AFFILIATION OF THE ALGONQUIN LANGUAGES. 21 



while in the Algonquin it generally precedes it. There are, however, 

 suffix particles that take the place of adjectives in the latter class, and 

 in most cases they are represented by verbs. The Malay-Polynesian 

 adjectives are often hard to distinguish from substantives and verbs. 

 The sign of comparison precedes the adjective in Algonquin, but 

 follows in Tonga. But the accusative or object of the verb follows it 

 in both Algonquin and Polynesian, and this separates them from the 

 Turanian languages. Tense is designated by special marks in each 

 case. These are Algonquin perfect ki, gi, future ka, ga; in Tonga 

 present gooa, perfect na, future te. A larger acquaintance with 

 Algonquin and Malay-Polynesian forms might reduce the differences 

 between these. In the Tonga the index of tense is placed before the 

 personal pronoun which precedes the verbal root, e.g., makee, give; 

 na-oo-makee, / gave; na-ger-makee, thou gavest; te-oo-makee, / shall 

 give; te-ger-makee, thou wilt give. In Algonquin the temporal 

 indices come between the pronoun and the verbal root, e.g., makew, 

 give; ni-ki-makew, / gave; ki-ki-makew, thou gavest; ni-ki-makew, 

 / shall give; ki-ka-makew, thou wilt give. In spite of the difference 

 in the order of pronoun and temporal index, the two classes agree in 

 placing both these before the verbal root, thus entirely disagreeing 

 with the Turanian languages in their Ural-Altaic and Dravidian 

 divisions. The possessive pronoun or its equivalent precedes in the 

 Algonquin, and either precedes or follows in the Malay-Polynesian 

 languages. These languages also agree in dispensing with the relative 

 pronoun. The forms of the demonstrative in Cree and Tonga are not 

 unlike ; Tonga, this aheni, that ahena ; Cree, this anah, that na.ha. 

 The same is true of the interrogative ; Tonga ahai, coeha who, which; 

 Cree awewe, kekway. The Polynesian languages have an article, 

 and have on account of it been affiliated with the Bantu or Caffre 

 languages of Southern Africa. Duponceau and other writers have 

 insisted that the initial M of many Algonquin nouns, which generally 

 precedes those that are not in a construct state, is the article. Others 

 as firmly deny the statement, but have not accounted for the frequent 

 dropping of this letter, e.g., mistik, a tree; meyw-atik, a good tree; 

 much-atik, a had tree; face, mikwakun; my face, ni-kwakun. Un- 

 doubtedly there is some analogy here with the common Bantu prefixes 

 mo, ma, me, and the Tagala article ang. The Caffre analogies, apart 

 from language, with the Algonquins are striking. One important 

 point of resemblance between the Algonquins and the Malay-Poly- 



