1. The identity-extremism nexus in virtual groups: the impact of online group alignment on radicalisation towards violence
- Author
-
Ebner, J, Schroeder, R, Neumann, P, Whitehouse, H, and Kavanagh, C
- Subjects
Radicalization ,Online chat groups ,Political manifestos ,Terrorism ,Violence ,Identity (Psychology) ,Linguistic analysis (Linguistics) - Abstract
Can would-be perpetrators of extreme violence be reliably identified by the linguistic traces they unintentionally leave behind in their online communications? Previous research has shown that identity fusion, a visceral feeling of oneness with the group, can motivate violent self-sacrifice when the in-group is threatened. This thesis tests the hypothesis that a higher degree of identify fusion in online groups, coupled with a range of mediating and moderating variables, increases the likelihood of individuals carrying out violent pro-group action. To date, there has been no systematic analysis of the relationship between identity dynamics within virtual communities and radicalisation towards violence. This research project seeks to fill this gap by investigating the relationship between identity fusion in the digital arena and violent extremist activities in the real world. Using a mixed methods approach, the project combines quantitative natural language processing (NLP) analysis with quantitative text analysis and digital ethnographic research. The first part of the thesis presents a new coding scheme that traces the narratives and linguistic markers found across the written statements published by terrorists prior to launching an attack. A total of 4,000 pages by 15 authors, on a spectrum from violent terrorist to non-violent political manifestos, were analysed in a comparative manifesto analysis and tested via intercoder reliability (ICR) analysis. The statistical and ethnographic findings indicate that linguistic proxies for identity fusion and other relevant variables, such as violence-condoning group norms and dehumanising vocabulary, can be reliably identified and are significantly more prevalent in the documents of would-be terrorists. A weighted score, the so-called “Violence Risk Index”, was created, drawing on the statistical findings of the manifesto analysis. This new language-based violence risk assessment framework was then applied to eight online groups varying in their degree of extreme ideologies, verbal commitment to violence, and real-world links to terrorist activities. A total of over one million messages were collected from online forums and messaging apps and investigated with a view to determining the scale and nature of violence-predicting narratives and language in each of the groups. The calculated Violence Risk Index for each group mirrored the level of real-world engagement in extreme violence and terrorism, thus providing preliminary evidence that the assessment framework offers an accurate estimation of violence risk associated with online groups. This thesis argues that psychologically grounded linguistic markers are a more reliable predictor of extreme violence than taking violent threats at face value. While this project’s findings should be treated with caution, they may contribute to complementing and improving existing early warning systems used by security and intelligence services.
- Published
- 2023