Wick R. Miller

The Classification of the Uto-Aztecan Languages Based on Lexical Evidence

0. Introduction. There has been a notable lack of agreement among informed scholars on the classification of the Uto-Aztecan languages. The problem revolves around the family-tree approach versus the wave or mesh approach (see Bloomfield 1933:311-18 and Swadesh 1959). The family-tree approach assumes sudden splits within a dialect-free parent, while the wave approach assumes a dialect continuum which dissolves into distinct languages and in which the newly budded languages reflect the earlier dialect interrelationships. The wave principle operated to a greater extent in Uto-Aztecan than some other families (e.g., Indo-European). The vexing and interesting problems for Uto-Aztecan are two: first, to what extent did the wave principle operate; and second, how are we to describe or represent the relationships that are difficult or impossible to represent by the traditional family-tree classification? The Uto-Aztecan family consists of about thirty languages, located in two main geographic areas: the northern one in southern California, the Great Basin, and nearby areas; and the southern one stretching from southern Arizona, through northwest Mexico, into central Mexico and beyond (see fig. 1). Those favoring greater importance for the family-tree approach recognize three branches: Shoshonean, Sonoran, and Nahuatl or Aztecan. A variant of this approach would group Sonoran and Aztecan into a single branch called Southern Uto-Aztecan (SUA), with Shoshonean then renamed Northern Uto-Aztecan (NUA). Those favoring greater importance for the wave approach view Uto-Aztecan as being composed of eight or more independent branches. The so-called Shoshonean, then, is viewed as consisting of four branches, Sonoran of three or more (the particular number varying somewhat among different investigators), with general but not universal agreement by both groups that Aztecan forms an independent branch.

Preaspiration and Gemination in Central Numic

The Numic (Uto‐Aztecan) languages are well known for consonant gradation, which each language shows to some degree. Three consonantal series have been reconstructed for Proto‐Numic: Geminating, Nasalizing, and Spirantizing. The Central Numic languages Timbisha, Shoshoni, and Comanche have preserved these three consonantal series and added a fourth, Aspirating. The Aspirating series is historically derived from the Geminating series, but it is synchronically distinct from it. On the basis of verb class behavior in Central Numic, we show that the Central Numic Aspirated series is a straightforward consequence of Proto‐Uto‐Aztecan stress patterns as reflected in pre‐Proto‐Central Numic.

Proto-Keresan Phonology

1. Keresan is spoken in seven varieties at seven Indian Pueblos in New Mexico. Five of the Pueblos, Cochiti, Santo Domingo, San Felipe, Santa Ana, and Zia (listed from north to south) are located in or near the Rio Grande Valley. The two remaining Pueblos, Acoma and Laguna, are situated about seventy five miles to the southwest of the main Keresan area. The languages, or dialects, are closely related, and have a time depth of about five hundred years. No sharp internal subdivisions are found in the Keresan language family. The greatest differences are found between the two most distant Pueblos, Acoma and Cochiti; the two dialects are mutually unintelligible except to speakers who have had an opportunity to become accustomed to the speech of the other Pueblo.1 It is possible to reconstruct the phonological structure of Proto-Keresan in great detail. The changes from the proto-language to the daughter languages are minimal, and reconstructed forms do not differ radically from the present-day forms. But since the Keresan languages have a complex morphophonemic system, it is possible to do some internal reconstruction, and achieve a deeper time depth. The reconstruction of Proto-Keresan is based on the comparison of Acoma, Santa

A Note on Extinct Languages of Northwest Mexico of Supposed Uto-Aztecan Affiliation

A number of languages in northwest Mexico, for which we have little or no linguistic data, became extinct in the early colonial period. Most of the extant languages in this area are Uto-Aztecan, and thus it is often assumed that many or most of the extinct languages were also Uto-Aztecan: a reasonable assumption, but one which does not automatically make them Uto-Aztecan. The problem has received passing attention from a number of scholars, with a tendency for one scholar to utilize (with or without full citation) the results of an earlier one without careful checking. I reviewed the situation recently, while preparing a paper on Uto-Aztecan (Miller, forthcoming b). It became clear that one of the few (perhaps the only one) to evaluate the original sources carefully was Sauer (1934). It would do no harm to reexamine the original materials, but this would require a look at archival material in a number of places, a task for which I have neither the means nor taste. But a careful reexamination of Sauers article proved rewarding. This note is based principally on that work, along with an occasional look at other material.

Agent in Passive Sentences in Yaqui and Guarijío

FLEMING, ILAH, AND RONALD K. DENNIS. 1977. Tol (Jicaque) phonology. IJAL 43:121-27. GRANBERRY, JULIAN. 1987. Antillean languages and the prehistoric settlement of the Bahamas: a working hypothesis. Proceedings of the Bahamas 1492 Conference, Freeport, November 1987, ed. Charles A. Hoffman (in press.) , AND GARY S. VESCELIUS. Ms. a. The toponymic method and the derivation of Taino morphemes. , AND GARY S. VESCELIUS. Ms. b. A lexicon and grammar of the Taino language. HAHN, PAUL. 1960. The Cayo Redondo culture and its chronology. Ph.D. dissertation, Yale University.

Lionnet's Article on the "Intensive" in Tarahumara

The Jesuit missionary linguists of Chihuahua have made Tarahumara one of the better documented Uto-Aztecan languages. Their publications are well known to Uto-Aztecanists, but a remarkable article by Andres Lionnet (1968) seems to have gone unnoticed, one in which the author reconstructs an old reduplication pattern rendered opaque because of recent sound changes. Lionnet internally reconstructed the pattern, and then through the use of a 1683 grammar by Father Tomas Guadalajara achieved a still greater time depth. With the availability of data from the closely related Tarahumaran language Guarijio, I shall do the opposite; I shall trace the pattern from Proto-Tarahumaran to the present. Initial consonants are lost if the second syllable begins with the same consonant: