640 THE AMERICAN JOURNAL OF SOCIOLOGY

It cannot be anything else. The fault of " subjective interpreta- tion" as an arbiter of method is that it is likely to be too little the mind's organization of elements observed in the object. It will consequently be too much the mind's fiction stimulated by certain impressions received from the object, but completed by extraneous material. The report of the object proves, then, to have in it relatively too little of the object and relatively too much of the subject. This danger is inevitable in the long process of deriving universals. It may be averted only by curbing the impertinences of the subjective presumption.

Sociology is essentially an effort to find more adequate categories with which to conceptualize social details, and to organize the contents of these categories into a universal con- ception. It is dangerous, however, to think anything in cate- gories which cannot be observed, but have to be imputed. In applying such categories we are likely to interpret by deduction from unauthorized impressions that fill the mind in the absence of adequate analysis of the object.

The whole argument of these papers is virtually upon the problem here presented. As the essentials involved will be dis- cussed in various relations, further detail may for the present be postponed.

E. The desirable combination of metftods.^ — It may be said in

general that men who have tried to explain social life have

tended to vibrate between two extremes. On the one hand they

have exaggerated fragments, sections, phases, ahsirdiCiions, disjecta

membra of human activities and conditions, and have neglected

the containing whole ; or they have adopted a presumption of

the whole which took away their freedom so to investigate the

parts that more appropriate conceptions of the whole might

result. Our thought about human affairs has consequently been

a farrago of snap judgments, partial formulations, and promotions

of narrow generalizations to the rank of universals. In order

' Among recent contributions to this subject the following deserve special notice : BosANQUET, " Relation of Sociology to Philosophy," Mind, January, 1898 ; Cald- well, "Philosophy and the Newer Sociology," Contem. Rev., September, 1898 ; Baldwin (F. S.), " Present Position of Sociology," Pop. Sci. Monthly, October, 1899 ;

GiDDiNGS, "Exact Methods in Sociology," Pop. Sci. Monthly, December, 1899.