Linguistic repercussions of COVID-19: A corpus study on four languages

Abstract

The global reach of the COVID-19 pandemic and the ensuing localized policy reactions provides a case to uncover how a global crisis translates into linguistic discourse. Based on the JSI Timestamped Web Corpora that are automatically POS-tagged and accessible via SketchEngine, this study compares French, German, Dutch, and English. After identifying the main names used to denote the virus and its disease, we extracted a total of 1,697 associated terms (according to logDice values) retrieved from news media data from January through October 2020. These associated words were then organized into categories describing the properties of the virus and the disease, their spatio-temporal features and their cause–effect dependencies. Analyzing the output cross-linguistically and across the first 10 months of the pandemic, a fairly stable semantic discourse space is found within and across each of the four languages, with an overall clear preference for visual and biomedical features as associated terms, though significant diatopic and diachronic shifts in the discourse space are also attested.

Keywords

• covid-19
• french
• english
• dutch
• german
• collocates
• semantic categorization
• pandemic discourse