An Evolutionary Robotic Model of Explicit Agency Detection
Leticia Cruz Bárcenas and Tom Froese
Studying social cognition implies to deal with two different approaches. On the one hand, the individualism point of view mainly used by cognitive science, where social interaction is the result of individual cognition capacities. On the other hand, the interactionist perspective focused on that the resultant behavior of two or more individuals relies on collective dynamical interaction mechanisms. However, the interaction role has not been considered as an important factor in social cognition. Some of the difficulties are related to the identification of qualitative and quantitative essential characteristics of the overall processes (Lenay and Stewart, 2012). In order to have better analytical tools, Auvray et al. (2009) proposed a minimal cognition model that reduces of social cognition to the most basic elements. The aim of this experiment was to identify the underlying mechanisms due to the recognition of an intentional subject. The results showed that the behavior of the individuals encouraged the interactions with the other participant, as well as the discrimination of the objects due to individual’s oscillatory movements. Additionally, there were developed works focusing on the discrimination from online interactions and recordings (Di Paolo et al., 2008) and dynamic stability (Froese and Di Paolo, 2011). In order to continue these efforts, we replicated the Auvray’s psychological experiment using the Evolutionary Robotics methodology. We developed a synthetic model to investigate the interaction dynamics of explicit agency recognition between agents. The experiment demonstrates that there is an engagement between the agents in the interaction with each other in spite of the obstacles encountered in the environment. Our results were similar compared with the original experiment: the clicks between agents 66%, when meeting the shadow objects 28% and the static objects 5%. Regarding the source of stimulation: 69.7% due to the other agent, 17.12% for the shadow object and 12.7% in the case of the static object. We show that the explicit recognition of the other is not just based on individual capacities of the agents but on the dynamics of the joint perceptual activity itself. Accordingly, we confirm the crucial relevance of interaction as a key explanatory factor in order to have a deep understanding of social cognition. Additionally, we consider this experiment as a novel contribution to understanding in a more holistic perspective the explicit recognition of an intentional subject.
Auvray, M., Lenay, C., and Stewart, J. (2009). Perceptual interactions in a minimalist environment. New Ideas Psychol. 27, 79–97.
Di Paolo, E., Rohde, M., and Iizuka, H. (2008). Sensitivity to social contingency or stability of interaction? Modelling the dynamics of perceptual crossing. New Ideas Psychol. 26, 278–294.
Froese, T., and Di Paolo, E. (2011), “Toward minimally social behavior: social psychology meets evolutionary robotics,” in Advances in Artificial Life: Proceedings of the 10th European Conference on Artificial Life, eds G. Kampis, I. Karsai, and E. Szathmary (Berlin, Germany: Springer Verlag), 426–433.
Lenay, C. & Stewart, J. Minimalist approach to perceptual interactions. Front. Hum. Neurosci. 6, doi:10.3389/fnhum.2012.00098 (2012).