The evolution of altruism through war is highly sensitive to population structure and to civilian and fighter mortality Mark Dyblea,1 aDepartment of Anthropology, University College London, WC1H 0BW London, United Kingdom Edited by David C. Queller, Washington University in St. Louis, St. Louis, MO, and accepted by Editorial Board Member James F. O’Connell February 1, 2021 (received for review May 31, 2020) The importance of warfare in the evolution of human social Bowles’ model suggests that warfare between groups could, in behavior remains highly debated. One hypothesis is that intense theory, select for both parochialism (out-group hostility) and warfare between groups favored altruism within groups, a hy- within-group altruism when individuals form small and geneti- pothesis given some support by computational modeling and, in cally differentiated groups that occasionally go to war with one particular, the work of Choi and Bowles [J.-K. Choi, S. Bowles, another and where success in these wars is determined by the Science 318, 636–640 (2007)]. The results of computational models proportion of parochial altruists in each group. are, however, sensitive to chosen parameter values and a deeper As set out in other work by Bowles (2, 13), differences in the assessment of the plausibility of the parochial altruism hypothesis frequency of altruists between groups is critical to the coevolu- requires exploring this model in more detail. Here, I use a recently tion of altruism and war—if individuals frequently migrate be- developed method to reexamine Choi and Bowles’ model under a tween groups or if groups are large, altruistic individuals are much broader range of conditions to those used in the original unlikely to become sufficiently concentrated. This raises the paper. Although the evolution of altruism is robust to perturba- question of how much population structuring is necessary for tions in most of the default parameters, it is highly sensitive to altruism to evolve in Choi and Bowles’ model and how this group size and migration and to the lethality of war. The results compares to empirical estimates of population structuring in show that the degree of genetic differentiation between groups contemporary small-scale societies. Although previous work on ANTHROPOLOGY F ’ ( ST) produced by Choi and Bowles original model is much greater parochial altruism estimated that FST (a measure of genetic than empirical estimates of FST between hunter-gatherer groups. variation explained by differences between groups) was ∼0.08 When FST in the model is close to empirically observed values, between contemporary hunter-gatherer populations (2), these altruism does not evolve. These results cast doubt on the impor- estimates were based on a wide variety of genetic markers in- tance of war in the evolution of human sociality. cluding some which are poor indicators of whole-genome genetic differentiation (26). Subsequent estimates based on differences altruism | war | population structure | parochial altruism | agent-based in autosomal data suggest that differences between groups who modeling could plausibly compete suggest that it is much lower than this (27–29) and similar to that seen in chimpanzees (26). This raises hile humans are capable of cooperation, tolerance, and two questions for the Choi and Bowles model. First, is the degree Wgenerosity toward others, we are also capable of prejudice, of population structure produced by the model similar to em- violence, and war. Although superficially at odds, these two sides pirical estimates? Second, does altruism in the model evolve of human behavior are sometimes closely related, with warfare when population structure is similar to the empirical estimates? promoting within-group solidarity and acts of individual sacrifice. The association between intergroup conflict and intragroup al- Significance truism has led evolutionary theorists including Darwin (1) to – “ hypothesize that the two may have coevolved (2 5). The pa- Many evolutionary theorists have suggested that the human ” rochial altruism hypothesis as typically conceived (6) holds that capacity for altruism was forged in war, with cohesive and if groups containing more altruistic individuals were able to out- altruistic groups outcompeting their selfish neighbors. Assess- compete groups containing fewer altruistic individuals, this could ing this “parochial altruism” hypothesis relies largely on com- have provided positive selection for both within-group altruism putational modeling. Here, I reexamine a well-known model “ ” – and out-group hostility ( parochialism ) (3, 6 8). that explores the coevolution of altruism and war. As well as The plausibility of the parochial altruism hypothesis depends clarifying the importance of factors such as the lethality of war partly on the likelihood that warfare was commonplace during to fighters and civilians, the results show that the evolution of human evolutionary history, a claim lent some support by ar- altruism in this model relies on a degree of genetic differenti- chaeological evidence of mass killings (9, 10) and ethnographic ation between groups that exceeds that seen among hunter- – data from contemporary or historic small-scale societies (11 16) gatherers. Furthermore, when the model produces a more re- – but which remains highly debated (17 20). However, even if alistic population structure, altruism does not evolve, casting warfare were commonplace in human evolutionary history, this doubt on the plausibility of the parochial altruism hypothesis. would not necessarily mean that it was an important force in selecting for within-group altruism. Since direct evidence of past Author contributions: M.D. designed research, performed research, analyzed data, and selection pressures on altruism and war are unavailable to us, we wrote the paper. rely on exploring the coevolutionary dynamics of parochial and This article is a PNAS Direct Submission. D.C.Q. is a guest editor invited by the altruistic behaviors using mathematical or computational mod- Editorial Board. eling. Several models exploring parochial altruism have been Published under the PNAS license. advanced (2–4, 13, 21) as part of a wider literature on the pos- 1Email: [email protected]. sible impact of warfare on the evolution of human sociality This article contains supporting information online at https://www.pnas.org/lookup/suppl/ (22–25). Of these models of parochial altruism, arguably the doi:10.1073/pnas.2011142118/-/DCSupplemental. most influential is a model by Choi and Bowles (3). Choi and Published March 8, 2021. PNAS 2021 Vol. 118 No. 11 e2011142118 https://doi.org/10.1073/pnas.2011142118 | 1of6 Downloaded by guest on September 29, 2021 As set out by Rusch (6), answering these questions is critical to δc, and all dead agents are replaced by the offspring of surviving our assessment of the plausibility of the parochial altruism hy- members of the winning group. When reproduction occurs, new pothesis for the evolution of human altruism. agents mutate to a random phenotype with probability μ. Each Another reason to explore the Choi and Bowles model in generation, a proportion of agents from each group determined more detail is that while computational modeling can be highly by parameter m migrate to a random group (although note that informative, the results of all models will be sensitive to the since agents may replace dead members of other groups during choice of initial conditions and default parameters. While some war, this migration is not the only way for genes to move between parameters can be grounded in ethnographic data, others will be groups). In addition to the original model, I added a third “gene” too abstract to ground empirically, and in all cases it is important with six alleles that is inherited and mutates with the same to explore the impact that each parameter has on model out- probability as the “altruism” and “parochialism” genes but which comes (in this case the evolution of altruism). Most computa- has no effect on fitness. This “neutral” gene allows the mea- tional models do this by using a fix-all-but-one approach in which surement of population structure from locus that is not under one parameter is varied while all others are kept at their default selection (33). values. Choi and Bowles use this fix-all-but-one sensitivity anal- ysis for five of their model parameters. However, the fix-all-but- Results one method reduces the exploration of the model outcomes to a I replicated Choi and Bowles’ original model and carried out small part of parameter space and limits our understanding of 60,000 simulations. In each simulation, parameters were set the relationship between each parameter and the model out- randomly within the ranges listed in Table 1 and model outcomes come and of interactions between parameters (30). Although the were recorded. Across these simulations, the mean proportion of fix-all-but-one approach employed by Choi and Bowles was the altruists in the population (f A) was strongly correlated with standard approach used at the time, methods have subsequently group size and migration rate, moderately associated with the been developed to explore model parameter space more fully lethality of war to both fighters and civilians and with the costs of (30–32). Here, I use a Fitting to Idealized Outcomes (FIO) altruism in the public goods game, and weakly associated with method developed by Gallagher, Shennan, and Thomas (30) to the payoffs of tolerance toward neighboring groups (Table 2). Of reexamine the results of Choi and Bowles’ model of parochial the 60,000 simulations, 29,158 (48.6%) resulted in a mean pro- altruism in order to 1) explore the results of the model in more portion of altruists in the population across all generations (f A) detail and under a broader range of conditions and 2) calculate of >0.5. Histograms of the parameters that produced these the degree of population structure produced by the model and 29,158 simulations are shown in Fig. 1.