Emotional text mining and multilingual corpora: The analysis of #Covid-19 on Twitter

Abstract This study proposes a method to analyze the meanings conveyed by a hashtag in multilingual corpora. In particular, it investigates the #Covid-19 sensemaking process in the first phase of European contagion. Each language can be considered as a system of codes and symbols characterized by its semiotic and semantic aspects, which are difficult to translate. Nonetheless, their investigation is crucial to study the complex symbolic-cultural system conveyed by a specific term, i.e., the hashtag, in a specific language. Accordingly, we use a text mining approach that does not need previously classified features, Emotional Text Mining (ETM), to identify the set of collective representations and symbolic categories conveyed by the hashtag in four European languages (Italian, French, Spanish, German) analyzing almost three million messages produced on March 4–11, 2020, the first phase of the European contagion. The interpretation of the results made it possible to compare the cultural-symbolic system highlighting similarities and differences between languages characterizing this phase of the pandemic.