As we have seen, the first of these alternative views, if consistently carried out, will not allow us to regard anything as real of which anything can be said, since all predication is founded on relation of some kind. It therefore naturally leads to the second. All that we in fact count real turns out to be determined by relations. Feeling may be the revelation or the test of the real, but it must be feeling in certain relations, or it neither reveals nor tests anything. Thus we are obliged to recognise a reality, at least of that kind which in our every-day knowledge and action we distinguish from illusion, in what is yet the work of the mind, or at any rate must be held to be so until relations can be accounted for without a relating act or that act referred to something else than the mind. Hence with those who adhere to the opposition between the real and the work of the mind, and who at the same time cannot ignore the work of the mind in the constitution of relations, there arises a distinction between reality in some absolute sense—the reality of things-in-themselves, which are supposed to be wholly exempt from any qualification through relating acts of the mind, but of which, for that reason, nothing can be known or said—and the empirical reality of that which we distinguish from illusion, as standing in definite relations to the universe of our experience. (§ 30 ¶ 1)