G4: Loyalty, Heresy, and Betrayal

The brave one and the sellout

Imagine two politicians, on the same afternoon, each break with their own party. The first belongs to the party you usually vote for. She stands up and says that on one issue you care about, the other side has been right all along and her own party has been wrong. The second belongs to the party you usually vote against. He stands up and says the same thing in reverse: on that issue, your side has been right, and his own party has been wrong.

Answer quickly. Which of them is showing courage? Which of them is cashing in, angling for a book deal, or settling a private score?

The tempting answer is that the one who crossed over toward you is brave, and the one who crossed over away from you is an opportunist. Yet the act is identical. Each broke ranks, in public, at some cost. If that was your answer, your verdict depended only on the direction of travel.

This is not a thought experiment for long. In the space of two weeks in early 2022, a state Democratic Party and the Republican National Committee each formally censured members of Congress from their own party for breaking ranks. The details are further down this page. Before you read them, notice which of the two you already think deserved it.

Now a smaller version. Suppose a friend who has shared your views for years posts something agreeing with the other side. Before you read their reasons, what is your first feeling: curiosity, or a small sting? You would not feel the sting if a stranger from the other side posted the same words. You expect strangers to disagree.

That sting, and that split verdict, are this grammar at work in you.

What the grammar is

Every group that holds beliefs in common has members who drift from them. The grammar is the habit of treating that drift as a moral offense rather than as a disagreement: of reading a dissenter inside the group as a traitor, punishing them harder than an outside opponent, and then treating the punishment itself as proof of who was right.

Its deep structure has two steps. First, dissent from inside is treated as more dangerous than opposition from outside, so the heretic is punished harder than the enemy. An outside opponent can be dismissed as not understanding; an insider who says the group is wrong cannot, and so threatens its coherence. Second, the expulsion gets read as evidence. The group takes it as proof that the expelled were wrong: good members do not leave, so whoever left must have been bad. The expelled take it as proof that they were right: they told the truth, and truth-tellers get thrown out. Neither inference follows. Being expelled tells you that a group objected to someone. It does not tell you whether the objection was correct.

This site calls a recurring pattern of argument like this a grammar: a shape the same mistake takes across many different subjects, so that once you can see it in one argument you can see it in others. This one has two poles, and each is a real failure.

Pole A, the enforcers, mistake internal agreement for truth. They treat the fact that everyone in the group holds a view as evidence for the view, and so they expel the very members who could have corrected the group's errors. The people best placed to notice a group's mistakes are the ones who share its goals and know its arguments from the inside. The enforcer removes them.

Pole B, the professional heretics, turn contrarianism into an identity. Having been pushed out of a group, or having walked out, they treat the expulsion as proof of being right rather than as one possible explanation among several. People are expelled for telling uncomfortable truths, and also for being wrong, or cruel, or tiresome. The professional heretic stops being able to tell these apart, because the role of brave outcast has become who they are.

The two poles feed each other. Each expulsion by an enforcer produces a potential professional heretic, and each professional heretic gives the enforcers fresh evidence that dissenters were never really loyal.

Taking the pattern apart

The grammar works by merging several questions into one. Separating them does most of the work of undoing it.

Is the claim true? This is the question the grammar is best at hiding. Whether a dissenter is right about the issue they raised is a question about the issue. It is not answered by their loyalty, their motives, their history with the group, or the reception they got.

Who is saying it, and does that matter here? Sometimes it does: a person's record and incentives can be fair evidence about how much to trust their testimony. But judging who made an argument is a different act from judging the argument, and the grammar slides from one to the other. This site's term for the distinction is argument screening versus source screening.

What is the agreement made of? When everyone in a group believes something, that may be because each of them looked at the evidence and reached the same conclusion independently. Or it may be because disagreeing was costly, so the people who doubted kept quiet or left. Agreement of the first kind is evidence. Agreement of the second kind is mostly a record of who was willing to stay.

What else could explain the expulsion? If someone was thrown out, list the explanations. They were right and the group could not bear it; they were wrong and the group corrected them; they were right but behaved badly; the group faced a real threat. More than one can be true. Both poles pick one and stop.

Would I judge this the same way if the direction were reversed? If a defection toward your side looks like courage and an identical defection away from it looks like betrayal, you are not judging the defection.

Why it is so persistent

Groups need some loyalty to function at all. A movement whose members abandon it at the first disagreement accomplishes nothing, so some suspicion of defectors is not irrational. The grammar begins when that suspicion is applied to ideas rather than to conduct, so that disagreeing becomes a form of disloyalty in itself.

Social psychologists study the first half of the pattern under the name black sheep effect. In the 1988 paper that coined the term, José Marques, Vincent Yzerbyt and Jacques-Philippe Leyens predicted, as their abstract puts it, "that judgments about both likeable and unlikeable ingroup members are more extreme than judgments about outgroup members." Notice the word both: the claim is not simply that people punish their own harder, but that they judge their own more intensely in either direction. One of their three experiments, the abstract reports, showed that the effect "occurs only when the judgmental cues are relevant for the subjects' social identity." (This site has read the abstract, not the paywalled paper.)

Several later studies concern members who broke rules of conduct, not members who disagreed. A 2018 paper by Dominic Abrams and colleagues, including Marques, summarizes nine of the authors' earlier studies of rule-breaking by leaders and ordinary members, such as a team captain swearing at a referee or a leader cheating to win, and reports that "transgressive ingroup leaders were evaluated and punished less harshly than identically transgressive ingroup members or outgroup leaders or members." The researchers call this "transgression credit," and they found limits to it: the leader had to seem to serve the group, not cross serious moral lines, and belong to a fairly small group.

The studies that tested disagreement itself come closer to this grammar. In a 2000 paper, Abrams, Marques and two colleagues found, in one of two studies, that members of one's own group whose opinions ran against its norm were judged more harshly than outsiders with the same opinions, "even though both held identical attitudes." (Again, this site has read the abstract only.) The 2018 paper also describes earlier experiments on attitudes to asylum policy, in which a member who deviated from the group's position, in either direction, and was about to become a leader was "evaluated more positively, granted more freedom to innovate and rewarded more than other deviants." So the research supports a qualified version of this grammar's central claim: in the laboratory, people judge an ordinary member who disagrees with their group more harshly than an outsider saying the same thing, and can give incoming leaders more room. How far that carries over to parties, churches and movements, these studies do not show.

A related idea is tribal epistemology, meaning a way of deciding what is true by asking which side a claim comes from. The journalist David Roberts used the phrase in its political sense in a 2017 Vox essay, defining it as a habit in which information "is evaluated based not on conformity to common standards of evidence or correspondence to a common understanding of the world, but on whether it supports the tribe's values and goals and is vouchsafed by tribal leaders." Roberts aimed it chiefly at the American right, writing that it was "disproportionately animating one of America's two political parties." This site uses his definition but not his targeting: the test of the concept is whether you can find it on your own side. Once truth is sorted by side, a member who changes their mind is not updating; they are switching teams.

How former members talk about the group they left is itself a battleground of this grammar. In the American argument over "cults," critics of new religious movements relied on ex-members' testimony, and scholars sympathetic to those movements questioned it. James R. Lewis, who studies new religious movements, reported in a 2009 conference paper that in his research "most voluntary defectors were ambivalent or even positive about their former religion," while people who had been "intensively counseled by anti-cultists" described their former groups "in terms of popular negative stereotypes about 'cults.'" (Such counseling, called deprogramming or exit counseling, aimed to talk a member out of a group, and was sometimes carried out after the member had been taken by force.) He quotes the anti-cult writer Michael Langone making the opposite charge: that sympathizers "tend . . . to accept at face value cultists' reports while doubting the accuracy of ex-cultists and their reports, sometimes pejoratively referring to them as 'apostates.'" Each side discounts the other's witnesses, and neither is neutral: Lewis calls the counselors "anti-cult vigilantes," and Langone wrote from the anti-cult side. Lewis's research, known here only from his own summary, suggests that the picture of leavers as a group's fiercest critics may fit the leavers an audience hears from more than leavers in general; this site has not found a study that settles it.

Source still being found. a study, read in the original, comparing former members' hostility toward a group with that of people who were never members

How the argument has been had before

Religious excommunication. On July 27, 1656, the Portuguese Jewish congregation of Amsterdam expelled the 23-year-old Baruch de Spinoza. The writ, in the translation of Asa Kasher and Shlomo Biderman, cites "the evil opinions and acts of Baruch de Spinoza," "the abominable heresies which he practiced and taught" and "his monstrous deeds." It says the community's governors had tried "by various means and promises, to turn him from his evil ways," and had heard witnesses in his presence. But it never says what the heresies were or answers them, and it ends by forbidding anyone to "read any treatise composed or written by him." Yet the community's position was not simply foolish. At a public symposium in 2015, the chief rabbi of the Portuguese Jewish community, Pinchas Toledano, declined to lift the ban, explaining, as JTA reported, that doing so "would clearly imply we share the views of Spinoza that state that God does not exist" and that "the law of Moses is no longer relevant." A religious community has to be able to say what it believes. What the ban could not do was show that Spinoza was wrong. Three and a half centuries later, JTA noted, his books were on sale in the synagogue's gift shop.

McCarthy-era loyalty programs. In March 1947, President Truman's Executive Order 9835 set up a loyalty program for federal employees. It allowed removal when there were "reasonable grounds" for "belief that the person involved is disloyal," and it listed among the evidence "sympathetic association with" any organization the Attorney General designated as subversive. In April 1951, Executive Order 10241 lowered the bar to "a reasonable doubt as to the loyalty of the person involved." The burden had moved from evidence of disloyalty to doubt about loyalty. In April 1953, Executive Order 10450 revoked Truman's order and replaced it with a security program. According to a history quoted by the Truman Library, between 1947 and 1956, across these programs, "over five million federal workers underwent screening, resulting in an estimated 2,700 dismissals and 12,000 resignations."

The fear was not invented. Even Walter and Miriam Schneir, sharp critics of the historians they were reviewing, wrote in a review in The Nation, a left-leaning magazine, of the decoded Soviet cables known as Venona, that "no reasonable person" could doubt "that in World War II Washington some employees of government agencies were passing information that went to the Russians." The failure was in treating association as guilt. At the Army–McCarthy hearings on June 9, 1954, Senator Joseph McCarthy announced that the Army's lawyer, Joseph Welch, had in his firm "a young man named Fisher" who had belonged to "an organization which is named, oh, years and years ago, as the legal bulwark of the Communist Party." Fred Fisher, Welch explained, had belonged to the Lawyers Guild in law school and for some months after, and was by then, in his own words as Welch recounted them, "Secretary of the Young Republicans League in Newton." Welch's reply, in the transcript: "Have you no sense of decency, sir, at long last?"

Sectarian left splinters. In March 1937, two months after the second of the Moscow show trials, Joseph Stalin told a meeting of the Soviet Communist Party's Central Committee that "contemporary Trotskyism is not a political trend within the working class, but an unprincipled and intellectually devoid band of wreckers, diversionists, intelligence agents, spies, and killers" (text). A disagreement inside the movement had been redefined as treason. Leon Trotsky's followers, organized as the Fourth International, answered in the same vocabulary, though not by the same means. Their 1938 program, the Transitional Program, called its rivals, from the French socialist leader Léon Blum to Stalin himself, "social betrayers," and predicted that, though its members "will inevitably feel themselves isolated" when war came, they "will be found at the head of the revolutionary tide." Each side defined itself by the other's betrayal, and the smaller one expected its isolation to be temporary and its vindication certain. The two sides were not equal: both spoke the language of betrayal, but only one of them held state power. Its version ended in the purges and, in August 1940, in Trotsky's murder "by an assassin acting on Joseph Stalin's orders," as AFP reported on the anniversary.

Talk-radio hunts for "RINOs." The acronym RINO, "Republican In Name Only," has its earliest recorded print use, according to the dictionary site Word Spy, in a December 31, 1992 column in New Hampshire's Union Leader. (A Democratic mirror, DINO, followed by 1994.) In an August 26, 2016 article by Oliver Darcy for Business Insider, the conservative talk-radio host John Ziegler described what happened to a host who would not accept two conspiracy claims about Hillary Clinton and Barack Obama: "it's almost as if you're a sellout. You're a RINO. You're somehow part of the liberal elite. It's nuts." That is one host's account, not a study of talk radio, but it shows the label doing this grammar's work: doubting a claim the audience wanted believed was recast as leaving the team.

In each case, the enforcers let the act of expulsion stand in for an argument about who was correct. All of these episodes show the enforcer's failure; this site has not yet found and sourced a historical case of the professional heretic's: someone who took their own expulsion as proof of being right.

Example still being found. a sourced historical episode in which someone expelled from a group offered the expulsion itself as proof of being right

How each pole fails

The enforcer's failure starts with treating agreement as the measure of truth. A member who raises a doubt is not seen as offering information but as revealing disloyalty, and the response is aimed at the person rather than the doubt. Once that member has gone, the group's agreement looks even stronger, because only those who agreed are left to count.

Here are the two censures the opening promised. On January 22, 2022, the Arizona Democratic Party's executive board censured Senator Kyrsten Sinema, who had voted against changing the Senate's filibuster rules to pass voting-rights bills. The party's chair, Raquel Terán, said the board acted "as a result of her failure to do whatever it takes to ensure the health of our democracy" (ABC News). On February 4, 2022, the Republican National Committee censured Representatives Liz Cheney and Adam Kinzinger for serving on the House committee investigating the January 6, 2021 attack on the Capitol. The resolution said they "support Democrat efforts to destroy President Trump more than they support winning back a Republican majority," and that the party should "immediately cease any and all support" of them as Republicans (NPR).

Before reading on, ask yourself which of the two looked like a party protecting itself and which like a party punishing courage. If the answer lined up with your politics, go back to the two politicians at the top of this page.

Held to the same standard, the two are closer than either side's supporters think. Each turned a member's judgment on a contested question (whether the filibuster should bend for one cause, whether a party member should help investigate the leader of her party) into a failure of loyalty, and each put the member's standing in the party at stake: the Republicans by ending the party's support, the Arizona party by its earlier threat of a "no confidence" vote if she did not change her stance, while Democrats in the state raised money for a primary challenger, ABC reported. Terán had even contrasted Sinema with Arizona's other senator, Mark Kelly, who "courageously voted to defend the voting rights of millions of Arizonans," according to the left-leaning site Talking Points Memo: the opening's vocabulary exactly. There is one real difference. The Arizona chair addressed Sinema's stated reason, the defense of the filibuster, if only to reject it: "In the choice between an archaic legislative norm and protecting Arizonans' right to vote, we choose the latter." The Republican resolution did not address the two members' stated reasons at all; it attributed a motive. On that point, the Republican resolution was the harsher of the two.

The professional heretic's failure is the same machine run backward. Where the enforcer counts agreement as proof, the heretic counts rejection as proof. Any correction can be absorbed as one more attack, and because the role depends on being opposed, the heretic drifts toward positions that provoke opposition, right or not.

No one should be diagnosed as a professional heretic from a single statement, and this page does not do so. What can be shown is the inference itself: the step at which hostility is offered as a credential. In November 2025, after months of public disagreement with President Trump, Representative Marjorie Taylor Greene was cut off by him. Trump wrote that she had gone "Far Left" and said, "She's lost a wonderful conservative reputation" (AP, via PBS). That is the enforcer's move: disagreement read as defection to the other side. The following week Greene announced her resignation from Congress. Her statement objected to "being called a traitor" and argued that "Loyalty should be a two-way street." But it opened with the other pole's move: "I've always represented the common American man and woman as a member of the House of Representatives, which is why I've always been despised in Washington DC and never fit in" (full statement, as reprinted by Mediaite). Being despised is offered as evidence of whose side she is on. That is a looser form of the inference than this page defined: hostility from Washington, not expulsion by her own side, offered as proof of loyalty to ordinary people, not of being right about an issue. It is consistent with her claim, and with other explanations the sentence does not weigh.

Liz Cheney's statement before the vote to censure her, which NPR reported, included: "I do not recognize those in my party who have abandoned the Constitution to embrace Donald Trump. History will be their judge." It offers nothing as proof; it shows only how close the poles sit, since "I do not recognize" does to her party what her party was about to do to her.

Both cases come from people who broke with the American right. This site has not yet found a clear statement of the inference from someone leaving a left-leaning group. That is a gap in the research, not a finding that the habit belongs to one side, and until it is filled this section shows the professional heretic's failure from one side only.

Example still being found. a real, quoted, dated instance of someone who left or was pushed out of a left-leaning group presenting that departure as proof of being right

Both poles share one move: each lets the social fact of expulsion answer the factual question of who was right.

Where it borders other grammars

With G1: Purity and Contamination it shares the purity test. A belief treated as a contaminant makes anyone who holds it a source of contamination, and expelling them becomes a kind of cleansing rather than a disagreement.

With G10: Victimhood as Standing it shares the professional heretic's credential. Being expelled can become a form of injury that grants authority to speak, so that "I was cast out for saying this" is offered as though it supported what was said.

With G11: Contested Epistemic Authority it shares the question of when consensus is evidence. A field or community that punishes dissent makes its own agreement harder to trust.

With G6: The Hidden Hand it shares the suspicion of infiltration. A dissenter who cannot be explained as honestly mistaken can be explained as secretly working for the other side, as Stalin's speech did to Trotsky's followers and as McCarthy did to Fred Fisher.

When naming the pattern becomes a weapon

This page's vocabulary can be used to end arguments as easily as any other.

In one direction, "orthodoxy," "purity test" and "cancel culture" can dismiss a group's response to a member's conduct by recasting it as punishment for dissent. In March 2021, after several prominent New York Democrats called on Governor Andrew Cuomo to resign over allegations of sexual harassment and, for some of them, an attorney general's report that, they said, found his administration had hidden data on nursing-home deaths, he told reporters that resigning would be "bowing to cancel culture" (Rolling Stone, left-leaning and openly critical of him). After a report commissioned by New York's attorney general found the women's claims credible, Cuomo, denying that he had touched anyone inappropriately and calling the report politically motivated, resigned in August 2021. In his first speech after leaving office, on March 6, 2022, he said: "With 'cancel culture' we now live in a frightening New World" (New York Now). Whatever one concludes about the allegations, which this page does not adjudicate, his party's response was to accusations about conduct, not to a heretical opinion.

In the other direction, attacks on who a former insider was can stand in for answering what they say. On January 29, 2019, the day a former White House aide, Cliff Sims, published a memoir of his time there, President Trump wrote that it was "based on made up stories and fiction," then turned to the author: "He pretended to be an insider when in fact he was nothing more than a gofer," calling him "a low level staffer that I hardly knew" (as quoted by the conservative-leaning Fox News). Trump's tweet does not use this page's vocabulary; it makes the same move with a put-down about rank. Labels such as "grifter" can do the same work; this site has not yet recorded a dated example.

Example still being found. a real, quoted, dated instance of "grifter" or "professional contrarian" being used to dismiss a former member's specific criticisms without answering them

A person who left a group may know things about it no outsider does. That they benefit from saying so is a reason to check their claims, not to ignore them.

Both weapons do the same thing as the grammar they claim to name: they settle a question about whether someone is right by pointing to their relationship with a group.

Catching it in yourself

When someone from my side changes their mind, do I ask what they learned, or what happened to them?

Which of my beliefs would I be afraid to question out loud among my friends? If others around you share that fear, their agreement on that point tells you less than it seems to.

Have I ever taken being disagreed with as a sign I was right? That feeling is the professional heretic's habit in miniature.

Who did my group stop listening to, and were they ever right? Think of someone your side wrote off. Ask whether anything they said has since turned out to be true.

Tools that help

• C2: The Steelman Protocol. State the dissenter's position so well they would sign it, before deciding what you think of it. Against this grammar it forces the question the grammar hides, which is whether the heretic's argument is any good, and the professional heretic must state their former group's case fairly too.
• F3: The Charity-with-Audit Protocol. Read an ambiguous statement in its most reasonable sense first, and let a demonstrated pattern, not a single remark, justify a harsher reading later. It slows the leap from one deviation to a verdict of betrayal.
• A4: The Mirror Check. Ask whether you scrutinized the claim or simply liked where it came from. It catches the split verdict in the opening of this page, where the same defection reads as courage in one direction and treachery in the other.
• F5: The Exit Ramp. Recognize when an exchange has become a loyalty performance for an audience, and leave without writing anyone off. It gives both the group member and the dissenter a way to stop a fight without converting the other person into a traitor or a persecutor.

Where it shows up now

As of October 2026, the site has not yet examined this grammar in a live argument through a full article. The 2022 party censures and the 2025 break between President Trump and Marjorie Taylor Greene on this page are examples from the recent past, not a study of current debates. The following are listed as places to look, not as findings.

• Primary-election purity dynamics. Not yet examined.
• Exvangelical deconstruction and its backlash, meaning former evangelicals publicly leaving the faith, and the responses to them. Not yet examined.
• "Gender-critical" expulsions from left spaces. Not yet examined.
• MAGA loyalty tests. Not yet examined.
• Fandom civil wars. Not yet examined.

Related concepts

• The black sheep effect. Judging members of one's own group more extremely than comparable outsiders.
• Transgression credit. Judging rule-breaking leaders less harshly than ordinary members who break the same rules.
• Tribal epistemology. Deciding what is true by asking which side a claim comes from.
• The apostate effect. People who leave a group becoming its fiercest critics, and the group treating them worse than outsiders. Contested; see above.
• Argument screening versus source screening. Judging an argument on its merits versus judging it by who made it.
• When consensus is evidence and when it isn't. The difference between agreement reached independently and agreement produced by the cost of disagreeing.