Signed Networks, Triadic Interactions and the Evolution of Cooperation

Introduction. In this paper we examine how cooperation can be supported by negative ties and triadic interactions. Previous research has demonstrated that cooperation is more likely to evolve in social dilemma games, such as the Prisoners Dilemma, if the game is played in networks [1, 4, 5, 12]. Besides sparseness [3, 4, 5], the structure of the network is also important because on top of direct ties, indirect relations also control behavior and contribute to the establishment of cooperation through reputation mechanisms, such as image scoring [2]. These ideas have been fruitful in explaining why and under which conditions can cooperation evolve and be stable among humans. In these studies however, only positive relations between players are analyzed. Negative ties could potentially be better enforcers of cooperation than positive ones. At the dyadic level, social psychological mechanisms of vengeance and anger are manifested through the use of powerful trigger strategies in the repeated Prisoners Dilemma game [7, 8, 9, 10]. Another mechanism is selectivity that prescribes cooperation with those who are liked and defection with those who are disliked. Experiments demonstrate that subjects play differently in social dilemma games with related and unrelated others (e.g., [6]). There is also evidence of homophilous selection of partners that can stabilize cooperative regimes [11]. Furthermore, we hypothesize that when interactions happen on a triadic basis, cooperation can increase. For instance, if A and B are friends, and C is their common enemy, the presence of the negative ties towards C could strengthen cooperation between A and B. Signed relations could reliably signal cooperative or non-cooperative intentions, and hence could guarantee the spread of appropriate reputational information that result in confirming expectations about cooperation.