The danger for him in this laudable endeavor seems twofold. First, one must remark that the language of definition is inevitably the language of generality because only the generalizable is definable. Singulars and individuals can be described but not defined; _e.g._, one can define man, but one can only describe Abraham Lincoln. The greater, then, his solicitude for the factual and the concrete, the more irresistibly is he borne in the direction of abstract language, which alone will encompass his collected facts. His dissertations on human society begin with obeisance to facts, but the logic of his being a scientist condemns him to abstraction. He is forced toward the position of the proverbial revolutionary, who loves mankind but has little charity for those particular specimens of it with whom he must associate. In the second place and more importantly, the definition of non-empirical terms is itself a dialectical process. All such definition takes the form of an argument which must prove that the _definiendum_ is one thing and not another thing. The limits of the definition are thus the boundary between the things and the not-thing. Someone might inquire at this stage of our account whether the natural scientists, who must also define, are not equally liable under this point of the argument. The distinction is that definitions in natural science have a different ontological basis. The properties about which they generalize exist not in logical connection but in empirical conjunction, as when “mammal,” “vertebrate,” and “quadruped” are used to distinguish the genus _Felis_. The doctrine of “natural kinds” thus remains an empirical classification, as does the traditional classification of elements.[156] Consequently the genus _Felis_ has a reality in the form of compresent positive attributes which “slum” cannot have. The establishment of the genus is not a matter of negating or depriving other classes, but of naming what is there. On the other hand one could never arrive positivistically at a definition of “slum” because its meaning is contingent upon judgment (and theoretically our standard of living might move up to where Westchester, Grosse Point, and Winnetka are regarded as slums). Thus “slum” no more exists objectively than does “bad weather.” There are collections of sticks and stones which the dialectician may call “slums,” just as there are processions of the elements which he may call “bad.” But these are positive things only in a reductionist equation. Of course, the natural scientist works always with reductionist equations; but the social scientist, unless he is an extreme materialist, must work with the full equation. It is a grave imputation, but at the heart of the social scientist’s unsatisfactory expression lies this equivocation. Remedy here can come only with a clearer defining of province and of responsibility. II