The 'A Priori' and the Empirical tion of ourselves). Obviously it must be admitted that I can have ideas of things which I have never experienced by my senses, e.g. centaurs, but it was contended that such ideas were always formed by combining in some way ideas thus derived, e.g. in this case the idea of a horse's body and that of a man's head, or at least by comparing such ideas or abstracting elements in them. But there must be at least one idea that is not so derived. Inference is impossible unless we have the idea of logically necessary connection between the premises from which we start in our inference and the conclusion we reach, and this idea is clearly quite different from any which could be given in sense-experience. Other ideas which have been, at least plausibly, supposed not to be derived from sense-experience are God, good, duty, the laws of logic, substance, cause, and the ideas of geometry such as that of a perfect circle or square. It is particularly difficult to suppose that moral ideas could be derived from sense-experience, but the place for discuscion of them is a book on Ethics.