GENERAL SUMMARY A.VD COXCLUDIXG REMARKS. 415 



Philology," therefore, I briefly traced the principles of 

 language growth, so far as these are now well recognized by 

 all philologists. It was shown, as a matter of classification, 

 that the thousand or more existing languages fall into about 

 one hundred families, all the members of each family being 

 more or less closely allied, while members of different families 

 do not present evidence of genetic affinity. Nevertheless, 

 these families admit of being comprised under larger groups 

 or "orders," in accordance with certain characteristics of 

 structure, or type, which they present. Of these types all 

 philologists are agreed in distinguishing between the 

 Isolating, the Agglutinating, and the Inflectional. Some 

 philologists make a similar distinction between these and 

 the Pol)'synthetic, while all are agreed that from the agglu- 

 tinative the Incorporating type has been derived, and from 

 the inflectional the Analytic. 



Passing on from classification to phylogeny, we had to 

 consider the question of genetic relationship between the 

 three main orders, inter se, and also between the Polysyn- 

 thetic type and the Agglutinating. The conflict of authori- 

 tative opinion upon this question was shown to have no 

 bearing upon the subject-matter of this treatise, further than 

 to emphasize the doctrine of the polyphyletic origin of 

 language — the probability appearing to be that, regarded as 

 types, both the isolating and the polysynthetic are equally 

 archaic, or, at all events, that they have been of equally 

 independent growth. In this connection I adduced the 

 hypothesis of Dr. Hale, to the effect that the many apparently 

 independent tongues which are spoken by different native 

 tribes of the New World, may have been in large part due to 

 the inventions of accidentally isolated children. The curious 

 correlation between multiplicity of independent tongues and 

 districts favourable to the life of unprotected children — in 

 Africa as well as in America — seemed to support this hypo- 

 thesis ; while good evidence was given to show that children, if 

 left much alone, do invent for themselves languages which 

 have little or no resemblance to that of their parents. 



