See Caird, Op. cit., Vol. II, p. 402. See also Hinton, The Law Breaker, p. 287: "The real meaning of the difficulty about a word for "regard for others" is that we do not want it. It would mislead us if we had it. It is not a regard for _others_ that we need, but simply a _true_ regard, a regard to the facts, to nature; it is only a truth to facts in our regard, and its nature is obscured by a reference to "others", as if that were the essential point.... It is not as being for others, but as being _true_, that the regard for others is demanded."
Some ethical writers have gone to the other extreme and held that all benevolence is a disguised or an enlightened selfishness, since having a necessary reference to self. The reference to self must be admitted; unless the action springs from an interest of the agent himself the act may be outwardly useful, but cannot be moral. But the argument alluded to inverts the true relation involved. If a man's interests are such that he can find satisfaction only in the satisfaction of others, what an absurdity to say that his acting from these interests is selfish! The very fact of such identity of self with others in his interest is the proof of his unselfishness.
See Leslie Stephen, Science of Ethics, p. 241, for an admirable discussion of this difficulty. When it is said that your pain is painful to me, he says, the inference is often "insinuated that I dislike your pain because it is painful to me in some special relation. I do not dislike it _as_ your pain, but in virtue of some particular consequence, such, for example, as its making you less able to render me a service. In that case _I do not really object to your pain as your pain at all_, but only to some removable and accidental consequences." (And see his whole treatment of sympathy, pp. 230-245). The whole question is shown to come to this: Is my interest in, my sympathy with, your joy and sorrow as such, or in your joy and sorrow as contributing to mine? If the latter, of course the interest is selfish, not being an interest in others at all. But if the former, then the fact that such sympathy involves one's own satisfaction is the best proof that man is not selfishly constructed. When Stephen goes on to say that such sympathy does not involve the existence of a real unity larger than the individual, he seems to me to misread his own facts, probably because he conceives of this unity as some abstract or external thing.