On Sept. 21, Liberal Currents, a left-leaning magazine, published a poll of 683 Democratic voters conducted by the Center for Strategic Politics, a polling group that calls itself part of the progressive community. Its author, Chance Phillips, summed up one result with relief: “Thankfully, on net Democrats are not pro political violence.”
He was describing the only scenario that polled underwater. Asked how they would feel if they woke up to news that President Donald Trump had been assassinated at a rally, 31% of respondents said positive and 43% negative.
A few lines up the same chart sits a different headline: “To ‘safeguard the Constitution,’ a group of U.S. military leaders have removed President Donald Trump from power and assumed control of the United States.” Fifty-nine percent said they would feel positive about that one; 43% said very positive. The pollster’s own announcement did not hedge: of the five routes it tested, it wrote, “a plurality of Dems support four of them, including a military coup.”
TRUMP DIDN’T ‘SETTLE FOR LESS’ ON GREENLAND. HE GOT EXACTLY WHAT HE WANTED
More respondents reacted positively to a military coup than to the president dying peacefully in his sleep.
Measured by outcome, the coup is the more radical scenario. An assassination kills a president; the Constitution survives it, and the vice president takes the oath. In the headline respondents read, the generals hand power to no one. They “assumed control of the United States.” That is not impeachment with tanks. It is military rule. What made it palatable was the costume. One scenario is a man with a rifle. The other is an institution with a mission statement — uniformed, collective, and invoking the document it is suspending.
Coups rarely announce themselves as coups. They arrive as councils and caretakers, with the constitution in the first sentence of the communiqué. In counterintelligence, one of the first lessons is that people rarely consent to violence. They consent to violence that arrives with paperwork.
The poll’s wording supplied the paperwork. Every American officer swears to “support and defend the Constitution of the United States against all enemies, foreign and domestic.” The headline borrows that oath and turns it into a license to remove the elected commander in chief.
The crosstabs make the point sharper. The coup did not poll best at the party’s radical edge. Among self-described socialists and progressives, 56% reacted positively. Among self-described liberals, 66% did, and 53% of liberals said very positive. Socialists and progressives were more at ease with the rifle: among them, assassination broke even, 39% positive to 39% negative. Each wing picked its costume. Respondents who scored above the median on the poll’s social-trust questions — about trusting government, the press, neighbors, and strangers — were more positive about the coup (65%) than those below it (55%).
The uniform and the word “Constitution” did what such props are for. They made an unconstitutional act feel legitimate to the people who trust institutions most.
The coup number does not stand alone. Majorities want a future Democratic administration to prosecute Trump officials over immigration enforcement, the boat strikes, and the Iran war; 69% want the next Democratic president to expand the Supreme Court. Phillips’s article ends by saying that what unites the base is a demand to “impose consequences for the rampant lawlessness” of the last two years. A few sections earlier, it reports that the base would be happy with a military coup.
Now the other column.
This is not a probability sample. It was drawn from text messages and online panels, and the stated margin of error of 3.75 points is hard to interpret for that kind of sample; subgroups are noisier still. Conservative commentators from Gov. Ron DeSantis (R-FL) to Jonathan Turley have taken the pollster at its word and called the result support for a coup. The question measured something narrower: an anticipated feeling about a headline, not a willingness to act or to authorize anything. The constitutional routes scored far higher: 86% positive for impeachment and removal, 84% for the 25th Amendment.
Nor is this a Democratic disease. In a UC Davis survey of more than 8,000 adults last year, 52% of MAGA Republicans said political violence was usually or always justified to advance at least one of 20 political goals, compared with 32% of strong Democrats. Between that survey’s 2024 and 2025 waves, several such measures fell among MAGA Republicans and rose among strong Democrats. The appetite for shortcuts seems to follow power out the door.
That is why the rhetoric of the party out of power matters. The week the poll ran, Kamala Harris told an NAACP audience in Detroit that Republicans are “cheating” through redistricting, relocated polling places, and purged voter rolls. Her critics said she was defending non-citizen voting; she said no such thing. But a leader who tells voters the machinery is rigged against them is teaching the lesson Republicans taught their own voters after 2020: that when the other side runs the machinery, the result is suspect.
Americans saw one ending of that lesson at the Capitol on Jan. 6, 2021. I have seen another. In the winter of 1991–92, I was a university student in Tbilisi when Tengiz Kitovani’s forces shelled the Parliament building where Georgia’s first democratically elected president, Zviad Gamsakhurdia, was barricaded. Two weeks of fighting left about a hundred dead and the center of Rustaveli Avenue in ruins. One afternoon, a stray round found one of my classmates. He survived. He should never have bled.
The men who drove Gamsakhurdia from power formed a Military Council. They said they were rescuing Georgia from an authoritarian. The headline looked like a rescue. The hand at their elbow was Moscow’s. In December 1992, Gen. Sufian Beppayev, then deputy commander of the Transcaucasus Military District, Moscow’s garrison headquartered in Tbilisi, wrote in Moskovskiye Novosti that his units had helped the rebels, and that without that help Gamsakhurdia’s side would have won. I laid out that record in these pages in July.
What followed did not look like one. Gamsakhurdia’s supporters fought on in western Georgia through 1993. By May 1992, Kitovani, one of the council’s leaders, was defense minister; in August, he led Georgian troops into Abkhazia. A year later, that war was lost, and the country’s conflicts had driven some 250,000 people from their homes. That fall, with Russian tanks helping against the rebels, Tbilisi agreed to join the Russian-led Commonwealth of Independent States it had refused to enter in 1991.
Russian “peacekeepers” took up posts in Abkhazia and in South Ossetia — Samachablo, as Georgians call it — and were still there when Russian armor rolled through in August 2008. I was there too, in uniform. Kitovani, by then, was living in Moscow.
The poll asked Democrats how they would feel about a headline. It did not ask about the morning after, or the decade after. The headline is the only part of a coup that ever looks clean.
TO THE TRUMPS, IT’S A TOWER. TO GEORGIA’S OLIGARCHS, IT’S A STAMP OF APPROVAL
Phillips is right that most Democrats in his sample do not want Trump shot. That was never the test. The test is whether the word “Constitution” can still be borrowed by the people suspending it.
In this poll, it could.
Emzari Gelashvili is a former senior Georgian counterintelligence officer and a former member of the Parliament of Georgia. He is a veteran of the 2008 Russia-Georgia war.