We operationalize on a robot a subset of social behaviors as described by Hall’s proxemics theory and Kendon’s observations of greetings. Our hypothesis is that basing robot behaviors on the social science of such human behaviors will make the robot appear to convey social intelligence. Specifically, we track the location and orientation of a Nao humanoid robot relative to a person, and programmed the robot to engage in a distance salutation, approach, close salutation and transition as described by theory. Overall, our design appears effective in simulating social intelligence, especially with respect to eye contact. However, mechanical limits affects the robot’s ability to express necessary social nuances, including seemingly fine distinctions such as the robot’s slow speed in moving into position, or its inability to direct gaze independent of head position. Our findings suggest that HRI design must consider detailed nuances of how particular expressions of social theory are realized as robotic behaviors.