English

When Words Divide: Diachronic Ideological Polarization in Political Discourse on Social Media

Computation and Language 2026-08-02 v1 Social and Information Networks

Abstract

Political polarization has become a defining feature of online discourse, yet its long-term evolution remains poorly understood. We present a longitudinal analysis of ideological polarization in Reddit discussions by measuring semantic differences in the language used by opposing political communities. We construct temporally aligned community-specific word embeddings and quantify ideological polarization as the semantic divergence of political concepts over time. Our analysis shows that ideological polarization has increased substantially during the study period, both at the concept- and topic-level. Unlike prior computational work, which has largely focused on cross-sectional analyses or affective dimensions of polarization at a single point at time, our approach captures the evolution of ideological differences in semantic framing. The proposed framework provides a scalable method for studying the temporal dynamics of ideological polarization in large-scale social media discourse.

Keywords

Cite

@article{arxiv.2608.01176,
  title  = {When Words Divide: Diachronic Ideological Polarization in Political Discourse on Social Media},
  author = {Roy Yitzchak and Noa Lavie and Ella Rabinovich},
  journal= {arXiv preprint arXiv:2608.01176},
  year   = {2026}
}