CLASSIFICATION. 293 



ology, Chemistry, Physiology, &c. ; he goes on to sub-divide 

 his books on Physics, into those which treat of Mechanical 

 Motion, those which treat of Heat, those which treat of Light, 

 of Electricity, of Magnetism. 



Between these two modes of classification, note the essen 

 tial distinctions. Arrangement according to any single con 

 spicuous attribute is comparatively easy, and is the first that 

 suggests itself : a child may place books in the order of their 

 sizes, or according to the styles of their bindings. But ar 

 rangement according to combinations of attributes, which, 

 though fundamental, are not conspicuous, requires analysis \ 

 and does not suggest itself till analysis has made some pro 

 gress. Even when aided by the information which the author 

 gives on his title page, it requires considerable knowledge to 

 classify rightly an essay on Polarization ; and in the absence 

 of a title page, it requires much more knowledge. Again, 

 classification by a single attribute, which the objects possess 

 in different degrees, may be more or less serial, or linear. 

 Books may be put in the order of their dates, in single file ; 

 or if they are grouped as works in one volume, works in two 

 volumes, works in three volumes, &c., the groups may be 

 placed in an ascending succession. But groups severally 

 formed of things distinguished by some common attribute 

 which implies many other attributes, do not admit of serial 

 arrangement. You cannot rationally say, either that His 

 torical Works should come before Scientific Works, or Scien 

 tific Works before Historical Works ; nor of the sub-divi 

 sions of creative Literature, into Fiction, Poetry, and the 

 Drama, can you give a good reason why any one should take 

 precedence of the others. 



Hence this grouping of the like and separation of the un 

 like, which constitutes Classification, can reach its complete 

 form only by slow steps. We saw (First Principles, 36) 

 that, other things equal, the relations among phenomena are 

 recognized in the order of their couspicuousness ; and that, 

 Dtner things equal, they are recognized in the order of their 



