Making People Laugh like a Pro: Analysing Humor Through Stand-Up Comedy

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 5206–5211 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 5206 Making People Laugh like a Pro: Analysing Humor Through Stand-Up Comedy Beatrice Turano, Carlo Strapparava University of Trento, Fondazione Bruno Kessler Trento, Italy beatrice.turano@studenti.unitn.it, strappa@fbk.eu Abstract The analysis of humor using computational tools has gained popularity in the past few years, and a lot of resources have been built for this purpose. However, most of these resources focus on standalone jokes or on occasional humorous sentences during presentations. In this paper we present a new dataset, SCRIPTS, built using stand-up comedy shows transcripts: the humor that this dataset collects is inserted in a larger narrative, composed of daily events made humorous by the ability of the comedian. This different perspective on the humor problem can allow us to think and study humor in a different way and possibly to open the path to new lines of research. Keywords: Computational humor, stand up comedy, humor recognition 1. Introduction Among the communicative activities that humans en- gage in, humor is one of the most powerful and com- plex: it holds the power of grabbing attention (Wanzer et al., 2010), establishing rapport with the audience (Stauffer, 1999), and plays a central role in friendship (Gray et al., 2015). Studying humor from a computa- tional perspective is no laughing matter: together with many practical applications that could benefit from such studies, from better recommendation systems to writing-aid tools for writers and comedians (Winters, 2021), computational humor systems could help lin- guists and researchers collect data to build theories of humor and deepen our understanding of the cognitive processes behind it. Furthermore, it might actively con- tribute to deepening the bond between humans and ma- chines (Binsted, 1995). When it comes to Artificial Intelligence, humor is con- sidered an “AI complete problem” (Stock and Strappa- rava, 2003), because the nature of humor itself is hard to grasp and, consequently, to model computationally. Moreover, factors like gender, culture and religious be- liefs are understudied variables that play a role in how we perceive and appreciate humor (see, for an exam- ple, Schweizer and Ott (2016)) . Hence, while studying standalone jokes will not lead us far on the path to Nat- ural Language Understanding, a full understanding of humor requires the study of the linguistic and cultural contexts in which it appears, - especially considering the pivotal role that the context of a joke has for the joke to be considered funny. One way to tackle this issue is studying whole nar- ratives built with the purpose in mind of making the audience laugh, and stand-up comedy offers that kind of narratives. Indeed, comedians performing stand-up usually tell a story that resembles, or is inspired by, real-life experiences, sometimes not comedic at all (as per the famous quote, “Comedy equals tragedy plus time”1), and the jokes they tell are interweaved in a routine composed of bits that are …

Content truncated to 500 words