swAXToxl THE TLINOIT AND HAIDA LANGUAGES 475 



and in the same laiiuuafje we tind a suffix with ronnectives indicatinjf 

 motion in the particular ])lace ju.st referred to. Both have a distributive 

 suffix used after luuuerals and a few other words. 



In the verb we tind mai'ked differences. Tlingit verbs have so few 

 affixes that, includintr a pi'etixed auxiliary and three substantives very 

 frei[ueiitiy used as preti.ves, so far 27 only have been enumerated, 

 while in Haida there are as many as 115. Their distribution with 

 reference to the stem is also very unequal. Of the Tlingit affixes 18 

 come before the stem and 9 after it; in Haida 60 precede and .55 fol- 

 low. In actual usage, however, the diti'erence is still more marked. 

 In Tlingit more than one suffix is rarely found in the same verb, and 

 suffixes are often entirely wanting, wiiile the ratio of suffixes to 

 prefixes in any one Haida verb is at least as two to one. 



Coming to the varieties of ideas expressed by affixes, we find in 

 Haida, bt-sides 16 affixes of miscellaneous character, a series of mor- 

 phological affixes, an instrumental seiies, locatives, temporals of two 

 orders, modals, and 10 suffixed auxiliaries. In Tlingit, however, the 

 first three are entirely lacking, locatives being expressed by adverbs, 

 and auxiliaries are represented by the causative only. This is as much 

 as to say that the bulk of Tlingit affixes are temporal and modal. 

 Nouns are taken up into the Haida verb much more easily than into 

 the Tlingit verb, and the combination of two verb stems, which is very 

 conmion in Haida, is unknown in the other language. This great 

 divergence between the verl)s is mitigated only by the close reseml)lance 

 among several Haida affixes and the consequent possibility that thej' 

 have been difl'erentiated from one original, and, by the great independ- 

 ence of those affi.xes, suggesting word composition rather than a real 

 subordination of elements. 



■1. Sfrucftt/v of xentenre. In Haida the verb — or a breathing repre- 

 senting it — almost invarialily stands at the end of the sentence, but, 

 while this is commonly the case in Tlingit, the nominal subject or 

 oiiject or a noun and postposition frequently follows. Adverbs pre- 

 cede the verb, usually inunediately, in Iwth languages, and are closely 

 relat(>d to the postpositions, of which each language has a great 

 abundance. 'I'hey also agree in placing adjectives after the nouns they 

 depend upon and usually in placing the nominal subject before the 

 nominal object. More imjwrtant is the fact that the pronominal sub- 

 ject in Haida and the pronominal subjective prefix in Tlingit follow 

 the corresijonding objective forms. 



5. Lexical Kiiinlar (ties. Having considered the kinds of phonetics 

 used by Tlingit and Haida, the tj'pes of stems and roots, and the com- 

 position of words and sentences, we come to the most important point 

 of all — actual similarities in form. This, w hich is ordinarily the first 

 crit(>rion appealed to, is resei'ved until tlie last, because such similari- 

 ties are confessedly scanty and because such as exist nmst be viewed 



