Conspiracy theory discourse: a corpus-driven cross-linguistic analysis of Telegram data
The present study will provide a corpus-driven analysis of anti-scientific conspiracy theory discourse in a computer-mediated communication setting (CMC), by comparing data gathered from a selection of Italian-, English-, and German-speaking Telegram channels opposing COVID-19 vaccination and denying both climate change and round-Earth theory.
After carrying out an interdisciplinary overview of conspiracy theory literature, a corpus of conspiratorial Telegram posts was built to verify the presence of the linguistic phenomena that literature identifies as indicators of conspiracy theory discourse, possibly expand their classification, and compare them across languages and user communities.
In order to build a balanced corpus, we selected three Telegram channels – one for each of the
conspiracy theories under scrutiny – for each of the three languages we focused on, for a total of nine channels. In order to work on relevant data, human annotators were asked to classify posts as either conspiratorial or non-conspiratorial using the Taguette[1] tool, a user-friendly and open-source online annotation environment. Only conspiratorial posts were included in the corpus, which was then uploaded on the Sketch Engine platform (Kilgarriff et al., 2014) to identify potentially relevant indicators of conspiracy theory discourse exploiting its automatic natural language processing functionalities.
Preliminary findings agree with literature, showing that, among the most relevant linguistic phenomena taking place at the interface between semantics and pragmatics, we can find: (1) belief narratives contrasting an accepted insider group and a threatening outsider group (Holur et al., 2022); (2) a peculiar use of markers of epistemic stance and evidentiality (Catenaccio, 2022); (3) a creative debunking vocabulary made up of expressions that explicitly refer to the fact that another version of the truth lies behind the official one orchestrated by the media (Ebling et al., 2013); (4) the instrumental use of conceptual metaphors to convey biased content as non-literal meaning (Danesi, 2023).
References
Catenaccio, P. (2022). A corpus-driven exploration of conspiracy theorising as a discourse type. Lexical indicators of argumentative patterning. In Conspiracy Theory Discourses (pp. 25-48). Amsterdam /Philadelphia: John Benjamins Publishing Company.
Danesi, M. (2023). Politics, lies and conspiracy theories: A cognitive linguistic perspective. Oxon/New York: Routledge.
Ebling, S., Scharloth, J., Dussa, T., & Bubenhofer, N. (2013). Gibt es eine Sprache des politischen Extremismus? In Die da oben. Sprache, Politik, Partizipation (pp. 43-69). Bremen: Hempen Verlag.
Holur, P., Wang, T., Shahsavari, S., Tangherlini, T., & Roychowdhury, V. (2022). Which side are you on? Insider-Outsider classification in conspiracy-theoretic social media. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (pp. 4975-4987). Stroudsburg: Association for Computational Linguistics.
Kilgarriff, A., Baisa, V., Bušta, J., Jakubíček, M., Kovář, V., Michelfeit, J., Rychlý, P., & Suchomel, V. (2014). The Sketch Engine: Ten years on. Lexicography, 1(1), 7-36.
[1] https://www.taguette.org/