32 THE AMERICAN JOURNAL OF SOCIOLOGY
as a kind of action, since any kind of action, from saying "Mamma" to building a ship, is imitation, provided it is occa- sioned by this particular type of relation to an antecedent similar action. The fact that all these heterogeneous forms of social activity are spoken of as instances of the "process of imitation" is an illustration of the fact that actions the most heterogeneous in outcome may have an intrinsic similarity on the side of origin, and with respect to origin are unified into a single class. The word " association " itself, if it is a name of activity, is a name for all kinds of activity, however diverse, which, after all, are unified by virtue of this peculiar relationship of occasioning or being occasioned by the activities of associates. It is its origin in this peculiar conditioning that all social activity has in common. With reference to this mutually occasioning relationship association is unified and distinct from all other phenomena.
Even on the side of their outcome, the social activities, differ- ent as they are from each other, are also different from all other phenomena, and thus set apart from all other phenomena as one general class by themselves; while on the side of their origin they are seen to be the offspring of types of occasioning relations that are common to them all. At first, and so long as attention is mainly fixed upon their practical outcome, the greater methodological advantage may be secured by emphasizing their differences and analyzing the study of association into economics, politics, ethics, etc. But when we pass on to the deeper genetic task, the task of investigating their rise, and the methods by which they are occasioned, it may appear that the same types of rise and of occasioning are common to them all; that on the side of origins the social activities constitute one unified field of research ; and that methodological advantage is secured by recog- nizing that society constitutes one complex of causal condi- tions, and that the same forms and methods of causation are effective throughout the whole range of social phenomena. Indeed, though it may be impossible to identify by its outcome any general social process distinguished from the processes to be investigated by special social sciences, already considerable