There can be no doubt that Aristotle’s definition is right, in the main, so far as he says that it is an ‘habitual disposition’ to perform certain actions: this is one of the marks by which we should distinguish a virtue from other things. But ‘virtue’ and ‘vice’ are also ethical terms: that is to say, when we use them seriously, we mean to convey praise by the one and dispraise by the other. And to praise a thing is to assert either that it is good in itself or else that it is a means to good. Are we then to include in our definition of virtue that it must be a thing good in itself? Now it is certain that virtues are commonly regarded as good in themselves. The feeling of moral approbation with which we generally regard them partly consists in an attribution to them of intrinsic value. Even a Hedonist, when he feels a moral sentiment towards them, is regarding them as good-in-themselves; and Virtue has been the chief competitor with Pleasure for the position of _sole_ good. Nevertheless I do not think we can regard it as part of the definition of virtue that it should be good in itself. For the name has so far an independent meaning, that if in any particular case a disposition commonly considered virtuous were proved not to be good in itself, we should not think that a sufficient reason for saying that it _was_ not a virtue but was only _thought_ to be so. The test for the ethical connotation of virtue is the same as that for duty: What should we require to be proved about a particular instance, in order to say that the name was wrongly applied to it? And the test which is thus applied both to virtues and duties, and considered to be final, is the question: Is it a means to good? If it could be shewn of any particular disposition, commonly considered virtuous, that it was generally harmful, we should at once say: Then it is not really virtuous. Accordingly a virtue may be defined as an habitual disposition to perform certain actions, which generally produce the best possible results. Nor is there any doubt as to the kind of actions which it is ‘virtuous’ habitually to perform. They are, in general, those which are duties, with this modification that we also include those which _would_ be duties, if only it were possible for people in general to perform them. Accordingly with regard to virtues, the same conclusion holds as with regard to duties. If they are really virtues they must be generally good as means; nor do I wish to dispute that most virtues, commonly considered as such, as well as most duties, really are means to good. But it does not follow that they are a bit more useful than those dispositions and inclinations which lead us to perform interested actions. As duties from expedient actions, so virtues are distinguished from other useful dispositions, not by any superior utility, but by the fact that they are dispositions, which it is particularly useful to praise and to sanction, because there are strong and common temptations