1. Introduction. The name" Penutian" was coined by AL Kroeber and Roland B. Dixon in their ambitious and influential classification of California Indian languages (Dixon and Kroeber 1913) as a label for the genetic relationship that they proposed among five Central California" stocks"-Costanoan, Miwok, Maiduan, Yokuts, and Wintuan. 1 This grouping was based primarily on lexical evidence; the term" Penutian" derives from two stems for the numeral'two'(pen and uti), one or the other of which Dixon and Kroeber found in all five stocks. In the final presentation of their material Dixon and Kroeber (1919: 55-69) published a list of 171" cognate stems" and some other lexical resemblances, from which they derived a set of rudimentary sound correspondences. In his 1921" Bird's-Eye Classification" of North American Indian languages Sapir-who had begun working on comparative Penutian as early as 1915-extended the Penutian relationship (and the term) to a number of other languages and families to the north of California. In 1929 he added Mixe-Zoque and Huave (" Mexican Penutian") to this construct. Other scholars have subsequently used" Penutian" as a label for even broader proposed relationships (most recently Joseph Greenberg 1987, who subsumes under it the Mayan, Yukian, and the Gulf families among with the California and Northwestern stocks). In most contexts, however-certainly in most textbook treatments of North American linguistic classification-the Penutian relationship is the one outlined by Sapir in his 1920-21 and 1929 statements. In its earlier version (Sapir 1920; 1990: 98) it is as follows: