Capturing Order in Social Interactions

Following Aristotle, " Man is by nature a social animal; an individual who is unsocial naturally and not accidentally is either beneath our notice or more than human " (Politika, ca. 328 BC). This is more than an abstract philosophical statement if, twenty five centuries after the great Greek philosopher, domains as diverse as psychology, physiology and neurology, just to mention a few, still investigate how humans are the perfect machines for social interaction: the muscles of our faces are aimed at expressing our subtlest feelings and emotions to others [1], our ears are tuned to human voices more than to any other sound [2], specific brain structures (the mirror neurons) are aimed at imitating and learning from others [3], and the list could continue. As humans appear to be literally wired for social interaction, it is not surprising to observe that social aspects of human behavior and psychology attract interest in the computing community as well [4][5]. The gap between social animal and unsocial machine was tolerable when computers were nothing else than improved versions of old tools (e.g., word processors replacing typewriters), but nowadays computers go far beyond that simple role. Today, computers are the natural means for a wide spectrum of new, inherently social, activities like remote communication, distance learning, online gaming, social networking , information seeking and sharing, training in virtual worlds, etc. In this new context, computers must integrate human-human interaction as seamlessly as possible and deal effectively with spontaneous social behaviors of their users. In concise terms, computers need to become socially intelligent [6]. Such an ambitious plan of filling the social intelligence gap between humans and machines starts from a fundamental problem, namely how to make social phenomena accessible to computers when the only evidence these have at disposition about the world are signals captured with devices like microphones and cameras. The consequent question is: " Do social phenomena leave physical, machine detectable, traces in signals captured with sensors? " One possible answer comes from the findings of human sciences (sociology, anthropology, social psychology, etc.) showing that social phenomena, while appearing unconstrained and spontaneous, are governed by principles and laws and give rise to ordered and predictable behavioral patterns [7]. For example, during social interactions, people tend to mirror postures and facial expressions of individuals they like, play with pencils and other little objects when they are uncomfortable, avoid exchanging mutual gaze with …