Page:Popular Science Monthly Volume 84.djvu/497

Rh us; there is no whole and completed development in the world's history; the beginning and the end are unknown; the origin is shrouded in darkness; before the future there hangs the veil of Mayo; we can observe the pupa or cocoon, but not the caterpillar and moth. Under such conditions every explanation must be subjective in character.

Leaving then to one side the search for laws of social dynamics, the historian contents himself with disclosing the causal relations of the successive movements in the evolution of human society, and this is the sole aim of history; but even here the science is subject to important limitations for the use of experiment is impossible, so that the method must be that of observation. At best many disadvantages confront the observer, which are not encountered by the experimenter, so that his results form a very insecure basis for induction, unless, as in other sciences, his observations can be often repeated and the human senses aided by sensitive instruments. But repetition and the use of instruments are not for the historian, who works over the observations of the untrained minds of the past. In seeking the cause or causes of any phenomenon the natural scientist views it as a type of a large class; and even in the case where causation is determined by a single experiment, there always exist numerous phenomena of the same kind or else the particular phenomenon offers itself to the possible repetition of the observation, so that the assurance of the opportunity of repeating the test case plays an important part in the induction. The scientist abstracts from the occurrence all individual variations and finds the cause of the typical phenomenon, which is generalized in thought so as to cover all individuals of like kind. Thus are obtained causal relations, which have objective truth. Such a method of abstraction is inapplicable in history, for, as we have already seen, from the point of view of the historian each phenomenon is exceptional and can not be classified to find types, and also the same conditions and events never recur.

In the search for causal relations how far is it possible to make use of the canons of inductive logic? On account of the reasons stated above it is impossible to find two events which agree or disagree in all respects except one. Therefore the canons of agreement and difference are of no assistance in historical research. The impracticability of these canons in history has always been acknowledged, and yet the literature of history as well as of sociology and economics is filled with errors arising from their unscientific use.

Of the inductive canons there remain those of residue and of concomitant variations, neither of which is a very safe criterion of causal relations and both of which can to a limited extent be employed by the historian. When there are general propositions proved by other sciences, such as psychology, sociology and economics, which will establish the needed partial causation, the canon of residue can be used. The value of such reasoning will depend on the reliability of the general