80 BUEEAU OF AMERICAN ETHNOLOGY [bull. 40 



changed into a w; rp becomes hw, etc. While in some languages 

 these phonetic changes do not occupy a prominent place, the}'^ are 

 exceedingly important in others. They corresponcfin a way to the 

 laws of euphony of Sanskrit. 



Just as much variety as is shown in phonetic systems is found in 

 the use of grammatical devices. In discussing the definition of the 

 word, it has been pointed out that in some American languages the 

 word-unit seems to be perfectly clear and consistent, while in others 

 the structure of the sentence would seem to justify us in considering 

 it as composed of a number of independent elements combined by 

 juxtaposition. Thus, languages which have a polysynthetic char- 

 acter have the tendency to form firmly knit word-units, which may be 

 predicative sentences, but may also be used for denominative pur- 

 poses. For example, the Chinook may say, He runs into the water, 

 and may designate by this term the mink; or the Hupa may say 

 They have been laid together, meaning by this term a fire. On the 

 other hand, there are innumerable languages in America in which 

 expressions of this kind are entirely impossible. 



In forming words and sentences, affixes are used extensively, and 

 we find prefixes, as well as suffixes and infixes. It is not absolutely 

 certain that cases occur in America where true infixing into a stem 

 takes place, and where it might not be better explained as an insertion 

 of the apparently infixed element into a compound stem, or as due to 

 secondary phonetic phenomena, like those of metathesis; but in the 

 Siouan languages at least, infixion in bisyllabic stems that are appar- 

 ently simple in their origin occurs. Otherwise, suffixing is, on the 

 whole, more extensively used than prefixing; and in some languages 

 only one of these two methods is used, in others both. There are 

 probably no languages in which prefixing alone occurs. 



Change of stem is also a device that is used with great frequency. 

 We find particularly that methods of reduplication are used exten- 

 sively. Modifications of single sounds of the stem occur also, and 

 sometimes in peculiar form. Thus we have cases, as in Tsimshian, 

 where the lengthening of a vowel indicates plurality; or, as in 

 Algonquian, where modality is expressed by vocalic modification; 

 and, as in Chinook, where diminutive and augmentative are 

 expressed by increasing the stress of consonants. Sometirnes an 

 exuberance of reduplicated forms is found, the reduplicated stem 

 being reduplicated a second and even a third time. On the other 



