Engaging to deceive: strategies of synthetic involvement in fake tweets

Abstract This study investigates the lexico-grammatical and graphemic features of synthetic involvement employed by fake news writers to engage audiences on Twitter (now X). Conceptually, it draws on the linguistic notion of involvement as a form of phatic communication but extends it to include features of digital discourse produced with the intent to manipulate and deceive – hence, synthetic involvement. Methodologically, the study adopts a comparative corpus-based approach to identify features of involvement in a dataset of 13,000 tweets produced by fake accounts operated by the Russian Internet Research Agency. In addition to language markers such as the frequent use of proper nouns, third person present tense reporting verbs associated with standard English, interjections, and personal and possessive pronouns, the study also identifies graphemic features that reflect deliberate attempts to fabricate authenticity, for example, the use of asterisks for inconsistent self-censorship to obscure offensive language. By revealing this wider range of features, the study contributes to the growing body of linguistic research on online disinformation, offering insight into how synthetic involvement is strategically constructed and deployed to deceive and manipulate public opinion.