Page:Mind (Old Series) Volume 12.djvu/449

 436 M. H. TOWEY : mere index to an indefinite number of other distinctive attributes. If a large part of these qualities are unknown, and " infinite, so far as we are concerned," what grounds have we for affirming that the Natural Kind possesses them ? How can we build a class on an a priori supposition ? Further, how can we be justified in framing a class upon such a changeable and subjective point as our own ignorance? Why should that enter as a factor into Divisions of Things ? Surely the number of properties belonging to a group and our hopelessness of discovering them are two points that are wholly alien to the question of the group-forma- tion. That, Mill has told us, is regulated by the quantity and importance of statements concerning the group which shall be applicable to the members. If many and important identical statements can be made concerning such and such things, group them. Good. But what statements can be made about an unknown multitude of attributes? Hitherto we have classed things on account of their recognised resemblances, not on account of their assumed and as-yet-unfound ones. It is undeniable, of course, that of things agreeing in only one bond of likeness (e.g., colour, shape, specific gravity, &c.), only one general assertion and its corollaries are possible. And that of a Kind, as Horse, or Animal, or Sulphur, many general assertions are possible. But the one class is no whit less a merely intellectual creation than the other. Yet it is juster, it will be said. More useful to us, doubtless, but not more objectively true. More useful to us is what underlies Mill's remark, that it would be a palpable absurdity to investigate the common properties of all white things. But Nature has in reality neither the class White Things nor the class Horse. We made both. Mill, however, would say that in the latter case there is a distinction answering to our class. Well, then, so there is, in the former. There are a quantity of things in the universe, alike in point of being white ; there are a quantity of things alike in points a b c, &c. = Horses. The properties are not found by the Kind, but the Kinds are formed by the properties. OBJ. 3. The second criterion of Kinds is that they wholly differ from each other, whilst non-natural Kinds differ only in finite and determinate particulars. Roses and Brambles are not natural Kinds, because a rose does not seem to differ from a rubus, or the Urnbelliferae from the Banunculaceae in much else than the charac- ters botanically assigned to those genera or those families. All Kinds, Mill says, must have a place amongst classes, but all classes in a natural arrangement cannot be Kinds, for the distinc- tions of Kind are not numerous enough to make up the whole of a classification. "The great distinctions of Vascular and Cellular, Dicotyledonous or Exogenous, and Monocotyledonous or Endoge- nous plants are perhaps differences of Kind. The lines of demar- cation which divide those classes seem, though even in this I would not pronounce positively, to go through the whole nature of the