BOAS] HANDBOOK OF AMERICAN INDIAN LANGUAGES 47 



The comparison of vocabularies offers peculiar difficulties in 

 American languages. Unfortunately, our knowledge of American 

 languages is very limited, and in many cases we are confined to col- 

 lections of a few hundred words, without any information in regard 

 to grammatical forms. 0\\'ing to the strong tendency of many 

 American languages to form compound words or derivatives of various 

 kinds, it is very difficult in vocabularies of this kind to recognize the 

 component elements of words, and often accidental similarities may 

 obtrude themselves which a thorough knowledge of the languages 

 would prove to be of no significance whatever. 



Setting aside this practical difficulty, it may happen quite often 

 that in neighboring languages the same term is used to designate the 

 same object, owing, not to the relationship of the languages, but to 

 the fact that the word may be a loan word in several of them. Since 

 the vocabularies which are ordinarily collected embrace terms for 

 objects found in most common use, it seems most likely that among 

 these a number of loan words may occur. 



Even when the available material is fuller and more thoroughly 

 analj^zed, doubt may arise regarding the significance of the apparent 

 similarities of vocabulary. 



Mutual Influences of Languages 



In all these cases the final decision will depend upon the answer to 

 the questions in how far distinct languages may influence one another, 

 and in how far a language without being subject to foreign influ- 

 ences may deviate from the parental type. While it seems that the 

 time has hardly come when it is possible to answer these questions ' 

 in a definite manner, the evidence seems to be in favor of the existence 

 of far-reaching influences of this kind. 



Fhonetic Influences 



This is perhaps most clearly evident in the case of phonetics. It 

 is hardly conceivable why languages spoken in continuous areas, and 

 entirely distinct in vocabulary and in grammatical structure, should 

 partake of the same phonetic characteristics, unless, by imitation, 

 certain phonetic traits may be carried beyond a single linguistic 

 stock. While I do not know that historical evidence of such occur- 

 rences has been definitely given, the phenomenon as it occurs in 

 South Africa, among the Bantu and Hottentot, admits of hardly 



