This article examines the linguistic characteristics of propagandistic texts based on the Propaganda and Disinformation corpus annotated according to themes (narratives) and propaganda techniques. The aim of the study is to reveal how linguistic means are employed in constructing manipulative communication and ideological influence. The analysis is grounded in discourse analysis and text linguistics methods, which enable the investigation of lexical, syntactic, semantic, and rhetorical strategies used in propagandistic texts.
The article describes the Propaganda and Disinformation corpus, its principles of compilation and structure, and presents the annotation process, applied criteria, and methodological decisions. The findings show that morphological annotation of the most frequent words (part-of-speech analysis) is not sufficiently effective for identifying most propaganda techniques, although content words are useful for distinguishing propaganda themes (narratives). It is also established that for some propaganda techniques, neither lists of characteristic lemmas nor n-gram generation are sufficient, as the same lexical items frequently recur across different narratives.
Based on lexical analysis, the most readily identifiable propaganda techniques include emotional appeal (emotive and/or evaluative language, labeling), simplification, and vagueness (use of pronouns, modality markers). Lexical items with limited semantic scope are characteristic of the Flag-Waving and Reductio ad Hitlerum techniques, while collocation analysis proves particularly important for identifying simplification through clichés and slogans. The study further demonstrates that some propaganda techniques overlap, which means that their identification requires a combined analysis of multiple linguistic features. The findings noted in the study contribute to a deeper understanding of the mechanisms of propagandistic discourse and may be applied in academic research as well as in media literacy education.
This work is licensed under a Creative Commons Attribution 4.0 International License.