The truth is, I think, and it is characteristic of Mill's modes of analysis, that he is making an impossible abstraction. He is separating parts of a single process and treating them as independent. If actions are bad because they have bad consequences, motives are bad because they are causes of bad actions. You cannot suppress the effect without suppressing the cause, and therefore the cause of the cause. Mill relies chiefly upon one argument. The same conduct will produce the same consequences whatever the motives. That is undeniable. It is the same to me whether I am burnt because the persecutor loves my soul or because he hates me as a rebel to his authority. But when is conduct 'the same'? If we classify acts as the legislator has to classify them by 'external' or 'objective' relations, we put together the man who is honest solely from fear of the gallows and the man who is honest from hatred of stealing. So long as both act alike, the 'consequences' to their neighbours are alike. Neither is legally punishable. But if acts are classified by their motives, one is a rogue and the other virtuous; and it is only then that the question of morality properly arises. In that case, it is idle to separate the question of motive and consequences, because the character determines the motive and therefore the action. Nobody should have seen this more clearly than Mill as a good 'determinist.' Conduct and character are related as the convex and concave of the curve; conduct is simply the manifestation of character, and to separate them is absurd.
Why did he not see this? For reasons, I think, which illustrate his whole method. From a scientific point of view, the ethical problem raises the wide questions, What are the moral sentiments? and, What functions do they discharge in regard to the society or to its individual members? We might hold that morality is justified by 'utility' in the sense that the moral rules and the character which they indicate are essential to the welfare of the race or its individual constituents. But to Mill this proposition is interpreted as identical with the proposition that conduct must be estimated by its 'consequences.' We are to consider not the action itself, but its effects; and the effects are clearly independent of the motive when once the action has been done. We may therefore get a calculus of 'utility': general rules stating what actions will be useful considered abstractedly from their motives. The method, again, might be plausible if we could further assume that all men were the same and differed only in external circumstances. That is the point of view to which Mill, like Bentham, is always more or less consciously inclining. The moral and the positive law are equally enforced by 'sanctions'; by something not dependent upon the man himself, and which he is inclined to suppose will operate equally upon all men. Such language could be justifiable only of an average and uniform 'man,' a kind of constant unit, whose varying behaviour must always be explained by difference in circumstance. We have sufficiently seen the results elsewhere, and in this ethical doctrine they are especially manifest.