Parameterization of manipulative media discourse: possibilities and problems of automatic diagnosis.

The article focuses on identifying and analyzing linguistic markers of manipulativeness in media texts using automatic diagnostics and machine learning, specifically within the media environment of Kazakhstan. Employing methods such as system analysis, content analysis, computer modelling, and comparative analysis, the study utilized PYTHON software to detect language features like subjunctive verb forms, rhetorical questions, double negatives, and frequent use of negation particles as indicators of manipulative discourse. A corpus of 42 Russian-language media texts was analyzed, revealing that pronouns, adverbs indicating degree, and function words are significant markers, while colloquial and invective vocabulary are less common. The research demonstrates the potential and challenges of formalizing and automating the detection of manipulative language in media, emphasizing the need for further development of algorithms and expanded linguistic resources to improve machine recognition of manipulative content.