Classifying Laughter: An Exploration Of The Identification and Acoustic Features of Laughter Types

This thesis seeks to improve the classification of laughter by uncovering its purpose in communication, identifiability, and acoustic features. Reviewing the existing literature, this paper identifies three main types of laughter: affiliative, de-escalative, and power. Consulting with research assistants, this paper then classifies 113 instances of laughter from 62 Congressional Committee meetings published on C-SPAN. The interrater classification agreement suggests individuals can identify and categorize the different types of laughter with context. Additionally, 14 participants were recruited to complete exercises designed to elicit archetypes of the three laughter categories. These study recordings, which included 124 laughter bouts, were analyzed for acoustic features (pitch (Hz), energy (dB), duration, and proportion of voiced laughter vs. silence). The audio analysis indicates acoustic features of laughter are not overall significantly different amongst the three categories and therefore suggests social context, including proximal language and visual cues, predominantly explains the identifiability of the laughter types.