A reversible trial can answer a question that discussion keeps enlarging.

Some questions do not get smaller with discussion. Each round adds a new consideration, everyone becomes more invested in their own position, and a decision that could have been tested in a fortnight turns into a standing argument about theory. When the cost of being wrong is low and the step can be undone, trying it produces better information than more talking.

Take the debate you keep reopening and design the smallest reversible version of it. Agree in advance how long the trial runs and what would count as a result. Then run it, and let what actually happens end the argument.