Echoes across the divide: incivility and the fragmentation of issue publics in immigration discourse

Abstract

As immigration remains one of the most polarizing issues in U.S. politics, public discourse around it has become increasingly uncivil, especially on social media. This longitudinal study examines how incivility spreads and escalates across ideologically opposed issue publics on X. Drawing on conflict spiral theory, we conceptualize incivility as both a symptom and a driver of deepening conflict between pro-and antiimmigration publics. Additionally, we consult insights from research on networked publics and networked social influence to understand how online interactions shape the diffusion of hostile rhetoric. Using largescale X data, we deploy large language models to classify users' immigration stances and measure linguistic incivility. We also construct dynamic mention networks and apply graph-based machine learning to trace how incivility spreads over time and across ideological boundaries. The findings illuminate how conflict dynamics manifest through group network structures and how networked interactions spread incivility across groups with different immigration stances.