The move: proof as a mirage
The pattern has three beats. A criterion is set: "show me one customer who asked for this." The criterion is met: here are three. The criterion is revised as though it had always been the real one: "well, three isn't a trend — show me it moves revenue." Each individual demand sounds like rigor. The sequence is the tell: whatever arrives, the finish line recedes by exactly one step.
The phrase is British, borrowed from goal-based sports, and settled into print as a named argumentative pattern in the 1980s — largely as a business expression, which is fitting, because the modern goalpost lives in project reviews, promotion criteria and definition-of-done debates.
Why it's a core denial technique
Researchers of science denial rank goalpost-moving among the five fundamental techniques — in John Cook's FLICC taxonomy it appears as impossible expectations: demanding a standard of proof the domain cannot supply, then treating the shortfall as refutation. Decades of climate data meet "but models aren't perfect"; a vaccine trial of forty thousand meets "but not on people exactly like me." The demand is calibrated to be unmeetable, and recalibrated if accidentally met.
The workplace version is smaller but structurally identical — and often not even conscious. Status-quo defenders rarely experience themselves as moving goalposts; each new criterion feels like prudence. That is what makes the written record so clarifying: it separates the honest raiser-of-standards (who will happily amend the recorded criteria in daylight) from the tactical one (whose standards only ever move after being met).
Recognizing it
- ✗Success is always one more study away — every met criterion spawns a successor, with no acknowledgment that the last one was met.
- ✗Retroactive rewording — "that's not what I meant by proven" arrives only after the proving.
- ✗Asymmetric standards — the demanded rigor applies to your position, never to theirs.
- ✗No stateable finish line — asked directly what evidence would suffice, they cannot or will not say.
The counter
The verbal counter is to fix the line before running at it: "Before we gather anything — what evidence would change your mind?" Get it specific, get it acknowledged, and the later move becomes visible as a move. If the criteria genuinely should rise — new information does that honestly — the raise gets named as a change, not smuggled in as a clarification.
This tactic simply does not survive a written record. In the Argumentree Method the framed question states what is being decided and by what criteria, before the arguing starts; the decision record keeps that framing on the page. Meeting the criteria is then a checkable fact, and changing them is an explicit edit with a timestamp and an author — which legitimate revision can proudly be, and tactical revision cannot afford to be. The goalposts can still move; they just can't move silently.
Where it connects
Goalpost-moving is the criteria-level twin of motte-and-bailey — one revises the standard, the other revises the claim, and both need the fog of unrecorded talk. Its energy source is status-quo bias and motivated reasoning: the conclusion is fixed, so the criteria must float. And it is a first cousin of sealioning's bottomless evidence demands — the difference is that the sealion never states a bar at all, while the goalpost-mover states one and then disowns it.