Steven Pinker on Deaths by Violence in Pre-state Societies

Steven Pinker seems to have come in for a lot of criticism over his view that in terms of per capita death rates by violence, stateless societies were (or are) actually much more violent than modern states, even in the 20th century.

R. Brian Ferguson has a scholarly refutation of Pinker’s views on prehistoric war in Ferguson (2013a) and (2013b), but these are in a book yet to be published, so who knows if his criticisms have force.

Other critical reaction to Pinker’s data can be found here:

Christopher Ryan, “Steven Pinker’s Stinker on the Origins of War,” March 29, 2011.

Christopher Ryan, “Pinker’s Dirty War on Prehistoric Peace,” Huffington Post, January 9, 2012.

A damaging criticism in these articles is that Pinker’s examples of hunter gatherer per capita death rates (from Keeley 2001) are mostly from societies that are not hunter gatherers: most are partly agriculturalists and not the strict “nomadic (immediate-return) hunter-gatherers” who were “most representative of human prehistory.”

But, if we actually turn to Pinker’s book (Pinker 2011: 49, Figure 2-2; and 53, Figure 2-3), we find that the data compiled there is different from the table Ryan cites, and Pinker does not seem to have made the errors that Ryan points out.

I wish to review what Pinker says in his book about violence in stateless and state societies. The purpose of this post is summarise what Pinker actually argues in his book (not in the popular summaries of it).

First, some background. Late Pleistocene humans were stateless hunter gatherers. Later these hunter gatherers started to domesticate plants and animals, and pastoral and horticultural/agricultural populations arose. Obviously, agriculturalists tend to the sedentary, but hunter gatherers could be nomadic or sedentary (or seasonally one and then the other).

So the sequence of human societies (in a long-term historical sense) runs as follows:

(1) stateless hunter gatherers (nomadic and/or sedentary);

(2) stateless agricultural/pastoral populations (the first generally sedentary), eventually leading to rural populations; at the same time hunter-horticulturalist societies also developed and persisted in some regions of the world;

(3) cities;

(4) state societies;

(5) more and more complex state societies (incorporating any or all of the above types of societies).

What is being claimed by Pinker is that as populations moved into (4) and (5) above, the per capita level of death rates by violence fell as a long-term trend.

Notice how (according to Pinker’s theory) it is not just hunter gatherers who supposedly had a high per capita level of death by violence, but also even stateless agricultural/pastoral populations (a point we should remember).

I list the important details of Pinker’s argument below:

(1) Pinker (2011: 4) notes that “total war” (meaning war where civilians are often attacked and killed in great numbers) is simply not an invention of the 20th century. Many wars in the past have involved massacres of the male population of a community and the enslavement of the women and children. We need look no further than Homer’s Iliad and Odyssey and the Old Testament – some of the foundational cultural and religious books of Western civilisation – to see this: both texts know “total war,” and speak of it as a normal activity (Pinker 2011: 4–12). Indeed god even orders it more than once in the Old Testament (Pinker 2011: 7–8). I might add that even though the events the Bible imagines here – the conquest of Canaan and many other things down to the time of Solomon – are simply fiction (as noted by Pinker 2011: 10), it is the cultural significance of the passages that is important: they indicate that what we would call a type of “total war” was quite acceptable.

(2) Pinker (2011: 31–36) briefly sketches two historically important theories of primitive man: that of Thomas Hobbes (the “Leviathan state” theory) and of Jean-Jacques Rousseau (“noble savage” theory). Pinker comes down on the side of Hobbes, and notes how some modern anthropologists hold a view related to that of Rousseau (Pinker dubs them the “anthropologists of peace”): they think that

(1) prehistoric humans were not generally inclined to kill fellow humans;
(2) that war is a “recent invention”, and
(3) that fighting in indigenous pre-modern societies was largely ritualistic and often harmless (Pinker 2011: 31–36).

Pinker rejects this, and I agree. Pinker also draws attention to an unhealthy political correctness amongst the so-called “anthropologists of peace” who have attempted to unfairly oppose research that disputes their own ideas on the origin of war and violence (Pinker 2011: 43).

(3) First, human evolution and the evidence from primates does not support the ideas of the “anthropologists of peace.” Some have argued that the bonobo (Pan paniscus) was possibly like the common ancestor and that humans developed aggression and war-like tendencies subsequently. But Pinker notes the problems with this. First, bonobo non-aggression may be exaggerated, as violence has been observed in wild bonobos (Pinker 2011: 39). Secondly, even if we assume that bonobo aggression is minimal, bonobos are so anatomically different from normal chimps and other apes (through the evolutionary process of “neoteny”) that it is likely that they diverged from an animal more like the common chimps rather than the common ancestor (Pinker 2011: 39).

Now Chimpanzees are violent, and fight inter-communal wars (Pinker 2011: 37). In some chimp communities, it has been observed that more than 33% of males die by violence (Pinker 2011: 38). Violence is also observed in other primates. It is not unlikely that it was a trait in the common ancestor and was passed on to Australopithecines and early Homo species.

Prehistoric humans certainly had real motives for war: (1) struggle over and appropriation of resources or even theft; (2) what we would now call “credibility” or pre-emption: the need to demonstrate to potential enemies that there will be a price for any violence; and (3) revenge or honour (the use of violence to extract revenge for crimes or insults, real or perceived) (Pinker 2011: 46–47; 56).

(4) While Homo sapiens is anatomically 200,000 years old, behaviourally modern humans (with art, clothing, tools, and rituals) are probably about 75,000 years old.

Agriculture developed about 10,000 years ago independently in China, India, the Fertile Crescent, West Africa, Mesoamerica, and the Andes (Pinker 2011: 40).

Before this humans were all hunter gatherers, but of course the hunter gatherer lifestyle persisted for many thousands of years in some regions, and indeed up until the 20th century in some areas.

A crucial point that Pinker makes is that human societies were still stateless after the invention of agriculture, and it was some 5,000 years after the appearance of agriculture that effective states developed (Pinker 2011: 41). Therefore we should not be remotely surprised that stateless or loosely organised agricultural societies should also have high levels of violence. Even the violence of tribal sedentary societies with petty chiefdoms does not really refute Pinker’s thesis. For, quite simply, these societies are not subject to a higher “Leviathan state” that enforces laws and policies behaviour.

(5) Pinker draws on a body of work against the naïve modern anthropological successor of Rousseau’s ideas: LeBlanc (2003); van der Dennen (1996); Keeley (1996); Thayer (2004); and Wrangham and Peterson (1997).

The “anthropologists of peace” ignore the cumulative causalities of battles (not insignificant), the murderous raiding that some stateless human societies engage in, killing of non-combatants, and the other violent activities (Pinker 2011: 44).

(6) Pinker admits that the 20th century in terms of absolute numbers of violent deaths is the most bloody of any century (Pinker 2011: 47). But the linchpin of Pinker’s analysis is his contention that really it is per capita death rates by violence for a society that demonstrate how violent that society is. Certainly this is how modern social science decides such a question: we use per capita homicide rates (per 100,000) to compare one country with another.

The crucial question is: if I were a member of such-and-such a society, what would my chances be of dying a violent death in this society? For that, we need per capita death rates by violence (both war and murder).

(7) Pinker admits that war was not universal amongst foraging hunter gatherer groups, but nevertheless frequent for many, as data suggest that 65 to 70% of hunter gatherers are at war at least every 2 years (Pinker 2011: 52). Above all, we should look at the per capita death rates for hunter gatherers, hunter-horticulturalists, and other essentially stateless tribal groups, whether agricultural or pastoral populations.

Here is Pinker’s data for the proportion of deaths of the population in a number of sample societies:

Percentage of Deaths in Warfare

Prehistoric Archaeological sites
Crow Creek, South Dakota, 1325 CE: 60%
Nubia, site 117, 12,000–10,000 BCE: c. 46%
Sarai Nahar Rai, India, 2140–850 BCE: c.29%
Br-Columbia, 30 sites, 3500 BCE–1674 CE: c.23%
Volos’ke, Ukraine, –7500 BCE: c. 21%
Vasiliv’ka III, Ukraine, 9000 BCE: c. 20%
Illinois, 1300 CE: 16%
Northeast Plains, 1485 CE: c. 15%
Vedbaek, Dnk., 4100 BCE: c. 13%
Bogebakken, Dnk., 4300–3800 BCE: c. 12%
Ile Teviec, France, 4600 BCE: c. 11%
Brittany, 6000 BCE: c. 8%
Ctl. California, 1400 BCE–235 CE: c. 8%
Skateholm I, Sweden, 4100 BCE: c. 7%
S. California, 28 sites, 3500 BCE–1380 CE: c. 5%
Kentucky, 2750 BCE: c. 4%
Ctl. California, 1500 BCE–1500 CE: c. 3%
Calumnata, Algeria, 6300–5300 BCE” c. 3%
Ctl. California, 2 sites, 240–1770 CE: c. 3%
Nubia, nr. site 117, 12,000–10,000 BCE: c. 3%
Gobero, Niger, 14,000–6200 BCE: 0%
Avg. 21 prehistoric archaeological sites: 15%

Hunter-gatherers
Ache, Paraguay: 30%
Murngin, Australia: c. 28%
Hiwi, Venezuela-Colombia: c. 15%
Ayoreo, Bolivia-Paraguay: c. 13%
Modoc, N. California: c. 11%
Tiwi, Australia: c. 8%
Casiguran Agta, Philippines: c. 55
Anbara, Australia: c. 4%
Avg. 8 hunter-gatherer societies: 14%

Hunter-horticulturalists & other tribal groups
Waorani, Amazon: c. 58%
Jivaro, Amazon: c. 30%
Gebusi, New Guinea: c. 29%
Montenegro, Europe: c. 23%
Yanomamo-Shamatari, Amazon: c. 18%
Mae Enga, New Guinea: c. 15%
Dugum Dani, New Guinea: c. 12%
Yanomamo-Namowei, Amazon: c. 12%
Huli, New Guinea: c. 11%
Anggor, New Guinea: c. 10%
Avg. 10 hunter-hort. & tribal groups: 24.5%

States
Ancient Mexico, before 1500 CE: 5%
World, 20th C (wars & genocides): 3%
Europe, 1900–1960: 3%
Europe 17th century: 2%
Europe & U.S., 20th century: 1%
World, 20th century (battle deaths): 0.7%
World, 2005 (battle deaths): 0.0004% (Pinker 2011: 49, Figure 2–2).

Some data on Pinker’s sources: on prehistoric violence, Pinker cites Bowles (2009); Keeley (1996); McCall and Shields 2008; Walker 2001; and Thorpe 2003.

On hunter-gatherers, see Bowles (2009). On hunter-horticulturalists and other tribal groups, see Gat (2006); and Keeley (1996).

From Pinker’s data, we can see that death rates were quite high in hunter-gatherer societies, and were even higher amongst stateless hunter-horticulturalists and other tribal groups.

That might suggest that violence spiked once agriculture was invented, perhaps because sedentary agriculturalists have more to fight over.

But the clear trend is apparent once the modern state appeared: a downward trend.

(8) The 20th century saw about 40 million battle deaths (soldiers and civilians directly killed in combat). Pinker estimates that 6 billion people died in the 20th century (from all causes natural or otherwise), so that 0.7% died in battle deaths (Pinker 2011: 50).

What is more striking is that, if we include all state-based violence, such as genocides, purges, man-made famines, and other human-caused disasters, the percentage rises to 3% (Pinker 2011: 50). That is still lower than sample death rates in stateless tribal societies, whether hunter-gatherers or horticulturalists (or mixtures of the two).

(9) So what was the watershed? Pinker argues that it was the “Leviathan state.” The rise of states that incorporated many communities and that imposed law and order and peace upon the populations. But Pinker does not deny that many of these early states were socially and economically stratified, had authoritarian rulers and peace was kept with some very brutal law codes and punishments (Pinker 2011: 57). Pinker is not giving a moral endorsement to early states or all states, but is merely identifying a trend and process that they caused: reduction in violence per capita.

(10) Shifting to a calculation of deaths from war per 100,000 in 27 non-state societies, Pinker comes up with an average of 524 per 100,000 (Pinker 2011: 52).

But then we only need look at deaths from the worst and most violent states to see the difference:

Average for 27 non-state societies: 524 per 100,000
Germany (20th century): 144 per 100,000
Soviet Union/Russia (20th century): 135 per 100,000
Japan (20th century): 27 per 100,000
United States (20th century): 3.7 per 100,000
Entire world (20th century; state-based violence)*: 60 per 100,000

* that is, war, genocide, purges, man-made famines (Pinker 2011: 53, Figure 2-3).

A good scan of the whole table can be found over at Gene Callahan’s blog.

Finally, we can compare some data below on the worst regimes of the 20th century:

Average for 27 non-state societies: 524 per 100,000
Khmer Rouge regime (1975-1979): 8,160 per 100,000
Nazi Germany (1939-1945, European victims): 1,008 per 100,000
Soviet Union (1917-1987): 450 per 100,000
Communist China (1949-1987): 120 per 100,000 (Cooney 1998: 58).

There is no doubt that the Khmer Rouge and Nazi Germany stand out as outliers, the most terrible regimes of the 20th century. But what is astonishing is that even murderous tyrannies like the former Soviet Union and Communist China (in its worst days) were less violent than some non-state societies.

(11) Pinker also notes that, even if we are to assume that prehistoric hunter gatherers really had low deaths per capita from violence, that does not contradict his thesis.

For Pinker is comparing people in a state of anarchy with state-based societies (where the government has a monopoly on force or at least a significant role in keeping the peace). Those groups living in stateless societies include hunter-gatherers, hunter-horticulturalists, pastoralists, and some agricultural communities.

Pinker is not simply comparing prehistoric hunter gatherers with every other society.

(12) One important outlier in the 20th century was Cambodia under the Khmer Rouge, possibly the most murderous regime of the 20th century. It is unclear whether this is included in Pinker’s data for 20th century wars and genocides (though I would assume so). Estimates of the total number of deaths range from 1.4 to 2.2 million in a population of around 7 million, and a death toll of perhaps 20-31.42%. But let us assume an even worse upper estimate of 33%. But even that shocking death toll was exceeded by some stateless societies: e.g., the Amazon Waorani (c. 58%). Some stateless societies came close: the Amazon Jivaro (at c. 30%) and New Guinea Gebusi (c. 29%).

As a final point and matter of interest, the Amazon Waorani, a semi-nomadic group, seem to have the highest homicide rate ever seen in any modern society (Beckerman et al. 2009): some sources report that 60% of deaths were due to violence over the past century. That is pretty shocking.

Further Links
I have realised that there is a lot of good discussion of this topic at Gene Callahan’s blog, where you can get a feel for the type of (mostly silly) libertarian rebuttals of Pinker:

Gene Callahan, “It Ain’t the State,” La Bocca della Verità, November 20, 2011.

Gene Callahan, “The Violence of Pre-State Warfare,” La Bocca della Verità, January 4, 2013.

Gene Callahan, “State Policies Never Work?!,” La Bocca della Verità, January 19, 2013.

Other Criticisms of Pinker:

Douglas P. Fry, “Peace in Our Time,” Bookforum.com, December/January 2012.
I have, incidentally, no idea how the author of this review knows that prehistroic humans did not suffer from “chronic warfare, torture, slavery, and exploitation.”

BIBLIOGRAPHY

Beckerman, S., Erickson, P. I., Yost, J., Regalado, J., Jaramillo, L., Sparks, C., Iromenga, M. and K. Long. 2009. “Life Histories, Blood Revenge, and Reproductive Success Among the Waorani of Ecuador,” Proceedings of the National Academy of Sciences of the United States of America 106.20: 8134–8139.

Bowles, Samuel. 2009. “Did Warfare among Ancestral Hunter-Gatherers affect the Evolution of Human Social Behavior?,” Science 324: 1293–1298.

Cooney, Mark. 1998. Warriors and Peacemakers: How Third Parties Shape Violence. New York University Press, New York and London.

Dennen, J. van der. 1996. The Origin of War: The Evolution of a Male-Coalitional Reproductive Strategy. Origin Press, Groningen.

Ferguson, R. Brian 2013a. “Pinker’s List: Exaggerating Prehistoric War Mortality,” in Douglas P. Fry (ed.), War, Peace, and Human Nature: The Convergence of Evolutionary and Cultural Views, Oxford University Press, New York. (forthcoming).

Ferguson, R. Brian 2013b. “The Prehistory of War and Peace in Europe and the Near East,” in Douglas P. Fry (ed.), War, Peace, and Human Nature: The Convergence of Evolutionary and Cultural Views, Oxford University Press, New York. (forthcoming).

Gat, Azar. 2006. War in Human Civilization. Oxford University Press, Oxford and New York.

Harris, Marvin. 1980. Culture, People and Nature: An Introduction to General Anthropology (3rd rev. edn.). Harper and Row.

Keeley, Lawrence H. 1996. War Before Civilization: The Myth of the Peaceful Savage. Oxford University Press, New York and Oxford.

Keeley, Lawrence H. 2001. War Before Civilization. Oxford University Press, New York and Oxford.

Lacina, Bethany and Nils Petter Gleditsch. 2005. “Monitoring Trends in Global Combat: A New Dataset of Battle Deaths,” European Journal of Population 21: 145–166.

LeBlanc, Steven A. 2003. Constant Battles: The Myth of the Noble Savage and a Peaceful Past. St. Martin’s Press, New York.

McCall, G. S. and Shields, N. 2008. “Examining the Evidence from Small-Scale Societies and Early Prehistory and Implications for Modern Theories of Aggression and Violence,” Aggression and Violent Behavior 13: 1–9.

Pinker, Steven. 2011. The Better Angels of our Nature: Why Violence has Declined. Viking, New York, NY.

Steckel, R. H. and Wallis, J. 2007. “Stones, Bones, and States: A New Approach to the Neolithic Revolution,” http://www.nber.org/confer/ 2007/daes07/steckel.pdf

Thayer, Bradley A. 2004. Darwin and International Relations: On the Evolutionary Origins of War and Ethnic Conflict. University Press of Kentucky, Lexington, KY.

Thorpe, I. J. N. 2003. “Anthropology, Archaeology, and the Origin of War,” World Archaeology 35: 145–165.

Walker, Phillip L. 2001. “A Bioarchaeological Perspective on the History of Violence,” Annual Review of Anthropology 30: 573–596.

Wrangham, Richard and Dale Peterson. 1997. Demonic Males: Apes and the Origins of Human Violence. Bloomsbury, London.