The project developed a custom fine-tuned big data model (a so-called XLM-RoBERTa model) capable of identifying multiple emotional categories—anger, fear, disgust, sadness, joy, and “none of them”—at the sentence level in political texts. The model was made freely available for research purposes.
Using the AI model, MORES analysed the patterns and trends of the emotionalisation of political discourses in Hungary and Germany—with analysis in France and Poland underway. We identified the factors behind increased emotionalisation. For instance, it appears that opposition parties use more emotions than parties in power, and that the increase in the saliency of policy issues (like that of immigration after 2015) also increases the emotional content of political speeches.
The model also helped analysis on the emotional content of social media posts during the 2024 European elections. It seems that context matters: comparing Hungary and Germany, we found that both the uses of emotions and the reactions given by the public to emotional communications differed across the countries.
In the study of political TV series, the research highlights how anger is used to portray the rise of new, populist leaders who claim to speak for "the people" against a corrupt elite. These series often reflect a populist logic and use anger as a key emotional device.
In a groundbreaking approach, the project uses psychological inoculation as a method to increase emotional regulation—that is, awareness of and resilience to the emotional appeal of populist rhetoric. While psychological inoculation has been previously used to increase resilience to disinformation, MORES is the first research endeavour to apply it to strengthen emotional self-awareness and regulation.
MORES is also developing a normative model of democracy that considers the increased role of emotions in politics. We argue that normative democratic theory has two choices:
Either to stick to the classical normative pillars of democracy, reject identity- and emotions-based politics as part of the imperfect reality, and focus on those elements that limit the power of identity and emotions (e.g. fostering deliberative processes, de-polarising mechanisms, awareness raising on political manipulation); or develop a new normative conceptual framework.
MORES pursues the first path while keeping the second under consideration. Possible approaches include evaluating the normative rightness of identities; evaluating the consequences of identity-based choices; and appreciating the deeper existential meaning and the implications for a “good life” of the identity in question.