Speaking like ordinary people, representing ordinary people?

The rise of populism in Western democracies and beyond has garnered significant attention from researchers. While the prevailing view in the literature conceptualizes populism as a “thin ideology” (Mudde 2004), other scholars contend that it is better understood as a “style of political communication” (e.g., Jagers and Walgrave 2007; Moffitt 2016; Ostiguy and Moffitt 2021). According to these researchers, a defining characteristic of populist politicians is the way they speak. Not only would they be inclined to adopt rhetoric that pits virtuous citizens against corrupt elites, but they would also tend to use simple, informal language (Moffitt and Tormey 2014). This style would reflect a desire to connect with “ordinary” people and to set themselves apart from conventional politicians. Yet, while numerous studies have examined differences in how populist politicians communicate, we still know little about citizens’ attitudes toward the language style of political candidates. Do voters prefer politicians to speak more casually, with simple words and phrases, or do they expect those seeking office to use more formal language? Do all voters share the same expectations regarding how political candidates express themselves? So far, the emerging literature on this topic has focused primarily on the effect of language complexity on voter attitudes. Bischof and Senninger (2024) demonstrate that the level of complexity of a political candidate’s language affects how citizens perceive their socioeconomic status. Kittel (2024), on the other hand, finds that German voters appear less inclined to support candidates who use simple language compared to those who use average complexity language. However, these studies exclusively explore evaluations of written content and thus cannot address how citizens respond when they hear candidates speak. It seems essential to investigate perceptions of spoken language, as the criteria for evaluating oral communication differ from those used for written content, and citizens are far more likely to hear candidates than to read them. This article will examine how language-based judgments shape the public image of political figures. Using two randomized survey experiments conducted in the United States of America, I will analyze the effects of candidates’ language styles on voter attitudes. Unlike previous research centered on the evaluation of written content, I will focus on language registers. Though registers are inherently connected to the level of complexity of spoken or written productions, they carry a more significant political dimension, as they account for the level of formality in communication and are generally associated with specific socioeconomic groups. The first study will examine the effect of language register when political candidates make non-controversial statements, while the second study will focus on the effect of register when candidates use populist rhetoric. The rationale for conducting two studies is to determine whether language register influences citizens’ attitudes independently of the type of message conveyed by the candidates. The first study constitutes a more isolated test, in which only variation in register can be associated with the populist style. The second, by contrast, serves as a more conservative test, wherein two features potentially linked to populism—register and message content—are manipulated simultaneously. In both studies, brief audio recordings representing different experimental conditions will be randomly assigned to respondents. The level of formality—and consequently, the level of complexity—of the scripts of the speeches will vary from one recording to another. After listening, respondents will rate the candidates on various personal qualities and indicate whether they feel the candidate can understand their concerns and represent their interests. They will also assess the likelihood that they would vote for the candidate.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.