A diachronic cross-platforms analysis of violent extremist language in the incel online ecosystem

Abstract

The emergence and growth of incel subculture on the internet has triggered a considerable body of research to date, most of which analysing its worldview or mapping its position and connections within the broader manosphere. While this research has considerably enhanced our understanding of the incel phenomenon, it tends to offer a somewhat static, one-dimensional portrayal of what is – like all online subcultures and communities – a highly dynamic and multi-layered environment. Consequently, we lack insufficiently nuanced answers to what is arguably the critical question for law enforcement and academics alike: is this a violent extremist ideology? Using a uniquely extensive corpus covering a range of online spaces constitutive of the incelosphere (forums, sub-Reddits, Instagram channels, blogs, wiki pages) spanning several years, we analyse the evolution of incel language across both time and platforms. Specifically, we test whether this language has grown more extreme over time as online spaces shut down and others emerged. Our findings demonstrate that, while levels of violent extremist language do vary across the incelosphere, they have steadily increased in the main online spaces over the past 6 years. Further, we demonstrate that, while activity on these online spaces is responsive to offline events such as major acts of incel-inspired violence and the COVID-19 lockdowns, the impact of these on violent extremist ideation is not uniform.