The Tackling Hate Lab’s Matteo Vergani and Rouven Link have co-authored a new open-access article with colleagues that takes a close look at how the “hate element” is defined in behaviours such as hate crime and hate speech. The aim of the paper is to map the full range of definitions used across countries and fields, and see whether these definitions cluster into clear types. The idea is to bring some order to a space where terms are used often, but not always consistently.
To do this, the authors pulled together 423 definitions of hate crime, hate speech and related concepts published between 1990 and 2021. The definitions came from academic writing and grey literature across ten countries in North America, Europe and Oceania, and in five languages. Every definition was coded using 16 categories derived from theory, covering things like motives, target identities, impacts and social context. These were then converted into a binary matrix and clustered using cosine distance and agglomerative hierarchical clustering. The team used the elbow method to work out the optimal number of clusters and settled on six, which they then grouped into five broader definitional types.
The data show that definitions fall into five main approaches: those that define hate based on the perpetrator’s motives, those based on the hatred shown in the act, those that emphasise intentions, those that draw on historical or structural contexts, and those that focus on the physical or psychological impacts of the behaviour. Each of these approaches can apply at three levels: universal (any group), protected attributes (for example race, religion or gender), or group specific (for example antisemitism or homophobia). Put together, this creates 15 combinations that cover almost every definition in the sample. One clear pattern is that hate crime definitions usually rely on the perpetrator’s motive, while hate speech definitions more often depend on the expression itself, its impacts, or the broader social setting in which it occurs.
The policy implications are direct. A motive-based definition may be straightforward on paper but is hard to prove in court, especially when offenders act on mixed motives. Definitions based on manifestations or impacts may help capture more subtle forms of hate but can become too broad without clear bias indicators. Structural definitions give important context but are harder to operationalise in legal frameworks. The typology helps policymakers see these trade-offs and choose definitions that suit their aims, whether for legislation, policing, victim support or public education.
For research, the framework offers a way to compare definitions across countries and fields and to test which types are most workable in practice. It sets up new questions about which approaches improve reporting, build trust, shape prosecutions, or help communities recognise and respond to hate-driven behaviour.