to other Branches of Knowledge, 323 



these principles which have not yet been stated, as far as I 

 know, in a systematic manner. 



It is the prevalent opinion of philologists, that the most 

 extensive relations between languages and those which are 

 the least liable to be effaced by time and foreign intercourse, 

 are the fundamental principles of construction. Grammatical 

 construction, or the laws which govern the relations of words 

 in sentences, appear to be very enduring and constant, since 

 it extends to whole classes of languages which have few 

 words in common, though it is supposed that they originally 

 had more. But beyond this, there is a cognate character in 

 words themselves which pervades the entire vocabulary of a 

 whole family of languages, the words being formed in the 

 same manner, and according to some artificial rule. This 

 may be exemplified by the monosyllabic structure of the 

 Chinese and Indo-Chinese languages, and by the principle 

 of vocalic harmony pervading the languages of High Asia, to 

 which I shall have occasion again to advert. Of grammati- 

 cal analogy, or correspondence in the laws of inflection and 

 construction, we have a specimen in the Aboriginal languages 

 of the New World, whose structure is known to be very com- 

 plicated and artificial, and at the same time common to all 

 the idioms of America which have been examined. 



Another example of a more definite character is afibrded 

 by the grammatical structure of the languages of High Asia 

 and Great Tartary, and a still more striking one by that of 

 the Indo-European idioms. 



Connected with the subject of the formation of words is 

 the remark, that in the various branches of particular fa- 

 milies of languages which spring by gradual development 

 from the same root, the elements of words, consonants and 

 occasionally vowels, are found to undergo changes accord- 

 ing to certain rules. Particular classes of consonants in one 

 language are substituted for other classes in another lan- 

 guage of the same family. One European idiom, for ex- 

 ample, substitutes palatine letters for sibilunts ; another re- 

 jects them both, and substitutes labials in their place. 

 When corresponding phenomena can be traced through a 

 great part of the vocabulary of two languages, we recognise 



