Attention has been drawn, particularly since Kant, to propositions which can not have negative instances. They used to be called a priori, axioms, first principles. Today, they are usually called postulates—C. I. Lewis uses both the old and new terminology—because there is a growing recognition of the fact that at least some of them are not “necessary” in the traditional sense. Kant placed a limitation on the apriorism of the continental rationalists. Current epistemologists and logicians have outstripped Kant in the (...) direction signalled by him. Both he and they, however, concur in the opinion that apriori propositions are subject to negation of a sort. The Kantian dinge-an-sich, though they can not contradict these propositions by being negative instances of them, do nevertheless stand to them in the relation of insubordination, thereby exerting on them a “negative action” of a sort. And, according to current views, there are instances which, though they fall outside the range of significance of a proposition or system and hence can not contradict it, do nevertheless tend to bring about a revision of it by a negative action of some sort. The purpose of this brief essay is to coin a convenient name for such instances and to describe certain properties of their peculiar negative action. (shrink)