The second weak point in the doctrine lies in the extreme vagueness of the terms, protective and self-regarding.  The practical difficulty begins with the definition of these terms.  Can any opinion, or any serious part of conduct, be looked upon as truly and exclusively self-regarding?  This central ingredient in the discussion seems insufficiently laboured in the essay on Liberty.  Yet it is here more than anywhere else that controversy is needed to clear up what is in just as much need of elucidation, whatever view we may take of the inherent virtue of freedom—­whether we look on freedom as a mere negation, or as one of the most powerful positive conditions of attaining the highest kind of human excellence.

To some persons the analysis of conduct, on which the whole doctrine of liberty rests, seems metaphysical and arbitrary.  They are reluctant to admit there are any self-regarding acts at all.  This reluctance implies a perfectly tenable proposition, a proposition which has been maintained by nearly all religious bodies in the world’s history in their non-latitudinarian stages.  To distinguish the self-regarding from the other parts of conduct, strikes them not only as unscientific, but as morally and socially mischievous.  They insist that there is a social as well as a personal element in every human act, though in very different proportions.  There is no gain, they contend, and there may be much harm, in trying to mark off actions, in which the personal element decisively preponderates, from actions of another sort.  Mr. Mill did so distinguish actions, nor was his distinction either metaphysical or arbitrary in its source.  As a matter of observation, and for the practical purposes of morality, there are kinds of action whose consequences do not go beyond the doer of them.  No doubt, you may say that by engaging in these kinds in any given moment, the doer is neglecting the actions in which the social element preponderates, and therefore even acts that seem purely self-regarding have indirect and negative consequences to the rest of the world.  But to allow considerations of this sort to prevent us from using a common-sense classification of acts by the proportion of the personal element in them, is as unreasonable as if we allowed the doctrine of the conservation of physical force, or the evolution of one mode of force into another, to prevent us from classifying the affections of matter independently, as light, heat, motion, and the rest.  There is one objection obviously to be made to most of the illustrations which are designed to show the public element in all private conduct.  The connection between the act and its influence on others is so remote (using the word in a legal sense), though quite certain, distinct, and traceable, that you can only take the act out of the self-regarding category, by a process which virtually denies the existence of any such category.  You must set a limit to this ’indirect and at-a-distance argument,’ as Locke called a similar plea, and the setting of this limit is the natural supplement to Mr. Mill’s ’simple principle.’