582 THE AMERICAN JOURNAL OF SOCIOLOGY
At all events we may confidently assume that a thinker of the rank of Simmel cannot have failed to perceive how difficult, how almost impossible, is the problem which he has formulated. Psychology is still far from ability to show with sufficient exact- ness the relation in any case between known and unknown psychical factors. Daily experience shows us that, even in the case of a given concrete occurrence, the actors in which stand before us, it is by no means always possible to prove either conscious or unconscious agency. Countless examples are furnished by the criminal courts, where accountability and consciousness are practically identical ideas. How much more difficult is this then in the case of historical persons and actions. It is the more to be regretted that, in addition to proposing the problems, Simmel did not attempt to indicate their solutions, or at least the points of view, not merely methodological, but also concrete, according to which solutions may be found.
Nevertheless the positive results of Simmers investigations in this province are still very significant. So far as I know Simmel has here for the first time clearly spoken out the thought that the Kantian principle, according to which the thinker con- strues experience with his own forms of conception and reflec- tion, is applicable to the psychical sciences, and particularly to history. Here, also, there is an a priori, and indeed a richer and more complete one than in the case of natural science, since the a priori with which we are now concerned is the whole ego of the investigator. "Psychology is the a priori of historical science"
(P- 33).
This pregnant thought seems to me to render impossible for
all time all sorts of naive objectivism in historical investigation. The historians who emphasize the value of being impersonal will never write interesting and intelligible history. Only the chron- icler can be "impersonal," and he only to a certain degree. Whoever seeks on the contrary a reasonable and intelligible unity, a correlation in history which may be scientifically com- prehended, must in the first place live history over again in order to be able to narrate it. This does not cause historical
