ation of such procedure. At this point we immediately assume, per analogiam (in accordance with the theory of the historic method, which we have elaborated above), that the procedure itself is something older and earlier than its utilisation in punishment, that this utilisation was introduced and interpreted into the procedure (which had existed for a long time, but whose employment had another meaning), in short, that the case is different from that hitherto supposed by our naïf genealogists of morals and of law, who thought that the procedure was invented for the purpose of punishment, in the same way that the hand had been previously thought to have been invented for the purpose of grasping. With regard to the other element in punishment, its fluid element, its meaning, the idea of punishment in a very late stage of civilisation (for instance, contemporary Europe) is not content with manifesting merely one meaning, but manifests a whole synthesis "of meanings." The past general history of punishment, the history of its employment for the most diverse ends, crystallises eventually into a kind of unity, which is difficult to analyse into its parts, and which, it is necessary to emphasise, absolutely defies definition. (It is nowadays impossible to say definitely the precise reason for punishment: all ideas, in which a whole process is promiscuously comprehended, elude definition; it is only that which has no history, which can be defined.) At an earlier stage, on the contrary, that synthesis of meanings appears much less rigid and much more elastic; we can realise how in each individual case the elements of the synthesis change their value and their position, so that now one element and now another stands out and pre-
