Expressively vulgar: The socio-dynamics of vulgarity and its effects on sentiment analysis in social media

Proceedings of the 27th International Conference on Computational Linguistics, pages 2927–2938 Santa Fe, New Mexico, USA, August 20-26, 2018. 2927 Expressively vulgar: The socio-dynamics of vulgarity and its effects on sentiment analysis in social media Isabel Cachola∗‡ Eric Holgate∗† Daniel Preot¸iuc-Pietro♦ Junyi Jessy Li† ‡Department of Mathematics, †Department of Linguistics, The University of Texas at Austin {isabelcachola@,holgate@,jessy@austin.}utexas.edu ♦Computer and Information Science, University of Pennsylvania danielpr@sas.upenn.edu Abstract Vulgarity is a common linguistic expression and is used to perform several linguistic functions. Understanding their usage can aid both linguistic and psychological phenomena as well as ben- efit downstream natural language processing applications such as sentiment analysis. This study performs a large-scale, data-driven empirical analysis of vulgar words using social media data. We analyze the socio-cultural and pragmatic aspects of vulgarity using tweets from users with known demographics. Further, we collect sentiment ratings for vulgar tweets to study the re- lationship between the use of vulgar words and perceived sentiment and show that explicitly modeling vulgar words can boost sentiment analysis performance. 1 Introduction Vulgarity is common in language use, with vulgar words appearing with a frequency estimated be- tween 0.5% and 0.7% in day-to-day conversational speech (Jay, 2009; Mehl et al., 2007) and 1.15% in Twitter (Wang et al., 2014). Vulgarity can employed for multiple goals: as an intensifier for subjective opinions, as a way to offend or express hate speech towards others, to describe vulgar activities or as a way to signal an informal conversation. Understanding the expression of vulgarity in naturally occur- ring text is thus important for several disciplines such as linguistics, which aims to better understand the pragmatics of vulgarity, for computer scientists which can explicitly model vulgarity in downstream NLP applications and for psychologists who study the socio-cultural factors of profanity. Social media offers researchers access to vast volumes of naturally occurring user-generated content, with language use on social media containing a high level of expression of thoughts, opinions and emo- tions (Java et al., 2007; Kouloumpis et al., 2011). Furthermore, it is a platform for observing written interactions and conversations between users (Ritter et al., 2010). Thus, social media is a ideal medium to observe and study the expression of vulgar words and, in addition, allows us to study the socio-cultural context of this phenomenon. Given the fact that most thoughts can be rephrased to not include vulgarity, the use of vulgar words indicates a purposeful attempt of performing a specific function. Table 1 includes examples of tweets containing vulgar words that perform different functions. Vulgarity is often employed used to express emotion in language and can be used either to express negative sentiment or emotions or to intensify the sentiment present in the tweet (Wang, 2013). In one of the examples, ‘I am stupid as f*ck’ conveys more intense anger, while ‘I am stupid’ conveys a less emotional expression of irritation. Hence, understanding vulgar words is expected to have practical implications in sentiment analysis on social media. Furthermore, vulgar word usage is dependent on the …

Content truncated to 500 words