All science leads necessarily to a _system_; and system signifies here, not the proceeds of observation or a route followed by analysis or synthesis in order to arrive at the knowledge of the truth. System here means, a mode of classifying beings or observed facts, a mode essentially in connection with the science which treats of these beings or of these facts, and often applicable to itself alone. A perfect system can only be really established _à posteriori_, after the knowledge has been acquired of all the phenomena which are to be classed. This is absolute. In practice, a system can only be observed _à priori_, by reference to a certain number of facts which it is destined afterwards to embrace; it is only true that the more facts we acquire, the more chance has a system of being exact, without our ever having the right of proclaiming it to be absolutely good; it may be satisfactory, and remain so for a long time, but one fine day a new fact may prove it to be false. “I am of opinion,” said Étienne Geoffroy, “that a perfect system cannot exist; it is a sort of philosopher’s stone, impossible to be discovered.”[327] A science being given, it does not at all follow that there already exists a proper method for classifying in a natural series the phenomena which manifest themselves to us in this branch of human knowledge. If we have not yet succeeded in discovering a true anthropological system, if Camper, Prichard, and Morton have been foiled, it is because the science of mankind is still too new.