Social Media and Political Radicalization: Case Study of Indian Twitter/X Discourse

Abstract

Social media has reshaped political communication in India by creating a fast-moving environment in which political narratives, identities, disagreements, and emotionally charged messages circulate at scale.  This study proposes a computational framework for examining whether and how political discourse on Twitter/X displays measurable patterns of rhetorical escalation associated with online radicalization. Rather than equating ideological disagreement or polarization with extremism, the study conceptualizes radicalization as a multidimensional discourse phenomenon involving hostility escalation, out-group derogation, dehumanizing expression, exclusionary framing, and legitimization of coercive or violent action. A major research gap lies in the limited availability of India-focused studies that jointly model linguistic signals, temporal transitions, user interaction structures, and multilingual or code-mixed political communication. To address this gap, the proposed research develops a Radicalization Discourse Intelligence Framework (RDIF) combining transformer-based language representations, contextual emotion analysis, stance detection, temporal sequence modelling, and social-network features. The framework is designed to distinguish routine political disagreement from increasingly antagonistic discourse and to examine how such discourse changes around politically salient events without assigning ideological labels to individuals.