the edit, vol. 54
the holdouts
On September 11, Representative Mario Diaz-Balart of Florida told reporters he was "very concerned" about the administration's handling of the government shutdown. He did not vote against the continuing resolution, he did not hold a press conference — he expressed concern and returned to his office.
This is the dominant mode of Republican dissent in 2026: the carefully worded statement that stops one step short of consequence. It is worth examining what happens in the cases where the step was actually taken — and what happened to the people who took it.
Three weeks before the midterms, the Republican Party is navigating a split that is visible in the polling, audible in the fundraising calls, and legible in the voting records of a small but growing number of members who have concluded that the political cost of silence now exceeds the political cost of speech. This piece is about what those members did, what happened to them, and what the contrast between their experience and the experience of those who stayed silent tells us about where the party — and the country — actually is.
The Ones Who Spoke
Byron Donalds, the Florida congressman and former Trump surrogate who spent two years as one of the administration's most reliable advocates, gave a floor speech in late August that drew more national attention than anything else he had said in four years in office. He did not attack the president. He did not call for investigations. He said, in measured language, that the deployment of National Guard units at polling places would be "a step too far" and that he could not support it "as a matter of principle." The speech lasted six minutes. His office received more than four thousand calls in the following forty-eight hours. Roughly two-thirds were supportive. The other third were not.
Ashley Hinson of Iowa, in a competitive district that Trump carried by four points in 2024, voted against the administration's proposed budget framework in committee — one of three Republicans to do so — and issued a statement citing the impact on rural healthcare funding in her district. She did not hold a press conference. She did not appear on cable news. She voted no, put out a statement, and let the record speak. Her campaign fundraising in the following reporting period was the strongest of her congressional career.
Maria Salazar of Florida, who has broken with the administration on immigration enforcement language twice this year, framed both breaks in explicitly moral terms — invoking her family's exile from Cuba and arguing that the rhetoric being used to describe certain immigrant communities crossed a line she was not willing to stand behind. She was primaried in 2024 and survived. She is being primaried again. She has not changed her position.
These three are not a movement. They are data points. But the pattern across all three is the same: a specific, limited break on a specific issue, framed in terms of principle rather than politics, followed by a political consequence that was either survivable or — in Hinson's case — actively beneficial. None of them called for impeachment. None of them appeared on MSNBC. None of them left the party. They drew a line, stated why, and remained.
The Ones Who Didn't
For every member who spoke, there are a dozen who considered it and did not. The evidence for this is not circumstantial — it is in the private conversations that have made their way into reporting at Politico, the Atlantic, and the New York Times over the past six months, in which Republican members describe, with varying degrees of candor, a calculation that comes down to a single variable: the primary.
The primary threat is real. It has taken out members before, and the members who survive primaries to lose general elections are not uncommon in this cycle. But the calculation is being made, in many cases, on the basis of 2022 data in a 2026 environment. Trump's net approval among Republican primary voters has dropped fourteen points since January. The share of Republican primary voters who describe themselves as "strong Trump supporters" has fallen from 62 percent to 47 percent in the same period. The primary electorate that punished dissent in 2022 and 2024 is not the same primary electorate that will vote in November.
The members who stayed silent made a bet on an electorate that may have already shifted beneath them. The members who spoke, in the limited and specific way that Donalds and Hinson and Salazar spoke, appear to have read the room more accurately. Whether that reading turns out to be correct will be answered in twenty-three days.
What the Silence Costs
There is a version of this story that treats silence as a neutral act — as the absence of a choice rather than a choice itself. That version is not supported by the evidence.
When a member of Congress declines to comment on the deployment of military personnel at civilian polling places, they are not withholding judgment. They are signaling to their constituents, to their colleagues, and to the administration that the action is within the range of things they are willing to accept. When a member expresses concern in private and votes with the leadership in public, the public vote is the one that counts. The concern does not appear in the record. The vote does.
The silence has a second cost that is harder to measure but no less real. Republican members who have spent this year watching the administration test the floors that Vol. 36 described — the press ban, the military deployment threat, the financial disclosure — and have chosen not to respond have trained their constituents to expect nothing from them on those questions. An electorate that expects nothing from its representatives on fundamental questions of democratic governance is an electorate that has already begun to accept that those questions are settled. The silence is not neutral. It is instructive.
This is what makes the holdouts significant beyond their individual votes. They are demonstrating, in real time, that the calculation can be made differently — that a Republican member of Congress can draw a specific, principled line on a specific issue and remain in office. The question the next three weeks will begin to answer is whether enough of their colleagues were watching.
What Comes Next
The midterm elections on November 3 will be read, by every major news organization, as a referendum on the Trump administration. That framing is not wrong, but it is incomplete. The elections are also a referendum on the Republican members who chose silence — on whether the bet they made on voter passivity and primary fear was correct.
If the wave materializes — if the historical pattern at this approval rating holds and Democrats make significant gains in the House — the members who broke will have been right, and the members who stayed silent will have to explain why they made the choice they made in a chamber they no longer control.
If the wave does not materialize, the conclusion that most commentators will draw is that dissent was punished and silence was rewarded. That conclusion will be available to the next administration that wants to test the floors. It will be cited. It will be acted on.
The holdouts took a risk. The ones who stayed silent took a different risk. In twenty-three days, we will have a clearer picture of which calculation was right — and of what American voters decided to reward when they had the chance to decide.
