t K'-U


;j^ .rf

^1 [fi


Collected Works


Edward Sapir


The Collected Works of Edward Sapir Editorial Board

Philip Sapir Editor-in- Chief

William Bright

Regna Darnell

Victor Golla

Eric P. Hamp

Richard Handler

Judith Irvine


Collected Works


Edward Sapir


American Indian Languages


Volume Editor

Victor Golla


Mouton deGruyter

Berlin New York

Mouton de Gruyter (formerly Moulon. The Hague) is a Division of Waller de Gruyter & Co., Berlin.



@ Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability.

Library of Congress Cataloging in Publication Data

(Revised for vol. 2)

Sapir, Edward, 1884-1939.

American Indian languages.

(The Collected works of Edward Sapir : 5 6)

Vol. 1 edited by William Bright.

Vol. 2 edited by Victor Golla.

Includes bibliographical references and index.

1. Indians of North America Languages. I. Bright, William. 1928- . II. Golla, Victor. III. Sapir, Edward, 1884-1939. Works. 1990 ; 5-6. IV. Title. V. Series. PM108.S26 1990 497 89-13233

ISBN 0-89925-654-6 (v. 1) ISBN 0-89925-713-5 (v. 2)

Deutsche Bibliothek Cataloging in Publication Data

Sapir, Edward:

(The collected works]

The collected works of Edward Sapir / ed. board Philip Sapir

ed. -in-chief. ... Berlin ; New York : Mouton de Gruyter.

ISBN 3-11-010104-1 (Berlin)

ISBN 0-89925-138-2 (New York) NE: Sapir. Edward: [Sammlung]

6. American Indian languages. 2. Vol. ed. Victor Golla. 1991

ISBN 3-11-012572-2

(f) Copyright 1991 by Walter de Gruyter & Co.. Berlin 30.

All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Printing: Gerike GmbH, Berlin. Binding: Luderitz & Bauer, Berlin. Printed in Germany.

Edward Sapir, 1937

University of Michigan, Ann Arbor, Michigan, taken by Kenneth Pike

( Courtesy of Sapir family )

Edward Sapir (1884-1939) has been referred to as "one of the most brilliant scholars in linguistics and anthropology in our country" (Franz Boas) and as "one of the greatest figures in American humanistic scholarship" (Franklin Edgerton). His classic book, Language (1921), is still in use, and many of his papers in general linguistics, such as "Sound Patterns in Language" and "The Psychological Reality of Phonemes," stand also as classics. The development of the American descriptive school of structural linguistics, including the adop- tion of phonemic principles in the study of non-literary languages, was pri- marily due to him.

The large body of work he carried out on Native American languages has been called "ground-breaking" and "monumental" and includes descriptive, historical, and comparative studies. They are of continuing importance and relevance to today's scholars.

Not to be ignored are his studies in Indo-European, Semitic, and African languages, which have been characterized as "masterpieces of brilliant associa- tion" (Zellig Harris). Further, he is recognized as a forefather of ethnolinguistic and sociolinguistic studies.

In anthropology Sapir contributed the classic statement on the theory and methodology of the American school of Franz Boas in his monograph, "Time Perspective in Aboriginal American Culture" (1916). His major contribution, however, was as a pioneer and proponent for studies on the interrelation of culture and personality, of society and the individual, providing the theoretical basis for what is known today as humanistic anthropology.

He was, in addition, a poet, and contributed papers on aesthetics, literature, music, and social criticism.

Note to the Reader

Throughout The Collected Works of Edward Sapir, those publications whose typographic complexity would have made new typesetting and proofreading difficult have been photographically reproduced. All other material has been newly typeset. When possible, the editors have worked from Sapir's personal copies of his published work, incorporating his corrections and additions into the reset text. Such emendations are acknowledged in the endnotes. Where the editors themselves have corrected an obvious typographical error, this is noted by brackets around the corrected form.

The page numbers of the original publication are retained in the pho- tographically reproduced material; in reset material, the original publication's pagination appears as bracketed numbers within the text at the point where the original page break occurred. To avoid confusion and to conform to the existing literature, the page numbers cited in introductions and editorial notes are those of the original publications.

Footnotes which appeared in the original publications appear here as footnotes. Editorial notes appear as endnotes. The first endnote for each work contains the citation of the original publication and, where appropriate, an acknowledgment of permission to reprint the work here.

All citations of Sapir's works in the editorial matter throughout these vol- umes conform to the master bibliography that appears in Volume XVI; since not all works will be cited in any given volume, the letters following the dates are discontinuous within a single volume's references. In volumes where unpublished materials by Sapir have been cited, a list of the items cited and the archives holding them is appended to the References.


Frontispiece: Edward Sapir 6

Preface 13

Introduction to Volumes V and VI 15

Section Six: Athabaskan and Na-Dene Languages

Introduction 21

Notes on Chasta Costa Phonology and Morphology (1914) 27

Corrigenda to Father Morice's "Chasta Costa and the

Dene Languages of the North" (1915) 95

The Na-dene Languages, a Preliminary Report (1915) 105

The Sino-Dene Hypothesis

[excerpts from a letter to A. L. Kroeber] (1921) 133

Athabaskan Tone (1922) 141

A Type of Athabaskan Relative (1923) 143

The Phonetics of Haida (1923) 151

Pitch Accent in Sarcee, an Athabaskan Language (1925) 169

The Similarity of Chinese and Indian Languages (1925) 191

Review of Berard Haile, Manual of Navaho Grammar (1926) 193

A Summary Report of Field Work among the Hupa,

Summer of 1927 (1928) 195

The Concept of Phonetic Law as Tested in Primitive Languages

by Leonard Bloomfield [excerpt] (1931) 199

Two Navaho Puns (1932) 203

Problems in Athapaskan Linguistics 205

1 0 Contents

Review of A. G. Morice, The Carrier Language (1935) 207

Internal Linguistic Evidence Suggestive of the Northern Origin of the Navaho (1936) 209

Cornelius Osgood, The Distribution of the Northern Athapaskan Indians [contribution by Sapir]: Linguistic Classification within the Northern Athapaskan Area (1936) 221

Section Seven: Penutian Languages

Introduction 225

Preliminary Report on the Language and Mythology of the

Upper Chinook (1907) 231

Franz Boas, Chinook [contributions by Sapir] (1911): Diminutive and Augmentative Consonantism in Wishram Post-positions in Wishram Wishram Text and Analysis Modal Elements 243

A Characteristic Penutian Form of Stem (1921) 263

A Chinookan Phonetic Law (1926) 275

L. S. Freeland, The Relationship of Mixe to the Penutian Family

[with notes by Sapir] (1930) 283

(with Morris Swadesh) Coos-Takelma-Penutian Comparisons (1953) .... 291

Comparative Penutian Glosses 299

Section Eight: Wakashan and Salishan Languages

Introduction 319

The Rival Chiefs, a Kwakiutl Story Recorded

by George Hunt (1906) 323

Some Aspects of Nootka Language and Culture [excerpt] (1911) 353

Abnormal Types of Speech in Nootka (1915) 357

Noun Reduplication in Comox (1915) 381

The Rival Whalers, a Nitinat Story (1924) 435

Contents 1 1

Nootka Baby Words (1929) 465

Morris Swadesh (ed.), Salish-Wakashan Comparison (1949) 467

Section Nine: Other American Languages

A Tutelo Vocabulary (1913) 471

Review of B. Bibolotti, Moseteno Vocabulary and Treatises (1918) 475


A. G. Morice, Review of Sapir, Notes on Chasta Costa

Phonology and Morphology (1915) 481

A. G. Morice, Chasta Costa and the

Dene Languages of the North (1915) 485

A. G. Morice, Misconceptions Concerning Dene Morphology:

Remarks on Dr. Sapir's Would-be Corrigenda (1917) 499

E. Sapir, Corrigenda and Addenda to Takelma Texts (1914) 513

Phonetic Key to Publications of Edward Sapir 515

References 525

Index to Volumes V and VI 543


Volumes V and VI of The Collected Works of Edward Sapir are devoted to shorter works on American Indian languages (mainly of North America), including some previously unpublished material. Volume V, edited by William Bright, contains papers of a general nature on typology, classification, and phonetic notation, followed by work on Hokan languages, on the Uto-Aztecan family, and on the relationship of Algonkian, Wiyot, and Yurok. Volume VI, edited by Victor Golla, contains articles on Athabaskan and Na-Dene lan- guages, on Penutian, and on the Wakashan and Salishan families, plus two short papers on languages of other groups. Appendices in both volumes con- tain papers written by other authors which were discussed in papers by Sapir. A combined index to Volumes V and VI appears in the latter.

The editors of these two volumes have worked together in planning the entire sequence. Two possible ways of organizing the material were considered. One would be purely chronological, without considering topic; the other, adopted here, separates the articles into topical divisions and then arranges them chron- ologically within each division. This has the advantage, we believe, of making it easier for the reader to consult related papers in close proximity.

In addition to the articles contained in these two volumes, a number of arti- cles which discuss one or more specific American Indian languages appear in Volumes I through IV of The Collected Works. These are listed below, orga- nized by language or language group. The volume in which a paper is to be found is indicated by the appropriate roman numeral in brackets.

Athabaskan Languages: 1923c, A Note on Sarcee Pottery [IV]; 1924d, Per- sonal Names among the Sarcee Indians [IV]; 1933c, La realite psychologique des phonemes [I]; 1935b, A Navaho Sand Painting Blanket [IV]; 1936e, Hupa Tattooing [IV]; 1936h, Kutchin Relationship Terms [IV]; 1930, A Note on Navaho Pottery (with Albert G. Sandoval) [IV].

Comox: 1939e, SongsforaComox Dance Mask (edited by Leslie Spier) [IV].

Nootka: 1913b, A Girls' Puberty Ceremony among the Nootka Indians [IV]; 1915h, The Social Organization of the West Coast Tribes [IV]; 1919e, A Flood Legend of the Nootka Indians of Vancouver Island [IV]; 1933c, La realite psy- chologique des phonemes [I].

Southern Paiute: 1910d, Song Recitative in Paiute Mythology [IV]; 1933c, La realite psychologique des phonemes [I].

Takelma: 1907b, Notes on the Takelma Indians of Southwestern Oregon [IV]; 1907d, Religious Ideas of the Takelma Indians of Southwestern Oregon [IV].

14 VI American Indian Languages 2

Tsimshian: 1915g, A Sketch of the Social Organization of the Nass River Indians [IVj; 1920c, Nass River Terms of Relationship [IV]; 1921c, A Haida Kinship Term among the Tsimshian [IV].

Yana: 1908a, Luck-Stones among the Yana [IV]; 1916g, Terms of Rela- tionship and the Levirate [IV]; 1918j, Yana Terms of Relationship [IV]; 1922d, The Fundamental Elements of Northern Yana [IX]; 1923m, Text Analyses of Three Yana Dialects [IX]; 1928j, The Unconscious Patterning of Behavior in Society [III].

Volumes VII-XV, which contain Sapir's work of monographic scope on American Indian languages and cultures, also include some shorter, closely related articles containing lexical inventories and textual analyses. Note that Sapir s Language: An Introduction to the Study of Speech (Volume II) cites some thirty American Indian languages, and his 1916 monograph, Time Per- spective in Aboriginal American Culture (Volume IV), one-third of which is devoted to "evidence from linguistics," cites dozens of American Indian lan- guages or language groups. It should also be noted that all references to specific languages in each article are listed in the indices of each individual volume, as well as in the comprehensive index in Volume XVI.

Preparation of this volume was supported in part by grants from the Phillips Fund of the American Philosophical Society, the National Science Foundation (grant no. BNS-8609411), and the Andrew W. Mellon Foundation.

The editor also acknowledges the contributions to the preparation of this volume by Jane McGary and the help of Dr. Marie-Louise Liebe-Harkort, editor in chief of Mouton de Gruyter.

Introduction to Volumes V and VI

It has often been said that Franz Boas is to be considered the father of anthro- pological linguistics in North America, and in particular the initiator of serious research on American Indian languages. But surely Edward Sapir, who began his career as a student of Boas, became the most influential scholar of the twen- tieth century in both these fields . Consider the diversity of the Native American languages on which Sapir did original research Chinook, Takelma, Yana, Southern Paiute, Nootka, Sarcee, Navajo, and others; or the language families in which he did ground-breaking comparative work Hokan, Uto-Aztecan, Algonkian, Athabaskan, and Penutian; or the types of studies he carried out descriptive, historical, comparative, ethnolinguistic, and what would now be called sociolinguistic. Even before his untimely death, Sapir's achievements were monumental; after 1939, his stature as an Americanist only grew, as many of the materials he left in manuscript were edited and published by his students. His stature grows yet more in subsequent volumes of these Collected Works, with the publication of several major collections of texts (Sarcee, Kutchin, and Hupa) and other important longer manuscripts, now edited by students of his students.

It is possible to attempt some general comments about the overall course of Sapir's work on North American Indian languages as it is reflected in the pres- ent pair of volumes. Publications from the period 1906-1910 are primarily descriptive, including the first results of field work on Wishram Chinook, Ta- kelma, and Yana. In 1911, typological interest emerges in "The Problem of Noun Incorporation in American Languages" (1911c) and is pursued most nota- bly in the two reviews (1917k, 19171) of works by Uhlenbeck. Comparative lin- guistic research, aimed at establishing relatively remote linguistic relationships on the basis of both lexical and grammatical comparisons, comes to the fore in 1913 with "Southern Paiute and Nahuatl, a Study in Uto-Aztekan" (19131, 19151) and "Wiyot and Yurok, Algonkian Languages of California" (1913h). During the following half dozen years, Sapir's enthusiasm for tracing remoter rela- tionships is manifest in such papers as "The Na-Dene Languages" (1915d), "The Hokan and Coahuiltecan Languages" (1920b, written in 1915), and "A Charac- teristic Penutian Form of S-tem" (1921b, written in 1918). This interest reached its culmination in a drastic proposal to reduce 58 North American "stocks" (as formulated by John Wesley Powell in 1891) to just six "great groups. " This classi- fication, based on grammatical and typological rather than lexical corre- spondences, was presented in a lecture at Chicago in 1920 (the notes for which are published here in "Materials Relating to Sapir's Classification of North American Indian Languages"). With little change, this formed the core for Sapir's influential Encyclopaedia Britannica article on "Central and North

15 VI American Indian Languages 2

American Languages" (not published until 1929, 1929a). After the early 1920s, Sapir s interest in these problems seems to have cooled; however, his last major work in this genre, "The Hokan Affinity of Subtiaba in Nicaragua" (1925b), argues for a Central American extension of the far-flung Hokan (-Coahuiltecan) group, and presents what is perhaps Sapir's most detailed argu- ment for the importance of "submerged" structural features in recognizing remote linguistic relationship.

Sapir's sixfold classification and the methodology supporting it constituted, during his lifetime, the most controversial part of his work on North Amer- ican languages (it was never accepted, for instance, by his onetime teacher Boas). It should be remarked, however, that what Campbell and Mithun (1979: 26) have called the "reductionist zeal" of this classification was not unique to Sapir. Large-scale genetic regrouping of North American languages was initiated by Alfred L. Kroeber and Roland B. Dixon, who, in a series of papers beginning in 1913, proposed assigning most of the Powellian language families of California to one or the other of two new "stocks," Penutian and Hokan (Dixon and Kroeber 1913, 1919). Sapir joined in this work only after the groundwork had been laid, and at Kroeber's urging (GoUa 1986: 178). Sapir brought to the task a thorough famiharity with the methods and data of Indo-European comparative philology, and after a brief period of skepticism he became convinced that a rigorous application of philological principles to American languages would yield important new insights. He moved from one bold synthesis to another, and his comprehen- sive classification of 1920 must be regarded as little more than a report on work in progress. It is noteworthy, however, that Sapir did relatively little after 1920 either to support or to revise that classification. His 1925 paper on Subtiaba, while introducing some new structural arguments for Hokan, is based on essentially the same group of cognate sets as in his earlier work, and it refers only briefly to the larger Hokan-Siouan grouping introduced in his 1920 lecture.

In contrast with the wide-ranging comparative work that had absorbed him during the preceding decade, Sapir s research during much of the 1920s focused narrowly and intensively on one group of languages: the "Na-Dene" stock of his 1920 classification (comprising Tlingit, Haida, and the widespread Athabaskan family). As early as 1906 he had worked briefly, during his Takelma field work, with a speaker of Chasta Costa, an Oregon Athabaskan language; in preparing this material for publication (1914c), he saw Athabaskan as a family having the diversity and the relatively good documentation to make it a match for his skills as a comparativist. He was soon embroiled in controversy with older Athabaskan scholars (e.g. , Father Morice, 1915c, see Volume VI and Appendix to Volume VI); this was exacerbated by his 1915 proposal (1915d) of a genetic relationship among Athabaskan, Tlingit, and Haida. Sapir concluded that only through extensive field work of his own could he hope to accumulate the evi-

Introduction 17

dence necessary to convince his critics. His feeling about the necessity of such work became even stronger when, around 1920, he came to suspect that an intercontinental genetic connection between Na-Dene and Sino-Tibetan was a distinct possibility.

Sapir's plan for Na-Dene field research was extraordinarily ambitious, and it was never completed. Except for a foray into Haida phonetics (1923d), his field work was entirely devoted to Athabaskan, involving four major inves- tigafions: Sarcee, in 1922; Kutchin and Ingalik, in 1923; Hupa, in 1927; and Navajo, principally in 1929. Only the Sarcee work is significantly represented in Sapir's bibliography; even here the major published study was prepared in collaboration with his student Li Fang-Kuei (Li 1930, see Volume XIII). A good deal of the material collected by Sapir has been published post- humously, but the definitive grammar of Navajo which Sapir planned (and was working on even during his last illness) will never be written. Of his comparative insights into Athabaskan, Na-Dene, and Sino-Dene, we have only fragmentary notes.

Sapir's active research career extended from 1905 to 1938, or 33 years. During the first two decades of this period until his move from Ottawa to a teaching post at the University of Chicago he was engaged almost exclusively in Amer- ican Indian research, the bulk of it descriptive linguistics. After 1925 his inter- ests began to turn toward other types of study, particularly the psychology of culture; and his linguistic field research virtually came to an end when he moved from Chicago to Yale in 1931. He remained, nonetheless, a central figure in American Indian linguistics, second only to Boas in status and pre-eminent in intellectual influence. Nearly all his important students took up the study of American Indian languages. It was left to them, and to their scholarly progeny in turn, to continue the many facets of his research. We will do no more here than mention the names of Harry Hoijer, Morris Swadesh, George Trager, Stanley Newman, Li Fang-Kuei, Benjamin L. Whorf, Charles F. Voegelin, and our own teacher, Mary Haas. All these scholars have transmitted to their own students not only an enthusiasm for American Indian linguistics, but, even more important, Sapir's commitment to the study of language within the broad- est context of human understanding.

William Bright Victor Golla

Section Six: Athabaskan and Na-Dene Languages


While Sapir began his involvement with Hokan and Penutian languages as a field linguist, turning to comparative and classificatory studies only after com- pleting several major descriptive works, with Athabaskan the reverse was the case. Except for some unsystematic notes on Chasta Costa collected during his Takelma field work in 1906 (published as 1914c; see below), Sapir had had no first-hand experience with Athabaskan languages when he launched what came to be the most absorbing of his historical hypotheses: the genetic relationship of Athabaskan, Tlingit, and Haida, and the deep connection of this "Na-Dene" family to the Sino-Tibetan stock.

Neither the Na-Dene hypothesis nor the possibility of connections with Asi- atic languages were, strictly speaking, original to Sapir (see the discussion of earlier speculation in Krauss 1973), but there is little doubt that Sapir was the first to explore these questions in the light of modern comparative linguistics. The evidence from Sapir's correspondence and manuscripts is that he probably took up this work late in 1912 or early in 1913, shortly after completing a detailed study of Uto-Aztecan (Sapir to Kroeber, 23 December 1912, in GoUa 1984: 71). In the ensuing months he apparently read through all the extant mate- rial both on comparative Athabaskan and on Tlingit and Haida. In the spring of 1913 we find him complaining in a letter to Robert Lowie that Pliny Earle God- dard, a leading student of comparative Athabaskan, was over-cautious:

He does not . . . seem to me to get very much beyond descriptive Athabascan sketches cast in parallel grooves. The unifying reconstructive spirit, the elimination of secondary features and the emphasis on essential ones, seem to be lacking, on the whole. It seems to me that with all the experience that Goddard has had with Athabascan he would have felt irresistably drawn by this time to a serious consideration of Haida and Tlingit as possibly genetically related, though remotely, to Athabascan. This may not seem eventually as far-fetched as it does now. But I am afraid that Goddard is rather timid in these matters. (Sapir to Lowie, March 1913.)

Three months later he wrote to A. L. Kroeber:

A propos of larger linguistic units, which seem to be somewhat in favor just now, I may say that I have been occupying myself of late with Athabaskan, Tlingit, and Haida, and that I have collected enough evidence to convince myself at least of the genetic relationship of these three . (Sapir to Kroeber, 30 May 1913, in Golla 1984: 104)

It was during this period that he went back to his Chasta Costa notes and worked them up into a short monograph with a distinctly comparative emphasis. Notes on Chasta Costa Phonology and Morphology (1914c). The publication of this work immediately brought him into the small circle of serious Athabaskanists, of whom the most established and productive was the French-Canadian Oblate priest, A. G. Morice. Father Morice reviewed Sapir s Chasta Costa monograph in glowing terms (Morice 1915a), but followed his review with a more critical

22 V7 American Indian Languages 2

appraisal, "Chasta Costa and the Dene Languages of the North" (Morice 1915b). Sapir replied to Morice's strictures in "Corrigenda to Father Morice's 'Chasta Costa and the Dene Languagesof the North'" (1915c), to which Morice gave rejoinder in "Misconceptions Concerning Dene Morphology: Remarks on Dr. Sapir s Would-be Corrigenda" (Morice 1917). Morice s review and papers are reprinted in the Appendix to this volume. In the assessment of at least one scholar, "in the long run . . . the interaction between Morice and Sapir turned out to be productive," with the two showing "grudging respect for each other" (Krauss 1986: 153-154). The relationship, stormy or not, was essentially between two generations of scholars, and between a man thoroughly familiar with the concrete details of several Canadian Athabaskan dialects and a com- parativist primarily interested in historical reconstruction. As the nature of his Athabaskan work shifted in the 1920s and 1930s from comparative to descrip- tive, Sapir s appreciation of Morice's extensive knowledge of Athabaskan struc- ture grew, as is evidenced by his short but appreciative review (1935c) of Mor- ice's massive grammar and dictionary of Carrier (1932).

Between 1913 and 1915 Sapir continued to devote much of his research time to the Na-Dene project, combing the published documentation of Northern, Pacific Coast, and Southwestern Athabaskan languages for material to be com- pared with Swanton's descriptions of Haida and Tlingit. By 1915 he had amassed about 300 lexical comparisons and had begun writing a "systematic presentation" of the material (Sapir to Radin, 17 July 1918, quoted in Krauss 1986: 156; see Dallaire 1984: 169, letter no. 263). At this point, at Goddard's request, he prepared the shorter paper reprinted here, "The Na-Dene Lan- guages, a Preliminary Report" (1915d), iox pub\\cdi\\on'mi\\Q American Anthro- pologist. The manuscript of the full study has unfortunately been lost, although the ledgers in which Sapir entered his Na-Dene lexical comparisons have sur- vived (manuscript 497.3 B63c Na20a.3, vols. 1, 3, and 4, American Philosoph- ical Society Library).

Even after the publication of Sapir's evidence Boas and Goddard remained skeptical about Na-Dene, and Boas challenged Sapir's methods in a heated exchange at the annual meeting of the American Anthropological Association the following December. The strength of Boas's opposition (which reached print in 1920 in a scathing attack on the misuse of genetic classification), together with the appearance of Boas's own descriptive study of Tlingit (Boas 1917), seems to have taken the wind out of Sapir's sails, at least temporarily. He devoted hardly any time to Athabaskan or Na-Dene from 1916 through 1920, which was in general a period during which he was more occupied with literary and artistic matters than with linguistic research.

Late in 1920, as he was completing his general book Language (1921d), Sapir experienced what he described to Kroeber as a "considerable recrudescence of interest in linguistics" (Golla 1984: 347), particularly in classificatory work. On the one hand, this led to Sapir's working out of a general classificatory scheme for all North American languages, grouping most of them into six "great

Six: Athabaskan and Na-Dene Languages 23

Stocks" (1921a). But it also led him to reconsider Na-Dene. As he put it to Kroeber:

I am just now interested in another big linguistic possibility. I tremble to speak of it, though I've carried the germinal idea with me for years. I do not feel that Na-dene belongs to the other American languages. I feel it as a great intrusive band that has perhaps ruptured an old

Eskimo-Wakashan-Algonkin continuity In short, do not think me an ass if I am seriously

entertaining the notion of an old Indo-Chinese offshoot into N.W. America. ... I have already carefully gone over two Tibetan grammars (Jiischke and Foucaux) and find in Tibetan pretty much the kind of base from which a generalized Na-dene could have devel- oped, also some very tempting material points of resemblance. (Sapir to Kroeber, 4 October 1920, in GoUa 1984: 350; reprinted in volume V: 81-83).

In the ensuing months Sapir delved deeply into Chinese and Sino-Tibetan lin- guistics, working to some extent under the guidance of the anthropologist and orientalist Berthold Laufer. By the end of the summer of 1921 Sapir had developed the outlines of his "Sino-Dene" hypothesis, and he discussed the matter at some length in a letter to Kroeber (a copy of which was also sent to Laufer), dated October 1, 1921, and printed here almost in its entirety as "The Sino-Dene Hypoth- esis." The complete text of this and a short follow-up letter can be found in Golla (1984: 374-384).

More pressing than the need to acquaint himself thoroughly with the Asiatic side of the relationship, however, was the need Sapir felt for a more complete and accurate documentation of Athabaskan, Tlingit, and Haida. Seeing them now through a Sino-Dene lens, it was clear to Sapir that many important aspects of Na- Dene phonology and grammar had been missed by previous investigators. Late in 1921, he drew together extensive comparative evidence on the morphophonology of relativization in Athabaskan. This paper, "A Type of Athabaskan Relative" (1923n), calls attention to the distinctive clause-like nature of Athabaskan poly- synthetic structure and, as Sapir wrote to Kroeber, "insidiously prepares for far bigger things than its ostensible theme" (Sapir to Kroeber, 24 November 1921, in Golla 1984: 386).

Especially important in Sapir's eyes was accurate information on the presence of tonal systems, which the Sino-Dene connection made him certain was fundamen- tal throughout Na-Dene. A pitch accent had been reported for Tlingit by Boas (1917), and Sapir himself, working briefly with a speaker of Haida in March 1920, had noted tonal and other phenomena of potential historical importance in Haida, later described in "The Phonetics of Haida" (1923d). Tone had not, however, been definitely reported in Athabaskan, and Sapir was certain that this was an over- sight. To his delight, the first Athabaskan language he had the opportunity to work on, Sarcee, turned out to have a well developed system of pitch accent. In an immediate announcement in the American Anthropologist, "Athabaskan Tone" ( 1922a), he flatly stated that in the light of his Sarcee observations "it is well nigh inconceivable that [tone] should be absent in any other Athabaskan dialect." This view was reiterated in Sapirs full analysis, "Pitch Accent in Sarcee, an Athabaskan Language" (19251), which is less a descriptive study than a general theory of Athabaskan tone illustrated with Sarcee data. The paper, moreover.

24 VI American Indian Languages 2

ends with a list of questions that Sapir felt could be answered satisfactorily only from the standpoint of Na-Dene (1925f: 204-205).

His wife's declining health forced a postponement of Sapir s planned visit to the Hupa in 1923, but he was able to spend much of the summer of that year doing productive field work with two young Alaskan Athabaskans, speakers of Ingalik (Anvik) and Kutchin, who were working at a camp in Pennsylvania. In the turmoil following his wife's death early in 1924, and his subsequent move from Ottawa to Chicago in 1925, Sapir carried out little Athabaskan work dur- ing the next two years. That the Sino-Dene hypothesis still strongly attracted him, however, is shown by an interview he gave to Science shortly after arriving at the University of Chicago, printed under the title "The Similarity of Chinese and Indian Languages" (1925o).

From the beginning of his teaching at Chicago until his death 14 years later, Sapir's Sino-Dene research had to vie for time with his many other involve- ments. The evidence of his manuscripts is that he did little further with the larger historical questions, although he continued the serious study of Tibetan and Chinese. Certainly his publications after 1925 show little direct concern with the Sino-Dene relationship, or even with Na-Dene, except insofar as the stock was represented in his general classification of Central and North Ameri- can languages (1929a). The major area in which he continued his comparative research was Athabaskan. "A Summary Report of Field Work among the Hupa, Summer of 1927" (1928i) and the Athabaskan portions of "The Concept of Phonetic Law as Tested in Primitive Languages by Leonard Bloomfield" (1931b) give brief glimpses of Sapir's progress in working out the intricate details of comparative Athabaskan phonology in the late 1920s. (In the latter paper, written in 1928-29, Sapir equates his Athabaskan work with Bloomfield's Algonquian in a methodological discussion.) Even here, however, his work slowed considerably in the following decade. He offered a course in Com- parative Athabaskan twice during his teaching at Yale (in 1931-32 and again in 1936), and from the students' notes that survive there is little evidence that Sapir's views had evolved much after 1930. A short survey of "Problems in Athapaskan Linguistics" found among Sapir's papers, apparently dating from about 1932 and published here for the first time, contains little not found in his earlier published work. The statement of "Linguistic Classification within the Northern Athapaskan Area" (1936i), which Sapir provided Cornelius Osgood for inclusion in his ethnographic survey of the Northern Athabaskan Indians, is also unremarkable.

From a brief foray into Chasta Costa in 1906 through major field work on Navajo beginning in 1929, Sapir collected a very large corpus of descriptive data on several Athabaskan languages. It was the analysis of this material that came more and more to occupy Sapir's attention after 1925, much of it done in collab- oration with his students, Fang-Kuei Li and Harry Hoijer (also, more briefly, Mary Haas), and with his Navajoist colleague. Father Berard Haile. Only frag- ments of this work saw print during Sapir's lifetime, and some has remained unpublished to the present day.

Six: Athabaskan and Na-Dene Languages 25

The materials from Sapir's Sarcee field work in the summer of 1922, his first extensive synchronic study of an Athabaskan language, have been more fully published than most later materials. Sapir himself, in addition to preparing a largely comparative paper on the Sarcee tone system (1925f, see above), had nearly completed a volume of Sarcee texts when he left Ottawa in 1925. This manuscript is being published for the first time in Volume XIII of The Collected Works. Also in that volume are reprinted two studies based on Sapir s mate- rials: Fang-Kuei Li's "A Study of Sarcee Verb-Stems" (1930), originally written as a masters thesis at the University of Chicago under Sapir's direction, and Harry Hoijer and Janet Joel's "Sarsi Nouns" (1963). Sapir also wrote two short papers on ethnographic aspects of his Sarcee work, "A Note on Sarcee Pottery" (1923c) and "Personal Names among the Sarcee" (1924d); both are found in Volume IV.

Sapir's Kutchin materials, collected from John Fredson at Camp Red Cloud, Pennsylvania, during the summer of 1923, are much more poorly represented in his published work. Sapir extracted the kinship terms for inclusion in Cornelius Osgood's Ethnography of the Kutchin (Sapir 1936h, printed in Volume IV), but otherwise published nothing of his Kutchin data. As with Sarcee, he had begun preparing a volume of Kutchin texts while still at Ottawa, and during the 1930s Mary Haas, as his research assistant, worked on a Kutchin stem list. In 1961-62 Victor Golla completed a preliminary stem list but did not publish it. The texts and a stem list are being published in Volume XIII of The Collected Works.

The Anvik (Ingalik) notes that Sapir obtained from Thomas Reed, also at Camp Red Cloud in 1923, are far less extensive than his Kutchin materials. Essentially a wordlist, of no great descriptive or comparative interest, the material was never utilized by Sapir and is not published in The Collected Works.

Sapir collected extensive Hupa data during a northwestern California field trip in the summer of 1927, during which he also worked more briefly on Yurok and Chimariko. He was accompanied by his Chicago student, Fang-Kuei Li, who carried out his own work on Mattole and Wailaki. (For Sapir's lively description of this trip see Volume IV [1927b].) Other than a short "summary report" of his linguistic findings (1928i) containing, inter alia, the (to Sapir) distressing information that Hupa lacks a tonal system and the Hupa data incorporated into papers on the comparative method in American Indian lin- guistics (1931b) and on the northern origin of the Navajo (1936f, sec below), Sapir published only one paper based on his Hupa work, a largely ethnographic study of Hupa tattooing (1936e) written for A. L. Kroeber's Festschrift. Sometime during the 1930s he began work on a volume of Hupa texts, with extensive ethnographic notes, but it was far from complete at the time of his death. The texts and notes, edited by Victor Golla, are being published in Vol- ume XIV of The Collected Works, together with a lexical index to Sapir s data.

Sapir regarded his Navajo work, begun with a native speaker in Chicago in 1926 but largely carried out in the field in 1929 and in later collaboration with Father Berard Haile, as "by far the most extensive and important linguistic

26 yf American Indian Languages 2

research" he ever accomplished (Sapir to Boas, 12 April 1938, quoted in Krauss 1986: 166). The data he collected, particularly lexical and paradigmatic mate- rial, was extraordinarily rich, and his relationship with his principal consultant, Albert G. (Chic) Sandoval, was especially close. Sapir prepared a large collec- tion of texts for publication , and his correspondence indicates that he was plan- ning to write a full Navajo grammar. In the years immediately preceding his death, he was actively working with Father Haile, a Franciscan missionary and scholar (see Sapir's review of Haile's earlier work [1926f]), in preparing literacy and language teaching materials for Navajo (Krauss 1986: 164-166). Despite all of this activity, at the time of his death in 1939 Sapir had in fact published hardly anything based on his Navajo work. A short paper, "Two Navaho Puns" (1932d), and a comparative-historical tour dc force, "Internal Linguistic Evidence Sug- gestive of the Northern Origin of the Navaho" (1936f), nearly exhaust the list, except for two brief ethnographic notes, "A Navaho Sand Painting Blanket" (1935b) and "A Note on Navaho Pottery" (with A. G. Sandoval, 1930), and a newspaper article describing the circumstances of the 1929 field trip (1929c). A volume of Sapir s Navajo texts was seen through the press in 1942 by Harry Hoijer, Sapir's principal Athabaskanist student, who much later published a Navajo grammar based on Sapir's materials as The Phonology and Morphology of the Navajo Language (Sapir and Hoijer 1967). These two publications are reprinted in Volume XV of The Collected Works, together with extracts from the voluminous correspondence on Navajo linguistics that Sapir and Father Haile carried on between 1929 and 1938 (Berard Haile Collection, University of Arizona Archives, Tucson). Besides work directly attributed to Sapir, the influ- ence of Sapir's materials and interpretation is strong in at least two other pub- lications: Father Haile's Learning Navaho (1941-48), the first volume of which incorporates many of Sapir's insights into Navajo phonology; and Harry Hoi- jer's A Navajo Lexicon (1974), which faithfully reproduces the organization of Sapir's lexical files.



In a large part of southwestern Oregon and contiguous territory in northwestern CaUfornia were spoken a number of apparently quite distinct Athabascan dialects. The terri- tory covered by tribes or groups of villages speaking these dialects embraced not only a considerable strip of Pacific coast^ but also much of the interior to the east (Upper Umpqua and Upper Coquille rivers, lower Rogue river, Chetco creek and Smith river) ; some of the tribes (such as Tolowa and Chetco) were strictly coast people, others (such as Galice Creek and Umpqua or J\kwa}) were confined to the interior. While some of the Athabascan dialects spoken south of the Klamath in California, particularly Hupa and Kato, have been made well known to students of American linguistics, practically nothing of linguistic interest has as yet been published on any of the dialects of the Oregon- California branch of Pacific Athabascan. It is hoped that the following imperfect and fragmentar}'- notes on one of these dialects may prove of at least some value in a preliminary way.^

' Outside of a few points in southern and southeastern Alaska (Cook Inlet, mouth of Copper river, Portland Canal) this is the only region in which Athabascan tribes have found their way to the Pacific.

* My ^ denotes nasalization.

* The material for these notes was secured in a very incidental manner. While the writer was at work on Takelma in the latter part of the summer of 1906, he was living with Mr. Wolverton Orton, a full-blood Chasta Costa Indian. At odd moments Mr. Orton and the writer whiled away the time with Chasta Costa.


28 VI American Indian Languages 2


The Chasta Costa (or Cis/ta qlwAs/ta) Indians, now gathered in Siletz Reservation in western Oregon, formerly occupied part of lower Rogue river; between them and the coast were other Athabascan tribes or villages of practically identical speech, above them to the east were the unrelated Takelma.'' Among these tribes of nearly or quite identical speech were the YiV^/gwl or Euchre Creek people, the Tee' /me dA/ne or "Joshuas" of the mouth of Rogue river, the Du/Vil dA/nl, the Ml/klu/nu"* dA/nl, and the GwA/sd. All these formed a linguistic unit as contrasted with the coast people {d/yds/ta "lower tribes") or, as they are now commonly called by the Indians of Siletz, "Sol Chuck" Indians, a Chinook Jargon term meaning "salt water, coast" people; the dialect of these coast tribes was probably identical to all intents and purposes with Chetco. While Chasta Costa and Coast Athabascan are thus more or less distinct, they seem to have been mutually intelligible without very much difficulty, the coast dialect sounding merely somewhat "strange" and "drawn out" to a speaker of Chasta Costa. At least three other Athabascan dialects of this region, however, seem to have differed so much from Chasta Costa as to be but partly understood, if at all^ by speakers of the latter; these are Upper Umpqua, Upper Coquille, and Galice Creek.

* It has already been pointed out (American Anthropologist, N. S., 9, p. 253, note 2) that there is reason to believe that J. O. Dorsey was incorrect in assigning the Chasta Costa villages above those of the Takelma (see his map in Journal of American Folk-Lore, III, p. 228). On p. 234 Dorsey gives a list of Chasta Costa villages.

Six: Athabaskan and Na-Dene Languages 29



The vowels of Chasta Costa are a, a, e (open as in Eng- lish met), e (long and open), o (close as in German Sohn), 0, u (apparently variant of o), u, i (generally open), I, and A (like u of EngHsh hut)\ 6 (short and open as in German voll) sometimes occurs after velars as variant of o {sxo/ld "five," cf. Hupa^ tcwo/la), a (as in English hat) occurs after velars as variant of e {tsxci/xe "child," cf. Carrier^ cezkhehkhe "chil- dren").

Vocalic quantity is of considerable importance in Chasta Costa, not so much etymologically as phonetically. On the whole, long and short vowels interchange on regular mechan- ical principles; open syllables (that is, syllables ending in a vowel) with long vowel regularly shorten this vowel when the suffixing of one or more consonants to the vowel makes the syllable closed. Examples of a thus varying with a are:

do/ydc/tla "I won't fly;" do/ydt/tla "we won't fly" (cf.

dd/yd/t!a "he won't fly") dad /da "he is sitting down" (cf. da/ 6 Ad /da "I am sitting

down") tcIdsL/se "he cries;" tcldch/se/Ve "I shall cry" (cf. tcla/Bil/se

"you cry") tc!a/ydsL/se "they cry" (cf. tc!a/yd/6il/se "we cry") nac/tlb "I swim" (cf. nd/tcll/tlo "you bathe")

' Hupa examples are taken from P. E. Goddar^l, "The Morphology of the Hupa Language," Univ. of Cal. Publ. Amer. Arch, and Ethn., 3.

* Carrier examples are taken from Rev. A. G. Morice, "The D^n^ Languages," Trans- actions of the Canadian Institute, I, pp. 170-212.