A Comprehensive Pragma-Linguistic Analysis of ‘Imminent Harm’ in Social Media Discourse: Speech Act Theory, Pragmatic Inference, and the Limits of Free Speech in Online Contexts

Abstract

This research explores the way in which social media narratives construct "imminent harm" through a pragma-linguistic lens. It argues that harmful meanings are rarely directly articulated and can often only be interpreted through indirect means (i.e., via context-sensitive means). The researcher will utilize Austin and Searle's Speech Act Theory and Grice's Pragmatic Implicature Theory to develop a comprehensive analytic framework for analyzing how illocutionary intent is encoded and understood within online contexts, integrating elements of Speech Act Analysis with Pragmatics and Selected Aspects of Critical Discourse Analysis. A corpus of 120 social media posts sampled from 3 different platforms (Twitter (X), Facebook, and Reddit) were analysed using a mixed-methods research design to determine what types of speech acts are represented, what pragmatic strategies are used to convey implicit meanings in the speech acts and how discourse-level features can be used to classify an individual speech act as being harmful. Indirect speech acts (i.e., implicit threats/directives) were found to be more harmful than, for example, direct speech acts due to their use of: implication (implicature), ambiguous interpretations and collective framing; as these factors often generated an excessive level of emotional intensity across a large number of people who perceived such speech acts as imminent. The findings illustrate that the distinction between an individual’s right to free speech and the individual’s right to not be harmed are not always clear-cut, but rather are formed at the intersection of language form, pragmatic inference and social context. Contributions from these findings are made to linguistic theory by establishing the complex nature of harmful discourse detection and regulation, which outlays the need for approaches to be pragmatically informed and context sensitive to capture the implicit meaning of deemed harmful speech and the scope of complexity associated with the very structure of digital communication.