¹School of Public Health and Community Medicine, Hebrew University Hadassah, Ein Kerem Campus, Jerusalem, Israel

Correspondence: Elihu D. Richter, School of Public Health and Community Medicine, Hebrew University Hadassah, Ein Kerem Campus, Jerusalem, Israel. E-mail: elihur@ekmd.huji.ac.il

"Sticks and stones will break my bones, but names will never hurt me", goes the childhood taunt.

Not so, according to Paul Rusesabagina, the hero of the movie Hotel Rwanda, concerning the Rwandan genocide. The big question, of course, is why did it all happen? Rusesabagina, after telling us "it happened because of ethnic hatred" directed against the Tutsis, descendents of the ruling classes of Rwanda, then asks: "What caused this to happen? Very simple. Words".

This is the punch line from Ruseabagina's book, No Ordinary Man (1), that inspired the movie, for those of us concerned with the prediction and prevention of genocide – the intentional destruction of a population in whole or in part. It recalls the statement of Abraham Joshua Heschel, the Jewish theologian who escaped from Europe, who wrote that it was not machinery, but words that killed at Auschwitz. Genocide and mass atrocities have been the ranking cause of violent deaths in the 20th century. The toll is estimated to be some 250 million, of whom 70 million have died since World War II. Ruseabagina's book is essential reading for everyone in public health and social medicine.

Rusesabagina tells the story of his relatively happy childhood in the countryside of Rwanda, the land of the thousand hills: brewing of banana beer, his father's role in shaping his personality, and his excitement about his work as a hotel manager for Hotel Milles Collines. He gives us a retrospective look into how the Hutu–Tutsi divide was rigidified by European rulers.

When his childhood Hutu friend, Gerard, was expelled from school in 1973 and had to flee during an outbreak of Tutsi violence against the Hutus, Rusesabagina had his first personal encounter with the ethnic hatreds embedded in Rwandan political life. He reminds us that the Hutu–Tutsi distinction was based on flawed theories of race and genetics of the late 19th and early 20th centuries. Belgian scientists with tape measures determined that a typical Tutsi nose was at least 2.5 mm longer than a Hutu nose. Similar flawed theories gave us eugenics, and later, Nazi medicine and "racial hygiene".

Rusesabagina recounts his work in managing to turn the five story luxury Hotel Milles Collines into a sanctuary for Rwandans escaping the genocide perpetrated by the Hutu elites on Tutsis and Hutu moderates in 1994. Clearly, he was heroic. During a 3–4-month period, death squads butchered some 800,000 people – old, young, children, infants, and pregnant women. During fewer than 100 days, using machetes, they killed at a rate of 8,000 lives per day or five lives per minute – a pace of killing comparable to that of the Nazi gas chambers during the peak of the Holocaust in World War II.

Although not the largest of the genocides of the 20th century, Rusesabagina reminds us that it was the fastest and most efficient. He protected what he estimates to be 4 hours worth of slaughter, some 1,200 persons – adults, children, mostly Tutsis – from certain death. Those on the outside had to watch as the bodies of their loved ones were cast aside like garbage before they too were hacked to death at roadblocks set up around Kigali and rural areas. During the genocide, identity cards with ethnic classification as Tutsi could become death certificates, although many Hutu moderates were butchered as well.

None of the refugees in his hotel, protected only by two or three UN guards and separated from the outside world by a symbolic fragile wire fence, was killed, beaten, taken away, or made to disappear. So we learn right away that symbols can be protective.

Rusesabagina was a tough operator who exploited his experience, street smarts, supplies of liquor, some cash, and connections with nearly everyone in the Rwandan establishment, including some of the big shots in the Interhambawe – the organizers of the genocide. And, via his concealed fax machine, he was able to connect to his superiors in the Sabena hotel chain, getting their support in protecting all those who sought refuge in the hotel. He manipulated, cajoled, bargained, bluffed, bribed, and negotiated. His insights on how to negotiate with genocidaires from a position of physical powerlessness are revealing. Even perpetrators of genocide, we discover, calculate their interests, and, in their moments of weakness, are sensitive to outside world pressures.

As with the Armenian genocide, but unlike the Holocaust, the media brought the gory details of this low-tech genocide to the attention of the outside world in virtual real time. Husbands killed wives and lifelong friends killed neighbors, and, as others have reported, mothers with babies in papooses hacked to death other mothers with babies in papooses.

In Bosnia, while on a 2007 visit to Srebrenica, where the Serbian Army machine-gunned some 8,000 victims, mostly men, over a several day period in July 1995, I heard the story of a teacher, a woman named Bella, who herself machine-gunned her Bosniak students in the classroom. In both Rwanda and Bosnia, and more recently in Darfur, the movers and shakers of Geneva, Paris, London, Berlin, Stockholm, New York, and Washington did nothing effective to stop the genocide.

There was nothing inevitable about the Rwandan genocide. It was planned by a ruthless leadership obsessed with notions of racial purity and afraid of losing its grip on power. The trigger event was the shooting down of the aircraft carrying Rwanda's President and that of Burundi. Rusesabagina describes the 20-year rule of a corrupt dictatorial leadership, whose failures led to increases in discrimination and socio-economic gaps along with poverty. Over these years, there was a rapidly growing population, overcultivation of the countryside, depletion of soil quality, and worsening population-resource ratios. And there were the memories of old conflicts. Rwandans classified as Tutsis had been favored by the Belgians. Were these so-called Malthusian pressures and the conflicts a sufficient explanation, thus rigidly deterministic in Rwanda and other genocides? The evidence for this deterministic explanation, argued most forcefully by Jared Diamond (2), is not persuasive, for reasons presented elsewhere (3).

Bystander indifference and inaction is the story of the Rwandan genocide from the outside. Rusesabagina passes harsh judgments on the dismal performance of all the European powers, notably the French and Belgian governments. The first was obsessed with protecting a Francophone regime; the second with establishing a colonial legacy, ruling by divide and conquer tactics. The US government under President Clinton, and notably the State Department, had been traumatized by the fiasco in Somalia and was preoccupied with Yugoslavia, and the US media was preoccupied with the OJ Simpson trial.

Rusesabagina is sarcastic concerning the dismal performance of the Roman Catholic Church at both the local and international level, and critical of Kofi Annan, who famously rejected the faxed request of General Romeo Dallaire, commander of the UN Peace Keeping Forces, for authorization to use force to stop the Interhamabawe as its gangs began their frenzy of mass butchery. The author of No Ordinary Man does not understand why Dallaire did not defy his superiors' orders. He remains skeptical about the accomplishments of Paul Kagame, now the President of Rwanda, the professional soldier who led the Tutsi conquest of the Hutu genocidaires. Perhaps Hutus fleeing Rwanda triggered wars in the Congo Republic in which millions have perished. Rusesabagina, however, praises Kagame's policy of setting Rwanda on a course of Truth and Reconciliation, and of banning ethnic status listing on identity cards – a revolutionary step in a country in which perpetrators and victims have to live with each other. But at the time of writing, Canada still requires Rwandans to fill out a visa request in which they have to tick off Hutu or Tutsi.

Hate language and incitements broadcast by Radio TV Libre Milles Collines – a private corporation set up by a consortium of higher-ups in the Rwandan government to bypass the prohibitions against governmental incitement of the Arusha accords – was at the center of the Rwandan genocide. Eight months elapsed between the onset of RTVLMC's incitement in August 1993 – at first, coy, and then ever more explicit – and the mass killing. Radio messages to "kill the cockroaches" ("nyenzi" in Rwandan), and "do your job" preceded the outbreak of the mass butchery in April 1994. RTVLMC began referring to individual Tutsi leaders, and then groups of Tutsis, as "nyenzi" in increasingly provocative language.

The use of terms such as "filth," "cancer," "parasites," and "microbes" in the media was a relatively late and specific warning sign of genocidal intent, and recalls Turkish propaganda during the Armenian genocide and, of course, Nazi propaganda directed against Jews during the Holocaust. The Nazis used the techniques of mass media perfected by Bernays for marketing tobacco, so well described by Devra Davis in her book, the Secret War against Cancer (4), to spread hate throughout Europe and the Mideast.

The Rwandan genocide reminds us that state-sponsored hate language and incitement are among the most toxic exposures of our times. Such hate language and incitement is the software that leads to the use of the hardware – machetes, guns, rockets, missiles, and WMD that kill, maim, and destroy. As with other toxic exposures, it is the children who are the most vulnerable and in whom the effects are most long lasting. Children exposed to hate language and incitement become programmed to act on its messages as they grow up, thereby ensuring intergenerational perpetuation of violence. It follows that to prevent its damaging effects to this and coming generations, we must apply public health models of surveillance to identify and ban the use of hate language and incitement.

Inciting of genocide is now a crime against humanity, as specified by Article 2(4) of the UN Genocide Convention and the Rome Statute of the International Criminal Code, Articles 6 and 25(3)(e). Journalists, radio stars, and educators who organized and broadcast the campaign of hate language and incitement from Radio TV Libre Milles Collines have been convicted by the International Criminal Tribunal-Rwanda of crimes against humanity for violating these clauses. New rulings, by the Canadian Supreme Court, uphold the conviction for the same offenses by Leon Mugasera, a high-ranking Hutu hard-line politician who fled to Canada. The Canadian court is clear that these laws against incitement can be used to convict potential perpetrators even if it cannot be proved that the incitement led directly to the crimes. Moreover, they can be prosecuted not only in their own countries, but also when visiting other countries in which their broadcasts have been heard (5).

The precedents established by the convictions of the Special Tribunal for Rwanda and the verdict of the Mugasera trial, are currently the basis of proposals for indicting the President of the Islamic Republic of Iran, Mahmoud Ahmadinejad, for his calls to wipe Israel off the map (6). He uses many of the same metaphors used by the Nazis. The latest of these inflammatory statements is that "Israel is a filthy corpse destined to disappear" (7). These proposals, urgent in and of themselves, are the template models for preventing future genocides by deterring their architects and perpetrators. As World War II taught us, the longer the latent period from the onset of state-sponsored incitement, the greater the costs associated with intervention. And as Rwanda taught us, delays on the order of months in stopping state-sponsored incitement translate into increased death tolls of the order of hundreds of thousands.

We now know, from studying the timelines of many genocides, that there are early warning signs, and that there are circumstantial, extrinsic predictors and intrinsic initiators, triggers, and tipping points (8). But as neat as all this sounds, we are left with the paradox of prevention in the real world: if there are costs associated with successful intervention, we only the count the former.

The proposal to criminalize and prosecute incitement by state authorities and their funded or protected surrogates is an example of applying the precautionary principle – environmental medicine's gift to genocide prevention (9). Policy towards genocide needs to move from interventions based on proof of intent after the event to actions to predict and prevent. The case for application of the precautionary principle to the prevention of genocide derives from the fact that the consequences of doing nothing – a false negative – are catastrophically greater than intervention triggered by early warning signs which turn out to be false positives.

Scholars have proposed odds ratios for certain circumstantial predictors and risk factors for genocide: past genocides, frequent internal wars and regime crises, authoritarian governments, exclusionary or racist ideologies, exclusion of political parties, a history of past warfare and defeats, vulnerable minorities with elite status, and low openness to external trade (10,11). Any of these may be warning signs calling for interventions that may employ a mix of political, economic, social, and military carrots and sticks.

And confronted with warning signs of impending genocide, continued movement toward genocide depends on outsider indifference. Gregory Stanton of GenocideWatch has described the eight stages of genocide, using a model that looks at it from the inside looking out, that is, the vantage point of the architects and perpetrators: classification, symbolization, dehumanization, organization, polarization, preparation, extermination, and denial (12). Denial of past genocide paves the way for the intergenerational perpetration of future mass atrocities.

One dismal example of intergenerational perpetration has been the continued persecution of Assyrian and Pontine Greek Christian minorities in the Middle East following denial of the Armenian Genocide. I and others have heard the story told by Professor Shevach Weiss, formerly Israel's ambassador to Poland. He recalls how, following the Nazi invasion of Poland, the general practitioner who lovingly took care of the children of his village, turned into a sadistic killer in a matter of weeks to months.

In the aftermath of the breakup of Yugoslavia, victims of Serbian genocide in Kosovo themselves then killed many belonging to Roma minorities, whom they described as "majupi", or lower than garbage (13).

Professor Israel Charny's study of 107 moviegoers entering Claude Lanzman's epic movie "Shoah" in 1990 found that 38% accepted or explained away the 1956 massacre of Arab men, women, and children coming home from work in the fields after curfew to Kfar Kassam, a village inside Israel, by Israeli soldiers. Those in whom acceptance of the killings was greatest were those considered to be weaker or lesser in status: youngest, oldest, females, less educated, immigrants, and Holocaust victims (14). If hate language and incitement has the potential for bringing out latent tendencies to genocide in all of us, Charny's research tells us something about the bystanders. The perpetrators are usually young males motivated by charismatic, xenophobic leaders using language such as "filth", "disease", and "cancer." But it is usually the elites who provide the ideology. Clearly some are predictably more vulnerable to becoming perpetrators and others to remaining bystanders.

Would the Rwandan genocide have been prevented had Dallaire been given approval to use his forces in January 1994, and, simultaneously, the UN, or NATO, had jammed Rwandan radio stations, and immediately indicted the Rwandan leadership for incitement to genocide? In retrospect, there were many missed opportunities for preventing the horrendous massacre (15). Similarly, timelines for death tolls in Darfur, where racist language was used to incite the Janjaweed, suggest that the tipping point occurred in September 2004, when the death toll had reached 40–50,000. Had the outside world imposed strong boycotts and no-flight zones, and warned the Sudanese leadership that it would be subject to prosecution for genocide, we probably would not be where we are today with 300,000 to 400,000 killed by the Janjaweed horsemen who plunger, rape, and murder. In Kenya, more recently, the post-election outbreaks of mass violence were not spontaneous, but the results of organized incitement by extremists on both sides. In Kenya, however, thanks to proactive outside intervention, mediated by a chastened Kofi Annan, both sides in the recent outbreak of mass violence drew back from a situation in which they were beginning to carry out genocidal attacks on each other. At the time of writing, in Burundi, there are chilling parallels to the situation in early Rwanda in 1994, and a compelling case for the outside world to apply the precautionary principle to put boots on the ground to protect members of the political opposition from mass killing.

What can we in public health and epidemiology do to prevent genocide?

First, epidemiologists should ensure that standards similar to those for rapid investigation of reports of communicable disease outbreaks are applied to reports of atrocity crimes and incitement – clearly the next role for WHO's Injury Prevention Programs. Previously, the UN's sloppy investigation of the genocide in Darfur had the effect of manufacturing doubt concerning the results of a far more rigorous investigation by the US State Department, a study that did find that there was a pattern of organized intent to destroy a population "in whole or in part".

Second, we in public health should be vigilant against the hijacking of medical metaphors to dehumanize – a highly specific early warning sign. Epidemiologists should lead the way in setting up an international surveillance network modeled after the systems for monitoring and reporting warning signs of epidemic diseases for monitoring hate language and incitement in state-sponsored media, textbooks, and places of worship. These systems need to be insulated from political pressures and use standardized diagnostic criteria that are defined in advance.

Third, we in public health should aim to broaden classic definitions of genocide to include ecocide – the destruction of life, reproductive capacity, and habitat of populations resulting from wanton or reckless industrial practices. Should ecocide be classified as a crime against humanity?

Fourth, we have a responsibility, as individuals, and as organized professionals, to not be passive bystanders, but to speak out publicly on genocidal threats. I take strong exception to the notion that speaking out on such threats, if based on a careful review of the evidence, is somehow a slide down the slippery slope of politicization. Not to speak out is to slide down that slippery slope. One lesson of the Holocaust: silence makes one a complicit bystander to genocide.

I wrote this commentary one day after my university commemorated the 93rd anniversary of the Armenian Genocide and on Holocaust Remembrance Day. Rusesabagina's book forces us to ask "How many more will perish from genocides in the 21st century before we stop tolerating the use of words to kill?".