Filters

Analyzing Dangerous Speech in Bangladesh

Prof. Aynul Islam of the University of Dhaka proposed a wide-ranging, detailed study of dangerous speech in Bangladesh, where it is rife, to the research team at the DSP. We agreed and in 2021 and 2022, worked with Prof. Islam and a team of professors and graduate students throughout Bangladesh to collect and study 124 diverse examples of dangerous speech. Together we analyzed the distinctive characteristics, messages, and sociopolitical contexts through which such speech builds support for collective violence against women, religious minorities, atheists, Rohingya refugees, and other marginalized groups.

This is the largest detailed study so far of dangerous speech in one context, and it generated important new ideas, including new hallmarks (characteristic rhetorical patterns) of dangerous speech. We are proud of this collaboration and hope this work will serve as a model for others projects in other countries, for comparative study.

SUMMARY REPORT

DSP Research Director Cathy Buerger introduces the Bangladeshi context in this paper intended to introduce the topic to readers with no knowledge of the country and its history. She explains the project methodology, identifies groups frequently targeted by dangerous speech, and offers illustrative cases targeting each of those groups.

RESEARCH TEAM’S REPORT

This report by Prof. Aynul Islam, head of the research team, offers a firsthand account of the local context and explores the hallmarks of dangerous speech that were identified in the dataset, offering examples of each hallmark with analysis. Prof. Islam also importantly describes cases in which dangerous speakers continue to spread their messages with impunity, and one case in which a target of dangerous speech successfully pushed back.

DATASET ACCESS REQUEST

The annotated dataset is available to researchers who wish to use it for further study. Simply submit a request.

SUPPLEMENTAL DATASET ON ELECTION-RELATED SPEECH

In December 2023, the research team produced a supplemental dataset of speech related to the national elections in January 8, 2024. Click the button below to download it.

Distinctive characteristics, messages, and sociopolitical contexts help speech build support for collective violence against marginalized groups.

DownloadRead More