Awareness of the surroundings as well as bodily cues is essential for effective social interaction. Nonverbal behavior such as gaze, facial expression, gesture and posture, provide cues for interpreting intent and relational status. These cues are also present as people traverse a physical setting, where locomotion and social expression integrate effortlessly. The goal of this study was to examine how human observers interpreted social behavior of virtual agents constrained by the physical environment of a movie theater with seating rows. An online survey compared videos of agents exhibiting only locomotion, and agents that also exhibited two different levels of social commitment. Results showed that people were able to interpret the social behavior correctly, indicating that adding it on top of complex maneuvering is possible. These results continue to build a bridge between the literature on agent navigation and agent social behavior.