The Seri-Salinan Connection Revisited

In 1925, Sapir proposed that Seri (Sonora, Mexico) and Salinan (central California) formed, with Chumash, a branch of the Hokan stock but did not provide any supporting argumentation. In 1956, Bright presented a small amount of data that in a preliminary way seemed to confirm that hypothesis, although these data and this presentation were not meant to be taken as the argumentation that was lacking. No further documentation has ever been produced, although the Seri-Salinan family reappeared in Greenberg's 1987 volume on languages of the Americas (without discussion), which was cited in the 1992 International Encyclopedia of Linguistics, which was the basis for the classification in the Ethnologue. In this paper, the Seri data are refined, the data from Bright (1956) are re-evaluated, and the proposal of a genetic relationship is put into doubt. It is suggested, in fact, that the hypothesis should be abandoned until proper evidence is presented.