How do Emojis Shape Conversations in Digital Environments?

• Main Authors: Li, Yiming, Schlager, Tobias, Hoegg, Joandrea
• Format: Conference Proceeding
• Language: English
• Subjects: Algorithms; Emojis; Probability; Social networks
• ISSN: 0098-9258

Conversations that take place online suffer from the lack of audio-visual cues. Using real conversation samples from Twitter and online experiments, the current study demonstrates that emojis shape online conversations such that people better understand each other and eventually engage in greater turn-taking behaviors (e.g., more conversation topics and turns). Emojis can be considered the most important linguistic development of the past decades. In 2015, the Oxford Dictionaries chose, for the first time, an emoji as the word of the year (Dictionaries 2015). Emojis are a textual paralanguage (Luangrath, Peck & Barger 2017), which is described as "visual textual paralinguistic elements". Conversations in digital environments (e.g., on social media, chatbots, or forums) lack the non-verbal cues that are common in face-to-face conversations, which led to the development of surrogates for those missing cues (Byron & Baldridge 2007; Ganster, Eimler, & Krämer 2012; Walther 1996). Emojis, as paralanguages in general, thus aid people in expressing themselves in digital environments. Research on the effect of emojis on conversations and especially on the efficiency of conversations is sparse. Previous studies in the area of emojis focus on people's interpretation of emojis (Na'aman, Pro venza & Montoya 2017; Miller et al. 2016; Pohl, Domin & Rohs 2017; Barbieri, Ronzano, & Saggion 2016), the sentiment of emojis, their similarity (Novak et al. 2015; Ljubešić & Fišer 2016; Eisner et al. 2016), as well as their impact on customer outcomes, such as service relationships and purchase intentions (Hill 2016; Smith & Rose 2020). Finally, Li et al. (2019) showed that emojis could significantly affect the consumers' perception of the warmth and competence of the service provider. However, it is still unclear how emojis affect conversations and the work hitherto primarily uses non-experimental methods. At the same time, researchers across disciplines call for more work on language (Kronrod & Danziger 2013; Schellekens, Verlegh & Smidts 2010; Sela, Wheeler & Sarial-Abi 2012). An important gap is the impact of emojis on conversations in digital environments, which are hitherto predominantly using text. Specifically, such conversations lack nonverbal (e.g., facial expressions) and vocal cues (Kiesler, Siegel & McGuire 1984; Culnan & Markus 1987; Walther 1992). While this does not seem to be a major drawback, at first sight, the lack of verbal cues can lead to inefficient conver