Deepening Inequalities or Bridging Divides? Tracing Racist Language in German News Media Over Time

As a structural and discursively constructed practice of exclusion, racism is not only embedded but further amplified in news media by recurring patterns such as racist and controversial language. Racist language, as well as the meta-discourse on it, can contribute to stigmatization and marginalization, reinforce societal barriers, and deepen polarization and social division, particularly considering that marginalized individuals too form part of the news media audience. Conversely, informative, inclusive and critical reporting has the potential to challenge these dynamics, raise awareness of structural inequalities, and contribute to increased social cohesion. As societal boundaries and meanings are constantly renegotiated by various actors, these patterns of both exclusion and awareness-raising may have evolved over time. This longitudinal study investigates the dynamics of racist and controversial language (RQ1) and explores the contexts in which these terms are embedded (RQ2) within three traditional newspapers (Bild, Die Tageszeitung, Die Welt) and three far-right alternative news outlets (Junge Freiheit, Politically Incorrect News, zuerst) in Germany over the period from 2010 to 2022. To explore these dynamics, we employ a combination of automated content analysis methods on data sampled by applying a validated search string reflecting common sub-forms of racism and their target groups in Germany. To examine the salience and dynamics of racist and controversial language, we apply a dictionary of relevant terms and phrases in the German discourse to our corpora. Constructing this dictionary involved a multi-step procedure, leveraging knowledge from diverse sources. In addition to lists and glossaries aggregated from manual hand-coding, academic literature and anti-racist organizations, we include terms identified through surveys inviting participants to submit terms they assess as racist to acknowledge the lived experience of affected individuals and those with heightened awareness of racism and marginalization. After ensuring validation using a manual baseline, we apply the final combination of terms to our corpora and study the temporal dynamics of the salience of the annotated terms in a comparative manner. To study the contextualization of the terms, we employ two more inductive procedures: direct co-occurrences of the terms and dynamic BERTopic models of the articles containing the terms. By combining the results of both analyses with manual inspection and validation, we study the contexts in which the terms from our dictionary occur and how the contexts shifted over time. The findings of this study offer valuable insights into the evolving role of media in shaping narratives around racism and marginalization, highlighting the potential to either reinforce social division or encourage cohesion. By identifying key patterns and shifts in racist language over time, this research also contributes to implications for public awareness, journalistic practice and ongoing efforts to combat discrimination and promote social cohesion.