Speech Acts in the Digital Sphere: A Corpus-Based Exploration of Interactional Dynamics on Social Media Platforms

Abstract This study was motivated by the need to understand the ways users perform speech acts on social media platforms, specifically Twitter, Facebook, and Instagram, and how these acts differ between public and private contexts. The purpose was to analyse the frequencies and types of speech acts (requests, apologies, and compliments) and identify the linguistic and pragmatic strategies employed. Using a mixed-methods research design, a corpus of 3 million posts was collected and analysed. Stratified random sampling ensured a balanced representation of speech acts, and both manual annotation and machine learning techniques were used for classification. Three major findings emerged: first, requests were significantly more frequent and direct in private messages than in public posts across all platforms; second, public apologies were more formal and detailed, while private apologies were concise and personal; third, Instagram had the highest frequency of compliments, with public posts being more explicit and enthusiastic compared to private messages. The study concluded that context and platform-specific features heavily influence communication strategies. These insights advance theoretical understanding and offer practical applications for optimizing social media communication.