A game of raids: Expanding on a game theoretical approach utilising the prisoner's dilemma and ethnography in situ

Abstract

In this commentary, we set out the specifics of how Glowacki's game theoretical framework for the evolution of peace could be incorporated within broader cultural evolutionary approaches. We outline a formal proposal for prisoner's dilemma games investigating raid-based conflict. We also centre an ethnographic lens to understand the norms surrounding war and peace in intergroup interactions in small-scale communities.

Cultural evolutionary theory can be used to investigate pathways to peace by examining norms for peace historically and across cultures. Researchers have applied game theory simulations to existing and historic interactions between nations extensively (see O'Neill, 1994, for a full review). Glowacki extends these models to understand the conditions that are required for the evolution of peace within small-scale communities. These interactions are raid based, with groups operating attacks for resources followed by retaliation. Building on Glowacki's framework, we describe specific considerations for a game theoretical approach, and propose that modelling (e.g., Bunce & McElreath, 2017; Cohen, Lewin-Epstein, Feldman, & Ram, 2021) alongside ethnographic research, can help determine causal relationships between different cultural norms which are thought to influence the establishment of net sum positive interactions, that is, peace.

Glowacki suggests the use of prisoner's dilemma game. Prisoner's dilemma games have been used since Axelrod and Hamilton's (1981) The Evolution of Cooperation. A crucial consideration when using game theory to determine cultural evolutionary traits is that it maps closely onto real life by capturing cultural complexities, ensuring ecological validity (Pisor, Gervais, Purzycki, & Ross, 2020). We propose that to represent intergroup interactions over time, iterated prisoner's dilemma games should be played. Iterated prisoner's dilemmas have been found to create tit-for-tat cooperative structures on all bar the last round, where the optimal strategy is to defect (Axelrod & Hamilton, 1981). It is impossible for “last rounds” to be known within traditional raids; therefore, iterated games can accurately match such events.

A traditional prisoner's dilemma where participants are anonymous allows for the simulation of real-life unclear threat of retaliation, just as it may be in raid-based warfare, where groups attack without knowing who was originally behind the raid. By allowing individual identities to be known to other players within games, using recipient identity-conditioned heuristics improves the validity of recreating decision making relating to resource acquisition within social networks (Gervais, 2017). As resources must be established for raids to be undertaken, and intragroup negotiation is often taken on the possibility of retaliation, there is clearly communication between and within groups. For the prisoner's dilemma to be representational, one group's choice would be communicated before the other group participates. This would simulate a raid being carried out, and the other group being aware that the choice to cheat had been made, before they in turn decided to cooperate or defect.

Using a sequential game to establish actions between communities could mirror the intragroup interactions mentioned by Glowacki. A prisoner's dilemma game could be played within a society negotiating a raid, where cooperation among the group would be akin to all agreeing to go on a raid and gain maximal resources, while cheating may lead to a smaller individual raid (Fig. 1). In this case, the cooperative act paradoxically results in warfare, and for peace to be established all members of the group must effectively defect by refusing to raid the adversary group and thus not acquiring potential resources. However, despite including a measure of intragroup collaboration, hierarchies are often in place within societies that game theory could miss. Certain additional games such as the pirate game could implement these hierarchies where members are ranked and the most prestigious chooses how to distribute the resources (Moulin, 1986).

Figure 1. Suggested model for a sequential prisoner's dilemma game featuring intragroup cooperation towards a raid to gain additional resources, the response from the target group (B) and the intergroup prisoner's dilemma matrix this results in, allowing for a more comprehensive understanding of the evolution of peace in small-scale societies.

There are other games that could also be applicable, including stag hunt and chicken. Stag hunt is similar to the prisoner's dilemma but allows for evolutionary stability within cooperation and defection. This can be seen as a positive within the models suggested above, as a game for intragroup cooperation towards resource acquisition (Skyrms, 2004). The chicken game could allow for the understanding of the norms of cowardice and appeasement between groups, as found similarly by Bornstein, Budescu, and Zamir (1997).

Glowacki draws on ethnographic data to inform his framework. This is a crucial step in research which is not always present. Long-term ethnographic fieldwork can inform the design of games such as the prisoner's dilemma, and help define the parameters set out in the models used to determine whether peace can be achieved. Ethnographic research exploring such factors as resource allocation, decision making, and most importantly the specific cultural norms which are alleged to facilitate and uphold peaceful interactions are needed to enable researchers to design more appropriate games and set more accurate parameters in the models. Several researchers have included community-situated research to achieve this (Henrich et al., 2005; Pisor et al., 2020) but these efforts could be strengthened by incorporating longer-term ethnography and collaboration with social anthropologists, who typically have more training and experience pertaining to gaining insider perspectives, and conducting research in a more explorative, less hypothesis-driven manner, which could reveal less well-studied aspects of societies which affect how their members cooperate.

Additionally, following Henrich et al. (2005), researchers could play games such as the prisoner's dilemma with their interlocutors, while learning about not just their choices but the context and meaning given to those choices by the players. This would give more strength and accuracy to interpretations of the choices made, and link back into improving the design of both the games and models. Ethnographic research can also be utilised in designing versions of games which are more understandable by the players (Henrich et al., 2005, 2006), which may reduce the likelihood that players take actions which appear to support or undermine theory based on misinterpretations of how the game works (Bayer, Renner, & Sausgruber, 2012), rather than the accuracy of the theory.

Altogether if these proposals are employed and their findings align with the predictions and theories Glowacki draws on, this would strengthen the validity of the concepts put forward, providing more substantial empirical evidence to support them. Game theory and ethnography can be used together to understand how people evolved to live peacefully.