The Road to Success: Assessing the Fate of Linguistic Innovations in Online Communities

Proceedings of the 27th International Conference on Computational Linguistics, pages 1591–1603 Santa Fe, New Mexico, USA, August 20-26, 2018. 1591 The Road to Success: Assessing the Fate of Linguistic Innovations in Online Communities Marco Del Tredici and Raquel Fern´andez Institute for Logic, Language and Computation University of Amsterdam {m.deltredici|raquel.fernandez}@uva.nl Abstract We investigate the birth and diffusion of lexical innovations in a large dataset of online social communities. We build on sociolinguistic theories and focus on the relation between the spread of a novel term and the social role of the individuals who use it, uncovering characteristics of innovators and adopters. Finally, we perform a prediction task that allows us to anticipate whether an innovation will successfully spread within a community. 1 Introduction Language is an incessantly evolving system, and linguistic innovations of different kind are continuously created. New variants originate in concrete communicative situations with particular speakers. After being introduced, some of them are adopted by other speakers belonging to the same community and possibly spread until they become community norms. In contrast, other innovations do not manage to make their way into the community conventions and just disappear after a certain period of time. This process raises many intriguing questions, which have been the focus of attention in Sociolinguistics since the late 1960’s (Weinreich et al., 1968; Chambers and Schilling, 2013). For example, at what linguistic levels (phonology, lexicon, syntax) do innovations arise and succeed? Who are the leaders of change and what are their social characteristics? In this paper, we investigate lexical innovation in online communities, focusing on the latter type of question, by means of a large-scale data-driven approach. We study the interplay between the birth and spread of new terms and users’ social standing in large online social communities, taking as starting point hypotheses put forward in the Sociolinguistics lit- erature (see Section 2). We present a longitudinal study on a large social media dataset, including 20 online forums and around 10 million users overall. We consider each forum as an independent social network, and investigate how lexical innovations arise and spread within a forum. We analyse the fate of approximately 8 thousand innovations (focusing on acronyms such as lol, phonetic spellings such as plis, and other linguistic phenomena usually collected under the term of Internet slang), and relate their spread within a social network to the role of the individuals who use them. We characterise users’ roles within a community by means of a novel, theoretically-motivated tie-strength measure, combined with well-known centrality measures. We show that (1) innovators (users who introduce a new term) are central members of a community, connected to many other users but with relatively low tie-strength, and (2) strong-tie users (who belong to cliques or sub-groups within the community) effectively contribute to the dissemination of a new term. This pattern is surprisingly consistent across the 20 online communities under investigation. In addition, we show that, by solely using information on speakers’ tie strength as predictor variable, we can anticipate whether an …

Content truncated to 500 words