Moving the goalposts is demanding new evidence once the evidence originally asked for has been produced. The standard that was supposed to settle the question is quietly replaced with a higher one, and the process can repeat indefinitely.
What makes it a fallacy is not the raising of standards, which is sometimes legitimate, but the pretence that the original standard was ever going to be sufficient.
“Show me one peer-reviewed study.” A study is produced. “One study proves nothing — show me a meta-analysis.” A meta-analysis is produced. “Who funded it?”
Each question is reasonable on its own. The sequence is the tell.
The honest version exists: encountering evidence can legitimately reveal that your original test was too weak. The difference is whether the new standard is stated as a correction or presented as though it had always been the bar.
Moving the goalposts is most damaging where it is hardest to notice, because each individual demand is defensible. Asking about funding is good practice. Asking for replication is good practice. The fallacy lives in the pattern, not in any one request, which means it can only be identified across a conversation rather than within a sentence.
It also has an asymmetric cost: the person meeting the demands does real work each time, while the person issuing them does not. An argument conducted this way can be sustained indefinitely by someone with no evidence of their own.
The defence is to fix the standard in advance and get agreement on it. “If I show you X, will that settle it?” is the single most useful question available, because it forces the standard to be stated before it can be revised — and an opponent who will not answer it has told you something important.
