The Tower of Babel in science communication on social media: An analysis of linguistic diversity in Twitter mentions of scientific publications

Abstract To unravel the linguistic dynamics of science communication on social media, this study presents a large‐scale, cross‐disciplinary analysis of language use in over 21 million Twitter mentions of 6.7 million scientific publications. While English dominates—accounting for 90.8% of all mentions and serving as a bridging language for the international dissemination of research—90 non‐English languages contribute to a rich and diverse multilingual ecosystem. A strong alignment is observed between the language of non‐English publications and their corresponding Twitter mentions, particularly for languages such as Japanese and Spanish, reflecting linguistic proximity and regional engagement. Importantly, non‐English tweets achieve user engagement levels comparable to those written in English, whereas tweets lacking meaningful textual content consistently receive lower interaction. These findings highlight the inherently multilingual nature of science communication on Twitter and underscore the importance of incorporating non‐English activities into altmetric analyses to ensure a more inclusive and equitable understanding of global scientific discourse.