agricultural societies, which implies that they are sedentary or at least have the ability to be sedentary (see previous characterization of the significance of agriculture). ↩

  15. I skipped two other tables that list violent death shares instead of rates, because shares are much harder to use for the comparisons I want to make later in this post. The "violent death share" is the number of deaths from violence divided by the total number of deaths; the "violent death rate" is the number of deaths from violence divided by the population. ↩

  16. I've generally seen New Guinea and the American Northwest identified with sedentary populations. From Lifeways: "We have less ethnographic information on nonegalitarian than on egalitarian foraging societies; these include Florida’s Calusa (Widmer 1988; Marquardt 2001; although the Calusa probably cultivated some plants such as gourds and chili peppers); various California foragers, with the Chumash of southern California the most heavily studied (e.g., Bean 1978; Arnold 2001a,b; 2004; Kennett 2005; Gamble 2008); the Northwest Coast (Ames 1995; Ames and Maschner 1999; Grier 2006); the Plateau region of the northwestern United States (Hayden 1992; Prentiss and Kuijt 2004); some New Guinean peoples (Roscoe 2006); and Japan’s Ainu (Watanabe 1968, 1972a,b)."  ↩

  18. The Tiwi figure cited above comes from the following quote in this book: "In one decade (1893-1903), at least sixteen males in the 25-10-45 age group were killed in feuding; either during sneak attacks or in arranged pitch battles. Those killed represented over 10 per cent of all males in that age category, which was the age group of the young fathers." No indication is given of how this was converted to a "violent deaths per 100k people per year" rate, and I'm also concerned that this might be someone noting a particularly violent (rather than representative) decade (although it might also understate violence if it is focused specifically on feuding). This has increased my interest in a more thorough examination of where all these figures are coming from. I haven't yet been able to track down a primary source for either of the two other figures cited above, as both come from out-of-print books. ↩

  19. For the archaeological studies specifically - which I am not focusing on for reasons noted above - it looks like they are generally looking for the sorts of injuries that would've been inflicted by other humans, such as parry fractures. ↩

    • There are of course some countries today with higher violence rates than are in the table above. But only one (Syria) has a rate higher than the Tiwi, and only about 20% of the listed countries have rates higher than the Ju/'hoansi - and this is using the more inclusive violence definitions that include e.g. road injuries.
    • The 20th century had much bigger, bloodier wars than are going on today, and those sorts of events can significantly increase total deaths from violence: I estimated that 20th-century atrocities (including e.g. estimated deaths from human-caused famines), spread out over the whole 20th century, accounted for about 81 violent deaths per 100k people per year. (The 20th century was unusually bad in this regard.) Even adding that figure in, though, would leave the modern world looking considerably less violent than 2 of the 3 nomadic foraging societies. ↩

  21. This is when the Neolithic Revolution caused a transition to agricultural - and therefore sedentary - societies. Though it's possible that there was a transition to sedentary societies before the transition to agricultural societies.

    I assume that most people were living in agricultural societies not too long after ~10,000 BCE, because of my impression (which I don't have a clean citation for) that the agricultural societies experienced much faster population growth than other societies.  ↩

  22. And as I stated above, I don't think we actually have a clean comparison between early non-state societies and early state societies - we only have early non-state societies vs. state societies after 1600 or so. ↩

  23. In theory, we can look at archaeological remains as well, though I am generally skeptical of this sort of evidence (more in appendix), and I haven't been able to find systematic examinations of this data for the "in-between period." ↩

  24. Though with the exception of Maryland and Virginia before 1650, they do look lower than the figures that include both nomadic and sedentary societies, and probably would be even if we added deaths from battles, accidents and executions. ↩