There remains one point which must not be ommitted in a complete description of the kind of questions which Ethics has to answer. The main division of these questions is, as I have said, into two; the question what things are good in themselves, and the question to what other things these are related as effects. The first of these, which is the primary ethical question and is presupposed by the other, includes a correct comparison of the various things which have intrinsic value (if there are many such) in respect of the degree of value which they have; and such comparison involves a difficulty of principle which has greatly aided the confusion of intrinsic value with mere goodness as a means. It has been pointed out that one difference between a judgment which asserts that a thing is good in itself, and a judgment which asserts that it is a means to good, consists in the fact that the first, if true of one instance of the thing in question, is necessarily true of all; whereas a thing which has good effects under some circumstances may have bad ones under others. Now it is certainly true that all judgments of intrinsic value are in this sense universal; but the principle which I have now to enunciate may easily make it appear as if they were not so but resembled the judgment of means in being merely general. There is, as will presently be maintained, a vast number of different things, each of which has intrinsic value; there are also very many which are positively bad; and there is a still larger class of things, which appear to be indifferent. But a thing belonging to any of these three classes may occur as part of a whole, which includes among its other parts other things belonging both to the same and to the other two classes; and these wholes, as such, may also have intrinsic value. The paradox, to which it is necessary to call attention, is that the value of such a whole bears no regular proportion to the sum of the values of its parts. It is certain that a good thing may exist in such a relation to another good thing that the value of the whole thus formed is immensely greater than the sum of the values of the two good things. It is certain that a whole formed of a good thing and an indifferent thing may have immensely greater value than that good thing itself possesses. It is certain that two bad things or a bad thing and an indifferent thing may form a whole much worse than the sum of badness of its parts. And it seems as if indifferent things may also be the sole constituents of a whole which has great value, either positive or negative. Whether the addition of a bad thing to a good whole may increase the positive value of the whole, or the addition of a bad thing to a bad may produce a whole having a positive value, may seem more doubtful; but it is, at least, possible, and this possibility must be taken into account in our ethical investigations. However we may decide particular questions, the principle is clear. The value of a whole must not be assumed to be the same as the sum of the values of its parts. (§ 18 ¶ 1)

A single instance will suffice to illustrate the kind of relation in question. It seems to be true that to be conscious of a beautiful object is of great intrinsic value; whereas the same object, if no one be conscious of it, has certainly comparatively little value, and it is commonly held to have none at all. But the consciousness of a beautiful object is certainly a whole of some sort in which we can distinguish as parts the object on the one hand and the being conscious on the other. Now this latter factor occurs as part of a different whole, whenever we are conscious of anything; and it would seem that some of these wholes have at all events very little value, and may even be indifferent or positively bad. Yet we cannot always attribute the slightness of their value to any positive demerit in the object which differentiates them from the consciousness of beauty; the object itself may approach as near as possible to absolute neutrality. Since, therefore, mere consciousness does not always confer great value upon the whole of which it forms a part, we cannot attribute the great superiority of the consciousness of a beautiful thing over the beautiful thing itself to the mere addition of the value of consciousness to that of the beautiful thing. Whatever the intrinsic value of consciousness may be, it does not give to the whole of which it forms a part a value proportional to the sum of its value and that of its object. If this be so, we have here an instance of a whole possessing a different intrinsic value from the sum of that of its parts; and whether it be so or not, what is meant by such a difference is illustrated by this case. (§ 18 ¶ 2)

There are, then, wholes which possess the property that their value is different from the sum of the values of their parts, and the relations which subsist between such parts and the whole of which they form a part have not hitherto been distinctly recognised or received a separate name. Two points are especially worthy of notice. (1) It is plain that the existence of any such part is a necessary condition for the existence of that good which is constituted by the whole. And exactly the same language will also express the relation between a means and the good thing which is its effect. But yet there is a most important difference between the two cases, constituted by the fact that the part is, whereas the means is not, a part of the good thing for the existence of which its existence is a necessary condition. The necessity by which, if the good in question is to exist, the means to it must exist is merely a natural or causal necessity. If the laws of nature were different, exactly the same good might exist, although what is now a necessary condition of its existence did not exist. The existence of the means has no intrinsic value; and its utter annihilation would leave the value of that which it is now necessary to secure entirely unchanged. But in the case of a part of such a whole as we are now considering, it is otherwise. In this case the good in question cannot conceivably exist, unless the part exist also. The necessity which connects the two is quite independent of natural law. What is asserted to have intrinsic value is the existence of the whole; and the existence of the whole includes the existence of its part. Suppose the part removed, and what remains is not what was asserted to have intrinsic value; but if we suppose a means removed, what remains is just what was asserted to have intrinsic value. And yet (2) the existence of the part may itself have no more intrinsic value than that of the means. It is this fact which constitutes the paradox of the relation which we are discussing. It has just been said that what has intrinsic value is the existence of the whole, and that this includes the existence of the part; and from this it would seem a natural inference that the existence of the part has intrinsic value. But the inference would be as false as if we were to conclude that, because the number of two stones was two, each of the stones was also two. The part of a valuable whole retains exactly the same value when it is, as when it is not, a part of that whole. If it had value under other circumstances, its value is not any greater, when it is part of a far more valuable whole; and if it had no value by itself, it has none still, however great be that of the whole of which it now forms a part. We are not then justified in asserting that one and the same thing is under some circumstances intrinsically good, and under others not so; as we are justified in asserting of a means that it sometimes does and sometimes does not produce good results. And yet we are justified in asserting that it is far more desirable that a certain thing should exist under some circumstances than under others; namely when other things will exist in such relations to it as to form a more valuable whole. It will not have more intrinsic value under those circumstances than under others; it will not necessarily even be a means to the existence of things having more intrinsic value; but it will, like a means, be a necessary condition for the existence of that which has greater intrinsic value, although, unlike a means, it will itself form a part of the more valuable existent. (§ 19 ¶ 1)