But a realist may try to escape this consequence. He may say: “No, the relation R is itself, to my mind, something sure, indeed, but transcendent. I do not regard causality in itself, or explanation in itself, as capable of being presented to the mind in any possible experience. What I say is, that the facts of the type exemplified by p are known to stand in a transcendent relation R to a transcendent basis x. This is sure. But R is as transcendent as x.”
I reply: Thus you but open the door to a fatal infinite progress. One asks you, again: What evidence can you give for this transcendent and unexperienced existence, beyond consciousness, of R, — say, of causation, or of some other form of explanatory relation? Afresh you must answer, if you still cling to the present line of argument: “Because the facts of experience demand, for their explanation, the existence of some such transcendent relation to transcendent realities.” But this new demand for explanation introduces a new relation, R' between the facts of experience, a, b, c, etc., and the first relation R, which was to be that relation to x whereby they were explained. All our questions as to R now recur as to R' the new mediator that is to bring us to the assumption of R. For instance, if you first had said: “The data of experience need causation to explain them,” one has now asked you, as above: What sort of causation? — the sort of causation known within experience, and, by its very definition, known as a datum of possible experience? Our realist is now supposed to have replied: “Not so. The causation whereby I explain the data of experience is itself a
