Textual manifestations of social network user frustration

Abstract

The work is devoted to the identification of differentiating textual signs of frustration as a reaction to subjective inability to satisfy a need. The linguistic realization of frustration in the texts of network communication is considered as emotive speech and is a useful tool for diagnosing emotional well-being both at the individual and at the group level. The data of a psycholinguistic empirical study carried out using the methods of artificial intelligence are presented. Texts of 100 social networks users were analyzed: posts and comments had been written in different emotional states — while calm and in frustration. Texts were processed using a linguistic analyzer that can automatically detect 177 text parameters. Clustering on the basis of changes caused by frustration in the measured text characteristics revealed the existence of two clusters of users. The first cluster encompasses people whose texts, under the influence of frustration, reveal an increase in the proportion of singular past tense verbs, 1st person verbs, and 1st person pronouns, the frequency of “pronouns-nouns,” the coefficient of action objectification (number of verbs relative to the number of nouns) and the number of punctuation marks relative to the number of words. The second cluster includes people whose texts reveal an increase in the frequency of nouns and words in the semantic role of predicate, and for the average word length, but a decrease for indicators from the first cluster. It is concluded that there are two linguistic patterns of frustration response: the “story about one’s own actions” and the “story about the case circumstances.”