The short version
Somebody makes a claim about a group. You produce a member of the group who doesn’t fit. Instead of the claim being wrong, the group changes shape: that person turns out not to have really been one after all.
The claim survives. It just isn’t about anything any more.
Where the name comes from
From the philosopher Antony Flew, in Thinking About Thinking (1975). A Scottish chauvinist reads in his Sunday News of the World that the Sidcup Sex Maniac has struck again, and says: “No Scot would do such a thing.” The very next Sunday the same paper carries the considerably more scandalous doings of one Angus McSporran, in Aberdeen. He says: “No true Scotsman would do such a thing.”
What it looks like
No conservative would support that bill.
Then he was never really a conservative.
Nobody on this team ships without tests.
Well, those two aren’t really part of the team.
Why it doesn’t work
Because it converts a claim into a tautology, and quietly, so that it keeps the authority of the original.
Before the move, “no conservative would support that bill” was a real claim about actual people. You could go and check it. It could be wrong. After the move, “conservative” means a person who wouldn’t support that bill — so the sentence is now true, guaranteed, forever, and tells you precisely nothing about anybody.
You’ve traded a claim that could be wrong for one that can’t be anything. That’s a bad trade even when you win it.
There’s a second cost, which is to the word. If membership gets revoked every time a member does something off-message, then the label stops describing a group of people and starts describing an ideal — and everyone still using it to talk about actual humans is now talking past everyone else.
This is not No True Scotsman
- Definitions that genuinely have requirements. “No vegetarian eats beef” isn’t a dodge — not eating meat is what the word means. Excluding a beef-eater from the category is just using English correctly.
- The gap between a label and a practice. “He calls himself one but does none of the things it requires” can be a completely fair observation, provided the requirements were on the list beforehand.
- Claims about a distribution. “That’s not what most of them think” is a checkable statement about a population, not a redefinition. Go check it.
- Groups with actual membership criteria. Licensed, certified, employed, enrolled — plenty of categories have real edges, and pointing at one isn’t a fallacy.
The whole thing comes down to timing: was the new requirement part of the definition before the counterexample showed up? If you’d have said it last week, it’s a definition. If it appeared four seconds ago, it’s the move.
How to call it out
Make the definition come first. That’s it — that’s the entire counter, and it works every time:
- “Before we go further: what makes somebody a real one? Give me the list, and then we’ll check him against it.”
- “Is that a claim about conservatives, or about the kind of person you’d be willing to call a conservative?”
- “Is there anybody who could support that bill and still count? If the answer’s no, then the claim isn’t telling me anything about the world.”
That last question is the one that ends it. If no possible person could ever be a counterexample, then the sentence was never a claim in the first place, and both of you can stop treating it like one.