are not intrinsically good, i.e. would not be good, if they existed quite alone; anything which is “ultimately good” or “good for its own sake” can contain no such parts. This, I think, is the meaning which we must assign to the expressions “ultimately good” or “good for its own sake,” if we are to say that our theory asserts pleasure to be the only thing “ultimately good” or “good for its own sake.” We may, in short, divide intrinsically good things into two classes: namely (1) those which, while as wholes they are intrinsically good, nevertheless contain some parts which are not intrinsically good; and (2) those, which either have no parts at all, or, if they have any, have none but what are themselves intrinsically good. And we may thus, if we please, confine the terms “ultimately good” or “good for their own sakes” to things which belong to the second of these two classes. We may, of course, make a precisely similar distinction between two classes of intrinsically bad things. And it is only if we do this that our theory can be truly said to assert that nothing is “ultimately
