Gendered Dimensions of Discourses Around Islamophobia on Twitter in French: A Domain-Topic Analysis

Against the backdrop of the expansion of far-right rhetorics across several French media channels, this paper offers to examine discourses circulating on the social media X (Twitter at the time of data collection) around the notion of islamophobia, from a gendered perspective. We study the discourses that emerge and the actors carrying them out through a systematic computer-assisted analysis of the 123k tweets published in French during the year leading to the 2022 French presidential elections (March 2021 to May 2022) and including at least one of the keywords: ‘islamophobie’, ‘touchepasàmonhijab’, ‘pasvosbeurettes’, ‘ntarajel’ and ‘letustalk’. Our analysis relies on domain-topic modeling, through whch we obtain a partition of the corpus’ documents (tweets) into 302 groups of similar documents (called domains), and a partition of the corpus’ vocabulary in 847 groups of related terms (called topics). We then qualitatively identify and explore the 12 domains that reflect distinct discourses around gender. For each domain, we explore the specific configuration of topics that gives rise to this disourse, the timeline of the political context, the most significant tweets and conversations and the profiles of key accounts. The most sizeable domain is by far the one labeled ‘veiled’ (3,747 tweets) that captures opinions as to whether the veil is oppressive. Another domain labeled ‘school’ contains variations of this controversy in relation to headscarves in school and sports settings (581 tweets). A third discursive tenet crystallizes around the hashtag ‘letustalk’ (1331 tweets in this domain) that refers to the experiences of forced veiling, notably in Iran, to weigh on the French debate. Another example is the domain ‘media’ (886 tweets) which uncovers how broadcasters are discussed in relation to the question of islamophobia. Overall, the paper presents an analysis of Twitter discourses around islamophobia from a gendered perspective, systematized through domain-topic modelling.