474 THE TLINOTT INDIANS Ieth. axx. 26 



to deal with a set of suffixes composed of single consonants, and it is 

 not always certain that a supposed stem cvc is not really cv plus 

 the suffix -c. On the other hand, there are many Haida forms of two 

 .syllables endinof in a weak vowel, this vowel being perhaps of no sig- 

 nificance. Supposing this to be the case, all such stems would resolve 

 themselves into single syllables of the type cvc and increase the pro- 

 portion of such stems just so much. 



To assure ourselves that a comparison of languages on the basis of 

 the types of stems which they contain is of real importance it would 

 be necessary to make similar classifications of the roots and stems of 

 .several other languages. This has not been carried out extensively 

 by anyone so far as the writer is aware, but an attempt to treat one of 

 the Athapascan dialects in this manner s<>enied to show that in this 

 re.spect Athapascan and Tlingit are much more closely related to each 

 other in this respect than is either of them to Haida. 



3. Word coiiipofitiim. In l)oth Tlingit and Haida nouns may be 

 stems, they may be formed b}' combining two other nouns, or they 

 may be taken from verbs or adjectives. In forming a compound two 

 nouns observe the same order as in English. These facts, however, 

 are world-wide and furnish no basis for comparison. The absence of 

 foi-mal gender from both is of little more significance. Tlingit appears 

 to differ from Haida in possessing a plural, but this is really a collec- 

 tive, and nouns in the plural do not require it. On the other hand, 

 certain Haida words referring to human beings have a special plural 

 or collective form which corresponds in a measure grannnatically, 

 though not morphologically, to the Tlingit. In both languages terms 

 of relationship take a special plural, but the Tlingit is formed by suf- 

 fixing the personal pronoun of the thii-d pei'son plural, or indefinite 

 person, while the Haida suffix is never employed independently. Cases 

 are lacking in Haida, case relations being expressed by postpositions, 

 and though what have been called cases exist in Tlingit these are 

 expressed bv suffixes, which differ from other j)o.stpositions in the lan- 

 guage only in the fact that they are single sounds. One of them, 

 moreover, occurs, with precisely the .same function, in an expanded 

 syllabic form. Tlingit nouns also take diminutive and intensive 

 suffixes absent from Haidsi, Init it will be shown later that they appear to 

 lie represented th(M-e in anothei' manner. Both languages distinguish 

 between an intimate, more or less reflexive, kind of possession and 

 one which is more remote, tending to become syntactic ; and the latter 

 is expressed in exactly the same manner by both, i. e., the objective 

 personal pronoun is prefixed to the noun and a special syllable suffixed. 

 To indicate intimate possession, however, Haida employs a distinct 

 suffix and dispenses with the pronominal prefix, while Tlingit employs 

 the ])i-oiioun and omits the suffix. Haida adjectives of shape and size 

 take a special jilural or distributive suffix not representee! in Tlingit, 



