46 BUREAU OF AMERICAN ETHNOLOGY [bull. 40 



compare the whole region with the neighboring districts. They 

 form a unit rather by contrast with foreign phonetics than when 

 compared among themselves, each language having its own peculiar 

 characteristics in a group of this kind. Thus, the Tlingit of the 

 North Pacific coast differs very much from the Chinook of Columbia 

 river. Nevertheless, wdien both languages are compared to a lan- 

 guage of southern California, the Sioux or the Algonquian, traits 

 that are common to both of them appear to quite a marked degree. 



What is true of phonetics is also true of grammatical form, and 

 this is evidently a characteristic trait of the languages of the whole 

 world. In North America particularly such groups of languages 

 can be readily recognized. A more detailed discussion of this prob- 

 lem will be given in another place, and it will be sufficient to state 

 here, that languages — like, for instance, the Athapascan, Tlingit, 

 and Haida — which are spoken in one continuous area on the north- 

 west coast of our continent show certain common characteristics 

 when compared with neighboring languages like the Eskimo, Algon- 

 quian, and Tsimshian. In a similar way, a number of Californian 

 languages, or languages of southern British Columbia, and languages 

 like the Pawnee and Iroquois, each form a group characterized by 

 certain traits wliich are not found in other languages. 



In cases where such morphological similarities occur without a 

 corresponding similarit}^ of vocabulary, it becomes exceedingly diffi- 

 cult to determine whether these languages may be considered as 

 descendants of one parent language; and there are numerous cases 

 in which our judgment must be suspended, because, on the one hand, 

 these similarities are far-reaching, while, on the other hand, such 

 radical differences are found that we can not account for them with- 

 out assuming the introduction of an entirely foreign element. 



Similar phenomena have recently induced P. W. Schmidt to con- 

 sider the languages of Farther India and of Malaysia as related; and 

 the same problem has been discussed by Lepsius, and again by Mein- 

 hoff, in reference to the relation of the languages of the Hottentot 

 to a number of east African languages and to the languages of the 

 Hamitic peoples of North Africa. 



Difficulties also arise in cases where a considerable number of 

 similar words are found without a corresponding similarity of gram- 

 matical forms, so that we may be reluctant to combine two such 

 languages, notwithstanding their similarities of vocabulary. 



