In 1649, Oliver Cromwell signed the death warrant of King Charles I. He was a religious radical and a military commander, a man who believed with absolute certainty that God had chosen his side. And, a handful of years later, Cromwell turned his cannons on his former allies - the Levellers, the Diggers, the Baptists - radicals who had fought beside him for a freer, more godly England. They were no longer “pure” enough. The English Revolution collapsed into factionalism and fear, and by 1660, the monarchy returned.
Revolutions often do this.
The purity spiral is older than Twitter. But digital networks give it new velocity and scale. And what worked as moral pressure in a tight-knit group collapses in the open field of internet politics. Every coalition is half cathedral, half trench warfare.
Purity tests can galvanize in small doses. But at scale, they can only destroy.
The Charm of Certainty
Purity tests feel good to those administering the tests. They create a sense of clarity, of rightness, of belonging. You know who your people are. You know who the enemy is. You know what counts as betrayal. It doesn’t even have to be formalized. Just a few agreed-upon positions, some shared shibboleths, a couple of banned terms, and you're in. You might get that hit of dopamine when someone you already dislike says the wrong thing and can be safely exiled. It simplifies the world, and simplicity is addictive.
There’s a reason Orwell made the final line of 1984 a surrender: "He loved Big Brother."
But the clarity offered by purity is fragile. The border between us and them quickly narrows. The field of acceptable opinion contracts. The number of acceptable allies shrinks. And you end up eating your own.
In the early 20th century, the German left was split between the SPD and the Communists. They agreed on 90% of things, but the 10% they disagreed on consumed them. In 1932, instead of forming a coalition against Hitler, the Communists helped topple the Social Democratic government in Prussia, believing the SPD were "social fascists."
By 1933, there were no leftists left to form a government.
Scale Is a Different Game
Purity tests work up to a point. Inside activist spaces or religious communities, they help enforce norms and prevent drift. That’s what they’re for. But they’re a terrible tool for larger organizing, because they assume a level of homogeneity that simply doesn't exist at population scale.
There’s a reason the Founding Fathers of the United States were lawyers, landowners, and merchants, not ideologues. Jefferson wanted an agrarian republic. Hamilton wanted banks and debt and cities. Madison was terrified of both mob rule and monarchy. They argued. They compromised. They built a working system - not by any means a pure one. That system survived because it allowed disagreement without fragmentation.
Contrast that with the French Revolution. The Jacobins began by purging monarchists. Then they purged Girondins. Then they purged their own. Robespierre, the architect of the Terror, died by the same guillotine he praised. Because once you've built a political project on the righteousness of purity, you're next.
Purity works in cults. It does not work in nations.
Internet Coalitions Are Not Friend Groups
The group chat is a special place. You can be mean. You can be sharp. You can hold each other accountable with a kind of “loving violence.” Everyone shares the same language and mostly the same priors. It's a bounded context. Norm enforcement works there because everyone opted in.
Twitter (or Threads, or Bluesky, or Reddit) is not that. It's a field full of overlapping publics. You can’t enforce purity at scale without turning discourse into a punishment machine. You stop building coalitions and start building inquisitions. And because scale requires strangers, and strangers always disagree about something, your coalition collapses under the weight of its own expectations.
The old idea of a "big tent" party depended on shared goals, not shared dogma. The civil rights movement included church leaders and Marxists, labor organizers and liberal lawyers. If they'd spent their energy on calling each other out for theological inconsistencies or insufficient anti-capitalism, nothing would have happened. But because they focused on clear, actionable shared outcomes, the coalition held.
Moral Narcissism as a Feature
Some purity spirals aren’t bugs. They are design features. If you’re trying to make a name for yourself in a crowded information economy, one way to stand out is to go harder than everyone else. The most brutal call-outs, the most aggressive standards, the most unyielding judgment - these win attention. And attention is currency.
It becomes a way of signaling: not only am I on the side of good, but I’m more committed to it than you. It’s a moral speedrun. And it works, right up until the movement starts losing actual power.
The historian Richard Hofstadter noted that American populist movements often eat themselves alive. The paranoid style, he argued, thrives on the idea of betrayal from within. When there’s no longer an external enemy that unifies the group, the energy turns inward. Who's insufficiently pure? Who's compromised? Who liked the wrong tweet?
The right does this too. Consider the current GOP, where voting to certify an election can end a career. The irony is that while the right often weaponizes purity as a loyalty test, the left is more likely to believe it.
Exit, Voice, Loyalty, and the Collapse of Disagreement
In Albert O. Hirschman’s classic framework, political actors faced with a failing organization can choose one of three paths: exit (leave), voice (speak up to fix it), or loyalty (stay and support it anyway). Purity culture tilts the balance hard away from voice. Disagreement looks like betrayal. Loyalty requires performance. Exit becomes the only option.
The problem: if everyone who disagrees quietly exits, the movement never evolves. It calcifies. It rots. Worse: it becomes so brittle that even good-faith disagreement looks dangerous. That’s not a coalition. That’s a porcelain figurine.
The people who can help you build power are often not the people you most want to sit beside at dinner. Politics is the art of getting over that.
There’s a line in The Crucible where John Proctor says, "God help me, I lusted, and there is a promise in such sweat." It’s an ugly truth. But it’s real. The play is about witch hunts and purity and the way guilt becomes a weapon. The same dynamic plays out online, except we’re burning credibility instead of witches.
If your goal is catharsis, then by all means: demand perfection, draw hard lines, and keep the circle tight. But if your goal is power, the math changes. You need people. And people are messy. They’ve voted for the wrong party. They follow someone problematic. They don’t get the language right. You can spend your time trying to fix all of that. Or you can find the overlap and build.
Compromise Isn’t Capitulation
There’s a tendency to treat any ideological dilution as a betrayal. But in practice, every successful movement has required compromise. Not surrender. Not appeasement. Just the boring, difficult work of figuring out what matters most and building structures that can hold more than one kind of person.
In software engineering, there’s the idea of a "minimum viable product." Maybe we need a minimum viable politics. Not a total system. Not an ideology that accounts for every edge case. Just enough cohesion to get something done.
With all apologies to Yeats, the center can hold. It just needs reinforcement.
The internet gives us reach, but it takes away intimacy. It virals outrage, but not trust. If you want to build something that lasts, purity can’t be your foundation. Coalitions aren't perfect. They aren’t clean. They require tolerance for disagreement, a stomach for ambiguity, and a willingness to say: this person is useful, even if they’re wrong about some things.
Purity tests scale badly. They turn politics into performance. And performance makes bad strategy.
Build coalitions. Or lose.