580 THE AMERICAN JOURNAL OF SOCIOLOGY
torical investigator is in danger of interpreting known psychical motives into the phenomena where, it may be, only undeter- mined motives are manifest ; and also the reverse of this process is possible. This is the more important since even in daily observation the boundaries between known and unknown run into each other, and, moreover, what was at first consciously done falls gradually by long practice beneath the threshold of consciousness, whence it still operates by way of limitation and impulse. These two critical requirements are urged by Simmel upon historiography : subjective reproduction by the historian of the unfamiliar psychical conditions ; and the obscure, perhaps inevitably obscure, relation between the known and the unknown psychical motives operating in the circumstances studied.
So much for the critique of knowledge. But how about the theory of knowledge ? Simmel has not proposed one. He offers only suggestions in that direction, and shows the way that the theory must necessarily take. He even doubts whether it is possible to reach a solution of the problems of cognitive theory which he has proposed. Thus he says with reference to the first point here raised : "This feeling of something which I still do not really feel, this reconstruction of a subjectivity which is again a subjectivity, but at the same time stands objectively over against the former that is the riddle of historical cognition, for the understanding of which our logical and psychological categories are still much too clumsy instruments " (p. 1 6 ) .
Nevertheless he attempts to solve this riddle to some extent by a very bold hypothesis. He refers this power of understand- ing psychical conditions long since experienced to a sort of inheritance. He says in this connection, among other things : "In order to regard this vast domain of comprehension of psychical occurrences not experienced by ourselves as not wholly miraculous, we may consider such intelligence as a kind of consciousness of latent inheritances" (p. 25). This hypoth- esis would seem to suffice in explanation merely of normal psychical occurrences. Only that which is general and average is usually perpetuated by heredity and propagation. If we have
