Framing Hate with Emotions? A Hybrid AI and Discourse Approach to Online Homophobia and Transphobia

Fabienne Baider^*Orthodoxia Mesaritou

• University of Cyprus, Nicosia, Cyprus

This qualitative study builds on previous research on homophobic and transphobic discourse by integrating critical discourse analysis with AI-based methods. Earlier studies identified the topos of threat as a shared feature of both homophobic and transphobic discourse. Whereas prior work primarily focused on cognitive framing strategies, the present study extends this line of research by explicitly examining emotional strategies—specifically the expression of contempt and disgust—alongside argumentative framing. The analysis is based on a corpus of annotated homophobic and transphobic online comments (approximately 1,000 comments). AI-based and lexicon-based sentiment analysis tools, together with statistical analyses, were used to support the qualitative investigation. Despite its limited size, this corpus allows for a fine-grained qualitative comparison of affective and argumentative patterns. To our knowledge, this is the first study to systematically contrast homophobic and transphobic discourse with respect to both emotional and framing strategies, with the aim of informing counter-discursive approaches tailored to each form of prejudice. The results show that contempt overwhelmingly characterizes discourse targeting both communities, while disgust appears predominantly in homophobic comments. In terms of framing, the misfit social frame is dominant in both datasets, whereas the medical frame is more strongly associated with transphobic discourse.