Psycholinguistic Analysis of Non-Professional Discourse on Corruption in the Social Network VKontakte

Abstract

The study focuses on the analysis of non-professional online discourse related to corruption. Unlike the institutional or expert approach, the focus on everyday, user-generated reflection allows for capturing the deep values, forms of expression and linguistic patterns of everyday morality, which is especially important in the context of the growing importance of social networks as a source of information and the formation of public sentiment. The objective is to identify the psycholinguistic features of non-professional discourse on corruption in the social network VKontakte, the psycholinguistic aspects of non-professional discourse on corruption: the degree of involvement of participants, emotional and rational components, evaluative characteristics, and thematic focus. The study used methods of text mining based on multi–level natural language processing using the TITANIS tool, with an emphasis on identifying linguistic and semantic markers reflecting the features of non-professional online discourse, such as emotional coloring, categoricality, stylistic orientation, and the presence of thematic vocabulary units. A total of 59,297 comments from the VKontakte social network were collected and prepared for analysis, distributed into three thematic subcorps: corruption topics (19,804 comments), medical topics (19,493 comments), and background, neutral topics (20,000 comments). The TITANIS tool was used to calculate text markers based on intelligent processing of user comments from social media, followed by a statistical comparison between the cases. To assess the significance of the differences, the Mann–Whitney U-test was used for normally distributed data and the chi-square criterion for abnormally distributed data. In the case of markers with an abnormal distribution, such as rare dictionary markers, the data were binarised (0 — absence, 1 — presence). The Bonferroni correction was used to prevent the first kind of error in multiple comparisons. As a result, high emotional tension, predominance of negative tone and focus on political aspects, such as the activities of authorities and law enforcement agencies, with low representation of economic topics were revealed in the online non-professional discourse on corruption. Discussions are characterized by decreased rationality and low communicative engagement, with frequent use of cliched expressions and radical sentences. It is shown that the unprofessional discourse on corruption demonstrates stable psycholinguistic features, indicating the moral and emotional positions of the authors. The digital morality of participants in discussions about corruption is characterized by pronounced condemnation and contradictory moral attitudes, where anger and pessimism are combined with radical calls for justice. However, low communicative involvement and avoidance of personal experience indicate that this morality serves more as a tool for public criticism and social identification than as a basis for deep moral analysis or dialogue.