Minorities bear the brunt of online hate, yet the language used to target them keeps shifting. Hatred Monitoring Centre launches the Online Hatred Lexicon, a research project mapping hate speech across 30+ minority communities, languages, and regions worldwide.
Why Minorities?
Minorities are disproportionately targeted by hate speech and hate crimes—making up 70% or more of those affected on social media—while also facing higher rates of content moderation restrictions and removals themselves. The language used to express hostility online is constantly evolving. New expressions emerge, existing terms acquire new meanings, and many forms of antagonistic or exclusionary rhetoric vary considerably across languages, cultures, and online communities. Such online rhetoric does not exist in isolation, it constitute hate crimes. Hate crimes are precursors of atrocity crimes, the United Nations has warned.
Lexicon project
As far as we are concerned, existing lexical resources often focus on specific regions, languages, or categories of hate speech, making comparative and multilingual research challenging. The Hatred Monitoring Centre is launching the Online Hatred Lexicon project, a research initiative that aims to document the vocabulary associated with online hatred across regions, languages, and targeted communities.
Rather than focusing solely on explicit slurs, the project records a broad range of expressions that may be used in hostile, discriminatory, or exclusionary contexts. These include slurs, stereotypes, coded language, conspiracy narratives, dehumanising metaphors, historical references, and other forms of identity-based rhetoric. The objective is to document linguistic patterns that have been observed in public discourse. The project will be extended to include analysis of data collected online to understand their practical use in in everyday online interaction.
Each entry in the lexicon includes information on the targeted community, language, country or region, known spelling variants, semantic meaning, category of expression, severity, contextual requirements, and examples of documented usage across online platforms. This structure is intended to facilitate comparative research while recognising that the meaning and function of many terms depend heavily on a specific context.
The initial Online Hatred Lexicon adopts a global perspective. It includes terminology associated with approximately 30 minorities, a wide range of communities across different continents and linguistic traditions, recognising that the forms and expressions of online hostility differ substantially between cultural and political contexts. Where appropriate, the lexicon also documents region-specific expressions that may have limited visibility outside their local environment but are significant within particular online ecosystems.
Communities researched
Communities included in the preliminary research:
| No |
Region |
||||
| Africa | Middle east | South Asia | East & Southeast Asia | Europe | |
| 1 | Banyamulenge | Kurds | Hazara | Uyghurs | Jewish communities |
| 2 | Batwa / Pygmy peoples | Yazidis | Rohingya (Myanmar) | Tibetans | Muslim minorities |
| 3 | Fulani / Peul pastoralists | Assyrians / Chaldeans | Dalits | Papuans | Irish Travellers |
| 4 | Somali minorities | Baha’i community | Baloch people | Hill tribe minorities | Sámi people |
| 5 | Amhara minorities outside their region | Palestinians | Tamil minorities (Sri Lanka) | Rohingya (regional Southeast Asian discourse) | Crimean Tatars |
| 6 | Tutsi (Africa Great Lakes region) | Bedouin minorities | Ahmadiyya Muslims | Roma / Romani people | Kurdish language speakers |
Digital hate is contextual
The socio-cultural context is a central principle of the project. Many expressions can have multiple meanings depending on the speaker, audience, historical setting, or intended use. Some terms may function as neutral descriptors, reclaimed identifiers, or historically significant references in one context while serving as hostile or discriminatory language in another. For this reason, the lexicon documents terms together with contextual information rather than treating individual words as inherently hateful.
The Online Hatred Lexicon is designed as a continuously evolving research resource. By documenting linguistic patterns across multiple languages and contexts, it provides a structured dataset that can support future research on online discourse, multilingual natural language processing, computational social science, and digital content analysis.
As language continues to evolve, so too will the lexicon. We hope that collaboration with researchers and regional experts will help expand its linguistic coverage, improve its accuracy, and strengthen its value as an open resource for the study of online hatred.
Author:
For further details, contact HMC team at hmcentre2024@gmail.com