BOAS] HANDBOOK OF AMERICAN INDIAN LANGUAGES 81 



hand, we find numerous languages in which the stem is entirely 

 unchangeable, excepting so far as it may be subject to phonetic 

 contact phenomena. 



The following grammatical sketches have been contributed by 

 investigators, each of whom has made a special study of the linguistic 

 stock of which he treats. The attempt has been made to adopt, so 

 far as feasible, a uniform method of treatment, without, however, 

 sacrificing the individual conception of each investigator. 



In accordance with the general views expressed in the introductory 

 chapters, the method of treatment has been throughout an analytical 

 one. No attempt has been made to compare the forms of the Indian 

 grammars with the grammars of English, Latin, or even among 

 themselves; but in each case the psychological groupings wliich are 

 given depend entirely upon the inner form of each language. In 

 other words, the grammar has been treated as though an intelligent 

 Indian was going to develop the forms of his own thoughts by an 

 analysis of his own form of speech. 



It will be understood that the results of this analysis can not be 

 claimed to represent the fundamental categories from which the pres- 

 ent form of each language has developed. There is not the slightest 

 doubt that, in all Indian languages, processes have occurred analogous 

 to those processes which are historically known and to which the 

 modern forms of Indo-European languages owe their present forms. 

 Grammatical categories have been lost, and new ones have developed. 

 Even a hasty comparison of the dialects of various American lin- 

 guistic families gives ample proof that similar processes have taken 

 place here. To give an example, we find that, in the Ponca dialect 

 of the Siouan languages, nouns are classified according to form, and 

 that there is a clear formal distinction between the subject and the 

 object of the sentence. These important features have disappeared 

 entirely in the Dakota dialect of the same group of languages. To 

 give another example, we find a pronominal sex gender in all the dia- 

 lects of the Salishan stock that are spoken west of the Coast range in 

 the states of Washington and in British Columbia, while in the dia- 

 lects of the interior there is no trace of gender. On the other hand, 

 we find in one of the Salish dialects of the interior the occurrence of an 

 exclusive and inclusive form of the pronoun, which is absent in all the 

 other dialects of the same stock. We have no information on the 

 44877— Bull. 40, pt 1—10 6 



