units of pleasantness, just as weight may be represented numerically, though depending not only on bulk, but on specific gravity? Now it may be admitted at once that it is not inconceivable that the intensity of pleasure should be registered numerically, nor that some numerical equivalent should be found for the qualities of pleasure. The feat has not been accomplished yet, but looking at the success with which non-numerical phenomena are reduced to figures, we have to admit it as conceivable. But we observe also that the higher we rise in the scale of complexity from mechanical to organic nature the more distinct is the growth of a principle of selection or distribution which the members of a combination must follow in order to produce a given quantitative result. We can combine together substances of any bulk and specific gravity without the substances affecting one another we can vary the substances in a thousand different ways. The number which represents the specific gravity acts as a mere multiplier of the bulk. The expansion of a body under heat depends upon a constant called the co-efficient of expansion, and specific to each substance. This constant does not indeed act as a mere multiplier, but still, if we have to expand by a certain amount a row of metal bars of a given length, we can alter the composition of the row by proper care indefinitely. When, however, we come to chemistry, the case is different we find that bodies select their partners by special affinities, they combine with each other only on certain terms. Accordingly, when we know the components of a substance and their proportions, we can represent the compound numerically, because we have equivalents for the separate elements. But the mere atomic equivalent of the compound tells us nothing of the composition. We could obtain the equivalent of sulphuric acid (98) in many ways, but we could only get sulphuric acid by combining together hydrogen and sulphur and oxygen, and in the proportions of 2, 1, and 4.

In organisms this principle of selection is carried still further. If we are determining what food is most nutritious for an animal, we have to look not merely to its bulk, but to the kinds of food wanted by the animal. We might express the nutritiousness of various foods by numbers, but the numerical equivalent of food would tell us nothing unless we knew the kinds of food which have to be combined with one another to suit the animal. We might produce the same numerical sum by a sufficient bulk of a quite indigestible food. In the same way, in the numerical phrase 'one man lives more than another,' we include not merely longer or more intense life, but a certain selection of functions.

20. Now to regard the greatest sum of pleasures as the test of conduct, supposing we could express it in units of pleasure, would be like declaring that when you had an atomic weight of 98 you had sulphuric acid. The numerical test would be useless till we knew what elements were to be combined, and in what proportion. Similarly till we know what kinds of activities (and therefore what kinds of pleasures) go with one another to form the end, the greatest sum of pleasures will give us only the equivalent of the end, but will not tell us what the composition of the end is, still less how to get it. Or to put the matter more simply, since acts depend on character, when we know what the characters of persons are, and how they are combined in morality, we can estimate the corresponding sum of pleasures: which will be the maximum happiness because it represents the solution of the problem how best to satisfy all the characters together. But if we merely know the sum of pleasures, we know nothing of the characters which are the condition.1

21. Thus we can reaffirm our conclusion that the sum

1 Though I have worked out the matter in my own way, the general result is the same as that of Mr. Stephen. See Science of Ethics, c. IX., ii., especially paragraphs 8-12. And also c. x., paragraphs 4-6. Mr. Stephen does not, however, recognise the distinction of pleasures in kind.

of pleasures considered simply in their intensity cannot give a criterion of morality at all. They cannot do so because pleasures which differ only in intensity have no existence, and the pleasures we know in fact combine in ways not determined by mere intensity. On the other hand, the end can always be expressed in a pleasureformula, but that formula is not independent of good conduct or character itself, for we can only describe its component pleasures by reference to the conduct or character which produces them. We have to say that those pleasures are good which are the pleasures of doing good actions, or are such as the good man likes. Nor, it must be added, is it a natural way of expressing the end, for whereas pleasure depends on conduct, we make prominent not the conduct but the pleasure. It is as if in a chemical formula, instead of writing the names of the elements with their proportions as suffixes, we wrote down the numbers of the proportions with the names of the elements as suffixes, e.g., sulphuric acid 2140 instead of H2SO. Finally, it is to be observed that the popular conception of happiness avoids all the difficulties and complexities caused by setting up pleasure as the end, because in that conception pleasures and pains are never considered in abstraction from the conduct to which they belong and the character which enjoys them.

22. Good conduct (that is, equilibrated conduct) is thus once more verified to be the real test of morality. But though the end as pleasure always implies the end. as good conduct, must we go on to say that the standard or criterion of goodness involves pleasure? If pleasure is only a function of the act, is it not merely repeating the criterion to include in it any reference to pleasure? And this is the idea in the minds of those who, following Aristotle's splendid treatment of the problem of pleasure, have regarded pleasure as merely a concomitant closely attached to conduct, but not involved in it. To this it might indeed be answered, that the bare reference to the

quality of the conduct, as consisting, say, in stopping at the second glass, or whatever it may be, does not ensure that the act is done for its own sake, that the act might still be external, and before it can be really good the agent must take pleasure in it. But this objection has been already met when it was shown that the outwardly good act was a different kind of act when it was done as a merely outward action, and when it proceeded from a right sentiment.1

To answer our question, let us note that it arises with regard to perfection also. Perfection by itself is not a criterion of goodness, but yet all good conduct is perfect. If, then, we want to know if an action is right, it is sufficient to ask if it is temperate, or just, or the like; and though if it is so it is also perfect, we need not refer to its perfection for an abstract criterion. It is the same with pleasure we do not need pleasure for an abstract criterion. To add perfection or pleasure to the criterion is only adding that what we are testing is built upon certain pathological data, and that it is conduct which is chosen. But at the same time the abstract statement of the quality of good action is only gathered from the concrete standard itselt, and it is this with which we compare a given concrete action. The difference between the two things is as it, wanting a wooden beam 20 feet long, we should say the criterion was 20 feet, or was 20 feet of wood. Pleasure, therefore, is part of the end by which in fact all conduct is judged, and belongs to it just because it is the end of conduct. Without the pleasure the standard of action would be something divorced from our experience, and if it is an abstraction to consider pleasure apart from the act, it is equally an abstraction to regard the act without its pleasure.

1 See Bk. I., ch. ii., pp. 51, 52. Observe that the usefulness of pleasure as a practical criterion is that it enables us conveniently to say whether the act is done for its own sake or not, when all we are assured of is that its external aspect is that of a good act. If the agent does not feel the appropriate pleasure, it is a sign that his act does not proceed from a good character, and is really different from what it appears to be.

23. (c.) Ethical and Pathological Pleasures.- When the standard of conduct is thus represented by a pleasureformula or law, in which the different parts of conduct are combined according to their pleasure-equivalent, the pleasures so used must be understood with a qualification. They are the pleasures proper to the volition itself, the pleasures of attainment, in distinction from the attendant. or consequent feelings which any action may involve. These pleasures of attainment may be called ethical pleasures, because it is these which carry ethical value: the other pleasures and pains are pathological, because though they have an important bearing on the choice of the act, they do not by themselves make the act either good or bad. Courage is the classical illustration of the contrast. The ethical pleasure is the enjoyment or attainment of the brave act, a pleasure made up partly of the satisfaction of the claims of patriotism, or honour, partly, of the actual pleasure in exercise and self-defence. Though so far superior in preferableness that it is this pleasure which is chosen, it is probably far inferior in intensity to the attendant circumstances of pain, against which also the attendant pleasures of reputation and admiration are of little account. An indefinite variation in these attendant feelings is possible according to the person, and they may be altogether absent; the act is to be done in spite of them. They subsist, but the duty of the individual is not performed unless he prevents their affecting his action. Such incidental pleasures and pains arise from the many points of attachment which a man's character offers to a proposed action. While it furthers some, the action may repress other parts of the character. The pains he suffers in doing right are the price he pays for the manysidedness of his nature. In the case in question, the prospect of danger by its contact with his inclination for life is painful; but the act when presented to the whole mass of moral sentiments (this whole mass of sentiments including, of course, the inclination