Page:IJAL vol 1.djvu/288

 THE HOKAN AND COAHUILTECAN LANGUAGES By E. SAPIR

In the general simplification of American Indian linguistic stocks which is at lasat being seriously undertaken by various investigators, two recently published articles are of particular interest. These are Kroeber’s Serian, Tequistlatecan, and Hokan and Swanton’s Linguistic Position of the Tribes of Southern Texas and Northeastern Mexico. The former adds to the Hokan stock recently determined by Dixon and Kroeber (Shasta-Achomawi, Chimariku, Karok, Porno, Yana, Yuman, Esselen; possibly also Chumash and Salinan), the Seri language of western Sonora and the Tequistlatecan or Chontal language of Oaxaca; the latter gives good evidence to show that a number of languages spoken along the Texas coast and back into the interior from it (Coahuilteco, Cotoname, Comecrudo; Karankawa; Tonkawa; and Atakapa), which have, according to Powell’s scheme, been classified into four distinct linguistic stocks, are best considered as genetically related. The full evidence for the validity of the Hokan stock has not yet been made public, but we have been promised it by Dixon and Kroeber. A comparative Hokan vocabulary insofar as it affects Yana has been kindly put into my hands by Dr. Kroeber; this, together with such descriptive or comparative grammatical and lexical Hokan material as has been published and such further comparative evidence serving to link Yana with Hokan as I have been able to gather from time to time, leaves small doubt in my mind of the correctness of the theory.

In going through Swanton’s comparative vocabularies, I was soon struck by a number of startling Hokan echoes. My interest having been actively aroused, I looked into the matter more carefully. The following comparative vocabulary of over a hundred stems and elements is the result. When we consider that only a very limited number of comparable terms were available for any two of the languages concerned, this result seems astonishing. It is difficult for me to suggest any alternative to the hypothesis of a common origin of the Hokan and Coahuiltecan languages. True, I have little morphologic evidence at hand, but the study of the problem thus newly opened up is confessedly in its infancy. As it is, the very imperfect sketch of Tonkawa given by Gatschet suggests a considerable number of Hokan-Tonkawa parallels in morphological elements.

In order not to complicate our problem, I have not listed in the table such Chumash and Salinan terms as seemed likely to be connected with Hokan words. These have been referred to in the notes to the vocabularies. A few Chumash-Coahuiltecan terms are noted at the end.

Kroeber’s, Dixon’s, Barrett’s, and Swanton’s