Abstract. The issue of radicalization is still a topical issue. This presentation introduces a general model of radicalization based on the self-categorization paradigm adapted to political violence, which has been tested in 34 campaigns of political violence over more than two centuries over four continents. It was derived from a detailed process tracing analysis from my access to sensitive data (and my personal interviews with perpetrators) and tested in historical campaigns by Ragin's method of Qualitative Comparative Analysis/Fuzzy Set. It is basically a two step process under three conditions and argues that political violence and radicalization is first and foremost political. It is a paradigmatic change from available theories of radicalization and has important policy implications. It implies a new theory of social influence and group dynamics. Its basic universal claims can easily be tested in a laboratory.