Animals’ understanding of the partner’s role in cooperation

Compared to the rest of the animal kingdom, human large-scale cooperation is exceptional and considered to be at the origin of our unusual cognition, technology and culture. One aspect that might make cooperation outside our immediate social group possible is likely that humans recognize not just the gains to be made from cooperation over years, but that they understand the roles of their partners and can flexibly adjust their own behavior by assessing the impact of the other partys actions on their outcomes. While many non-human species reach functionally cooperative outcomes, only if we elucidate the extent to which they can recognize each others roles can we assess what exactly is or is not unique about human cooperation and start to understand the decisions animals make in cooperative interactions. In recent years, researchers have used economic games to investigate the decision-making processes in humans to test when they are willing to cooperate or rather defect depending on various factors. Such studies have provided an enormous insight into our cooperative interactions and in the last years, researchers have also started to test monkey in similar set-ups. However, in most of these studies it is still unclear in how far the animals decisions are indeed based on an understanding of the partners role. We will break new ground by testing canines, wolves and dogs, to explore exactly that question using similar methods. While wolves and dogs cooperate with each other and also with humans, differences between the two species exist. A better understanding of their decision making processes will help us to 1) better understand these differences and 2) to elucidate the evolutionary origin of such skills. We have competing predictions based on two hypotheses for the origins of canine social skills. If dogs have been selected for increased cooperation during domestication, we expect dogs to outperform similarly raised wolves in their understanding of cooperative decision making processes. Alternatively, if their skills are due to their social-ecology, we predict that wolves will be more cooperative and attentive (at least towards conspecifics) than dogs, predicting a higher propensity for wolves as compared to dogs to understand the role of a partner in economic games. In a new collaboration (Priv. Doz. Dr. Range and Prof. Dr. Brosnan), we will test animals at the Clever Dog Lab and the Wolf Science Center. We will explore if they do, or do not, take their partners actions into account when making decisions in a cooperative assurance game and test the hypothesis outlined above by comparing the canines flexibility and success depending on the strategy played by their partner.

Compared to the rest of the animal kingdom, human large-scale cooperation is exceptional and considered to be at the origin of our unusual cognition, technology and culture. One aspect that might make cooperation outside our immediate social group possible is likely that humans recognize not just the gains to be made from cooperation over years, but that we understand the roles of their partners and can flexibly adjust their own behavior by assessing the impact of the other party's actions on their outcomes. In our project, we first investigated whether dogs actually pay close attention to the actions of their partner (dog and human) in a cooperative buzzer pressing task. The results were very surprising, showing that dogs very closely watch their partners' actions and adjust their own action accordingly. We then ran several experiments to test if wolves and dogs would recruit human partners to solve a cooperative task (and refrain from recruiting a partner if not needed to get access to the reward) and whether they would differentiate between potential partners based on their efficiency and willingness to cooperate. We found that both wolves and dogs recruit human partners when help is needed but do not seem to pay attention to whether the partner was efficient or even willing to cooperate. These results indicate that while dogs understand that they need a partner to cooperate, they do not seem to really understand how the actions of the partner may influence their own outcome. These results were very interesting and, using pet dogs, we investigated this further using an economic game paradigm. Here the animals had the choice between a cooperative apparatus (high-value reward) and an apparatus that they could solve alone (low value reward). Very few dog dyads converged on solving the cooperative apparatus and thus did not seem to take each other's actions into account. Another study showing that dogs might have a limited understanding that the behavior of the partner has consequences for themselves found that they do not seem to reciprocate actions from a human partner. Overall, these results suggest that while dogs (and wolves) can cooperate with humans, they do not seem to understand the contingencies of the actions and might act on simple cues. Currently, we are also testing capuchin monkeys in the same tasks to see how they perform in comparison. The results provide insights into animal cooperation and its limitations and beg for further studies. Because dogs are an integral part of our society, the results are not just interesting for scientist, but also for lay people. Accordingly, the results have been communicated in several public events as well as in seminars and talks designed for dog owners and dog trainers.