President Trump speaks to media representatives on the South Lawn of the White House in Washington on Wednesday, May 23, 2018, en route to a roundtable discussion on illegal immigration and gang violence. (Jacquelyn Martin/AP)

On Monday, the White House issued a news release doubling down on President Trump’s insistence that undocumented immigrants, including members of transnational gangs like MS-13, “aren’t people. These are animals.”

In the past, Trump has referred to immigrants as rapists and criminals. Many have decried the president’s conflation of immigrants with MS-13 and animals. Some observers have drawn parallels between the president’s speech and Nazi descriptions of Jews as rats and vermin before the Holocaust and Hutu extremists referring to Tutsis as snakes and cockroaches to prepare for the Rwandan genocide.

Dehumanizing propaganda is a common precursor to genocide and other mass atrocities. But is there really a direct link between dehumanizing discourse and participation in violence?

Simply hearing others called ‘animals’ or ‘vermin’ doesn’t cause violence.

Rather, recent research suggests that promoting negative views of others can go only so far in motivating people to kill. In Rwanda, for example, Hutu militants issued calls on the radio to exterminate inyenzi, the Kinyarwanda word for “cockroach.” But some Hutus refused to kill, saved Tutsis, or shifted stances from killing to not killing neighbors. This suggests that the decision to commit murder and other violence was difficult for Hutu civilians. Dehumanizing propaganda alone didn’t persuade Hutus to suddenly turn on their Tutsi peers.

Moreover, research that looks at the timing of radio broadcasts in Rwanda and when massacres occurred finds no correlation between dehumanizing discourse on the radio and individuals’ participation in genocide. Further, the dehumanizing propaganda wasn’t widely received in the majority of places where Tutsis were killed. Only 10 percent of Rwandans owned radio transmitters in 1994, and broadcast range had little reach in rural areas while genocide was national. “Hate radio” can’t explain why violence occurred almost everywhere.

Maybe, though, Rwanda is unique. What can we learn from the Holocaust, that other classic case in which dehumanizing propaganda preceded genocide?

Here, too, evidence suggests that calling Jews rats and vermin did little to motivate ordinary Germans to kill. In fact, historical studies indicate that “ordinary men” in the Nazi killing machine often felt disgusted and disturbed by their actions. From Belarus to Poland and throughout Soviet territory, not all Germans who participated in violence against Jews saw them as a dehumanized mass that was easy to murder; many needed liquor, drugs and, eventually, the concentration camp system to separate them (mentally, emotionally and physically) from their terrible tasks.

Dehumanizing discourse doesn’t cause violence, but it helps prepare the way.

Language is not innocent. Dehumanizing propaganda helps to normalize extreme perspectives on how to address social problems. It grants legitimacy to those who do believe that certain others are inherently threatening, dangerous and ought to be eliminated from the community.

Following this, dehumanizing propaganda alters norms of what is and isn’t perceived as acceptable views or behavior. Even when people don’t believe what they hear on the radio or on TV, dehumanizing propaganda might make them hesitate more to speak out against it. By raising the costs of dissent, dehumanizing propaganda further perpetuates the idea that the threat described is real.

Finally, when some people do perpetrate extreme violence, dehumanizing propaganda can serve as post-hoc justification for those actions — enabling people with no history of violence to justify their killings to themselves as moral. Although research has found that Rwandans and Germans often felt horrified the first few times they murdered another human being, over time, they became used to it. Dehumanizing language helps this desensitization process.

So does it matter that the president conflated the violent crimes of MS-13 with the risk posed to the country by all immigrants?

Along with the research described above, scholarship suggests that how local leaders respond to dehumanizing propaganda has the power to either encourage or disrupt these narratives and their impact.

In particular, what local law enforcement, police unions and Immigration and Customs Enforcement (ICE) officials — authorities with the ability to enact violence against immigrant communities — say and do now is crucial. In areas where local authorities pushed back against extremist discourse in Rwanda, no one committed genocide. When Catholic leaders in France protested Nazi violence during the Holocaust, public opinion shifted from complacency to anger, helping to legitimize the resistance and trigger actions to save Jews.

Moreover, sociologist René Flores’s recent study looks specifically at Trump’s vilification of immigrants during the 2016 presidential election, finding that such statements did affect American attitudes toward immigrants but had a greater impact among individuals who already held restrictionist views. The study also suggests that the animosity prompted by dehumanizing language is momentary; propaganda needs to be repeated regularly or its effects dissipate.

This means that every time the Trump administration and associates use dehumanizing language, it’s an opportunity for others — especially people of high status — to object, saying, “This has gone far enough.”

By speaking out, they can profoundly shape public opinion and decrease the possibility of future violence.

Finally, whether or not the president’s language contributes to escalated violence, immigrant communities — especially Latinx ones — already face steady harassment, imprisonment and deportation. In our current system, families are divided, livelihoods are interrupted, and physical-integrity rights are regularly violated. Dehumanizing discourse matters, but violence does not require it.

Aliza Luft (@alizaluft) is an assistant professor in the department of sociology at UCLA.

Daniel Solomon (@dan_e_solo) will be a PhD candidate in government at Georgetown University beginning in the fall of 2018.