58 T/fE AMERICAN JOURNAL OF SOCIOLOGY
among persons competent to criticise the assumptions so used. In the present state of knowledge it is safest for those sociolo- gists who approach the subject from the humanities side to let this border territory severely alone. The best work will be done there at present either by men whose sociological interest is hardly known to themselves, or by sociologists who have approached the problems of association from the physical side. There are certain uses in carrying biological speculations over into the field of human associations, just as there are certain uses in carrying psychological speculations back into the field of biology. The misuse of this method appears when sociologists duplicate the practice of those thrifty New England fishermen who used to send their young herring across the ocean and bring them back as French sardines. The biological generalizations which sociologists are apt to use are inventions of speculative philosophers, who have exported them into biology, and then have imported them into sociology as accredited scientific results.
For instance, versions of supposed laws of heredity, environ- ment, natural selection, have done service in sociological theory, while competent biologists have never sanctioned their use, except as hypotheses. They cannot be validly used in any other way in sociology. It is therefore safer and more economical in the end for sociologists to employ such hypothetical scientific data as little as possible, and to confine themselves to territory in which they can be more sure of their ground. The sociologist must know where his problems reduce to physical problems, but he must know that he is not, as a sociologist, equipped for their solution.
In the course of his ethical argument, 1 and frequently else- where, Spencer has adverted to the impotence of the idea of causation in most minds. His thought runs back to the prem- ises now under consideration. Knowledge of social conditions and movements involves intelligence about the physical setting in which associations occur, and of the physical forces of which human associations are in part the product. In practice this amounts to a demand that at every step in sociological theory
' E. g., Principles of Ethiit, Book I, chap, iv, et passim.
