Hilbe Christian, Hagel Kristin, Milinski Manfred
Program for Evolutionary Dynamics, Department of Organismic and Evolutionary Biology and Department of Mathematics, Harvard University, Cambridge MA, United States of America.
IST Austria, Klosterneuburg, Austria.
PLoS One. 2016 Oct 4;11(10):e0163867. doi: 10.1371/journal.pone.0163867. eCollection 2016.
Direct reciprocity is a major mechanism for the evolution of cooperation. Several classical studies have suggested that humans should quickly learn to adopt reciprocal strategies to establish mutual cooperation in repeated interactions. On the other hand, the recently discovered theory of ZD strategies has found that subjects who use extortionate strategies are able to exploit and subdue cooperators. Although such extortioners have been predicted to succeed in any population of adaptive opponents, theoretical follow-up studies questioned whether extortion can evolve in reality. However, most of these studies presumed that individuals have similar strategic possibilities and comparable outside options, whereas asymmetries are ubiquitous in real world applications. Here we show with a model and an economic experiment that extortionate strategies readily emerge once subjects differ in their strategic power. Our experiment combines a repeated social dilemma with asymmetric partner choice. In our main treatment there is one randomly chosen group member who is unilaterally allowed to exchange one of the other group members after every ten rounds of the social dilemma. We find that this asymmetric replacement opportunity generally promotes cooperation, but often the resulting payoff distribution reflects the underlying power structure. Almost half of the subjects in a better strategic position turn into extortioners, who quickly proceed to exploit their peers. By adapting their cooperation probabilities consistent with ZD theory, extortioners force their co-players to cooperate without being similarly cooperative themselves. Comparison to non-extortionate players under the same conditions indicates a substantial net gain to extortion. Our results thus highlight how power asymmetries can endanger mutually beneficial interactions, and transform them into exploitative relationships. In particular, our results indicate that the extortionate strategies predicted from ZD theory could play a more prominent role in our daily interactions than previously thought.
直接互惠是合作进化的一种主要机制。几项经典研究表明，人类应该能够迅速学会采用互惠策略，以便在重复互动中建立相互合作关系。另一方面，最近发现的ZD策略理论发现，使用敲诈策略的主体能够剥削并制服合作者。尽管预计这类敲诈者在任何适应性对手群体中都会成功，但后续的理论研究对敲诈行为在现实中是否能够进化提出了质疑。然而，这些研究大多假定个体具有相似的策略可能性和相当的外部选择，而在现实世界的应用中，不对称情况普遍存在。在此，我们通过一个模型和一项经济实验表明，一旦主体在战略权力上存在差异，敲诈策略就很容易出现。我们的实验将重复的社会困境与不对称的伙伴选择相结合。在我们的主要处理方式中，有一名随机挑选的小组成员被单方面允许在每进行十轮社会困境游戏后更换另一名小组成员。我们发现，这种不对称的更换机会通常会促进合作，但最终的收益分配往往反映了潜在的权力结构。几乎一半处于更有利战略地位的主体变成了敲诈者，他们很快就开始剥削同伴。敲诈者通过根据ZD理论调整他们的合作概率，迫使他们的共同参与者进行合作，而自己却不进行类似的合作。与相同条件下的非敲诈者相比，敲诈行为带来了可观的净收益。因此，我们的研究结果凸显了权力不对称如何危及互利互动，并将其转变为剥削关系。特别是，我们的结果表明，ZD理论预测的敲诈策略在我们的日常互动中可能比之前认为的发挥更显著的作用。
