ment, as we have seen, is an organization of instincts about certain objects of worship in order to conserve values of importance to man. Within what limits are possible modifications of the religious sent in the future conceivable? Or is it possible, since the religious impulse is sentimental and not instinctive, that the values now conserved by religion will be conserved in some other way? In these days we hear much of socialism, and of milder programs for the 'socialization' of our institutions. It is objected to these proposals that it is impossible to change human nature. Well, just what is human nature, and in what sense is it absolutely fixed, and in what respects can it be altered? Such questions as these very possibly can be answered, or at least new light can be thrown upon them, by the method of analysis here advanced.
It may seem that I am making sweeping claims for the psychology of instincts and sentiments, and the objection may be raised that it would be impossible by means of any a priori analysis of instincts and sentiments to distinguish what is alterable and what is inflexible in them. It must frankly be admitted that such problems as I have mentioned could be solved only by taking carefully into account all the available data in the evolution and history of the institutions involved, and in practical social experience at the present time. In view of this concession it might be objected that such a procedure would merely complicate the search for empirical data, by stating in a peculiar terminology what could be much more easily understood without it. My reply is, that this technical terminology and point of view is needed to enable the investigator to know for what facts to look, and how to coordinate his facts after he has found them. Such a mode of analysis could of course be overworked, like anything else. But it does, I maintain, offer promising possibilities, and it can afford to rest its claims for truth upon whatever pragmatic value it may reveal in actual employment. The conceptions, in the form that McDougall states them (and which have been utilized in this paper), are very simple, and Shand's not greatly different conceptions are little more complicated. They will, I believe, be found to be fruitful, and to repay the
