152 BUREAU OF AMERICAN ETHNOLOGY [bull. 40 



the accent is not strong, and in most cases does not belong to the 

 word, but to the sentence. 



Nouns and pronouns are clear cut. They are capable of calling up 

 definite and complete mental visions without the aid of associated 

 words and word-elements. The large number of monosyllabic nouns 

 in Hupa, and the still larger number in related languages, which do 

 iK)t occur in Hupa, points to the fact that the original form of Atha- 

 pascan nouns was monosyllabic. Monosyllabic nouns have given 

 place to pol3^syllabic ones in Hupa constantly for 3'ears, perhaps for 

 centuries. This may have been due to the pleasure which the Hupa 

 find in poetical descriptive names, but it was certainl}^ due, in part, to 

 the dropping of nouns out of the language at the death of persons 

 who had had them for names. These dropped words were replaced 

 by longer descriptive words coined for the purpose. 



Only one word has been found in the language which appears to be 

 reduplicated. The aboriginal flute is called milhnil or mihnil in 

 Hupa, and in related dialects hiilhul. It is possible that some 

 etymology will appear to explain this apparent exception. 



Very few words or word-parts seem to be onomatopoetic in their 

 origin. There is a verb, kyuwindil it rang, the root of which, -dil^ 

 no doubt represents the sound of striking metals. Another verb 

 closely resembling this is kyuwihhet^ which is used of the creaking of 

 trees. The sounds of nature which occur may be represented, but 

 the}^ have no other meaning. They do not stand foi' the thing or 

 animal which makes them: for example, d!l duwenne {dil it sounded) 

 is said of an arrow striking the sky; did duwenne {dvl it sounded), 

 of a ball of wood striking a wall of obsidian; and ka ka duwenne {ka 

 ka IT said), of the cawing of a crow. 



For the most part, both the monosyllabic words and the elements of 

 the longer words are to all appearances the ultimate facts of the lan- 

 guage. They express fundamental concepts and relations, which are 

 no more resolvable into parts than are the syllables which express 

 them. These elements, simple words, roots, prefixes, and suffixes, are 

 not very numerous (probably less than a thousand), but the combina- 

 tions of which they are capable are very great. Many combinations 

 theoretically possible are not logically possible, and of these only those 

 for which there was a frequent need in the life of the people reall}^ 

 existed as words. 



§88 



