Motte-and-bailey

Also written: motte and bailey, motte-and-bailey doctrine

Advancing a strong, contentious claim, retreating to a modest and obviously true one when challenged, then returning to the strong one once the challenge passes.

Escalation:: Sophisticated / covert
Type:: Rhetorical

Where the term comes from

Named by the philosopher Nicholas Shackel, who took the image from the medieval castle. The layout of that castle, and what he does and does not claim about the pattern, are quoted from his own writing below.

Whose definition this is

The one-line summary above is ours. The definition it compresses comes from a named source, quoted in full and linked below — where the two differ, the source is what counts. Everything else on this page that states a fact is sourced.

The name is an image from fortification. ^[1] The tower is unpleasant to live in; the point of holding it is that it lets you keep the good land around it.

Applied to argument: the bailey is the strong, interesting, contentious claim you actually want. The motte is the modest, obviously-true version you fall back to when challenged — not worth attacking, and not what you came to say.

It is not, strictly, a fallacy

This matters because the term is very often used as though it were one.

^[2]

^[3] So what is being described is a body of propositions rather than a single bad argument.

That distinction has a consequence people find counter-intuitive. A natural objection runs: different people advanced those two claims, so you cannot accuse anyone of anything. Shackel grants part of that — against a charge that one argument is fallacious, it would hold, since a fallacy needs a single argument in front of you. But he does not accept that it settles the matter: ^[4]

So the objection does not clear the ground. It changes what is being criticised — from a single person's reasoning to the adherents of a doctrine — rather than removing the criticism.

What it looks like

A claim stated confidently, then rendered unrecognisably modest under pressure
"Obviously I only meant…" where the modest reading would not have been worth saying
The strong version reappearing later, to a different audience, unqualified
Challenges answered by restating the definition rather than the evidence

Each of these has an ordinary explanation. People do overstate things and then correct themselves; a claim really can have a defensible core and a contested edge; and someone repeating a strong version to a different audience may simply not have been present for the earlier exchange.

The honest difficulty

This is a label that is easy to throw and hard to answer.

Someone who genuinely overstated something and is now being careful looks identical, in a single exchange, to someone retreating tactically. The pattern is only visible across time, and the accusation is close to unfalsifiable in the moment — whatever the person says next can be read as the retreat.

That asymmetry is a reason to use the idea to understand a conversation rather than to win one.

What to do

Pin the version. Ask which claim is being defended, and quote it. Most of the work here is preventing the two versions from being exchanged without anyone noticing.

Accept the retreat, out loud. If someone narrows to the modest claim, say that the modest claim is fine and that your disagreement was with the other one. That records the narrowing without requiring anyone to admit a tactic.

Do not argue about motive. Whether the retreat was deliberate is not knowable from outside, and disputing it is how the subject changes.

Not the same as

Clarifying what you meant: People routinely overstate a point and then narrow it under questioning. That is how thinking in public works. The pattern here is the RETURN — going back to the strong version once the challenge has moved on.
Holding a nuanced position: A view with a strong core and cautious edges is normal and defensible. The complaint is not about nuance but about which version is offered to which audience.

Responding at this level

Recognition matters more than response. The tell is the pattern across a conversation, never any single message, and the honest advice is to judge the exchange rather than the sentence.

Others at this level

Sealioning — Relentless, superficially polite demands for evidence that exhaust the target while preserving the appearance of good faith.
Whataboutism — Deflecting a criticism by raising a different alleged wrong instead of answering the one raised.
Conversational gaslighting — Denying what was plainly said or done until the other person doubts their own recollection.
DARVO — Deny, Attack, and Reverse Victim and Offender — a documented response pattern that recasts the accused as the wronged party.

Sources

[1] Nicholas Shackel. Motte and Bailey Doctrines, Practical Ethics, University of Oxford (2014-09-05) [archived]
“a stone tower on a mound (the motte) is surrounded by an area of pleasantly habitable land (the bailey), which in turn is encompassed by some sort of a barrier, such as a ditch”
[2] Nicholas Shackel. Motte and Bailey Doctrines, Practical Ethics, University of Oxford (2014-09-05) [archived]
“a fallacy is an argument that is invalid and equivocation is giving different meanings to the same terms when you should be keeping their meaning invariant. a doctrine, however, is a body of propositions, not an argument. so a motte and bailey doctrine cannot be a fallacy”
[3] Nicholas Shackel. Motte and Bailey Doctrines, Practical Ethics, University of Oxford (2014-09-05) [archived]
“i think it is clearly worth distinguishing the motte and bailey doctrine from a particular fallacious exploitation of it”
[4] Nicholas Shackel. Motte and Bailey Doctrines, Practical Ethics, University of Oxford (2014-09-05) [archived]
“however, when we get clear that we are speaking of a doctrine, different people who declare their adherence to that doctrine can be criticised in this way”