A medieval motte-and-bailey castle had two parts. The motte was a steep mound with a small stone tower on top: cramped, not much use day to day, but easy to defend. The bailey was the open courtyard below: the fields, the workshops, the place worth having. In peacetime everyone lived in the bailey. When raiders came, everyone scrambled up to the tower. Once the raiders gave up and left, everyone walked back down to the bailey and carried on.
The philosopher Nicholas Shackel borrowed the picture in a 2005 paper to describe a rhetorical move with the same shape. You hold a bold, valuable claim. When someone attacks it, you retreat to a small, defensible one. When the attacker leaves, you go back to the bold claim as if nothing happened.
The two claims
The whole move lives in the gap between two claims that get treated as one.
The bailey is the claim the speaker actually wants to advance. It is bold, interesting, and contested. It is the reason they are in the conversation at all.
The motte is the fallback. It is narrow, modest, and hard to argue with. Often it is something nobody would dispute. The speaker never really wanted to defend only this, but it makes a fine place to hide.
| What gets said in public (the bailey) | What gets defended when pushed (the motte) |
|---|---|
| "Diet is the only thing that matters for health." | "Diet affects health." |
| "Our app will replace your accountant." | "Our app does some accounting tasks." |
| "That study proves the policy works." | "That study is consistent with the policy working." |
The trick is treating a win in the right column as a defense of the left. You prove the modest thing, then act as if you proved the bold thing. The two claims are not the same, but if they share enough words, a fast-moving conversation will let the swap slide by.
Three examples
An academic claim
A researcher writes that a field is "structured entirely by power." That is the bailey: a strong, sweeping claim about how the field works. A critic pushes back and asks for the mechanism. The researcher answers that, after all, every field has some politics in it and that funding and prestige shape what gets studied.
That retreat is the motte. Almost no one disputes that funding and prestige influence research. But "shaped in part by incentives" is a much smaller claim than "structured entirely by power." If the critic accepts the modest version, the sweeping version quietly returns in the next paragraph, now treated as established.
A product claim
A startup's landing page says its assistant "understands your business like a partner." In a demo, a buyer asks whether it actually reasons about their numbers or just pattern-matches on text. The founder says it "uses advanced language models." True, and uncontroversial. Every comparable product uses advanced language models. The buyer wanted to know about "understands like a partner," the bailey, and got an answer about the model architecture, the motte. The bold claim was never defended. It was just protected from inspection long enough to move on.
A policy claim
Someone argues that a city should abolish all parking minimums because "cars are killing the planet." Challenged on the leap from local parking rules to planetary outcomes, they fall back to "we should care about the environment." Most people agree we should care about the environment. That agreement does nothing to support abolishing parking minimums, which is the claim on the table. The same structure runs in the other direction: a person who says "any regulation is theft" and retreats to "people should keep what they earn" when pressed is doing the identical move. The move has no politics. It is available to everyone, and everyone reaches for it.
How to spot it in your own writing
This is the part that matters, because the speaker usually is not lying. The swap tends to happen on its own. You believe the bold claim and you believe the modest one, so your mind treats them as interchangeable. They are not.
The tell is that your claim changes size depending on who is watching. In unguarded moments it is large and exciting. The instant someone pushes, it shrinks to something safe. Watch for the phrases that mark the shrink:
- "All I'm saying is..."
- "I just mean that..."
- "Obviously I'm not arguing that..."
- "Well, technically I only said..."
When one of those shows up right after a challenge, check what you said before the challenge. If the earlier version was bigger, you have a motte and a bailey, and you are standing on the wrong one.
The fix is to pick the claim you actually want to advance and commit to it. If you want the bold claim, defend the bold version with arguments that support the bold version, not with a modest claim that happens to use similar words. If the honest thing you can support is the modest claim, that is fine. Make the modest claim and stop reaching for the bold one when no one is looking. A claim you can defend at full size is worth more than a big one you can only hold by retreating.
How to address it when someone else does it
Keep both claims visible at once. The move only works while the two stay blurred, so name them. Something like: "Earlier you said the strong version. Are you defending that, or the narrower one?" forces a choice the speaker would rather avoid.
Then hold them to the pick. If they choose the modest claim, the bold one is off the table and should stay there. If they choose the bold one, the modest fallback no longer counts as a defense. Do not let the claim toggle between paragraphs. For more on raising this without turning the thread into a fight, see using fallacy.is in a discussion.
How it differs from nearby moves
Motte and bailey gets confused with three other patterns. The line is worth keeping straight.
A straw man misrepresents your opponent's claim, then knocks down the weak version. Motte and bailey is the reverse: the bold claim belongs to the speaker. They are not distorting anyone. They are swapping their own claim for a safer one and back again.
Equivocation shifts the meaning of a single word inside an argument, like sliding between two senses of "free." Motte and bailey swaps the whole claim, not one word. The sentences can stay clean while the position underneath quietly moves.
No true Scotsman rescues a generalization by redefining the category so counterexamples no longer count. Motte and bailey leaves the category alone and changes how much is being claimed about it. One edits the boundary; the other edits the ambition.
There is also a family link to circular reasoning. When the modest claim gets used as proof of the bold one they share words with, the argument starts leaning on itself, assuming the conclusion it was supposed to earn.
The move dies when you name it
A motte is only a motte for as long as nobody mentions the bailey. Point at the territory the speaker actually wants, the bold claim they keep slipping back to, and the retreat stops working. The tower was never the point. The fields were.
If you want a short link to drop into a thread the next time you see the swap, it is fallacy.is/motte-and-bailey, or fallacy.is/mab for short. The page does the explaining so you do not have to write the paragraph.