I am all for attempting to use evolutionary models to help understand how we humans fit in the world and if they can help us better understand why we are the way we are. However, there seem to be a lot of recent attempts to provide simple answers for really, really complex human realities via superficial comparisons to other animals.
Human aggression and violence are complex and very important. Understanding the how’s, why’s, and where’s of lethal violence is a critical aspect in the study of humanity. But we need to do it thoughtfully and scientifically, recognizing that simple answers are likely wrong (as with almost every aspect of the human existence). So when a major science journal (Nature in this case) publishes an article entitled “The Phylogenetic Roots of Human Lethal Violence” we should all read it. And if that article has significant errors or flaws, we should point them out and use them as learning points to make our discussion of violence better and more effective.
So, go read the article here, then come back and read the brief response I offer below.
Here are five points for your consideration. It turns out that I actually agree with much of what the authors’ data tell us, however not with how they interpret or represent what those data mean. This is a good example of how science should work… back and forth about what data might, or might not, mean.
1) “Aggression” is not a uniform or discrete trait (see here) and therefore cannot be modeled like one and thus is not amenable to a phylogenetic analyses as one “thing” across all animals. Also, aggression does not equal violence, and violence does not equal lethal violence. There are lots of problems with these definitions and how the data to assess violence are collected and represented. It is possible to use phylogenetic models (models that compare elements across different kinds of organisms looking for ancestral roots and patterns) using what we call “continuous characters,” things that range in their expression, and that is what the authors purport to do…however, in this case the kind of data they use for the huge range of types of animals vary from one another quite a lot, especially the fossil and archeological data they use for humans. Which leads to my second point…
2) I am confused as to the validity of the dataset presented as the authors offer examples of hamsters and horses as “peaceful” mammals…and in fact that does not jibe with rates and patterns of aggressive behavior in those mammalian taxa (nor do they clarify if they are referring to domestic or non-domestic forms) and they call out whales (and I assume orcas and dolphins) as having little violence (again problematic). And, their use of the archaeological dataset contains numerous omissions, they are missing a lot of the sites and decided to use a date range of 50,000-10,000 for the Paleolithic (not clear reason why).
3) The authors also misrepresent the current debates about the origins and causes of human violence. They state that some researchers think of lethal violence as a cultural trait (that it has nothing to do with evolution or biology). That is not accurate. The authors they cite see lethal violence as an outcome that emerges from a diverse and complex range of ecological, social, historical, economic, physiological and evolutionary interconnections, not just “culture.”
4) The author’s state that because “evolution has shaped violence” one can conclude that “violence can be seen as an adaptive strategy.” This is a nonsensical statement as almost everything has been shaped by evolution and very few things are adaptive strategies. One needs to demonstrate specific suites of benefits for a set of behaviors (lethal violence in this case) in an evolutionary context to make any claims of adaptiveness.
5) Finally, the authors argue for a two percent level of lethal violence across the vast majority of human evolution (which I think actually fits the available data we have to date) and then they note that it shoots up when we start to see increased complexity and inequality in the archeological record. Then they state that, as a percentage of total population numbers, it decreases in nation-state societies with the capacities to enforce broader social control. There is room to debate this, but it is not an outlandish assertion at all. Many researchers have made similar arguments. However, in the article the authors leap from this line of reasoning to the assertion that “humans have phylogenetically inherited their propensity for violence.” Meaning we are naturally lethally violent and we are so due to its origin in our deep past. But what their actual data show is a level of about two percent for the vast majority of our evolutionary history and then an increase when political and economic complexity and inequality ramp up. So really, they are actually making the case for a societally mediated augmentation of the ecological conditions that facilitate an increase in our ancestrally low levels of violence. That is sort of an argument that that they seem to deny in the introduction to their paper.
Clearly this article and the dataset deserve a broader and more comprehensive discussion. In my opinion there are a number of other errors and logical fallacies in the article, but it passed peer review and is lighting up the social media world. So on one hand that is good…debate and discussion about science and human evolution is important. However, I am increasingly worried that it is not always good science that gets articles into the major science journals. I get the feeling that nothing sells like naturalistic explanations for horrible violence, if an article argues it is in our genes it seems to get into press quickly. As my colleague, the anthropologist Marc Kissel, noted to me recently, if we tried to publish an article entitled “the phylogenetic roots of human peace” (which by the way would be way higher than two percent) would it get as much press? Not likely.