No True Scotsman

Someone says no Scotsman puts sugar on his porridge. A friend says his uncle Angus does, and Angus is a Scotsman. The reply comes back instantly: no true Scotsman puts sugar on his porridge. Angus was in the group until he spoiled the sentence. Antony Flew told that tiny story in 1975. The word "true" did not describe Angus. It removed him.

You do it without porridge. A real friend, a real engineer, a real fan would never do the thing this one just did. Instead of changing the claim, you change the membership. They were never really one of us. The new entrance rule exists to protect the generalization. It was not a rule you were using yesterday.

Groups do have edges. A license, a creed, or a skill can mark who counts, if the mark was there before the embarrassing case. The move becomes a fallacy when the edge is redrawn after the case, using the disputed behavior as the new entrance exam. "No careful analyst ignores base rates" is fine as advice. "She ignored base rates, therefore she was never an analyst" is the Scotsman.

• "No true friend would forget a birthday. So he was never really a friend."
• "Real programmers don't use that tool. The ones who do aren't real."
• "A genuine believer wouldn't do this. He must have been pretending."
• "No competent manager misses a deadline. Missing it proves she wasn't competent."
• "Actual fans understand. If you disagree, you're not a fan."
• "Socialists don't want that policy. Anyone who wants it isn't a socialist."
• "I don't fail. If I failed, it wasn't the real me doing it."

The birthday: he forgets, and the sentence you wanted was "friends remember." What you say is that he was never really a friend. The friendship does not have to be revised. The membership does the work instead.

The ship: an engineer you respected yesterday ships a bug you hate. By the afternoon she is "not a real engineer." Her commit history did not change at lunch. Your sentence did.

The mirror: you miss a promise and hear yourself say you don't do that sort of thing, so this must have been a special case, not you. The promise stays missed. The self-description stays clean.

You cannot learn from counterexamples, because counterexamples are defined out of existence. Groups become unfalsifiable: every failure belongs to an outsider, so the ideal never has to update. People inside the group get exiled for ordinary deviation, and people arguing with the group can never bring a case that counts. Applied to yourself, the move blocks a simple admission. You do not have to say you broke your own rule. You say the person who broke it was not you.

Identity claims feel like more than descriptions. "We are the kind who do not do that" is doing social work: pride, belonging, a boundary against outsiders. A counterexample threatens the work, not just the sentence. Redefinition is cheaper than grief or revision. It also flatters the audience still inside. They get to remain true members. Essentialist thinking helps: if the group has a fixed inner nature, then anyone who violates it was never of that nature.

Flew introduced the label in Thinking About Thinking (1975) as a picture of an ad hoc rescue. The logical point is older than the joke. A universal claim is supposed to be open to counterexample. If every counterexample revises the subject term instead of the claim, the claim has stopped saying anything checkable about the world. It only says that whoever does the thing is renamed.

This is a reasoning pattern, not a prevalence study. You will not find a percent of conversations that contain it. You can find it in a definition that changed during the argument and nowhere else.

• "True," "real," or "actual" showed up only after the awkward case.
• You cannot state the membership rule without mentioning the behavior under dispute.
• Yesterday the person counted. Today they do not, and the only new fact is the embarrassing one.
• The generalization would be false if the person stayed in the group, and that is why they cannot stay.
• You feel the category is cleaner after the expulsion, and no belief had to change.

Say the rule and the exception as two sentences. "Analysts should not ignore base rates. This analyst did." The second sentence is allowed to stand. If you need a narrower group, name the narrower group with a criterion you would have used last week: licensed, elected, trained, present at the vote.

When someone else does it, ask what would count as a member who still did the thing. If the answer is "no one," the claim is not about members. It is a definition of the virtue, wearing the group's name.

Are all definitions of "real" this fallacy?

No. "A real square has equal sides" is a definition, not a rescue. The fallacy needs a generalization that was offered as a fact, then saved by changing who counts after a counterexample arrived.

What if they lied about belonging?

Fraud is a different fact. If someone never met a standard that already existed, saying so is description. If the standard was invented to exclude them this afternoon, that is the Scotsman.

Let the counterexample injure the sentence. Sentences heal faster than groups do when you pretend they were never injured.

The friend

The profession

Yourself

Practice this pattern in the Reframing App - capture the trigger, label it (like No True Scotsman), check evidence, and write a more balanced thought.