Scholars continue to disagree as to how far contentious politics diffuses within and across states and by what mechanisms it does so. We use new data and empirical measures to test polarization during and after the Arab Uprisings of 2010-12. After authoritarian governments began to fall, populations in several states began to polarize between secularists and Islamists over what kind of regime was to replace the ousted one. We hypothesize that this Islamist-secularist polarization was triggered by catalytic events (such as Muslim Brotherhood electoral victory in Egypt) and diffused transnationally owing to social media and satellite television, dividing anti-status-quo actors throughout the region. To examine polarization over time, we collect a comprehensive dataset on elite and citizen Twitter accounts across Arab countries after the Arab Spring. Using item-response models, we model polarization as the difference in latent ideological position between elites, and we show how polarization within countries changes over time in response to exogenous political shocks. By doing so we are the first to offer compelling statistical evidence of the endogenous process of polarization across competing ideological groups and states.