Do You Really Want to Hurt Me? Predicting Abusive Swearing in Social Media

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 6237–6246 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 6237 Do You Really Want to Hurt Me? Predicting Abusive Swearing in Social Media Endang Wahyu Pamungkas, Valerio Basile, Viviana Patti Dipartimento di Informatica, University of Turin C.so Svizzera 105, 10149, Turin, Italy {pamungka, basile, patti}@di.unito.it Abstract Swearing plays an ubiquitous role in everyday conversations among humans, both in oral and textual communication, and occurs fre- quently in social media texts, typically featured by informal language and spontaneous writing. Such occurrences can be linked to an abusive context, when they contribute to the expression of hatred and to the abusive effect, causing harm and offense. However, swearing is multifaceted and is often used in casual contexts, also with positive social functions. In this study, we explore the phenomenon of swearing in Twitter conversations, taking the possibility of predicting the abusiveness of a swear word in a tweet context as the main investigation perspective. We developed the Twitter English corpus SWAD (Swear Words Abusiveness Dataset), where abusive swearing is manually annotated at the word level. Our collection consists of 1,511 unique swear words from 1,320 tweets. We developed models to automatically predict abusive swearing, to provide an intrinsic evaluation of SWAD and confirm the robustness of the resource. We also present the results of a glass box ablation study in order to investigate which lexical, syntactic, and affective features are more informative towards the automatic prediction of the function of swearing. Keywords: swearing, social media, abusive language detection 1. Introduction Swearing is the use of taboo language (also referred to as bad language, swear words, offensive language, curse words, or vulgar words) to express the speaker’s emotional state to their listeners (Jay, 1992; Jay, 1999). Not limited to face to face conversation, swearing also occurs in online conversations, across different languages, including social media and online forums, such as Twitter, typically featured by informal language and spontaneous writing. Twitter is considered a particularly interesting data source for investi- gations related to swearing. According to the study in Wang et al. (2014) the rate of swear word use in English Twitter is 1.15%, almost double compared to its use in daily con- versation (0.5 − 0.7%) as observed in previous work (Jay, 1992; Mehl and Pennebaker, 2003). The work by Wang et al. (2014) also reports that a portion of 7.73% tweets in their random sampling collection is containing swear words, which means that one tweet out of thirteen includes at least one swear word. Interestingly, they also observed that a list of only seven words covers about 90% of all the swear words occurrences in their Twitter sample: fuck, shit, ass, bitch, nigga, hell, and whore. Swearing in social media can be linked to an abusive con- text, when it is intended to offend, intimidate or cause emo- tional or psychological harm, contributing to the expres- sion of hatred, in its various forms. In …

Content truncated to 500 words