I think I agree - this is a kind of End-To-End Argument, just substitute abiguity in the place of transmision erros. By the way I think there should be some more abstract formulation of the argument.