The paper examines the origin and definitions of these descriptors and assesses their relationship with discrete patterns of political behaviour displayed by Islamist groups. The paper argues that many of the Arabic terms are problematic in the analysis of political behaviour and political violence, because they do not signal political preferences, they encapsulate phenomena that are really distinct, and they encourage essentialist analyses of jihadism. The paper then presents an alternative typology of Islamist actors based on revealed political preferences. The preference-based categories - such as “socio-revolutionary”, “irredentist”, “sectarian” etc - not only reflect the main patterns of behaviour of Islamist groups more accurately, but they also allow for more nuanced thinking about the causes of Islamist militancy, and they facilitate the study of jihadism in a comparative perspective. The findings of the inquiry have implications for two important theoretical debates in the study of Islamism, namely the question of the relative importance of politics and religion in determining the behaviour of Islamist actors, and the question of whether Islamism is essentially different from other religious and political phenomena.