In the past 16 weeks, more than 50 drivers have plowed into peaceful protesters all around the country. Armed militants shut down Michigan’s legislature. Unidentified law enforcement officers heaved demonstrators into unmarked vans. Security forces in Washington used low-flying helicopters to harass citizens decrying police brutality. Protesters and police alike have brutalized journalists. Ideologues from left and right have been accused of killing political opponents. Should Americans be worried about widespread violence?

Yes. Political violence in democracies often seems spontaneous: an angry mob launching a pogrom, a lone shooter assassinating a president. But in fact, the crisis has usually been building for years, and the risk factors are well known. The United States is now walking the last steps on that path.

Partisans who would never commit violence themselves are transforming from bystanders to apologists, making excuses for the “excesses” of their side while pointing fingers across the aisle. Particularly striking have been the inflammatory statements of Republican politicians, given the influence leaders’ words carry. Of course, they are simply mimicking President Trump, who is most responsible for setting the kindling aflame.

Political violence tends to strike in countries where it has happened before. It feeds on discrimination, social segregation and inequality — which provide reasons for grievance while making it hard for divided populations to understand each other. Polarization exacerbates these conditions while blocking societies from solving their problems.

All the ingredients are here: America’s political violence traces back to our Civil War, the causes of which were never really resolved. The Union won the war, but the Confederates prevailed in the peace. Attempting to undo nearly a century of segregation and discrimination with civil rights legislation in the mid-1960s ushered in the next national outbreak of violence. The wound of racism deepens America’s deep inequality and our political polarization.

* * *

The most deadly countries in the world, in fact, are not at war. They’re highly unequal, highly polarized democracies facing a cocktail of political, criminal and state violence. In these places, such as Colombia in the 1990s and early 2000s, political violence mixes with state repression, vigilantism and crime (which often gains community support by playing a political angle, exemplified by drug kingpin Pablo Escobar’s use of charity to ingratiate himself with the poor). South Africa’s brutal police, vigilante gangs and xenophobic pogroms may seem far from our society — but the pattern is similar to America’s problems with police brutality, white-supremacist vigilantes, and shootings in recent years at an immigrant-friendly Walmart in El Paso, a Pittsburgh synagogue and a Black church in Charleston, S.C.

Similar risk factors also exist in countries that aren’t slipping into violence. For instance, Spain fought a civil war in the 20th century and battled a domestic terrorist campaign into the 21st. It is one of the most unequal countries in Europe. Since 2008, it has faced political polarization so intense that its two-party system has split into four. Catalonia’s secessionist movement resulted in nine regional politicians receiving prison sentences for sedition just a year ago. Yet partisans are not killing one another in the streets.

That’s because countries can be resilient, too. Strong democratic institutions, such as trusted courts and functional legislatures, can keep violence at bay by enabling populations to resolve differences. Political parties that are shaped by mass membership and ideological agreement, rather than serving as vehicles for individuals, also make violence less likely.

Even more important is a healthy society. Norms against violence are crucial. Crosscutting identities — such as religious affiliations or strong ties to a location — help people come together around identities that are not polarized. Social trust is a nation’s immune system. When it weakens, countries are less able to fend off or recover from all manner of ills.

Consider industrializing England in the early 1800s, 150 years after its civil war. Extraordinary inequality and exclusion existed alongside a dysfunctional, corrupt electoral system. A failing economy, tariffs and taxes impoverished the poor. Luddites smashed industrial machines. In 1819, aristocrats were so frightened by working-class demands for the vote that the cavalry charged a peaceful political rally of 60,000 people, murdering nearly a dozen protesters. The government temporarily suspended rights to assembly and speech, instigating assassination plots against multiple ministers.

Yet Parliament, political parties and the courts were strong. The government worked to end slavery and enable unions; politicians didn’t want violence. While their attempts were highly imperfect, they tried to steer the state into a new economic epoch as best they could. England exited this rot into an era of bustling Victorian do-gooders rather than armed fighters in the streets.

* * *

In the United States, however, resilience has been weakening, even as our risk factors increase. We are among the world’s most polarized countries. Racial, ideological, religious and geographic identities reinforce one another, exacerbating partisan division rather than offering alternative connections. Trust in the government is at a near-historic low of 17 percent, and 79 percent of Americans think there’s too little trust between citizens. Meanwhile, institutional guardrails are eroding. Gridlock and decline define Congress. Opinions about the Supreme Court are increasingly driven by partisanship. The Republican Party has set aside professed values to rally behind a personality.

As resilience fades and risks rise, society begins to break. People lose faith in established parties and processes. They begin to look to outsiders who promise to “fix it.” Politicians who see a path to power by enabling violence pose the greatest danger. They test the system to gauge costs and benefits. If they are shut out of mainstream political parties, as the Brits, Finns and Belgians did to homegrown fascist leaders in the 1930s and 1940s, societies can move on peacefully. Republican voters similarly succeeded in preventing nativist Pat Buchanan from running on their ticket in 2000. But by 2016, the GOP, weakened by the tea party’s rise, couldn’t stop Donald Trump from taking over.

When opportunistic politicians get a foothold, they break institutional guardrails with astonishing speed. They undermine professional norms, as with Trump’s refusal to set aside his personal business interests. They politicize institutions of government, often demanding personal loyalty — as Trump has done with independent inspectors general, the Justice Department, the post office and countless senior appointments. Meanwhile, they accelerate political polarization and gin up violent sentiment.

An early warning sign is language that casts enemies as subhuman and threatening. The causal relationship between such language and violence is one reason scholars have been so alarmed at Trump’s consistent amalgamation of immigrants, people of color and political opponents as nonhuman and menacing.

Dehumanizing language reduces inhibitions to violence. Framing vilified populations as threatening enables perpetrators to justify aggression as self-protection. That’s what happened with Bosnian leader Radovan Karadzic, who compared Muslims to animals as he rose to power in the late 1980s and early 1990s. Repetition of the canard that Muslims were a demographic threat to Serbian identity helped Serbs rationalize murder to “defend” their way of life against neighbors whose lives were remarkably similar. Consider Trump’s 2016 Republican convention speech, in which he claimed that “nearly 180,000 illegal immigrants with criminal records, ordered deported from our country, are tonight roaming free to threaten peaceful citizens,” followed by countless repetitions of the theme that “these aren’t people — these are animals,” blurring distinctions between immigrants in general and gang members. The result has been a steep rise in violent hate crimes against Latinos, according to FBI statistics.

A country’s final line of defense is a professional, nonpolitical security service that refuses to use force against civilians. Even if the tinder has been laid and an institution-destroying politician sends sparks flying, the conflagration rarely catches until state violence is turned on peaceful opposition.

The United States has one of the world’s most professional militaries. But we also have 18,000 local police forces and myriad state and federal security agencies with different norms and levels of professionalization. Since the beginning of his presidency, Trump has recognized and applauded law enforcement agencies that felt underappreciated, such as U.S. Immigration and Customs Enforcement and local sheriffs. Using incendiary language and pardons for lawbreaking, he has encouraged violence against civilians, as when he told a group of law enforcement officers, “Don’t be too nice” to suspects. Many officers welcome the freer rein.

Decades of research on violence, ranging from 1960s protests in Germany and Italy to terrorist movements in North Africa today, has found that government repression is the turning point that shifts peaceful movements toward violence. Organizations trying to operate within the system look weak and foolish, while violent splinter groups swell with new recruits from the fearful and angry. Protesters and the state feed each other’s violence in a vicious spiral. Data shows that this is precisely what happened in Portland, Ore., where violence escalated after the deployment of federal security forces this summer.

Soon, polarized populations begin rationalizing, downplaying or even denying the violence from their side. Moderates begin to silence themselves, fearing for their jobs, social standing or lives. Even partisans once seen as extreme begin to face similar consequences if they try to stand against dehumanization, intimidation or violence. Why speak when David Shor, a liberal former Obama campaigner, can be fired for tweeting a study about the negative political effects of violent protest tactics? Even partisans once seen as extreme begin to face similar consequences, as when the evangelical conservative writer David French was attacked from the right for defending civility.

This is where America is now.

* * *

So what might our future hold? Civil war seems far-fetched; U.S. institutions are weakening, but they remain stronger than in many countries. Despite pervasive “law and order” rhetoric, overall violence is actually quite low: At just under 5 homicides per 100,000 people, today’s murder rate is about where it was in the late 1950s and early 1960s, according to the FBI’s statistics.

What we may be in for, though, is a return to the 1960s and 1970s. That might sound innocuous — except that 13,000 National Guardsmen were deployed to Los Angeles after 34 people were killed in the 1965 riots, the first of hundreds of protests and more than a dozen violent riots over the next three years that killed over 100 and left urban centers gutted for four decades. A president and presidential candidate, John F. and Bobby Kennedy, were assassinated, as were Malcolm X and Martin Luther King Jr. In 1972, another presidential candidate, George Wallace, narrowly survived an assassination attempt. A 1968 Gallup poll found that 81 percent of the country agreed that “law and order has broken down.” A University of Michigan researcher found in 1971 that nearly one-third of men thought that “police beating students” wasn’t violence at all — and even more felt the same about “police shooting looters.” Throughout the 1970s, America faced nearly 1,500 terrorist attacks, nearly all domestic in origin. The murder rate crept up steadily, nearly doubling from the early 1960s through the late 1970s.

To avoid this fate again, we must stop casting blame. Finger-pointing is seductive; it feels morally necessary and even essential to identifying the “real” cause of the problem. But at this point, partisans can’t agree on the real cause. The only way out is to sidestep worn arguments and begin rebuilding from points of agreement — as Colombia did when millions voted in favor of constitutional change in 1990, even if they couldn’t find common ground on whether left-wing rebels or right-wing paramilitaries were more to blame for the crisis.

America’s democracy is ailing, and its immune system is on life support. We are the only industrialized democracy with decreasing life expectancy. Deaths in our streets are augmented by those from covid-19 and deaths of despair. These ailments grow from a broken social contract, which we cannot repair from inside polarized bunkers. Instead, we must find some way to step back from the brink.

Twitter: @RachelKleinfeld