1 7 6
��INTERNATIONAL JOURNAL OF AMERICAN LINGUISTICS
��VOL. I
��tive attempts Sapir 1 presented voluminous evidence for a genetic relationship between the Shoshonean, Piman, Sonoran, and Nahuatl languages.
From now on the comparative investigations shifted to the north. In addition to verbal announcements made on several occasions by Sapir of his belief that a genetic relationship may be ultimately established between the Takelma and Coos languages of Oregon, he wrote a preliminary paper 2 on the ultimate re- lationship between Athapascan, Haida, and Tlingit. In my own field I have collected a mass of material establishing a probable com- mon origin for the Kusan, Siuslawan, Yakon- an, and (perhaps) Kalapuyan languages which will be presented in the near future either in the conclusion to my grammatical sketch of the Alsea (Yakonan) language, 3 or else in a separate paper. I have, furthermore, gath- ered voluminous data supporting previously expressed contentions concerning the genetic relationship between Lutuamian, Wailatpuan, and Sahaptin, which will be published as soon as additional material from the Sahaptin field will be made available; and I have also good material for a comparative study of Salish, Chimakuan, and Wakashan. The latest ef- forts towards a re-classification of the Indian languages were made by Swanton 4 in the southeastern field which, however, are still in an experimental stage, although the Natchez- Muskhogean relationship would seem a prac- tically established fact. 5
To be sure, a number of these reconstruc- tions are by no means new ideas. They were formulated by previous investigators who, for lack of suitable data, could merely indicate but not follow them up minutely. Thus, the Uto-Aztekan relationship was suspected long
2 Southern Paiute and Nahuatl a Study in Uto- Aztekan (Journal de la SocietS des Amtricanistes de Paris, NS., vol. x, pp. 379-425; and AANS, vol. xvii, no. I, pp. 98-120; ibid. no. 2, pp. 306-328).
- The Nadene Languages, a Preliminary Report
(AANS, vol. xvii, no. 3, pp. 534-558).
��ago by Buschman and Brinton; similar ideas were held concerning Haida, Tlingit, and Ath- apascan by Boas and Swanton; marked re- semblances between Siuslawan and Yakonan were first noticed by Latham and Gatschet; Gatschet and Hewitt were convinced of a genetic relationship between Lutuamian, Wai- latpuan, and Sahaptin; and Boas long ago called attention to the marked structural agreements between the Salish, Chimakuan, and Wakashan languages. Thus, it will be seen that, in most cases, the younger linguists merely tried to follow up and develop the de- ductions arrived at by their predecessors.
Concerning the merits of the relationships that have been promulgated thus far in print, not all of them are tenable when subjected to the acid test of minute and scientific criticism. This is especially true of the several proposed reductions in California. Thus Dixon's and Kroeber's presentation of Penutian and Hokan are exceedingly inadequate, both methodo- logically and in regards to subject matter. The conclusions arrived at by these two stu- dents are based upon such fragmentary ma- terial and presented so scantily that, while their Penutian and Hokan relationships may be probable, they are by no means a certainty. Comparisons presented of five or more stems and of a few formative elements and not backed by exhaustive grammatical sketches, can not be accepted as conclusive evidence, and all judgment concerning the correctness of these conclusions must be withheld pending the introduction of additional and extensive data. During a recent visit to San Francisco I was accorded access, for a brief time, to the comparative vocabularies of the so-called Penutian and Hokan languages that have been
1 Now in course of publication by the Bureau of American Ethnology as part of volume ii of the Hand- book of American Indian Languages.
4 Linguistic Position of the Tribes of Southern Texas and Northeastern Mexico (AANS, vol. xvii, no. i, pp. 17-40).
'Ethnological Position of the Natchez Indians (AA) vol. ix, no. 3, pp. 513-528).
�� �
