ABSTRACT
This chapter presents a relatively new but reliable method to analyze the symbolic and other linguistic functions served by facial displays. It presents the theoretical basis underlying the measure, along with related findings by other nonverbal researchers. The chapter outlines a system with which to analyze conversational displays and provide evidence for its reliability. Listeners can contribute effectively to the ongoing talk through their facial displays without interrupting the speaker. In Chovil, reliability of the scoring system was assessed by a second scorer who received training to code the facial displays. In order to identify specific sources of error that occurred in the scoring process, reliability was assessed in three steps includes: agreement that a facial display occurred, agreement about the general function, and agreement about the specific function. Conversational facial displays have attracted the attention of computer animators in developing more effective human-computer interactions.
