This post is an adapted version of a presentation given at ASHB2018 in early December 2018. This talk was an opportunity to discuss my current thinking around the interplay between different forms of human (particularly male) violence and human self-domestication–a process which implies selection for lower aggression and increased sociability.
Since ASHB, Richard Wrangham has extended the discussion of human self-domestication and violence in his book, ‘The Goodness Paradox’, which I will be reading shortly. Also, I came across this paper by Kissel and Kim (2018) which is very pertinent to the topic of human warfare, so I’ve added mention of it below. Further, I’ve referenced the idea of a Z-curve in human social equality, this was raised here by Peter Turchin, and is discussed in his book, ‘Ultrasociety’.
I start with an outline of domestication, domestication syndrome, and human self-domestication….
The most compelling evidence we have regarding domestication syndrome comes from an experimental fox breeding program conducted in Russia. This long-running trial clearly demonstrated that selection for lower stress reactivity, (or a dampened fight or flight response), leads to correlated behavioural, morphological, and physiological changes in the descendant population. Comparable results have been replicated in captively-bred mink, rats, and chickens.
These experiments confirm previous observational research which also identifies a suite of correlated traits common to domesticated animals, relative to their wild ancestors or relatives. These traits include behavioural docility, a tendency to paedomorphic morphology and behaviour, smaller bodies, changes in craniofacial morphology, shifts in oestrus cycling and fertility, reduced sexual dimorphism, curly tails, floppy ears, smaller teeth, and changes in pigmentation. I won’t go into detail here, but these changes are now thought to be caused by altered embryonic neural crest cell behaviour (see this link for the original paper by Wilkins et al.).
Having considered the effects of selection occurring under domestication in significant depth, Charles Darwin originally theorised that there must have been a period of ‘unconscious selection’ prior to any human understanding of deliberate breeding for desirable traits. That is, he reasoned that the capture and taming of animals had unintended effects before any deliberate selection began.
Based mainly on archaeological evidence, more recent authors have suggested this early period of domestication would mostly have involved reflex selection against particularly dangerous animals, especially males. This expectation is supported by genetically-focussed research showing substantial y-DNA bottlenecks in many domesticated lineages, consistent with early slaughter and castration of male animals as primary Neolithic management techniques.
Interestingly, a comparison of recent humans with fossil evidence from earlier Homo sapiens also shows multiple traits resembling domestication syndrome. This mightn’t be that surprising given our hyper-sociable, group-living nature, which must imply some selective pressure in favour of sociability and against reactive aggression.
Evidence of domestication in recent humans includes reduced body size, lower sexual dimorphism, smaller cranial capacity, reduced prognathism, declines in masculine facial morphology, overall skeletal gracility, and reductions in tooth size. Note that these traits can all be interpreted as reductions in masculinity since each is somewhat sexually dimorphic today.
Several modern authors use this morphological evidence to infer a process of human self-domestication via selection against aggressive reactivity and in favour of enhanced sociability. For instance, Thomas and Kirby identify self-domestication as a necessary process in the development of human language and complex communication. Further, Cieri and colleagues argue self-domestication led to an enhanced capacity for knowledge exchange and group collaboration, crucial precursors for all subsequent human civilisations.
Three primary mechanisms have been proposed as drivers of selection against reactive aggression leading to self-domestication in humans: Firstly, an improved capacity for social interaction and cooperation could lead to elevated fitness among group-living individuals. Second, sociable and cooperative groups might practice capital punishment or ostracism against particularly aggressive group members, thus dampening their reproduction. A third hypothesis posits female mate choice for relatively less-masculine men. This latter proposal was the focus of a study I recently published with Dr Geoff Kushnick which found evidence in support of this hypothesis, however, we also concluded that multiple selective mechanisms probably operated in concert to reduce aggressive reactivity in humans.
The above explanations have been explicitly addressed towards the self-domestication process, but other mechanisms proposed to explain human sociability, including alloparenting and group collaboration in the face of inter-group violence, could also have involved selection against aggressive stress reactivity, and so might also induce a process of human self-domestication.
At the mention of human inter-group violence, I’d like to raise a point compellingly argued by Richard Wrangham who discusses two primary forms of human aggression. The first is described as ‘reactive’ aggression, where individuals engage in violence triggered by explosive anger or fear response. The other is termed ‘pro-active’ aggression, a mode of violent action which involves deliberate behaviours intended to achieve desired goals. These two forms of aggression involve two very different proximate mechanisms.
Re-active aggression results from an autonomic response, whereas aggression calculated to achieve desired outcomes requires conscious forethought and, at least in the case of inter-group hostilities, a capacity for collaboration. This is because, in inter-group conflicts of the type we associate with coordinated human battles, there are groups of individuals on both sides, and both attackers and defenders must cooperate with their companions in order to fight effectively. This requires calculated and collaborative violence and is very different to reactive interpersonal rage.
Cieri and colleagues’ work suggests that the kind of cooperative human violence evidenced by large scale battles would be more likely following self-domestication via selection against re-active aggression. So, as counter-intuitive as it sounds, violence at the scale of armed conflict between two groups may provide evidence for enhanced human interpersonal sociability and cooperation. This point is also compellingly argued by Marc Kissel and Narn Kim.
The most convincing archaeological evidence for early warfare is dated to around the beginnings of the Neolithic—there may be evidence of interpersonal violence before this, but collaborative inter-group battles appear absent. Interestingly, Karmin and colleagues recently found that this period in history also tends to show extreme y-DNA bottlenecks in human lineages. That is, they found that y-DNA diversity repeatedly crashes at around the dawn of agriculture in multiple world regions.
A recent article by Zeng et al. argues this phenomenon can only be explained by patrilineally-related bands of males engaging in collaborative violence against males of other lineages. According to this hypothesis, victorious bands absorbed women from defeated groups, whilst many competing male lineages were eliminated.
However, if collaborative Neolithic violence provides evidence of increased cooperation and sociability made possible via self-domestication and selection against masculine aggression, it seems odd that Cieri and colleagues found craniofacial masculinity was higher in agricultural populations when compared to past and present hunter-gatherers—see the amended figure from their article, shown below.
Geoff and I discussed one possible explanation for this in our paper; it may be that women tend to have lower status in agricultural societies, and so are less able to exert any preference for less-aggressive males (see the paper for the full discussion).
Another possibility is that social norms specific to these groups might help to keep in-group violence in check. If this were so, social selection against masculine physiology could be somewhat relaxed. In fact, in the face of inter-group hostility, higher masculinity might be favoured (despite the unwanted social impacts of reactive aggression), so long as resulting violence could be culturally controlled and directed against other groups.
The observation of increased craniofacial masculinity among recent agriculturalists offers an interesting link to ideas on human social development, especially social inequality. Peter Turchin has usefully depicted a path of theorised change in human social equality over our evolutionary and social development using a “Z-curve of human inequality” (see figure below). The suggestion is that whilst various cultural mechanisms allowed foraging societies to constrain relative inequality, the emergence of archaic states promoted a concentration of wealth and power for the few. In more recent times (really only a heartbeat in human history), the expansion of certain social institutions and norms are thought to have paved the way for greater accountability among elites and for reductions in relative social inequality.
My interest here is in how the rise of social inequality associated with the emergence of early states might relate to Cieri et al.’s observations of increased facial masculinity in agricultural peoples. The concentration of wealth which new agrarian states allowed would likely have promoted a hierarchy of male dominance and exploitation. But, negotiating a competitive social hierarchy seems more likely to require calculated and ‘pro-active’ forms of aggression, rather than irrational and reactive rage. Wouldn’t Machiavellian manipulators be more favoured, and reactively aggressive individuals more ostracised, in bigger groups of capable collaborators?
Perhaps this curious reversal of a longstanding trend in facial masculinity highlights the complex interplay between social and bio-physiological processes occurring at a time when humans began living in increasingly large and increasingly stratified agrarian populations. In my opinion, the combination of sociopolitical hypotheses regarding human inequality (Turchin’s Z-curve) with genetic evidence of the Neolithic crash in y-DNA diversity, as well as morphological evidence of a partial re-masculinisation among agriculturalists, all point to a Neolithic increase in social control mechanisms. Perhaps we partially reverted towards a more masculine physiology, but we maintained capacity for social collaboration and cohesion due to novel hierarchical forms of social compulsion and control.
Now, I should note that the worldwide sample which Cieri et al. use (from Howell 1973, 1989, 1995) is of relatively recent agricultural people, few of which can be said to exist in large and hierarchical states. So, the assumption that higher facial masculinity in these people indicates a re-masculinisation that also occurred during classical state formation requires further examination. Perhaps there is something about agricultural populations whatever their scale of social organisation that would tend increase physiological masculinity. This is certainly implied by the findings of Cieri et al., though they do not discuss this observation in depth, or offer an explanation.
At this point, I’d like to briefly introduce some recent statistics on human violence. These data are from Australia, but they’re broadly indicative of cross-cultural trends in other parts of the world. In the slide shown below, the table of homicide statistics on the left indicates the dramatic preponderance of male offenders in cases of homicide. The chart on the right gives a rapid indication of how males are most often assaulted by a stranger, whereas women, disturbingly, are far more likely to be assaulted at the hands of a family member. As with homicides, the offender is predominantly male; usually the woman’s intimate partner, or ex-partner.
These kinds of statistics provide evidence of continuing sexual differences in interpersonal aggression and violence. With regard to human self-domestication, they suggest that any socio-sexual selection against reactive aggression should disproportionately impact men. Perhaps not all of the violence represented here involves reactive aggression, but a great deal of it certainly would.
So, now might be a good time to ask, ‘well, just how sociable and domesticated are humans really?’ We have fossil evidence of selection for less-masculine physiology, which we assume indicates long-term declines in reactive aggression, but is this trajectory continuing today? In truth, there are so many potentially influential factors, it is pointless to speculate. However, there are also a range of morphological indicators that might be used in empirical investigation of this question. In theory, reduction, or increase, across a range of measurable masculine traits should provide relevant evidence.
One of the fundamental implications of pre-historic reduction in masculinity is that it further challenges previously respected narratives regarding the evolution of ‘Man’.
This is important because although the language we use in evolutionary anthropology today is typically more inclusive, many underlying social preconceptions about the ‘manliness’ of human survival and fitness still remain unexamined.
A common assumption is that the existence of masculine morphology and behaviour (that is, the presence of any amount of sexual difference) is proof that masculinity must somehow be favoured by evolutionary processes. To some, this suggests that masculine attributes are reproductively superior to non-masculine ones (i.e. ‘alpha male’ type theories), which naturalises masculine behavioural predispositions and the social dominance of men.
But, a pre-historic trend toward declining masculinity which allowed for and promoted an explosion of human social and communicative complexity must create some uncertainty around these pervasive beliefs.
In closing, all humans clearly have the capacity for aggression and violence, but a significant male bias in these behaviours is glaringly apparent. Whilst previous depictions of the evolution of ‘Man’ tended to highlight masculine traits as intrinsic to human survival, prehistoric reduction in physiological masculinity played an important role in our apparent success (so far) by raising average sociability and allowing for the emergence of complex language and culture, thus promoting our sociocultural niche.
This being the case, the process of human self-domestication holds significant implications for anthropological science and for human society today, especially regarding our capacity for collaborative responses to the social and ecological crises we currently face.