Words as Weapons

(PDFs of World Policy Journal articles can be purchased through SAGESubscribe to WPJ here)

From the Spring 2012 Speaking in Tongues issue 

By Susan Benesch

Among friends and fans at his boozy 29th birthday party in March 2010, the South African youth leader Julius Malema cocked his right thumb, pointed his finger like a pistol and chanted “Dubulu iBhunu” (shoot the Boer). The crowd sang along merrily.     

Malema sang Dubulu iBhunu again a few days later at a rally at the University of Johannesburg, but this time it was aired on television and translated into Afrikaans, in which ‘Boer’ originally meant ‘farmer’ and is now a derogatory term for Afrikaner. Hundreds of agitated whites filed formal protests and a judge ordered Malema to stop singing Dubulu iBhunu until the matter could be decided in court. He went on anyway, saying he was only preserving an old anthem from the anti-apartheid struggle—a piece of cultural heritage not to be taken literally. He was singing about the Afrikaner-designed apartheid system, he said, not encouraging his listeners to shoot people.

A hailstorm of debate followed in South Africa, especially on the Internet. Whites especially feared that the song was inspiring black South Africans to kill Afrikaner farmers. In recent years, hundreds of white farm owners and managers had been murdered, mostly in connection with robberies, but often with gruesome violence, and sometimes with their wives or children. But the ruling African National Congress defended Malema, then head of its Youth League, and the song. Party spokesman Jackson Mthembu took responsibility for it on behalf of the ANC, saying it “was sung for many years even before Malema was born,” and must be understood in the context of the anti-apartheid struggle. South Africa’s minister of arts and culture, Lulu Xingwana, and other politicians joined in defying the judge by chanting the song on the 31st anniversary of an anti-apartheid fighter’s hanging.

The controversy was still raging months later, in February 2011, when U2’s Bono went on tour in South Africa. Asked about Dubulu iBhunu, he compared it to Irish Republican Army songs he had sung with his uncles as a child. “It’s about where and when you sing those songs,” Bono said. Indeed, speech and song can be powerful catalysts for human action of all kinds, but their meaning and impact depend tremendously on context and on who’s speaking and listening.

Afriforum, an Afrikaner civil rights group, took Malema to court over his song, and in September 2011, Judge Collin Lamont banned it under South African law prohibiting speech that demonstrates a clear intention to be hurtful, to incite harm, or to promote hatred. Minutes after the judge finished reading his thoughtful, hour-long verdict, a knot of Malema’s young supporters belted out Dubulu iBhunu a few feet away from the courthouse door, just as he had warned they would. The singers had a live audience—a row of police standing rigid in their riot helmets. Neither judge nor police could stop a song.

In his courtroom defense, Malema said the fault lay not with him but with journalists who had made Afrikaners aware of the song by translating its words. He failed to convince the judge, but unwittingly put his finger on a deep change taking place in communications worldwide, which demands new policy, especially since law is inadequate to deal with it alone.


People are increasingly privy to communication that they would not have heard (or read or seen) in the past, namely the internal language of other, disparate cultural communities—the songs that members of a group sing together, the jokes they tell to one another, and the words their leaders use to rally supporters, to teach fear and hatred of others outside the group, or to inspire violence.

Muslims from Afghanistan to Saudi Arabia learn that the prophet Mohammed has been lampooned (and to them, defiled) by cartoonists at a provincial Danish newspaper, feminists discover Facebook pages where men gather to trade rape jokes among themselves, and rural Afrikaners hear of a Zulu song that they fear may be catalyzing violence against them in the racially charged atmosphere of present-day South Africa.

Several factors have converged to influence diverse audiences and raise the stakes in many linguistic battlegrounds. In South Africa, a prime factor was the fall of apartheid, as Judge Lamont noted in his Dubulu iBhunu ruling. Since Nelson Mandela became president in 1994, blacks and whites in South Africa have been forced to hear one another, literally and figuratively, as never before. People who “were not accustomed to each other in any way commenced associating and interacting with each other” sometimes producing “extreme social conflict,” the judge observed. In a number of other countries, struggles against authoritarian regimes and sometimes, their fall, have broken walls between culturally or religiously disparate groups. This can cause greater understanding, as in the case of some Han Chinese netizens who have learned to listen to the grievances of Tibetans for the first time. Or it can lead to terror, as it did for Coptic Christians in Egypt.

Moreover, mass migration and shifting boundaries have moved culturally far-flung communities within earshot of one another, since they now share the same nationality—transforming once largely homogeneous societies into stewpots of diversity.The Danish Muslim community that reacted with outrage to the cartoons published in 2005, for example, hardly existed a few decades earlier. And worldwide, the Internet, text messaging, and social media are perhaps the greatest engines of audience diversification, since they allow communities to listen in to one another as never before. Not only do words and images travel faster and further, they hop lightly and quickly over historic boundaries between human communities.

Even language barriers are surmounted by means of technology. Though at times dangerously imprecise, translation can now be automatic and rapid.

It is difficult to overstate the importance of this shift. Since our primitive ancestors learned to speak, it has been a universal and common human practice to gather in a group and listen to a speaker—family elder, religious leader, military commander, or politician. Speaker and audience share a unique body of cultural knowledge and beliefs that shape their understanding of the language that passes between them, whether it is a song, cartoon, or a shouted phrase. When communicating within a group, speakers choose language familiar to their own people, not the “others”—feminists, blacks or whites, the other tribe, another party, or adherents of another religion. The speaker knows the audience’s particular fears, grievances, historical references—and uses that knowledge to make the speech powerful.

Since language travels so easily between communities, it is increasingly unusual for leaders to speak to their own people and not be heard by others. Some, like the Danish cartoonists, are surprised by the outraged and sometimes violent responses they provoke. Others seem indifferent or cravenly satisfied, like Malema or Terry Jones, the pastor who burned the Quran in Gainesville, Florida, in March 2011, ostensibly to “punish” the holy book. Jones persisted even after General David Petraeus, then the American commander in Afghanistan,  as well as the U.S. State Department, and the White House asked him to refrain. The news immediately reached Afghanistan and all too predictably, 22 people including seven foreign UN workers were killed by an enraged mob in Mazar-i-Sharif.


With each well-publicized case, producers of speech—whether preachers, cartoonists, or singers—are increasingly on notice that their words or images are likely to reach audiences of different cultural backgrounds and will be judged by varied standards. As Judge Lamont wrote, after Dubulu iBhunu was the subject of such a widespread furor in South Africa, it “would never be innocuous again.”

Some speakers can be induced to be more careful with their words when they know they will be heard by a wider group, which presents opportunities for new policy and advocacy. Where the audience is not already broadened, it can be useful to diversify it intentionally. This is an alternative to prohibiting certain kinds of speech, which is a bad option since it can impinge on freedom of expression and often fails in practice.

Kenya saw months of mounting inflammatory speech in 2007 when political leaders incited their own tribes against other tribes—too often using vernacular or tribal languages. In a country with 43 distinct ethnic groups, almost everyone speaks a “mother tongue,” usually in addition to Swahili or English, but no one can understand them all. At the end of 2007, after a disputed presidential election, parts of the country exploded into violence. More than 1,000 people were killed, some 500,000 displaced. Now, as Kenya is preparing for its next presidential election, its KTN television network has been experimenting with an audience-diversification technique.

When Kenyan politicians give inflammatory speeches to their own flocks in their own languages, speaking in terms that they would hesitate to use before a wider audience, KTN sometimes broadcasts such clips on national television, subtitled in Swahili or English. The network has chosen to diversify the audience as a way of discouraging or embarrassing those who would use inflammatory speech in their own narrow circles. KTN news director Linus Kaikai says of politicians who use more incendiary language when they speak to their own ethnic groups in their own mother tongues, “It makes them mind what they say. It tends to sanitize their language.”

This technique is related to the standard human rights advocacy technique of naming and shaming. True rogues may not be deterred by attempts to shame them, but some speakers can be influenced, even brought into line by members of their own community. The ANC, for example, expelled Malema from the party at the end of February. He had given the ANC many other reasons to discipline him, including personal corruption and his public criticisms of President Jacob Zuma. But the song, too, had embarrassed the ANC leadership.


Law, although blunt and unwieldy for this task, is the main tool to rein in dangerous speech without trampling on freedom of expression. In the past 15 years, international courts have convicted more than a dozen defendants, most of them Rwandan, for incitement to genocide, and accepted several guilty pleas. The International Criminal Court investigated Kenya’s 2007-2008 violence and is now prosecuting just four Kenyans for crimes against humanity, including a radio broadcaster whose role was limited to speech—a clear indication that the ICC is convinced of the importance of speech in catalyzing mass violence.

Joshua arap Sang, the broadcaster, is accused of inciting violence during his morning call-in show on KASS FM, a Kalenjin-language radio station that remains influential among members of Kenya’s Kalenjin ethnic group. If Sang is convicted, and perhaps even if he is acquitted, his case will become a landmark in international law on criminal speech. 

The ICC alone could not possibly deal with the large universe of speech that inflames tension and may spark violence. In addition, most such speech does not rise to the level of the grave international crimes to which the ICC’s jurisdiction is limited. A policymaking tool is needed to draw the line between speech that should be sanctioned and speech that must be tolerated in the name of freedom of expression, no matter how ugly it may be.

Diverse communities will never agree on which speech is inherently offensive. Is a drawing depicting Mohammed merely provocative, even amusing, or unbearably offensive? International law and policy should focus on a narrow but notable subset—inflammatory speech that precedes violence, especially outbreaks of mass violence like genocide. Before such outbreaks, leaders address their own group with language calculated to dehumanize another target group. Nazi propagandists referred to Jews as vermin and pests, Hutu propagandists spoke of cockroaches, meaning Tutsi people, Slobodan Milosevic described Muslims as black crows. In case after case, such speech is a precursor to mass violence, especially violence against defenseless civilian populations.

The Dangerous Speech Project has gathered typical hallmarks of speech that seems to catalyze just such mass violence and has developed guidelines for analyzing the level of danger posed by a particular turn of phrase: how likely it is to lead to violence in a specific context. This analysis can be applied to any form of expression—a drawing, photograph, or film—not just words.


One can estimate the likelihood that speech will spark violence in any given situation using just these five criteria: the speaker, the audience, the speech itself, the social and historical context, and the means of dissemination.

In each case, one or more of these criteria may be especially important. A speaker can have great influence over a particular audience, while certain audiences may be especially vulnerable, because of economic hardship, fear, or existing grievances. Certain language-related events—defined broadly to include such acts of expression as burning a book—can be particularly powerful. In some cases, it is the last criterion, the mode of dissemination, that is of paramount importance, especially when it is a form of new media. Text messaging is used increasingly to organize riots and massacres in many countries. For youths in developing nations, whose cell phones link them to the wider world and give them a sense of agency and power, a message may pop up on their screens like this one from the 2007, when ethnic violence broke out  in Kenya:

“No more innocent Kikuyu blood will be shed. We will slaughter them right here in the capital city. For justice, compile a list of Luo’s you know … we will give you numbers to text this information.”

In such a case, the mode of dissemination may be as influential as the content.

In the case of Dubulu iBhunu, Malema was an influential leader chanting his song to young followers who are suffering widespread economic hardship and are still disadvantaged compared to white South Africans. Some of the words of the song compare Boers to dogs, dehumanizing them. So it is hard not to construe the refrain, especially with the hand gestures Malema uses, as a call to attack.

A few months after Judge Lamont banned the song, in April 2011, Tokelo Nhlapo, a youth league member and student leader at the University of the Witwatersrand, told a gathering, “I am tempted to sing this song,” and then began chanting the banned song with a new refrain: “Dubulu Lekgoa” or “shoot the whites,” accompanied by his enthusiastic student audience, all on their feet.



Susan Benesch directs the World Policy Institute’s Dangerous Speech project, which is funded by the MacArthur Foundation, the U.S. Institute of Peace, and the Fetzer Institute. She also teaches at American University’s School of International Service.

[Photo: Jon Fife]

(PDFs of World Policy Journal articles can be purchased through SAGESubscribe to WPJ here)



Related posts