Pass we now from the view of those who make excellence of being the standard, to the view of those who make virtuousness of action the standard. I do not here refer to moralists who, having decided empirically or rationally, inductively or deductively, that acts of certain kinds have the character we call virtuous, argue that such acts are to be performed without regard to proximate consequences: these have ample justification. But I refer to moralists who suppose themselves to have conceptions of virtue as an end, underived from any other end—who think that the idea of virtue is not resolvable into simpler ideas. (§13 ¶1)

This is the doctrine which appears to have been entertained by Aristotle. I say, appears to have been, because his statements are far from consistent with one another. Recognizing happiness as the supreme end of human endeavour, it would at first sight seem that he cannot be taken as typical of those who make virtue the supreme end. Yet he puts himself in this category by seeking to define happiness in terms of virtue, instead of defining virtue in terms of happiness. The imperfect separation of words from things, which characterizes Greek speculation in general, seems to have been the cause of this. In primitive thought the name and the object named, are associated in such wise that the one is regarded as a part of the other—so much so, that knowing a savage’s name is considered by him as having some of his being, and a consequent power to work evil on him. This belief in a real connexion between word and thing, continuing through lower stages of progress, and long surviving in the tacit assumption that the meanings of words are intrinsic, pervades the dialogues of Plato, and is traceable even in Aristotle. For otherwise it is not easy to see why he should have so incompletely dissociated the abstract idea of happiness from particular forms of happiness. Naturally where the divorcing of words as symbols, from things as symbolized, is imperfect, there must be difficulty in giving to abstract words a sufficiently abstract meaning. If in the first stages of language the concrete name cannot be separated in thought from the concrete object it belongs to, it is inferable that in the course of forming successively higher grades of abstract names, there will have to be resisted the tendency to interpret each more abstract name in terms of some one class of the less abstract names it covers. Hence, I think, the fact that Aristotle supposes happiness to be associated wish some one order of human activities, rather than with all orders of human activities. Instead of including in it the pleasurable feelings accompanying actions that constitute mere living, which actions he says man has in common with vegetables; and instead of making it include the mental states which the life of external perception yields, which he says man has in common with animals at large; he excludes these from his idea of happiness, and includes in it only the modes of consciousness accompanying rational life. Asserting that the proper work of man consists in the active exercise of the mental capacities conformably to reason; he concludes that the supreme good of man will consist in performing this work with excellence or virtue: herein he will obtain happiness. And he finds confirmation for his view in its correspondence with views previously enunciated; saying—our notion nearly agrees with theirs who place happiness in virtue; for we say that it consists in the action of virtue; that is, not merely in the possession, but in the use. (§13 ¶2)

Now the implied belief that virtue can be defined otherwise than in terms of happiness (for else the proposition is that happiness is to be obtained by actions conducive to happiness) is allied to the Platonic belief that there is an ideal or absolute good, which gives to particular and relative goods their property of goodness; and an argument analogous to that which Aristotle uses against Plato’s conception of good, may be used against his own conception of virtue. As with good so with virtue—it is not singular but plural: in Aristotle’s own classification, virtue, when treated of at large, is transformed into virtues. Those which he calls virtues, must be so called in consequence of some common character that is either intrinsic or extrinsic. We may class things together either because they are made alike by all having in themselves some peculiarity, as we do vertebrate animals because they all have vertebral columns; or we may class them together because of some community in their outer relations, as when we group saws, knives, mallets, harrows, under the head of tools. Are the virtues classed as such because of some intrinsic community of nature? Then there must be identifiable a common trait in all the cardinal virtues which Aristotle specifies—Courage, Temperance, Liberality, Magnanimity, Magnificence, Meekness, Amiability or Friendliness, Truthfulness, Justice. What now is the trait possessed in common by Magnificence and Meekness? and if any such common trait can be disentangled, is it that which also constitutes the essential trait in Truthfulness? The answer must be—No. The virtues, then, not being classed as such because of an intrinsic community of character, must be classed as such because of something extrinsic; and this something can be nothing else than the happiness which Aristotle says consists in the practice of them. They are united by their common relation to this result; while they are not united by their inner natures. (§13 ¶3)

Perhaps still more clearly may the inference be drawn thus:—If virtue is primordial and independent, no reason can be given why there should be any correspondence between virtuous conduct and conduct that is pleasure-giving in its total effects on self, or others, or both; and if there is not a necessary correspondence, it is conceivable that the conduct classed as virtuous should be pain-giving in its total effects. That we may see the consequence of so conceiving it, let us take the two virtues considered as typically such in ancient times and in modern times—courage and chastity. By the hypothesis, then, courage, displayed alike in self-defence and in defence of country, is to be conceived as not only entailing pains incidentally, but as being necessarily a cause of misery to the individual and to the State; while, by implication, the absence of it redounds to personal and general well-being. Similarly, by the hypothesis, we have to conceive that irregular sexual relations are directly and indirectly beneficial—that adultery is conducive to domestic harmony and the careful rearing of children; while marital relations in proportion as they are persistent, generate discord between husband and wife and entail on their offspring, suffering, disease, and death. Unless it is asserted that courage and chastity could still be thought of as virtues though thus productive of misery, it must be admitted that the conception of virtue cannot be separated from the conception of happiness-producing conduct; and that as this holds of all the virtues, however otherwise unlike, it is from their conduciveness to happiness that they come to be classed as virtues. (§13 ¶4)