denying) would be impossible, and there would be no such thing as fancy, hallucination, illusion, or error.
I think it is evidently necessary to revise a vocabulary which lends itself to such equivocation, and if I keep the words existence and intuition at all, to lend them meanings which can apply to something possible and credible. I therefore propose to use the word existence (in a way consonant, on the whole, with ordinary usage) to designate not data of intuition but facts or events believed to occur in nature. These facts or events will include, first, intuitions themselves, or instances of consciousness, like pains and pleasures and all remembered experiences and mental discourse; and second, physical things and events, having a transcendent relation to the data of intuition which, in belief, may be used as signs for them; the same transcendent relation which objects of desire have to desire, or objects of pursuit to pursuit; for example, such a relation as the fact of my birth (which I cannot even remember) has to my present persuasion that I was once born, or the event of my death (which I conceive only abstractly) to my present expectation of some day dying. If an angel visits me, I may intelligibly debate the question whether he exists or not. On the one hand, I may affirm that he came in through the door, that is, that he existed before I saw him; and I may continue in perception, memory, theory, and expectation to assert that he was a fact of nature: in that case I believe in his existence. On the other hand, I may suspect that he was only an event in me, called a dream; an event not at all included in the angel as I saw him, nor at all like an angel in the conditions of its existence; and in this case I disbelieve in my vision: for visiting angels cannot honestly be said to exist if I entertain them only in idea.
Existences, then, from the point of view of know-
