Page:Mind (New Series) Volume 8.djvu/549

 J. P. DUBAND, Aperpus de Taxinomie Generale. 535 mother species and to his collaterals, and the great kingdoms of the living world will fail to be discriminated. E.g. the Pri- mordial Monera, the Vegetable Monera and the Animal Monera, would appear as three but slightly differentiated species. If then their natures are construed as class ideas including their descen- dants, the whole animal and vegetable world would be brought within a single genus. It appears to me that votaries of Natural Classification never make any real attempt, such as the author suggests, to combine genealogical and generic orders by erecting the names of common ancestors into names of kinds. And it seems plain that such an attempt would meet with the difficulty indicated above, and em- phasised by Dr. Edward Caird in dealing with the evolution of religion, viz., that by cutting down the essence of great class ideas to the actual given qualities of the least evolved species w : ithin them, all real grasp of their nature is sacrificed. On the other hand, to read into the names of undeveloped species the universal scheme of qualities which we have gathered from their higher descendants, seems to involve falsification of fact. It may be said, therefore, that to combine genealogy and generality, accord- ing to M. Durand's idea of the combination, is impossible. The trunk of the evolutionary tree must stand, as Huxley says, for a general plan of structure (Enc. Brit., art. Evolution). But if so, in what part of such a diagram are the common ancestors repre- sented ? We have just seen the difficulty of supposing the trunk to represent both actual species and a common plan. I do not think it results from what has been said that a generic classification, following the affinity of descent, is an illusion. It only results that the essence of such a classification cannot be expressed through proper or specific names strictly designative of common ancestors. It is plainly conceivable that the principle of heredity may be the true controlling universal, although in a given common ancestor its possibilities are not made explicit. No one who is familiar with a good natural classification will easily withdraw himself from the conviction that he is dealing in it with the differences of a true governing universal ; and his belief will be confirmed rather than weakened by the fact which M. Durand emphasises, that the tendency of such classification is to go behind obvious similarities in looking for presumptions of affinity ; to replace, as I should say, the superficial perception of resemblance by the analysis of identity in difference, pointing ultimately to the unity and differentiation of parts in a whole. In any case M. Durand's work is full of suggestions and remarkably lucid. Perhaps it is somewhat tinged with the idea that the laws of classification can be studied in advance, apart from the special investigation of the matter to be classified. This remark applies especially to the final chapter on the classification of the sciences. BERNARD BOSANQUET.