Mapping Pride in Political Rhetoric: An AI-Driven Emotion Manifestation Analysis of the Hungarian National Assembly Speeches

Although emotional political rhetoric has long been a major concern for the international scholarly community, research in this area still lacks insights from psycholinguistic and AI-assisted big data corpus analysis. This study seeks to address this gap by employing a theory-driven, fine-tuned XLM-RoBERTa NLP model for longitudinal emotion analysis. It explores the varied expressions of pride in politics—a core component of nationalism and political competition—in parliamentary speeches delivered from 1998 to 2022 by political parties represented in the Hungarian National Assembly. Adopting a functionalist perspective, we view displays of pride as a form of communicative behaviour signaling to observers, including members of the social group, the individual and collective accomplishments of the proud Member of Parliament. Such expressions convey that they, along with their group, merit a higher status within the political sphere. Moreover, the repeated objects of pride expressions provide valuable insights into the socio-political values upheld by Hungary's political elites. Given that pride often arises from significant achievements, particularly after overcoming challenges, and that anger can stem from wounded pride, this study examines whether anger-pride dynamics manifest in political communication or if other emotionalization techniques emerge in connection with pride-related expressions. We offer a single-country case study focused on Hungary which draws on a corpus of over 5 million sentences from parliamentary speeches, the analysis also uncovers distinctive features of pride rhetoric across opposition and pro-government parties. These include independent variables such as policy topics, issue ownership, and party positioning on the left-right political spectrum.