Within studies of extremism, extremist and non-extremist messages are generally treated as two sets of competing constructed narratives. However, some research has argued that these message forms are not dichotomous and that non-extremist narratives demonstrate overlap with extremist master narratives. The aim of this paper is to test this hypothesis empirically by comparing 250 extremist, 250 mainstream and 250 counter-extremist messages. The paper finds considerable overlap between extremist and non-extremist material. However, an analysis of underlying content suggests that this overlap may not be so much due to the extensive adoption of an extremist master narrative by non-extremist authors, but rather a question of resistance and positioning, specifically, who are authors resisting and why? The findings have implications for counter-extremism policy.