Almost all of these objections centre upon the supposed'egoism' of our theory. Now, as I shall show, the theory does not even raise the question of' egoism or altruism.' But before developing this point I wish to remind the reader that the view of desire which is commonly set against our so-called 'psychological egoism' is one whose weakness has already been exposed in the course of the foregoing analysis.

1 It is not denied that under certain conditions such a state, or a close approximation to it, may occur. But where it does, the impulse has not 'entered into the experience of a self-conscious subject.' It is, and is recognised by the subject to be, external to his experience as a subject, almost a 'foreign body.' Such cases are, of course, exceptional.

I refer to the view, especially associated with the name of Bishop Butler, that desire 'terminates on its object.' It may be granted readily that there is an element of real value in Butler's thought here, for the point which the phrase carried for him lay in its implicit repudiation of the fallacy of 'psychological hedonism.' The object of desire is certainly not necessarily, nor even generally, the pleasure expected to accrue from the attainment of some object: and it was important to place it beyond all reasonable doubt that there is no such 'ulterior' end in normal desire. But neither is the object in desire just 'object.' If I have been right, to take desire in this wise is to abandon all possibility of making a distinction between desire and impulse. It is impulse which 'terminates on its object.' In desire the 'object' is raised to a new level by being regarded in the light of its capacity to satisfy the self, charged by the reflective consideration that 'this is something which I want.' 1

It is the presence of this element of 'self-reference' in desire which explains the fact that the desires of a rational being show always a more or less well-marked unity. A man's desires bear the impress (as mere appetites do not) of their common relationship to a single self. This is accepted matter of fact, and we can see why it must be so. To consider the object (as happens in desire) in the light of its capacity to satisfy the self, is to consider it in the light of its compatibility with the other objects which the self is conscious of wanting. And the perception of compatibility will naturally heighten, while the perception of incompatibility will naturally weaken, the self's desire for the object. This is by no means to say that each particular desire will be automatically adjusted, through its reference to the good of the self, the common subject of many desires, to that precise degree of strength in which it will best harmonise with the self's other desires. To say this would be to lose sight of the fact that desire has its roots in instinctive impulse, and that the dominating determinant of the strength of desire is (prior to deliberate volitional effort towards adjustment) the intensity of the impulse. It is obvious that where the underlying impulse is strong, as it may be, e.g. in extreme hunger, it is not always enough to secure harmonious adjustment that the consideration of the object 'eating food' in the light of its capacity to satisfy the self should show (as it may under certain conditions) that 'eating food' is contrary to the self's interests. All that is claimed is that it does naturally promote some adjustment in the way of harmony; and that it is this 'self-referent' aspect of desire, accordingly, which explains why it is that the desires of each individual do as a matter of fact exhibit a certain characteristic unity. That consideration of the object of the impulse in the light of the self's good, in a situation where the conflict of the two is manifest, should fail to produce in the desire any modification of the initial strength of the impulse, seems unthinkable, save in pathological cases: save in cases, that is, where the self is not regarded as behaving like a 'rational being' at all.

1 I am far from wishing to decry the great services of Bishop Butler to moral philosophy, but it does seem to me definitely unfortunate that so few ethical writers, in pausing to bestow a well-earned compliment upon Butler's exposure of the fallacies of Psychological Hedonism, have not paused a little longer to point out the defect in Butler's own theory of desire. This defect has been pointed out on more than one occasion by Idealist writers - notably by E. Caird (Critical Philosophy of Kant, vol. 11. p. 213, note) - and it has been plainly enough indicated that a main point in the Idealist theory is to restore to desire that characteristic in virtue of which alone desire can be distinguished from animal appetite. It is a little disturbing, therefore, to find philosophers of eminence like Dr Broad speaking of the theories of Bradley and Green as though they belonged to a pre-Butlerian stage of thought, and had long been refuted in principle by Butler's arguments against Psychological Hedonism. It would be a great deal nearer the truth to say that these are the only genuine post-Butlerian theories, in that they have not only appreciated the element of value in Butler's view, but have also understood, and indicated clearly, the respect in which that view falls short of being an adequate theory of desire.

But we need not dwell upon the difficulty of accounting for the facts which must beset any attempt to evolve a positive theory of desire without recognition of the element of self-reference. For such attempts originate, almost always, as deliberate reactions against the supposed egoism of the self-referent theory; and if we can show that this egoism is a myth, by far the most serious objection felt against the theory disappears. In what follows I shall consider the problem chiefly in regard to 'motive,' since it is in regard to the special case of motive that the problem appears in its most crucial form. It is urged that if the motive of action is always a 'personal good' of the agent, then all action is ipso facto egoistic.