408 



SCIENCE 



[N. S. YoL. LIV. No. 1400 



Dr. K. Or. Matheson, president of the 

 Georgia School of Technology since 1906, has 

 resigned to become president of Drexel Insti- 

 tute. Dr. Matheson will go to Drexel next 

 spring, probably April 1. Until then the 

 institute will continue to be directed by the 

 administrative board, which took charge upon 

 the recent retirement of Dr. Hollis Godfrey. 



Dr. Franklin Stewart Harris was in- 

 stalled as president of Brigham Young Uni- 

 versity at Provo, Utah, on October 17. Dr. 

 Harris, who was formerly director of the 

 Utah Agricultural Experiment Station, suc- 

 ceeds President George H. Brimhall, who 

 becomes president emeritus. 



Dr. Frank Pierrepoint Graves^ formerly 

 head of the school of education of the Uni- 

 versity of Pennsylvania, who succeeds Dr. 

 John H. Finley as commissioner of education 

 of New York State, and president of the Uni- 

 versity of the State of New York, was in- 

 ducted into office on October 20. 



Dr. Harry W. Crane, assistant professor 

 of psychology at Ohio State University, has 

 been called to an associate professorship in 

 psychology at the University of North Caro- 

 lina. He will also act as psychiatrist to the 

 State Board of Public Welfare. 



DISCUSSION AND CORRESPONDENCE 



A BIRD'S-EYE VIEW OF AMERICAN LANGUAGES 

 NORTH OF MEXICO 



It is clear that the orthodox " Powell " 

 classification of American languages, useful 

 as it has proved itself to be, needs to be super- 

 seded by a more inclusive grouping based 

 on an intensive comparative study of morpho- 

 logical features and lexical elements. The 

 recognition of 50 to 60 genetically inde- 

 pendent " stocks " north of Mexico alone is 

 tantamount to a historical absurdity. Many 

 serious difficulties lie in the way of the task 

 of reduction, among which may be mentioned 

 the fact that our knowledge of many, indeed 

 of most, American languages is still sadly 

 fragmentary; that frequent allowance must 

 be made for linguistic borrowing and for the 



convergent development of features that are 

 only descriptively, not historically, com- 

 parable; and that our persistently, and rather 

 fruitlessly, " psychological " approach to the 

 study of American languages has tended to 

 dull our sense of underlying drift, of basic 

 linguistic forms, and of lines of historical 

 reconstruction. Any genetic reconstruction 

 that can be offered now is necessarily but an 

 exceedingly rough approximation to the truth 

 at best. It is certain to require the most seri- 

 ous revision as our study progresses. Never- 

 theless I consider a tentative scheme as pos- 

 sessed of real value. It should act as a 

 stimulus to more profound investigations and 

 as a first attempt to shape the historical 

 problem. On the basis of both morphological 

 and, in part, lexical evidence, the following 

 sis great groups, presumably genetic, may be 

 recognized : 

 I. Eskimo-Aleut 



r Algonkin-Wiyot- Yurok 

 II. Algonkin-Wakasliaii J Kootenay 



1 Wakashan-Salish 



III. Na-dene (Haida; Tlingit-Athabaskan) 



r Calif ornianPenutian 

 rV. Penutian J Oregon Penutian 

 I Tsimshian 



Yuki 



Hokan 



Coahuiltecan group 



Keres 



Tunica group 



Siouan-Yuchi-Musko- 



V. Hokan-Siouan 



VI. Aztec-Tanoan 



giaa 

 Iroquois-Caddoan 

 r Uto-Aztekan 

 1 Tanoan-KSowa 



This leaves the Waiilatpuan-Lutuami-Sahap- 

 tin group, Zufii, and Beothuk as yet unplaced. 

 The lines of cleavage seem greatest between 

 IV. and v., and between III., on the one hand, 

 and I. and II., on the other. Group V is 

 probably the nearest to the generalized " typi- 

 cal American " type that is visualized by 

 linguistic students at large. 



E. Sapir 

 Canadian Geological Survey, 

 Ottawa 



