146 L. T. HOBHOUSE : the concept to the experience out of which it arises and to which it refers. This experience we shall find constitutes an intelligible totality of a definite kind and within it the reference of the conception falls. General conceptions are based on resemblances in the order of experience. In such resemblances several factors are ordinarily involved. There are, first, two individuals, Callias and Socrates ; secondly, the common element, man- hood, in which they are alike, and thirdly, the differences which separate them. Further, on these two sets of at- tributes relations of resemblance and difference between the individuals are founded, and these relations are hardly to be identified with the attributes themselves. Lastly, the whole set of contents as finally analysed and compared form a systematised totality for thought with a character of its own. It is on this totality that the general concept, Man, is founded, and within this that its meaning is to be sought, but whether the concept as such should be taken to express the totality as a whole, or rather some element within it, is a point which we shall have to discuss later on. Meanwhile what we have said is enough to fix the general conditions under which conception arises. To form general concepts we must mentally bring points of resemblance in the empirical order together, that is, we must compare. And in order to compare we must in some degree break up the given order. This latter act is generally called abstraction and is made the central point of the work of conception. But here we are on the verge of error. For in the first place abstrac- tion is rather a means to conception than a part of concep- tion itself. That a content is general means that it is found in many instances. It may be as abstract as pure being or as concrete as this room with all its contents at this moment, but as long as it can be said to be realised in more than one case it is general and not merely particular. Of course in point of fact the richer your content is in detail the less likely . it is to be repeated point for point in any other case. So far, I suppose, the old doctrine of the ' inverse ratio ' is a truism. Hence, ordinarily, if you want resemblance, and especially exact resemblance, you must ignore differences. And to this extent abstraction as mere ' leaving out ' is a means of forming the general concept. Abstractness is not as such generality nor is generality abstractness, but the empirical order being what it is, a tangle of resemblances and differ- ences, one way of ascertaining and retaining resemblances is to leave differences out of account. So far abstraction figures as a means. But this is not all.
