Understanding dangerous speech
What is dangerous speech?
Dangerous speech is any form of expression (e.g. speech, text, or images) that can increase the risk that its audience will condone or participate in violence against members of another group. We have observed striking similarities in the rhetoric that leaders use to provoke violence in completely different countries, cultures, and historical periods. One of these rhetorical ‘hallmarks’ or recurring patterns in dangerous speech is dehumanization, or referring to people in another group as insects, despised or dangerous animals, bacteria, or cancer. More hallmarks are listed below. Rhetoric alone can’t make speech dangerous, though; the context in which it is communicated is just as important. One can capture that context, and analyze speech for dangerousness, by asking about five aspects of the speech:
- Speaker: Did the message come from an influential speaker?
- Audience: Was the audience susceptible to an inflammatory message, e.g. because they were already fearful or resentful?
- Message: Does the speech carry hallmarks of dangerous speech? The hallmarks are:
- Dehumanization. Describing other people in ways that deny or diminish their humanity, for example by comparing them to disgusting or deadly animals, insects, bacteria, or demons. Crucially, this makes violence seem acceptable.
- ‘Accusation in a mirror.’ Asserting that the audience faces serious and often mortal threats from the target group - in other words, reversing reality by suggesting that the victims of a genocide will instead commit it. The term ‘accusation in a mirror’ was found in a guide for making propaganda, discovered in Rwanda after the 1994 genocide. Accusation in a mirror makes violence seem necessary by convincing people that they face a mortal threat, which they can fend off only with violence. This is a very powerful rhetorical move since it is the collective analogue of the one ironclad defense to murder: self-defense. If people feel violence is necessary for defending themselves, their group, and especially their children, it seems not only justified but virtuous.
- Assertion of attack on women/girls. Suggesting that women or girls of the audience’s group have been threatened, harassed, or defiled by members of a target group. In many cases, the purity of a group’s women is symbolic of the purity of the group itself, or of its identity or way of life.
- Coded language. Including phrases and words that have a special meaning, shared by the speaker and audience. The speaker is therefore capable of communicating two messages, one understood by those with knowledge of the coded language and one understood by everyone else. This can make the speech more dangerous in a few ways. For example, the coded language could be deeply rooted in the audience members’ sense of identity or shared history and therefore evoke disdain for an opposing group. It can also make the speech harder to identify and counter for those who are not familiar with it.
- Impurity/contamination. Giving the impression that one or more members of a target group might damage the purity or integrity or cleanliness of the audience group. Members of target groups have been compared to rotten apples that can spoil a whole barrel of good apples, weeds that threaten crops, or stains on a dress.
- Context: Is there a social or historical context that has lowered the barriers to violence or made it more acceptable? Examples of this are competition between groups for resources and previous episodes of violence between the relevant groups.
- Medium: How influential is the medium by which the message is delivered? For example, is it the only or primary source of news for the relevant audience?
Why not use the familiar term ‘hate speech’ instead of dangerous speech?
The term ‘hate speech’ is vague, broad, and in practice, everyone defines it differently. This confusion tends to lead to expansive understandings of hate speech which, in practice, can jeopardize freedom of speech. Indeed, laws against hate speech or hateful speech are often misused to punish and silence journalists, dissenters, and minorities, recently in countries as varied as Hungary, India, Rwanda, Kazakhstan, and Bahrain. We focus instead on dangerous speech since it is a narrower, more specific category, and we define it by its link to something that almost everyone can agree on (unlike the definition of hate speech): violence against a group of people is a grave harm that should be prevented. Also dangerous speech can be quite different from hate speech: sometimes it doesn’t even express hate, and instead promotes fear. Definitions of hate speech differ, both in law and in colloquial use, and some aspects of the term remain undefined. For instance, what is hatred? How strong or how durable must emotion be to count as hatred? Another unresolved question is this: does the ‘hate’ in hate speech mean that the person speaking feels hate, or wants to convince someone else to hate, or wants to make someone feel hated in response to the speech? Generally, ‘hate speech’ refers to a message that vilifies a person or group of people, because they belong to a group or share an identity of some kind. Legal definitions of hate speech refer to various kinds of groups, defined by religion, race, or ethnicity; others add or omit disability, sexual orientation, gender, or even philosophy of life (Norwegian Penal Code, section 135a). Under all of those definitions, then, “I hate you” is not hate speech. In practice, the boundaries of hate speech are drawn by prevailing social norms and individual and collective interpretation, so that each person has an idea about what hate speech is, but one person’s notion of hate speech rarely matches another’s. Focusing instead on the narrower category of dangerous speech makes it easier to achieve consensus, and to respond effectively. For some examples of debate on what constitutes ‘hate speech’ see this article on the French newspaper Charlie Hebdo’s cartoons and text, and this New York Times report on students who felt that “Vote Trump 2016,” when written in chalk on their campus’ steps, was hate speech. And here is a thoughtful discussion on hate speech by the British writer Kenan Malik, who argues that hate speech should be morally condemned but not criminalized.
How is dangerous speech different from hate speech?
Hate speech is offensive, painful, and even threatening, but it does not, very often, inspire violence by those who are exposed to it. Conversely, dangerous speech isn’t always hateful. It often instills fear, which can be at least as powerful as hatred, in inspiring violence. So the two categories overlap only partly, as this diagram illustrates. Another distinction is this: the most familiar way in which hate speech harms is directly, by hurting the feelings, self-respect, or dignity of people it purports to describe, when they are exposed to it. By contrast dangerous speech does much of its damage indirectly, by persuading one group of people to fear, hate - and eventually to condone violence against - another group. Hate speech can also harm indirectly, by persuading one group of people to hate another group - here the categories overlap as illustrated.
Does dangerous speech include extremist or terrorist speech?
Yes, since dangerous speech can inspire people to commit or condone violence, and is often intended to do that. Extremists and terrorists seek to convince people that violence is not only acceptable, but virtuous and necessary - just as dangerous speech does. Countering violent extremism, or CVE, is a field with its own specialized scholars and practitioners. The Dangerous Speech Project is often in contact with them, but we are not expert in that field ourselves. We encourage research on the application of the dangerous speech ideas to cases of extremist and terrorist speech.
Does a message have to call for violence in order to constitute dangerous speech?
No. Early in this project, we defined dangerous speech very narrowly, to include only speech that was understood by the audience as a call to violence. We’ve since removed that limitation for several reasons: direct calls for violence usually come only after the ground has been prepared by less explicit speech, so focusing only on explicit incitement would often mean waiting too long to prevent violence. Second, when a message clearly calls for violence, its dangerousness is obvious; there’s no need for new terminology or analysis. Finally, we have realized that dangerousness can’t be correctly understood as a toggle or light switch, on or off, dangerous or not. Instead it falls on a spectrum: for example, Kenya’s Umati project, the first systematic effort to collect and study a country’s dangerous speech, sorted speech into three categories: offensive, moderately dangerous, and very dangerous.
What are some examples of dangerous speech?
- Before the national elections of 2007, Kenyan political leaders described other ethnic groups in terms that made their own followers despise and fear them. For example, they referred to other Kenyans as madoadoa (‘spots’ or ‘stains’) and kwekwe (‘weeds’).
- Influential Buddhist monks in Myanmar vigorously spread violent anti-Muslim speech both in person and online. Muslims in Myanmar have historically been the target of persecution and violence.
- After the American Civil War, whites described African-Americans as “noxious insects” and sexual threats to the purity of white women and girls. According to historian Leon Litwack, lynching participants saw themselves as “pest control” and fighters of a virus that would harm their communities.
Who is affected by dangerous speech?
Which kinds of groups can be targets of dangerous speech?
Any kind of group, including but not limited to groups defined by ethnicity, race, religion, gender, disability, and sexual orientation, can be the subject and target of dangerous speech. Unfortunately, the rhetorical techniques of dangerous speech are used against many types of groups, so we do not exclude any.
Can individuals be targets of dangerous speech?
Yes. The fact that speech targets an individual, as long as it makes reference to their membership in a group, does not disqualify it from being dangerous speech. As with any speech, to find out whether it is dangerous, one would conduct a dangerous speech analysis.
Do dangerous speech ideas apply to speech in the United States?
Yes. The United States is not currently at risk of genocide or other mass atrocities (though hundreds of Americans are killed in mass shootings each year, which happen every day on average), dangerous speech exists at one end of a spectrum in which each kind of speech, when it proliferates, can make the next, more inflammatory category of speech more acceptable and therefore more likely to become influential. In this way discourse norms (what people believe it is socially acceptable to say) can shift gradually toward speech that is more and more dangerous, increasing the risk of violence. This is a spectrum that we are continuing to study.
Do dangerous speech ideas apply to an attack by an individual (a “lone wolf”) or only to violence committed by groups?
They can apply to both. Some lone wolf attackers report that they were inspired to commit violence by the speech of others, such as Anders Breivik, the Norwegian who in 2011 massacred 77 people, including 69 children at a summer camp, and Dylann Roof, who murdered 9 African-American churchgoers in South Carolina in 2015. A lone attacker may be seen as a person who thinks and makes decisions quite differently from other people, so that s/he could be inspired to violence by speech that would not affect someone else that way at all. On the other hand, a lone attacker might be the member of a group who is most susceptible (e.g. isolated, distressed, agitated) to dangerous speech and therefore the first person to act. In all cases, whether only an individual reacts or a group does, dangerous speech has a variable capacity to influence and galvanize individuals.
How to respond to dangerous speech
Are there steps that can be effective against dangerous speech?
Yes. There are at least two methods we believe can be effective. The first is to educate people about dangerous speech before it happens, so that they can recognize and resist it. Radio La Benevolencija, an Amsterdam-based organization that pioneered this kind of effort with fictional radio dramas in Rwanda, named it “inoculation.” We use the same term, recognizing that dangerous speech, like bacteria, cannot be eradicated, so it is much more effective to help people to resist it. We used this to inform our work in Kenya with the popular TV show Vioja Mahakamani, which produced and aired four episodes about Dangerous Speech prior to Kenya’s 2013 presidential election – the first held since mass violence broke out between ethnic groups during their 2007 election. The second is counterspeech – responding to dangerous speech in a way that undermines it. Anyone can do this, and (just like dangerous speech itself) it is especially effective when done by people who have influence over the relevant audience. For example, German Chancellor Angela Merkel and President Joachim Gauck have repeatedly responded to dangerous speech targeting refugees, Muslims, and immigrants by denouncing the speech and calling for unity and tolerance. In response to anti-refugee protests that drew thousands of audience members, Merkel said, “There is freedom of assembly in Germany but there is no place here for incitement and lies about people who come to us from other countries,” and, “Everyone needs to be careful that they are not taken advantage of by the people who organize such events.” Influence does not need to be political, however, and it can exist at any level - local, regional, national, international, etc. In 2011, when a pastor from Florida threatened to bring anti-Islam protesters to the Arab International Festival in Dearborn, Michigan, mayor Jack O’Reilly wrote an open letter to residents urging them to “be agents of peace and forgiveness” and “ignore [the protesters] and their empty words”, because “no positive result” could come from confrontation. Counterspeech can be a spontaneous reaction to dangerous speech or a component of an organized counter-messaging campaign. See the U.S. Holocaust Memorial Museum’s guide Defusing Hate: A Strategic Communication Guide to Counteract Dangerous Speech for more information.
Should dangerous speech be censored or banned?
We’re skeptical of censorship as a means of countering dangerous speech, especially as a sole method, for several reasons. First, it can fail or even backfire. It’s usually impossible to suppress speech entirely, and efforts to do so can only amplify its message. Removing it from social media sites, for example, can push it onto other platforms that act as ‘echo chambers’ where people repeat inflammatory messages to each other, making them seem more convincing and therefore more dangerous. A second reason for skepticism is that rules restricting speech are often interpreted too broadly, leading to the suppression of more speech than necessary. Prosecution for dangerous speech can backfire, by publicizing it and even making it - and its disseminators - more popular. For example, Geert Wilders, a far-right populist politician in the Netherlands, was prosecuted for incitement to discrimination, and then found guilty in December 2016, but during the trial the Party for Freedom, which he founded, only climbed in the polls. In national elections in March 2017, the party won so many seats that it became the second-largest one in the Dutch House of Representatives. Censorship can easily infringe, then, on freedom of expression, which must be protected as a basic human right and a vital component of a democratic society. In addition, people use freedom of speech to seek peaceful resolution of their grievances and without it, may resort to violence.
Is dangerous speech a crime?
No. It is not codified as such in any body of criminal law, nor do we advocate for that. Criminal law is not a very effective way of suppressing speech or its harmful effects, and laws against specific types of speech, such as hate speech are often misused to suppress dissent or journalism. In addition, prosecuting someone for speech (like censorship) can backfire, since it tends to publicize that speech and can even lead to increased public support for the defendant. For example Geert Wilders, the Dutch leader of the far right Party for Freedom (PVV), saw his party make dramatic gains in the polls even as he was being prosecuted - and was then convicted - in 2016 for anti-Muslim hate speech. Most countries have laws criminalizing forms of speech, such as incitement or hate speech, which overlap with dangerous speech. In other words, there are examples of dangerous speech that would be considered crimes in the countries in which they are made, but no country criminalizes dangerous speech as such.
How have the dangerous speech ideas been used?
Dangerous speech ideas have been put into practice to help prevent violence in a diverse array of countries. Such projects include the Umati Project’s monitoring of dangerous speech before and during Kenya’s 2013 election, the Panzagar project to counter dangerous speech targeting Muslims in Myanmar (Panzagar means “flower speech” which stands for a commitment not to use or tolerate hateful or violent speech), PeaceTech Lab’s efforts to identify, monitor, and counter dangerous speech in South Sudan, and the Centre for Information Technology and Development’s dangerous speech monitoring project in Nigeria. For more information on these efforts, visit our page on countries where dangerous speech ideas are being used in efforts to prevent violence. Our ideas have also been influential in the study of free speech and violence prevention. For example dangerous speech has been featured in studies on Ethiopia and Sri Lanka and Rachel Brown, director of the violence-prevention NGO Over Zero, wrote the 2016 Defusing Hate: A Strategic Communication Guide to Counteract Dangerous Speech, for the United States Holocaust Memorial Museum. Internet companies have also made use of dangerous speech ideas. The Dangerous Speech Project is part of Twitter’s Trust and Safety Council, and we have worked with other platforms to restrict speech as little as possible while still protecting users, e.g. from violence.
Can you cite cases of individual(s) whose dangerous speech led to genocide or other mass atrocities?
Yes, though we caution that one generally can’t say that any individual's speech actually caused mass atrocities, since it is often hard to know exactly why a person acted, and people always act for many reasons: if you ask someone why he or she did anything, even something trivial like choosing what to eat for lunch, there are usually many reasons. This is why our definition of dangerous speech focuses on increasing the risk of violence, in context, not on ‘causing’ violence. That said, some individuals have conceded that their own dangerous speech led to genocide and mass atrocities. One of those is Jean Kambanda, who was prime minister of Rwanda during the 1994 genocide. Kambanda pleaded guilty before the International Criminal Tribunal for Rwanda (ICTR), to crimes including direct and public incitement to genocide, based in part on his inflammatory statements over the airwaves of the notorious radio station Radio Télevision Libre des Milles Collines (RTLM), which broadcast such statements by many people. Kambanda had called that station "an indispensable weapon in the fight against the enemy." (After he was imprisoned for life, Kambanda attempted to withdraw his guilty plea, but the tribunal did not accept that.) Another person who pleaded guilty to incitement to genocide, before the same tribunal, is Georges Ruggiu, a Belgian who was a broadcaster at RTLM.
Is there evidence that intervening where dangerous speech occurs is an effective way to thwart mass violence?
It’s hard to measure the effect of an intervention to prevent mass violence, since one can't replay to find out what would have happened without the preventive effort. However we have observed cases in which there was evidently a serious risk of violence which, after intervention, didn't happen. One example is Kenya before and after the presidential elections of 2013. Kenya saw a terrible eruption of violence right after a national vote in 2007, and in the months before the next national elections in 2013, many of the same risk factors were still present and Kenyans feared more violence. A stunning variety of individuals and organizations spoke out to prevent violence in many ways - on billboards, as part of a comedy on TV, in a music video with police dancing in uniform, PSAs by soccer stars, peace graffiti, public meetings, etc. In the end, that election went forward with minimal violence.