Hate Speech | Countering Hate Speech
Hate Speech in South Sudan
Since the outbreak of violence in the world’s newest country in December 2013, South Sudanese have called attention to the dangers of hate speech as a means of inflaming further violent conflict. Indeed, online hate speech was seen as a concern even prior to the onset of hostilities in December 2013. Diaspora communities around the world have increasingly voiced their grievances through social media, often using inflammatory language and images.
Analyzing and Monitoring Hate Speech in South Sudan
In response to this problem, PeaceTech Lab developed a project to identify, monitor, and help counter hate speech in South Sudan. The project consists of three main phases designed to contribute to building peace in South Sudan as well as contribute to the community of practice working to address online speech, media, and mass violence. These phases are summarized below:
Develop a lexicon of online hate speech.
By creating a lexicon of hate speech terms commonly used on social media in the South Sudanese context, an analytical foundation (qualitative and quantitative) will be available for use by local and international groups to more effectively monitor and counter hate speech.
Provide data visualizations and social media monitoring.
PeaceTech Lab will use software tools to create visualizations of hate speech ‘influence networks,’ as well as to create historic views of how hate speech terms are used online. This additional analysis will be presented in regular social media monitoring reports and featured on a web portal attached to the Lab’s Open Situation Room Exchange (OSRx).
Validate the lexicon and analysis through a ‘ground-truth’ process of dialogues with local actors.
Several sessions will be held with varied groups of South Sudanese to validate the context of the hate speech terms identified. Discussions will also center on how online narratives of hate can fuel violence on the ground. Findings from the dialogues will be incorporated into a final “lessons learned” report.
In order to compose the lexicon, project staff conducted an online survey of South Sudanese in the country and in multiple Diaspora communities worldwide to identify terms that are potentially contributing to the conflict.
The lexicon of hate speech terms includes definitions and contextual information identified by survey respondents. The terms are listed in order of frequency of appearance. For each term, the ‘Definition’ section contains information provided by respondents about the term’s origins, its general meaning, and related information. The section ‘Why it is offensive and inflammatory’ discusses information provided by respondents as to why they believe the terms were offensive and inflame the conflict, including past usages, historical references to past conflict, and other context. Finally, the ‘Alternative Term’ section has terms provided by respondents which they thought could be used in place of the offensive and inflammatory terms or to mitigate or counter those terms. Survey data was supplemented by additional contextual analysis provided by a small, but diverse, group of South Sudanese advisers.
These terms were used to develop "monitors" of hate speech in social media platforms using Crimson Hexagon's ForSight social listening platform. The monitors used the lexicon and enabled PeaceTech Lab and partners to explore trends surrounding hate speech in various social media platforms. Below are live visualizations drawn from this analysis.
Definition:According to some respondents, Equatorians use this relatively new term to describe Dinkas; others indicate that it’s used widely to create fear about Dinkas’ encroachment on other communities’ traditional lands and annihilation of those communities. It’s based on the slogan for the MTN mobile service provider: MTN is “everywhere you go.” According to one person, it’s “used to target Dinkas who are found all over the country, like MTN service. It targets Dinkas who have abandoned their lands and scattered all over other lands—and especially against Dinkas when they’re traveling. Vehicles are stopped, and drivers are asked whether MTN are in the cars.” This has reportedly happened to public transport on the Juba-Yei road. In the more recent conflict in 2016, the term has evolved to mean the coordination of operations against the Dinkas.
Why it is offensive/inflammatory: The term stirs fear by exaggerating the number and location of Dinkas within South Sudan, suggesting an increasing presence and pervasive (negative) influence throughout the country, specifically in competition for land, access to water, government services, and jobs. It’s a coded, action-oriented word: An MTN with “no service available” may mean a Dinka who’s unarmed and therefore may be attacked.
Other spellings and related references: ariooce.
Definition: While the first term is recognizable to English speakers, respondents said that Dinkas use both terms to refer to people of Equatoria. Combatants of Dinka, Nuer, and Shilluk communities, among others, believe Equatorians didn’t participate in the 20-year Second Sudanese Civil War, which liberated the south from Arab rule. Currently, the term may more generally refer to those who don’t take the government’s side in the recent conflict. It’s also used to justify the rowdy behavior of non-Equatorian people.
Why it is offensive/inflammatory: In reference to the 20-year conflict, it labels an individual or tribe as unpatriotic. One respondent noted the unintended consequences of using such speech: “It attacks an entire community [for] being cowards and could create an urge [for proof of the opposite] from the other.” Indeed, some of those in Equatoria have taken arms against their accusers.
Other spellings and related references: jienge • jiaang • arian jenge, Aryan jeinge, arian janke • jange • jeng • jengi • government of Dinkas, – of Jienges, – of bush persons • jonkos.
Definition: This term is used by Nuer, or those in Equatoria, to refer to Dinkas. There are many variations, including in Arabic, Juba Arabic, Murle, and Bari. However, Dinkas also use the traditional term Jieng (“the people”) to describe themselves; the term may have neutral cultural and historical roots related to the pastoralist backgrounds of many Dinkas. Arian jenge was a term developed in Juba in the 1970s that differentiated naked, pastoralist Dinkas from naked Mundari pastoralists. People in Juba used the terms government of Dinkas and government of bush persons after the CPA came into effect and many South Sudanese descended upon Juba; the South Sudanese, particularly SPLA soldiers, grabbed all manner of resources. The terms are now often used by people critical of Kiir and his government or by people who want to disassociate themselves from the Kiir government.
Why it is offensive/inflammatory: The term degrades Dinkas by associating them with cattle, characterizing the targeted person or group as illiterate, primitive, or barbaric. Specifically, it scapegoats the Dinka people generally for the behavior of government officials or soldiers.
Other spellings or related references: Nuer wiw • Nuer money.
Definition: Nuer for “money”; refers to Nuer who remained allied with the Kiir government.
Select a term and date range to update the visualizations below.
The topic wheel below displays the top themes of conversation from public social media content including the hate speech term(s) selected above. Use it to explore key topics (inner circle) and sub-topics (outer circle) in social media. In the field above, you can select one of four hate speech terms and a different date range to update the visualizations. Dataset: Crimson Hexagon. Content Sources: Twitter, Facebook, Blogs, Comments, Forums, YouTube.
This chart shows on which sites this term is most commonly used, based on our monitoring. Dataset: Crimson Hexagon. Content Sources: Twitter, Facebook, Blogs, Comments, Forums, YouTube.
The word cloud below displays words most commonly used in social media posts including the hate speech term(s) selected above. Use it to explore related words, themes, hashtags, and accounts in social media. In the field above, you can select one of four terms and a different date range to update the visualizations. Dataset: Crimson Hexagon. Content Sources: Twitter, Facebook, Blogs, Comments, Forums, YouTube.
This chart shows which sites this term is most commonly used based on our monitoring. Dataset: Crimson Hexagon. Content Sources: Twitter, Facebook, Blogs, Comments, Forums, YouTube.
The first line chart below displays the overall volume for social media content including any of the terms identified through the lexicon and relating to the South Sudan conflict. You can use this to monitor rises in the use of hate speech in social media.
The second line chart below displays the volume for social media content with a focus on each of the four terms described above (Jieng, Coward, MTN, and Nuer Wew). You can use this to compare volume over time for each of these terms.
Dataset: Crimson Hexagon. Content Sources: Twitter, Facebook, Blogs, Comments, Forums, YouTube.
This chart shows the volume of total social media content over the time period chosen for each of the four terms described above. Dataset: Crimson Hexagon. Content Sources: Twitter, Facebook, Blogs, Comments, Forums, YouTube.
This map shows the total volume of posts coming from each country in the world, filtered for content that is likely to be hate speech. Note: The map draws from geolocated content, and Twitter and Forums are the only content sources that enable geolocation. Given that most of the hate speech content we have identified is in Facebook, news comments, and forums, this map is not highly accurate. Dataset: Crimson Hexagon. Content Sources: Twitter, Facebook, Blogs, Comments, Forums, YouTube.
These visualizations show networks from two primary Facebook pages that have been used to spread hate speech online: Nyamilepedia and South Sudan News. The visualization was generated using the Facebook Fan Page Search function of NodeXL.
Nyamilepedia Facebook Page Social Network
South Sudan News Facebook Page Social Network