Societies in the wild: Cooperation, norms, and hierarchies
- Autores: Pablo Lozano Rodríguez
- Directores de la Tesis: Miguel Angel Sánchez Sánchez (dir. tes.), Alberto Antonioni (codir. tes.)
- Lectura: En la Universidad Carlos III de Madrid ( España ) en 2022
- Idioma: inglés
- Tribunal Calificador de la Tesis: Sandro Meloni (presid.), Francesca Lipari (secret.), Giulia Andrighetto (voc.)
- Programa de doctorado: Programa de Doctorado en Ingeniería Matemática por la Universidad Carlos III de Madrid
- Materias:
- Enlaces
- Tesis en acceso abierto en: e-Archivo
- Resumen
Societies are very complex systems. The current understanding about the interplay of its components—people, individual connections, emergent norms, institutions and structures—is far from complete. This thesis had the objective of shedding some light on the interaction between reputation, hierarchies, social norms, and cooperation, critical components in our societies.
The starting point of this thesis is reputation. Since the human species has had an enormous capacity to travel long distances, and settlements could accommodate a large number of individuals, the possibility of interacting with many different individuals was very high at a very early stage. Therefore, humans have found mechanisms to reduce the cognitive effort of interacting with numerous individuals that they have not yet met. By resorting to some prescription of how one should interact with or what to expect from other individuals, humans use reputation to guide their behavior in society. Clear examples appear in our everyday life, for instance: one does not buy from a doubtful place or people but instead goes to known places where more individuals have previously put their trust. In this manner, reputation becomes, in a certain way, a collective measure that individuals trust and use to choose the actions they want to make. Remarkably, it is not restricted to local behavior and even transcends to global scales, affecting how countries organize and choose their policies and trade agreements.
Reputation and hierarchies are two intertwined expressions of similar phenom- ena: dominance. Dominance is often mediated—both established and maintained— by a social norm that can be inherited and can evolve over time. Evidence from multiple studies shows that aggression pursuing rank and involving a certain cost- infliction or benefit-withholding strategy is a viable way of establishing social status in human societies. The reason for the study of these systems is that hierarchies, which usually arise from aggressive behavior, and cooperation seem to be incom- patible from the perspective of a rational agent: cooperating with individuals who will control access to the generated resources while not obtaining any benefit is not optimal. However, we observe cooperative hierarchies in nature; therefore, in our case as humans, some other social component must be helping hierarchies to exist, and social norms are a promising candidate mechanism. This suggests that the interaction of norms and dominance has played, and continues to play, a role in structuring our social environments and shaping our behavior. Experiments with real subjects will ultimately answer the question about such interaction.
In this spirit, we started by developing an agent-based model based on actual experiments to predict future behavior in situations involving reputation that are difficult to study in laboratories. This model is defined by rules and can explain the behavior of human groups in the context of dynamic networks, even when reputation is unreliable. With the model verified against experimental results, we considered larger systems, finding that similar behavior could be expected in them. The model proved that extensive experiments could be more statistically significant than run- ning many small ones to improve statistical significance. Additionally, the model showed that small-sized experiments, like those in classrooms and laboratories, cor- rectly replicate the behavior of people in large systems: keep links with others with good reputation and make new links with people similar to the already present.
However, such a significant increase in the number of links would imply that hu- mans have unlimited attention and cognitive abilities to behave that way, which is impossible as social link creation and maintenance are not free, requiring time and attention.
Subsequently, we turned to the question of norms and hierarchy. The majority of models studying social norms to date did not consider a social structure such as hierarchy, whereas most models studying social structures neither consider dynam- ical processes in the hierarchy nor take into account social norms as part of their behavior. We then proposed a model that combines the ideas of hierarchies and social norms, namely an agent-based model to assess if hierarchical structures arise and how that affects cooperation. Hierarchies turn out to be a robust feature in all the considered versions of the model, and, for unconstrained dominance scores, cooperation is limited to the individuals with lowest score. This behavior arises without the intervention of social norms, just because of purely material interest: agents try to maximize their payoff by generating resources even if they might be robbed in the conflicts. Cooperation spreads to all levels in the hierarchy when the score is constrained, both endogenously or exogenously, but in this case norms do play a role, more so in small groups. The main conclusion is that it is possible to transition from an egalitarian society to a hierarchy and maintain a high cooperation level.
Our model assumed specific hypotheses about human behavior that needed ex- perimental validation. We thus carried out experiments that show that individuals do compete for resources, irrespective of whether a social ranking is present, and that hierarchies arise when individuals compete for those resources. Surprisingly, those hierarchies exhibit higher levels of cooperation than the egalitarian groups.
Additionally, we found that individuals discriminate by the action they see in oth- ers: conflicts are more probable with non-contributing individuals, which supports the idea of indirect punishment. Our results results thus suggest possible mecha- nisms that early humans may have used to remain highly cooperative: recognizing cooperation, punishing non-cooperations, and adapting the behavior when other factors affecting the outcomes of conflicts change. Therefore, low-ranked individuals could have spread a cooperative social norm, making cooperation stable in later generations.
All in all, the main message of this thesis is that considering the interactions be- tween emergent phenomena, such as reputation and norms, with the social structure and the individual decision-making process leads to advances in the understanding of real-world contexts and problems. Examples of such advances are the understanding of the relationship between social structure and dynamics and reputation, and the proof that simple, fight-based mechanisms can explain the egalitarian-hierarchical transition.
