Our kind of people? Detecting populist references in political debates

Findings of the Association for Computational Linguistics: EACL 2023, pages 1227–1243 May 2-6, 2023 ©2023 Association for Computational Linguistics Our kind of people? Detecting populist references in political debates Christopher Klamm Ines Rehbein Data and Web Science Group University of Mannheim, Germany {christopher,ines,simone}@informatik.uni-mannheim.de Simone Paolo Ponzetto Abstract This paper investigates the identification of populist rhetoric in text and presents a novel cross-lingual dataset for this task. Our work is based on the definition of populism as a "com- munication style of political actors that refers to the people" but also includes anti-elitism as another core feature of populism. Accordingly, we annotate references to The People and The Elite in German and English parliamentary de- bates with a hierarchical scheme. The paper describes our dataset and annotation procedure and reports inter-annotator agreement for this task. Next, we compare and evaluate differ- ent transformer-based model architectures on a German dataset and report results for zero-shot learning on a smaller English data. We then show that semi-supervised tri-training can im- prove results in the cross-lingual setting. Our dataset can be used to investigate how politi- cal actors talk about The Elite and The People and to study how populist rhetoric is used as a strategic device. 1 Introduction The rise of populism in Europe and throughout the world has been noted not only in politics and the media but also has been the subject of many studies in political science and related areas (see, among others, Mudde (2007)). The concept of populism, however, is complex and vague and eludes a strict definition. So far, only limited agreement exists on the exact properties of the construct, despite numerous efforts to provide a clear definition. In the literature, populism has been described as an ideology (McRae, 1969; Mudde, 2004), a rhetoric (Abts and Rummens, 2007) or style (Mof- fitt, 2016), as a political strategy (Weyland, 2001, 2021; Hawkins and Kaltwasser, 2017) and as a dis- course (Laclau, 1977; Aslanidis, 2016), amongst others (see Aslanidis (2018) for a short overview). The Oxford Handbook on Populism (Rovira Kalt- wasser et al., 2017) groups existing work into three dominant approaches to analyzing populism, i.e., (i) the ideational approach of Mudde (2004), (ii) the socio-cultural approach (Ostiguy, 2017), and (iii) the political-strategic approach (Hawkins and Kaltwasser, 2017), each one capturing a different view on populism. Nevertheless, most studies agree that anti-elitism and people-centrism are amongst the core dimen- sions of populist rhetoric, and the two dimensions are therefore included as features in most sur- vey tools used to measure the degree of populism of political parties and actors (Polk et al., 2017; Rooduijn et al., 2019a; Meijers and Zaslove, 2020). One major drawback of surveys, however, is that they only provide us with one score for each party or actor and can not be used to study how pop- ulist rhetoric is used as a strategic tool in different contextual settings. As a result, more and more efforts have been made recently to measure populist and anti-elitist attitudes directly …

Content truncated to 500 words