270
��INTERNATIONAL JOUKNAL OF AMERICAN LINGUISTICS
��VOL. I
��HOOK ; and Tlingit cat, TO TAKE, TO PICK UP, TO SEIZE. The difference in meaning should be sufficiently apparent without comment.
The pronouns show but one resemblance sufficient to warrant comment. Tlingit has a third person singular form dn. Certain of the northern Athapascan dialects have a reflexive third person possessive pronominal prefix de. Here, again, is a possible borrowing, which has been responsible for the limited distribution of the form in Athapascan.
Among the numerals, THngit has Lr.v' ONE, with which Athapascan Kato in ONE, is com- parable.
Professor Boas has succeeded beyond expec- tation in isolating and defining the etymolo- gical parts of the verbs. The adverbial prefixes are of the same general sort as are found in Athapascan, but among these there are no correspondences of note. Professor Boas lists as an incorporated noun if it SPACE, used in such expressions as qudil'iik" IT is WET (weather, soil). Compare with this Kato kou.'zns*l IT WAS
HOT.
Tlingit has a set of classifiers seemingly enti- rely lacking in Athapascan verbs which classify the subject or object solely by the limited application of the stem.
Morphologically, Tlingit is very similar to Athapascan. The nouns in both stocks seem to have been originally monosyllabic. To these primary nouns certain suffixes to form diminu- tives and augmentatives, etc., were added. The verbs are similar in structure, having elements of the same character which take the same general order. First are adverbial elements of direction and position, and pronoun objects. The stems are toward the end, and are preceded by the subject pronouns. In Athapascan there are modal elements, some of which precede the subject, and others follow. Tlingit has modal prefixes preceding the subject, but with classi-
��fiers following it. Both Tlingit and Athapascan have suffixes for customary action, etc.
The most striking resemblance is the fact that each has a modification of the stem itself, which affects in Tlingit the quality and pitch of the vowel, and in Athapascan the quality of the vowel and modifies the final consonant. These modifications of the stems are connected in both instances with differences in mode and tense.
With this striking likeness in morphology, one would expect lexical similarity leading to the definite conclusion that the languages were originally one, or sprang from the same source. The comparisons made of the lexical content, however, do not justify this conclusion. The similarities are few, forming but a slight percent- age of the whole. They might all be attributed to accident were there not at hand a more acceptable solution. The few nouns that are common are probably due to borrowing. It would be a remarkable thing if fully the number noted had not been borrowed in the course of the generations that Tlingit and Athapascan peoples have been neighbors.
The large majority of Tlingit monosyllabic nouns, stems, and other elements making up the verbs, the pronouns, post-positions, and adverbs, are totally different from any known Athapascan words or elements having a similar meaning. Until some satisfactory explanations can be given for this mass of apparently unre- lated material, a common genetic origin cannot be admitted. Were a genetic relationship to be assumed, one of three possible explanations must be accepted :
1. That changes in the forms of the words and in their meanings have been so great and so general, that resemblances have disappeared without leaving discoverable phonetic shifts.
2. That the original parent language from which Tlingit and Athapascan have sprung had such a complete double set of names for com-
�� �
