296 Dr R. G. Latham on the 



it, or the facts by which it is supported ; principles and facts 

 which, when examined by himself, have convinced him that 

 most of the later movements in this department of ethnogra- 

 phical philology, have been movements in the wrong direc- 

 tion. 



There are two principles upon which languages may be 

 classified. 



1. According to the first, we take two or more languages 

 as we find them, ascertain cei'tain of their characteristics, 

 and then inquire how far these characteristics coincide. 



Two or more languages thus taken might agree in having 

 a large per-centage of words in common, or a large per- 

 centage of grammatical inflexions ; in which case they would 

 agree in certain positive characters. On the other hand, two 

 or more such languages might agree in the negative fact of hav- 

 ing a small and scanty vocabulary, and an inflexional system 

 equally limited ; whilst, again, the scantiness of inflexion 

 might arise from one of two causes. It might arise from 

 the fact of inflexions having never been developed at all, or 

 it might arise from inflexions having been lost subsequent 

 to a full development of the same. In all such cases as these, 

 the principle of classification would be founded upon the ex- 

 tent to which languages agreed or differed in certain exter- 

 nal chai'acteristics ; and it would be the principle upon which 

 the mineralogist classifies minerals. It is not worth while 

 to I'ecommend the adoption of the particular term mineralo- 

 gical, although mineralogy is the science that best illustrates 

 the distinction. It is sufficient to state, that in the principle 

 here indicated, there is no notion of descent. 



2. It is well known that in ethnographical philology (in- 

 deed in ethnology at large) the mineralogical principle is not 

 recognised ; and that the principle that is recognised is what 

 may be called the- historical principle. Languages are ar- 

 ranged in the same class, not because they agree in having 

 a copious grammar or scanty grammar, but because they are 

 descended (or are supposed to be descended) from some 

 common stock ; whilst similarity of grammatical structure, 

 and glossarial identity are recognised as elements of classi- 

 fication only so far as they are evidence of such community 



