j 2 8 THE SCIENCE OF LOGIC 



(in reference to Evolution), may considerably influence the whole trend and 

 significance of our attempts at classification. 



We have been looking for some tangible test whereby to 

 select &quot; important &quot; attributes on which to frame a classification for 

 general purposes, and we have at last, as it seems, found something 

 tangible, in the opening words of the extract just quoted from Mill. 

 The important attributes are those which make the members of a 

 class &quot; like one another and unlike other things,&quot; those which will 

 secure &quot; that the members of each class may resemble each other 

 in as many points as possible,&quot; 1 while, presumably, they will 

 differ as much as possible from those of immediate co-ordinate 

 classes, and more and more widely from those of more remote 

 classes. In other words, our guiding principle in natural classi 

 fication is to be Degree of Resemblance, Similarity, Affinity in the 

 objects to be classified. Individuals are to be brought together 

 into a group, and groups into wider groups, according to Degree 

 or A mount of Resemblance, Similarity, Affinity. 



Here we have a principle which is not merely intelligible, but 

 apparently at least easy of application. It appears, moreover, to 

 be suitable to our purpose : the gaining of general knowledge 

 about the objects classified. For, what other way have we of 

 advancing in our knowledge of things, than by comparing them 

 with one another, observing their points of agreement and differ 

 ence, arguing by analogy from the known characteristics of one 

 class to the existence of like characteristics in a similar class ? 



In applying the principle, however, we are constantly exposed 

 to the danger of what is known as &quot;judging by appearances&quot;. 

 The common advice, &quot;Do not judge by appearances,&quot; would, of 

 course, be unmeaning if it were interpreted literally ; for, after all, 

 what have we to judge by except appearances ? The real mean 

 ing, obviously, is that we must beware of hastily interpreting 

 points of resemblance (or difference) which may be only superficial, 

 and of no significance or importance towards the advance of our 

 knowledge by indicating profound and far-reaching affinities (or 

 divergences) between the things under examination. And in 

 this sense, the advice, though somewhat of a paradox in the form 

 in which it is expressed, is both sound and useful. 



67. WHY CLASSIFICATION FOR GENERAL PURPOSES is 



CALLED &quot;NATURAL&quot; CLASSIFICATION. When we thus en- 

 1 WELTON, ibid., p. 140 (italics ours). 



