Is Human Sacrifice Functional at the Society Level?



Join 36.9K other subscribers

An article published this week by Nature is generating a lot of press. Using a sample of 93 Austronesian cultures Watts et al. explore the possible relationship between human sacrifice (HS) and the evolution of hierarchical societies. Specifically, they test the “social control” hypothesis, according to which human sacrifice legitimizes, and thus stabilizes political authority in stratified class societies.



Their statistical analyses suggest that human sacrifice stabilizes mild (non-hereditary) forms of social stratification, and promotes a shift to strict (hereditary) forms of stratification. They conclude that “ritual killing helped humans transition from the small egalitarian groups of our ancestors to the large stratified societies we live in today.” In other words, while HS obviously creates winners (rulers and elites) and losers (sacrifice victims and, more generally, commoners), Watts et all argue that it is a functional feature—in the evolutionary sense of the word—at the level of whole societies, because it makes them more durable.

There are two problems with this conclusion. First, Watts et al. do not test their hypothesis against an explicit theoretical alternative (which I will provide in a moment). Second, and more important, their data span a very narrow range of societies, omitting the great majority of complex societies—indeed all truly large-scale societies. Let’s take these two points in order.

An alternative theory on the rise of human sacrifice and other extreme forms of structural inequality is explained in my recent book Ultrasociety. By “structural inequality” I mean more than just great differentials in income and wealth that characterize our modern, even democratic, large-scale societies. In addition to HS, these include ruler deification and more generally “despotism,” when there are no constraints on what rulers and elites can do to commoners; for example, kill them without any negative consequences for themselves.

Briefly, my argument in Ultrasociety is that large and complex human societies evolved under the selection pressures of war. To win in military competition societies had to become large (so that they could bring a lot of warriors to battle) and to be organized hierarchically (because chains of command help to win battles). Unfortunately, hierarchical organization gave too much power to military leaders and their warrior retinues, who abused it (“power corrupts”). The result was that early centralized societies (chiefdoms and archaic states) were  hugely unequal. As I say in Ultrasociety, alpha males set themselves up as god-kings.

Human sacrifice was perhaps instrumental for the god-kings and the nobles in keeping the lower orders down, as Watts et al. (and social control hypothesis) argue. But I disagree with them that it was functional in making early centralized societies more stable and durable. In fact, any inequality is corrosive of cooperation, and its extreme forms doubly so. Lack of cooperation between the rulers and ruled made early archaic states highly unstable, and liable to collapse as a result of internal rebellion or conquest by external enemies. Thus, according to this “God-Kings hypothesis,” HS was a dysfunctional side-effect of the early phases of the evolution of hierarchical societies. As warfare continued to push societies to ever larger sizes, extreme forms of structural inequality became an ever greater liability and were selected out. Simply put, societies that evolved less inegalitarian social norms and institutions won over and replaced archaic despotisms.

We now come to my second critique of the Watts et al. paper. I actually agree that Austronesia is a good laboratory for testing many (but not all) theories of social evolution. There is of course a serious deficiency in the dataset that Watts et al use: it’s static. The basis for their data are observations made in the “ethnographic present.” If you want to assign causality, however, you need to get at the dynamics—how things change with time (as causes typically precede effects, the ability to resolve socio-cultural trajectories in time is critical). Russell Gray’s group (AKA Watts et al) has come with a clever way to get around this problem by using linguistic trees. Some cultural evolutionists, like Joe Henrich, are quite critical of this approach. But even if it works, there is a much more important flaw in the Austronesian database.

The most complex society in their sample is Hawaii, which is not complex at all when looked in the global context. I am, right now, analyzing the Seshat Databank for social complexity (finally, we have the data! I will be reporting on our progress soon, and manuscripts are being prepared for publication). And Hawaii is way down on the scale of social complexity. Just to give one measure (out of >50 that I am analyzing), polity population. The social scale of Hawaiian chiefdoms measures in the 10,000s of population, at most 100,000 (and that achieved after the arrival of the Europeans). In Afroeurasia (the Old World), you don’t count as a megaempire unless you have tens of millions of subjects—that’s three orders of magnitude larger than Hawaii!

Why is this important? Because it is only by tracing the trajectories of societies that go beyond the social scale seen in Austronesia that we can test the social control hypothesis against the God-Kings theory. If HS helps to stabilize hierarchical societies, it should do so for societies of thousands, tens of thousands, hundreds of thousands, millions, tens of millions, and so on. So we should see it persist as societies grow in size.

If HS is destabilizing, then once societies become too large, HS would be selected out. HS is a liability at any size, but the larger a society becomes, the harder it is to hold it together. In this sense HS is a greater liability for megaempires than for chiefdoms. Moreover, if you live on a huge continental mass (and not in an isolated archipelago), once somebody figures out how to make hierarchical societies less unequal (and more cooperative), these people will conquer their more despotic neighbors, spreading egalitarian institutions—eventually—through the whole Afroeurasia.

Note that equality here is very relative—post-Axial megaempires (read Ultrasociety on the importance of the Axial Age) also had kings and nobles, and were quite unequal by today’s standards. But they were better than the archaic states. As time went on, they gradually accumulated various egalitarian institutions. After a few thousands of years this process resulted in pretty decent modern societies in which most of us live today. Even the worst ones, like North Korea, don’t practice human sacrifice. Cultural evolution is faster than genetic evolution, but it still needs time to cumulate useful, functional traits, such as the norms and institutions that promote equality, and therefore cooperation.

So we need to go outside the Austronesian material to test the predictions of the social control hypothesis versus the God-Kings hypothesis. This is what we are doing in the Seshat project, and I estimate that we will complete data collection on extreme forms of inequality by the end of summer or early Fall. So stay tuned.

An illustration of the ritual sacrifice by strangulation of 53 of young women (aged 15 to 25) at the Mound 72 burial of an important personage, now referred to as the "Birdman" because of the falcon shaped arrangement of beads around his body. Mound 72 is a ridgetop mound at the Cahokia Mounds Site, a large Mississippian culture mound center located in present day Madison County, Illinois.
An illustration of the ritual sacrifice by strangulation of 53 of young women (aged 15 to 25) at the Mound 72 burial of an important personage, now referred to as the “Birdman” because of the falcon shaped arrangement of beads around his body. Mound 72 is a ridgetop mound at the Cahokia Mounds Site, a large Mississippian culture mound center located in present day Madison County, Illinois. Source

But as a preview (which will not come as a surprise for those of you who read a lot of world history), preliminary results suggest that HS is very common (if not ubiquitous) at a certain stage of social development—and then it invariably fades away. HS must impose a pretty hefty price on social durability and ability to resist external enemies to generate such a clear-cut macrohistorical pattern.

Also, our preliminary results in the Seshat Databank indicate that there are many other kinds of human sacrifice than the one Watts et al. discuss, which targeted people at the low end of the social hierarchy. In some societies, it is aliens, such as captured warriors, are sacrificed. In others we see sacrifice of high-status individuals, such as children of the elites, or their military followers. Things are much more complex—an interesting!—than one might conclude on reading their paper.

A final thought, hearkening back to another paper Watts et al published last year, in which they argued that moralizing High Gods are not necessary for the evolution of political complexity. Same problem. Perhaps High Gods were not needed for complex chiefdoms in Hawaii and Samoa, but the jury is still out on whether megaempires cold survive without them. Again, stay tuned!

Notify of
Most Voted
Newest Oldest
Inline Feedbacks
View all comments

It seems that the constant is an invocation of supernatural properties to cohere society and confer divine legitimacy to hierarchies and minority leadership, whether the sacrifices were tribute for intercession, tribute to existing order with ritualized burial practices that affirmed status in all worlds, as well as the evergreen strategy of submission through mortal fear.


A notorious example of dysfunctionality and the way it is “selected out”. The Mexican massive human sacrifices seem to have involved not only the downtrodden but also many aliens, imported slaves, captive warriors, tribute hostages from neighbouring tribes… were used as an instrument of domination by the Aztecs, and maybe did provide a measure of symbolic social unity, as a mode of “living tragedy” among other ways. But their systematic nature provoked such a degree of resentment among their neighbours (who also practised human sacrifice and cannibalism by the way) that they were more than ready to join Cortés and the Spaniard invaders in order to overthrow the Mexican empire. The ritual was functional at a given stage of development and imperial expansion, but a new player in the game disrupted the play of forces.


One has to add that the functionality of human sacrifice in the Mexican culture was complemented with cannibalism—a meat-rich diet for the warriors and the élites which no doubt helped promote their predatory efficiency on neighbouring tribes, in the peculiar Mexican, in which humans were the only existing form of cattle and beasts of burden—yet another factor which changed with the arrival of the Spaniards.

Karl Frost

I have been reading some lately about John Haidt’s ideas of bi-modal moral foundations, roughly conservative vs liberal. Whether or not this really is a good description of the foundations of human social psychology cross culturally is still very much to be determined, but there is at least some evidence that it might be or might not be far off the mark. The conservative mode of morality gives much greater weight to fundamental moral principles of order, norms and norm enforcement, hierarchy, loyalty. It also seems very much driven by sustained fear. While one theory (mentioned by Penny, above), is that HS serves to reinforce hierarchy (at least for a while) through straightforward fear of being sacrificed, it could also be that the agitation and arousal around the ritual invokes a more conservative psychological mode, which could possibly help reinforce hierarchies. If such were the case, the HS events would be seen as righteous normative behavior, those sacrificed as righteously sacrificed. If so, perhaps they also are extreme and unstable forms of such conservativism-invoking rituals that later gave way to less extreme forms of sacrifice that were more stable, less pushing the edges of how far conservative normative behavior can be pushed.


A conservative in one historical social setting is a radical revolutionary in another and vice versa.

In the first few centuries of Christianity, it was extremely radical and far different from most other religions and social bodies before it.


You write that human sacrifice is not practiced in modern societies, and cite North Korea as an example. But North Korea, and many other societies practice capital punishment. Some societies, such as Iran and Saudi Arabia, even has public executions. Can this be viewed as a form of current human sacrifice? I get that the proximal justifications for the murders are different (to please the gods/god kings vs as a punishment for crimes) but could be that they ultimately serve the same functions? Capital punishment is something that currently is associated with despotism and unequal societies.

Ralph Haygood

The questions that interest me most about human sacrifice and other such absurdities and atrocities are whether there have generally been people who recognized them as absurd and atrocious at the time, as some of us recognize various still-prevalent practices to be; why there haven’t been more such people; and which hereditary and environmental factors tend to yield more such people.


Absurd and atrocious are relative.

To the people of that time and place, sacrificing humans was neither absurd nor atrocious. It was what was required to keep up the “good life”, as they defined it.

To people of another time and place, the vast sea of guns that Americans live in is absurd and atrocious. To people of a future time, that we all all use polluting machines that harm others to travel (cars) may seem absurd and atrocious.


“In fact, any inequality is corrosive of cooperation, and its extreme forms doubly so.”

I would be careful here. The reason is that inequality is inevitable due to the ambiguous and multifaceted nature of equality:
Equality of outcome
Equality to need
Equal rules applied to all
Equality of outcome to effort
Equality of outcome to actual value added
Equal to position within the hierarchy

The above are all valid ways of judging equality, but some of them are 180 degrees opposed to each other. When you add 7 billion people with different skills, values, cultures and so on, inequality is intrinsic in human interaction. In the end, it is how we balance these tradeoffs and definitions that matters most.

Ross Hartshorn

I immediately thought of the example JoseAngel mentions above, where Cortes and a very small band of Spaniards are able to muster a large force of Native Americans to fight against the Aztecs, who were pretty enthusiastic about the whole human sacrifice thing even by the standards of their peers.

But, more interesting would be to see a survey of wars between states that practice human sacrifice, and others that do not, across history (and continents). Perhaps Seshat or something like it will be able to help to add more data points to questions like this soon! Looking forward to Seshat’s big coming out party soon!


“So we should see it persist as societies grow in size.”

Maybe, unless this interacts with some other force like religion or technology. Axial religion displaced HS, our technocratic society seems to be re-introducing it. I’m trying to find the time to write my thoughts about Ultrasociety which has had a profound affect on how I think about war, religion and technology.

As a totalizing system, Liberalism as it has evolved from the enlightenment through industrialization has taken on a cultural importance much like religion had in the post Axial world. The military competition of Axial period likely did what you describe. What has been happening since industrialization seems to me different. The creation of a money/market based totalizing system has evolved into one that appears to be pursuing “human sacrifice”, not on the alters of a “god”, but in the extraction zones of material resources the techno/money complex can use to concentrate wealth.

If you look at the peripheral affects of technology economies, the accelerated mortality of “shock therapy” Russia in the 90s is reminiscent of what we are seeing for the non-college white middle class here where the needs of a system to continue to provide money returns have been allowed to become more important than the lives of marginal people. Similarly, the US wars of the post 9/11 world can be read as necessary to feed a profit machine which continues to see no negative consequences from continued chaos.


I think you may need a historical perspective that extends beyond a few decades.

If you read history, you’ll find that humans have been both far more exploitative and warlike/violent in the past and in current times in places that are less economically developed compared to the current developed world.

  1. Home
  2. /
  3. Cliodynamica
  4. /
  5. Regular Posts
  6. /
  7. Is Human Sacrifice Functional...

© Peter Turchin 2023 All rights reserved

Privacy Policy