Interpersonal interaction, especially in face-to-face circumstances, requires coordination (Clark, 1996). This involves many subtle behaviors, controlled carefully in the context of another person, from eye movements and gestures, to choice of words and beyond. Several new directions have pursued this microstructure of interpersonal interaction. First, with advances in sensing and computing techniques, we have the capability to process visual, audio and other sensory data collected from real-world interactions. Second, researchers in developmental robotics have investigated mechanisms of interpersonal coordination, to model and implement social systems. Third, research on virtual agents has developed new models of embodied human-agent interaction. Together these strands of research offer new insight into human social dynamics, and the means to implement and test theories in robotics and virtual agents. Bringing them together in one workshop is an opportunity to convey these new methods, and find shared interests and synergies among different approaches and different fields.