This document discusses modeling verbal and nonverbal communication using formal analytic methods. The research has two goals: 1) contribute to the science of human interaction and intersubjectivity, and 2) develop techniques for computers to detect states of intersubjectivity from video and audio data. It focuses on identifying the structure and sequence of implicit references, understanding, agreement, and sentiment between conversation participants. The approach analyzes video recordings to extract machine-perceivable cues that can detect these behaviors automatically.