126 THE SCIENCE OF LOGIC. 



of them by such classification ? The vast extent of the problem 

 constitutes its difficulty. What general classification of plants or 

 animals, for example, will give us the clearest and deepest insight 

 into their nature, constitution, characteristics, relationship, origin, 

 utility, and so forth? 



The duty of the logician here is merely to make suggestions 

 of a general nature. The value of these will be tested by the 

 facility or difficulty of putting them into practice, and by the 

 practical results reached by acting on them. 



(a) We are told, for example, that if we begin with the 

 summum genus, proceeding downwards, we should aim at selecting 

 as basis of division, at each stage of the process, the most im 

 portant attribute, 1 meaning by &quot; most important &quot; that which is 

 most important for the purpose in hand : and the purpose here 

 being the acquiring of general knowledge, the &quot; most important &quot; 

 attribute will be the one which will enable us to make a maximum 

 amount of aggregate assertion with a minimum number of pro 

 positions about the classes so obtained. 



The aim here is intelligible enough. The grouping together, 

 in a Trade Directory, of people s names beginning with S, does not 

 enable us to make a single other statement about them. The 

 arrangement according to streets will probably give us a basis for 

 numerous statements and inferences about those dwelling in a 

 given street ; the arrangement according to trades will be still 

 better in this respect, and so on. This example, from &quot;arti 

 ficial &quot; classification, shows how one method of classifying may 

 enable us to affirm economically about the members of the classes, 

 much more than another method does. So, too, will the 

 &quot; natural &quot; botanical classification, as compared with the analytical 

 key, and this again as compared with the alphabetical index. 

 And so, again, the classification of flowers on the basis of simil 

 arity of structure, as compared with that on the basis of similarity 

 of colour. 



But is the recommendation in question as practicable as it is 

 intelligible? That the degree of power and facility obtained 

 for embodying a maximum amount of knowledge about the 

 classified objects in a minimum amount of language, is a good 

 test of the value of any scheme of classification, when that scheme 

 has been made out, is undoubtedly true ; for the individual mem 

 bers or groups constituting each such class or sub-class will be 



1 Cf. what was said about Proprium, supra, 47. 



