Skip to main content

Full text of "The collected works of Edward Sapir"

See other formats


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. — 

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 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 

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 

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 

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 

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- 

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 

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. 


30 ^f American Indian Languages 2 


An example of c shortened to e is: 

7ies/ts!Al/l "I am seen" (cf. lie' /ts!Al/l "he is seen") 

Original long vowels may lose their quantity even in an 
open syllable, provided they are immediately followed or 
preceded by a syllable with relatively strong accent. Such 
are tela-, na-, and 7ie- in: 

tc!a/ya/dil/se "we cry;" tc!a/ydsL/se "they cry" 

ne/7id/ts!Al/l "we are seen" 

Id na/dit/t!d "don't bathe;" {na/dit/t!d is phonetically 

enclitic to strongly accented ld\ contrast 7id/dit/t!d/Ve 

"you will bathe") 

In general, however, stress accent cannot be said to be 
particularly well marked in Chasta Costa." Each syllable 
is a fairly well-defined phonetic unit tending to hold its own 
against others, so that an approximately level accentual flow 
with but few peaks results. Such writings as 7id/dit/t!d and 
tc!d/dil/se, with apparent accent preceded by long vowels, 
are doubtless but imperfect renderings of forms with level 
stress on first and second syllables (they might perhaps better 
be written nd/dU/t!d and tc!a/Bll/se with secondary accent 
on second syllable). It does not seem that every vowel in an 
open syllable is organically long; thus e in future -Ve and in 
-de of Vwl/de "everything" is regularly short. Many such 
cases are, however, probably only apparent, the short vowel 
being followed by a glottal stop; thus plural ya- of ya/dAl/ni 
"they make a sound" should doubtless be ya-. 

Short a of closed syllables is regularly reduced from long 
a; original short a becomes /i in a closed syllable. Examples 
of A thus dulled from original a are: 

VAc/yAc/Ve "I shall go" (cf. t't/dic/ya "I go;" -yAc = 
Hupa -yauw) 

' Weak stress accent seems characteristic of Athabascan generally. Father Moric6 
goes so far as to say, "there is no accent in Den^" {op. cit., p. 173). 

Six: Athabaskan and Na-Dene Languages 31 


nd/xAn/do "eight, two less" {;nd/xA- = Kato'^ nqk/ka" 

do/na/yAct/xwl "I do not vomit" (cf. na/yd/dAdt/xwf 

"I vomit") 
VAl/dAc "he runs" {-dAc= Hupa -dauw) 
jAn/na/'Ac "he will bring" {-'ac = Hupa -aiiw) 
Ve/A7i/yit/lAl "we are sinking" (cf. Ve/nit/lat' "we drown;" 

Hupa -lat, -la "to float") 

Not to be etymologically confused with this a is inorganic 
A. Whenever a consonant is not followed by a definitely 
determined vowel and yet, for some reason or other, is not 
phonetically appended to the preceding syllable, it must begin 
its own syllable and takes an inorganic, in other words ety- 
mologically meaningless, ^ -vowel after it. This syllable may 
either be .completed by a consonant of etymological value 
(such as first person singular c, verb class signs I, t, I) never 
followed by a definite vowel or, if it is immediately followed 
by a syllable beginning with a consonant, this consonant is 
borrowed to complete the inorganic syllable (-/ closes inorganic 
syllables preceding d-, tl-, dj-, tc!-, ts!-, tdf-, tc'-, lI-), so that 
a doubled consonant results of which the first half is of no ety- 
mologic significance. In some cases, however, as before 7-. 
and in rapid speech generally, this inorganic consonant is not 
always distinctly heard; yet in syllabifying words Mr. Orton 
completed such inorganic syllables with a consonant with 
mechanical regularity. These syllables with inorganic vowel 
and consonant are characteristic not only of Chasta Costa 
but also of Hupa and Kato and doubtless other Athabascan 
dialects as well. The general phonetic tendency to speak in 
definite syllables and the further tendency to limit short vowels 
to closed syllables explain these characteristic Athabascan 

* Kato examples are taken from P. E. Goddard, "Kato Texts," Univ. Cal. Publ. 
Amer. Arch, and Ethn., 5, 65-238; and "Elements of the Kato Language," ibid., II, 

' -a- may be secondarily lengthened from -a-. 

32 VI American Indian Languages 2 


developments. The quality of the inorganic vowel varies 
for different Athabascan dialects; it is a{u) in Kato as well 
as in Chasta Costa, i {u before voiced or voiceless w, il or e 
before post-palatal ^-sounds) in Hupa, apparently e in Galice 
Creek, (s (probably identical with our a) in Carrier. Chasta 
Costa xAt/VAl/lal "they sleep" is etymologically equivalent 
to x/V/lal; X-, third person plural prefix, cannot stand alone 
and is therefore followed by a and / borrowed from -/'-, while 
-/'- (verb prefix V- reduced from Ve-, therefore not capable of 
combining with x- into xaV-) in turn needs a syllabifying a 
followed by / borrowed from -lal. Other examples of inorganic 
A, with and without following inorganic consonant, are: 

VAc/yAc/Ve "I shall go" {t"A- = V- reduced from Ve-) 
dd/ya/xAt/t!a "they won't fly" {xAt- = x-) 
nd/xAt/dAl/nic "they work" {xAt/dA- = x/d-, d- reduced 

from de-) 
dd/xAn/uAt/VAc "they go to bed" {xAn/nAt- = x/n-) 
t'e/An/yAl/Ul "he is sinking" (jaI- = y-) 

Many syllables with final consonant and a- vowel must 
be considered as radical or at least unanalyzable elements. 
In not all such cases is ^ a reduced form of a; where a seems 
a primary vowel, as shown by comparison with other Atha- 
bascan dialects, it seems best to consider it an organic element 
in the syllable, though it remains plausible that at last analysis 
it is but a reduced form of some fuller vowel. Thus, while 
-yAc has been shown to represent an original -yac (Hupa -yaiiw), 
-Vac contains a primary a, as shown by comparison with Hupa 
-tuw "to lie down" (ultimately -Vac is doubtless -V, reduced 
from -Ve, and suffix -c). 

Inorganic a sometimes becomes palatalized to i, though 
there is not enough material available to make it certain just 
when this change takes place. Examples of this secondary 
i have been found before c (but not before its developments 
s and 6) and 5 derived from tc (but not before original 5 or its 

Six: Athabaskan and Na-Dene tankages 33 


development) when itself preceded by m, n, or 6 (preceding 
7, however, tends to preserve a). Examples are: 

mis/ki'' "gull" (cf. Kato butc/k'ai') 

nic/ya "I come" {nic- = cessative 7z- and first person 

singular c) 
nic/dac "I dance" 
t'e/nic/lat "I drown" 
t'e/dic/ya "I go" {6ic- = durative 6- and prone minal c; 

cf. t'ed/ya "he goes" without vowel after 6) 
tclAy/ye/6ic/ya "I eat" 

ye/ die /I "I saw him" (cf. c/ycd/l "he saw me") 
Bid/ si "I let him" 

With -die- contrast -OaO- (both from original *-sac-) in da/ 6 Ad /da 
"I am sitting;" with -0zc/- contrast -sasI- (from original -*sAd- 
and -*sAd- respectively) in tcId/sAsi/se "I am crying." -yic- 
was heard in yd/yic/tia "I fly," but as this is an isolated example 
(contrast -yAc- in nd/da/yAct/tlo "I bathe" and -yAd- in yAd/Az 
"I sneezed"), it seems possible that this form was misheard 
for yd/yAc/tla. Besides -nic- also -nAc- is met with: dd/nAc/t'Ac 
"I go to bed" and nd/nAc/ An "I stop him;" it is probable that 
in these forms -nA- is a reduced form of ne- (cf. Hupa tcin/ne/tuw 
"she goes to bed") and thus not directly comparable with -ni- 
of -nic-. Unaccented A, itself reduced from a, has in one case 
{-yAc "to go") been found further palatalized to i: do /Vac/ 
yic "I'll not go," Id/n/yic "don't go!" (cf. t'Ac/yAc/t'e "I 
shall go"); this -yic contracts with directly preceding t'A- 
into -Vac: do/ Vac "he won't go.^^ 

Original Athabascan ai has in Chasta Costa become 
monophthongized to i. Examples are: 

l/gl "white" (cf. Kato i/gai) 

'" Should probably be miskH'. 

''With ihis-t'Ac Kato ta/cac in dd/la/co* ta/cac "not anywhcrt' I went" (P. E. God- 
dard, "Kato Texts," Univ. Cal. Puhl. Amer. Arch, and Ethn., 5, No. 3, p. 182, 1. 17) is in 
striking agreement. 

34 yf American Indian Languages 2 


mis/kfi"- "gull" (cf. Kato biUc/k'ai') 
/zp'-^ demonstrative "that" (cf. Hupa hai) 

ail as organic diphthong seems to occur but rarely in Atha- 
bascan. If do "no!" (cf. Hupa dau) may be regarded as distinct 
from adverbial do "not" (cf. Hupa do), we would have an 
example of the parallel development of au to o in Chasta Costa. 
Certain contractions that take place between i of first person 
plural -it- and second person plural -o- with preceding vowels 
will be spoken of in discussing the pronominal prefixes. 

One of the most striking phonological characteristics of 
Chasta Costa is the disappearance of an original rj^^ or of its 
representative, nasalization of preceding vowel. Its former 
presence can always be proved by comparison with other 
Athabascan dialects that, like Hupa, still preserve it. In the 
case of all vowels but inorganic A nasalization has left no trace 
whatever, original q, (from arj), ^ (from et]), and i (from Iry) 
being reduced to a, e, and t; originally short vowels, on losing 
their nasalization and thus coming to stand in an open syllable, 
become lengthened, while originally long vowels in a closed 
syllable not only lose their nasalization but are shortened. 
Thus, a syllable si may represent an original sj (or sir]) or 
si (or sir]), while sil may go back as well to sll as to sjl. Examples 
of the absolute disappearance of an original 17 are: 

nd/xe "you paddle" {nd- = *nq-, cf. Hupa nun/ya "you 

are about") 
do/yCl/tla "you won't fly" {yd- = ^yq,-, cf. Hupa yum/ mas 

assimilated from *yun/mas "you are rolling over") 
tc!dl/se/t'e "you will cry" {tcfal- = *tc!q-l-; cf. tc!dd/se/t'e 

"I shall cry" with -c- "I" morphologically parallel 

to -.- "you") 

'* t is here shortened to i because of following glottal stop. 

•' i* denotes long I with weakly rearticulated parasitic i. Such "pseudo-diphthongs" 
sporadically occur in Chasta Costa in lieu of ordinary long vowels. 
'* i. e., ng of English sing. 

Six: Athabaskan and Na-Dene Languages 35 


id/na/yat/xwl "don't vomit!" {yat- = *yqt-, cf. yd- from 

*yq- in na/yd/dit/xwl "you are vomiting") 
nel/l "you are looking at him" {nel- = *n^l- ; -I = -'?*, 

cf. Kato -in' "to see") 
It "dog" (original Athabascan */?, "^li-q; cf. Hupa Lzw, 

Montagnais I'in, Hare //'/w, Loucheux /'m, Carrier 

//, old form loe^'n}^) 

Nasalized inorganic 4 seems to have acquired a palatal 
coloring i\ this i then regularly developed to I in open, i in 
closed syllables. It thus often seems as though Chasta Costa 
I, i is the morphologic equivalent, for instance in second person 
singular forms, of Athabascan 77, an equivalence, as has just 
been shown, due to secondary phonetic developments. Examples 
oil < i < 4 are : 

Ve/Bi/ya "you go" {61- = *S4- ; cf. Hupa na/sin/ya "you 

are going about") 
ni/dac "you dance" {ni- = *W4-; cf. Hupa nin/yauw 

3'w/u'w f/x/wi "you whistle" {di- = *d4-\ ci. Hupa da /din/ La 

"run!" assimilated from * da /din/ La 
na/tcll/ilo "you swim" {tcli-^'^'kylA-'}^ cf. Hupa na/kin/- 

yun "come eat!") 
yd/yi/tla "you fly" (7t- = *74-; cf. Hupa ye/win/ya "you 

are going in") 
yd/yi/tla "it flies" (7^- = *74-; cf. Hupa na/win/tau "it 

will settle down" assimilated from ^na/win/tau) 
Vi/lal "you are sleeping" (/'I- = */'4-; cf. Hupa tin/xaiiw/ne 

"you take along") 
verb stem -si "to make" (cf. Hupa -tcwiii) 

'^ Morice, op. cit., p. 210. Carrier has evidently undergone a development parallel 
to that of Chasta Costa. All northern Athabascan forms except Carrier (and Chipewyan) 
are taken from R. P. E. Petitot, " Dictionnaire de la langue Dcnc-Dindji^." 

'* k^l is "fortis" palatal k, Hupa ^1, Morice's q. 

36 VI American Indian Languages 2 


Examples, in closed syllables, of i<i<4 are : 

tc!d/Bil/se "you cry" {9il- = *s4-l-; cf. Hupa na/dil/we/- 

na/dit/t!d/Ve "you will bathe" {dit- = *d4-t-', cf. third 

person nd/dAt/t!d/t'e) 
t'd/yit/nd "you drink" {yU = *y4-t-; cf. third person 

yd/yil/gAd "you climb" {yil- = *y4-l-; cf. third person 

t'il/xwAd "you cough" (tHl- = *t'4-l-; cf. third person 

ne/cli/l "look at me!" {cil- = *c4-l-) 

Hupa -;7 (that is, our rj) seems at times to correspond to 
Chasta Costa -?z, but comparison with northern Athabascan 
dialects indicates that in such cases we are dealing with original 
-n. Thus, ?iAn "you," despite Hupa nin, is shown to have 
original -n by Montagnais nen and Loucheux 7ia7i; dAn/tdi 
"four," Hupa dink {=dir)k^!), does not go back to original 
*d4/k^!i but to *dAn/k^!i or *dA7]/k^!i (rj assimilated from n), 
as evidenced by Loucheux tan; la/cAn "black" corresponds to 
Loucheux del-zen; similarly, dAn "in, at" must have original -n 
despite Hupa din and Kato dun (original *d4 would have given 
Chasta Costa *dl). 


The consonantal system of Chasta Costa, like that of most 
Athabascan dialects, is characterized by a lack of labial stops, 
though m is common; b has been found in bo/di "cat," a loan- 
word from English pussy, but seems not. to occur in native 
words (yet cf. tcA/pd/yu "flower"). The consonants of Chasta 
Costa are: the labi'al nasal m; the dental stops /', d, t!, and 
dental nasal 7t; the back stops g, q' (or qx), q!, voiceless spirant 

" In Hupa n (or nasalization) disappears in closed syllables. In such forms Chasta 
Costa is etymologically more transparent than Hupa insofar as -i- is a reflex of original 
-A-, whereas Hupa -t- is the normal inorganic vowel. 

Six: Athabaskan and Na-Dene Languages 37 


X (as in German Bach) , and voiced spirant y (as in North German 
Wagen) ; the labiahzed back stops k'w, gw, q!w, and spirant xw 
(sometimes weakened to hw) ; the sibilants s, c (as in EngUsh 
ship), 6 (as in Enghsh thin), and z (voiceless lenis, intermediate 
between ^ and English z, heard in -az "to sneeze"); the affric- 
ative palatal consonants tc\ dj, and tcl; the affricative alveolar 
consonants ts, tsl, and affricative dental consonant tdl; the 
laterals /, I (voiceless spirantal /, with l, dorsal t followed 
by /, as variant), and l/; the glottal stop ('); the aspirate h 
(' at the close of a syllable) ; and the semivowels y and w. 

Of these /', q\ k'w, and /c' (English ch) are aspirated surds 
(^' is not found, k'w has been found but once and may be con- 
sidered of doubtful occurrence); (b), d, g, gw, and dj are voice- 
less but lenis, intermediate acoustically between surds and 
sonants^^ {dj is intermediate between English ch and j) ; //, 
ql, td, ts!, td!, and l! are so-called "fortis" consonants, in other 
words, they are pronounced with simultaneous closure of glottis 
but are released before the release of the glottal chords. q\ 
q!, gw, and q!w {g has not been found, but very likely exists) 
are velar consonants; k! has not been found, ^^ its place being 
taken by g/.^° Of secondary origin are syllabically final / and 
k, which may be considered as voiceless stops differing from 
/' and k' in their lack of aspiration; they are etymologically 
equivalent to d and g. It is highly probable that also w, which 
does not frequently occur, is but a secondary development 
or acoustic variant of y after o-vowels;-^ after o-vowels y be- 
comes labialized to y'^, in which both y and w elements are so 
weak that one is constantly in doubt as to whether he hears 

'" It is possible that these "intermediate" stops are sonant at their moment of releass. 

" Unless, as seems possible, k of mis/ki "gull" was misheard for k!. 

^° ql corresponds to Hupa kz, g is Hupa k-z. q! is by no means as forcible a sound as 
is, e. g., Chinookan q!. There is something decidedly illusive about it; the velar stop 
element seems to be reduced to a minimum, the glottal catch element is very strongly 
marked, and a weak x seems at times to precede the velar stop (e. g., '^qld/xAO "arrow"). 
Despite my familiarity with Chinookan q!, I did not often succeed in pronouncing Chasta 
Costa ql so as to satisfy Mr. Orton's car. It may well be that ql is really "fortis" or glot- 
talized x (x/); cf. Tlingit 5/. 

"' In Hupa 7 has become w in every case. 

38 VI Atnerican Indian Languages 2 


7 or w (thus do/ ye- becomes do/y'^e-, dd/'^we-\ similarly, 
what was heard as do/wa- may really be do/y'^a-). However, 
w occurs also in sa' /wAs/ts!e "sandhill crane;" wAs/xe "good." 

This consonant system is only in part a faithful repre- 
sentative of the original Athabascan system. Some conso- 
nants have become merged with others, while other consonants 
have kept distinct but have been changed in regard to place 
of articulation. Chasta Costa m, t\ d, tl, n, g, q! (kl), q!w, 
y, I, i, lI, ', h, and y seem in practically every case to correspond 
to these same Athabascan sounds. 

Athabascan k\ as also in Hupa, has become x in Chasta 
Costa : 

xd'/tc'ii "goose"-^ (cf. Hupa xa-^; Applegate Creek k'q'/- 

tc'u; Kato k'a') 
nd/xi "two" (cf. Hupa nax; Montagnais nak'e^^) 
ts!d/xe "woman" (cf. Carrier tsekhe-'") 
tsxd/xe "child" (cf. Carrier cezkhehkhe) 

Analogously to this change of k' to x, original Athabascan 
k'w has become xw (sometimes heard as hw) in Chasta Costa. 
This sound is preserved as such in Kato {k'w) and Chasta 
Costa (yXw), but seems generally to have fallen together in other 
dialects with original k\ Examples are: 

hwd "foot" (cf. Kato kwe*"; Carrier ne-khe; Loucheux 

na/yd/dAdt/xwi "I vomit" (cf. Carrier khu "vomiting") 

It seems, however, to persist as k'w in: 

k'wAs/Vd/7ie "six" (cf. Hupa xos/tan) 

Etymologically but not phonetically distinct, both in 
Hupa and Chasta Costa, from these secondary x and xw are 

^ -Ic u is augmentative. 

" See Goddard, "Kato Texts," note 32. 

" Petitot's ' represents aspiration. 

** Father Morice represents "fortis" stops by means of points belovv characters. 

Six: Athabaskan and Na-Dene Languages 39 


original Athabascan x and xw. A good example of the latter 

-xwAd "to cough" (cf. Carrier xwces "cough," as noun) 

Athabascan sibilants and sibilant affricatives {ts and 
tc sounds) have undergone various modifications in Chasta 
Costa. Original s has regularly become 6: 

Ba/^aI "grizzly bear" (cf. Carrier sces-e^oel "brown bear") 
t'e/dic/ya "I go" (cf. Hupa te/se/ya/te "I am going away") 
VeB/ya "he goes" (cf. Hupa tes/ya/te "it is about to come") 
-gAB "to climb" (cf. Hupa -k2s) 
-xwaB "to cough" (cf. Carrier xwces) 

Before / (or its variant l), however, 5 is regularly retained: 

ts!a/sASL/se^^ "I cry;" tddsL/'se "he cries;" tc!a/ydsL/se 
"they cry" (with these forms contrast tc!d/Bil/se 
"you cry") 
na/yesL/sl "he tells" (contrast nd/BU/sl "you tell") 
cAsl/sl "he lets me" (contrast Bid /si "I let him") 
cAsl/Val "he kicks me" (contrast Bid /Veil "I kicked him") 
qlwAi/dasL/na "it was lying on it" 

Athabascan is would, by analogy, have been expected to 
develop into IB (as in Chipewyan), but B seems to be regularly 
found instead: 

Bi "head" (cf. Carrier n-tsi "your head;" Montagnais 
-ihif Hare -kjwi\ Loucheux -tchir^. Kato -sV "head" 
seems to indicate that in Kato also, at least initially, 
5 and ts fell together. 

Ba/^^cl "hair of head" (cf. Montagnais ethi-pa^^) 

^* -SASL- is assimilated from *-sacl-, -s- being here prevented from becoming -6- be- 
cause of following -5- (before l) of same syllable. 

" i. e., -tei. Petitot's th is td. In Hare /5 (or its reflex 16) developed into what Petitot 
writes kfw, perhaps to be understood as k(t>, i. e., k plus bilabial/. 

'' Petitot's tch is our Ic. 

** Petitot's p is 7. 

40 VI American Indian Languages 2 


l/do "yellow, green" (of. Montagnais del-thop "yellow;" 
Hare de-kfwoy "yellow," Hupa u't-tso "green;" Kato 
L-tso "blue")' 

In some cases ts seems to have become s: 

se "stone" (cf. Kato se\ Hupa tse; Montagnais the; Hare 
kfwe; Loucheux tchi; Carrier tse) 

As might be expected, Athabascan ts! has regularly become 
td! in Chasta Costa: 

dd/de/dil/tB!i "we are sitting" (cf. Hupa na/ya/del/tse, 

i. e., -tsle, "they lived as before") 
teiAd/dd "story" 

Athabascan c is normally preserved as such (e. g., ct "I"). 
However, it is assimilated to s before 5 and tsl: 

s/tsli/de "my sickness" {c- "my") 

nes/tsUi/l "I am seen {-c- "I") 

As/se/t'e "I shall cry" (from *ac-) 

s I tsl An I na I 'Ac "he will bring it to me" (c- "me") 

Assimilation of *sac to sas has taken place in: 

tc!a/sAsL/se "I cry" (cf. tddci/se/Ve "I shall cry") 

Original *sAc>*sic, however, regularly developed to die: 
tclAy/ye/eic/ya "I eat" 

Original *sac, after being assimilated to *sas, regularly shifted 
to OaB, unless, as we have seen, it was protected by immediately 
following I: 

da/ 6 AS /da "I am sitting" (from *dd/sAc/da) 
Ve/BAB/lal "I have been sleeping" (from *re/sAc/lal) 
tcfed/t.'d "I swim across" (probably misheard for tcIe/dAd/tfo) 

Original s, when immediately following c, also causes it to 
assimilate; ss, which thus results, is then regularly shifted to 

yd/yA6/6el "I threw" (from '^yd/yAc/sel) 

Six: Athabaskan and Na-Dene Languages 41 


Athabascan tc (sometimes /cw?) is not retained in Chasta 
Costa, but appears regularly as s: 

l/sAk "red" (of. Kato L/tcik\ Loucheux ditssigY^ 

mis/k{!)i(') "gull" (of. Kato butc/k'ai*) 

sd'/wAs/tsIe "sandhill crane" (cf. Applegate Creek tcd'/- 

-si "to make" (cf. Hupa -tcwin; Kato -tcl; Chipewyan 

-se "to cry" (cf. Chetco -swe; Hupa -tcwen; Kato -ke'; 

Carrier -ssd) 

Chasta Costa sx is found in: 

sxo/ld "five" (cf. Hupa tcwo/la; Chipewyan sa/so/la/yai^) 

Athabascan tc! remains, tc! often being shifted, however, 
to ts! (or 5'22) : 

tele- verb prefix "across the water" (cf. Hupa tee-, i. e., 
td.e-, "down to the beach, out of the house;" Kato 
tce-\ Chipewyan ts'e- "to a body of water") 

-ts!An "toward, to" (cf. Hupa -tcin, i. e., -tdit); Kato 
-tc'uiV; Chipewyan -ts'un) 

ts!i/de "sickness" (cf. Loucheux tssik, i. e., tsUk) 

-s'at' "to be hurt" (cf. Hupa -teat, i. e., -te!at, "to be sick, 
to become ill") 

There is still another set of sibilants in Chasta Costa, 
which go back to original palatalized (anterior palatal) ^-sounds 
(gy, k^, k^!). In Kato, Navaho, Apache, Chipewyan, and 
other Athabascan dialects, as in Chasta Costa, these have 
become affricative sibilants, without, however, falling together, 
as a rule, with the original Athabascan te- consonants. In 
Chasta Costa, k^ has become te\ k^! has become tc! (this te! 

*> Petitot's tss is our tsl. 

" Chipewyan forms are taken from P. E. Goddard, "Analysis of Cold Lake Dialect, 
Chipewyan," Anthr. Papers Amer. Mus. Nat. Hist., vol. X, pt. II. Chipewyan forms 
taken from Petitot are referred to as Montagnais. 

^- It is quite likely that tcl and ts! are here merely auditory variants of ts! (i is mid- 
way between 5 and c). In Kato tc' , ts' and 5' also interchange. 

42 VI American Indian Languages 2 


does not vary, apparently, with ts!) ; for g^ I have no examples. 
Chasta Costa and Chipewyan are largely parallel in their 
development of Athabascan ts, tc, and k^ sounds: 



Chasta Costa 



d^, d 




te, d 




te!, 6' 
























There are thus three distinct series of sibilant affricatives (and 
of sibilants) in Chasta Costa and Chipewyan, none of which 
is in direct accord with the original Athabascan sounds; Hupa, 
it is highly important to note, reflects the original sounds almost 
exactly .^^ Carrier, it would seem, has also preserved the ky- 

Examples of Chasta Costa tc' from original ky are: 

dL,tcd/yi "big thing" (cf. Hupa -kya/o "large;" Kato 
-tcay, -tea' "to be large") 

''In his "Analysis of Cold Lake Dialect, Chipewyan," Goddard treats Chipewyan 
ts and tc as though they were one sound corresponding to Jicarilla and Navaho tc (p. 86). 
Examination of the various illustrative forms scattered throughout the paper, however, 
soon convinces one that Chipewyan ts, dz, and ts! correspond respectively to Hupa, 
Jicarilla, and Navaho fc(w), dj, and tcl; whereas Chipewyan tc, dj, and tc! correspond 
respectively to Southern Athabascan ts, dz, and ts! and to Hupa k^, g^, and k^!. Thus, 
the Southern Athabascan ts- sounds represent both original ts- sounds and k^- sounds; 
perhaps there is a phonetic difference that does not come out clearly in the orthography. 

As for Kato, Goddard finds no difference between tc- sounds that go back to original 
tc- sounds and those that correspond to Hupa k^- sounds (" Elements of the Kato Language," 
pp. 16, 51). However, deictic tc'-, corresponding to Hupa tc!-, varies with ts' and s', thus 
suggesting ts! as the true sound ; on the other hand, tc'- (to indicate indefinite third personal 
object) corresponding to Hupa k^!- occurs consistently as tc' (contrast examples of tc'-, 
ts^-, s'- on p. 50 with those of tc'- on p. 51). It seems plausible, then, that in Chipewyan, 
Chasta Costa, and Kato original k^- sounds became true tc- sounds, while original tc sounds 
were shifted to ts- sounds (which are apt to be heard as either ts- or tc- sounds). 

Six: Athabaskan and Na-Dene Languages 43 


-tc'u augmentative suffix (e. g., W/tc'ii "horse," literally 
"big dog") (cf. Hupa -kyo; Kato -ted) 

Examples of tc! going back to Athabascan k^I are: 

dAn/tdi "four" (cf. Hupa dink, i. e., di-qkyf) 

sic! At I de "seven" (cf. Hupa xo/kit, i. e., -k^Ht) 

tddsL/se "he cries" (cf. Hupa kya/teL/tcwii "it cried, 

i. e., k^Ia-) 
tc!- verb prefix indicating indefinite object (cf. Hupa 

k-, ky-, i. e., k^!-; Kato tc'-) 

Athabascan possessed sonant sibilants (2, j) and sibilant 
affricatives {dz, dj). Of these sounds z has been found in Chasta 
Costa -Az "to sneeze;" dj is illustrated in several forms, but, 
as we shall see in a moment, does not in these go back to Atha- 
bascan dj. dz has not been found, though it may exist. 7, as 
in Kato and Hupa, has become c: 

ia/cAn "black" (cf. Hupa Lu/hwin<*-cin; Kato i/cmi^; 
Jicarilla Ll/zl; Nav. Ll/jin; Chipewyan del/zun; 
Loucheux del-zen) 

Chasta Costa dj results from / (unaspirated) plus y: 

qlwAt/tc/At/dja "table" {<*qIwAt/tc!At/ya "whereon one 

eats;" -ya "to eat") 
ya/da/yit/dja "we are ashamed" {<*ya/da/yit/ya; cf. 

yAc in ya/dAcl/yAc "I am ashamed") 

Of the lateral consonants, only three (/, I, and lI) have 
been found in Chasta Costa. Original dl may have been pre- 
served also, but Athabascan did was heard rather as / (unas- 
pirated) plus Id: 

yAct/lo "I laugh" (cf. Chipewyan -did, -dldk' "to laugh") 
-/- is very probably third modal -/- here; while -did really 
appears as -Id. After c and s, I becomes /: 

nd/dAcl/nic "I work" (cf. nd/dAl/nic "he works") 
nd/xwAcl/ye "I play" (cf. nd/xwAl/ye "he plays") 
qlwAt/dasi/nd "it was lying on it" 

44 VI American Indian Languages 2 



Independent personal pronouns: 

ci "I" ne "we" (probably contracted 

from *ne/he; cf. Hupa ne/he) 
nAn "you" nd/ne "you" (plur.) 

yu "he, that one" yu/ne, yun/ne "they, those" 

(really demonstrative) (really demonstrative) 

Examples of possessive pronouns are: 

cic/la "my hand" {cic is independent cl combined with 

possessive prefix c-\ literally, "I my-hand") 
nAn /la "your hand" (that is, nAn n-, "you your-hand") 
hi la "his hand" {hi is demonstrative) 

c/na/yd "my eyes" 

s/tsH/de "my sickness, I am sick" 

n/ts!l/de "you are sick" 

nd/ts!l/de "our sickness, we are sick" 

no/tsll/de/ha "your (pi.) sickness? are you (pi.) sick?" 

{-ha is interrogative) 
xd/ts!l/de "their sickness, they are sick" 

Many nouns, when limited by preceding possessive pronouns, 
suffix -e, as regularly in Athabascan. Thus, from mAn "house:" 

cic/mAne "my house" 
nAn/mAne "your house" 

A noun followed by another with suffixed -e is to be under- 
stood as genitively related to it. Examples are: 

dAne' III tele "person's dog" {ll/tde from ll "dog," with 


Six: Athabaskan and Na-Dene Languages 45 


voicing of /- to /-; cf. Hupa Lin "dog," xo/lin/ke, 

i. e., xo/lirj/kyfe "his dog") 
tAkAc^'^ hlci^/le "bowstring" (literally, "bow's string;" cf. 

Chipewyan Vul "rope," possessed form Vu/le) 
ga/yu ts!i/de "baby's sickness, baby is sick" 

As reflexive possessive is used xd/dAt- (with -d/dAt- cf. 
Hupa a/d- ; Carrier cedced-) : 

xd/dAt/lt/tc!e "his own dog" (used reflexively) 

Of demonstrative pronouns there have been found: 

hi* "that, he" (cf. Hupa hai, indefinite demonstrative and 

article); hi'/tli "that thing" 
yii "that one" (cf. Hupa yd "that") 
yu/ne, ytin/ne "those, they" 
m- "it" (cf. Hupa m-; Kato b-): niAl "with it" 

de seems to be used as relative in: 

de ucL/Ve "what I want" 

This element is perhaps demonstrative in force and related to 
Hupa de in ded "this," hai/de "this." 

Totality is expressed by Vwl "all, everything" (cf. Hupa 
a /tin "all"). Compounded with this element are: 

Vwl/de "everything" {-de is very likely related to Hupa 
di- in dl/hwd "something," dl/hwe/e "nothing") 

do I Vwl/de "not everything" 

Vwi/dAn "everywhere" (literally, ^'all-at;" cf. Hupa a//iw/- 
din "every place") 


Primitive non-descriptive nouns, as in all Athabascan dia- 
lects, are relatively frequent in Chasta Costa. Monosyllabic 
nouns are: 

Body Parts. 

la "hand" (cf. Hupa -/a; Kato -/a*) 

** Probably to be understood as lAk/gAc. 

46 yi American Indian Languages 2 


hwd "foot" (cf. Kato -kwe*; Chipewyan -ke) 
61 "head" (cf. Kato -si'; Chipewyan -61, -t6i) 
-ya "hair" (in 6 Ay a "head-hair;" cf. Kato -ga' "hair;" 
Chipewyan -ca, i. e., -ya) 

tdac "bird" 

ll "dog" (cf. Hupa Lin; Chipewyan Ll) 

Natural Objects. 

se "stone" (cf. Hupa tse; Kato se) 
cd "sun" (cf. Hupa hwa; Kato ca) 
lAt "smoke" (cf. Hupa hit; Kato uit) 

Culture Objects. 

mAn "house" (cf. Hupa diminutive min-tc "hut") 

Llel "matches" (originally doubtless "fire-drill;" cf. Chipe- 
wyan L'eL "fire-drill") 

gd6 "camass" (cf. Hupa kos "bulbs") 

LloH-e "(its) string" (cf. Hupa Lol "strap;" Chipewyan 
L'ilL "rope") 

Primitive, at any rate not easily analyzed, nouns of more 
than one syllable are: 


dAtt/ne, dAne' "person, man" (cf. Chipewyan de/ne, dun/ne; 

Carrier tcene) 
ts!d/xe "woman" (cf. Carrier tsekhe; Kato tc'ek) 
dis/ne' "male" (with -ne cf. probably -ne of dAne') 
sd/sAs "white man" 
tsxd/xe "child" (cf. Carrier oe'zkhehkhe; Kato skl-k "boys, 

kel/'e "boy" (perhaps misheard for k!el/'e; cf. Kato 

k'il/lek "boy") 
gd/yu "baby" 

Body Part. 

na/y& "eye" (cf. Hupa -na; Kato -7ia*\ Chipewyan -na/ca, 

Six: Athabaskan and Na-Dene Languages 47 



BAyAl "grizzly bear" (cf. Carrier sces-e^cel "brown or cross 

mis/ki!)i{') "gull" (cf. Kato biitc/k'ai') 

dis/dac "fawn" 

dA/meV /ke "pelican"^^ 

nat/qH "duck" (cf. Kato nd'/qH'') 

ml/tc'd/tsUl/fil "deer" 

Ve/q!d/lec/re "mink" 

sd'/wAs/ts!e "sandhill crane" (cf. Applegate Creek tcd'/- 

tclal/tchis/dje "ruffled grouse, 'pheasant' " 

6d/gi "kingfisher" 

dAs/nAl "red-shafted flicker" 

teiAB/nd/yal/tetde ' 'hummingbird' ' 

0c I id e "bluejay" 

nd/ts!d/le "horned lark" 

so's/ga/ga "robin" 

ts!d/ts!uk "wren" 

kAsis "barn swallow" 

ga/lal/'e "crow" 
Many of these animal names, as well as some of those that 
follow, are probably descriptive verb forms that have become 


tcA/pd/yii "flower"'" 
mt/tlal/tdAd "arrow- wood" 

do'/de "tar- weed" (probably compounded with Atha- 
bascan do' "grass;" cf. Hupa Lo/da-itc "an herb") 
tc!Al/yat/ts!s "sunflower (?)" 

"This word is humorously used to refer to Democrats, Democrat and dA/mel' /ke 
exhibiting some similarity in sound. 

'* This form was obtained independently. 

" This word is remarkable as containing p, a sound that is normally absent in Atha- 

48 ^f American Indian Languages 2 


dAl/si "pine" (cf. Kato dul/tcik "yellow pine," from -tclk 

nd/ife "pine-nut" 
dA/nAc "manzanita" (cf. Hupa din/nuw; Kato tun/nuc 

"manzanita berries;" Galice Creek de/rec) 
mAt/tcfi "cat-tail" 
cAc/dd' "oak" 

Culture Objects. 

xAnAd "canoe" 

at/ tea "pipe" 

tclA/BA/gAl "sandstone arrow-shaft scraper" 

tA/kAc "bow" (probably VAk/gAc\ cf. Kato gqc "yew") 

*q!d/xA6 "arrow" 

det/t!e "arrow-point" 


tdlAd/dd "story" 

yA/wls "whistling" (cf. Carrier ytiyuz "whistling," as 

tsli/de "sickness" (used with possessive pronouns to indi- 
cate "to be sick") 

Several animals are designated by words ending in -tc'u, 
an augmentative suffix, "big" (cf. Hupa and Kato animal and 
plant names in -kyo and -ted respectively). Such are: 

ii*/te'u "horse" (literally, "big dog;" cf. Chipewyan 

xd'/te'ii "goose" (cf. Applegate Creek k'q'/te'u. These 

words are formed from Athabascan xa: Chipewyan 

xa "goose;" Kato ka') 
dAe/te'u, des/te'ti "grouse" (cf. Kato due /ted, dus/ted 

Vet/mo/tc'u "pigeon" 
eu/dc' /tc'u "bald eagle" 
BAe/dA/li/tc'u "owl" 
H/tcIe/tc'u "red-headed woodpecker" 
ga/sd'/te'ii "raven" 

Six: Athabaskan and Na-Dene Languages 49 


Nouns ending in -til or -tllni denote "one who has so and 
so." -ni is, Hkely enough, related to -ne of dAn/ne "person;" 
-ne or -n is found in many Athabascan dialects as suffix denot- 
ing "person." Examples of -t!i{ni) are: 

U/t!l/ni "dog-owner" 

do I ail till ni, do J at I ill "bachelor" (literally, "not-wife-hav- 

ing-person." do- "not;" at-, i. e., ail "wife," cf. 

Hupa ui "wife," Kato aV "sister") 

Examples of noun compounds consisting of two noun 
stems are: 

6a /yd "head-hair" (shortened from 61 "head" and -ya 

"hair." Cf. Chipewyan 6l/Ga) 
ga/lal gwd/yu "red-winged blackbird" (literally, "crow('s) 

brother-in-law." With this cf. Chipewyan da/tsa/- 

tcel/le "a small crow," literally, "crow younger- 


An example of a compound noun consisting of verb and 
noun is: 

aI/Az dAn/ne "sneezer" (literally, "he-sneezes person") 

An example of a compound noun consisting of noun and 
adjective is: 

iclac l/66/e "bluebird" (literally, "bird blue") 

A characteristic type of noun in Athabascan is formed by 
verbs which, while remaining strictly verbal in form, are used 
to refer to objects, in other words, are logically nouns. As 
has been already noted, several nouns of more than one syllable 
listed above as unanalyzable are doubtless, strictly speaking, 
verb forms. Quite clearly verbal in form are: 

nd6/Ll6 "paper" (cf. nal/ilo "he writes") 

ql'wAt/da6t/gAc "table-cloth" (literally, "it lies or is thrown 

^ Goddard, op. cit., p. 1 10. 

50 VI American Indian Languages 2 


down on top;" cf. Hupa -k2as, i. e., -gas, "to throw," 
and wes/kas "it lay"^^) 
qlwAt/tclAt/dja "table" (literally, "thereon it is eaten") 
mAl/Ve/tc!At/ts!Al/lec "smoking materials" (hterally, "there- 
with it is smoked") 


1. la, Wjca (cf. Hupa La\ Kato La/ ha') 

2. nd/xi (cf. Hupa nax] Kato nqk/ka'); nd/xi la "two 


3. t'd/yi (cf. Hupa tak, i.e. Vak!; Kato tak'; Chipewyan 

ta, ta/ce 

4. dAn/tdi (cf. Hupa dink, i. e. dirik^!; Chipewyan 


5. sxo/ld (cf. Hupa tcwo/la; Chipewyan sa/so/la/cai*) 

6. k'wAs/Vd/ne (cf. Hupa xos/tan) 

7. slclAt/de (cf. Hupa xo/kit, i. e. -K^z7) 

8. nd/xAn/do ( = "it lacks two, two less") 

9. Ian/ do ( = "it lacks one, one less") 
10. hwe'/de 

Of numeral adverbs there were recorded: 

Idt/dAu "once" (cf. Hupa na/din "twice," min/ Lun/din 

"ten times") 
la/me/q!e/ca "in one time" 


Of adjectives, or verb stems with adjectival significance, 
there have been found: 

wAs/xe, wAs/xd "good;" wAs/xe li "dog is good" 

txAs/xe/la "rich" {-la is verbal sufhx) 

du/An/de "bad" (evidently verbal in form, du-, do- is 
negative; -de probably misheard for -t!e "to be, exist;" 
cf. Hupa un/te, i. e. An/t!e "there is") 

"Goddard, op. ciL, p. 281. 

Six: Athabaskan and Na-Dene Languages 5 1 


AL/tcd/yt "big thing" (cf. Hupa -kya/o "large;" Kato 

l/gi "white" (cf. Hupa -L/kai; Kato -h/gai) 

la I c An "black" (cf. Kato -L/cun^\ Chipewyan del/zun) 

l/sAk "red" (cf. Kato -h/tclk) 

l/do "yellow, green" (cf. Hupa lit/tsd "green;" Kato 
-L/tso "blue") 
"White," "black," "red," and "yellow, green" are characterized 
by prefixed I {a)-, which is common as adjectival prefix also in 
other Athabascan dialects. 


Adverbs of place are: 

xun "there" (cf. Hupa third personal pronoun xon?): 

xiin t'e/Bi/ya "there you go" 

hV xiin Ved/ya "there he goes" 
txun/la "where?": 

txun/la VejBi/ya "where are you going?" 
do/dAt "nowhere" (cf. Hupa -dit- in hai/dai/dit/din 

"where;" do- is negative) 
(i^y^/ge "up" (cf. Hare /eg^): 

dAk/ge Bicl/Vdl "I kicked him up" 
md'^/dAn "on edge" (-dAn is postposition "at;" md^-K 
*mq-<*marj-; cf. Hupa ntL/man "each side") 

Adverbs of time are: 

xat "then" (cf. Hupa xat "yet, right") 

xd "quickly" (cf. Hupa xa "yet") 

xun/de "tomorrow" (cf. Hupa yis/xiln/de "tomorrow"): 
xun/de do/wa/yAc/l "I'll see him tomorrow" 
xun/de td/Ad/dd nAl ndcl/si "tomorrow story to- 

ycu I-shall-tell" 
xun/de t'Ac/yAc "tomorrow I'll go") 

52 VI American Indian Languages 2 


t'wt/dAn "always" (literally, "all-at"): 

Vwi/dAti t'Al/dAc "he always runs" 
t'wt/dAn As/se "I always cry" 
xAi/tsH/dAn "this evening" (doubtless misheard for xal!-', 
-dAn is postposition "at." Cf. Hupa xu/Le "in the 

xAL/tsH/dAn do/wan/yAc/l "I'll see you this 

Modal adverbs are: 

do negative (cf . Hupa do) : 

do/t'Ac "he won't go" 

do/rAc/yic "I'll not go" 

do/yd/t/a "he won't fly" 

dd/As/se "I'm not crying" 

do/nd/dACL/nic "I'm not working" 

dd/yAc/l "I didn't see him" 

do/ned/l "I'm not looking at him" 

dd/ucL/t'e "I do not want" 

do/na/yAct/xwl "I do not vomit" 
. la prohibitive: 

Id "don't!" 

Id/n/yic "don't go!" 

la/yi/i "don't see him!" 

Id/nd/xwil/ye "don't play!" 

ld/na/dit/t!d "don't bathe!" 

Id/na/yat/xwi "don't vomit!" 
doldd/qle "unable" 

dol LAn "not much" (cf. Hupa Lan "much," do/ian "little") 
do I wi I la "of course" (cf. Hupa ddii "it is," he I don "at 

do lid emphatic negative (really verbal inform, "to cease;" 
cf. Hupa -lan, -lun with negative prefix do- "to quit, 
leave, desist") : 

do /Id c/yl/i "you didn't see me" 

Six: Athabaskan and Na-Dene Languages 53 


cd'*/djl "all right" (cf. Hupa nil/hwdn/^x "properly"?) 
cAl/q!we "to be accustomed to": 

cAl/qlwe na/dAct/t!d "I'm used to bathing" 
t!V/xun "to keep on:" 

tlV/xun ne/cAl/l "he keeps looking at me" 
BAk/gwe "in fragments" 
ho future prefix (more properly intentive) : 

hd/ya/yic/t!a "I'll fly" 

hd/tc!AsL/se "he wants to cry" 

ho I ill I yit/lo "stop laughing!" 
do/wa future prefix (probably with dubitative coloring) : 

do/wa/c/yl/i^/Ve "you'll see me" 

s/ts!l/de do/wa/ aI/W "I'll get sick" (Hterally, 
"my-sickness will-become") 

do/wa/ncL/yan/nAl "he will upset them" 

dd/wa/it'dt/nTii "they will go to pieces" 


Athabascan is characterized, among other features, by the 
use of a considerable number of postpositional elements of 
chiefly local force. They are appended to nouns or pronom- 
inal, numeral, or adverbial stems; less often to verb forms, in 
which case they have subordinating force. Chasta Costa 
examples are: 

-dAn "at" (cf. Hupa -din): 

xAL{!)/tsH/dAn "this evening" 
Vwi/dAn "everywhere" (hterally, "all-at") 
Idt/dAn "once" (cf. la- "one") 
md^/dAti "on edge" 

al/dAc/ni/dAn "when I tell him" (hterally, "I- 
tell-him at")*« 

"Similarly in Hupa -miL "when," as verb suffix, is doubtless simply pronominal 
■mi- plus postposition -/. "witli." 

54 VI American Indian Languages 2 


4 "with, to" (cf. Hupa -l; Kato -l): 

xAUAd/l 7idc/xe "I paddle canoe" (literally, "canoe- 
with I-paddle") 

tdlAd/dd riAl ndcl/sl "I tell you story" (literally, 
"story you-with I-make") 

td/Ad/dd caI na/yesL/si "he tells me story" 
(literally, "story me-with he-makes") 

mAl/Ve/tc!At/ts!Al/lec "wherewith it-is-smoked, ma- 
terials for smoking" {mA-l- "therewith;" cf. 
Kato buL "with it;" Hupa mil "with, in") 

This same -I is probably also found attached to verbal prefix 
a- (used in verbs of saying) : 

al/dAc/ni/dAfi "when I tell him" (cf. Hupa ah/- 
tcit/den/ne "he talked to") 
-tslAn "toward" (cf. Hupa -tcin "toward;" Kato -tc'un* 
"to, toward"): 

s/tslAn/na/'Ac "to-me he-will-bring-it" 
-me "in" (cf. Hupa -me "in;" Kato -hV "in"): 

mAn/me "in house" 
-me/q!e "in, around in" (compounded of -me and -q!e\ 
cf. Chipewyan -k'e "on"): 

mAn/me/q!e "around in house" 

la/me/q!e/ca "all in one time" (cf. la, Wjca "one") 


As in other Athabascan dialects, the typical Chasta 
Costa verb consists of one or more adverbial prefixes, which 
may be followed in order by a deictic or third personal ele- 
ment, a first modal prefix, a second modal element, a first or 
second person subjective element, and a third modal element or 
"class" sign; these, not all of which need of course be present, 
are then followed by the Terb stem itself. The stem often 
ends the verb form, but may be followed by one or more enclitic 
elements of modal or syntactic force. The verb form is fre- 

Six: Athabaskan and Na-Dene Languages 55 


quently preceded by an adverb or postposition which, while 
best considered as a non-integral part of the verb, forms a 
rather close syntactic unit with it. A pronominal object, 
if present, comes after an adverbial prefix but before a first 
modal element. Thus, the verb form Vd/yd/Bot/nd/hd "do 
you (plur.) drink?" consists of seven elements: Vd-, an ad- 
verbial prefix referring to water; yd-, a second adverbial ele- 
ment; B-, a second modal element of durative significance; 
-6-, second person plural subjective pronominal element; -/-, 
a third modal element, probably intransitive in force; -nd, 
verb stem "to drink;" and -hd, an enclitic interrogative element. 
The various elements that go to make up verb forms will be 
taken up in the order indicated. 

Adverbial Prefixes, a-, a-, 'a- used with verbs of say- 
ing, doing, and being (cf . Hupa and Kato a-) : 

d/dJAn "he says" 

al/dAc/ni/dAn "when I tell him" (for -1-, see 

under Postpositions) 
dd/dAt 'An/ tie "there is not anywhere" 

This a- is probably equivalent to an indefinite object, "some- 
thing," indicating what is said or uttered without definitely 
referring to it. This comes out rather clearly on comparison 
with a form like yu/wls dAcl/ni "I whistle" (literally, "whistling 
I-utter"), where no indefinite object a- is required, what is 
uttered being specifically referred to by yu/wls "whistling." 
That a- is somewhat in a class by itself as compared with other 
adverbial prefixes is indicated by its being followed in forms 
with indirect object by postpositive -/-. 

yd-, ya- "up (in the air)" (cf. Hupa ya-\ Kato ya^-): 
yd/yAcl/gAd "I climb" 
yd/ y Ad /del "I threw" 
yd/yic/lla "I fly" 

It is not clear what significance is to be attached to ya- in: 

ya/dAcl/yAc "I am ashamed" 
ya/da/yit/dja "we are ashamed" 

56 VI American Indian Languages 2 


ye- "into enclosed space (including mouth)" (cf. Hupa 
ye-] Kato ye'-, yV-) 

ye/ydt/ne/la "he bit it" 
da-, da- "sitting or lying on something above ground" 
(cf . Hupa and Kato da-) : 

da/ 6 Ad /da "I am sitting down" 

dd/de/dil/tB!i "we are sitting down" 

dd/nAc/Vac "I go to bed" 

q!wAt/dadt/gAc "it lies thrown down on top," 

i. e. "table-cloth") 
qlwAt/dash/nd "it was lying on it" 
Ve- "in the water" (cf. Hupa ie-; Kato /e'-): 

Ve/An/yAc/lAl "I am sinking in the water" 
Ve/nic/lat "I drown" 
Vd- referring to water (cf . Hupa and Kato ta-) : 

Vd/yAct/nd "I drink" 
tde- "across a stream" (cf. Hupa tee- "out of;" Kato 
tc'e- "out of;" Chipewyan ts'e- "used of approach 
to a body of water") : 

tcle/Bii/tlo "I swim across" 
An- implies disappearance or undoing (cf. Chipewyan 
*a-, an- "away," implies "desertion or abandonment"): 
Ve/ An/yAc/lAl "I am sinking in the water" 
do/wd/ An/nd/yan/nAl "he will upset them" 
an- "back, hither" (cf. Chipewyan '5- \in-, ai- "back, 
toward home"): 

an/yi/al "come here!" 
tdd-, tela- of unknown significance (cf . Hupa kya-"^^) : 

tcld/ sAsh/se "I cry" (cf. Hupa kya/teh/tcwe "she 

heard it cry") 
tddsi/se "he cries" 
se'- used with verb of smiling: 

se' /yAt/lo "he smiles" (cf. yAt/ld "he laughs") 

^' Goddard lists forms in kya-, i. e., k^ta-, under ky-; see op. cit., p. 90. It seems better 
however, to keep them apart. 

Six: Athabaskan and Na-Dene Languages 57 


tclo- of unknown significance (cf . Hupa kyo-^^) : 

tcfo/yit/siL/la "he pointed with his finger" 

ne/tc!uc/lec "I'll bet you"« 
u- of unknown significance (cf. Hupa verbs in o-**) 

dd/ucL/t'e "I do not want" 

de/ucL/t'e "what I want" 
nd-, na- indefinite movement on surface of ground or 
water; horizontality (cf. Hupa and Kato na-): 

fiAn/ndd/yd/la "he went around it" 

s/tslAn/na/'Ac "he'll bring (it) to me" 

yAn/na/Ac/t'e "he will bring here" 

nd/ni/An "stop him!" 

nd/ya "is going about, living" 

i/t'i yAn/na/'d "he brags" (literally, "high, im- 
portant he-has") 

nd/xwAcL/ye "I play" 

nd/dAch/nic "I work" 

do/wa/nd/yan/nAl "he will upset them" 

ndcl Lib "I write" 

xatiaBU ndc/xe "I paddle canoe" 

nac/t!d "I swim, bathe" 

na/tc!il/de "you wash" 

nd/dAcl/de "I washed myself" 

nd/dAt/t!d/Ve "he'll bathe" 

nd/xAt/dAl/el "they'll bathe" 

tdlAd/dd caI nd/Bil/sl "story to-me "you-told, 
na- "back again" (cf. Hupa and Kato na-), followed by 
third modal -/-: 

na/yd/dAdt/xwi "I vomit" 
yAn- of uncertain significance (cf. Hupa wun- "to pursue 

*^ Goddard lists forms in kyo, i. e., kyjo-, under ky-; sec op. cit., p. 90. Perhaps k^lo 
is compounded of k^l- and o-. 

♦' This tclii- is probably better explained as deictic tcl- followed by future imperative 
M-; see note 86. 

"Goddard, op. cit., p. 115. 

58 VI American Indian Languages 2 


or seek something; to attempt something by per- 
sistent effort"): 

yAn/na/Ac/t'e "he will bring here" 
i/tH yAn/na/'d "he brags" (perhaps "important 
ne- of unknown significance: 

necl/l "I'm looking at him" (cf. ye/ die /I "I saw 

nes/tslAl/l "I am seen" 
xw{a)- of unknown significance: 
nd/xwAci/ye "I play" 
xwAcl/I "I believe (it)" 

xwAn/ne/diL/ya "you win" (see under first modal 
yd- ya-, (y) refers apparently to "mouth": 

ye/ydt/ne/la "he bit it" {yat- may, however, 
have been misheard for yAt-, with second 
modal 7-; see note 92) 
na/yd/6Adt/xwt "I vomit" 
id/na/yat/xwl "don't vomit!" 
t'd/yd/6it/nd "we drink" 

Vdlyit/nd "you drink" (or is y- here second 
modal prefix? 

Verbal prefixes of local force which are doubtless primarily 
postpositions and which are prefixed to adverbial prefixes proper 

nAn- "around" (cf. Hupa -nat\ Kato -nd)\ 

nAn/ndS/yd/la "he went around it" 
qlwAt- "on, on top" (cf. Hupa -kut "on;" Kato -k'wut' 

qlwAt/tclAt/dja "whereon one eats, table" 
q!wAt/daBt/gAc "it lies thrown down on top, 

qlwAt/dasi/nd "it was lying on it" 

Six: Athabaskan and Na-Dene Languages 59 


Deictic Prefixes. Under this head are grouped a small 
number of quasi-pronominal elements of third personal refer- 
ence which regularly come after adverbial prefixes, if any of 
these are present. They cannot be grouped with first or second 
personal subjective elements, as their position is quite distinct 
from these; first and second modal prefixes may come between. 
Of deictic elements there have been found: 

td- denotes lack or indefiniteness of object of transitive verb 
(cf. Hupa ^-, ky-, i. e. k^!-\ Kato tc'- ): 

tdAy/ye/dic/ya "I eat" (i. e. without specific 
object being designated; cf. Hupa yik/kyu/- 

win/yan "it ate") 
q!wAt/tc!At/dja "whereon one eats, table" 
na/tc!il/L!d "you write" (cf. Hupa na/kis/Lon, 

i. e. na I k^ lis I hi on "she made baskets") 
nd/tc!l/t!d "you swim, bathe" 
na/tcHl/de "you wash" (cf. Kato te'/na/tc'ih/deG 

"he washed it") 
tdAt/Vit/dAl "we wash ourselves" (cf. Hupa wa/- 

kin/nin/seL "it was heated through") 
mAl I V e I tcl At I tsl aI I lee "wherewith it is smoked" 

(somewhat doubtful, as id- here follows first 

modal prefix Ve-\ but see note 77) 
td All ltd "he sucks" (cf. Kato tc'lh/Vot "[make] it 


It is possible that in this last example td- is third personal 
subjective (cf. Hupa tc-, i. e. td-; Kato tc'-, ts'-, 5'-), as sug- 
gested by Ad/tlo "I suck" with its lack of td- prefix. No other 
plausible case, however, of third personal subjective td- is 
available, so that its existence in Chasta Costa must be con- 
sidered doubtful as yet. 

Generally third person singular subjective forms are dis- 
tinguished by the lack of any pronominal prefix, but in certain 

60 ^^ American Indian Languages 2 


cases deictic elements are found which are clearly third personal 
(subjective) in value. These are: 

dj- (cf. Hupa tc-, i. e. tcf-; Kato tc'-, ts'-, s'-): 

d/dJAn "he says" (verb-stem -n; cf. Hupa ai/- 

tcit/den/ne "he talked to them") 
djAn/la "he says" (cf. Hupa tcin "they say;" 
Kato tc'in) 

It is quite likely, however, that dJAn is to be explained as from 
*dyAii {dy, as we have seen, becomes dj), in which d- is first 
modal prefix (cf. dt/nl "you make a sound") and *yAn is reduced 
from *yen {ye- as below; -n to say). 

ye-, ya- (cf. Hupa y-, yl- referring to other than adult Hupa; 
Kato yl-) : 

uaI na/yei/si "to-you he-tells" (contrast ndcl/sl 

"I tell") 
caI na/yesi/sl "to-me he-tells" (with second 
modal prefix s-, 6-; contrast nd/dU/sl "you 
ya/qled/ya'^^ "he eats" (contrast tclAy/ye/6ic/ya 
"I eat") 

tcIe/xAd/tfo "he swims across" (contrast tcfe/- 
Bit/t!d "you swim across") 

This X- seems to have no parallel in Hupa or Kato (is it con- 
nected with third person objective xo- of Hupa, kw- of Kato?). 
Were it not that -t!d "to swim, bathe" is Used only of singular 
subjects, one might surmise that nc-'is really plural xA- (see 
below) . 

Among deictic elements are further to be reckored certain 
prefixes that serve to indicate either plurality as such or more 
specifically third personal plurality. These are: 

ya- (cf . Hupa ya- ; Kato ya^-) : 

yu/wis ya/dil/ni "we whistle" (contrast yil/- 
ims dAcl/ni "I whistle") 

" qlc- was very likely misheard for te-. 

Six: Athabaskan and Na-Dene Languages 61 


yu/wls ya/dAl/ni "they whistle" (contrast yn/wls 
dAl/nl "he whistles") 

ya-, ya- 

tc!a/yd/dil/se "we cry" (contrast tcfd/sAsL/se 

"I cry") 
tcIa/ydsL/se "they are crying" (contrast tcfdsL/se 

"he cries") 
tc!d/ydl/se/t'e "you (pi.) will cry" (contrast Icldl/- 

se/t'e "you (sing.) will cry") 

XA- third person plural (apparently not found in either 

Hupa or Kato; but cf., without doubt, Chipewyan 

he- "used for dual or plural of verbs in third person"): 

yd/xAy/yi/tIa "they fly" {yd/yi/tia "it flies") 

dd/yd/xAt/t!a "they won't fly" {do/yd/tfa "he 

won't fly") 
XAs/se/re/ha "will they cry?" {As/se/t'e/ha "will 

he cry?") 
do/xAs/se "they're not crying" {do/As/se "he's 

not crying") 
nd/xAt/dAl/nic "they work" {nd/dAl/nic "he 

c/xA/yee/l "they saw me" {c/yeS/l "he saw me") 
na/xAt/da/yAl/el "they are bathing" 
t"e/An/xAy/yAl/lAl "they sink in the water" 

{Ve/An/yAl/lAl "he sinks") 
xAt/t'Al/lal "they are sleeping" {VAl/lal "he is 

dd/xAn/nAt/VAc "they went to bed" {dd/nAt/t'Ac 

"he went to bed") 
x At /VAI/xwaB'' they cough.'' {VAI/xwaB'' he coughs') 
xa/AI/az "they sneeze" (aI/Az ''he sneezes") 

First Modal Prefixes. Under this term are comprised 
a small number of rather frequently occurring elements which 
regularly come after both adverbial prefixes and deictic ele- 
ments, but precede another set of modal elements (second 

62 VI American Indian Languages 2 


modal prefixes) which are to be taken up shortly. Their mean- 
ing is rather colorless. Besides their position they have this 
peculiarity in common, that they lose their vowel in indefinite 
tense forms (such as have no second modal prefixes: d-, y-, 
or W-) and are thus reduced to single consonants. They are: 

t'e- (definite tenses), t'- (indefinite tenses) seems to indicate 
durative activity (cf. Hupa te-; Kato te-, t-): 

re/dic/ya "I go;" indefinite: do/rAc/yic "I'll 

not go;" t'l/yAc/t'e "you must go" 
VAci/dAc "I run" (indefinite) 

tclAt/Vo/dAl "you (pi.) wash yourselves" (indefi- 
t'/yi/l "he looks around" (indefinite; but see 

note 69) 
Ve/BAd/lal/la "I've been sleeping;" indefinite: 

VAc/lal "I'm sleeping" 
VAcl/xwAd "I cough" (indefinite) 
mAl/t'e/tc!At/ts!Al/lec "wherewith it is smoked" 

(as following td- is deictic, it is more likely 

that Ve- here is adverbial prefix, not first 

modal; see note 77). 
de- (definite tenses; da- before 7-), d- (indefinite tenses) 
meaning unknown (cf. Hupa d-, du-\ Kato de-, d-)\ 
al/dAc/ni/dAn "when I tell him" (indefinite) 
yu/wis dAcl/ni "I whistle" (indefinite) 
c/na/yd di'/s'aV "my-eyes hurt" (definite; cf. 

Hupa du I win Ileal "it got sick") 
nd/da/yAct/l!d "I bathe;" indefinite: nd/dAct/- 

Ud'/Ve "I'll bathe" 
na/da/yil/el "we are bathing;" indefinite: nd/- 

dil/el "we'll bathe" 
nd/dAci/nic "I work" (indefinite) 
na/dAcl/de "I washed myself" (indefinite) 
ya/dAcl/yAc "I am ashamed" (indefinite) 
ya/da/yil/dja "we are ashamed" (definite) 
dd/de/dil/lB!i "we are sitting down" (definite) 

Six: Athabaskan and Na-Dene Languages 63 


ye- (definite tenses) , 7- (indefinite tenses) meaning unknown : 
tclAy/ye/dic/ya '1 eat" 
ya/q!eB/ya "he eats" {q!e- is probably misheard 

for ye-) 
ye I Bill "you saw him;" indefinite: do/wa/yl/i/- 

Ve "you'll see him" 
yedt/lo "he breaks into laughter;" indefinite: 

yAt/lo "he laughs" 
an/yi/al "come on!" 

This ye-, y- should not be confused with second modal 7-, 
which will be taken up presently. Two first modal prefixes 
(/'- and 7-) occur in V/yi/l "he looks around;" that 7- is not 
second modal here is indicated by parallel definite forms with 
ye- (see ye/di/l above), further by weak form V- of first prefix 
(definite tenses require t'eY"^ 

ne- (definite tenses), n- (indefinite tenses) meaning unknown 

(cf. Hupa ne-, n-; Kato ne-, n-; Chipewyan we-, nil-): 

xwAn/ne/dii/ya "you win" (cf. Kato kun/ne/sli/- 

yan "you win") 
dd/nAc/VAc "I go to bed" (indefinite; cf. Hupa 

definite: tcin/nes/ten ''hQlSiy') 
n/dd" "it is, stays" (indefinite; cf. Kato definite: 

tc'n/nes/dai "he sat down") 
nd/nAc/An "I stop him;" nd/ni/An "stop him!" 

n/do "it is not" (indefinite; cf. Kato n/dd*/ye 
"there is none") 

This ne-, n- is not to be confused with second modal n-, which 
occurs only in definite tenses. 

Second Modal Prefixes. These comprise three conso- 
nantal elements {B- or s-\ 7-; and 71-) which are used only in 
definite tenses and which have reference, as far as any definite 

*' Moreover, te- in definite tenses seems regularly followed by second modal 6-, not 
7-. Yet -I- of yl- causes difficulty; see note 69. 

64 yf American Indian Languages 2 


significance is ascertainable at all, to what may be termed range 
or span of activity, but not to tense as such. 6- {s- in certain 
forms) is durative or continuative in force (cf. first modal 
t'e-, which is regularly followed by 6-) ; w- is cessative, marking 
the end of an activity or marking an activity which is con- 
ceived as the end point of a previous activity (e. g., "to come" 
as contrasted with durative "to go"); y- is the most uncertain, 
being apparently inceptive or momentaneous in some cases, 
but clearly not so in others/^ They are, it seems, mutually 
exclusive elements. In practice their use seems largely deter- 
mined by the prefixes that precede, n- and y- always begin 
their syllable, being completed either by -i- ( < *-ir}-) or by 
subjective pronominal or by third modal elements, which are 
joined to them by means of -i- or inorganic -A-; 6- (s-) is similarly 
joined to following subjective pronominal elements, if one is 
present, otherwise it forms part of the preceding syllable. 
Exrmples illustrating 6- (s- before I, l) are: 

t'e/dic/ya "I go;" reO/ya "he goes" (cf. Hupa 

te/se/yai "I went away") 
nAfi/ndd/yd/la "he went around it" 
tdAy/ye/dic/ya "I eat;" ya/q!ed/ya "he eats" 

(contrast Hupa yik/kyil/win/yan "it ate" 

with w-) 
tcfe/dit/tfo "you swim across" (contrast na/da/- 

yit/tfo "you bathe") 
dad /da "he is sitting down" (cf. Hupa sit/dai 

"he lived") 
dd/de/6il/td!i "we are sitting down" (cf. Hupa 

de/soh/tse/te "you will stay") 
tc!d/6U/se "you (sing.) cry" (contrast Hupa winj- 

tcwu "you have cried") 

'^ Goddard somewhat doubtfully assigns inceptive force to its Hupa cognate w-; in 
Kate its cognate g- seems clearly inceptive only in certain verbs; while in Chipcwyan 
Goddard ascribes continuative value to g-. It would be worth while making a somewhat 
extcnilcd comparative study of the second modal prefixes of Athabascan, which form one 
of the most difficult but at the same time important chapters of its grammar. 

Six: Athabaskan and Na-Dene Languages 65 


ye /die /I "I saw him" (cf. Hupa te/suw/ifi "I am 

going to look") 
na I QUI SI "you told story" (cf. Hupa na/sei/tcwen 

"I made") 
Ve/Bi/lal/la "you've been sleeping" (cf. Hupa 

nit/te/sil/lal/le "you would go to sleep") 
Bid/ Veil "I kicked him" (contrast Hupa ye/tcu/- 

wii/taL "they landed" with w-) 
na/yd/6it/xwi "you vomit" 
ye/dAdt/lo "I break into laughter" 
Lad/d/la one was { = Ld "one" plus d/'d/la; cf. 

Chipewyan ^e/'d "was there") 
q!wAt/dadt/gAc "it lies thrown down on top, 

table-cloth" (cf. Hupa wes/kas "it lay") 
q!wAt/dasL/nd "it was lying on it" 
xwAn/ne/diL/ya "you win" (cf. Kato kun/ne/- 

slL/yan "I win") 

Examples illustrating n- are : 

nic/ya "I come" (cf. Hupa nei/yai "I came") 
nic/dac "I dance" (cf. Kato nuc/dac "I will 

Ve/ni/lat "you drown" (cf. Kato tc'n/ mil flat 

"it floated there") 

Examples illustrating y- are: 

yt/dac "he dances" (cf. Kato tc'/gtm/dac/ kwaii 

"he had danced") 
yd/yAd/gAd "I climb" (cf. Hupa ya/wii/kas 

"he threw up") 
yd/yi/tfa "it flies" (cf. Hupa na/win/tau "it will 

settle down") 
nd/da/yAct/t!d "I bathe" 
na/da/yil/el "we are bathing" (cf. Chipewyan 

ni/i/oinL/ilL "take through the water") 

66 ^/ American Indian Languages 2 


fe/An/yAc/Ul "I'm sinking in the water" (of. 

Hupa da/na/wil/laL "it was floating there") 
yAci/Az "I've been sneezing" 
tc!d/yit/siL/la "he pointed with his finger" 
ya/da/yit/dja "we are ashamed" 

Subjective Pronominal Prefixes. There are three per- 
sons and two numbers (singular and plural), making six persons 
in all. The third persons, as we have seen, are indicated either 
by the absence of a pronominal element or by deictic prefixes 
which come between the adverbial prefixes and the first modal 
elements. There thus remain four persons (first person singular 
and plural, second person singular and plural) for treatment 
here. In the definite tenses the pronominal elements are 
appended to the second modal elements, with which they form 
a syllable, an inorganic a or i, if necessary, serving to connect 
them. In the indefinite tenses the pronominal elements are 
appended to whatever element (adverbial prefix, deictic ele- 
ment, or first modal prefix in reduced form) happens to precede 
them. They never begin their syllable except in the com- 
paratively small number of cases in which the verb form, in- 
definite in tense, has nothing preceding the pronominal element 
or, in the case of the third person, nothing preceding the third 
modal prefix or verb-stem. When this happens, the second 
person singular and plural and the first person plural stand 
at the very beginning of the verb; the first and third persons 
singular, however, begin with an inorganic vowel A-. 

First Person Singular -c- (cf. Hupa -w-; Kato -c-] Chipe- 
wyan -s-) : 

Ve/dic/ya "I go" 

dicL/Val "I kicked him" 

nic/ya "I come" 

dd/nAc/VAc "I go to bed" 

Ve/An/yAc/lAl "I am sinking in the water" 

y Act /Id "I laugh" 

Six: Athabaskan and Na-Dene Languages 67 


ya/yAcl/gAd "I climb" 
VAcL/dAc "I run" 
nac/tlo "I swim, bathe" 
Acl/t!6 "I suck" 
acl/Az "I sneeze" 

In definite tenses with 6- or n- as prefix the inorganic vowel 
connecting these elements with -c- is regularly i\ this is evi- 
dently due to the palatal quality of the -c-. In definite tenses 
with 7- as prefix, however, the normal inorganic vowel, a, 
is found, due, no doubt, to the velar position of the prefix. 
In the indefinite tenses the connecting vowel, if required, is 
always a. Where we have nAc- we are dealing with first modal 
ne-, reduced to n-, plus -c-, not with second modal n- plus -c-\ 
contrast definite nic/ya with indefinite da/nAc/VAc. 
Before s- sibilants -c- is assimilated to -s-: 

do I AS I se "I'm not crying" {<'^Aclse) 

Sic- goes back to original *sic- or *sac-. When -c- came to 
stand before a dental consonant {d, t, /), it was assimilated to 
-S-, and the inorganic vowel preceding it assumed the form a ; 
this *sAs- then regularly became 6 Ad-: 

da /e Ad /da "I am sitting" 
ye/BAQt/lo "I break into laughter" 
na/ya/BAdt/xwi "I vomit" 
Ve/dAd/lal/la "I've been sleeping" 

Before third modal -1-, die- seems to be regularly retained (cf. 
Bicl/t'dl above; Sid/ si "I make"). Secondary sas-, not shifted 
to OaO-, is found, however, before 1{l) when this element is 
secondarily changed from third modal -/-: 

tc!d/sAsL/se "I cry" 

That sasl- here is equivalent to *sasI-<*sacI- is indicated by 
tc!d/6il/se "you cry;" contrast Bil/t'cd "you kicked him," 
eicl/Val "I kicked him." 

68 VI American Indian Languages 2 


It is to be carefully noted that -c- (or its reflexes -s-, -$-) 
is in Chasta Costa found in both definite and indefinite tenses. 
There is no trace of an element corresponding to the Hupa 
.^. (-e-), Kato -t-, Chipewyan -1-, which are found in forms of 
definite tenses. It is quite probable that the -c- of the indefinite 
forms was extended by analogy. 

Second Person Singular i- (cf. Hupa -77-, i. e. -r)-\ Kato 
-W-; Chipewyan n-, ne-, or nasalization of vowel: 

Vil/dAc "you run" 
Vd/yit/nd "you drink" 
na/tcHl/Ud "you write" 
na/da/yit/t!d "you bathe" 
nd/dit/t!d/t'e "you'll bathe" 
yd/yil/gAB "you climb" 
ya/dil/yac "you are ashamed" 
nd/dil/sl "you made, told" 
nd/xwil/ye "you play" 
xwil/l "you believe it" 
xwAn/ne/BiL/ya "you win" 
VU/xwaB "you cough" 

In all these cases the -i- connects a following third modal ele- 
ment {-1-, -/-, or -/-) with a preceding prefix. Examples of 
-/- beginning its own syllable are: 

tl/Az "you sneeze" 
hd/il/l "stop!" 

If there is no third modal element, the -i-, lengthened to close 
-1-. closes its syllable: 

dd/ni/t'Ac "go to bed!" 
t'i/lal "you are sleeping" 
dd/Bi/dd "you are sitting" 
yd/yi/t!a "you fly" 
do/l/se "you do not cry" 

Six: Athabaskan and Na-Dene Languages 69 


This i-, 1-, is only secondarily the second person singular 
subjective element. The original element was doubtless -17- 
(cf. Hupa). which was reduced to nasalization of preceding 
vowels; the inorganic vowel, when nasalized, took on i- timbre 
Finally, when nasalization disappeared, the /- timbre alone 
remained as the reflex of original -rj-. Where, in many indefi- 
nite tense forms, the nasalized vowel was other than an inorganic 
one, there was nothing left of the -77-: 

do/yd/tfa "you won't fly" 
nd/xe "you paddle" 

nei/l "look at him!" (cf. ned/i "I'm looking at 

In such cases the second person singular fell together with the 
third, as in dd/yd/t!a "he won't fly." 

First Person Plural (i)t-*^ (cf. Hupa it/d-, -d-\ Kato d-; 
Chipewyan -/-, -d-^'^) : 

Vit/lal "we are sleeping" 

tclAt/Vit/dAl "we wash ourselves" 

Ve/ nit flat "we drown" 

da /nit /Vac "we went to bed" 

ye/dit/l "we saw him" 

ya/yit/t!a "we fly" 

Ve/An/yit/lAl "we are sinking in the water" 

dd/it/se "we are not crying" 

In Hupa and Kato regularly, and in Chipewyan often, the 
first person plural subjective pronominal prefix begins its 
syllable; in Chasta Costa it regularly ends its syllable, unless 
it has to stand at the beginning of the verb form, when it consti- 
tutes a syllable by itself (cf. dd/it/se above; do "not" is inde- 
pendent adverb rather than prefix). 

" / is here unaspirated, and is thus etymologically identical with d. 
'" In Father Legoff's Montagnais paradigms -id- or -//- often, in fact regularly, appeals; 
-i- seems, as in Chasta Costa, to be organic. 

70 ^f American Indian Languages 2 


If the prefix preceding the pronominal element ends in 
a vowel, the -i- disappears: 

do/ydt/tla "we won't fly" 

This does not mean, however, that this -/- is to be considered 
an inorganic vowel, as is the case in Hupa it/d-. If -it- is followed 
by third modal -/-, both -/- elements combine into a single -/-, 
and all that is left of the pronominal prefix is the -i- : 

t'd/yd/dit/nd "we drink" (contrast t'd/yddt/nd 
"they drink") 

If the third modal element is -I- or -/-, -/- disappears and -l- 
is changed to -/-; thus the first person plural of I- verbs and 
/- verbs is always formed alike. In Hupa and Kato third modal 
-f- regularly becomes -/-, but d- is preserved; hence Hupa 
dil-, Kato dul-. In Chipewyan, however, as in Chasta Costa, 
-/- not only becomes -/-, but -/- disappears. For Chasta Costa 
this means that the second person singular and first person 
plural of /- verbs is identical, provided, of course, that there 
is no deictic prefix of plurality in the latter and that the verb 
stem does not change for the plural. Examples of /-verbs are: 

dd/de/Bil/td!i "we are sitting" (cf. Chipewyan de/dll/d'l 

"we are sitting") 
nd/dil/nic "we work" (cf. nd/dil/nic "you work") 
t'il/xwAd "we cough" (cf. VU/xwaB "you cough") 
nd/xwil/ye "let us play" (cf. nd/xwil/ye "you play") 
il/Az "we sneeze" (cf. tl/Az "you sneeze") 
na/da/yil/el "we are bathing;" nd/drl/el "we'll bathe" 
tc!a/ya/dil/se "we cry" (cf. tda/Bil/se "you cry") 

Examples of /- verbs are: 

yu/wls ya/dil/nl "we whistle" (cf. third person plural: 

na/tclil/Ud "we write" (cf. na/tdil/dd "you write") 

Six: Athabaskan and Na-Dene Languages 71 


If, in an indefinite tense form, the pronominal element is pre- 
ceded by a prefix ending in a vowel and is, besides, followed by 
third modal -I- or -/-, both -i- and -/- have to disappear and 
there is nothing left of the pronominal element except, in the 
case of /- verbs, the change of -/- to -/-; 

nel/l "let us look at him!" (cf. nel/l "look at him!") 
tcld/yel/se, very likely misheard for tc!d/ydl/se "we'll cry" 
(cf. definite: tcfa/yd/dil/se "we cry") 

Second Person Plural 0- {ci-Ylupa. o'-; Katoo'-; Chipewyan 

rS/lal "ye sleep" 

t'e/do/lai "ye have been sleeping" 

t'e/nd/lat "ye drown" 

yd/yd/ t!a "ye fly" 

t'e/An/yd/Ul "ye sink in the water" 

t'd/yd/dot/nd "ye drink" 

dd/o/se "ye are not crying" 

No aspiration was heard after o in Chasta Costa. This does 
not seem due to faulty perception, as /- verbs keep their -/- 
after o-, whereas, under similar circumstances, Hupa, Kato, 
and Chipewyan change -/- to -I- {o'-l- becomes ol-). Indeed, 
in Chasta Costa I- verbs change their -I- to -/- after second 
person plural 6-. Examples of o- before /- verbs are: 

nd/dol/nic "ye work" 

na/da/yol/el "ye bathe;" nd/dol/el "ye will bathe" 

t'dl/xwAd "ye cough" 

nd/xol/ye/le "ye play" (for -xwol-) 

ol/Az "ye sneeze" 

Examples of -/- becoming -/- after 3- are: 

ya/ddl/7il "ye utter, make a sound" (cf. third person plural 

ne/xd/ol/l "ye look at him" (cf. nci/i "you're looking at 


72 VI American Indian Languages 2 


When, in an indefinite tense form, o- is preceded by a prefix 
ending in a, a and o contract to long a (which, it would seem, 
remains long even in closed syllables) : 

do/yd/ t!a "ye won't fly" {<*yad'-; cf. third person singular 

do/yd/tla with original yd- ; and second person singular 

dd/yd/t!a < *yq- < *yarj) 
tc!d/ydl/se/t'e "ye will cry" (cf. definite: tcIa/yd/Odl/se 

"ye cry;" and contrast tcld/ydl/se "we'll cry" with 

short -a-) 

Third Person. As already noted, the third person, apart 
from possible deictic prefixes, is marked by the absence of any 
pronominal element. If the element preceding the third modal 
prefix or the stem consists of a consonant which must begin 
its syllable, an inorganic -a- is found between the two; if a 
third modal prefix is absent, the syllable preceding the stem 
is closed by a consonant borrowed from the first consonant of 
the stem. Examples of third persons with -a- before a third 
modal prefix are: 

ya/dAl/yAc "he is ashamed" 
VAl/dAc "he runs" 
nd/xwAl/ye "he plays" 
yd/yAl/gAd "he climbs" 
yu/wls dAl/ni "he whistles" 
Vd/yAt/nd "he drinks" 

Examples of third persons with -A- followed by an inorganic 
consonant are: 

dd/nAt/VAc "he went to bed" (-/- is not third modal; cf. 

second person singular dd/ni/VAc) 
Ve/Ati/yAl/lAl "he is sinking in the water" (-/- is not third 

modal; cf. second person singular Ve/An/yl/lAl) 

First modal n-, reduced from ne-, however, has in several cases 
been found without following inorganic vowel and consonant. 

Six: Athabaskan and Na-Dene Languages 73 


In such cases it closes the preceding syllable, which may even 
belong to another word. Examples are: 

dd/dAt 'An/tie "not-anywhere there-is" (cf. Kato qn/t'e 

"it is;" Hupa un/te "there is") 
Idn/do "nine" (really la n/do "one is-lacking") ; nd/xAn/do 

"eight" (reduced from na/xi n/do "two are-lacking") 

(cf. Kato 71 /do' /bun "it will not be," but also niit/do' 

"all gone") 
cic/mAnen/dd' { = ci c/mAne n/da') "I my-house is" (cf. 

Chipewyan ne/da "she sat") 

If the verb form consists, properly speaking, of the stem alone, 
without prefix of any kind, an inorganic a- completed by a 
consonant that depends for its form on the first consonant of 
the stem is prefixed for the third person: 

As/se/Ve "he must cry" (<*j^//V; cf. dd/As/se "I'm not 

crying" <*Ac/se) 
do / wa / aI / W "he will become" (<*/^; do/wa is adverb 

not influencing form of verb proper. That -/- is here 

no third modal element is shown by forms like Hupa 

o/le "let him become") 

This /I- at the beginning of a third personal form appears also 
when the verb begins with a third modal element: 

aI/Az "he sneezes" 

In this respect Chasta Costa differs from Kato, which need 
have nothing preceding the stem; with As/se compare Kato 
tee' "he cried." 

In the third person of definite tenses with second modal 
7- or n- prefix this element is followed by -I, in case there is no 
third modal prefix present. This goes back, without doubt, 
to nasalized -i- or -A-, in turn reduced from original -it)- (or 
-At]-). This nasal element, characteristic of definite third 
personal forms (except such as have 6-, Athabascan s-, as second 

74 VI American Indian Languages 2 


modal prefix) is found also in Hupa (-in-), Kato {-un-), and 
Chipewyan (-«-, -in-). Examples are: 

Ve/ni/lat "he drowns" (cf. Kato tc'n/nul/lat "it floated 

there," nul- assimilated from nun-; Ve/ni/lat also 

"you drown") 
yi/dac "he dances" (cf. Kato tc' /gun/dac/kwan "he had 

yd/yi/t!a "it flies" (cf. Hupa na/win/tau "it will settle 

down;" ya/yi/ila also "it flies") 
di*/s'aV "it pains" {dl*-<*dir]-, contracted^^ from *de/- 

yir}-; cf. Hupa du/win/tcat "it got sick") 

Rather hard to understand is: 

Ve/An/yAl/lAl "he is sinking in the water" 

One would have expected -7!-, not -yAl- (as seen above, -/- 
is not third modal, but inorganic). Is yA- reduced from first 
modal ye-, this form being indefinite in tense? 

In Hupa this -in- does not seem to be found before third 
modal prefixes; in Chipewyan -n- {-in-) may, however, occur 
before -I- and, as inferred from Father Legoff's Montagnais 
paradigms, also -/-. As for Chasta Costa, what examples are 
available on this point show that -i- does not occur before -/- 
(e. g. yd/yAl/gAB "he climbs"). For I- verbs I have no safe 
example. Before -/- it seems that -i- is present in some cases, 
not in others: 

tcld/yit/sih/la "he pointed with his finger" 

but, without -i-'. 

na/da/yAt/t!d "he is bathing" 

On the whole, it seems possible that Athabascan -Ar]- (or 
-An-) was originally a more freely movable element than it has 

" Parallel in form to Hupa verbs belonging to Class I, Conjugation 1 D, in which 
prefixed first modal d- or deictic k^l- contracts with -it}, w- (Athabascan 7-) being lost. 
See Goddard, op. cit., p. 113. 

Six: Athabaskan and Na-Dene Languages 75 


become in e. g. Hupa, being required 'by certain verbs in their 
definite tenses, but not by others. This is suggested also by 
Father Legoff's Montagnais paradigms. 

Third Modal Prefixes. There are three of these: -/-, 
-/-, and -t-\ they always complete a syllable immediately before 
the stem. -I- is characteristic of many verbs which are either 
transitive or, at any rate, imply activity directed outward; 
in some cases, however, this significance is not obvious. As 
we have seen, this -I- becomes -/- in the first and second persons 
plural. Examples of -1-{-l-) are: 

Bid/Val "I kicked him" 

tB!AB/da caI nd/6il/sl "story to-me you-told, made" 

necl/l "I'm looking at him" 

ndl/de "he washes (something)" 

nail Lib "he writes" 

tc!Al/t!d "he sucks" 

yu/wis dAl/nl "he whistles;" yu/wls dAcl/nl "I whistle"'^ 

xwAn/ne/diL/ya "you win" 

If nothing precedes this element, it seems (unlike -/-) to begin 
its word without preceding inorganic A-: 

dd/wa/l/t'dt/fiAl "they will be broken" {do/wa is merely 

l/tH "he is important" 

Verbs in -/- are regularly intransitive; they denote states 
of mind or bodily activities that may be thought of as self- 
contained, not directed outwards. A reflexive meaning is 
sometimes apparent. After first person subjective -c- (s-) 
it always appears as -/-. Hence the first person singular, the 
first person plural, and the second person plural of -/- verbs 
and /- verbs are always alike (but contrast did- < *sAd- with 
5Asi<*sAd-). As -/-, when standing after s, becomes -I- also 
in the third person, the second person singular alone remains 

'- This verb is irregular, inasmuch as -/- does not occur in the second person singular: 
yA/wis di/nt "you whistle." 

76 VI American Indian Languages 2 


as an infallible criterion of whether a verb belongs to the l- 
class or /- class. Examples of -/- are: 

ya/dAl/yAc "he is ashamed" 

xwAcl/1 "1 believe;" xwil/i/ha "do you expect?" 

VAl/dAc "he runs" 

yd/yAl/gAd "he climbs" 

nd/dAl/de "he washed himself" 

aI/Az "he sneezes" 

VAI/xwaB "he coughs" 

na/xAt/da/yAl/el "they are bathing" 

nd/xwAl/ye "he plays" 

nd/dAl/nic "he works" 

tddsh/se "he cries;" tela I HI I se "you cry" 

tc!dcL/se/Ve "I'll cry;" tc!dl/se/Ve "you'll cry" 

hd/yAci/i "I stop, cease;" hd/il/i "stop!" 

Verbs in -/- are also intransitive. Examples are: 

Vd/yit/nd "you drink" 

ye hat I ne I la "he bit (it)" 

tdd/yit/sii/la "he pointed with his finger" 

qlwAt/dadt/gAc "it lies thrown down on top, 

yAct/ld "I laugh" {-t/lo may, however, represent 

original Athabascan -did "to laugh") 
ya/da/yit/dja "we are ashamed" 

There may be a passive significance in : 

q/wAt/tclAt/dja "whereon it is eaten, table" 

With iterative na-: 

na/yd/6Adt/xwl "I vomit" 

Verb Stems. The stems that have been determined for 
Chasta Costa are: 

-'d,-'a "to have position, to be" (cf. Hupa -ai, -a; Kato 
-*ai*, -'a'): d/'d/la "(one table) was" 

Six: Athabaskan and Na-Dene Languages 77 


-'AC "to bring" (cf. Hupa -an, -uii, -auw "to transport 

round objects"; Kato -*qn, -^qc): yAn/na/'Ac "he 

will bring it" 
-'An "to bring to a halt, stop" (perhaps another form of 

preceding stem): 7td/ni/An "stop him!" 
-Az "to sneeze": acl/Az "I sneeze" 
-ai "to come" (cf. Chipewyan -^as, -*az, -*ais "to travel, 

used of two persons only"?) an/yi/'al "come on!" 
-ci, -el "to bathe (plur. subject)" (cf. Chipewyan -el, -eL, 

-uL "to move on the surface of water"): na/da/yil/el 

"we are bathing" 
-'I "to see, look at" (cf. Hupa -en, -in; Kato -*ln' "to 

look"): ye/dic/l "I saw him" 
-i "to stop, cease" (cf. Hupa -en, -in "to do, to act, to deport 

one's self"?): hd/yAch/i "I stop (laughing)" 
-I "to believe, expect": xwAcl/i "I believe" 
-ya, -yAc, -yic "to go, come" (cf. Hupa -yai, -ya, -yauw; 

Kato -yai, -ya, -yac): Ve/6ic/ya "I go;" t'Ac/yAc/t'e 

"I must go" 
-ya "to eat" (cf. Hupa -yan, -yun, -yauw; Kato -yan^ -ytl*): 

tclAy/ye/Oic/ya "I eat;" qIwAt/tcfAt/dja(<-t/ya) 

"whereon one eats, table" 
-ya "to win" (cf. Kato -yan, "Kato Texts," p. 146, 1. 13; 

not listed in "Elements of the Kato Language"): 

xwAn/ne/dii/ya "you win" 
-yan "to upset" (cf. Kato -yan "to clear off"?): do/wd/An/- 

nd/yan/nAl "he will upset them" 
-ya, -yAc "to be ashamed" (cf. Kato -yafi "to be ashamed"): 

ya/dAl/yAc "he is ashamed;" ya/da/yit/dja {<-t/ya) 

"we are ashamed" 
-ye "to play" (cf. Hupa -ye "to dance"): nd/xwAl/ye "he 

-lal "to sleep" (cf. Hupa -lal, -lai; Kato -lal, -Iqi): t'cd/- 

lal/la "he's been sleeping" 
-IaI "to sink in water;" -lat "to drown" (cf. Hupa -lat, 

78 VI American Indian Languages 2 


-la "to float"): re/Ari/yAc/Ul "I am sinking;" fe/- 

nic/lat "I drown" 
-le "to become" (of. Hupa -len, -lin, -hi, -le; Kato -lin% 

-le): do j wa I aI / le' "it will become" 
-lee "to wager, bet": ne/tclilc/lec "I'll bet you" 
-lee "to smoke": mAl/Ve/tc!At/ts!Al/lee "wherewith it is 

-Id "to laugh, smile" (cf. Chipewyan -did, -dlok'): yit/lo 

-lIo "to write" (cf. Hupa -Lon, -Ld, -Low, -Loi "to make 

baskets, to twine in basket-making;" Kato -Loi, 

-Ld, -Lon): nail Lib "he writes" 
-na "to drink" (cf. Hupa -nan, -nun-, Kato -7iqn): Vd/- 

yAct/nd "I drink" 
-na "to lie" (cf. Kato ndLjiin/na'' "were left"?): qlwAt/ 

dasL/nd "it was lying on it" 
-ne "to bite, seize with one's teeth": ye/ydt/ne/la "he 

bit it" 
-nl, -71 "to make a sound, to say" (cf. Hupa -ne, -n "to 

speak, to make a sound;" Kato -nl, -ne, -n, -nee, -nlL): 

yu/wls dAl/ni "he whistles;" a/dJAn "he says" 
-nic "to work": nd/dAl/nic "he works" (cf. Montagnais 

-7ii "exprime Taction des mains"^^ 
-dAl "to wash oneself (plur. subject)" (cf. Hupa -sel, -scl 

"to be or to become warm;" Kato -sll "to steam," 

-sUl, -suL "to be warm"): tc/At/t'it/dAl "we wash 

-del "to throw": yd/yi/del "you threw" 
-se "to cry" (cf. Hupa -tcwu, -tcwe "to cry, to weep;" Kato 

-tcec, -tee'): tcIdsL/se "he cries" 
-61 "to cause" (cf. Hupa -tcwen, -tcwin, -tewe "to make, to 

arrange, to cause;" Kato -tcin, -let, -tclL): ndcl/sl 

"I cause" 
■sil "to point with one's finger": tc!d/yit/siL/la "he pointed 

with his finger" 

" Father L. Legoff, "Grammaire de la Langue Montagnaise," p. 139. 

Six: Athabaskan and Na-Dene Languages 79 


-da', -da "to sit, stay" (cf. Hupa -dai, -da; Kato -da, -dai): 

da I Bi I da "you are sitting" 
-dAc "to run" (cf. Hupa -dai, -daL, -dauw "to pass along, 

to go, to come;" Kato -dac "to travel"): VAl/dAc 

"he runs" 
-dac "to dance" (cf. Kato -dac "to dance"): nt/dac "you 

-de "to wash (sing, subject)" (cf. Kato -dec, -de'): nd/- 

dAl/de "he washed himself" 
-t'di "to kick" (cf. Hupa -taL, -tul, -tuL, -tal "to step, to 

kick;" Kato -tal\ -tqL): dicl/t'dl "I kicked him" 
-t'Ac "to lie down, go to bed" (cf. Hupa -ten, -tin, -tuw "to 

lie down;" Kato -tin, -tuc): dd/nAc/VAc "I go to bed" 
-Vat "to break, go to pieces" (cf. Chipewyan -taL, -till "to 

break"): dd/wa/l/t'dt/uAl "they will be broken" 
-Ve "to want" (cf. Hupa -te "to look for, to search after"?): 

dd/ucL/t'e "I do not want;" de/uci/t'e "what I want" 
-t'J "to be, make valuable" (cf. Carrier til/thi "thou makest 

him valuable, treatest him as important"): l/Vi 

yAn/na/'d "he brags" 
-t.'a "to fly" (cf. Hupa -tau; Kato -t'ac, -t'a'): do/ydc/tla 

"I won't fly" 
-tie "to be of (that) sort" (cf. Hupa -te; Kato -t'e): do/dAt 

'An /tie "there is not anywhere (one like him)" 
-t!d "to swim, bathe (sing, subject)": nac/t!d "I swim, 

-t!d "to suck" (cf. Kato -t'ot): tclAi/t/d "he sucks" 
-tdfi "to sit (plur. subject)" (cf. Hupa -tse; Chipewyan 

-d'i): da/de/dil/tdii "we are sitting" 
-tshit "to hurt, pain (intr.)" (cf. Hupa -teat, -tea "to be 

sick, to become ill"): di*/s'at' "(my eyes) hurt" 
-7« "to go about, live" (cf. Hupa -wai^ -wa "to go, to go 

about;" Kato -ga, -gai): nd/ya "he goes about, lives" 
-xe "to paddle" (cf. Hupa -xen, -xuw "to float, used only 

of plural objects;" Kato -ke' "to bathe (plural only);" 

80 VI American Indian Languages 2 


Chipewyan -kl "to paddle a canoe, to travel by canoe") : 

ndc/'xe "I paddle" 
-xwAd "to cough" (cf. Kato kos "cough," as noun; Carrier 

xwces): VAl/xwAd "he coughs" 
-xwl "to vomit" (cf. Carrier khti "vomiting," as noun): 

na/yd/dAdt/xwi "I vomit" 
-gAd "to climb" (cf. Hupa -kas "to throw"): yd/yAl/gAd 

"he climbs" 
-gAC "to throw"? (cf. Hupa -kas "to throw"): qfwAt/dadt/- 

gAc "it lies thrown down on top, table-cloth" 

It will be observed that several verb stems are restricted 
in their use as regards number of subject (or object). This 
trait is characteristic of Athabascan, as also of other American 
linguistic stocks. 

Definite and Indefinite Tenses. My material on 
Chasta Costa is not full enough to enable me to give a satis- 
factory idea of its tense-mode system. It is clear, however, 
that absolute time (present, past, future) is quite subordinate 
to whether activities are thought of as taking place at some 
definite time (generally present or past) or are more indefinite 
as to time occurrence. Indefinite forms are apt to be used 
for general statements that apply irrespective of any particular 
time, for future acts, for negative (particularly negative future) 
acts, and regularly for imperative and prohibitive forms. The 
contrast between definite and indefinite present forms comes 
out in: 

definite: nd/da/yAct/t!d "I bathe" (i. e. am. now engaged 

in bathing) 
I indefinite: cAl/qlwe na/dAct/t!d "I'm used to bathing" 
[ (here bathing is not restricted as to time) 

f definite: tc!d/sAsL/se "I cry" 
[indefinite: Vwi/dAn As/se "I always cry" 
definite: yeSt/lo "he breaks into laughter" (i. e. laughs 

at one particular point of time) 
indefinite: ydt/lo "he laughs" 

Six: Athabaskan and Na-Dene Languages 81 


[definite: xAt/Ve/lal/la "they have been sleeping" (may 
\ be said of them at moment of waking up) 
[indefinite: xAt/VAl/lal "they sleep" 

Futures, as we shall see, are explicitly rendered by suffixing 
-Ve to present (generally indefinite) forms; but simple indefinite 
forms, particularly with adverbs pointing to future time, may 
often be used as futures in contrast to definite present forms. 
Examples are: 


definite: na/da/yil/el "we are bathing" 

indefinite: nd/dil/ei "we'll bathe" 

definite td!AB/da caI na/dU/si "story to-me you-told" 

indefinite: xiin/de id! Ad /da uaI ndd/sl "tomorrow story 
to-you I-tell" 

definite: n/ye/Bic/l "I saw you" 
I indefinite: xAL/tsH/dAn do/wan/yAc/l "this-evening I'll- 
■[ see-you" 
I definite: Ve/dic/ya ''I go'' 
[indefinite: xun/de Vac/yAc "tomorrow I'll-go" 

Negative presents or futures are regularly expressed by pre- 
fixing do "not" to indefinite forms; when more explicitly future, 
-Ce is suffixed to them. Examples of indefinite forms preceded 
by do are: 

definite: tdac yd/yi/t!a "bird is-flying" 

indefinite: dd/yd/t!a "he won't fly" 

definite: na/yd/SAdt/xwi "I vomit" 

indefinite: do/na/yAct/xwi "I do not vomit" 

definite: tc!d/sAsL/se "I cry;" tc!d/dil/se "you cry" 

indefinite: do/As/se "I'm not crying;" do/l/se "you're not 

definite: ye/Bic/i "I saw him;" c/ye/Bi/l "you saw me" 
indefinite: do/yAc/t "I didn't see him;" do/ld/c/yl/l 
"you didn't see me" 

[definite: Ve/Bic/ya "I go;" VcBjya "he goes" 
indefinite: dd/VAc/yic "I'll not go;" do/VAc "he won't 

I go" {<*VAlyAc) 

82 VI American Indian Languages 2 


Imperatives are simply second person subjective indefinite 
forms. Examples are: 

yi/l "see him!" c/yi/l "see me!" 

nel/l "look at him!" (identical with indefinite present: 
nei/l "you're looking at him"); ne/cil/l "look at me!" 
nd/ni/An "stop him!" 

Prohibitives are simply imperative forms preceded by ia: 

la/yt/l "don't see him!" 

First person plural indefinite forms may have hortatory signifi- 

nd/xwil/ye "let us play!" 

As regards form, definite tenses are primarily distinguished 
from indefinite tenses by the presence of second modal prefixes 
in the former, often also by the appearance of the first modal 
prefixes in a fuller form than in the latter; the presence of 
-I- or -i- in certain third person definite forms may also be 
recalled. It seems, further, that certain adverbial prefixes 
which have a short vowel (even though in an open syllable) 
in definite forms lengthen it in corresponding indefinite forms: 

definite: nd/da/yAct/t!d "I bathe;" na/da/yit/t!d "you 

bathe;" na/da/yAt/t!d "he's bathing" 
indefinite: nd/dAct/t!d' /Ve "I'll bathe;" nd/dit/t!6/Ve "you'll 

bathe;" nd/dAt/t!6/Ve "he'll bathe" 
definite: na/da/yil/el "we are bathing;" na/da/ydl/ cl 
"ye are bathing;" na/'xAt/da/yAl/el "they are bathing" 
indefinite: nd/dil/el "we'll bathe;" nd/dol/el "ye will 

bathe;" nd/xAt/dAl/el "they'll bathe" 
[definite: tc!a/yd/Bil/se "we cry;" tc!a/yd/ddl/se "ye cry" 
undefinite: tc!d/ydl/se "we'll cry;" tdd/ydl/se/t'e "ye will 
' cry" 

These changes of quantity, however, are doubtless only second- 
arily connected with change of tense, as indicated, e. g., by 

Six: Athabaskan and Na-Dene Languages 83 


tela- in definite singular forms: tcId/sAsL/se "I cry;" tcld/- 
dil/se/ha "do you cry?" It is very likely that we are dealing 
here primarily with considerations of syllabic and quantitative 
rhythm or balance.^* 

In Hupa Goddard has exhaustively shown that verb stems 
often assume different forms for different tenses and modes. 
This is very likely also true to a considerable extent of Chasta 
Costa, but I have but little material bearing on this point. 
A quantitative change is found in: 

j definite -el: na/da/yil/el "we are bathing" 
(indefinite -el: nd/dil/el "we'll bathe" 

-c characterizes indefinite forms in : 

definite -ya: Ve/Bic/ya "I go;" Ve/Si/ya "you go" 
indefinite -yAc: t'Ac/yAc/Ve "I must go;" Vl/yAc/Ve "you 

must go" 
negative indefinite -yic: d5/VAc/yic "I'll not go;" Id/t'l/- 

1 yic "don't go!" 

f definite -ya: ya/da/yit/dja{<-t/ya) "we are ashamed" 

^ indefinite -yAc: ya/dAcl/yAc "I am ashamed" 

Pronominal Objects. Pronominal objects are regularly 
prefixed to the verb. They come before deictic and first modal 
elements, but after adverbial prefixes. Thus, while not as 
thoroughly immersed in the verb form as the subjective pro- 
nominal elements, they cannot well be considered apart from 
it. The third person singular object is not designated. In 
form the objective elements are, on the whole, identical with 
the possessive pronominal prefixes of the noun. They are: 

Singular 1. c- Plural 1. 7i6- 

2. n-, ne- 2. no- 

3. — 3. xo- 

" Hardly stress accent as such. I cannot help feeling that such rhythmic phenomena 
will turn out to be of fundamental importance for Athabascan generally. 

84 VI American Indian Languages 2 


"He — them" or "they — them" is expressed by means of xl-. 
c- and n-, when standing at the beginning of a verb form, take 
no inorganic a- before them (contrast subjective Ac-). 

The definite forms of yed/l "he saw him" with combined 
pronominal subject and object are: 

With first person singular object: 

Sing. 2. c/ye/di/l "you saw Plural 2. c/ye/do/l 
3. c/yed/t 3. c/xA/yed/i 

With second person singular object: 

Sing. 1. n/ ye/Sic/ 1 "I saw Plural 1. n/ye/dit/l 
3. n/ yed/l 3. u/xa/ yed/l 

With first person plural object : 

Sing. 2. no/ye/dd/l "you saw Plural 2. no/ye/do/l 
3. no /yed/l 3. no /xa/ yed/l 

For no/ye/dd/l "you saw us" one would have expected *«<?/- 
ye/dl/l. It seems that "ye saw us" has been extended in its 
usage to embrace also "you (sing.) saw us." It may indeed 
be that my data on this point rest on a misunderstanding, but 
there seems to be something analogous in Hupa. "You (sing.) 
are picking us up" would be expected in Hupa to be *yun/nd/- 
hil/luw {Ml- assimilated from hin-). Instead of this form, 
however, Goddard lists yun/no/hol/luw, which is not identical 
with but seems, as regards its second 6- vowel, to have been 
influenced by ymi/nd/ho/luw "ye are picking us up."^^ 

With second person plural object: 

Sing. 1. no /ye /die /I "I saw Plural 1. no/ye/dit/l 
you (pi.)" 
3. no /yed/l/ 1 1 3. no /xa/ yed/l /la 

"Goddard, op. cit., p. 186. 

Six: Athabaskan and Na-Dene Languages 85 


With third person singular object: 

Sing. 1. 7^/^^cA "I saw him" Plural 1. ye /Bit/ 1 

2. ye /Si/ I 2. ye/ 66 /I 

3. ye6/l 3. xi/yeB/i 

With third person plural object: 

Sing. 1. x6/ ye/ die/ 1 "I saw Plural 1. xo/ye/dit/l (heard 
them" also as xo/we-) 

2. ye /Bo /I 2. xo/ye/Bo/l 

3. xl/dA/yeB/l 3. xi/yeB/i/la 

Here again, one would have expected *x6/ye/Bi/l for "you 
(sing.) saw them." As it is, "you (pi.) saw him" seems to be 
used also for "you (sing.) saw them," both forms being logically 
parallel in that both involve a second person — third person 
relation, only one of the two persons, however, being plural. 
Objective forms of indefinite tenses of this verb are: 

With first person singular object: 

do/wa/c/yl/i^/Ve "you'll see me" 
dd/ld,/c/yi/i "you didn't see me" 
c/yi/l "see me!" 
c/ya/i*/Ve "he'll see me" 

With second person singular object: 

do/wa/n/yAc/l "I'll see you" 

With third person singular object: 

do/wa/yAc/l "I'll see him" 
do/yAc/i "I didn't see him" 
do/yAc/l/Ve "I won't see him" 
do/wa/yl/i/Ve "you'll see him" 
yi/l "see him!" 
la/yi/i "don't see him!" 

Objective forms of indefinite tenses of ne-l-'l "to look at" are: 

With first person singular object: 

ne/cU/i "look at me!" 
ne/cAl/l "he looks at me" 

86 VI American Indian Languages 2 


With third person singular object: 

neci/l "I'm looking at him" 

do/ned/l "I'm not looking at him" 

nel/i "you're looking at him;" "look at him!" 

nel/l "let's look at him!" 

ne 1x6 1 oil I "you (plur.) look at him" 

This last form may, likely enough, have been mistranslated 
for "you (plur.) look at them" (of. xd/ye/Bo/l above). 
Other forms with first person singular object are: 

cAsl/sl "he lets me, causes m.e to" 
cAsl/t'dl "he kicked me" 

With second person singular object: 

ne/tc!ilc/lec "I'll bet you" 

Passives. As in Hupa, pronominal subjects of passive 
verbs are objective in form. From ne-l-'l are formed: 

nes/ts!Ai/i "I am seen" ne/no/tslAl/t "we are seen" 

nen/tslAl/l "you are seen" ne/no/tslAl/l "ye are seen" 

ne' I is! All I "he is seen" nelxblislAlll "they are seen" 

IsIaI-, which appears in these forms, probably contains third 
modal -I- preceded by deictic tsl- implying indefiniteness of 
logical subject: "man sieht mich." Apparently connected 
with this IsIaI- is IsIaI- in: 

mAllt'eltclAtltslAlllec "wherewith it is smoked, smoking 

Verbal Suffixes. A number of enclitic elements of tem- 
poral or modal significance are found rather loosely suffixed 
to verb forms. These, so far as illustrated in our material, are: 

-t'e future particle (cf. Hupa -te, -tei; Kato -tei, -telle): 
aclI AzIVe "I shall sneeze" 
VAclldllVe "I shall sleep" 
naldAcililo' iVe "I shall bathe" 
nedlllVe "I'll look at him" 
dolyAclilVe "I won't see him" 

Six: Athabaskan and Na-Dene Languages 87 


nd/dAci/nic/t'e "1 shall work" 
do/nd/dACL/nic/t'e "I shall not work" 
tcIdcL/se/re "1 shall cry" 
do/As/se/re "I'll not cry" 
tc!dl/se/t'e "you will cry" 

-t'e seems to imply obligation to some extent, as well as simple 
futurity, as is shown by its translation as "must" in some 
cases : 

As/se/t'e "he must cry" 

rAc/yAc/re "I must go" 

t'l/yAc/t'e "you must go" 

All forms with suffixed -t'e, it will be noticed, are indefinite; 
none has been found that is definite. 

-nAl seems to be used for future acts: 

dd/wd/An/nd/ya?i/nAl "he will upset them" 
do I wa HIV at I nAl "they will be broken, go to 

-ha, -hd interrogative: 

nd/xwil/ye/ha "are you playing?" 
Ve/Bo/lal/ha "have ye been sleeping?" 
no/ye/Bd/l/ha "did you see us?" 
nel/i/ha "did you look at him?" 
tcld/dil/se/ha "did you cry?" 
do/o/se/ha "are ye not crying?" 
As/se/Ve/ha "will he cry?" 
tcId/ydl/se/Ve/ha "will ye cry?" 
fd/yit/nd/hd "do you drink?" 

ha seems to both precede and follow in : 

ha/xwil/i/ha "do you expect?" 

-/a probably inferential (cf. Hupa -xo/laii, -xo/liiii): 

t'e/OAd/lal/la "I've been sleeping" (said on wak- 
ing up) 
Ldd/d/la "there was one (table)" 
txAs/xe/la "(evidently) rich" 

88 VI American Indian Languages 2 


Probably also in : 

do I wi I la "of course" 

-la seems also to be used of simple narrative in past time, with 
very weak, if any, inferential force: 

tc!d/yit/siL/la "he pointed with his finger" 

ye/ydt/ne/la "he bit it" 

nAn/ndd/yd/la "he went around it" 

no/yeB/l/la "he saw you (plur.)" (cf. no /yed/i "he saw us") 

no/xA/yeB/i/la "they saw you (plur.)" (cf. no/xA/yeB/l 

"they saw us") 
xl/yeB/l/la "they saw them" (cf. xi/yeB/l "they saw him") 

I do not know whether the contrasts in person and number 
found in the last three pairs of forms are real or only apparent. 

-le of unknown significance 

nd/xol/ye/le "you (plur.) play" 

Syntactic Combination of Verbs. Two verb forms some- 
times combine syntactically, one depending on the other. The 
second verb is subordinate to the first in: 

dd/ucL/Ve nd/xwACL/ye "I-do-not-want I-play," i. e. "I 

don't want to play" 
hd/yACL/l yAct/lo "I-stop I-laugh," i. e. "I stop laughing" 
hd/il/l yit/lo "stop laugh!" i. e. "stop laughing!" 
ha/xwil/i/ha yAn/na/Ac "do-you-expect he-will bring?" 

It seems that sometimes the first verb, which is then a third 
personal form, acts as a sort of complementary infinitive to 
the second: 

yd/yi/t!a Bicl/sl "he-flies I-make-him," i. e. "I let him 


ya/yi/t!a cAsl/sl "he-flies he makes-me," i. e. "he lets me 

l/t'i yAn/na/'d "he-is-important he-has-for(?)," i. e. "he 

brags about him" 

Six: Athabaskan and Na-Dene Languages 



l/t'l" yAn/na/'a^** xa/dAt^" 

Make important he has his own 

'An/t!e«2 na/ya" a/dJAn.«^ 

is like him moves about," he says. 




"What I want 



hi/t !i«6 

that thing 

** Wolverton Orton claimed not to know any regular Chasta Costa myth texts. The 
following, which is merely an English joke anecdote taken from a popular periodical that 
happened to be lying about and translated into Chasta Costa by Mr. Orton, will at least 
serve to give some idea of Chasta Costa word order and sentence construction. 

"^, third modal element, -/'t, verb stem. Cf. Carrier tU/thi "thou makest him 
valuable, treatest him as important." 

**7/in- and na-, adverbial prefixes, -'o, verb stem. For na/'a- "to have," cf. Hupa 
naii/a/te "you will have." "He has his dog made valuable, treated as important," i. e., 
"he brags about his dog." Indefinite tense, because statement is general and does not 
refer to any one point of time. 

" X-, third personal pronominal element. -d/dAt, reflexive possessive element. 

*" Possessed form of it "dog." Observe change of I- to /-, and suffixing of -tele. Cf. 
Hupa -lin/k{yi)e; Chipewyan Ltn/k'e. 

" do, negative adverb. -dAl, postpositive element. 

" 'a-, reduced from 'a-, prefix used with verbs of saying, doing, and being. It is 
probably equivalent to indefinite demonstrative: "(there is of) that (kind)." -n-, first 
modal element, -//e, verb stem. Cf. Hupa Mn//c "there is;" Kato qn/^'e "it is;" Chipe- 
wyan an/t'e/hi/k'e "it was." Indefinite tense, because statement is general. 

" nd-, adverbial prefix, -ya, verb stem. "Moves about," i. e., "is living, is to be 
found": "there is no (dog) like him anywhere." Cf. Hupa na/wa "they were there;" 
Kato na/ga/kwqn "he had walked;" Navaho na/oa, i. e., na/ya, "he is going about" 
(quoted from Goddard, Analysis of Cold Lake Dialect, Chipewyan). Indefinite tense; 
general statement. 

" d-, prefix used with verb of saying; see note 62. dj-, third personal deictic prefix; 
or perhaps dJA- = *dyA-, reduced from *dye-, first modal prefix d- and third person deictic 
prefix ye-, -n, verb stem. Probably definite in tense, though it shows no second modal 
prefix; cf. Hupa present definite third singular a /den. 

" de, apparently relative in force, u-, adverbial prefix, -c-, first person singular 
subjective element. -L-, third modal prefix, -t'e, verb stem. Indefinite tense; general 

•• ht, demonstrative stem, -tli, suffix applying, it would seem, to things. Perhaps 
ht/tii is assimilated from "hx/tla; for -tla, cf. Chipewyan t!a "that; often used to point 
out one of several persons or things characterized by a descriptive phrase or clause." 


90 VI American Indian Languages 2 


s/ts.'An/na/'Ac" a}/dAc/nl/dAn.«» do t'wi/de«» La a/dJAn. 

to me he brings when I tell him." "Not everything," one says. 

H/tli/ni^" mAn/me/q !e" t'/yl/i." xat q!wAt/tc!At/dja" 

Dog-owner around in house he looked around. Then table 

Ld^/a/la^^ mAn/me" wAs/xe q !wAt/da^/gAc" na^/L!6" 

one there was in house. Good table-cloth, paper 

hi q!wAt/dasL/na^» i/yi/tc'u^" mAl/t'e/tc!At/ts!Al/lec.»« 

that was lying thereon, that wherewith it is smoked. 

" S-, assimilated from c-, objective (or possessive) first person singular pronominal 
element. -Is! An, postposition, no-, adverbial prefix, -'^c, verb stem. Cf. Hupa do/- 
xo/lin/na/ta/auw "he won't carry." Indefinite tense; general statement. 

** a-, as in notes 62 and 64. -1-, postposition; refers to implied third person indirect 
object of verb, -d-, first modal prefix, -c-, as in note 65. -nl, verb stem. -dAn, post- 
position; here used to subordinate verb. Cf. Hupa Hn/niL/duw/ne "I am telling you." 
Indefinite tense; general statement. 

^' t'wi, pronominal stem denoting totality, -de, indefinite demonstrative stem. 

'" /t "dog." -tli/nt "one who has;" evidently contains common Athabascan suflfix 
-n, -nl "person." 

" m/4n "house." -me/q!e, compound postposition. 

" /'-, first modal prefix reduced from t'e-. 71-, first modal prefix 7- reduced from ye-, 
-I- remaining unexplained. -I, verb stem. According to this analysis, t' /yt/i is indefinite 
in tense; this seems hard to understand, as it refers to one act in past time. Another 
analysis seems more likely: /'-, instead of or misheard for t'e-, form regularly used in 
definite tenses; 7-, second modal prefix; -?-, definite third person ending for 7- verbs. 
Cf. Hupa tcii I te I we I in I il "he looked about as he went along." 

" q/wAt-, postposition "upon" used as adverbial prefix, k!-, deictic prefix here indi- 
cating indefiniteness of object. -/-, third modal prefix presumably with passive force. 
-dja, from -ya after -/-, verb stem "to eat." "It is eaten thereon," i. e., "table." 

^* Ld, numeral "one," to which verb proper, 9/d/la, is attached. 6-, second modal 
prefix, a-, verb stem, -la, verb suffix. Definite past tense, because referring to definite 
point of time in narrative. Cf. Chipewyan de/*a/ht/k'e/lai "(lake) was there." 

" -me, postposition. 

^^ qlwAt-, as in note 73. da-, adverbial prefix. 6-, second modal prefix. -^, third 
modal prefix. -gAc, verb stem. Verb form ("it lies thrown on top") used as noun. 

" nd-, adverbial prefix, d-, second modal element. -lIo, verb stem. Verb form 
("whereon there is writing") used as noun. 

'* qlwAt-, as in note 73. da-, adverbial prefix, s-, second modal prefix. -L-, third 
modal prefix; doubtless original -/- changed to -1-, -L-, because of preceding s-, which in 
turn is prevented by it from changing to d-. -na, verb stem. Definite past tense. 

" Analysis uncertain, presumably demonstrative in force. 

^ ntAl- "therewith" consists of pronominal stem m- followed by postposition -1-. 
t'e-, adverbial prefix, tc!-, deictic prefix indicating indefiniteness of object. -/-, consonant 
borrowed from following -ts!, to complete syllable begun by k!-. tsUl-, apparently passive 
in force, -lee, verb stem. Verb form used as noun: "smoking materials." 

Six: Athabaskan and Na-Dene Languages 91 


txAs/xe/la^i d6/at/t!L82 q!wAt/tc!At/dja xa s/tslAn/na/Ac 

He was rich bachelor. "Table quickly he'll bring to me," 

dJAn/la«3 H/t!i/ni. d6/LAn»^ xwacl/P» djAn/la da/at/t !i/n!. 

said dog-owner. "Not much I believe it," said bachelor. 

ne/tc!uc/lec.»« c6"/dji." an/yl/aLS* djAn/la H'/t!i/ni tclo/yit- 

"I'll bet you." "All right!" "Come here!" said dog-owner, he pointed 

/siL/la«« H q !wAt/tc !at/dja lat/dAn'" nAn/na^/ya/la'^ 

with his finger. Dog table once he went around. 

xat ye/yat/ne/la^^ ma^/dAn.^^ la djAn/la d6/at/t!i/ni 

Then he bit it at edge. "Don't!" said bachelor, 

t'wi/de do/wa/na/yan/nAl.^" do/wl/la»*djAn/laq!wAt/tc!At/dja 

"everything he will upset." "Of course," he said, "table 

^^txAs/xe, adjective stem "rich;" perhaps related to wAs/xe "good." -la, verb 
suffix of probably inferential value. 

^"^ do, negative, at = at! "wife." -t!i, noun suffix denoting "one who has." "One 
who has no wife," i. e., "bachelor." 

*^ dJAn, as in note 64. -la, verb suffix. 

^* do, negative, lah, adverb "much." 

** XW-, adverbial prefix, -c-, first person singular subjective pronominal element. 
-/.-, third modal prefix; from -/-, because of preceding -c- (cf. note 98). -I, verb stem. 
Indefinite present, negative adverb preceding. 

^ne-, second person singular objective pronominal element, tclu-, adverbial prefix; 
very likely really compound of deictic element tcl- (indicating lack of specified object, 
namely wager) and modal 6-, il- denoting future imperative, -c-, as in note 85. -lee, 
verb stem. Indefinite present, because of future or slight hortatory meaning: "let me 
bet with you!" Cf. Chipewyan tus/be "let me swim." 

*' With co"-, cf. Hupa -hwon "good;" Kato -con "to be good." 

*' an-, adverbial prefix. 7-, first modal prefix, -i-, second person subjective pro- 
nominal element. -aL, verb stem. Indefinite tense, used as imperative. 

*^ tclo-, adverbial prefix; perhaps compound of deictic element tcl- (object pointed 
out is not specified) and first modal 6- of unknown significance, y-, second modal prefix. 
-t-, connecting element between second and third modal elements, characteristic of third 
person of definite tenses with 7-. -/-, third modal prefix, -sit, verb stem, -la, verb 
suffix. Definite past; marks point in narrative. 

^ Numeral adverb of la "one." -dAn, postposition. 

" HAn- and na-, adverbial prefixes. -0-, second modal prefix, -yd, verb stem, -la, 
verb suffix. Definite past; refers to definite point of time in narrative. 

^ ye-, adverbial prefix, -ya-, second adverbial prefix. -/-, third modal prefix, -ne, 
verb stem, -la, verb suffix. According to this analysis, this verb is indefinite in tense, 
which is difficult to understand. More plausibly, yat- may be considered as misheard 
for yAt-; 7- second modal prefix. In that case, it is definite past. 

^^ ma*^-, noun stem "edge." -dAn, postposition. 

" dd/wa, proclitic adverb indicating futurity, probably not with absolute certainty. 
na-, adverbial prefix, -yan, verb stem. -nAl, verb suffix. Indefinite in tense, because 
future in meaning. 

" Adverb containing inferential -la. 

92 VI American Indian Languages 2 


do/wa/An/na/yan/nAl9« t'wi/de do/wa/1/t'at/nAl" 0Ak/gwe 

he will upset, everything will go to pieces, in fragments 

s/tslAn/na/Ac. ha/xwil/i/ha^^ AL/tca/yi»« yAn/na/Ac'"" 

he will bring to me. Do you expect big thing he will bring here 

}a/me/q!e/ca.'"' wAs/xe If. na/ni/An'"' na/nl/An djAn/la 

all in one time? Good dog." "Stop him, stop him!" said 

d6/at/t!i/ni. d6/da/q!e>03 na/nAc/An'o" djAn/la H/tlT/ni 

bachelor. "Unable I stop him," said dog-owner, 

t'wi yAn/na/Ac/t'e'"^ xwAn/ne/^iL/ya'"« djAn/la d6/at/t!T/ni. 

"all he will bring here." "You win," said bachelor 

** As in note 94, except that another adverbial prefix, ah-, is present. 

" do/wa and -tiaI, as in note 94. i-, third modal prefix. -I'at, verb stem. 

'* ha, interrogative adverb, xw-, adverbial prefix, -i-, second person singular sub- 
jective pronominal element. -/-, third modal prefix. -I, verb stem, -ha, interrogative 
suffix. Indefinite present in tense. 

" A-, of unknown significance. L-, prefix common to several adjectives, -tcd/yi, 
adjective stem "big." 

^'"^yAtt- and na-, adverbial prefixes, -ac, verb stem. Indefinite tense, because point- 
ing to future time. 

"" ia, numeral stem "one." -me/q!e, compound postposition, -ca, found also with 
la alone: Id'^/ca "one." 

'"' nd-, adverbial prefix, n-, first modal prefix, -l-, second person singular subjective 
pronominal element. -An, verb stem. Imperative mode. 

103 Adverb containing negative do-. Perhaps -qle is postposition (cf. -me/qle). 

^^ no.-, n-, and -ah, as in note 102. -c-, first person singular subjective pronominal 
element. Indefinite in tense, because of preceding negative adverb. 

"* As in note 100. -Ce, future suffix; here used because idea of futurity is more explicit. 

">* xtf /in-, adverbial prefix, we-, first modal prefix. 6-, second modal prefix, -i-, 
second person singular subjective pronominal element. -L-, third modal prefix, -ya, 
verb stem. Definite present in tense. 

Six: Athabaskan and Na-Dene Languages 93 


A few Galice Creek words were obtained from Mrs. Punzie, 
a few Applegate Creek words from Rogue River Jack. These 
two Athabascan dialects are probably practically identical. 
5 indicates something acoustically midway between 5 and c; 
r (tongue-tip trilled) and / occur as reflexes of Athabascan n; 
nasalization (indicated by ') seems to occur, k' and k'w are 
found as contrasted with Chasta Costa x and xw. 

Galice Creek. 

ya'/k'ds "seeds (sp.?);" said to be called bdnax or b4yu 

in Chinook Jargon 
tcla/ba/d/k'wa's "brush used for medicinal purposes (sp.?)" 
L!d'/ddi "tar- weed" (cf. Chasta Costa ild'/de; Hupa 

yel/yat/ts!ai/yt "sunflower" (cf. Chasta Costa, tc I Al/y at /ts/e) 
giis "camass" (cf. Chasta Costa god; Hupa kos "bulbs") 
ddl/si "pine" (cf. Chasta Costa dAl/si; Kato dul/tclk) 
Id I lH "pine-nut" (cf. Chasta Costa nd/de) 
del res "manzanita" (cf. Chasta Costa dAJnAc; Hupa 

din/nuw; Kato tun/nuc) 
md'/ts!i "cat-tail" (cf. Chasta Costa mAt/tdi) 
sds/da' "oak" (cf. Chasta Costa cAc/dd') 
i/dd/ge "acorn" (perhaps misunderstood; cf. Kato i/taG 

"black oaks") 

Applegate Creek. 

k'q'/tc'u "goose" (cf. Chasta Costa xd'/tc'u; Kato ka') 
dAc/tc'u "grouse" (cf. Chasta Costa dAc/tc'u; Kato dAcf- 


94 y^ American Indian Languages 2 


dAclVe' Itc'u "bob-white, quail" 
k!ai'/dic/tca/we "ruffed grouse, pheasant" 
k'dn/ta/tc'u "pigeon" (cf. Kato kwl/ylnt) 
do/s'An/tsfa/ya "screech-owl" 
si lick lies "kingfisher" 
tddlkeldi "red-headed woodpecker" 

tea lwdcltc{!)e "sandhill crane" (cf. Chasta Costa sd'lwAsl- 

Editorial Note 

Originally published in University of Pennsylvania, Anthropological Pub- 
lications 2(2), 271-340 (1914). Reprinted by permission of the University 
Museum, University of Pennsylvania. 

The Athabaskan languages of southwestern Oregon — which are now (1989) 
virtually extinct — constitute a distinctive subgroup whose relationship to the 
northwestern California subgroup is surprisingly distant, although the two are 
usually classified together as "Pacific Coast Athabaskan" (Hoijer 1960). The 
Oregon group seems to have comprised four distinct, though closely related, 
languages: Upper Umpqua, Galice-Applegate, Chetco-Tolowa, and Lower 
Rogue River. The last was spoken in a variety of local dialects, the best known 
of which are Chasta Costa (on Rogue River about 30 miles upstream from the 
mouth), Tututni (near the mouth of the river). Euchre Creek (along the coast to 
the north of the river mouth), and Upper Coquille (still farther north, in the 
inland area behind Coquille Bay). No full grammatical study exists of these or 
of any Oregon Athabaskan language. Besides Sapir's description of Chasta 
Costa, the fullest published documentations of Oregon Athabaskan are of 
Galice (Hoijer 1966), Tututni (Golla 1976), and Tolowa (Bright 1964; Bom- 
melyn et al. 1989). Considerable manuscript documentation also exists, partic- 
ularly of Galice (by Melville Jacobs; see Seaburg 1982) and of Lower Rogue 
River (by J. P. Harrington; see Mills 1981: 69-76). Some of Harrington s Lower 
Rogue River material (including a number of aluminum disk recordings) was 
obtained from Wolverton Orton, Sapir's Chasta Costa consultant. 

Corrigenda to Father Morice's "Chasta Costa and the 
Dene Languages of the North."' 

In view of the fact that Father Morice has reviewed my Notes on 
Chasta Costa Phonology and Morphology in so evidently a friendly spirit 
it may seem a bit churlish to point out what seem to me to be either 
slips or misunderstandings in his recently published paper on Chasta 
Costa and more northern Athabaskan dialects. If, nevertheless, I 
venture to do so, it is not because of any desire to minimize the value of 
Father Morice's paper or to attach an overweening importance to my 
own very scanty contribution to Athabaskan linguistics, but to help 
advance our understanding of the problems of Athabaskan phonology 
and morphology. The chief value of Father Morice's paper seems to 
me to lie in the further light it throws on the Carrier language, of which 
previous papers have already shown he has an admirable mastery. I 
earnestly hope that Father Morice will not be content with the rather 
sketchy papers he has hitherto given us on the Carrier language, but will 
eventually publish a complete presentation of the intricacies of its 
phonetics and grammatical structure. 

I. " Dr. Sapir's a is my as, almost the sound of u in ' but,' more 
exactly that of e in the French ^e, te, le" (p. 560, footnote). My a of 
Chasta Costa words is to be pronounced like a of- German Mann and thus 
in sound corresponds to Carrier a, not (b. My A is another vowel alto- 
gether, though often phonetically reduced from original a. It is prac- 

> See American Anthropologist, N. S., 17, 1915, pp. 559-572. 


VI American Indian Languages 2 


tically identical in sound with u of but and doubtless corresponds phone- 
tically, largely also genetically, to Carrier o^. 

2. "This [C. C. tclac 'bird'] is evidently none other than the 
Carrier tscrz [Morice's s and 2 are described by him in a letter as sibilants 
midway between 5 and c, and z and j, respectively; they correspond histori- 
cally to Chipewyan^and t? < Ath, s and z], which in that language means 
not ' bird ' but 'feather-down' " (p. 560, 1. 15). This is not so evident. In 
fact it is phonologically impossible. Father Morice's Carrier form is 
clearly cognate with Chipewyan -d'Hd (Goddard; my -dUd) "soft feathers," 
Montagnais, tthoepdh (Petitot; tdlA-yd^ in my orthography) "plume 
servant k broder, cote de plume " (possibly mistranslated for " duvet "), 
Hare kkwew (my k^\ew), Loucheux tthaw (my td\aw), Navaho ts'os 
(Franciscan Fathers; my /5!o5) "down feather." These forms imply 
original Athabaskan *ts\ez (reduced *ts\ez, *ts\es) and *ts\os "down." 
This would be expected to appear in Chasta Costa as td\Ad (very likely 
actually found in td\Ad-nd-yal-td\od " humming-bird," perhaps literally 
"soft-feathers fiy-about-making-a-humming-sound"; for verb stem 
•tdlod cf. Hupa -tsots, -tsos " to make a kissing-Iike noise, to smack one's 
lips," Nav. beets' OS " whistle " as noun). C. C. tc\ac can have nothing 
to do with this tdlAd, but must go back to Ath. *k^\ac or *tc\ac} 

3. " This possessive [C. C. -lUc\e '(his) dog '] entails in Chasta Costa 
the accretion of a sort of suffix which he writes tele, the equivalent of 
my tse. Now litse means in Carrier, not somebody's dog, but she-dog! " 
(p. 561, 1. 3). Father Morice's quandary is of his own seeking. -ts\e 
of his Carrier form is not at all connected with -tele of my Chasta Costa 
one. Carrier li-ts\e is simply compounded of li "dog" and ts\e 
" woman." This ts\e (Ath. *tc\e) is found in Carrier tsekhk, Montagnais 
ttsh-kwi, ttse-k'^ (i. e., tsle-k'e), Hare ttse-Une, ttse-k'u, Loucheux ttse-ndjo, 
Kato tc'ek, C. C. ts\dxe. Carrier, Montagnais, Kato, and Chasta Costa 
point to Ath. *tc\e-k'e " woman." C. C. -tele of -li-tcle, however, goes 
back, not to Ath. *-tcle, but to Ath. *-k^le with glottalized anterior palatal 
k^l; cf. Hupa -liii-ke " (his) pet " (read- k^le). How this peculiar possess- 
ive *-k''le (cf. normal -e in, e. g.. Hare se-llin-e " my dog ") arose I can 
not say. Its isolated character stamps it as probably archaic. Perhaps 
-jfe"!- of Ath. *-liiiki'\e is connected with -g-, -k of Ath. *lik, *hg-, which is 
parallel to *ltrj; cf. Anvik hlik (Chapman; my lik), Loucheux V egce-tsellce 

> Since this was written Dr. Goddard has kindly called my attention to Hupa 
klyawM (read k^lauW) " birds " which corresponds exactly to Ath. *k''!ac. This elimi- 
nates Ath. *lc!ac as possible prototype and still more effectively disposes of Father 
Morice's analysis. 

Six: Athabaskan and Na-Dene Languages 97 


" petit chien " (i. e., legA-). (Incidentally, Petitot gives for " chienne " 
Montagnais I'intse, Loucheux Ventsi. Are these forms errors for -ttsi, 

4. " The- . . . , at least in the north, does not merely mean ' in the 
water,' as Sapir would have it . . . but it hints at the ' bottom of the 
water ' " (p. 561, 1. 25). C. C. t^e- was defined as " in the water." 
That, in Pacific Athabaskan, it has no necessary reference to " the 
bottom of the water " is showji by several of Goddard's Hupa and Kato 
examples. Thus, some Hupa verbs with t^e- are translated " in the 
water it seemed about to tumble," " in the water she floated back," 
" when he put his hand in it (i. e., the water)," cf. Kato <'e'- verb " she 
washed them." Evidently Goddard's definition of Hupa /'e- as re- 
ferring " only to motion into water and under its surface " seems quite 
adequate. In Carrier thentltlat " thou art precipitately brought to the 
bottom, thou sinkest," second modal -n- (to use Goddard's terminol- 
ogy) is terminative in function, so that t^e-n- . . . -tlat necessarily 
denotes " to sink to the bottom of the water," even if Ce- by itself be 
assumed to mean " in the water." A wider range of significance for 
Ath. *t^e- than Father Morice would assign it is implied also by such 
Navaho forms as tqe-li "water horse"; Navaho tqe- (i. e., my txe-) 
regularly corresponds to Ath. *Ce-. Navaho tqe- forms are likely to be 
archaic, as Ath. *Ce- has been almost entirely superseded in Navaho by 
Ath. */'a- (Navaho tqa-); see below. Even in northern Athabaskan I 
do not find Father Morice's remark to apply without qualification, even 
if correct for Carrier. In Anvik Ath. *t^e- appears as te-, tH- (Chapman's 
orthography); note teyldoutel "floating in the water." In Loucheux 
Ath. *Ve- appears as tchi- (Petitot's orthography), i. e., tcH- (Loucheux 
Ve-, i. e., /'e-, does not correspond, according to Loucheux phonetic laws 
that I have worked out, to Ath. */'e-, but to Ath. *t'a-)\ note tchi-dhitlle 
"gtre k flot." 

5. " Tha- . . . , which he gives as ' referring to the water ' has 
really that signification [i. e., ' in the water '] " (p. 561, 1. 27). Ath. 
*Ca- is undoubtedly found employed locally (" in the water ") in prac- 
tically all Athabaskan dialects; indeed it seems in some dialects (e. g., 
Navaho, to some extent apparently also Chipewyan) to have extended 
its sway at the expense of parallel Ath. *t^e-. There is, however, plenty 
of evidence to show that Ath. *Ca- is frequently substantival, not local, 
in force, that it is, in fact, an old noun stem (" water, wave, sea," parallel 
to more wide-spread Ath. */'o " water ") that has become a verb prefix. 
A local meaning is impossible, for instance, in the numerous verbs of 

98 VI American Indian Languages 2 


drinking found in Pacific Athabaskan (e. g., Hupa ta-win-nan " he 
drank it," Kato ta-ya'-o-nqn " let them drink," C. C. t'd-yAct-nd " I 
drink "). A direct substantival meaning, moreover, is obvious in such 
compound nouns as Navaho tqd-bd " shore " (lit. " water-edge "), 
while in several Mackenzie Valley dialects Ath. */'a even occurs as un- 
compounded noun stem (Montagnais t'a, i. e., my t^a, " flot, onde "; 
Hare t'a; Loucheux t'i:). 

6. " As to the verbal stem -al, which he believes to mean ' to come,' 
I more than suspect that it is but a corresponding form of -ya, which he 
represents as expressing the idea of ' going, coming,' and should be 
-yal " (p. 562, 1. 11). That C. C. an-yi-al really means "come on!" 
and not, like Carrier 'oen inyal, "go on!" (as suggested further on by 
Father Morice, 1. 26) is conclusively proved by two facts. In my brief 
Chasta Costa text (p. 337, 1. 3) an-yi-aL occurs as translation of English 
" come here! " (this text, it should be remembered, was translated from 
English) under circumstances in which " go on!" would be quite out 
of place. Further, during my residence among Chasta Costa Indians 
in Siletz Reserve, Oregon, in 1906, I distinctly remember that an-yi-al 
was often used by elders in calling children to them. As to whether 
recorded C. C. an-yi-al, an-yi-aL is to be understood as an-yi-yal, as 
claimed by Father Morice, or as an-yi-'al, as I had assumed, I now incline 
to think that -yal is correct. .This is because of such forms as Hupa 
ivin-yaL " come on " and Kato gUn-yaL " walk," which seem to cor- 
respond exactly to C. C. -yi-{y)al. Should -'al prove to be correct, it 
would probably have to be compared with Loucheux -'a " to go," Ath. 
*-ac " two go " (with dualic -c). 

7. " Dr. Sapir furthermore quotes the verb stem -t\o as denotive of 
the act of swimming, while, according to him, that of paddling is rendered 
by the radical -xe. Now, in most northern Dene dialects, the former 
refers to paddling, while the latter indicates the act of navigating, or 
moving about in a canoe" (p. 562, 1. 28). True, but there is plenty of 
evidence to show that Ath. *-k^e, *-Ven, *-k'^er}, *-kH frequently refers 
to or implies paddling. Thus, Goddard translates Chipewyan -ki by 
" to paddle a canoe, to travel by canoe " (e. g., ta-kt-hwu " when he 
paddled "). This is confirmed by Father LegofT, who defines 'ke'l 
(i. e., -k'el): " est le progressif de 'ki, et signifie proprement naviguer, en 
ramant " (e. g., pes-'ke'l " je rame, j'avance en ramant "). Similarly, 
for " ramer " in Hare Petitot gives not only e-ttoh, but also e-k'e. And 
in Anvik I find -kan, -kahl often translated as " to paddle " (e. g., xH-kahl 
'" he is paddling," ti-qH-kan " he paddled on "). For C. C. -t\o " to swim, 

Six: Athabaskan and Na-Dene Languages 99 


bathe " I find no exact parallels. It may, as Father Morice suggests, 
have primarily meant " to paddle " (denominative verb from Ath. 
*t\os " paddle ") and changed its force dialectically. 

8. " The verbal stem -lal . . . , to which our author attributes the 
sense of ' to sleep,' has in Carrier the value of ' to dream of ' (with a 
complement). Might not Dr. Sapir's informant have misunderstood 
his questioner and thus unwittingly misled him? " (p. 563, 1. i). I 
think we can manage without this hypothesis of misunderstanding. In 
Hupa -lal, -laL means not only " to dream," but also " to sleep " (e. g, 
nit-te-sil-lal-le " you would go to sleep "). Both meanings are given also 
for Kato -lal, -IqL (e. g., n-to-lqL " let him sleep "). Turning to northern 
Athabaskan, we find that in Chipewyan (Montagnais) -lal, -lal, -lal, 
-lal regularly means " to sleep, to fall asleep, to put to sleep " (e. g., 
Goddard's hi-teL-lal " he is asleep"; Petitot's in-t'es-l'al s. v. " dormir 
debout," es-l'al s.v. "endormir"; Legoff's in-'tes-la'l " je m'endors "), 
while for " rever " Petitot gives quite another stem in Montagnais 
and Hare. C. C. t'e-d-lal-{la) " he's been sleeping," with its prefixed 
elements t'e- and (in definite forms) -6-, corresponds remarkably to 
Hupa -te-sil-lal- quoted above, Kato {n-)te-s-laL " he went to sleep," 
and Chipewyan {ln-)te-d-Lal " he is asleep." In Hupa and Kato -lal, 
-lal, when meaning " to dream " seems regularly preceded by 7ia- (e. g. 
Hupa kin-na-is-lal " he dreamed," Kato u-na-s-laL " he dreamed 
about "). 

9. . . . '^ -lal, or rather -tlal. ... As may be seen by Sapir's 
rendering: t'Ulal. The double consonants tl and ts are of frequent oc- 
currence in Dene and form as many indivisible groups. The syllables 
of all Sapir's verbs in the first person plural are wrongly cut up: the t 
which he attributes to the penultimate syllable should commence the 
last one: -tlal, -tsoel, -tlat, -thcec, etc. Hence several of his verb stems are 
incomplete. For instance, -se, 'to cry,' should be -tse (Carrier -tso)\ 
-si, ' to cause,' should be -tsi (Carrier -tsi); -lo, ' to laugh,' cannot be 
understood without its /. . . . Were he familiar with the Dengs' syl- 
labic way of writing their own language, he would have been spared this 
little inaccuracy " (p. 563, 1. i and footnote i). There are several com- 
ments I wish to make on this passage. 

(a) That, in Chasta Costa, not -tlal but -lal is to be considered as 
true verb stem is quite unmistakably evidenced by such forms as t'l-lal 
" thou art sleeping," I'o-lal " ye sleep." Were -/- part of the stem, 
there would be absolutely no reason for its disappearance in these forms 
(cf., for phonetic comparison, t'd-yit-nd " thou drinkest," I'd-yd-dot-nd 

100 VI American Indian Languages 2 


" ye drink," in which third modal prefix -t- is preserved between personal 
element -i-, -6-, and stem consonant n). These remarks apply also to 
C. C. -lat, not -tlat, as Father Morice would have it. Comparison with 
Hupa and Kato, also Chipewyan, abundantly confirms my own analysis. 
(6) There is no point whatever in quoting C. C. t'it-lal (not t'ltlal, 
as Father Morice has it) " we are sleeping " as evidence of a stem -tlal. 
In Chasta Costa -{i)t- regularly appears as first person plural subjective 
prefix, except, as in Chipewyan, before third modal -/- and -/- (e. g., 
ye-dit-'i " we saw him," where Father Morice would hardly claim that 
-t'i, not -'i, is the true stem). That this -t- is indeed an organic element 
in the first person plural prefix is shown by comparison with Hupa -d- 
(e. g., na-dtL-Le " we are painting "), Kato -d- (e. g., na-ddl-yic "let 
us rest "), Chipewyan -it-, -d- (e. g., e-gU-'t " we saw it "), Loucheux -di- 
(e. g., i-di-kwoll " nous amenons "). What has apparently misled Father 
Morice in his analysis of Chasta Costa is that Ath. first person plural 
(and dual) *-{i)d{e-)- has in several dialects, perhaps by analogy of third 
modal -I- and -/- verbs where -d- regularly disappeared (I am inclined 
to think that in Hupa and Kato -d- has in these cases been restored by 
analogy), been replaced by -i-, -i-, e. g.. Carrier nd-i-ta " we are both 
sick " (Carrier first personal plural prefix tsoe-, i. e., ts\A-,'^ does not seem 
to be a widespread Athabaskan element, though Father Jette has 
recorded for Ten'a ts- forms of like meaning, perhaps to be understood 
as ts\-); Hare i-ssi "nous deux faisons " {ya-issi "nous faisons"); 
Navaho ch'i-ne-i-ka " we two carry milk out " {ch'cB-de-i-kd " we carry 
milk out "). Even in these dialects, however, survivals of the old -d- 
are found in such cases as allowed of its carrying over into the stem 
syllable, e. g.. Carrier nt-tas " we two walk," i. e., nl-tlas, morphologically 
equivalent to n-U-'as (cf. noe-hce-'as "they walk two together"); Hare 
witta " nous deux allons," i. e., un-tla (Ath. *zi-tlas), morphologically 
equivalent to w-it-'a (Ath. *z-id-'as); Navaho da^-dt-t'a^ "we put a 
round object on " (equivalent to da^-d-it-'a'), bd^-ni-'nil " we two put 
several things on " (equivalent to ba^-n-U-nil), l-gye " we two marry" 
(equivalent to id-ye, cf. i-ye " he marries "). There can be no talk in 
such cases of stems -t\as, -t\a\ -'nil, and -gye but only of phonetic resul- 

1 1 strongly suspect, as is indeed indicated by Father Morice himself (see "The 
Dene Languages," 1891, p. 193) that this /foj- is really impersonal in meaning, ta 
begin with; with Carrierna;-/5a5-/»7 "we walk"; cf. nas-tsi-ya "one walks." For reasons 
of modesty, perhaps, it may have become customary to say "people (in general) do 
so and so" instead of " .ve do so and so." This indefinite Carrier tsce- is quite possibly 
cognate with Hupa U-, i. e., tc!-, of third personal forms applying to adult Hupa (e. g., 
tce-xauW "he is catching," originally perhaps "one catches"). 

Six: Athabaskan and Na-Dene Languages 101 


tants of first person plural -d- plus stems -as, -'a\ -nil, and -ye (Ath. 

(c) As regards syllabic division, I must emphatically disagree with 
Father Morice. It is quite clear that in many Athabaskan dialects the 
-/- {-d-) of the first person plural belongs to the following syllable, as we 
have just seen. This proves nothing, however, for Chasta Costa, nor 
do I see how familiarity with the Carriers' syllabic way of writing their 
own language would have materially helped my ear in its perception 
of Chasta Costa sound combinations. As a matter of fact, my informant 
Wolverton Orton was particularly careful in syllabifying and I was prac- 
tically never in doubt, in recording his forms, as to where syllabic di- 
vision was to be placed. Hence t'ii-lal is not to be " corrected " to 
tH-tlal, t'e-nit-lat to t'e-ni-tlat, yit-lo to yi-tlo, tclAt-t'it-dAl to tc\At-Ci-tBAl 
(Father Morice's -tscel). Father Morice's -thcec is simply another ortho- 
graphy for my -t'^Ac. I may point out that were we dealing in these 
Chasta Costa forms with such consonantal groups (affricatives) as // 
and tS, preceding i would appear lengthened to I, because completing its 
own syllable. That I have consistently misheard both vocalic quantity 
and syllabic division in these words (e. g., -iHt-dAl for -Cl-tdAl) I cannot 
admit. Naturally I do not deny that td and td\ also occur in Chasta 
Costa as true affricatives, though Id seems to have regularly developed to 6. 

As to C. C.-t-lo (I have no example of C. C. affricative il, dl, nor 
does it seem to occur in Hupa or Kato), I am uncertain, as I pointed out 
in my " Notes," whether this is to be understood as directly representative 
of Ath. *-dlo (as preserved, e. g., in Carrier, Chipewyan, and Navaho) 
or as analyzable into third modal -t- and stem -Id. For Ath. dl: I, cf. 
Ath. *-dlo, *-dli " to be cold ": Montagnais e-llu " cold," Kato Id " frost." 
C. C. -Id " to laugh " might well correspond to Kato -Id " to deceive " 
(primarily ".to laugh at"?). 

(d) As regards C. C. -se " to cry " and -si " to cause," Father 
Morice is going altogether too far when he says these stems are incom- 
plete for -tse and -ist. As I have already stated in my " Notes " (p. 
287), Ath. tc\ of which Carrier ts is reflex (Ath. /c' is preserved as such 
in Navaho, Jicarilla Apache, Kato, and, as labialized tc'w, in Hupa), 
has always been simplified in Chasta Costa to 5 (Ath. tc^>ts>s; cf. 
Ath. ts>td>d). Hence to " correct " C. C. -se to -tse is as justifiable 
as it would be to " correct " French chef to *kep because, as no one 
denies, derived from Latin caput. 

ID. " If Dr. Sapir will allow me, I will also observe that the desinence 
-tc'ac [probably misprint for -t^Ac], which he gives as a distinct verbal 

102 VI American Indian Languages 2 


element, is nothing else than the plural stem of the same [Carrier] verb 
nanisthi, whose derivative nthcesthih effectively means ' to lie down, go 
to bed ' " (p. 563, 1. 7). Father Morice then goes on to compare C. C. 
-t^Ac directly with Carrier -this, -thez, -thcES (plural stem). These state- 
ments are misleading. It is evident from my Chasta Costa material 
(e. g. dd-nAc-t^Ac " I go to bed ") that -t^Ac applies to singular subjects. 
This is confirmed by comparison with Hupa -tiiW (which corresponds 
regularly to C. C. -t^Ac) and Kato -tUc (e. g., Hupa tcin-ne-tiiW " she 
goes to bed," Kato na-nHn-tilc " lie down again "). This Ath. *-t'ec is 
characteristic in Pacific Athabaskan of " indefinite " ^ forms as contrasted 
with " definite " Ath. *-t^in, *-t^er}, *-t'i (Hupa -ten, -tin; Kato -tin; 
Carrier -thi; Montagnais -Vi; Navaho -tqi); in most Athabaskan dialects 
*-i'e or *-t^el is used in " indefinite " forms (Hupa imperative -te; Mon- 
tagnais eventual -Ve; Navaho present -tqe, future -tqel). " Indefinite " 
-c forms are in general apparently characteristic of Pacific Athabaskan. 
C. C. -t^Ac is phonetically practically identical with Carrier -thces, but 
not morphologically. Carrier " proximate future " -thces is reduced 
from present -thes; in other words -5 (Ath. -c) is here found in all forms, 
" definite " and " indefinite." This plural (and dual) stem Ath. *-t^ec, 
*-t'ej, *-t^ec (cf. Hupa -tetc; Montagnais dual -/'ez; Navaho -iec; Jicarilla 
Apache -kec) contains dualic -c (cf. Ath. -'ac " two go "; Navaho -'esh, 
-'ezh, -tsh " to act upon two animals "). We now see clearly that Father 
Morice was misled by a phonetic convergence of morphologically distinct, 
though genetically related, forms. 

II. " Unless I am very much mistaken, what he adduces as the equi- 
valent of ' I am seen, you are seen, he is seen,' etc., really means simply: 
' people see me ' (French: on me volt), ' people see thee,' etc." (p. 563> 
1. 22). Father Morice is, in my opinion, quite right. In fact he merely 
repeats what I had already pointed out (p. 332): " tslAl-, which appears 
in these forms, probably contains third modal -I- preceded by deictic 
^5!- implying indefiniteness of logical subject: ' mann sieht mich.' 
Surely Father Morice knows that German man sieht mich is identical in 
force with French on me voit. 

I take this opportunity of modifying my analysis of C. C. ts\A-. I do 
not now think that it is comparable to C. C. deictic tc\-, which denotes 
lack or indefiniteness of object, not subject, of transitive verbs; this tc\- 
goes back to Ath. H!(e)-, which is preserved as such in Hupa. C. C. 
ts\- (Ath. *tc\-) of such forms as nes-ts\Al-'i " one sees me, I am seen," 

' I use "definite" and "indefinite" in Goddard's sense. "Definite" tenses are 
present definite and past definite, all others are "indefinite" (including present in- 
definite, imperative, eventual, and other forms). 

Six: Athabaskan and Na-Dene Languages 103 


however, is clearly subjective and impersonal in force and very probably 
corresponds to Carrier indefinite ts{oe-), see 9 b above. Goddard's 
Chipewyan ts'- is probably identical with this Chasta Costa and Carrier 
impersonal prefix (e. g., ts'e-Lu " he was caught," i. e., " one caught 
him "); this means that Goddard's comparison of Chipewyan ts'- with 
Hupa ^(^!)- is incorrect (see p. 133 of his " Analysis of Cold Lake Dialect, 
Chipewyan "), and indeed we should in that case, as in Kato and Chasta 
Costa, have expected tc\-. I do not know how to reconcile with these 
subjective impersonal forms (Ath. */c!-> Kato tc\-, ts\-, s'-; C. C. ts!-; 
Carrier ^5!-; Chipewyan ts\-) Petitot's Montagnais, Hare, and Loucheux 
indefinite tse- forms. Could he have throughout misinterpreted is\e- 
(in his orthography Use-) as tse-? E. Sapir 

Editorial Note 

Originally published in American Anthropologist 17, 765-773 (1915). 
Reprinted by permission of the American Anthropological Association. 

This reply to Morice's criticism of Sapir's Chasta Costa work (Morice 1915b) 
prompted a rejoinder (Morice 1917). Both of Morice's papers, together with his 
original review of Sapir's monograph, are reprinted in the appendix to this 


1"^HE problem attacked in this paper is that of the genetic 
relationship of Athabaskan, Haida, and Tlingit. Important 
morphological, to a less extent also lexical, resemblances 
between Haida and Tlingit have long been pointed out by Boas and 
Swanton, resemblances which have led them to assume, though 
rather hesitatingly, genetic relationship between these languages. 
Boas has also somewhat vaguely hinted at fundamental resem- 
blances in structure between Athabaskan and Haida-Tlingit, but 
no concrete evidence has been given on this point. A full presen- 
tation of the comparative lexical, phonological, and morphological 
evidence that serves to show, beyond all reasonable doubt, that 
Athabaskan, Haida, and Tlingit are indeed but divergent represent- 
atives of a common prototype is given in an extensive paper on 
"The Na-dene Languages" now in course of preparation as a 
memoir of the Anthropological Series of the Geological Survey of 
Canada. The present sketch, prepared at the request of Dr P. 
E. Goddard, is merely a rapid abstract of some of the leading 
points involved. I wish expressly to emphasize the fact that it 
does not present all the evidence at my disposal. While, however, 
it does not constitute the complete demonstration of my thesis, I 
believe that enough is here given to remove this thesis beyond the 
realm of the merely probable. The term "Na-dene," which has 
been chosen to designate the hypothetical prototype of Athabaskan, 
Haida, and Tlingit, will be explained in the latter part of the 

I. Morphological Features 

Stem and Word Form. — The most typical and doubtless his- 
torically primary type of stem form found in the Na-dene languages 
is the monosyllabic stem consisting of consonant plus vowel; in 

1 Read in substance before the Anthropological Association at Philadelphia. 


106 VI American Indian Languages 2 


Haida (H.) the consonant may be replaced by a cluster of two 
consonants which, in cognate words, appears contracted to a single 
consonant in Athabaskan (Ath.) and Tlingit (Tl.). Examples 
are: Ath. *-tsi^ "daughter," *t'o "water," *tl!o "grass," *-k'e 
"foot," *-ne "to speak," *-ya "to stand (plur. subj.)," *-'a "to 
find," *-ya "for," *-na "around," *na- "again;" H. tc/W "cedar," 
q/a "harpoon," st/a "foot," Iga "rock," t'a "to eat," q!a "to 
sleep," xa "to follow," sa "above," gu "at," q/o- "by means of 
the teeth;" Tl. t^a "stone," nu "fort," xa "enemy," ha "to dig," 
g'a "to say," ci "to hunt for," t!a "behind," k^a "on," dJL- "quick- 
ly." Many, perhaps all, elements consisting of a single con- 
sonant (or cluster of two consonants) are phonetically reduced 
owing to the loss of a vowel; e.g., Ath. *-n, *-77 "person" <.*-ne; 
H. stU- "with the fingers" <stl!a "hand;" Tl. / "to" <de. 

In all Na-dene languages, however, a large number of stems is 
found consisting of consonant plus vowel plus consonant; e.g., Ath. 
*-ts!en "bone," *-tIas "to cut;" H. k'un "point," sgol "to hide;" 
Tl. dis "moon," tsi'n "to be strong." In a very large number of 
cases there is clear internal evidence to show that the final consonant 
is an old suffixed element whose original meaning has doubtless 
generally been lost. Examples of such "petrified" suffixes are: 
Kato hts "clay," Navaho h'c "dirt, ground," cf. Navaho h'- 
in compounds; Anvik /'aZ "bed" (< Ath. */V/), Hupa-f'e/c "several 
lie," cf. Hupa -/'c "to lie (sing.)," past definite -t^en; H. (Masset) 
s'ail "to weep," cf. s'ai-ga; H. xal- "by means of fire acting from 
without," cf. xai "sunshine;" Tl. t'i'n "to see," t'i'sl "to look 
for." While a considerable number of such stem finals corres- 
pond in Athabaskan and Haida or Tlingit (e.g. Ath. *-del "several 
go," H. dal "many persons go by land;" Ath. *-k!an "to burn," 
Tl. qia'n "fire"), numerous cases are found of stems that cor- 

1 Forms given as Ath. are reconstructed on the basis of the actual forms found in 
various Athabaskan dialects. The general methodology of linguistic reconstruction 
and the sounds reconstructed for Athabaskan specifically are dealt with in the longer 
oaper above referred to. The phonetic system employed in this paper is the one 
worked out by the Phonetic Committee of the American Anthropological Association; 
this report will be published in the near future. 

* Indicates reconstructed forms. 


Six: Athabaskan and Na-Dene Languages 107 


respond according to regular phonetic law except for the final 
consonant; sometimes two of the three Na-dene languages agree 
as against the other; often the simple vocalic stem is found in 
one or two, but extended by a final consonant in the other. Exam- 
ples of these cases are : H./'a7 "year," Tl. Z'o"^ "year;" Chipewyan 
dial "moss," Louch. tdek, Tl. slate (Chip. d!a- and Louch. tde- 
point to Ath. *ts!a-, which points, with Tl. sla-, to Na-dene *ts!a-) ; 
Ath. *-lad "end," H. tlan "end;" Ath. *-yel "night passes," H. 
ga-l "night," Tl. get "to get dark;" Ath. *-ca, *-cal "to catch with 
a hook," H. djd "bait," Tl. cat "to seize;" Ath. *xes "mountain," 
Tl. ca "mountain;" Ath. *-t'an "to eat," H. t'a "to eat;" Ath. 
*-t!o "to shoot," Tl. t!uk "to shoot." Examples of this sort 
make it fairly obvious that many of the stems with final consonants 
that are yielded by a purely descriptive analysis are ultimately 
reducible to vocalic stems followed by what was originally a suffixed 
element. That all Na-dene stems with final consonants are of 
such origin cannot be demonstrated, but it does not seem at all 
improbable. The characteristic Na-dene stem may thus be 
symbolized by cv, of which c and cv-c are further developments. 

Reduplication is a grammatical process that is conspicuous 
in Na-dene by its absence. It is found neither as a word-forming 
nor purely grammatical device. The only possible widespread 
Athabaskan exception that I have been able to find is the demon- 
strative stem didi "this," alongside of unreduplicated di. With 
this it is interesting to compare the probably reduplicated Haida 
interrogative stem gu'gu-s "what?" A negative feature of this 
sort is not in itself very indicative, but gains in weight when the 
Na-dene languages are contrasted with the Tsimshian, Kwakiutl- 
Nootka, and Salish languages to the south, in all of which redupli- 
cation plays an extremely important part. 

The typical Na-dene word is built up of a number of monosylla- 
bic elements (in most cases of form cv), one of which is the main 
stem, about which cluster a number of subsidiary etymological and 
grammatical elements that may be termed prefixes and suffixes. 
The various elements of a word, aside from certain ones that are 
perhaps best considered as proclitic and enclitic particles, make up a 

108 yi American Indian Languages 2 


coherent enough morphological unit, but are far from welding to- 
gether in a manner suggestive of such form units as we are accus- 
tomed to in Indo-germanic or are found also in manyAmerican Indian 
languages (e. g., Kwakiutl, Eskimo, Yana, Southern Paiute). Most 
of the elements preserve a considerable share of individuality, while 
many can, indeed, be shown to be identical in origin with or special- 
ized forms of independent stems. Thus, an Athabaskan word like 
Kato t^aya'o'nat] "let them drink" readily falls apart into four 
perfectly distinct elements: the main stem -narj "to drink" 
and three subsidiary elements that may be described as prefixes, 
but which are far from fusing either among themselves or with this 
stem into a close morphological unit; the prefixes are /'a-, an 
element having reference to water (cf. independent forms Kato 
t^o' "water," Montagnais ^ t^a "billow"), a demonstrative element 
ya'- indicating the plurality of the implied (but not definitely ex- 
pressed) third personal subject of the verb, and o'-, a hortatory or 
"future imperative" modal element: " water- they-shall-drink." 
The "word" Caya^o'tia-q feels decidedly like an old "sentence" of 
monosyallabic constituents, t^a ya' o' na-n, the first three elements 
of which have lost their absolute independence and all four of 
which have settled down to a rigidly prescribed order relatively to 
each other. This same type of sentence-word (we may either 
think of it as very closely knit sentence or, as we are more accustom- 
ed to do, as relatively loosely knit word composed of easily analyzed 
elements) can be abundantly illustrated also in Haida and Tlingit. 
A Haida example is tWalgu'ldayarj "(she) had put (it) (on her son) 
as a blanket," consisting of a primary stem -Igii'l- "to go around, 
to be wrapped about," an instrumental prefix tl- denoting activity 
with the hands (identical with the verb stem tl "to touch"), a 
classifying nominal prefix t!a- which defines the object of the verb 
as belonging to the class of flexible things thought of as crossing or 
coiled (cf. //a-da " to wear ; " an original Haida stem *^^a "blanket" 

'Another name for Chipewyan. I use "Montagnais" to indicate that I am 
quoting from Petitot, "Chipewyan" from Goddard. All other Athabaskan forms 
are quoted from Goddard and The Franciscan Fathers; all Haida and Tlingit forms, 
from Swanton. 

Six: Athabaskan and Na-Dene Languages 109 


becomes very probable on reference to Ath. *tfe "blanket"), 
a causative suffix or auxiliary verb stem -da " to cause, to have as, " 
and two temporal-modal elements, a perfective -y- (or -i-) and a 
suffix -aj] indicating that the statement is not made on one's own 
authority: hand-blanket-be wrapped about-cause-d-as experi- 
enced by others. Finally, a typical Tlingit example is afforded by 
q!a-o-dt-sa "(he) blew (upon the raft)." As in the Athabaskan 
example given above, the main stem, sa "to blow," comes at the 
end of the word ; it is preceded by three elements, an instrumental 
prefix qia- "with the mouth" (identical with the noun stem q!a 
"mouth"), a modal element 0- whose exact significance is unknown 
(it seems to be found only in active past temporal forms with third 
personal subject), and another modal or "aspect" suffix rfi- of 
apparently inceptive or momentaneous force: mouth — in past time 
(?) — momentaneously — blow. 

One of the incidental consequences of this type of structure is 
that, while the analysis of the word into its parts is in most cases 
easily undertaken, a just idea of the actual value or content of the 
word as a whole cannot be obtained by merely summing the values 
of the analyzed elements. There is, equally in Athabaskan, 
Haida, and Tlingit, a great deal of idiomatic usage involved; in 
many cases all we can say is that it is customary for a certain 
perfectly definite idea to be expressed by a stem of fairly wide 
range of significance preceded by such and such not always evidently 
applicable prefixed elements. Thus, the purely etymological 
analysis of the Hupa do' ma{k)-k!ai na-si-ri-a "not-it-after over 
surface of ground-continuously-thou-have in possession" conveys 
as good as no notion of the actually well determined idea conveyed : 
"thou didst not want to (go home)." Similarly, the Haida verb 
k^wa-lgi'-sta-sga- "in a stream-large cylindrical objects-remove 
from (a place)-toward an open place" really means " (olachens) 
run in a stream toward the sea." 

Noun and Verb. — The relation betwen noun and verb is quite 
parallel in all three languages. While verbal and substantival 
forms are throughout clearly kept apart (verb forms may be substan- 
tivized in various ways), the radical element of a word may often 

110 VI American Indian Languages 2 


be indifferently used as predicating or denominating stem. Thus, 
the Haida stem na indicates both "house" and "to dwell," go't 
is used either as a noun meaning "buttocks" or an adjectival verb 
"to be last." In Tlingit this elasticity of usage is apparently less 
marked, though examples occur (e.g., sa "voice, name; to name, 
call;" CL "song; to sing"). Denominative verbs of this sort are 
particularly common in Athabaskan, e.g., Kato k!ari' "withes," 
-k/ar]' "to twist;" Chipewyan xal "club," -xal "to use a club;" 
Hupa tlfo "grass," -tUo "to make baskets;" Chipewyan t'an 
"ice," Kato -t^arf "to be cold;" Navaho si'l "steam," Kato -si'l 
"to steam." Under these circumstances it is perfectly natural 
that stems which are found used only as nouns in one of the Na-dene 
languages have become specialized as verbs in another. Examples 
are: H. xao "liquid," Ath. *-k^a "liquid has position;" Tl. qia'n 
"fire," Ath. *-k!an "to build a fire, to burn;" Ath. *tc'el "steam," 
H. sgal "to steam;" H. das "live coals," Ath. *-das "to burn, to 
singe;" Ath. *ts!ai, *ts!a- "dish," Tl. sliql "dish," H. sq.'ao "to 
put in a dish." 

A peculiarity of many Na-dene verb stems is that they are 
limited in their range to a particular class or number of objects. 
The simplest type of these is formed by verbs applying specifically 
to a singular, dual, or plural subject or object; e.g., Tl. gu "to go 
(one person)," at "to go (plur.);" H. q'a "to go (one person)," 
dal "to go (plur.);" tHa "to kill one person," tUda "to kill several;" 
Ath. *-ya "one person goes," *-del "several go;" Hupa -yen "to 
stand (sing.)," -ya (plur.); -'a "one object is in position," -'e/^' 
"several objects are in position." Still more characteristic are 
distinctions based on the shape of the object affected; e.g., Tl. fan 
"to carry a long thing," t\ "to carry a round thing;" Hupa -t'an 
"to handle or move a long object," -k^os "to handle or move any- 
thing that is flat and flexible;" Navaho -\c "to lead (by a rope) 
a pair of animals," -los "to lead (by a rope) a single animal;" 
-tW "to act upon an animate object," -djol "to act upon such 
objects as hay, wool, or hair. " In Haida such verbs do not seem to 
be found, but it is interesting psychologically to observe that 
corresponding classifications are here expressed by another means, 

Six: Athabaskan and Na-Dene Languages 111 


namely by the use of a long series of classifying nominal prefixes; 
e.g., /c/1,5- "cubic objects, such as boxes, " sq!a- "long objects, like 
sticks and paddles," ga- "flat objects." 

Verb Structure. — In all Na-dene languages the verb consists of 
a series of elements, which may be grouped into certain classes that 
have fixed position in the complex relatively to each other; the 
verb stem gravitates towards the end of the complex, particularly in 
Athabaskan and Tlingit. The typical Athabaskan verb may be 
analyzed as consisting of: adverbial prefix (including original noun 
stems, ^ local postpositions,^ petrified demonstrative stems of 
chiefly objective reference,^ and certain other elements which do 
not occur in other connections'*) + objective pronominal prefix 
+ demonstrative element referring to subject of verb + "first 
modal" element + "second modal" element^ + pronominal 
subject + "third modal" element^ + verb stem + temporal- 
modal sufifix + syntactic suffix (these are best considered as enclitic 
particles). Any of these elements but the stem may, in a particu- 
lar form, be missing; two or more of the same general type may be 
exemplified in a single form. The order of elements as given above 
varies slightly for different dialects. 

Quite similar in its general features to the structure of Atha- 
baskan verb forms is that of corresponding forms in Tlingit. The 
analysis may be given as: pronominal object (best considered as 
proclitic to verb form) + nominal prefix of instrumental signifi- 

1 E. g., Hupa t'a- "water," sa'- "mouth." 

* E. g., Hupa ye- "into," wa- "to," xa- "after." 

' E. g., Hupa a- used as indefinite object with verbs of saying and doing, xa- 
"same as before." 

* E. g.. Hupa da- "resting on," no'- "coming to rest," na'- denoting indefinite 
movement over surface, na- "again." At least some of these may be independent 
verb or other stems in origin. With da- of. Ath. verb *-da "to sit (sing, subject)"; 
na- may be identical with *na- "two" (found, alongside of absolute *nak'e, in, e.g., 
Hupa na-diT) "twice," na-nin "two men"). 

'These two sets of "modal" elements are not easy to define. They are best 
considered as indicating certain "aspects," i. e., as defining range of activity with 
reference to such notions as inception, continuation, distribution, cessation, and 
indefiniteness of object. 

« These "modal" elements also are difficult to define and, like "first" and "second 
modal" elements, are largely bound up in usage with idiomatic factors. Their primary 
significance is to define voice, i. e., such notions as transitive, intransitive, and passive. 

112 VI American Indian Languages 2 


cance* + "first modal" prefix ^ + pronominal subject + "second 
modal" prefix + "third modal" prefix' + verb stem + quasi- 
temporal suffix + syntactic suffix. 

Differing more widely from the Athabaskan pattern of verb 
structure is that of Haida. In Haida the pronominal subject and 
object are not as closely welded into the verbal framework as in 
Athabaskan and Tlingit and are best considered as independent 
elements of speech. . However, as they occupy definitely deter- 
mined positions immediately before the verb form proper, 
their structural difference from the corresponding elements of 
Athabaskan and Tlingit is more apparent than real. There is 
involved here merely a difference of degree of coalescence of originally 
distinct elements. The Haida verbal scheme may be represented 
as follows: pronominal object -+- pronominal subject + instru- 
mental prefix (most of which are in origin noun and verb stems 
capable of being used independently) -j- classifying nominal prefix 
(several of which, perhaps all, are old noun stems) + prefixed 
adverbial element '' -f- main verb stem -f- auxiliary verb stem (doubt- 
less independent verb stems in origin which have become specialized 
as quasi-suffixes) + adverbial element (in origin independent 
noun, verb, adjective, or adverb stems)* + locative suffix + tem- 
poral-modal suffix. 

This analysis of the Haida verb is not complete. It should 

• E. g., qfa- "mouth," lu- "nose, point." 

* These elements do not form a well-defined class. They embrace such notions 
as causation, aspect, voice, tense, and indefiniteness of subject. They correspond, in 
Athabaskan, partly to certain adverbial prefixes, partly perhaps to "first modal" 

'These two sets of "modal" prefixes seem primarily to define various aspects 
(perfective, progressive, completive, inceptive, repetitive, momentaneous, transi- 
tional). They correspond to Athabaskan "first modal" and more particularly to 
"second modjil" elements. As far as known, Athabaskan "third modal" elements 
find no counterpart in Tlingit. 

♦ These elements, of which Swan ton lists four, are termed by him "stems in initial 
position." It does not seem to me that there is any real necessity for the setting up 
of this class. Two of the elements are best regarded as nominal classifiers, one as an 
instrumental prefix, the other as a verb stem regularly compounded with other stems 
(see below). 

' Swanton classifies these into three groups of "stems in terminal position," but 
this sub-classification, even if justified, is of no particular consequence here. 

Six: Athabaskan and Na-Dene Languages 113 


be observed, first of all, that the order of elements fluctuates in 
accordance with their logical relation to each other; thus, locative 
suffixes directly follow the main stem and are followed by auxiliary 
verb stems, if the locative element does not logically apply to the 
latter (e. g., q'a-dl-da " to take aboard," literally " to go-into 
canoe-cause"). Secondly, two or more independent verb stems 
may combine into a compound verb which is held together by the 
preposed pronoun (or pronouns) and the suffixed temporal-modal 
suffix (or suffixes) at the end of the complex. Each member of the 
compound may be itself attended by derivative prefixes or suffixes 
(including even certain temporal-modal elements, like continuative 
-gar]-). If we assume, as internal Haida evidence makes more than 
probable, that all auxiliary verb stems and suffixed adverbial 
elements are nothing but compounded originally independent 
stems, we may reduce the above analysis of verb forms to: pronom- 
inal object + pronominal subject + I. -f- II. + • • • -f- temporal- 
modal element, in which I., II., • • • stand for complexes of type: 
instrumental prefix -j- classifying nominal prefix + verb stem + 
locative sufifix + continuative suffix). 

Naturally, in any given verb form only a comparatively small 
number of theoretically possible positions are filled. I. generally 
contains the predominant stem of the whole verb form. Haida 
verb composition in its present form is doubtless largely a special- 
ized development, though probably based on Na-dene processes. 
For this reason the typical Haida verb form in its older form must 
be defined, eliminating II., . . . , as: pronominal object -\- pronominal 
subject + instrumental prefix + classifying nominal prefix + 
verb stem -+- locative suffix -f temporal-modal element. This 
scheme, despite its peculiar features, more nearly resembles the 
Athabaskan and Tlingit schemes than the one first given. ^ 

Comparing the three verbal analyses given, we find that the 
Na-dene languages have several important traits of verb morpho- 
logy in common. These are: 

• I do not, of course, mean to imply that all instrumental and classifying nominal 
prefixes are older, as verbal elements, than all "stems of terminal position." Analogy 
always operates to feed a type already in existence. 

114 VI American Indian Languages 2 


1. Noun stems are included as prefixes in the verb complex. 
They are partly of instrumental (or local) significance, partly, 
more particularly in Haida, general classifiers of subject or object. 

2. Both pronominal subject and object elements regularly 
precede the verb stem. Of these, the object comes first in the comp- 
plex. In Haida the degree of coalescence of pronominal elements 
with the verb complex is much less than in Athabaskan and Tlin- 
git. In these languages the subjective pronominal element is an 
integral part of the verb-form, being often separated from the 
objective element by an adverbial prefix. 

3. Local affixes are found in both Haida and Athabaskan, 
though they are suffixed in the former, prefixed in the latter. At 
least some of the Athabaskan local prefixes are postpositions in 
origin; these, as regards their position after pronominal objects, 
offer striking analogies with corresponding elements in Haida and 
Tlingit, as we shall see later. 

4. Athabaskan and Tlingit possess a large number of prefixed 
"modal" elements, which define adverbial notions, to a less extent 
temporal ideas, but primarily aspects. They are divisible into 
several position-classes, according to whether they precede or 
follow pronominal object and subject. These elements are in some 
respects the most characteristic of Na-dene morphology, though 
their presence is hardly traceable in Haida. 

5. The verb stem is a generally monosyllabic element clearly 
marked off from the rest of the verb complex. It is nearly always 
preceded by a number of originally independent modifying elements. 
In Athabaskan it undergoes internal phonetic and morphologic 
changes as it passes from one tense (present and past, definite and 
indefinite) to another. Such changes have not been indicated by 
Swanton for either Haida or Tlingit. Dr Boas, however, on the 
basis of material recently secured from a Chilcat Indian, informs me 
that internal stem changes for tense, analogous to those found in 
Athabaskan, are characteristic also of Tlingit. 

6. A series of temporal-modal elements is found suffixed to the 
verb stem. Some of these are firmly united with the verb stem 
(e. g., continuative -/ in Hupa tc!u-wi-l-Ce-l "he was bringing;" 

AM. ANTH., N. S., I7 — 35 

Six: Athabaskan and Na-Dene Languages 115 


usitative -tc in Tlingit u-q^ox-tc "he kept coming in;" perfective 
-y- in Waida. su' -da-y-agam "had said"), others are ni.ore in the 
nature of enclitic particles (e. g., emphatic -ht in Hupa do' a-du- 
win-ne-he "don't say that!" imperative -de in Tlingit na-at-de 
"(for firewood) go!" Haida hortatory -dja-q in I gLti-gat-dja-q "let 
me adorn [you]!"). 

7. Still more loosely suffixed, in most cases, to the verb form is 
a series of syntactic particles, largely used to subordinate it in 
various ways. In part these elements, as we shall see, are post- 
positions in origin. 

8. While compounding of verb stems is most luxuriantly de- 
veloped in Haida, indications are not lacking of the presence of 
the process also in Tlingit and Athabaskan. Thus, the Tlingit verb 
stem CL "to desire" may be prefixed to another verb stem to form 
its desiderative (e. g., CL-t^an "to desire to pick berries"); cf. such 
Haida compounds as gi'da-yu'an-SLtj-ga " to wish to give much food," 
in which the auxiliary verb stem -sir} "to wish," however, is suffixed 
to the main stem. Nearer the Haida type is the class of Athabaskan 
verbs in *-tsfe (cf. independent verb stem *-ts!e "to hear"), indicat- 
ing that the action of the main verb is heard or otherwise perceived 
(e. g., Hupa -Cuw-tsle "to hear one splitting logs," Chipewyan 
-ni-d'e "to hear one say"). Differing morphologically from, yet 
psychologically similar to, the Haida type of verb composition is 
the use in Tlingit and Athabaskan of two independent verb forms 
to form a logical unit; e. g., Tl. gax gax-yi'-sa-tH "cry you-will-be," 
i. e., "you will cry" (cf. such Haida compounds as t^a'-ga "to 
eat-be," i. e., "to eat"), and Hupa. tc/i[n]-nL- rj-y a wt-n-tfe "he-came 
he-was- (thus)," i. e., "he always came." 

9. A highly important feature found in all Na-dene languages is 
the use of subjective or objective pronominal elements, according 
to the nature of the verb, to indicate its logical subject. This 
feature will be referred to again in connexion with the pronouns. 

More important than any of the specific features we have named 
is the similar manner in which the various elements going to make 
up a verb-complex are linked together. The resulting structure 
may be termed a sentence in miniature, not only psychologically, 
but, as is much less often the case in America, also morphologically. 

1 16 VI American Indian Languages 2 


Noun Structure. — There are a large number of monosyllabic 
noun stems, which may be used absolutely, in all three Na-dene 
languages. These are both of type cv and cvc. 

They often enter into composition, the qualifying noun regularly 
preceding (e. g., Chipewyan tUo-bec "grass-knife, mowing machine;" 
Tl. xa't-sla'x"* "root-hat"). In Haida, however, simple composition 
of noun stems does not seem to be often found (in such compounds 
as Masset i'tHadas 'ai "chief's blood," possessive -i or -a seems to 
have contracted with the final vowel of the second noun). Another 
type of composition which is particularly characteristic of Na-dene 
is the suffixing of a possessive element (Ath. *-e, *-ye, *-ye; Tl. -i, 
-71, -w, -wu; H. -ga, -l, -a) to the second member of the compound; 
this element indicates that the second noun governs the first, in 
other words that the first is genitively related to it (e. g., Chipewyan 
k^a-t'uwe "goose-lake;" Tl. slate a'n-i "moss town;" H. xo'ya 
tlu'-ga "raven's canoe, beans"). It is important to observe that 
finite verb forms may be nominalized or turned into relative 
clauses in Athabaskan and Tlingit by the suffixing of this possessive 
(better relative) element (its different forms in Athabaskan and 
Tlingit are due to phonetic factors) ; cf. Montagnais td!ai gay-e 
"plate which-is-white" with Tl. at-ci'-yt "those who can sing." 

Possessive pronouns are prefixed to nouns; they are identical 
in form with the objective forms used with verbs. Most nouns but 
terms of relationship and, generally speaking, those indicating parts 
of the body are in Athabaskan followed by the relative particle 
discussed above (e. g., Chipewyan be-di' "his head" but be-'ay-e 
"his snowshoes"). Similarly in Tlingit the relative suffix is regu- 
larly used with a possessive prefix except in the case of terms of 
relationship and, though not consistently, body part nouns (e. g., 
Tl. du-tla "his mother" but du-tc^u'n-i "his dream"). Haida 
possessive constructions are on the whole analogous to those of 
the other two Na-dene languages, the relative suffixes -ga and -i, 
-a being used in some cases, omitted in others (e. g.,/ dja'-ga "his 
wife," Masset / tla'l "her husband," Masset / k!u'g-i "its heart"). 

Derivative affixes (aside from nominalized verb forms) are quite 
sparingly used in Na-dene. A diminutive suffix is found in both 

Six: Athabaskan and Na-Dene Languages 117 


Athabaskan and Tlingit (e. g., Hupa (/je/o'-Zc "small storage-basket," 
Tl. a'-^/" "little lake"). Noun plurals (aside from Tl. collective 
-q!) are not often formed, though special plural forms for terms 
denoting human beings are found here and there (e. g., Kato 
skH'-k'^ "boys," Tl. du-k^a'ni-yen "his brothers-in-law," H. kfwai- 
ga-lar) "elder brothers"). 

Pronouns. — There are two classes of pronouns in Na-dene. 
Subjective pronouns are used as the subjects of active verbs (in 
Athabaskan of most verbs generally) ; objective pronouns as the 
objects of transitive verbs, subjects of neuter verbs (which may 
best be interpreted as objects of impersonal verbs), and possessive 
prefixes with nouns. While the two series are distinct as such, not 
all the respective forms are etymologically unrelated; in Tlingit 
there are one or two minor differences between the objective and 
possessive series. In both Athabaskan and Tlingit, as we have seen, 
the subjective and objective pronominal elements are integral parts 
of the verb complex, the possessive elements of the noun complex; 
in Haida the pronominal elements may be considered as inde- 
pendent words or, at least in part, as proclitic elements. A third 
series of pronouns is found in Athabaskan and Tlingit; these are 
independent denominative terms, which, however, have no influence 
on the form of the verb or noun. 

The employment of objective pronouns with verbs denoting 
states has been rather obscured in Athabaskan by the spread of 
subjective forms, but there are enough cases to make it clear that 
the impersonal verbs with objective pronominal elements char- 
acteristic of Haida and Tlingit were at one time better represented 
also in Athabaskan. 

The contrast between verb forms with subjective and objective 
pronominal subject is exemplified, e. g., by Tl. gu-x-t^u-si't "we 
will cook it" (with subjective pronominal element t^u- "we") and 
ha-k"-gu-wa-t!a "we will be warm" (with objective pronominal 
element ha- "us": "it will be warm to us"); by H. I q^a-t!al-gan 
''I got off" (with subjective pronominal element I "I") and di' 
skfistl-djdi'-ga " I am truly full" (with objective pronominal element 
di' "me": "it is truly full to me"); and by Hupa o-r]-xai "thou 

118 VI American Indian Languages 2 


art buying" (with subjective pronominal element -77- "thou") 
and ni[k]-^!-o-u<ar} "go to sleep!" (with objective pronominal 
element « I- "thee": "let it sleep to thee!"). That the verb forms 
with objective pronominal subject are indeed impersonals with pro- 
nominal object is made clear by comparing them with such transi- 
tive forms as Tl. ha-u-si-ne'x "they have cured us;" H. di' dala-q 
tl-gaxa-gd-ga "you tire me with your handling;" and Hupa yafn]- 
ni-l-t\r] "he picks thee up." The possessive use of objective pro- 
nominal elements is illustrated by Tl. ha'-q'aha'gu "our eggs;" 
H. di' gi'da "my daughter"; Hupa nL[t]-t''ai "thy paternal uncle." 

Postpositions. — Very characteristic of Haida, Tlingit, and Atha- 
baskan is a set of local and relational elements which regularly 
follow the noun or pronoun that limits them (e. g., H. st'al-ai st'a 
"the cliff from;" Tl. xa'na-de " evening-towards ; " Hupa nLfi-tcht] 
"ground-toward"). These postpositions offer remarkable morpho- 
logical and etymological analogies in the three languages. No less 
than about thirty-five Athabaskan postpositions and local verb 
prefixes (which, as w^e shall see, are in all probability postpositions 
in origin) can be more or less confidently stated to be cognate with 
corresponding Haida, Tlingit, or Haida-Tlingit elements. Out of 
twenty-five Hupa postpositions listed by Goddard, at least fifteen 
seem to be related to similar elements in Tlingit, Haida, or both. 
These facts show that the postpositional elements of Na-dene reach 
back, aside from certain later dialectic developments, into the 
earliest period of Na-dene linguistic history that it is impossible to 
arrive at by comparative evidence. 

In some cases it is possible to show that postpositions are 
nouns in origin, the complex of noun -+- postposition forming 
originally a compound noun. Thus, Hupa and Kato -lai, Chipewy- 
an -laye "on top of" is simply the noun stem for "end, top" com- 
pounded with the preceding element; Chipewyan -bQ. "around," 
the similarly employed stem for "edge" (Ath. *mar], *man). With 
Tlingit t!a "behind," Haida t!al "behind, back of," and Tl. k'a 
"on" compare respectively Chipewyan -t!a-ze "back" (body-part), 
Kato -/.''a "tail," and Navaho -^'a "surface." Hence it is intelligible 
that the same noun stem may in some cases have developed inde- 

Six: Athabaskan and Na-Dene Languages 119 


pendently into distinct postpositions in different Na-dene languages; 
e. g., Ath. *man, Tl. wan "edge" means "around" as Athabaskan 
postposition, "close to" as corresponding Tlingit element. The 
nominal origin of postpositions is further made very probable by 
the fact that they are frequently preceded by possessive pronouns: 
Hupa mi-ye "under it" (originally perhaps "its bottom") like 
mL[n]-nLr] "its face;" Tl. hasdu-q^a'nax "after them" (originally 
perhaps "their following") like hasdu-cayi'tia-yi "their anchor;" 
H. di' ga "to me" (originally perhaps "my vicinity") like di' 
go'ri-ga "my father." Whether we shall ever be able actually to 
demonstrate the nominal origin of all Na-dene postpositions is 
doubtful, but there can be little doubt of the correctness of this view. 

Postpositions often occur compounded among themselves. In 
some cases the analysis is evident (e. g., Chasta Costa -me'-qU 
"inside of" < "therein-at;" Tl. -k'a-q! "on" < "on-at;" H. 
gei-st^a "out of" < "in-from"); in others the two (or more) ele- 
ments have grown into a unit that can be analyzed only by com- 
parative evidence (e. g., Tl. t^a'yi "under" contains Tl. yi' "down 
in," but /'o* does not occur alone; comparison with Tl. t^a'-k "in 
the middle of," t\-n "with," and particularly Ath. *-/'a "among," 
shows fa'yi to have originally meant "down among"). 

Postpositions combine with verb forms in two ways, as local or 
relational prefixes and as syntactic suffixes. We have already 
indicated that several of the local prefixes of Athabaskan are merely 
postpositions in origin that have become somewhat firmly attached 
to the verb complex. Thus, Hupa xa- in xa-n-t^e "look for it!" 
is evidently etymologically identical with -xa in no'-xa "after 
us." In some cases the postposition comes after elements which 
can hardly be disconnected from the verb form, e. g., a-ya-l-tc! L[t\- 
-dtn-nt "he told them" (here -I "with" appears immersed in the 
verb, which demands the indefinite objective a- "it" as constant 
prefix; morphologically parallel is Tl. da- "to" in verbs of saying, 
e. g., ye da-ya-dii-q^a "thus to-him-spoke" like Ath. a-l-, which is 
doubtless identical in origin with postpositive -de "to," Masset da 
"to," Ath. *-d, *-de, * de-n "to, at"). These facts are not sur- 
prising when we bear in mind that the indirect object, nominal or 

120 VI American Indian Languages 2 


pronominal, precedes the verb and is followed by its postposition 
(e. g., Hupa xo'-xa t^(-r]-\n-t'e "him-for thou-wilt-look;" Tl. a-da 
a-o-H-t'aq! "it-around they-drifted;" H. la-gei la sk' U-nana-q-xida- 
i-as "it-into he began-to-chop-up"). It needs only the removal 
of the object (which then remains understood) from the postposition 
to bring the latter into closer touch with the verb. In the last 
Hupa example the removal of the expressed object {xo'-) leads to a 
form like the xa-n-tU first quoted. In Tlingit this use of the 
postposition as verb prefix with unexpressed object does not seem 
to be common, but examples abound in Haida, e. g., gei la q^a'-tdi- 
gan "into he went-in." In the last example gei is morphologically, 
as well as etymologically, parallel to Hupa verb prefix yt- "into" 
(cf. Chipewyan postposition -ye "in"). Here again we observe that 
Haida has allowed distinct elements to coalesce to a less extent 
than Athabaskan. In view of the tendency in Athabaskan for 
postpositions to become specialized as verb prefixes, it is not sur- 
prising if we find cases of such prefixes, no longer used as post- 
positions, corresponding etymologically to Haida postpositive 
elements. Such an example is Ath. *tse-, *tsi-r}- "away from, out 
of;" H. st'a "from, after" (for H. st': Ath. ts, cf. also H. st'a- 
classifier for ring-shaped objects: Navaho tsa-bg.'s "ceremonial 
hoop," yo-s-tsa "ring"). 

A verb form as such is sometimes conceived of as nominalized 
and is followed by a postposition which serves to subordinate it. 
Thus, in a Chipewyan form like hi-l-tc!t-t!a "because he was angry," 
postpositive -t!a "with, on account of" (cf. he-tla "with it") 
nominalizes and subordinates hi-l-tde "he was angry" ("he-was- 
angry because-of," "because of his being angry"). Such syntactical 
developments have taken place independently in the Na-dene lan- 
guages, to some extent even in the various Athabaskan dialects. 
This is indicated, among other things, by the fact that even where 
two Na-dene languages have employed the same postposition for 
syntactic purposes, the use to which it is put is different (e. g., Ath. 
*-de "if" and Tl. -/ "in order to" both go back to Na-dene post- 
positive *-da "to, at"). Tl. -ya and -«, which make subordinate 
clauses out of verb forms (e. g., has a-ga-ca'-n "when they marry;" 

Six: Athabaskan and Na-Dene Languages 121 


a-t^e'-x-ya "when she slept"), are doubtless identical with post- 
positive -ya "in the neighbourhood of" and -n "with, at."^ In 
Haida temporal clauses are formed by nominalizing verb forms by 
means of suffixed demonstrative {g)ai "the," these being then 
followed by postpositive dliC. Subordinate clauses formed by 
means of postpositions without preceding {g)ai also occur (e. g., 
Masset / k!ota'l-an sd-e-t "after he died," literally "he died place- 
the-to"). The degree of coalescence of postposition and verb 
is again much less in Haida than in Tlingit and Athabaskan. 

Summary. — It has become evident that the morphologies of 
Haida, Tlingit, and Athabaskan present numerous and significant 
points of comparison. Despite not unimportant differences of 
detail, the same fundamental characteristics are illustrated in all 
three. In not a few cases elements (or even processes) which are 
thoroughly alive in one of the languages linger on merely as sur- 
vivals in another (e. g., -xa, freely used in Haida as distributive 
suffix with numerals, postpositions, and nouns, lingers on in Tlingit 
as compounded -na-x after numerals and as sporadic noun plural 
*-k\ *-k'e, *-k'ai in Athabaskan). 

Considerable specialization must, of course, be allowed for. 
Peculiar to Haida are the development of a large class of nominal 
classifiers, a great exuberance of composition of verb stems, the 
development of a set of local suffixes in the verb, and greater 
looseness in the treatment of pronominal elements and postposi- 
tions. The synthetic tendency has gone farthest in Athabaskan, in 
which, e. g., pronominal subject and "modal" element often unite 
inextricably (there are, however, analogies to this in Tlingit). 
Tlingit, on the whole, seems to have the smallest number of purely 
distinctive morphologic features. It shares with Athabaskan a 
lesser degree of independence of pronominal elements, a great 
development of verb prefixes denoting aspects and, it would seem, 
the employment of internal stem changes for tense differences. 
As in Haida, the distinction between active verbs with subjective 
pronouns and static verbs with objective pronouns is better pre- 
served than in Athabaskan. 

> My interpretation of Tl. -/, -ya, and -n as syntactically specialized postpositions 
diffeis from Swanton's, at least as far as expressed in his grammatical sketch. 


VI American Indian Languages 2 


II. Comparative Vocabulary 
The lexical evidence bearing on the genetic relationship of 
Athabaskan with Haida and Tlingit comprises, at the moment of 
writing, over three hundred distinct Athabaskan stems and gram- 
matical elements which can be, with greater or less probability, 
assigned to the reconstructed Na-dene language. Only a selection, 
comprising less than one third, of this lexical material is here 
presented. The arrangement is alphabetical, from the point of 
view of Athabaskan. 

I. a- demonstrative stem 


a- dit. 

-ade "elder sister" 

3. -ca-rj obligatory future 

4. -ca, -cal "to catch with a 


5. -d, -de "at, to" 

6. da "what?" 

7. -do, -da/ "to go, to travel" 

8. -da-f) "to drink" 

9. -das "to burn" 

10. -de/ "several go" 

11. del "crane" 

12. di "this" 

13. -dja hortatory 

14. djarj "mud" 

15- -ga?7 "to be mouldy" 

16. -go " toward" 

17. -gid, -yid "to dive" 

18. -7a "for," -7an "to" 

19. -ya "to go" 

20. -ye, xe "grease" 

21. -yed "to run" 

22. -yel "to be dark, night 


-sa-rj infallible future 

djd "bait" 
-da "to" 

-dal "to move along" 

da'dj, {das) "live coals" 
dal "several go by 

di'la dit. 

dei "just that way" 
-dja-T) dit. 
tc^a'n dit. 
gu'na "decayed" 
gtia, gui dit. 
gi' dit. 

ga "to," gan "for" 
-ga "to go in order to" 

ga't dit. 
ga'l " night" 


a- dit. 

a't "father's sister, 

father's sister's 


cat "to seize" 
-/, -dt "to" 
da' "what?" 

dana "to drink" 

at "several go" 

du'l dit. 
de "now" 

ga "for" 

-ya "to go to" 

e'x "grease" 

get "to get dark' 

Six: Athabaskan and Na-Dene Languages 




23. -ye, -yel "to kill, to fight" 

24. -ywo "tooth" 

25. hai "that" 

26. he-, xe- "they" 

27. X0-, he- "he, him" 

28. -^'a "liquid has position" 

29. -k'an, -k'a "to fish with a 


30. -k^a "on" 

31. -^'e personal noun plural 

32. k'ene "friend" 

33. kla "arrow" 

34. kleri "withes" 

35. -klan "to burn" 

36. -k!e "on" 

37. -k'os, k'es "to tie" 

38. 4 "with" 

39. 1-, la negative 

40. la "one" 

41. -la "to jump" 

42. -lad "end" 

43. lo', lok' "fish" 

44. -///a "butt; behind" 

45. me- "he, it" 

46. man "edge" 

47. mes "cheek" 

48. -n, -77 local postposition 

49. -na "to die" 

50. -nan "to drink" 

51. -ne, -n "person, people" 

52. -neg, -leg "to relate" 

53. -ne "to play" 

54. -ni "to touch, to do with 

one's hands" 

55. no "place of retreat, is- 


56. -onay{e) "older brother" 

57. 5- durative verb prefix 

gail "to fight" 

hao " that " 

xao "liquid" 
xao "to fish" 


-xa distributive suffix 

q!a' "harpoon' 
qian "grass" 

k^u "to tie" 
al "with" 

ila'- "the first' 
tla' "to dive" 
tlan "end" 

-dlga "after" 
wa- "that" 

-n, -rj general post- 

ntal, nil dit. 

na "to live; house" 

nar] dit. 

ux "tooth" 
he " this " 
has "they" 
hu "he" 

-k'a "on" 

-na-x distributive num- 
eral suffix 
xo'n "friend" 
q!a "point" 

q!a'n "fire" 
-q! "at" 

I negative 
tie'- "one" 

tUu'k! "cohoes" 

we "that" 

wan "edge" 

wac "cheek" 

-n "with," also local 

na "to die" 

na "people" 
nik "to tell" 

ni "to put" 

mi "fort" 

hu7ix "man's older 

s- modal prefix 


VI American Indian Languages 2 


58. -sen "to hide" 

59. sil "steam," -sil "to 


60. -/'a "among" 

61. /'a "wave" /'a- verb prefix 

referring to water 

62. -t'an "to eat" 

63. -t'e "to look for," -t^an "to 


64. -/'e "to be cold," /'en "ice" 

65. -t'e "to lie" 

66. i^ez "night" 

67. -tla "because of" 

68. tla "feather," -t!a "to fly" 

69. -tla "tail," tIa-Tj "back- 


70. -ties "to step" 

71. -tlo "to shoot" 

72. -tlod "to rub" 

73. tsa "ring-like object" 

74. tslai "dish" 

75. -tslen "bone" 

76. tsli "again" 

77. -/c't "grandfather" 

78. -/c'ot; " mother-in-law " 

79. tela "hat" 

80. tcio "fir, spruce" 

81. -ifa "for" 

82. xa- "up, out of" 

83. xfl' "goose" 

84. xa-T) "quickly" 

85. -xan, -yan "to grow up," 

xan "old age" 

86. xin "song" 

87. -ya "to stand ' (plur.) 

88. -yan "to eat" 


sil "to steam" 
t^a-oan "alongside of" 
t^a'ria "sea-water" 
fa dit. 

/'a/ "cold" 
/'at, /'t "to lie" 

-tla dit. 

tiagun "feather," tlao- 
" feather-like object ' 
-tial "back of " 



Sin "to hide" 
si't "to cook" 

-Ca'-k^ "in the middle 

/'a'g/ "to chew" 
tiln "to see" 

t^a'dj "cold" 

/'at "to lie," /'a "to 

/'c*/ "night" 

tia-wu "feather" 

•tla "behind" 

tla "to step" 

st^a- "ring-shaped ob- 

sqlao "to put in a 

tc\n dit. 
djo'n dit. 

tcju "cedar" 

k'wagi "above" 
xaha "mallard" 
xao- "to do a thing 

ga "to stand" 

tluk "to shoot' 
tlus "to rub" 

shql "dish" 

sla'q "bone" 
tslu "again" 

tc^a'n dit. 
sja'x" "hat" 

-xa-n "to" 
k^e "upward" 

can "old person" 
ci' "song" 
ya'n " to eat" 

Six: Athabaskan and Na-Dene Languages 







ye- "that," y- "he" 


'this, that, the" 

ya "this" 


-ye "at the foot of, under" 

-yi' "down in" 


-ye personal noun plural 

ye "supernatural being," 
-yen "to practice sha- 

-ye-n plural of terms 

of relationship 
ye'k "supernatural 



-ye "in" 



•ge' "inside of," -yi-k 


-ye suffix making relative 

-7i suffix making rela- 
tive clause 


yo "that yonder" 



yu "that yonder" 


-'a "to go" 


'to walk" 


-a "to tell, to sing" 

g'a "to say" 


-'in "to see, to look" 



gen dit. 

III. Phonology 

The phonetic systems of Athabaskan, Haida, and Tlingit, 
despite a good many differences of detail, present important points 
of similarity. Three types of stops are found in each : intermediate 
(or sonant), aspirated surd, and glottalized (fortis). Sibilants and 
sibilant affricatives, k-spirants, and laterals are well developed. 
A remarkable phonetic feature held in common by the three Na- 
dene languages is the paucity of labials; b, p\ and p! were clearly 
not found in Na-dene (b and />' are rare Haida sounds), m existed 
only doubtfully (Ath. m, whence b in certain dialects, is not equiva- 
lent to Haida m, but to Haida-Tlingit w), while w was certainly 
found. Athabaskan has lost the old velar series of stops as such, 
while Haida and Tlingit have preserved them; on the other hand, 
the Na-dene anterior palatals, best preserved in Haida, have been 
lost as such in Tlingit. 

In the more elaborate paper on the Na-dene languages in course 
of preparation, the historical relationship of the Athabaskan 
sounds to their Na-dene prototypes and Haida and Tlingit corre- 
spondents is systematically worked out on the basis of all the 
evidence available. Here it will suffice to point out some of the 
more important correspondences, referring to the numbered entries 
of the comparative vocabulary for illustrative examples. 

126 VI American Indian Languages 2 


Stopped Consonants 

1. Ath. d: H. d, -t: Tl. d, -t (nos. 2, 5, 6, 7, 8, 9, 10, 11, 12, 21) 

2. Ath. /': H. /": Tl. /' (nos. 60, 61, 62, 63, 64, 65, 66) 

3. Ath. //: H. //: Tl. // (nos. 67, 68, 69, 70, 71, 72) 

4. Ath. g: H. g: Tl. g, -k (nos. 15, 16, 52) 

5. Ath. /fe": Tl. k' (no. 30) 

6. Ath. /fe": H. x: Tl. y (nos. 28, 29, 31, 32) 

7. Ath. x: H. jfe'(M^): Tl. k' (no. 82) 

8. Ath. k': H. fe' (no. 37) 

9. Ath. 3/,(^): H.^ (nos. 17, 87) 

10. Ath. ife/pH. g/PTl. g/ (nos. 33, 34. 35, 36) 

11. Ath. ': H. g': Tl. g' (nos. 96, 97, 98) 

12. Ath. 7: H. g: Tl. g (nos. 18, 21, 22, 23) 

13. Ath. y: H. g: Tl. g, 7 (no. 93) 


14. Ath. m: H. w: Tl. w (nos. 45, 46, 47) 

15. Ath. n, -7)-. H. n, -77: Tl. n (nos. 3, 8, 14, 15, 32, 34, 35, 46, 48, 49, 50, 51 

52, 53- 54. 55- 56, 58, 63, 78, 85, 88, 98) 

16. Ath. /, I: H. /, I: Tl. I (nos. 4, 7, 10, 11, 22, 23, 38, 39, 59) 

17. Ath. s: H. s, dj: Tl. 5 (nos. 9, 57, 58, 59) 

18. Ath. c (j): H. 5, dj: Tl. c (nos. 3, 4, 47) 

19. Ath. x{> c in most dialects): Tl. c (nos. 85, 86) 

20. Ath. y: H. g: Tl. y (nos. 88, 89, 95) 

21. Ath. y (before front vowel): H. g (g): Tl. 7 (nos. 90, 91, 92, 93, 94) 

22. Ath. x: H. X, x: Tl. x (nos. 20, 81, 83, 84) 

23. Ath. h, x: H. h: Tl. h (nos. 25, 26, 27) 

24. Ath. 7: H. g: Tl. 7, -x (nos. 19, 20, 24, 56) 


25. Ath. I: H. /^: Tl. // (nos. 40, 41, 42, 43) 

26. Ath. dj: H. dj, tc' (nos. 13, 14) 

27. Ath. /c': H. tc\ dj: Tl. tc' (nos. 77, 78) 

28. Ath. tcl: H. /c/: Tl. s! (nos. 79, 80) 

29. Ath. ts!: Tl. /i/, 5/ (nos. 75, 76) 

30. Ath. tsl: H. til: Tl. tl! (e. g., Ath. *-/5.'e "penis": Tl. tllel dit.; Ath. *-tsIi 

"to sit [plur. sub].]": H. tl!a-{o-) dit.) 

Consonant Clusters. — The study of Na-dene sibilants and sibilant 
affricatives is rather involved and presents several difficulties. The 
summary given above (17, 18, 19, 26, 27, 28, 29, 30) exhibits some of 
the main developments. An important group of cases is afforded by 
Haida stems or elements beginning with consonant clusters whose 
first element is a lateral (I, tl!, dl) or a sibilant (5). Swanton, in 

Six: Athabaskan and Na-Dene Languages 127 


commenting on these clusters, surmised that they were perhaps 
due to the prefixing of an old morphological element (e. g., 1-, s-). 
There is, however, no evidence whatever to support this. On the 
other hand, I have at my disposal upwards of twenty such examples 
which point clearly to the inference that these Haida clusters were 
found in Na-dene and correspond to lateral and sibilant affricatives 
in Athabaskan and Tlingit. The following relations can be estab- 
lished : 

A. 31. H.lg-: Ath. ///- (?) 

32. H.lq'-: Ath. tU- 

33. H. dig-: Ath. tU- (no. 44) 

34. H. It'-: Ath. /c'- 

35. H. H.'x-: Ath. tcl- (?) 

36. H. tUd-: Ath. fsl- (cf. 30.) 

B. 37. H. sg-: Ath. Is- (?) 

38. H. sq!-: Ath. ts!-: Tl. si- (no. 74) 

39. H. sg-, (sk'w-): Ath. tc'-: Tl. tc'- 

40. H. sqlw-: Ath. tcl- 

41. H.St'-: Ath. ts- its!-): Tl. t'- 

42. H. St'-: Ath. tc'- 

43. H. St!-: Ath. id- 

Vowels. — The great majority of vowel correspondences is per- 
fectly intelligible; a certain number of unsolved problems still 
remain. In comparing Tlingit with Haida and Athabaskan. forms, 
it is necessary to bear in mind that, under as yet undetermined 
circumstances, Tlingit a has developed to e (e. g., Tl. xa "to eat": 
a-xe'-x "he ate"; Tl. de'x "two": daxa-ducu "two plus five, 
seven"; Tl. /'a "stone": /V-5.' "stones"). A more important prob- 
lem is presented by Ath. e (doubtless open in quality), which is not 
to be directly compared with Tl. e. It is clear, both from internal 
Athabaskan evidence (e. g., Ath. *-t^e and *-i^a "to look for"; 
Ath. *t'e'- "in the water," *t'a- "water") and, still more, from 
comparison with Haida and Tlingit, that Ath. e (which must be 
assumed for the earliest Athabaskan period) has developed from 
Na-dene a; less frequently Ath. e goes back to Na-dene i. Under 
what phonetic circumstances, however, Na-dene a has remained as 
such in Athabaskan or become e is not clear for the present. This 
I believe to be one of the most important problems of Na-dene 

128 VI American Indian Languages 2 


Some of the more important vocalic correspondences are: 

44. Ath. a: H. a, a: Tl. a, a', a, {e) (nos. i, 2, 3, 4, 6, 7, 8, 9, 13, 14, 18, 19, 

25. 28, 29, 30, 33, 35, 40, 41, 42, 44, 46, 49, 60, 61, 62, 67, 68, 69, 73, 
74, 79, 81, 82, 83, 84, 85, 87, 88. 96, 97) 

45. Ath. e (sometimes reduced to e^)'- H. a, a, «: Tl. a, a, a, (g) (nos. 5, 

10, 20, 21, 22, 26, 31, 34, 45, 47, 51, 53. 64, 65, 66, 70, 75, 89, 92) 

46. Ath. i: H. i, t, i, ei: Tl. t", , e e (nos. 12, 17, 54, 59, 77, 86, 98) 

47. Ath. 0: H. M, 0': Tl. m, «• (nos. 16, 24, 27, 37, 43, 55, 56, 71, 72, 78, 80. 95) 

Whether or not Na-dene possessed pitch accent must remain un- 
decided for the present. Its presence in Tlingit and a few remarks 
by Morice and Legoff as to its possible existence in Athabaskan 
make this not improbable. Should this prove to be the case, some 
of the phonological difficulties in Athabaskan and Tlingit vocalism 
may be solved (e. g., Ath. e < Na-dene d, Ath. a < Na-dene a). 
All this, however, is quite vague as yet. 

IV. Conclusion 
The main conclusion to be derived from the selected morpho- 
logical, lexical, and phonological evidence that we have passed in 
review is, I believe, obvious. Athabaskan, Haida, and Tlingit 
must be considered genetically related. The correspondences are 
of so intimate a character that mutual borrowing of words and 
morphological features seems out of the question. It is, however, 
no less obvious that each of these languages is very distinctive and 
represents a highly differentiated form of the Na-dene prototype. 
In no sense can Haida, Tlingit, and Athabaskan be said to form a 
continuum comparable to that of the Athabaskan dialects when 
these are compared among themselves. Each Na-dene language has 
evidently passed through a very long period of development in 
linguistic isolation from its sister languages. It would be rash, in 
the present state of our knowledge, to dogmatize on the relative 
conservatism of the Na-dene languages. I would venture to sug- 
gest, however, that Haida has remained the most faithful to the 
original sound system of Na-dene, but that, on the whole, the 
original morphological features are best preserved in Tlingit. 

' This weak vowel is differently colored in different dialects; e. g., Hupa i (less 
frequently a). Carrier a, Navaho ». 

Six: Athabaskan and Na-Dene Languages 129 


Several facts suggest that Tlingit and Athabaskan may have had 
a common Hnguistic history after Haida had become differentiated, 
Imt too much should not be made of this. 

The name that I have chosen for the stock, Na-dene, may be 
justified by reference to no. 51 of the comparative vocabulary. 
"Dene," in various dialectic forms, is a wide-spread Athabaskan 
term for "person, people"; the element *-ne (*-w, *-ri) which 
forms part of it is an old stem for "person, people" which, as suffix 
or prefix, is frequently used in Athabaskan in that sense. It is 
cognate with H. na "to dwell; house" and Tl. na "people." The 
compound term "Na-dene" thus designates by means of native 
stems the speakers of the three languages concerned, besides con- 
tinuing the use of the old term Dene for the Athabaskan branch 
of the stock. 

An important ethnological consequence of our linguistic results 
is that a demonstration is at last given of the northern provenience 
of the Athabaskan-speaking peoples. So long as Athabaskan was 
counted a separate linguistic stock, there was no conclusive k 
priori reason for considering its Pacific and Southern branches as 
having spread out from the northern group. Under the present 
circumstances a southern drift of Athabaskan-speaking tribes can- 
not seriously be doubted. The center of gravity of the Na-dene 
languages is clearly in the northwest, in southern Alaska and 
adjacent parts of northern British Columbia and southern Yukon 
Territory. Owing to the great linguistic gulf separating Haida and 
Tlingit, I should be inclined to consider the coast of southern 
Alaska, the present home of the Tlingit Indians, as the most likely 
region in which the Na-dene languages developed. The Athabaskan 
branch of the stock undoubtedly formed a relatively undifferentiated 
unit long after Haida and Tlingit had become differentiated from 
each other. The Athabaskan dialects have so many distinctive 
traits in common that it is perfectly evident they have had a 
long history in common. They may be considered a specialized 
interior offshoot, just as Haida is a specialized island offshoot. 

130 Vf American Indian Languages 2 

Editorial Note 

Originally published in American Anthropologist 17, 534-558 (1915). 
Reprinted by permission of the American Anthropological Association. 

Nineteenth-century Russian linguists speculated about a possible rela- 
tionship among Athabaskan, Tlingit, and Eyak, an isolate of the Copper River 
area (Krauss 1964; Pinnow 1976). Sapir was almost certainly unaware of this 
work, and saw data on Eyak only quite late in his career, after Birket-Smith and 
de Laguna's ethnographic field work in 1933. The germ of Sapir's Na-Dene 
hypothesis, as he himself notes (p. 534), lay in the speculations of Boas and 
Swanton (especially Swanton 1908: 472-485) about a historical connection 
between Haida and Tlingit. It was the position of Haida, rather than that of 
Athabaskan or Tlingit, that was central to Sapir's view of the relationship. 
Haida, with its relatively uncomplex morphology, seems an improbable con- 
gener for languages so thoroughly "polysynthetic" in spirit as Athabaskan and 
Tlingit, but in Sapir's view the convoluted morphosyntax of Athabaskan- 
Tlingit was a late development, and the more open structure of Haida words 
represented either an archaic situation or a quite different development from 
the "isolating" Na-Dene proto-language. Most of Sapir's lexical comparisons 
with Haida depended on the assumption of morphosyntactic changes of this 

Sapir's Na-Dene proposal was not well received. Both Boas (1920) and God- 
dard (1920) disparaged Sapir's use of a genetic hypothesis to explain 
resemblances — particularly morphological similarities — that they felt could 
better be explained as borrowings. The issue of Athabaskan-Tlingit-Haida 
"morphological borrowing" remained alive until at least the 1950s, when Dell 
Hymes in an important series of papers proposed a technique of "positional 
analysis" to assess the historical value of structural resemblances of this sort 
(Hymes 1955, 1956). 

Research in recent decades has clearly established the genetic relationship of 
Athabaskan and Eyak (Krauss 1964, 1965). A genetic relationship between 
Athabaskan-Eyak and Tlingit is accepted by most scholars, but is seen as dis- 
tant and problematic (Krauss 1968, 1969; Krauss and Leer 1981; Pinnow 1966). 
In the recent literature, it is this grouping that is usually referred to as "Na- 
Dene." Krauss has called Sapir's belief that Haida belongs in Na-Dene "an 
illusion" based on "mistranscription, misanalysis, mistranslation, and/or mis- 
interpretation" (1979: 841). Levine, in an influential paper (1979), has given a 
detailed critique of the "spurious" evidence on which Sapir based his hypoth- 
esis, tracing many of the supposed errors to Swanton's deeply flawed descrip- 
tion of Haida (1911). In his recent general treatment of linguistic relationships in 
the Americas, Greenberg (1987: 321-330) rebuts Levine and wholeheartedly 
accepts Sapir's more inclusive version of Na-Dcne , which he sees as an intrusive 
group unrelated to any other American Indian linguistic stock. Pinnow also 
regards the relationship of Haida to Athabaskan-Eyak-Tlingit as likely to be 

Six: Athabaskan and Na-Dene Languages 131 

genetic, and has published considerable lexical evidence for the relationship 
(Pinnow 1985a, b,c, 1986a, b, 1988, 1990), as well as a comprehensive history (in 
German) of Na-Dene research from 1798 through 1976 (Pinnow 1976). 
Sapir's ms. corrigenda on his own offprint are as follows: 

Original For: Read: 

p. 545, 1. 32 -ga -ga 

p. 546, 1. 6 ga-laer) §a-laer) 

p. 553, entry 63 t'i!n t'in 

p. 553, entry 79 sja-x" s!a-x" 

p. 553, entry 80 tcju tc!u 

ga ga 

p. 553, entry 87 

On p. 535, lines 1-3, the text from "the consonant ..." to "Tlingit (Tl.)." is 
bracketed with a question mark by Sapir. On p. 541, footnote 3, Sapir indicated 
the last sentence of the note with the word "revise!" On p. 547, last paragraph, 
Sapir questions the statement "Chipewyan-/7fl. . . *man)/' On p. 549, end of first 
paragraph, Sapir questions the sentence "Such an example . . . 'ring')." In the 
Comparative Vocabulary, Sapir questions the Tlingit forms in entries 22 and 
33 and the Haida forms in 60, 74, 89, and 95; adds Tlingit na 'to drink' to entry 
50; and recalculates the total entries from 98 to 95. On pp. 555-556, Sapir 
questions nos. 30, 33, and 38 in the list and states that "This paragraph [headed 
"Consonant Clusters"] should be eliminated. Swanton is right." On p. 557, first 
text paragraph, Sapir questions the sentence "Should this prove . . . Na-dene 
^/)." On p. 558, Sapir notes, beside the first three lines, "More certain now." 

The Sino-Dene Hypothesis 
[excerpt from a letter to A. L. Kroeber] 

I have long wanted to write you about Nadene and Indo-Chinese, but my 
evidence accumulates so fast that it is hard to sit down and give an idea. Let me 
say this for the present. If the morphological and lexical accord which I find on 
every hand between Nadene and Indo-Chinese is "accidental," then every anal- 
ogy on God's earth is an accident. It is all so powerfully cumulative and inte- 
grated that when you tumble to one point a lot of others fall into line. 1 am now 
so thoroughly accustomed to the idea that it no longer startles me. For a while I 
resisted the notion. Now I can no longer do so. 

The chief stumbling-blocks in the way of a general acceptance of the syn- 
thesis would be: 1. Failure to realize the very exceptional type of language to 
which Nadene belongs. It is really quite alone in America, so far as I can see. 
The contrast between it and Eskimo, Wakashan, and Algonkin is tremendous. 
2. Failure to realize that the Nadene languages are not one-third as synthetic as 
they look. Go at analysis half-way decently and get into a critical perusal of 
connected text and you soon realize that the complex verb breaks down into a 
cluster of very live elements, each of which has a syntactic or positional value, 
not merely as "affix" but as radical element. Haida in particular, I find, is 
extremely analytic. It knows no "prefixes" and "suffixes" at all except for certain 
important survivals that Swanton does not even mention (e.g. causative and 
denominative s- in s-kit to 'handle a club, to club' < kit 'club', cf. Tlingit sl-, 
transitivizing prefix; also Haida /- = Tlingit it- = Athabaskan "3d modal" /-, 
also primarily transitivizing). What Swanton calls affixes are all independent 
stems entering into composition, or even little verbs. There is no "tense suffix" 
in Haida, merely a series of enclitic phrases consisting of demonstrative + 
particle verb of being. His "continuative" -gAn, e.g., is simply 'that-is 
(duratively)', his "imminent future" -asan is really -'a-sa-i} 'this-will-be 
(durative)', and so on. It all crumbles to pieces at the least touch. I think the 
same will prove true of Athabaskan-Tlingit, though here the integration is 
more thorough. But I no longer seriously believe we have the right to consider 
anything preceding the "second modal" elements (Ath. 7-, n-, and v-) as part of 
the verb and am beginning to doubt if even these elements and the subjective 
pronominal "prefixes" are part of the true verb. I think it more than likely that 
such an Ath. form as *yasectk'os (purely theoretic form ad hoc\ Not to be mis- 
taken for genuine Ath.) 'I picked up a flexible object' is to be analyzed as *ya se 
c l-k'os 'up it-is (that) I handlc-a-flexible-object'. *l-k'os is the verb; the rest is a 
series of somewhat reduced independent elements that follow in a definite 
order. 3. The third prejudice to overcome is the nature of Indo-Chinese itself. 

134 VI American Indian Languages 2 

Modern Chinese is a very secondary development. The most typical represent- 
ative of the earlier stage is Tibetan — which is startlingly Nadene-like. It has 
those fundamentally important "3d modal" elements of Ath., Tlingit, and 
Haida (e.g. du- 'to be together': s-du- 'to cause to be together, to assemble'; in 
fact, 4 of its more important verb prefixes, which are "voice" elements, seem to 
me to correspond in form and meaning to Nadene elements — s- to Tlingit- 
Haida s-; r- to Tlingit-Haida-Ath. /-; d- (medio-passive) to Tlingit-Ath. d-, 
survivals also in Haida; nasal prefix to Ath. ''3d modal" -n-, -r)-, of mysterious 
value but probable active intransitive). Moreover, Tibetan has vocalic ablaut in 
its verbs (e.g. X]-gex]-s 'to fill', perfect b-kax], fut. d-gax}, imperative /c'oij). 
Again, the transitive verb is really passive, as in Tlingit. In both, for instance, 
you would say 'Man-by horse kill' = 'The man killed the horse', Tlingit agen- 
tive -tc corresponding exactly to Tibetan instrumental -s. In both Indo-Chinese 
and Nadene, postpositions are of extreme importance and serve to subordinate 
preceding verbs and clauses. Indeed, reading Tibetan text gives you precisely 
the same feeling as reading Haida text. I wish I had time to illustrate. In both 
groups the fundamental element is really a noun, the verb a kind of 
denominative structure. In brief, I should say that the similarity in feeling 
between Tibetan and Nadene is at least as close as between Latin and English, 
probably closer. Thus the theoretical road to a synthesis is clear. And the lexical 
evidence is startling. You would be amazed at some of my material. Things like: 

1. Tlingit k'a 'surface'; Navaho k'd 'surface': Tibetan k'a 'surface' 

2. Chinese fan 'charcoal': Haida s-fan 'charcoal' 

3. Old Chinese ti 'this': Ath. di 'this' (Ath. di really means ti) 

4. Old Chinese ti 'pheasant': Ath. di 'partridge' 

5. Nadene k'u 'hole' (TI. k'u-q'" 'hole', t'a- tu-k'^ 'cave' = 'rock-interior- 
hole', yoL-k'o 'to fall into a hole'; Nav. k'o, e.g. ts'e-k'o 'rock-hole' = 'canyon'): 
Indo-Chinese k'u 'hole' (dozens of forms, e.g. Tib. /:'w-r) 'hole', Karen k'u, 
with falling tone, 'to dig a hole'). 

These are only a drop in the bucket. Naturally it is a big problem and there 
are going to be hundreds of knotty points to unravel. But I do not despair. My 
present plan is to proceed as follows. First, to prepare part 1 of a Nadene com- 
parative study, to consist of my present lexical material (about 300 comparable 
radical elements, to which I add constantly). In this I would give reconstructed 
Athabaskan but also actual Ath. dialectic forms. Before publishing parts 2 and 
3, on morphology and phonology, which need much preliminary work, I intend 
to publish special papers on selected portions of Nadene grammar, e.g., certain 
archaic post-positions; or demonstrative stems; or general points of syntax. In 
this way I shall be keeping the problem live and accumulating experience for 
the definitive Nadene study. Of course I shall have to do Ath., Haida, and 
Tlingit in the field. I want particularly to gather a large amount of purely lexical 
material. People do not realize how scanty is our material, and for my purpose, 
which is comparative, I need stacks of it. What Goddard gives us is a miserable 
pittance — and wretchedly analyzed or not analyzed. Did you read his attack on 
my Nadene? You can have no idea of the laughable errors he commits. It is the 

Six: Athabaskan and Na-Dene Languages 135 

work of an utter groundling that does not know his own material. What do you 
think of a man who expects you to unravel the complex phonology of Nadene by 
drawing up an alphabetical list of Tlingit ''stems" and "matching" them with 
random "stems" from Kato or Ten'a ad libitum'? Very much as though you 
"matched" Sanskrit words now with French, now with Portuguese. Great 
method, what? And what do you think of a man who rules out comparisons 
because he does not "know" the Ath. form in question? Particularly when the 
form occurs in a book (Chapman's Ten'a) which he has "edited"! I may reply to 
Goddard, but it is really no use. He is a man of no more than average linguistic 
ability, completely at the mercy of his local sentimental memories, and abso- 
lutely without vision as to the older drift of Ath. He probably imagines his lists 
of stems are the last word on the whole subject. The degree to which he has 
failed to analyze his material is shocking in the extreme. 

I shall not broach the Indo-Chinese part of the problem till I have moored 
myself more completely in Nadene. The final plan is: 1. a Nadene comparative 
grammar to be published in 3 parts (possibly an Ath. etymological dictionary as 
a side-show); 2. a Nadene-Indo-Chinese demonstration; 3. a more general 
treatment of the evolution of the whole group, showing how old types have been 
replaced by new ones. As a starter, I am at work now on a paper on Haida 
phonetics, which may interest you when you see it. 

P.S. I cannot resist the temptation to give a somewhat livelier idea of the 
remarkable way in which lexical elements are interwoven in Nadene and Indo- 
Chinese. I have some cards along, so don't need to trust to memory. I shall give 
an idea of the richness of some of my entries by dealing with a group of related 

In Ath. we have a stem *fu, post-vocalic *-Iu, which may be rendered as 'coiP 
or loop'; e.g. Nav. to loop', as vb.: Nav. -lo 'to catch with a rope', Jic. Apache 
-lo' 'to lasso', Chipewyan -lu, -tu 'to be caught in a net or noose'. So far, so 
good. Here our friend Dr. P. E. Goddard would end. But it is difficult to believe 
that Ath. */76>/'rope, strap' (found in all dialects) is unconnected. How? With- 
out going into details (it would take too long), I may say that I feel justified in 
analyzing *t'lo-l into *t-'lo--i. How *-'lo-, -'lu is related to *{u, -lu I cannot yet 
tell, but I strongly suspect Nadene had both / and '/, and in related stems. As 
you will see from my Haida paper, Haida has both / and '/; in Tlingit '/ probably 
became /'. I should guess that 7 is causatively related to /,- Ath . *///, -lu is intr. : 
'loop; to lie coiled', *- '/w would be 'to cause to be coiled, to make a loop, to tie 
around' (possibly -lo' is a secondary form of *-'lu). Now -/ we know to be con- 
tinuative or usitative; and /- is medio-passive. Hence *t-'lo-l is what is always 
caused to loop around, what loops about something', in other words, rope, 
strap'. We learn important things from such an analysis: that "3d modal" ele- 
ments were welded with verb stems and appear in nominal derivatives; that 
there was an old alternation /.• '/ whose significance remains to be discovered. 
That we are on the right track is confirmed by another common Ath. stem 

136 ^^ American Indian Languages 2 

whose formation is precisely parallel to that of *t'to-i. This is *t'iei 'fire-drill'. 
Fortunately we are here not dependent on Goddard's material alone. From 
Petitot we learn that in Hare and Loucheux there is a verb *-/e (-d-le, -t-le, -l-le) 
meaning 'to revolve' (words involving it are: 'virer au cabestan', 'cylindrique', 
'tourbillon', 'tourner', 'se tourner'). Hence *r-7e-/ is 'what keeps turning itself, 
what revolves drill-like'. This parallelism of *r7o-/ 'rope' to */7e/ 'fire-drill' is, of 
course, highly suggestive. It shows that many of Dr. Goddard's "stems" may not 
be pure father-Adam radicals. And we see that Ath. fl fails to correspond to 
Haida and Tlingit fl for a reason. It is a secondary development in probably all 
3 groups. Such a sound as /' appears in cognate words throughout; not so fl — 
which fact alone casts a reflex light on our analysis. 

Let us proceed. To Ath. *hi, -lu is clearly related Ath. *-lui 'to wrap around': 
Hupa -loi 'to tie, to wrap around', Kato -//' 'to tie up' (old Ath. form possibly 
causative *-'lui). And further, having once allowed Ath. t'{ to analyze itself into 
medio-passive t- (d-) + '/, we do not feel we are doing anything ungodly to 
analyze Ath. *fVw' 'grass' into *t-'lu' 'what is wound (in basketry)'. This analysis 
of 'grass' is helped by denominative Ath. verbs, e.g. Hupa t'h-n, -t'lo, -t'lo-W, 
-t'lo-i 'to make baskets, to twine in basket-making'; Nav. -t'lo, -t'lg-, -t'lo-l 'to 
tie (e.g. the hair)'. 

We may summarize all this as follows: 

Ath. *lu, -lu 'coil, loop' 

— *lu 'to be caught in a noose' 

— *'lu 'to catch in a noose' 

— *'lui 'to wrap around' 

r-7w' 'what is twined' > 'grass' (denominative vb.: 'to twine in basketry; to 

braid hair') 

*t-'lo--l 'what is always looped' > 'rope, strap' 

Now comes the fun. Indo-Chinese lu as follows: 

Tibetan lu-i] 'a strap, slung over the shoulder or round the waist, for carrying 

Miao /o-r) 'bridle' (close o) 

Tib. metaphorically: caus. s-lu 'to cause to be snared', i.e. 'to ensnare, 
beguile, seduce' 

Angami Naga te-rhu 'sly' {rhu < h-ru; Tibeto-Burman hi-, hr- > Ih-, rh- is 
exceedingly common; h- is common as causative prefix, e.g. Tib. s-l- often par- 
allel to Ih-) 

T'ai group: Siamese roi^^^ 'enfiler' (numbers indicate tones in H. Maspero's 
orthography); White Tai roi, loi 'enfiler' 

And now Chinese: ///- group: lui]^-^ (numbers for tones according to usual 
Modern Pekinese system) 'a cage, to snare' < Old Chinese (i.e. 7th Cent, 
forms, which I have worked out carefully from Karlgren's tables) lu[o]x]- (- = 
level; / = rising; \ = falling tone). luv]<-^ (in hi^^^> t'ou(-> 'halter', i.e. 'snare- 
head') < Old Chinese /w/o/r)-;/w^-^^ 'girdle gem' <luo\ (words with falling tone 
seem frequently to be old passive derivatives: 'what is looped around one's 
waist'?); /m^^'' 'thatched hovel' < luo-; lu<^> 'hempen thread' < luo-; lo^^^ 'conch, 

Six: Athabaskan and Na-Dene Languages 137 

spiral, screw' < lud- {a is a dark-timbred a- vowel); lo^-^ 'lines in the palm' < lud 
?lei'^^ 'to creep, cling to', same character also read lei^^^ 'series, connected' 
(words evidently refer primarily to creeping vines) < ludi-, ludil. These Chi- 
nese lu, lud forms are paralleled by another set in lui- (cf. Ath. *-lui above): 
lei^^^ 'to bind' < lywie- (< *lwie-);lei''''> 'acreeper (asof melonor pea)' (not sure 
of reconstruction, but phonetic element in character suggests initial lu or lywi); 
lii^^^^ 'silken thread, a hank' < lyhi/. And metaphorically we have luj}^'^^ 'foolish, 
to impose on' < lu[o]r\\; lii^'^^ 'deceitful' < />'h'/o\ (i.e. 'winding about, ensnar- 
ing with blandishments'; cf. Tibetan s-lu above). 

Other probably connected Tibetan words are: Ihub (i.e. causative h-lub) 'to 
bind, tie, fasten (e.g. ornaments to the ear)'; k-lub 'to cover (e.g. the body with 
ornaments)'; Iwa-ba 'a woolen blanket'. 

Observe how well the Ath. forms integrate with the numerous Tib. and Chi- 
nese forms. But we are not done. Very likely connected with Ath. *-lu is Ath. 
*-lu-s 'to drag an animal by a rope': Nav. -Ids, -lo-z, -Ids classifier verb "denot- 
ing a single animal as an object: the inference is that the animal is led by a rope" 
(Franciscan Fathers); Hupa -los 'to drag, to pull along'. With these forms I feel 
inclined to compare (though here I feel far more hesitant) Chinese lo^^> tsz<^^ 
'mule' {tsz<-^^ is merely 'son', often used to make nouns) < lud-; also lu 'donkey' 
< lywio-. The parallelism between Ath. and Chinese would be a convergence 
from related radicals rather than a specific etymological parallel. 

But we are far from finished. Perhaps related to Ath. *-lu is an important 
classifier verb *-/e, *-la\ Nav. -/e, -la, -let 'to handle a long, flexible object, as a 
rope, quirt, leather, hide, etc' (Franciscan Fathers). Parallel to this is a set of 
Indo-Chinese forms in *la, *le (Ath. e is often parallel to a; just how related I do 
not yet understand, possibly reflex of old alternation a: [a): 

Tib. causative s-le, Ihe (< h-le) 'to twist, plait, braid the hair, to make a 
basket, to knit'; s-le 'a coarse basket'; s-le-po, s-le-ba, s-le-bo 'a flat basket' 
{-po, -ba, -bo are "articles"); Iha-s, Ihe-s < h-la-s, h-le-s 'braid, wicker-work, 
texture; twisted cake or bun'; Ihe-s-ma < h-le-s- 'the act of twisting, plaiting' 
(-ma is "article"); /a«-/?w 'braid, plait, tress of hair' (-/?« is diminutive); lan-ts'ar 
'ornaments worn in the hair'; le-brgan 'diapered design of woven fabrics'; le-na 
'the soft downy wool of goats below the long hair; fine woolen-cloth'; Ida-ldi 'a 
kind of ornament of silk or cotton, a fringe or tassel' < d-la- {dl- regularly > Id- 
in Tib.; very easy to illustrate); Idan-mgo 'the yarn-beam of a loom' < d-lah- 
{mgo 'head'); Idem-Idem < d-lem 'flexible, supple, elastic, pliant'; Ideb- < d- 
leb- 'to bend round or back, to turn round, to double down'. 1 am not so certain 
of this last, which brings us into a large set of forms in la- and lo- referring to 
'turning, turning back', which may well be related to our present set but which I 
prefer, for brevity's sake, not to go into just now. 

Now Chinese. We have two series: *la and *//. Based on *la are: lan'-> 'basket 
with handle' < Idm-; lan^^^^ 'rope, hawser' < Idml; lao'-^> (tezi) 'netted case' < 
Idk, same character also read h^-^^ 'joined, to tic up'; lei^'K also read h<'---'^> 'to 
rein in, to strangle' < hk (perhapsbcttcr to /fv- scries above); /o'^'^ 'net, sieve' < 
ld-\ lo<^) 'shallow open basket' < Id-. Based on *// are: li<~> 'ornamented girdle' 

138 ^I American Indian Languages 2 

< bie- (?); W^^ 'basket' < />'/-; //^- "^^ 'rope to tie a boat', character also readi/'^-^'' 
'a well woven gauze' < xyil (?), which probably means older *h-li (loss of post- 
consonantal / is now well established for Chinese). 

Is it not impressive that Ath. and Indo-Chinese *lu seem to have reference 
chiefly to "looping," Ath. */e, la and Indo-Chinese *la, li to "handling a long 
flexible object, twining, basketry"? Now let us return to Nadene. In Tlingit we 
have //r' 'fine basket'. Remember that Tlingit has no voiced /, only voiceless /. 
Further, experience shows abundantly that many Tlingit forms in final conso- 
nant (particularly if glottalized) are compounds. Hence we may suspect that lit' 
really means old *li-f(a). Reference to Ath. gives us *-t'a 'receptacle': Hupa -fa 
'sack'; Kato -fa 'pocket, blanket fold'; Nav. -dzis-fa 'pocket' {-dzis 'hollow, 
semi-tubular' ?). This somewhat unsatisfactory parallel is buttressed by Haida 
fao-fa 'box' (almost certainly misheard for tao-fa 'food-receptacle'). Tlingit ///' 
is therefore probably li-fa 'receptacle (for small objects) of twined basketry'. 
We have further /'/ 'woolen blanket' < *// or *//' (final glottal stops seem regu- 
larly to affect the first consonant; I have good evidence for this). Based on *la or 
*lia we have also in Tlingit fleq' 'tentacles of squid' *t-'le-q' (with medio-pas- 
sive r-; Tlingit e is always related to a- forms) < 'what is long and twists itself? 
(cf. fundamental meaning of Ath. *-/e, *-la above). 

And in Haida we have dli-n 'arm of devil-fish', which I analyze as d-li--n {d- 
medio-passive; -n perhaps causative, to which there are good Ath. and Tibetan 
analogies; -//•- < Im- ?) < 'long things that twist themselves about'? 

I am quite likely to have misinterpreted here and there, but the Nadene and 
Indo-Chinese parallels seem highly suggestive to me. Of course, this sort of 
group-parallelism is not isolated. 

I shall refer to another interesting group, without going into details. I spoke 
of Ath. *-/e 'to revolve' and its derivative *flel 'fire-drill'. To these forms belong 
Haida hi 'to surround, move around'. In Indo-Chinese we have a well-cham- 
bered Tibetan set: re and ril (e.g. res 'change, turn, time, times'; i}-g-re 'to roll 
one's self ; causative s-g-re 'to roll' ; ril 'round' ; causative s-g-ril 'to wind or wrap 
round; to roll, wrap, or wind up; to wag (the tail); to roll (a stone)'. Here may 
also belong Haka kut-hrer 'ring' {kut- 'hand'; h- common Tibeto-Burman 
causative prefix), though Conradi sets this to Tib. forms in la-, le- referring to 

Editorial Note 

Originally published as part of letter 332 (October 1, 1921) in Victor Golla 
(ed.). The Sapir-Kroeber Correspondence. (Survey of California and Other 
Indian Languages, Report 6.) Berkeley: Department of Linguistics, University 
of Cahfornia (1984). 

Six: Athabaskan and Na-Dene Languages 139 

Sapir's manuscript Sino-Dene "dictionary," including more than a hundred 
lexical comparisons, is in the Library of the American Philosophical Society 
(manuscript 497.3 B63c Na20a.3, vol. 2; cf. Kendall 1982: 28). Nearly all the 
entries date from the early 1920s, and it is clear that Sapir went little further 
with the idea than the speculations contained in this letter. Of these specula- 
tions (see also 1925o, later in this volume) Krauss has written: "Sapir was . . . 
carried far beyond any objectively justifiable conclusions by his enthusiasm for 
the idea" (Krauss 1973: 963). Few would challenge the accuracy of this assess- 
ment. The connection is, however, a plausible one, both on linguistic and 
anthropological grounds, and it continues to attract attention. Robert Shafer, a 
Sino-Tibetanist, published two papers in support of the Sino-Dene hypothesis 
(Shafer 1952, 1957), but, being based neither on Sapir s Sino-Dene files nor on 
extensive comparative Athabaskan or Na-Dene data, his work cannot be con- 
sidered an authoritative statement. Greenberg, Turner and Zegura (1986) have 
proposed that Na-Dene (including Haida) represents the language of an Asi- 
atic population that entered the New World about 7,000 years ago, and that it is 
distinct, both physically and linguistically, from all other American Indian pop- 
ulations. While this is consistent with Sapir's Sino-Dene hypothesis, Green- 
berg, in his recent world-wide linguistic classification, does not link Na-Dene 
with Sino-Tibetan or with any other linguistic group, considering it to be one of 
15 fundamental linguistic families in the world (1987: 332-337). Pinnow (1990), 
relying mainly on Greenberg 's own evidence, believes that a good case can be 
made for viewing Na-Dene as intermediate between Old World languages 
(especially Sino-Tibetan) and New World languages (especially Greenbergs 
"Almosan-Keresiouan" phylum). 

Athabaskan Tone 

A part of the summer of 1922 was spent by the writer at Sarcee Reserve, 
Alberta, in studying the language of the Sarcee Indians. A series of texts was 
obtained as well as supplementary grammatical material. The most important 
single point that appeared was the fact that Sarcee has a well-developed system 
of pitch accent. Fundamentally this system has a striking resemblance to the 
Tlingit tonal system described by Dr. Boas, though secondary developments 
have complicated the Sarcee system considerably. The tonal resemblances 
between Tlingit and Athabaskan constitute an important further argument in 
the Nadene theory recently put forward. Hints on Athabaskan tone are to be 
found also in Father L. Legoffs study of Chipewyan (Grammaire de la Langue 
Montagnaise). Father Morice makes a few isolated references to tone in Car- 
rier, where it is almost certainly a feature of importance judging from brief MS 
linguistic notes taken by C. M. Barbeau among Carrier Indians at Hagwelgate. 
Some years ago P. E. Goddard noted tonal differences between otherwise iden- 
tical second person singular and third person subjective forms in the Hupa 
verb; these observations, based on the study of [391] Rousselotgraphs, agree 
with the Sarcee results obtained. So fundamental is tone to Sarcee morphology 
that it is well nigh inconceivable that it should be entirely absent in any other 
Athabaskan dialect. 

Editorial Note 

Originally published in American Anthropologist 24, 390-391 (1922). 
Reprinted by permission of the American Anthropological Association. 

Sapir's references to earlier work are to Boas (1917) for Tlingit, to Legoff 
(1889) for Chipewyan, and to Goddard (1907) for Hupa "Rousselotgraphs." 
Morice later commented on Carrier tone more extensively in his full treatment 
of the language (1932). 

At the time of Sapir s Sarcee work, tone systems had been described for many 
African and Asian languages but only rarely noted among American Indian 
languages (Sapir's own description of a pitch accent in Takelma being one of 
these instances; see 1912h, Volume VIII). We now know tone to be a wide- 
spread feature in the Americas, particularly in eastern North America and in 
lowland South America. Ironically, comparative evidence in Athabaskan (and 
Athabaskan-Eyak) now indicates that tone was not present in the proto-lan- 
guage but developed in several Athabaskan subgroups (but by no means in all) 
as syllable-final consonants were simplified or lost (Krauss and Golla 1981: 

142 Vf American Indian Languages 2 

69-70). Sapir, in later fieldwork, encountered at least two Athabaskan lan- 
guages without tonal systems, Anvik and Hupa. In the latter case, where Sapir 
had gone to the field with every expectation of confirming Goddard s earlier 
reports, he was clearly nonplused (see Sapir 1928i, reprinted in this volume, 
and Sapir s letter to Kroeber, 28 June 1927, quoted in Krauss 1986: 163). 


As is well known by students of Athabaskan 
linguistics, the Athabaskan adjective is in form 
a verb. Even the simplest, non-pronominal or 
third-personal, form regularly contains either 
a " first modal " prefix - (generally de- ^, d; 
ne-, n-; t'e-, t'- ; or ie-, i-), a " second modal " 
prefix {se-, s-), a " third modal " prefix (gene- 
rally -/- or -i-), or a first (or second) modal 
prefix followed by a third modal prefix (forms 
in de-l-, ne-l-, nl-i- are typical). It is unneces- 
sary to give examples here, as they may be 
readily supplied from the special dialectic 
monographs (see, e.g., Goddard, Legoff, Mo- 
rice, Petitot), 

Father Legoff, however, in his Grammaire de 
la Langue Montagnaise \ calls attention to an 
interesting class of adjectival derivatives 
(" nouns formed from'adjectives ") in which 
the prefixed elements are lost and the bare stem 
appears in the relative ^ form. Some of his 
examples are : 

1. I follow Dr. Goddard's convenient terminology, 
without thereby committing myself in the least as to the 
term ""modal ". 

2. I use e as a formula for the reduced or " pepet " 
vowel which has different forms according to dialect or 
according to varying phonetic circumstances in one 
dialect (e.g. i, less frequently a, e, a. or u, in Hupa ; a 
in Kato and Chasta Costa ; e, a in Chipewyan). It may 
often be shown to be a reduced form of an older Atha- 
baskan e or /. 

5. Montreal, 1889. See p. jo. Petitot and Legoff use 
*' Montagnais " in the sense of Chipewyan. 

4. By " relative '" I mean the form assumed by nouns 
when they are qualified by preceding elements (nominal 
or pronominal) and by verbs when they are used as rela- 
tive clauses. The fundamental Athabaskan relative suffix 
is probably -e, -e (-ye, -y^), but the actual dialectic 
forms are often involved by the operation of various 

di-l-ba (to be) gray : bay-z the gray one 
dz-l-gai (to be) white : gay-z the white one 
m(d)-du-i (to be) short : du-z the short one 
dz-yzl (to be) squat : yzl -s the squat one 
dz-bal (to be) round : bao the round one 
dz-l-zzn (to be) black : :^zn the black one 
dz-yo' (to be) shaggy : yo' the shaggy one 
dz-l-Bo' (to be) yellow : Oo' the yellow one. 

Such forms as bg^, ;(en, yo' , and 60' look for 
all the world like unmodified stems, but there 
is every reason to believe that they are relative 
forms, like gay-z and yel-z, that have either 
fallen together with or that differ in some res- 
pect from the parallel stem forms found in the 
adjective-verbs. Possibly the phonetic record 
is defective ^. 

Legoff says of these forms : " This kind of 

phonetic laws. Frequently the e, e is dropped but is 
then apt to leave a trace in the voicing of the prece- 
ding consonant, now final (e.g. the relative form of 
Hupa -fa AMONG < Ath. *-t'ax is -t' au < -t' av.\ the 
regular Hupa development of Ath. *-t'a-{, reduced from 

5. I write ^a^-e rather thad ba-y^, though Legoff speaks 
of a suffixed -ye, because comparative evidence demons- 
trates the existence of the stem-form -bay- a.s well as -ba' . 
-ba'-i and Petitot's Montagnais -ba-a seem to presuppose 
Ath. *-mah-e. Navaho -bat, e.g. gim i-bai gray hawk, 
is Ath. *-may-i; d. further Hupa -nmi, Kato -bai. I am 
modifying Legoff's and Petitot's orthography and that 
of the Franciscan Fathers so as to bring them into har- 
mony with current Americanist usage. 

6. Goddard writes di-l-{7.ti to be black, but tia^iiz- 
^an' BLACK K0.\ (Texts and Analysis of Cold Lake Dialed, 
Chipewyan, apamnh, X, 1 10). Does this mean that -iin 
is the relative form of -inn ? The relative form of the 
corresponding Navaho -/t« black is -/««, with lengthened 
vowel. The formation of the relative in Athabaskan has 
never been properly studied. It is one of the fundamental 
problems of the language. 


VI American Indian Languages 2 


substantive is hardly used except as sobriquets 
which men give one another or as names which 
people give to animals, in order to distinguish 
them. And then they are always followed by 
the word ya:(^t, which means little". As a 
matter of fact, the type is illustrated in other 
connections, e g. : 

d--k\oTO BE BALD I t^i-k'zit THE BALD-HEADED 

(h)o-rt-iur '< to be slippery (Pet.) : t' an'-xurz 

SLIPPERY ice (Leg.) 

These two examples differ in an important 
respect. In t'xn'-:(urs and numerous other com- 
pounds of its type the first member (ice) desig- 
nates the properly denominating concept of 
the group, which is then qualified by a relative 
form (jur-i) of Legofi's " sobriquet " type. In 
t^i-k\lt the first member of the compound 
(head) is not the properly denominating con- 
cept of the whole, though it is itself qualified 
by a relative form (^k\l-t), again of the sobri- 
quet type. The group tHi-k\l-., taken as a unit, 
is to be understood as qualifying a third, un- 
derstood, noun. We can express tliis by saying 
that while slippery is relative to ice and ice 
not relative to another concept, bald is relative 
to head and bald-head to person. As far as 
such a form as tOi-k'eoi is concerned, it makes 
no difference whether the qualified noun is 
expressed or not. Obviously the difference bet- 
ween t'oin'-iure. and thi-k'zle is analogous to 
the English difference between red breast and 
(robin) redbreast. Whether there is a pro- 
sodic difference (one of stressor pitch) between 
the two Athabaskan types does not appear 
from the evidence, but it is possible that they 
are not formally identical. 

As there is no genuine line of demarcation 
in Athabaskan between " adjective " and 
" verb ", one may expect that forms both of 

7. Petitot writes 0- for /;o- (see Goddard and Legon). 
-r- is a postvocalic form ol d- in Chipewyaii. 

the type bays, the gray one and t'ixn'-:(tm 
slippery ice may be based on " verb stems ". 
This is exactly what we find. A few examples 
are : 
Mont. -Y? TO MELT ; ds.-l--(j to be melted 

(Pet.) : dlts '(in-z grease easy to melt, 

meltable grease (Leg.) 
Mont, -ti'a^ to cry : ts'ay-e. the whimpering 

one, grumbler (Pet.) 
Hare -k'^i to lie habitually : k'^t-t liar (Pet.) 

It scarcely needs to be pointed out that 
LegofT's " sobriquets " are simply qualifying 
terms in the relative form, the noun referred' 
to being unexpressed. It is not a far cry from 
compounds like t^i-k'tlt the bald-headed one 
and dUs -^in-e. meltable grease to such clipped 
forms, say, as k'toe. the bald one and -(ins. 
WHAT melts (easily). Such forms are in type 
identical with LegofFs hayt the gray one. Ot 
ts'a-^t grumbler, Petitot remarks that it is deri- 
ved from ts'ay tears; we would then have to 
interpret ts'ayz as the one with tears rather 
than as the whimpering one. Petitot's analysis 
is perfectly credible, for forms of the type of 
bayz may be directly formed from noun stems, 
as I shall show from Navaho evidence. But I 
hope also to show that the difference between 
-ts'a-( TO CRY and ts'a-; tears is purely a matter 
of translation, not of intrinsic Athabaskan 

In Navaho there are a great many prefixless 
or radical qualifying terms in relative form. 
They are freely used with or without preceding 
nouns and are based on " adjectival ", " nomi- 
nal ", or " verbal " stems. They all denote 
permanent or characteristic attributes and easily 
take on the character of sobriquets or regular 
clan or personal names. Examples of " adjec- 
tival " origin are ^ : 

8. My Navaho examples are taken from tiie Vocabuhry 
of the Navaho Language of the Franciscan Fathers (2 vols., 
St. Michaels, Arizona, 1912). 

Six: Athabaskan and Na-Dene Languages 



n-jun-i NICE (relative form) ; «i-/p" to be gocd : 

jun-i NICE 
n-nt'i TO BE LONG : nts, nf^-i long 

n-t't'lro BE WIDE : t'd WIDE 

h-ts'o to BE YELLOW : h ts'o-i YELLOW EARTH 

(place name) 
ii-jin TO BE BLACK : hwo' ji'n-i tooth-black, 

DECAYED TOOTH J ya'' ji''n black LOUSE 

a-i-i'sg's-i to be slim (also -t'sgs, -t'sos-i, 
-t'so's) : ts'i' t'so's-i head-slim, the slim- 
haired ONE (man's name) 

hwo' c-gi'j-i TOOTH wHicH-is-MissiNG (c- assimi- 
lated from S-, " second modal " prefix) : hwo' 
gi'j-i MISSING TOOTH ; bo-hwo'' gi'j his-tooth 


di-tc'oc-i stubby (relative form) : ts'i' tc'o'c-i 
head-stubby, the stubby-haired ONE (man's 

a-gud-i SHORT ; Montagnard i-gor-z (Pet.) : ga 

gu'd-i ARM-SHORT, ARMLESS ; k\ gud-i SHORT- 

ia-gai to be white : ga' ts'o gai' rabbit-large- 


Examples of *' verbal" origin are : 
-t'lo, -t'ig', -t'ioi to weave : Hog-i 9 grass- 
weavers, SIA INDIANS 
-y^'d, hwud, -hwul to run : t'o' hwul rapid 
WATER (place name) 

This type is doubtless actually well repre- 
sented in Navaho, but the material is scanty or 
not easily accessible. To it belongs probably 
Tucson (Arizona place-name), said to mean 
BAD-SMELLING WATER ; the Navaho (or Apache) 
form is probably something like to tc'a (cf. 
Nav. -tc'iUy -tc'a, tc'ji to smell). 

Examples in which the related word is a 
noun are quite numerous. The reference is not 
to the concrete content of the noun as such 
but to a person or object, expressed or implied, 
that IS conceived as the possessor of or as in 

9. Or directly to t'io' grass? 

some way related to the thing defined by the 

noun. Examples are : 

k'ai WILLOW : k'a''-i those who have (or are 


yo' BEAD : yo-0 those who have beads, bead 


fca' HAT : t'cah-i he with the hat (man's 

gic CANE : gic-i HE WITH THE CANE (man's name) 
ca'c KNOT : ca'j-i knotty 
xvi burden, bi-yf'l his burden : ye'l-i the one 

BACKED (man's nick-name) 

fcoc WORM : hwo' fco'c-i tooth -wormy, a hol- 
low TOOTH 


(place name) 
t'o WATER : na' t'o'-ho enemy that is connected 


cac BEAR : na' cdc-i ennemy that is bear, bear 


Relative forms like t'cah.-i bear an obvious 
formal similarity to such English derivatives 
in -ed and -y as bearded and knotty. But the 
resemblance is more apparent than real. The 
relative form of the noun is not a true " adjec- 
tival " derivative of the noun, as is shown by 
the fact that morphologically parallel forms are 
built on stems conventionally set down as 
" adjectival " or " verbal ". The genetic 
relation between these Athabaskan relative 
forms and the possessed form of the noun (e. 
g. Montagnais t'sah-i his cap [Leg.], si-t'sa-a 
MY CAP [Pet.]; Hupa hWm song : xo-hW\.n-[n^'. 
HER song ; Chasta Costa c-man-e my house) on 
the one hand and the subordinated form of the 
verb (e.g. Mont, i^-yt one grows up : {s--ye- 
h= GROWTH, m-^^-i-l-dyd we are afraid : ns—,"- 
t-l-dyed-i we who are afraid [Leg.] ; Hupa 
-lal TO FLOAT continuously : na-na-t-lnl-B it 
floating; Nav. di-c-Jnuuc i yell : d\.-la-hwuc-i 


VI American Indian Languages 2 


ONE WHO YELLS MUCH, howler) OH the Other 
is obvious. In Navaho the relative form with 
final vowel (generally -/ ; old -e assimilated to 
-cJ after radical o)is probably no longer felt as 
identical with the possessed nominal form with 
final consonant (cf. bi-ye-'l his burden with 
yz'l-i above), but the general consensus of 
Athabaskan evidence makes it higly probable 
that such alternations as -yi'l and yd-i go back 
to Athabaskan alternations of type *^z'l-e: 
*'(el-e. In other words, even in Navaho the 
possessed forms of the nouns are simply redu- 
ced relatives (perfectly analogous to xe'i : -yfl 
is d:(d MOUNTAIN : yo-l-gai' d^i'l shkll-white 
mountain). We have already seen that there 
are analogous doublets in Navaho for the rela- 
tive forms of " adjectival " stems (cf. gi'j-i 
and gi'j .missing, jm-i and ji'n black above). 

It is not the purpose of this paper to discuss 
the functions and the fundamental significance 
of the Athabaskan relative. I hope to show in a 
future paper that it is a feature that goes back, 
both in general form and as an actual phonetic 
element, to the Nadene period and that it con- 
sists in essence of an old particle, probably a 
demonstrative stem, that could be freely added 
to any word or group of words to relate it to 
an expressed or understood person or thing. 
The primary function of the Athabaskan (and 
Nadene) relative is thus an exceedingly wide 
one, of which the particular usages listed in our 
grammars are but specific applications or rather 
English (or French) translations. The lengthen- 
ing of the stem vowel and the voicing of a 
final voiceless spirant '° are merely secondary 
phonetic phenomena due to the presence of 

10. In part no doubt retention of originally final voiced 
spirant. 1 believe it to be probable that in such Athabas- 
kan alternations as '/Vo/ strap : *tlol-e, *t'iol-e STRAP 
OF (one) it is the -/- which represents the old conso- 
nant, preserved because of the following relative element, 
and that the -i of the absolute form is du^ to a secondary 
unvoicing of'the old -/. 

the relative element. As this element became 
reduced to zero, these secondary phenomena 
tended to take over the properly relative func- 

Just as we have the alternation of final -i : 
-/-, of -X (-') : -Y-, of -X (-') : -y-, of -c : -/-, 
and of -s : -;(-, so also these alternations occur 
initially " ; e. g. Mont, iue fish : se-llut my 
FISH (Pet.), Hare xi burden : se-yzl-t my bur- 
den, Hupa hi smoke : m'.[l]-h[t]d-t his smoke, 
Nav. Sin song : bi-yi'n his song (in Athabaskan 
terms *xen.: *-yen-i ; cf. Mont, cen : -yzn-Zy 
Hare ci : -yin-e). Here too the alternation 
could only have been due to phonetic circum- 
stances to begin with. If a word was closely 
connected, in thought and in position, with a 
preceding word or element, the voiced spirant 
(say /) was retained or the voiceless spirant was 
voiced. Athabaskan *xei burden : *cB-yel-e my 
burden thus originally meant no more than 
that two radical elements {*ce me and *xel bur- 
den) were united into a phrase with the help 
of a following denominating element *-e, *-e : 
me-burden the, i.e. my burden. The *-e 
preserved the -/ of *xel, ordinarily *xei, 
while the voiceless x passed to -y- in intervo- 
calic position '^ It is very doubtful if there 
was any specific function connected with the 
X- : -y- interchange. In time, however, there 

11. In final position note also -r; (or nazalization) : 

12. Possibly an old y- was here preserved, but passed 
to unvoiced x- when not protected by an immediately 
preceding vowel. The alternations listed in the text na- 
turally apply to primary Athabaskan. In certain dialects 
some of these alternations ceased to operate freely be- 
cause phonetic laws divorced the consonants that had 
originally belonged together. Thus, in Hupa the old 
Athabaskan interchange of x- : -y- (preserved in Chipe- 
wyan, Hare, and Navaho) had ceased to be a live pro- 
cess because of the falling together of Athabaskan .v and 
k' into Hupa x and the change of Athabaskan y to Hupa 
w ; the corresponding final alternation of -x : -y lingered 
on as -' : -w, -u. Note also that in Pacific Athabaskan 
voiced sibilant spirants have been leveled with voiceless 
sibilant spirants. 

Six: Athabaskan and Na-Dene Languages 



can be no doubt that the voicing of initial spi- 
rants came to be felt as intrinsically, not me- 
rely mechanically, connected with the relative 
function. Hence such detached forms as "^^eUe 

THE ONE WITH A BURDEN (Nav. yel-i '^), "jUH-e 

THE GOOD ONE (Nav. jun-i), *jen-e the black 
ONE (Mont. :(enn). It is probably because of 
the intrinsically " relative " significance of 
" adjective " stems that these regularly begin 
with a voiced spirant if the initial consonant is 
a spirant. 

Such forms as Hare ^'^'e-e liar and Navaho 
t'^'I WIDE bring home to us the highly impor- 
tant fact that the actual " radical elements " 
of Athabaskan verb and adjective forms are 
more freely isolated than one might at first be- 
lieve to be possible. A careful study of all the 
available material would tend to show that 
these radical elements have a considerable mo- 
bility, that they are not far removed from the 
status of independent monosyllabic " words ", 
and that the complex " word " of our Atha- 
baskan texts and paradigms feels a great deal 
more like a closely knit phrase or sentence 
than has yet been suspected or, at any rate, 
explicitly demonstrated. I hope to show in 
due time what is the true nature of the various 
" prefixes " and " suffixes " that render the 
morphology of the Athabaskan verb so complex 
in appearance. It will appear that each and eve- 
ry one of these elements is a relatively self- 
contained unit in the sentence, either a deter- 
minative or an actually predicating element. 
Much of the " vagueness " of meaning or 
function that we feel to attach to many of these 
elements is simply an index of our inability to 
carry over the Athabaskan manner of expres- 

13. Athabaskan y is preserved in Navaho only before 
a. It appears as y before i and e, labialized to w before 
0. The alternation x : y, however, leads one to suspect 
that this " y " is not phonetically identical with the y 
that corresponds to common Athabaskan y. 

sion into precisely equivalent English (or 
French) form'"*. 

For the present I shall content myself with a 
few random examples suggesting the mobility 
and essential concreteness "> of the verb stem. 
The Hupa verb stem for '' to flow " is -Un, 
-hr„ which has numerous cognates in other 
dialects. Forms like t'cewishnt't it will flow 
OUT and nzdhn it always flows do not seem 
to suggest the possibility of combining the bare 
stem freely with other concrete elements. I 
have found no examples in Hupa of -/sn or -hr, 
so used, though they may 'of course exist. But 
this stem (Athabaskan *-lin, *-leri) is clearly 
related to Athabaskan *-// (e.g. 'Mont, -^a-i-d-li, 
Hare yi-d-li couler a terre [Pet.]), probably 
also to Athabaskan *-lej (tears) flow (e. g. 
Mont, dc-l-hz, Hare dt-l-W [Pet.]). Athabas- 
kan *-li appears in Hupa as -k (reduced to -/) ; 
it is found, without formal prefixes of any 
kind, in certain compounds : no-U dam, wa- 
terfall (lit. down-flow or halt-flow), iz-l- 
d'.r, together-flow-place (village name). Si- 
milarly, -xa{u)W, an " indefinite "form of-.vd 
liquid has position, is directly compounded 

14. Very much as one who tries to see in a French lo- 
cution like Qu'est-ce qu'il a ? an exact equivalent of our 
English What has he ? would find it difficult to get the 
proper form-feeling of the elements est, ce and qu\ 

15. In a large number of Athabaskan verb forms it is 
impossible to assign to the radical element (that is, the 
" verb stem " proper) the kind of concreteness that would 
in our eyes yield the actual concrete significance of the 
form. It does not follow that the stem has not a defini- 
tely concrete significance of its own, clearly apprehen- 
ded by the native form-feeling.. All that we have a right 
to say is that the summing of significances inherent in 
the theoretically independent elements of the complex 
does not seem naturally to lead to the precise idea that 
we express in such and such a way. The true difficulty 
from our naive standpoint is in such cases not that the 
stem is a « vague » element but that it seems inapplic- 
able, just as to one unacquainted with English idiom 
the notion of " fall " in such a sentence as The friends 
had a falling out is inapplicable. A a matter of fact, I be- 
lieve that there are few languages in America that feel 
their " stems ",.and elements generally, as definitely and 
as concretely as Athabaskan, Tlingit, and Haida. 


VI American Indian Languages 2 


with sa MOUTH : sa-xa(u) H^ liqjjid which has 


vaho, again, the verb stem -ua to live (e. g. 
X'.-n-c-na 1 live) may be used as an unmodi- 
fied element in a compound ; k'a'-na''-ni 


name). Such examples could be multiplied con- 

If the monosyllabic *' verb stem " may thus 
be isolated in practice as a more or less freely 
movable element, capable of conveying a definite 
notion in its own right, we cannot but con- 
clude that the purely formal difference between 
verb (and adjective) stem and noun stem be- 
comes a tenuous one. What is to prevent us 
from interpreting the -h of Hupa no-lz as a 
noun meaning flowing or current, no-h and 
it-l-d\.rj meaning properly down-flowing (not a 
secondarily aominalized form of an inherently 
verbal to down-flow) and reciprocal-cur- 
rent-locality ? Might not the Navaho k'a'- 
na'-ni be just as well interpreted arrow-life- 
people ? As a matter of hex., I cannot see that 
anything seriously stands in the way of such 
an explanation, and its adoption would at once 
make clearer a number of morphological pecu- 
liarities. Among such peculiarities are : 1. the 
ease with which a great many evident nouns 
are transformed into " verb stems " (e. g. Hu- 
pa ^'fl DRESS : -k'a TO wear a dress ; Kato djit] 
DAY : -djvr^ TO BE day) ; 2. the frequency with 
which " verb stems " with aclearly defined verbal 
lorce, if we may trust all appearances, take on, 
when isolated, an abstract or concrete nominal 
significance (e. g. Athabaskan *-yan to pass 
THROUGH life : *xan old age, maturity ; *-dlo' 
TO LAUGH : *dlo' laughter ; *-lo to snare, to 


3. the fact that a number of verb stems refer 
not to specific activity but to a class of objects 

1 6. Voiceless spirants initially, voiced spirants in post- 
vocalic, or originally postvocalic, position, according to 
the typical Athabaskan rule. 

(e. g. *-'a« TO HANDLE A ROUND OBJECT ; *-^'oj 

TO HANDLE A CLOTH-LIKE object). I hope later 
to take up this fundamental question and to 
show that in a verb form it is not the " verb 
stem " that is the distinctively verbal element 
but, where found, the" third modal " element; 
that all " verb stems " are in fact nouns not 
only in theoretical origin but in actual usage ; 
and that verbs translated according to the forms 
TO DIE, TO BE SEEN, and TO KILL fall iuto patterns 
more accurately rendered by death is, sight 


interpretation is correct, an element like -an is 
not properly a " verb stem " indicating some 
kind of activity or state with reference to a 
single round object but is actually a noun which 
means, or originally meant, a round object. The 
three classes of verbal usage listed above would 
fall into a single category applicable to all other 
verb forms as well. Indeed, it will appear that 
this theory of the, essentially nominal character 
of all " adjective " and " verb " stems simpli- 
fies enormously the whole aspect of Athabas- 
kan (and Tlingit) morphology. 

Meanwhile, whether or not we are willing to 
go so far in the present stage of our knowledge 
as to accept the nominal theory of verb radicals, 
it is clear enough that the Athabaskan relative 
forms discussed in this paper belong together. If 
a Navaho form like Ccah-i is to be interpreted 


we may venture to interpret a verbal deriva- 
tive like Hare k''^t-z as the one having false- 
hood, an adjectival derivative like Navaho 
jun-i as the one having goodness. There is 
certainly no serious point of morphology that 
would make such an interpretation impossible. 
As it is, it is sometimes an arbitrary matter 
whether we assign a given relative form to an 
adjectival or to a nominal source. In Navaho 
ga k\s-i one-armed, armless, k\s-i short, 
crippled, DEPRIVED OF may be looked upon as 
an adjectival (or verbal) formation (cf. gnd-i 

Six: Athabaskan and Na-Dene Languages 



above). Identical, however, with Navaho k\s, 
which I have not found in its bare form, is 
Anvik Ten'a k'xf) piece (oF)(Athabaslian *k'es). 
Clearly it makes little or no difference, from 
the Athabaskan standpoint, whether we ana- 
lyze ga k'ls-i verbally as the one whose arm is 
CUT OFF(cf. Navaho verb -k'z, -ki'\ -k\ to cut 
WITH A knife), adjectivally as the one who is 
SHORT OF AN ARM, or nominally as the one 


the same way, it seems an indifferent matter 
whether we interpret Montagnais ts'ay-t ver- 

nominally as the one who (always) has tears, 
WEEPING ; or Navaho yil-i (a sobriquet for a 
hunchback) verbally as the one who is wont 
TO carry (a burden) (c(. Navaho -xe or -xd, 
-yr, -xi-i, TO carry on one's back, to handle 
A burden), or nominally as the one who (al- 
ways) HAS A BURDEN. 

Editorial Note 

Originally published in International Journal of American Linguistics 2, 



I. Consonants. 

The Consonant System. 

The Intermediates. 

The Unaspirated Hard Stops. 

The Aspirated Surds. 

The Glottalized Stops and AfFricatives. 

The Voiceless Spirants. 

The Nasals and Voiced Spirants. 

The Glottalized Nasals and Voiced Spirants. 

The Laryngeal Consonants. 

Secondary Consonantal Processes. 

Initial Consonant Clusters. 
II. Tht' Syllable. 

Syllables with 1-Vowel. 

Syllabic Nasals. 
III. Vowels. 

Qualitative Changes. 

Vocalic Quantity. 

W. Stress and Pitch. 


The following notes on the sounds of tlie 
Skidegate dialect of Haida are based on mate- which I was fortunate enough to secure 
iVom Peter R. Kelly, a well educated Haida 
Indian who is at present engaged in missionary 
work among the Indians at Nanaimo, Vancou- 
ver Island. Mr. Kelly visited Ottawa in March, 
1920, as member of an Indian deputation to 
the Canadian Government and was too much 
occupied to give me more than a few hours. 
In spite of the brevity of my notes I believe 
the insight gained into Haida phonetics is suf- 
ficient to warrant this paper. I cannot, of course, 
give an adequate account of the Haida sound- 

system, but purpose merely to present data 
supplementing Dr. Swanton's brief statement '. 
The phonetic system employed in this paper is 
explained in « Phonectic Transcription of In- 
dian Languages » (Smithsonian Miscellaneous 
Collections, vol. 66, n° 6). 

A remark or two on the general impression 
produced by Haida maybe of interest. I took 
several opportunities to have Mr. Kelly speak 
Haida connectedly and was thus enabled to 
hear it long enough to form a definite image 
ot its acoustic quality. It is one of the most 
remarkable languages that I have ever heard. 
Indeed, I cannot recall having at any time 
heard connected speech that appeared more 
definitely possessed of individuality. The great 
frequency of nasal consonants {n, r,), the con- 
stant occurrence of sonorous ^-sounds, the pro- 
fusion of /-syllables (see below, p. 152), and 
the musical cadences are probably the chief 
determinants of this individuality. Haida is 
very far from being a harsh language. On the 
contrary, it was voted a beautiful language by 
all who heard Mr. Kelly's recital of a Raven 
myth. Several of us in Ottawa heard connect- 
ed Mohawk, Tsimshian, Nass River, Thomp- 
son River, Shuswap, and Danish at the same 
time. If we were asked to rate these seven 
languages on the score of acoustic appeal, I 
believe the consensus of opinion would be a 
division into loin' groups : Haida as an easy 

I. See pp. 210-215 of I. R. Swanton, HuiJu, tin Illiis- 
lialive Sketch (Bureau o( American luhnology, Bulk-tin 

^O, pt. I, pp. 20)-2S2 [lc>10]). 


VI American Indian Languages 2 



first; Mohawk as a fairly pleasant, but none 
too close, second ; Tsimshian, Nass River, and 
Danish as a moderately uneuphonious third ; 
and Thompson River and Shuswap as an exe- 
crable last. These remarks are of no great scien- 
tific value, but they may be of some interest 
none the less as serving to bring home the fact 
that the « harshness » of certain West Coast 
languages results from the printed page rather 
than from their actual articulation. 

I. — Consonants. 

Thk Consonant system. — Swanton recog- 
nized 28 organically distinct consonants in 
Haida. I believe his table errs in two respects : 
in not including a number of sounds which he 
recognizes as existing but does not consider 
as elements of the fundamental sound pattern 
of the language ; and in neglecting to take ac- 
count of certain sounds that he did not hear. 
To the former class belong the anterior palatals 
and the labialized gutturals and velars. To the 
latter class belong the glottal stop ' and a set of 
glottalized nasals and semivowels. 

As to the anterior palatals, Swanton remarks, 
" An anterior palatal series might be added to 
these, but the sounds to be so characterized 
seem only palatals followed by a close [read 
" front "J vowel. " I do not know if this is 
originally true or not, but I think there can 
be no reasonable doubt that the anterior pala- 
tals are felt as a primary series. They occur 
before /-, »-, and rt-vowels, which last they co- 
lor to ci. It is true that in such a form as xdi 
ARM one sometimes hears a slight z-glide (x'di), 
but I do not think we have the right to 
conclude that xdi is felt as a secondarily modi- 

1. lam not referring to the glottal stop, if it is a 
glottal stop, of Masset Haida, which corresponds to 
Skidegate velar ^, but to a primary glottal stop in Skide- 
"ate itself. 


fied form of xiai, though it is of course pos- 
sible that it may go back to such a form. At 
any rate, I prefer to consider the anterior pala- 
tal series as a phonetically well defined prima- 
ry group of consonants and the /-glide, when 
it occurs, as the secondary fact. I do not hear 
gia''da.i the blanket, for instance, but gd''da.i. 
As a matter of fact, Mr. Kelly's ear proved ex- 
tremely sensitive on the diflfrence between the 
k- and ^-series. The word for eyebrows was at 
first recorded sk'd-'dji, but this pronunciation, 
in spite of the a-vowel, did not satisfy him. 
The correct form is sk'd''dji. There is even 
reason to believe, if my record is to be trusted, 
that there is a difference between the ^- and the 
^-series before /-vowels. Thus, I heard go'xxg'H- 
gx TO BE starting TO BURN, not -^i/- (misheard 
for -^J- ?) ; t'a'gr'gx TO be through eating, 
not -W"- as in p'sdi'r, two blanket-like 
OBJECTS. However, I have not enough evidence 
on this point to be justified in speaking with 

As regards the labialized gutturals and velars, 
Swanton sometimes writes according to the 
form Java (i.e. k'zva) to strike, at other times, 
according to the form sgod'nsir, (i. e. sgwa''- 
nsvr^ one. Here again I think there is no rea- 
sonable doubt that we must look upon the la- 
bial element (whether written lu, u, or in 
Swanton's material) as a constituent element 
of a primary labialized ^-sound. Counting the 
anterior palatals, the two sets of labialized k- 
sounds, and the new glottal and glottalized 
consonants, we have 47 primary consonants in 
Haida. They may be arranged as follows (see 
p. 145). 

Of these consonants, I did not myself obtain 
an example of aspirated p but set it in the 
table because of Swanton's form djA'pAt to 
SINK SUDDENLY, which I interpret as dju'p'at. 
It may, however, be really djxp'at (i.e. djxp -\- 
\t; cf. X3.T,\' eye), in which case p' would 
have to be removed from the table. Aside 

Six: Athabaskan and Na-Dene Languages 



Aspira- Glottal- 
Intermediate ted ized 
Stop Surd Stop Surd Stop 

Voice- Glottalized 

Glottalized less Voiced Voiced 

Nasal Nasal Spirant Spirant Spirant 












Palatal Sibilant 





Anterior Palatal 










Labialized Gut- 









Labialized Velar 




Lateral ' 















from w and 'w, which are common, and 
m, which is not rare as syllabic final, labials 
are very uncommon in Haida. Most, if not all, 
examples of initial m and 'w may be suspected 
of occurring in Tsimshian loan-words. Besides 
the 47 primary consonants that 1 have tabulat- 
ed, three secondary ones must be noted : syl- 
labically final p and /, which are true surds but 
unaspirated, and spirantal velar y ; p and / are 
secondary forms of b and d, while 7 is some- 
times heard as a variant of g between vowels. 
It is barely possible that we should distinguish 
between guttural q and velar r,, but I did not 
hear the latter sound. 

The INTERMEDIATES. — By " intermediates" 
are meant unaspirated " voiceless lenes, " con- 
sonants pronounced with the voicelessness of 
typical French surds (/?, /, k) but with the les- 
ser energy of our sonants {b. d, g). To the ears 
of English-speaking persons they are more apt 
to sound like sonants than surds, while French- 
men would be almost certain to set them down 
as typical unaspirated surds. The Haida inter- 
mediates are identical with the corresponding 
series heard in Iroquois, Athabaskan, Takelma, 

I. First three laterals are affricative. 

Yana, Achomawi, and Miwok. A recent oppor- 
tunity to study Mandarin Chinese phonetics 
has made it clear that these typical American 
" intermediates" are absolutely identical with 
the unaspirated surds of Chinese, which sound 
distinctly " softer "than the surds of French or 
Italian. It is possible that the Haida interme- 
diates are sonant at the moment of release, but 
their general effect, if carefully heard, is cer- 
tainly not that of sonants. They are here written 
as sonants because it is convenient to adhere 
to establisjied usage and because it is advisable 
to. keep p, t, ^ for the " harder" secondary 
forms that may result from them under appro- 
priate circumstances. The iv of giu and giv 
and the / o( dl are fully voiced. I supect that 
the d o{ dl'is at least partly voiced also. A word 
as to g. This is generally a very firmly pro- 
nounced stop, hut it seemed to me that a faint 
uvular trill could sometimes be detected as a 
glide between the g and a following vowel. 
Examples of intermediates are : 


dxr, THOU 

sdx'lISvr, FOUR 

k'u'dx LIP 



VI American Indian Languages 2 


dja-'dy. WOMAN 

t'/r EAR 

digud/x'rga MY DAUGHTER 

t'a-'gx TO EAT 

giva'i ISLAND 

ga''.'na bucket 

ga'xa'' CHILD 

gtva- sea-bird (sp.) 

dla'i peace, quietness 

After an accented short vowel Swanton 
heard a / before dj and d. Hence he writes 
klA'tdju and xA'tdjii small ' for what I heard 
as k'xdjir ; similarly, klA'tdAla small ones ^ 
The firm, voiceless attack of the dj after a 
markedly short vowel created the illusion of 
a syllable-closing /. As his classifiers k !At- and 
xAt- are illustrated only before dj, it is almost 
certain they should be read k'a- and xol- (Swan- 
ton's x and .V, it should be remembred, are my 
X and X respectively). He himself writes xA- 
in xA'dAla, the plural of xA'tdju '. That our 
analysis is correct is demonstrated by k'oLSgwan- 

S\:r^ ONE little OBJECT, k\sdir, TWO SMALL 

objects. The point is of some interest for 
N.idene, as it leaves Haida ^'a-, classifier for 
small objects, identical with its Tlingit cognate 
-k'x to be s.mall. 

The unaspirated hard stops. — When b 
and d appear at the end of a syllable, they sound 
much more like our/) and / than do the ordi- 
nary intermediate ^ and d. They impreess the 
ear as the normal English p and/; in which the 
breath release has been suppressed. Examples 
are : 

dx'pdju TO BE VERY SHORT "» 

1. Op. cit , p. 231. 

2. Op. cit., p. 241. 

3. Op. cit., p. 276. 

4. Swanton gives t.'Ap- as a classifier for short and 
protruding objecis (op. cit., p. 234), which does not 
correspood to my data. He does not give dAp- as a clas- 

di''gltga MY CHILD 

q'e't SPRUCE (cf. q'e-'d-x to be spruce) 

Sget RED 

^i''x\t to. PICK UP (a canoe) 

That the final stops are not aspirated is a 
noteworthy fact in view of the common Ame- 
rican Indian rule that final stops, whether in- 
termediate or aspirated surd in origin, are re- 
leased by a markedly audible breath (e. g Oji- 
bwa, Tlingit, Nootka, Comox, Takelma, Yana, 
Paiute). In this respect Haida difl"ers from its 
remote relative Tlingit and is in accord, it 
would seem, with certain Athabaskan dialects. 

Goddard writes, e.g., Hupa Lit smoke, while 
the Franciscan Fathers write lid in Navaho. It 
is likely that 'an unaspirated surd is tneant in 
each case. If I may trust my memory of Chasta 
Costa, there too final stops are unaspirated, e.g. 
l-t'at TO GO to pieces. 

The aspirated surds. — These are "hard" 
and markedly aspirated surds, much like our 
English p, t, k when initial before vowels ex- 
cept that they are even more strongly aspirated. 
They are identical in every respect with the 
aspirated surds of Athabaskan, Tlingit, Takel- 
ma, Yana, and Chinese. 

Examples are : 

t'a- TO eat 

st\ from 

tc'i'nga grandfather 


k'd'xwx OUTSIDE 



k'wa'i HIP 
q'a''dji' head 


sifier. Apparently there was some confusion between 
these rhyming classifiers. 

5. Swanton defines the classifier tci- (op. cit., p. 227) 
in terms of « such objects as full sacks and bags, pillow, s, 
etc. ». Mr. Kelly stated Ic'i- referred primarily to the 
blown-up stomach of a seal used asm float in fishing. 

Six: Athabaskan and Na-Dene Languages 



q wa'i ROPE 


tc'a'tidju'gx TO BE FAT 

Swanton remarks, " It is doubtful whether 
d and / [i. e. /'] and dj and tc [i. e. Ic'] really 
exist as recognizedlys eparate sounds " ' . I do not 
see how there can be any reasonable doubt on 
this point, da' thou and t'a' to eat ; dji--, 
classifier for cleft objects (like hands), and 
tc'i'-, classifier for blown-up objects ', are as per- 
fectly distinct as g and k' or g and q\ (Swanton 
does not list dji'- as classifier. He either did not 
isolate it as a classifying element or he confused 
it with tc'i'-. Examples of its use, besides dji''- 
djw already quoted, are sila'i dji'-sgwa''nsir, 
ONE hand [literally, hand cleft-one], stki'i 
dj'.-sdi'r, TWO hands, k'u'dx dji'-sd\r, TWO lips. 
There is no doubt that there are plenty of 
examples of dji'- in his material. Note, for 
instance, dji'wAl [Ma.sset] roots of fallen 

The glottalized stops and affricatives. 
— These are the well known stops and affri- 
catives pronounced with simultaneous glottal 
and oral closure and with glottal release follow- 
ing upon the oral release. Swanton says, 
" Some speakers bring these out very forcibly, 
while others p.iss over them with considerable 
smoothness ^ ". Mr. Kelly pronounced them 
quite as smoothly as any other consonants. 
Their essential nature is certainly not to be 
explained as due to " urging more breath against 
the articulating organs than can at once pass 
through 5 ". If there is a true " fortis " series 
in Haida and Athabaskan, it is the aspirated 
surds, which are indeed pronounced with an 
excess of breath. 

Examples of glottalized surds are : 

1. Op. cit., p. 210. 

2. Op. cit., p. 210. 

3. Op. cit., p. 210. 

t'a' gun feather 

t'a''r,OLl TONGUE 

st'a'i FOOT 

/Vw red cedar 

qoLuCcdoL cheek 

q'a't'car, to be going in 

^'d'll leg * 

k'i''dji seal-stomach 

k'a't DEER 

k'w'dx LIP 

k'wd'i TO WAIT 
q'a'ri HEMLOCK 

q'e'sdir, two spherical-like objects 


fia''doLn gorge 

C'tll WE ' 

I may note that/V tends to move front in its 
position — either to that of an nnterior palatal 
c- sound (/V) or even to that of an .f- sound (e. 
g. i''knCsui3. dla'' msdir, two gigantic people ; 
but also q'y.'nt'c:dx cheeks. Does t^adx corres- 
pond to Swanton 's -djit (op. cit., p. 260) ? 

The VOICELESS spirants. — These require 
no special comment, x is pronounced like ch 
of German ich. As Swanton remarks, s often 
interchanges with dj^, which does not normally 
occur as a syllabic final. Swanton states that " s 
becomes dj before most vowels ". Inasmuch as 
both dj and s occur between vowels (e. g. i'dji' 
to BE : isdxr, TO BE CAUSING TO BE ; but also 
" participial " -asi, -as'.\ would it not be better 
to say that -dj- becomes -s just as -b- and -d- 
become -p and -/ .'' We would then have one 
final (-j) representing two distinct intervocalic 
consonants (s- and -dj-). 

Examples" of voiceless spirants are : 

sdw'nsOLT^Xy. EIGHT 
xdi ARM 

4. See p. 152 regarding syllabic /- sounds. 

5. Op. cit., pp. 214, 215. 


VI American Indian Languages 2 


t'a-'xidlgJ. TO BE ABOUT TO EAT 

xa-'ya sunlight 
xwi' to be cold 

Xa' DOG 

X7Lr,\' eye 

xd neck 

xiua-'igcr a thing that is loose 

la\ ia, il i 

igiinnl THREE 

The nasals and voiced spirants. — Of the 
three nasals, n and c may occur as either sylla- 
bic initial or syllabic final. In such a word as 
'l(i''(ar,a THEIRS the guttural nasal q must be 
considered as belonging to the final syllable, in 
such a word as xoc'r^'c eye to the first, m, as we 
have already seen, is rare as an initial but not 
uncommon as a final. 

Examples of final m are : 

t^'mdjtr TO BE something THIN AND ROUNDED 

fy-m lice (Mr. Kelly considered this word 
as connected with the classifier t\m-. This mat 
be only a folk etymology, however.) 

dld''nidjn' to be a gigantic person 

ga''mdju- TO be a wide thing 

^d'mdjir TO be a large (canoe, blanket) 

Several Haida syllables ending in m seem to 
belong to a set of classifiers : 

t'xm- THIN and rounded 

dlam- gigantic, corpulent (?) 

ga'm- wide and rounded 

^iim- large (in reference to canoes and 

igum- (Swanton) large and roundish (e. 
g. rattles) 

Fa;//- (Swanton) small and roundish 

js,'rt'm- (Swanton) large around (?) 

dla'm-, ga'm-, and gam- are not listed by 
Swanton. My ga'm-, however, may be the 
same as his ga-w- (op. cit., p. 235) ; I am cer- 
tain of the velar g, for the element was also 
recorded zs -(am-, e. g. ^w -(am-sgiva'nsu, one 
big ear. 

Although the evidence is far from complete, 
I would risk suggesting that this set, of which 
there are doubtless other, members, contains 
a common element -;//- indicating something 
like rounded, all around. This view is strongly 
supported by the fact that several of the m- clas- 
sifiers are clearly related to other classifiers 
without -m or with final -p. Thus, to t'am- 
may correspond t'a- (Swanton) coiled and 
flexible; to ga'm- evidently corresponds ga'- 
FLAT (e.g. pr -(a'sgiva-'nsir, one flat ear); 
^dm- is clearly related to p'- canoe, blanket- 
like (e.g. tin' p'sgiua'fiS'.T, one canoe and 
^d''at grsdi'r, two blankets : tiw ^d'mdjus 
A large canoe, tiw gd-'mdxh sdvr^ two large, 
SPREADY CANOES, m'' ul gamdoiJa gi'sdir, two 
large blankets) ; igxm- probably belongs to 
iga- (Swanton) branching objects, k'xm-, as 
Swanton himself points out, is derived from 
k'x- SMALL ; and slla'm- belongs with sthp- 
(Swanton) slim. 

Parallel to the -m- series is a -p- series 
(fo^p-, doip-, stixp- [Swanton], t'ioip [Swanton], 
audita/)- [Swanton] = probably sk'oip-^. Of 
these, tap- may go with t\m- ; stiap-, as we 
have seen, with slia'iii-; t^ixp-, judging from 
Swanton's one example, which refers to the 
surface of the moon, goes well with his L!- 
thin and FLAT (as I shall show later, this must 
be interpreted as t'il-, a reduced form of /'fa-, 
t'ia-); and skap-, which in his one example 
refers to the curled tail of a dog, may go with 
sk'a'- roundish (used of eyes, water-drops, 
berries). The meaning of the -p- series is more 
obscure than that of the -m- forms ; possibly 
stubly protruding covers the case. Ond" sur- 
mises that the complete set of Haida classifiers 
is a more complex subject than Swanton's data 
represent. Further, that the i- and s- which 
begin so many of the classifiers (and other noun 
and vtrb stems) are vestiges of an older classi- 
ficatory system that was related to the /i- and 
s: classifiers discovered by Boas in Tlingit. The 

Six: Athabaskan and Na-Dene Languages 


NOS. 3-4 

whole subject of Haida classifiers needs a rene- 
wed and intensive study. 

Examples of n and r, are so numerous in the 
forms scattered in this paper that I do not need 
to give further examples here. 

Unglottalized w does not seem to be as com- 
mon a sound as glottalized w. Examples of w, 
y, and /.are : 

ha- Wit HURRY ! 


yd-'goLlari ancestors 
ga''yir sea 
q'a'fcaya''g<xn went in 
t'a'lx'rj WE 

xd NECK 

sk'a''s'oI%r, A round thing 

The glottalized nasals and voiced spi- 
rants. — The sounds coming under this head- 
ing are identical with the corresponding sounds 
in Kwakiutl, Nootka, Nass River, and Tsim- 
shian. I have not heard the Kwakiutl sounds 
of this type, but the Nootka series ('w, '«, 'w, 
'j), the Nass River and Tsimshian series ('w, 
'w, 7, 'w, y), and the Haida series ('w, *w, 'r,, 
'w, 'y), sound perfectly analogous to me. 
I hear no difference, for instance, between the 
'i^ of Nootka W"}7' HIGH HILL, of Nass River 
'wr GREAT, and of Haida 'wxsi' that thing. 
I speak of this because Boas has differentiated 
the 'm, for instance, of Kwakiutl from the 'w 
of Tsimshian, which he writes m ! and consid- 
ers a '' fortis " m analogous to " fortis " p ! . 
(our p'). In classifying these sounds with the 
glottalized stops and affricatives (*' fortes ") of 
Tsimshian I believe he is perfectly correct, for 
they all belong together psychologically, but 
this grouping applies fully as well to Nootka. 
When I first taught a Nootka Indian to write 
phonetically and explained the meaning of such 
symbols as p ! and / ! (for which I now write 
/)' and /'), I was interested to find that he wrote 
m ! of his own accord where I was in the habit 

of writing 'm. He seemed puzzled to find that 
I was not using an analogous orthography for 
the glottalized stops and affricatives on the one 
hand and the glottalized nasals and semivowels 
on the other. This instance demonstrates pretty 
clearly, it seems to me, that the native phonetic 
feeling of Nootka finds the essential peculiarity 
of the " fortes " in their glottalization and not 
in their supposedly '* increased stress of arti- 
culation , " for in such Nootka sounds as 'tn 
and '« there seems to be no increase of stress. 

In the glottalized stops and affricatives the 
closing of the glottis lasts during the whole 
oral articulation of the consonant and beyond. 
In the glottalized nasals, semivowels, and voi- 
ced lateral, however, the glottis is closed simul- 
taneously with the oral contact but released 
instantly thereafter, the voiced continuant the- 
reupon becoming fully audible. The acoustic 
effect, therefore, of such a sound as m is very 
nearly of a glottal stop followed by ni-, yet 
not quite, for a conscious compounding of '-|- 
m nearly always fails to satisfy the Indian's ear. 
That the glottal element is felt to inhere in the 
consonant is clear from the syllabification. In 
such words as Haida ga''tia bucket and fa'^na' 
child the glottal stop belongs to the second 
syllable, not to the first. In neither Haida nor 
Nootka, as a matter of fact, can a syllable end in 
a glottal stop. 

So far as I know, '•/; has not yet been record- 
ed for any other Indian language. It is not a 
common sound, nor are '/// and ';/ frequent in 
Haida;'///, as I have already indicated, probably 
occurs mainly in Tsimshian loan-words. On the 
other hand, \i\ '\, and '/ are exceedingly com- 
mon sounds, appearing in some of the most 
important stems in the language (e.g. \L'a- 
that, 'j)7r- big, la he). How important is the 
distinction between 'v and v, for instance, may 
be seen from the fact that when I pronounced 
Swanton's yu'An big as yir'xn, Mr. Kelly had 
not the remotest idea what it meant; it should 


VI American Indian Languages 2 


have been ';y7r''an. The finding of these sounds 
in Haida was unexpected. They are not found 
in Tlingit and seem also to be absent in Atha- 
baskan — with one interesting exception. In 
Navaho the Franciscan Fathers have recorded 
■;/. It is the regular correspondent there of an 
etymological t -\- n. 

Examples of glottalized nasals and voiced 
spirants are : 

'ma' (exclamation of pain) 

'madjx OCHRE (Probably borrowed from 
Tsimshian. Cf. Tsimshian niES- reddish, wE5- 
ans OCHRE ; perhaps Boas' m'Es- is to be read 


ga''na bucket 

W,a'g3,'r,ga'r, to vie with one another 

'wa'ngx TO LOOK for food at ebb tide 

'tt'aJi* THAT THING 
'yu'"oin BIG 

'la' GOOD 

la', 'h, 7/ HE 

The laryngeal consonants. Every stem that 
apparently bet;ins with a vowel really begins, 
in all probability, with either a glottal stop or 
with /;. Examples ol glottal stops beginning 
syllables are : 

'fl'7/ PADDLE 
'a'uga .MOTHER 

'a''tlgxu HERE (Swanton writes d'LgAn, 
but I heard //, not dl. This may be a mishea- 
ring on my part.) 

'(tllod-'a OURS 

'l-7'.Y] MAN 

7 'i''i:r,x ifc'dju'gi HE IS A STOUT MAN 

na'g"'i -77 'a in the house 
yif'txn BIG 
{gu'nuf three 
iH-fu'nui si.K 
tia'"ai ten 

sk'a's'olxT, A round thing 
In a few cases the ^lottal ston was not record- 

ed, e.g. t*7tr,/"iu/a 5iii-/; two MEN, ga'xa. i'hri<x 
CHILD-MALE, BOY, but these are either mishear- 
ings or secondary slurrings on the part of the 

There are, however, at least two important 
elements that begin, or seem to begin, with a 
vowel unpreceded by a glottal stop. These are 
the demonstrative u, o (w , o') and the verb 
i''dii\ i''dj'., a. Examples of the demonstrative 
are : 

di' sk'ddj u' v'dji it is my eyebrows 

§d'gxn na'i {y)u my house (Note the glide 
)', indicating clearly that no glottal stop has 
been slurred, gd'gxn is probably assimilated 
from gd'gxr,, see Swanton, op. cit., p. 259, 
who wites gia'gAfi) 

'la'o' he <C 'la o' 

ia'o- I <iiao' 


na''gai yf ia q'a'tcar, i go into a house 
a'k'os u this thing" 
'ivx'sga.i sd:r, fs'.'zvx'gx those two are 


The reason for this absence of the glottal 
stop has been indicated by Swanton. The gene- 
ral demonstrative of reference is haii (S>vanton's 
hao), often contracted to ir , u, 0. The vocalic 
hiatus without glottal stop is thus the etymo- 
logical equivalent of an old h. 

Examples of the verb i''dj\., '.5 are : 
la 'o' na'.t -;a c'dj-.-ri i am in a house 

tllgU".'' U tx isdxr, I CAUSE IT TO BE AWAY, 1 

See also the first example under demonstra- 
tive u, 0. The consistent absence of the glottal 
stop in this verb may possibly be explained as 
a slurring, but I think it more likely that it is 
to be interpreted as due to an old /; that has 
disappeared. I would suggest that i'dj\, is is a 

I. a- slurred from 'a-, the demonstrative stem this 
corresponding to ';tw- that. 

Six: Athabaskan and Na-Dene Languages 



later form of hi'dj'., his, and that this verb is to 
be understood as composed of a demonstra- 
tives stem hi'-, hi- (cf. hi''-dll-goLl come here ! 
and Swanton's hi-t!A-gA'n then, hi nan 
only). Justasthis /;/- parallels /;fl-(cf. Swanton's 
hx-n LIKE, AS FOLLOWS ; general demonstrative 
hao^if , u,o; " article "*hai, parallel 10 gai, in 
e. g. *na hat house-the > na'i), so *hi-dji, 
*hi-s TO BE parallels an old *ha-s, preserved as 
Swanton's " participial " -as, -is, -es. Such a 
phrase as Swanton's nxn gaxd'gas one who was 
A child is to be interpreted, it seems to me, as 
«ar( gaxa-'-g a-s one child-be it-is, contracted 
from an older gaxa'-ga has. Such endings as 
-^-5? (S wanton) are probably to be interpreted 
as -ils\ <Z -l-y-S'; i-asi, which occurs as phonet- 
ic parallel. See below on /- syllables. 

Though /; has demonstrably dropped out in 
certain cases in intervocalic position, it is a 
clearly articulated consonant when preserved, 
e.g. : 

hv'dllgy.1 COME here ! 

haw't HURRY ! 

Secondary consonantal processes. Final 
vowels are, as a rule, released without breath. 
This is in keeping with the phonetic forms -p 
and -t instead of/)' and t . A few cases of breath 
release have been noted, however, in absolu- 
tely final position : 

na'' house (but na t'c'-sdir, two houses) 

go'gx' (it) is on fire 

go'dxgy.' sets fire to 

Rather frequent is the spirantal voicing of g 
to velar y between vowels : 

dx'r,^d^a yours 

'la''(a'r^a ^d-ra theirs 

'zua'n-^oi. TO GO down to look for food at 

da la '(O'da you start the fire! 

na''gai yc into a house 

I. Op. cit., p. 254. 

ga-t'a-' 'hfai"';a something-eat keeper, 

d j c' go' '.;7i seven 

gw -(%'sgu<a''nsvr^ one (flat) ear 

q'o''-(a rock 

Less often g is spirantized and unvoiced after 
initial / : 

ixa' stone (for iga') 

Before s, t\, n, and i there is sometimes assim- 
ilation of r, to n : 

sdx'nsir, four <C *sd7:r,-sd<.r, TWO-TWO 

r'hnfsuix dla''msdvr, two gigantic people 
(cf. i-'hT,a man) 

dxn II I THEE. . . (Irom dxr, iV) 

gd'gxn na'i my house 
But this rule is not invariable : -y; is preserved, 
e.g., in : 

'r'/i-/; fsda. dx'psdir, TWO STUBBY MEN 

On the other hand, n does not assimilate to 
r, before g : 

tc'i'nga grandfather 
'llgx'nga cousin 

Initial consonant clusters.— Every Haida 
syllable begins with one of the forty-seven 
consonants listed in the table, or with a cluster 
of two consonants, or with a vowel (originally 
preceded, it would seem, by /;). The first 
element of the cluster is always i- or s-. The 
second element is an intermediate, aspirated, or 
glottalized stop or (after 5) a //, never, I believe, 
a sibilant affricative or a spirant ; ix- is merely 
a secondary form of /^^-. Examples are : 

igi'- (classifier for fairly big, roundish objects) 

iga' STONE 

igu'nid three 
Ik'a'i chin 

sdir, TWO 

sget red 

sgiva''ns:r^ ONE 

st'v'wai THE sea-egg 

sk'a'- (classifier for round objects) 

sq'aii GROUSE 


VI American Indian Languages 2 


St a i FOOT 

sq'a- (classifier for long objects) 

I obtained no example of i before a glottalized 
stop; Swanton gives Ik'.A'mAl needle of co- 
N'lFEROUS TREE. I suspcct that his in- is really 
llr-, as it would be an isolated example off 
(or s) -\- nasal. Swanton fails to distinguish 
initial sd- from j/'-. He writes both st-. An- 
other error is his group sL !-. It may exist, but 
his instrumental sL!- with the fikgers and 
sL!di HAND are really stll- and stia'i, probably 
related, with ^-prefix, to ///- to touch (Swan- 
ton's L-, p. 226). 

Swanton's clusters with initial /, dl, li, and 
/'/ do not really exist. These /-sounds are in 
every case to be interpreted as syllabic //, 'II, 
dll, til and /'// (see below). 

II. — The syllable. 

Before taking up the vowels, it will be con- 
venient to define the Haida syllable. There is 
no doubt that the language, like Athabaskan, 
has a strong and well-defined feeling for the 
syllable as an integral phonetic and psycholo- 
gical unit of speech. It therefore becomes 
important to understand its structure. Aside 
from the secondary loss of h and the slurring 
of ', every syllable begins with a consonant 
or an 5- or /-cluster of two consonants. It 
may end in a vowel (long or short), a diphthong 
(long or short, but I suspect that all long 
diphthongs are felt as the equivalent of two 
syllables of form *-ahai, *-ahau), an /-vowel, 
a consonant, or a cluster of two consonants 
ending in / or s (these clusters are likely, 
however, to be secondary forms of older disyl- 
labic forms, e. g. i'djins [Swunion] <^i'djiT,7Ls). 
The final consonant can be only rn, n, yj, 5, 
unaspirated p or /, /, /, or //. All of Swanton's 
examples of final dl and /'/ are to be inter- 
preted as dll and /'//. Swanton speaks of non- 

vocalic stems like " L to touch " or '' sL to 
PLACE IN A certain DIRECTION. " These ele- 
ments are syllabic : ///, stiL Many of his non- 
vocalic groups are even disyllabic, e. g. /'//'// 
(Swanton's Z,//). 

Syllables with l-vowel. — In normal 
English pronunciation the second syllable of a 
word like w^/a/ consists of a consonant follow- 
ed by an /-vowel. In other words, there is 
no pure vowel in the syllable at all, not even 
a " mid-mixed " d. The tongue does not 
release its stop position but merely adjusts 
itself on the spot to a lateral articulation. Such 
words as metal, medal, flannel are phonetically 
mi'tl, mz'dl, fld'nl. 

The peculiarity of the Haida /-syllables, of 
which there are a vast number in the language, 
is that they always begin with a lateral con- 
sonant, voiced or unvoiced. The following 
table shows the relation between the simple 
laterals and the syllables with /-vowel : 

Syllable luith y.-vcnvel Syllable with l-vowel 

I- n a-, 

*/« '// ('/•, '/) 

dU dll {dl, df) 

h il 

tix til 

fix t'il 

The syllables in the second, column are 
reduced forms of those in the first, with which 
they vary in Swanton's orthography (e.g. Lga 
or LA'ga land, both to be understood as 
lil'ga). They also interchange actually accf>rd" 
ing to accentual or other phonetic circumstances 
with full syllables in a', a, a (e. g. la', 'la in 
'In' .va'r/'." HIS EYES, na''i •;'. 'la ga't'car, HE is 
GOING INTO THE HOUSE : '/ 'i''ii.r,7. tc'r'djw HE 
IS A BLOWN-UP (obese) MAN ; /fl I : // ; -dxloL 
adjective plural : That Swanton too 
heard, though he did not explicitly record, /- 

Six: Athabaskan and Na-Dene Languages 



syllables is shown by such accentuations as Z,'-, 
/'-, I'-andL!'-. 

Examples of /-syllables are : 

k^all LEG 

da''ll RAIN (recorded as dal by Swanton, 
which failed to satisfy Mr. Kelly's ear. This 
word is not only clearly disyllabic, but the 
accented vowel is higher in pitch than the /- 
syllable. Cf. Swanton's ddla-ge'ii-si rain fell, 
BBAE 29 : 12, 1. 8.) 

'a'll PADDLE 

qui'' FOREHEAD (should have been record- 
ed q'u'Il or qui') 

'IJgx'nga COUSIN 

'h;a'i'^';a keeper 

dx'n U YOU I. . . (not to be confused with 
dxnl swelling) 

Sge-'il TO CRY 

titil TO RUB one's hand against 

tlc'll five 

q'w'dll to go aboard a canoe 

hi'dllgal come here! 

tilga' earth 

tilgWV' AWAY 

til^u'nui SIX 

t'il'^d' to soak (a dry salmom) 

'i''t'ilgd-fa ours 

When an /-syllable beginning with /, 7, 
or dl is pronounced with very weak stress, 
the length of the / is of course reduced, though 
it never loses its syllabic character. We may 
then write simply /, '/, dl. The syl'able -ilr, 
(reduced from -/ay;) was heard as -ir,, with syl- 
labic q, in lia'^ait, (i. e. tia'"al avj) sgu'a''ris:T, 
gou ten it-is one missing, nine. 

In interpreting the phonetics of Swanton's 
texts, there is no genuine syllabic ambiguity 
in the case of initial L-, L-, L!-, and /- before 
consonants and of final -L and -L/, which 
necessarily represent /-syllables. It is different, 
unfortunately, with initial / before consonants 
and with final -/, -/, and -L, which may repre- 
sent non-syllabic or syllabic laterals. It is im- 

possible to tell offhand whethersuch orthograph- 
ies as igeda-i and ga'lxua represent igc'daJ 
or ilge'da.i, ga''lxu'a or go'lhwa. It is hardly 
conceivable that the morphology of Haida can 
be adequately understood without an exact 
knowledge of its syllabification, for the Haida 
" word " is essentially a group of significant 
syllables. The recognition of /-syllables is likely 
to put many points of grammar in a new light. 
Thus, we may surmise that the two forms 
of the first person plural (subjective fa-la-r, 
and objective ''/'-t'il) are closely related, ''/'-til 
being a reduced form of *\''-t\xya. The true 
basic forms would then be fa and 'c-fa, -la 
and -Ix-r, being pluralizing elements. 

Syllabic nasals. — It is quite possible 
that we should recognize also syllables of type 
n and syllabic q, alternating on phonetic 
grounds with syllables of type w. and r,a. The 
alternation of-r; (-n) and -/;« in cases like i"7.-/; 
(or i''i:.n) fsdx sdvr, two men 

i''hr,x sgiua''nsiri one man 
should perhaps be interpreted as an alternation 
of i"'/iy;y; {'.''hnn by assimilation) and i'7tr,a. 
This consideration may explain the constant 
interchange in Swanton's material of forms like 
-gAn and -gAni, -agAri and -agAni, -gin nnd 
-gtni, in which -/ (i. e., -i) can hardly be a 
" perfective " element, as Swanton assumes. 
For the present, I cannot say whether Haida 
distinguishes final -n and -r, from -n and sylla- 
bic -Y). All I can say is that I feel strongly that 
while Haida reduces the quantities of syllables 
freely (e. g., in such a gamut as 7^/', 'la, 7a, 
'//, '/). it resists the extinction of syllables, and 
that if syllables ar.^ actually lost, they are lost 
phoneticallybut not psychologically. 

HI. — Vowels. 
I am able to give only a preliminary idea 


VI American Indian Languages 2 


of the Haida system of vowels. The vocalic 
nuances seem to be due primarily to secondary 
phonetic causes rather than to basic etymolo- 
gical differences. It is quite probable that there 
are only three organically distinct vowels : 
a, z, //. Each of these runs through a gamut 
of quantities and qualities that give the lan- 
guage a far greater vocalic variety than the 
simplicity of the fundamental vowel scheme 
would argue. 

Qualitative changes. — The a-vowel 
seems to be the most liable to phonetic 
change. Its fundamental quality is that of 
German a in Mann, e. g. da''ll rain, but 
when short, whether accented or not, it very 
frequently assumes the duller timbre of a. 
Thus, -ga TO BE and fa i vary with -gx and 
// (i. e. /x) in : 

ia ga l'a''s[a i am eating something 
a'k'os u il l'a-''(xsgx i shall go and eat 

THIS thing 

The a-timbre is particularly common before 
nasals, e. g. -gx-r, (continuative of -ga^ and 
-^a-n('.) (past definite oi -ga), /'aw lice. 

(It seems fairly evident to me that Swanton's 
temporal -gxn [op. cit., p. 247], future -sga 
[p. 249], and continuative -gah are merely com- 
binations of " declarative " -ga to be or to be 
so and certain particles that he has not properly 
isolated : -r,, -r,a to be continuously ; -s, -sa- 
ABOUT to bk ; and -n -««. to be at a given 
MOMKNT OF TIME, TO HAPPEN. It is clear that 
they occur also without -ga. Boas has already 
pointed out the analysis oi -s-ga- [see p. 249]. 
I hope to show at a future opportunity that the 
whole tense-modal system of Haida is nothing 
but a loose compounding of demonstrative 
elements and particle verbs and that the synthe- 
tic nature of this scheme is more apparent than 
real. Thus, Swanton's " infallible future 
-asah [p. 249] is merely a verb phase : 'a-sa-r,- 
[a] this-will-be [durativkly]). 

After anterior palatals and y, a (a) appears 
as a (as in English mat), perhaps even as z (as 
in English met), e. g. : 

k'd'xwx outside 

^'d'll LEG (also heard as k'e'l) 

Xdi ARM 

t'il'^d- to soak (a dry salmon) 
sk'd-'dji eyebrow 

^d'gxn MINE 

^d''da.i THE blanket 

^dm- wide and sprhady (classifier) 

yd''gxlxq ancestors 

After sibilant afFricatives and also before r, 
there is a tendency for x, particularly if unac- 
cented, to pass into i (English i of bit). 
Before yj this seems to take place particularly 
after laterals (/, /) and sibilants. Examples are : 

sdvr, TWO (d. sdxnsi-f] four, dissimilated 
from sdx'n-sdvr, <C *sdxr, -sdxr,) 

daUr,^d''(a yours (cf. dxixr, ye) 

There is no doubt that many Haida syllables 
in I are old a-syllables, as variants prove ; 
e. g. 'r'hrix MAN is evidently to be understood 
as an original 'i''ixr,x (perhaps actually pro- 
nounced 'i''iir,x, as Swanton's variant ff/^a 
seems to indicate), as shown by certain of 
Swanton's forms, e. g. iia'n-, iia'h-. 

After anterior palatals there is good reason 
to believe that an original a is sometimes com- 
pletely palatalized to r, i. This is not a mere 
matter of nuancing, like the change of a to a 
or of X to I, but a definite phonetic process that 
disconnects the new vowel from its old catego- 
ry. Under what circumstances this change 
takes place I do not know. I suspect that an 
old a' is merely colored to d' {a) after anterior 
palatals, but that the corresponding short 
vowel was completely palatalized to i (secon- 
darily also /■). A good example is ^i'- (clas- 
sifier for blanket- like objects), which is almost 
certainly related to gd'-, gd- of gd''-d- blan- 
ket and ^d-m- (see page 148). The original 
quantitative relations were probably as follows : 

Six: Athabaskan and Na-Dene Languages 



^a- > gi-- 

ga-d- >■ gd'-d- 

^a'-m- >> gdui- (closed syllable, hence with 
shortened vowel) 

Again, g'rw-ai the ear : ^w ear can be 
best explained as palatalized from an old ''gaw-, 
final -an o{*gaii contracting to gw . This inter- 
change of -/'tf- : -w seems to take place also 
after s, e. g. Skidegate sfi (Swanton) to say : 
siw-As (contrast Masset sdw-^. A basic saw- 
best explains the alternations. Another alter- 
nation of similar form is that of tin' canoe : 
tilw-a'i the canoe (Swanton always writes 
Lu-a'i). This again is doubtless the reflex of an 
old tlaii : tlaiu-. 

For the /-vowel, /, i, and e were heard as 
variant timbres. The t-timbre is quite frequent, 
perhaps normal, for the long /- vowel, e. g. : 

'c'til^dya OURS 

A-a'r/f EYE 

tilglUf' AWAY 
djr'gO'-(OL SEVEN 
'.■ ilTfT. MAN 

Examples of the less common ^-timbre are : 

c'dji-q IS 

xei REGION OF THE NECK (cf. xd neck). 

The characteristic timbre of the u-vowel is 
u (as in English put), here written u. The cor- 
responding long vowel was heard partly as o' 
(e. g. go'dx TO START A eire), partly as w 
(as in English fool). The latter seems to occur 
chiefly after sibilants and anterior palatals 
(e. g. -dju- TO BE so AND SO, gU' EAR, 'y/r"Qt« 


A glide a was noted in 'lyai^-^a keeper. 
In ^d''at BLANKET, the long a vowel, modified 
to d after g, reasserts its proper quality before 
the final / and thus appears as a broken vowel. 

Vocalic quantity. — Quantity is a diffi- 
cult matter in Haida. It is likely that there are 
etymological quantitative distinctions, but it 
is impossible to be certain, as the actual quan- 

titative variations are clearly largely due to 
secondary lengthenings and shortenings of 
the fundamental vowel. Thus, as already 
pointed out, 'la he may be shortened to 7a, 
'II, '/or lengthned to 'la' . The determinants of 
this quantitative variation are probably pho- 
netic rather than morphological, but I doubt 
if stress accent is the only or even the decisive 
factor, as both the long and the short vowel 
may occur in an accented or unaccented syl- 
lable. I suspect that the distribution of quanti- 
ties is the resultant partly of inherent quanti- 
tative distinctions (e. g. long a' in yd''gxhr, 
ANCESTORS : short a, a in k\'dju' to be short), 
partly of a tendency to establish a rhythmic 
equilibrium. This equilibrium seems to depend 
on several factors, chief among which are the 
nature of the syllable (a closed syllable tends 
to be short, an open one long), the place of 
he accent, and the grouping of the syllables 
in phrases. It is impossible to give rules at 
present ; the subject is evidently complex. At 
the same time I do not feel that the quantities 
are distributed ad libitum, rather that they 
remain to be discovered. 

The Haida type of quantitative variation is 
somewhat similar to that of Athabaskan, if 
I am not mistaken. Presumably the Tlingit 
feeling for quantity is analogous to that of 
Haida. Boas merely remarks that " the quan- 
titative value of Tlingit vowels varies conside- 
rably " '. In the body of his grammar he unfor- 
tunately pays no attention to the subject of 
quantity. The quantitative peculiarities of the 
Nadene languages are in striking contrast to 
those of the neighboring languages. In Es- 
kimo, Wakashan, and, I believe, in Tsimshian 
as well, the inherent quantitative value of a 
vowel is clearly felt and, on the whole, rigidly 
preserved. In these languages vocalic quantity 

I. F. Boas, GiamiiiatiCid Notes on the Laiu:;iiaQy of llie 
Tliu^it Indians, Univ. of Pa. Anthr. Publ., vol. VIII, 
no. 1, 191 7, p. ; I. 


VI American Indian Languages 2 


is as much a matter of etymology as is the 
consonantal framework of the word. Phonetic 
variations in the quantity of vowels such as 
are found in Haida would be well nigh un- 
thinkable in Xootka. 

Examples of quantitative alternations are : 

a — g^i'-y^'' CHILD : ga'Xx i''i':r,x CHILD 

MAi.K, boy; gaxx .ija'dx child female, girl 
s^'d''dji' EYEBROWS : di' sk'ddj w v'dji' 


sda-'nsar,xj. eight : sdy.'ns'.r, four 
na-' house; na'ga'i yfl 'a in the house : 
na I'e' 'sgwa'nsvr, one house 

sgwa''nsi.T, OYiE ; q'a-'dji q'e'sgwa-'nsir, one 
head; k'un dji'sgwa-'iisir, one nose; k'nn 
k'xsgwa-'uy.r, ONE LITTLE NOSE, POINT ; ou' 

''^d-sgwa-'ns\r, ONE (flat) ear : gir -'amsgwa'n- 

S'.r, ONE BIG EAR ;^M' k' x'sgwanS:r, ONE LITTLE 

E\R;m-'dai ^rsgwa'nsir, one blanket, shawl 
i — ; q'a''dji head, hair of head : q'a'dji 
7/ — ; la-'o nw'gai vt- q'a't'car, he is going 

INTO A house : na''i vi 'la q a't'car, he is 

GOING into the HOUSE 

dji-'djir TO BE SPLIT up, CUT UP : stia'i 
djisgwa-'ns'.r, one band ; stia'i djisdir, two 


/•V/V" TO BE, e''djirt TO BE (duratively) ; 

ia' '.sdxr, I CAUSE TO BE, PUT 

// — sk'a''dju' TO BE SMALL AND ROUNDISH ; 
^i'dju-'gx (it) is LONG AND MASSIVE, SPREADY; 

tilwa'i u gdmdju'gxn the canoe is large : 
xd k'lf'dJH neck IS short; dx'pdjti very 
short; litr gaiiid/us a large canoe 

di'sk'ddjw MY eyebrows; ia''o' la'gd^'a na'i 

godxgx I PUT HIS HOUSE ON FIRE ; a'k'os 
U {( l'a''gx I AM EATING THIS ; '■lVx!s\' gO'Xx- 

Diphthongs. — There are two diphtongs 
in Haida, (7/ and an. Each of these exists in 
two quantitatively distinct forms — short (a/, 
often contracted to e' ; an, ou^ and long (a'/, a.i, 
sometimes heard simply as ^Z; a'o\ a'o, ao'). 
The latter type seems to result from contraction. 

The uncontracted short ai is well illustrated 
by the enclitic articles -ai (probably <; *hai), 
when appended to a stem ending in a con- 
sonant, and gai, e.g. : 

St'i''iuai THE SEA-EGG 

na''gai y<.' o into a house 

According to Swanton ', the Skidegate ai 
appears in Masset as e', but Mr. Kelly, a Skide- 
gate Haida, pronounced a number of fl/-forms 
with ^"-vowe!, e.g. : 

qe'- classifier for large roundish things (cf. 
Swanton s q!ai-, p. 232) 

yd''ge' PARENT <^ yd'' gai (cf. T^\\iv.yd''gxlxr\) 

t'e'- classifier for bulky lying objects (cf. 
Swanton's t'ai-, p. 227) 

q'e't SPRUCE (cf. Swanton's^flf/TREE, p. 271) 

tic'il FIVE (probably contains iia- ; cf. tlx, 
til- TO touch, s-tla hand) 

Possibly the Masset-like g'-forms are gaining 
currency among the younger people at the 
expense of the older az-forms. The contracted 
e' is, of course, not to be confused with the e' 
which is merely a variant of r. 

The long a'i (a.i) probably always results 
from the contraction of fl -f- ai, e.g. : 

na'i THE house <C na-(h)ai 

sCa'i foot <C st'a-{h)ai (cf. Swanton's ins- 
trumental sl!a- by kicking) 

dia'i auiETNESS < dla-(h)ai 
Just as the primary ai may contract to e' , so 
it is likely that a'i may contract to ai. Thus, 
it seems probable that ^d''da.i the blanket is 
to be understood not as gd't -\- ai but rather 
as contracted from ^d'da-{hyn ; q\r'-(ai the 
ROCK, as contracted from q'o'''(a rock-|- (h)ai. 
Similarly, xdi arm >< xd'i < old *xa-(h)ai ; *xa 
itself passed to xi (Swanton's xi^) as *^a passed 

The short ^/<-diphthong is written ao by 

1. Ot). cii., p. 213. 

2. Op. cit., p. 226. 

3. See page 155. 

Six: Athabaskan and Na-Dene Languages 



Swanton. I heard it as an ordinary a -\- u and 
did not feel that the labial element was low- 
ered to o-position. Examples are : 

'a'uga MOTHER (Swanton's aoga^ 

sq'au GROUSE (Swanton's sqao) 

It seemed to me that it was rather the a that 
tended to become modified in the direction of 
u ; thus, I heard gm to be wanting (in tia-"air, 
sgwa'' ns'.Tt gJu nine) and 'o''tiga, almost 'o''ga, as 
variant of'a'uga. This tendency o( au to con- 
tract to a monophthong is carried through, 
e.g., in ga''yu' sea (cf. Swanton's gayao, 
p. 272), ^tf ear (*^aw, *gau, and in w, 0', u, 
as reduced form of demonstrative hau. 

The long a'o, also heard as wo' and ao', is 
markedly distinct from- short au. It seems to 
result always from the contraction of a -|- (/?) 
au, e.g. : 

'la'O', 'la'o HE (emphatic form ; «< 'la-[h\aii) 

ia'o', lao' I (emphatic form; <C la-\h\au) 

IV. — Stress and Pitch. 

I was not able to arrive at definite conclu- 
sions in regard to Haida stress and pitch and 
their interrelations. It is clear that a given 
word is regularly accented on a certain syllable. 
Generally this is the first, e.g. r7'.v;a man, 
go''xxgxr, (it) is flaming, 'yw'djir to be large ; 
but the accented syllable does not need to be 
the first, e.g. ga'xa'' child, ha'tv'.'t hurry! 
Further, it is evident that the stressed syllable 
may lose its stress, as it reduces its quantity, 
in a given setting in the sentence or when the 
addition of one or more syllables changes the 
rhythmic pattern of its syllables. Thus, beside 
go''xoLgxr, we hive go'xxgx'r,i^oi. (rr) is burning, 
go-yxgilgx (n) is starting to bukn ; beside 
/'fl"'^a TO EAT and a'k'os ti it t'a-'xidigxn i was 
ABOUT TO EAT THIS we have i'l ga t' a'g'r'gyji i 
HAVE FINISHED EATING. These alternations of 

stress may be purely rhythmic phenomena for 
the most part, but I doubt if they are entirely 
so. Functional alternations seem to be illus- 
trated in g(i''yu' SEA : ga'yw' smoke ; q'a''(ij\ 
HAIR OF HEAD : q'a'djx HAIRY. The same -a' 
PROVIDED WITH is fouud in 'madja ochre and 
nia'doL HAVING mountain-sheep, but we find 
also 'a''dja smeared with soapberries. 

The question of stress is complicated by that 
of pitch. The stressed syllable is higher in pitch 
than the other syllables of the word. At the 
same titne it seemed to me that a low-pitched 
syllable might very well bear a secondary stress 
so strong as to bring up the question whether, 
after all, what I have been making stress in 
this paper is not primarily a matter of relatively 
high pitch, only secondarily one of stress. 
Though I should not like to commit myself 
at present, I consider it very likely that pitch 
alternations are a primary factor in the dyna- 
mics of Haida, though it is perfectly clear that 
a given syllable is not uniformly high-pitched 
or low-pitched. The actual musical effect ot 
Haida is so marked and the sonority of the 
more weakly stressed syllables so great that the 
operation of a pitch principle is to he looked 
for. Yet I cannot say that I succeeded in finding 
one beyond doubt. 

A few sentences will have to do for the 
present to illustrate the alternations of pitch 
observed. A high pitch is indicated by an acute 
accent over the vowel (a), a low one by a 
grave accent over the vowel {a). A fiilling tone 
(^) was observed only in 'nid' (exclamation of 
pain). If there are significant pitch alternations 
in Haida, they are probably of the simple Tlingit 
high-low type observed by Boas. In the follow- 
ing e\amples, the un^iarked syllables following 
a high-pitched syllable did not impress the ear 
as being as definitely low-pitched as the one so 
marked. The pitch of unmarked syllables after 
a low-pitched syllable is low. A stressed syllable 
wliosc tone was not markedly high or low is 


VI American Indian Languages 2 


inaicatcd by an acute accent after the vowel (<7'). 
la\i^ii'';a tia-i ge 'la (]'(rt\:ayn-'gy.ii he went 


}:agan iia-i ()•)» go'g^-' my house is on 


'la-0- gligxii na-igo-dxgx' he puts my house 

ON riRE 

hro- la-gd'(a na't go-drgx' i put his house 


dd in ■;d-dx you start the fire! 

girxxgrrigx IT IS BURNING 

gi'rxxgilgy. it is starting to burn 

'vjxsi go-xxgUgxr, that is starting to 


iid-gdi '(d \i IN the house 

ido- nd'i -.-a c'djir^ i am in a house 

nd'gai ^'Z o la qa't'car, i GO into a house 

'Id-Q na-'gai yf q'd'fcxr, he is going into 


nd-iyi 'Ja q'd'fcar, he is going into the 


id ga t'd'gd i am eating 

a'k'os u U t'd-gd i am eating this 

tilgiui u ly. isdxTt it put it away 

'lUTiSga.i sdir, o t'sVwigx those two are 

cedar BOARDS 

tlVwdi U djl'-qgxr, THE CANOE IS LONG 

X'.'l sq'd'sdir, TWO (long and narrow) 


Editorial Notes 

Originally published in International Journal of American Linguistics 2, 

143-158 (1923). 

There are two, highly divergent dialect areas of Haida: Northern Haida, a 
group of closely related dialects spoken in the northern Oueen Charlotte 
Islands, particularly at Masset, and in adjacent parts of Alaska; and Southern 
Haida, a dialect chain, the best attested member of which is Skidegate. Sapir's 
informant spoke Skidegate, which was the focus of Swanton s study (1911) and 
has more recently been the subject of extensive work by Levine (1977). The 
Masset-Alaskan dialects (Lawrence and Leer 1977; Enrico 1986) have a mark- 
edly different phonology from Southern Haida, including phonemic tone and 
two pharyngealized consonants (see Krauss 1979: 839-840). 

Sapir's ms. corrigenda to his copy are as follows (note that we use the char- 
acter g for Sapir's slashed or underlined g): 

Six: Athabaskan and Na-Dene Languages 





144, col. 2, 1. 41 



146, col. 1, 1. 19 



147, col. 2, 1. 16 



147, col. 2, 1. 21 



148, col. 1,1. 11 

n and c 

n and r) 

148, col. 1, 2nd last 



148, col. 2, 1. 36 



148, col. 2, 1. 36 



149, col. 1, 1. 11 



149, col. 1, 1. 16 



149, col. 2, 1. 26 

and t'd'na' 

and Nootka t'ana' 

150, col. 1, 1. 15 



150, col. 1, 1. 36 

nag^ . . . 

na-ga . . . 

150, col. 1, 1. 38 



151, col. 1, 1. 23 



153, col. 1, 1. 14 



153, col. 2, 1. 18 



153, col. 2, 1. 25 



154, col. 1, 2nd last 



154, col. 2, 1. 25 



155, col. 1, 1. 19 



155, col. 2,1. 5 



155, col. 2, 1. 20 



155, col. 2, 1. 23-24 

they remain 

they obey rules which 


156, col. 1,1. 13 

t'e- 's... 


156, col. 1, 1. 20 

i — ; 

i — 

156, col. 1, 1. 21 



156, col. 1, 1. 25 

one band 

one hand 

156, col. 1, 1. 27 



156, col. 1, I. 34 

di-s . . . 

di- s . . . 

156, col. 1, 1. 34 

la-'o-'la- ... 

la-'o- 'la-... 

156, col. 1, 1. 35 



158, col. 2, 1. 2 



158, col. 2, 1. 9 

it put 

I put 




In the summer of 1922 the author of this paper undertook a field 
study of the language of the Sareee Indians, now located at a reserve near 
Calgary, Alberta, Canada. The Sareee and the Kiowa Apache were the 
only Athabaskan (Dene) tribes that descended into the Great Plains and 
assimilated the distinctive buffalo-hunting culture of that area. Both of 
these tribes were affiliated with more powerful neighbors — the Kiowa 
Apache with the Kiowa, the Sareee, in a less formal way, with the three 
Blackfoot tribes. Like other Athabaskan languages, Sareee shows prac- 
tically no influence from neighboring languages of alien stock. The num- 
ber of Blackfoot, Gree, or Stoney (Assiniboine) loanwords in Sareee is 
practically nil, while the morphological influence exerted by these lan- 
guagjes is entirely nil. Sareee is thoroughly Athabaskan in its sound-sys- 
tem, its morphology, and its vocabulary. It has developed distinctive 
dialectic peculiarities, but these do not remove it appreciably from the 
linguistic companionship of Navaho, Hupa, Carrier and Chipewyan. 

Perhaps the most interesting fact that emerged from the summer's study 
w^as the presence in Sareee of a well-developed system of pitch accent or 
tone. The interest of this fact far transcends Sareee itself, for so funda- 
mental is tone to the phonetic and morphological understanding of Sareee 
that it is inconceivable that it should not be shared by the other Atha- 
baskan dialects as well. As I propose to show in a moment, we have 
some good evidence in the literature on this point, though very little 
explicit information has been given on tone in the large body of Atha- 
baskan grammatical, lexical, and text material so far published. That 
Athabaskan is a tone language is of great comparative interest because 
of the existence of tone in Tlingit (S. Alaska), ^ a language to which 
Athabaskan is genetically related, though only remotely so '-. 

1. See F. Boas, Grammatical Notes on the Language of the Tlingit Indians (University 
of Pennsylvania Museum, Anthropological Publications, vol. VIII, n° I, 1917). 

2. See E. Sapir, The Na-dene Languages, a Preliminary Report (American Anthro- 

1 70 ^f American Indian Languages 2 


As early as 1876, Father E. Petitot * remarked of certain apparent 
Chipewyan monosyllabic homonyms: « Tout depend de I'accent ou esprit 
que Ion donne aux consonnes et du ton qui accompa^ne la prononcia- 
tion. Par exemple, sha [i. e. 6a1, prononce sur un ton eleve, signifie 
longtemps, tandis qu'il veut dire martre lorsqu'on le prononce sur un ton 
plus has. » This is clear and suj^gestive enough, but unfortunately Father 
Petitot ignores tone in his grammatical sketch and dictionary . A later 
student of Chipewyan, Father L. Legoff '^-^ is even more explicit. He 
states : ^ « Tout mot, meme toute racine, en montagnais, a sa note ouson 
ton bas ou eleve qui en determine la signification. Et ces mots, ces racines 
conservent ce ton, meme lorsqu'ils s'allient a d'autres mots, k d'autres 
racines, a d'autres elements pour former des mots composes. II faut 
done, de toute necessite, en suivant les conversations, s'^fforcer de saisir 
cette note ou ce ton, afin de ne pas detonner soi-meme en parlant, et de 
ne pas s'exposer ainsi ou k n'etre pas compris, ou a faire entendre tout le 
contraire de ce qu on voudrait dire. Caril y a des mots, et beaucoup, qui 
se prononcent tantot sur un ton bas, et tantotsur un ton eleve, et qui, de 
la sorte, servent a nommer deux choses, Prendre deux idees contradictoires , 
suivant le ton qui accompagne la prononciation. » But Father Legoff too 
contents himself with a general statement and neglects to provide his 
Chipewyan syllables with their appropriate tones. Hence it is impossible 
to gather from his material anything definite as to the tone-system of 
Athabaskan. Here and there iri his grammar he throws out a suggestive 
hint, but that is all. Thus, he notes that nan-[\. e. ng,-] as a verbal prefix 
meaning " across " is pronounced with a high tone, but that as an inde- 
pendent low-toned word it means " thy mother " ^. Again, the verb stem 
kkeih [i. e. -k\^] is high in the present ("imperfective") forms, low in the 
past ("perfective") forms'. 

It is a pity that Fathers Petitot and Legoff, who evidently had consi- 
derable feeling for Chipewyan, did not treat tone with Ihe detail that it 

pologist, N.S., vol, 17, 1915, pp. 535-558). I hope to publish fuller studies on the 
Nadene languages (Athabaskan-Tlingit, Haida) in the future. Since the preliminary 
report was published much new evidence has come to light. 

1. In his monumental Diclionnaire de la Langue Dene-Dindjie, dialectes Montagnais 
ou Chippexuyan, Peaux de Lievre et Loucheux. See p. xiii of the " Avant-Propos. " 

2. See his Grammaire de la Langue Montagnaise, Montreal, 1889. The French mis- 
sionaries use the term " Montagnais " for the more current Chipewyan. This Atha- 
baskan " Montagnais " is not to be confused with the Algonkian Montagnais, of the 
lower St. Lawrence. 

3. Op. cit., pp. 10, 11. 

4. Op. cit., p. 120. 

5. Op. cit., p. 321. 

Six: Athabaskan and Na-Dene Languages 171 


deserves. As it is, most other recorders of Athabaskan have either over- 
looked it or made light of it, so that at this late day it probably comes 
as a surprise to most linguists that Athabaskan is indeed a tone lan- 
guage. Father A. G. Morice, our authority for Carrier, is not entirely 
unaware of the existence of the tone problem. " Change of meaning by 
intonation or vocal inflection ", he remarks, "obtains also among some 
— not all — of the Dene tribes. Some of these intonations are even 
proper to fractions of tribes only. Thus ya which means 'sky' in almost 
all the dialects becomes 'louse' to a Southern Carrier when pronounced 
in a higher tone. Northern Carriers have another vocal inflection which is 
combined with the final hiatus and is also peculiar to them ". ' This pas- 
sage suggests dialectic variability in tone. As Father Morice gives no fur- 
ther information on Carrier tone, I asked my colleague Mr. C. M. Bar- 
beau, who was carrying on ethnological researches in 1921 among the Git- 
kshan Indians near Hazelton, B. C, to try to secure some linguistic 
material from the neighboring Carrier Indians of Hagwelgate. Mr. Bar- 
beau kindly furnished me wnth a small amount of Carrier material in 
which the attempt was made to note the high tones. Very suggestive. 
among others, are the following forms : - 


I am eating 


you are eating 


eat ye ! 


I have eaten 


he has eaten 

The tones of these forms are no doubt accurate in the main, for they 
have close Sarcee analogues. Sarcee too has a large number of verb stems 
whose present or "imperfective" form is high-toned, perfective low-toned 
(Father LegofT's Chipewyan -k'io : -k'ki and Mr. Barbeau's Carrier -'af : 
-'d'l correspond to the Sarcee pattern -ts^os : -ts'o':( " to handle a 
cloth-like object '" ; -^ii : :(l- I "to be warm" ; - k^as: -ka-^ " to be cold 
weather " ; -la:-la "to do"; -tc'o : -Ic'o "to take, get hold of"). Again, 
the high tone on the first syllable of "you are eating" is undoubtedly 
correct, for in Sarcee, as we shall see, and in other Athabaskan dialects 
as well, the second person singular of the verb frequently demands a 
high tone on the syllable preceding the stem. 

i. A. G. Morice, The Dene Languages (Transactions of the Canadian Institute, 
vol. I, 1889-90, pp. 170-212; see p. 182). In Sarcee, " louse " [ya] has the low tone 
with a final glottal stop, as in Northern Carrier. 

2. ' = high tone; ' = low tone ; •=: long vowel. 

172 ^f American Indian Languages 2 


Dr. P. E. Goddard's extensive and fundamental researches in Athabas- 
kan linguistics (comprising materials for Hupa, Kato, Chipewyan, Sar- 
cee, Beaver, Apache, and other dialects) nowhere include a treatment of 
tone '. Dr. Goddard has recently informed me, however, that upwards 
of lifteen years ago, when working on Rousselot tracings of Hupa words, 
he found that in apparently homonymous forms for the second and third 
person singular .(e. g. fczn:r,yai "you have gone out" or "he has gone 
out" ; mtndxt "you are finishing" or "he is finishing' ~) the second 
person has a higher tone on the .syllable preceding the verb stem than the 
tliird person. This important observation of Dr. Goddard's has apparent- 
ly not been published before. 

The Franciscan Fathers have given us an excellent dictionary of Na- 
vaho, 3 but not a word is said about tone in their introductory gram- 
matical notes. However, they are careful to mark the main stress of 
each word. It seems very likely that many, perhaps most, of these stress- 
ed syllabes are really high-toned (the English and German speech habit 
associates greater stress w^ith higher tone and tends to consider the form- 
er the primary factor). It is remarkable that in a great many verbs the 
stem is stressed in the present, unstressed in the perfective. These stress 
alternations are analogous to, and probably identical with, the tonal 
alternations of Sarcee (and of Chipewyan and Carrier) verb stems. To 
take but one instance, the Sarcee alternation of imperfective -ts'os with 
perfective -/5'o';{ is clearly the same phonetic and morphological fact as 
the Navaho alternation of stressed -ts^os (present) with unstressed -ts'o's 
(perfective), e. g. namsts^o's ''I give you (a buckskin)": na'n'ils^o's ^'' I 
have given you (a buckskin) ;" xflj^fj'o'j "I take out(a blanket)" : xa'its''o'S 
"I have taken out (a blanket)" *. Again, the verbal paradigms given by 
the Franciscan Fathers indicate that, as in Sarcee and Hupa, the second 
person singular sometimes dilTers from the homonymous third person 
in bearing a stress (i. e. high tone) on the syllable preceding the verb 

1. For a passing reference to the possible significance of tone in Chipewyan, see 
P. K. Goddard, Analysis of Cold Lake Dialect, Chipewiyan (Antliropological Papers of 
the American Museum of Natural History, vol. X, pt. 2, 1912), p. 83. 

2. See P. E. Goddanl, The Morphology of the Hiit>a Language (University of Cali- 
fornia Publications in American Archaeology and Klhnology , vol. 3, 1905), pp. 117, 


3. The Franciscan Fathers, A Vocabulary of the Navaho Language, 2 vols., Saint 
Michaels, Arizona, 1912. 

4. Op. cit., vol. I, sub" give " and " lake ". I have adapted the orthography of 
the Franciscan Fatiiers to my own. In my orthography accents following vowels are 
indicative of stress; wlien placed on the vowel, of tone. 

Six: Athabaskan and Na-Dene Languages 173 


stem, e. g. i'yi "you marry": iyt "he marries;" txadi'gis "you wash 
yourself" : txa'digis "he washes himself" *. 

I shall now proceed to a brief discussion of the Sarcee tone system. It 
is impossible in a brief compass to do justice to the subject, which is 
diflicult and involved. Sarcee tone cannot, properly speaking, be discuss- 
ed as an isolated chapter of the phonology. It interpenetrates each and 
every aspect of Sarcee morphology. None of the elements of the language, 
whether radical, prefixed, or suflixed, can be accurately understood with- 
out a consideration of its tone and of the tone shifts which it undergoes 
in combination with other elements. All that I can attempt here is a 
preliminary treatment, designed to give some insight into the tonal 
peculiarities of an Athabaskan language. 1 hope to give a full and system- 
atic study of Sarcee tone in a later work on the phonologv and mor- 
phology of this language. My method at the present time will be, in 
the main, to list and discuss a series of apparently homonymous sets of 
forms. This informal and inductive method will probably convey more 
to a public necessarily but little acquainted with Athabaskan morpho- 
logy than a more compact and generalized style of treatment. It will be 
convenient at times to refer to the general or reconstructed Athabas- 
kan form of a word. Aly reasons for the reconstructions - can hardly be 
given here but must be reserved for later comparative studies. In any 
event, the Athabaskan dialects do not dilTer greatly on most points; the 
degree of ditTerentiation is comparable to that of Romance, Slavic, Semi- 
tic, or Bantu rather than of Germanic, Indo-Iranian, or Hamitic. 


A few preliminary remarks on Athabaskan phonetics are required. The 
Sarcee sound system does not reflect the original Athabaskan system as 
well as do those of Chipewyan, Hupa, and Navaho. There have not only 
been important shifts of vowels and consonants, but levelings of certain 
consonants (e. g. the old ^-sounds and r-sounds ^ have combined into a 
single i-series, the Sarcee c-series being, in the main, of anterior 

1. Op.cit., vol. I, pp. 13, 14. 

2. In my comparative Athabaskan researclics I have used chiefly the following 
dialects and authorities : Chipewyan (Petitot, Logoff, Goddard); Hare (Petitot) 
Loucheux (Petitot); Beaver (Goddard); Sarcee ((ioddard, Sapir); Kulchin (Sapir) 
Ten'a of central Alaska (Jetle^ ; Ten'a of Anvik, hjwer Yukon ((-liapnian, Sapir) 
Carrier (Morice) ; Hupa (Goddard) ; Kato (Goddard) ; Chasta Costa (Sapir) ; Navaho 
(Franciscan Fathers) ; and Jicarilla Apache and Apache proper (Goddard). 

3. My c is Lepsius' s. 

174 yf American Indian Languages 2 


palatal origin ; Athabaskan y and y are distinguished only before origi- 
nal a) and vowels (e. g. old i and i, the pepet vowel, have become 
leveled to i) ; weakenings or- disappearances of certain syllabically final 
consonants (e. g. -n drops if preceded by a short vowel, which was pro- 
bably nasalized at one time ; -g and -d are weakened to -', the glottal 
stop), with resulting interchange of lost, or weakened, and retained con- 
sonant (e g. Ath. *-en : *en-C> Sarcee-i : -in-a ; Alh.* -e^: *-td-e^ Sarcee 
-a': -ad-i) ; and contraction of vowels (e. g. Ath. -ai-, -e'e-, -e'a-, -fe-, 
-ee- may all now appear in Sarcee as -a'"-, see below). Once the phonetic 
changes peculiar to Sarcee have been worked out, however, it is seen 
that this dialect is a perfectly regular, if somewhat disturbed, represen 
tative of its prototype. 

The Sarcee sounds are : 

I. Vowels : 

a. Short: a (a), fl, i, u {o) 

b. Long (or half-long) : a' , a' , i", o' {w) 

c. Over-long (with glide-like rearticulations) : 

a-", a-', '.••■, o-" («•«) 

d. Diphthongs (<original vowel -|- -y, -7): 

a{-)i, a{-)i, u[-)i 

II. CoiNsoNANTs : 

a. Stops and AfFricatives : 

Intermediates : b (very rarely), d, g, d^, dj 
Aspirated Surds : t' , k\ is\ tc' 
Glottalized Surds : /', k\ t's, t'c, 

h . Spirants : 

Voiceless: s, c., x, x, x, ^w, h (-') 
Voiced: i, y, i, 7-, y, T^ "^ (t"') 

c. Laterals: /, /, dl, ti, fi 

d. Nasals : m, n 

Of the vowels, a is a duller form of short a and is pronounced very much 
like u of English but ; a is a velarized, dark-timbred a with 7^-glide after 
preceding stopped consonants (aside from'), aflricatives (including J/, 
ll, H), and sibilant spirants (5, :(.'c, j)\ t is open, as in English it; u, 
open as in English put, varying with close (French eau). Of the two a- 
sounds, which are difficult to distinguish only in certain positions and 
which must be clearly kept apart in theory, a generally corresponds to 
Ath. e (open) and to Ath. a in certain cases, g. nearly always to Ath. a 

Six: Athabaskan and Na-Dene Languages 175 


(e. g. Sarcee -^'fl " to be so"<Ath. * -/'e, Sarcee fa' " feather, wing " 
<Ath.* fax). Vocalic series c. is particularly common in open syllables 
and results either from contraction (e. g. sa"'thk'a " my brother or sister " 
<isi-ati:k'a) or from short vowel -(- syllabically final -/- (e. g. xa-'^m " it - 
tastes " <i xa-l-ni, cf. Nav. ha-l-ni). 

The " intermediate " stops [b, d, g) and the stopped part of the inter- 
mediate afTricatives [d^, dj, dl) are essentially voiceless, or voiced only at 
the moment of release, unaspirated, and of lesser stress than the normal 
voiceless stops of French or Slavic ; they probably correspond to the 
" tonlose Medien " of many German dialects. The aspirated surds are 
more strongly aspirated than the voiceless stops of English. The glotta- 
lized surds, afTricatives, and lateral (/', k\ fsyfc, ^7) are pronounced with 
synchronous oral and glottal (or epiglottal) closure and with a sudden oral 
release which precedes by a perceptible moment of time the release of 
the closure in the larynx. The intermediates and aspirated surds seem 
respectively identical to the surds {t,-k, ts, tc,) and aspirated surds (/', k\ 
/5', tc'), of Chinese; the glottalized consonants (sometimes known among 
Americanists as " fortes ") are peculiar to many aboriginal languages of 

In this system c and ; are used to indicate the voiceless and voiced 
sibilants of French choix {Eng\is\shoe) andjoie; dj, tc\ /V are the correspond- 
ing affricatives of the series. The voiceless series of spirants x-, x, 
X, 'zv and, correspondingly, the voiced series y [y), y, y, w/ [y^) are 
conditioned by the preceding vowel, x (as in German ich) and y (close 
to English y, yet not strictly identical with it before i, where it is more 
definitely spirantal than English y ; I have written y, however, as there is 
no possibility of confusion) occur before i, y also before a; x and y (mid- 
guttural spirants) occur before a ; x and v (velar spirants) before a; 'w 
(approximately like English wh inwhen as pronounced in America, but with 
more nearly spirantal : ^w, almost xtv) and tv (properly, labialized y : 
Y"' or yw) before (and after) u. i represents a voiceless spirantal /, much 
like Welsh //. 

In pronouncing Sarcee, or any Athabaskan language, it is important 
to give each syllable its due weight. Ihe syllable as such has a more 
clear-cut phonetic distinctness and a more individual morphological signi- 
ficance than the syllable of English or French, even if it is not always 
possible to assign it a concrete meaning. The syllables of a word do not 
generally differ greatly in stress, though the stress is of course not uni- 
form. Certain elements, like third person plural gi- " they " and the 
radical syllable of the verb, are more strongly stressed than others (e. g. 
iuk'agits'a' '■ they cease to cry ", ^i- and the stem -ts'a' are more strong- 

176 yi American Indian Languages 2 


ly stressed than k'a- '' off "> " to cease ", which, however, must not 
be slurred over). Variations of stress do not seem to be coordinated with 
differences of tone. 

Sarcee syllables are of five types : 
a. Open syllables ending in a short vowel (e. g. wi, t'a). 
h. Closed syllables with short vowel followed by a voiceless consonant 

or, far more rarely, by two voiceless consonants (e. g. nd, Va\ the 

only finals allowed are -t\ -is', -tc' , -', -s, -c, -x,-\ -i, -st\ -ctc\ -lie' \ 

more rarely -k\ -t's, and -ft). 

c. Heavy syllables w^ith long, or half-long, vowel followed by a voiced 
consonant (e. g. ni'l, t''a'^\ the only finals allowed are -:(, -/, -/, -« ; 
historically speaking, syllables in -r [<C Ath. *-r7, *-i"y, *-£>', *-^y]> ^" 
-0' (<C Ath. *-o'y), and in -a'i, -a'i, -wz [< Ath. e' , a', o , -\ — y or 
-y,J also belong here) 

d. Heavy open syllables with over-long vow-el (e. g. «■.•', /'d"") result- 
ing from contraction of vowels or reduction of type a -(- /. 

e. Heavy syllables with over-long vowel followed b}^ a voiceless conson- 
ant (e. g. m''s, fa'^s) resulting from contraction of type a -|- type b. 
The first three types of syllables are old Athabaskan types, the last 

two are largely of dialectic origin. 

The point of syllabic division of a non-final, open syllable, particu- 
larly if the vowel is short, lies /// the following consonant, which 
therebv becomes greminated. Thus, the forrh k'>£'Js^a'' jjiven above is 
to be read k' xg' \.f s'' a'' {-g' zzz -kg-, i. e. unaspirated yoiceless ^releasingin 
intermediate g). We shall not indicate these purely mechanical gemina- 
tions. A final glottal stop, whether an old Athabaskan element or of Sar- 
cee origin, is absorbed in the following geminated consonant, e. g. 
a"'sdiy'.nC " he has said to him" -j- inferential particle la becomes 
a'^sd'-j'-U'd'a. There are other important rules of sandhi which we must 
neglect here. 


The tone system followed in this paper is that used by Father Schmidt 
in " Anlhropos " and by Westermann and other Africanists. In this sys- 
tem a indicates a high tone, a a low tone, and a a tone of intermediate 
pitch. The intlections are indicated by combinations of these symbols : a 
represents a tone falling from high to low, a falling from high to middle, 
a falling from middle to low, a rising from low to high, a rising from 
low to middle, a rising from middle to high. In writing Sarcee I have 
found it convenient to leave the middle tone unmarked [a = <i) ; my a 

Six: Athabaskan and Na-Dene Languages 177 


therefore does not mean an indeterminate or unknown lone, but one 
that is midway between a and a. 

The primary Athabaskan tone system may prove to be one of register 
alone, possibly recognizing but two grades, high and low. In Sarcee, 
however, it was found absolutely necessary to distinguish also a middle 
tone and inflected tones. The middle tone may have originated as a pho- 
netic compromise betw^een two conflicting tone principles, one of which 
demanded that the syllable receive a high tone, the other that it be pro- 
nounced on a low tone. However that may be, there are many verbs 
w^hich require that in certain aspects the radical element be pronounced 
on a middle tone, e. g. dlsi'ya " I have gone ", in which the -ya " to 
go " falls from the preceding high syllable but does not drop to the low 
tone of a syllable like -tia in dstla '' 1 have done ". With discya con- 
trast also dlc'd " 1 shall go " <C *d\-s-yd. Here the -yd has a definitely high 
tone, maintained throughout the " imperfective " forms (cf. di'yd " he 
will go ", hut di'ya " he has gone "), 

The syllables of the language belong to two tonal categories. Certain 
syllables have a fixed or inherent tone value of their own ; e. g., '('/ 
" pine bough with needles ", xdi "' burden ", dxi '' crane ", mii " moth, 
sleep ", ts'in " dirt ", tii " dog ", ts'i " red clay ", ts'd " rock ", t'6 
" water "', ts'Vl " snowdrift ", ^di " saliva ", Id " sore ", -dd'l " seve- 
ral go " (perfective), -t^d'^ " two go " (perf.), -yd " one goes 
(imperfective), -i " to act " (imperf.), nd- " again, repeatedly ", k'u- 
"■ inside, into ", and d- (demonstrative prefix with verbs of doing, being, 
becoming, and saying) have a fixed high tone ; t'cis " powdered char- 
coal ", {sl-:{ " a boil ", tlifs " clay ", k'ii " fire ", th " smoke ", mas 
" hoop ", did' " laughter ", ya" " louse ", wa' " war-raid ", did' " gum ", 
so " star ", ts'a " ordure ", md " snare ", -ddl " several go " (imperf.), 
-Cas " two go " (imperf.), -In " to act " (perf.), d^nd. da- (prefix indicat- 
ing distributive subject of verb) have a fixed low tone ; and -la (inferent- 
ial sufTix), -V' (suffix of relative clause implying absence of subject), -ya 
" to go " (perf.), -^t. " to call " (continuative), -k\' " to throw a cloth- 
like object " (perf.), and -w. " to use the hand '' (imperf.) have a fixed 
middle tone. Other syllables have no inherent tone of their own, but adopt 
a tone in a particular form in accordance with certain rules of contrast 
or of morphology. These syllables have no assignable tone as such, but 
this does not mean that they do not possess a defined tone in a given word. 
On the contrary, the intonation of syllables with variable tone is impor- 
tant, but the rules governing their tone ar£ not always easy to define. 
Thus, such elements as perfective-durative .ri-, perfective yt-, and possess- 
ive-objective s\- " my, me ", vary their tones from word to word, but 
not optionally. This is illustrated by the three forms : 

178 ^l American Indian Languages 2 


sit'\ " the calf of my \e^ 

sit^i " you are lying- (in position) 

ji/'i " he is lying 

The syllables -/'I " calf of the leg " and -t'i " to lie " (perf.) have inhe- 
rent tones, low and high respectively. The syllables ji- " my " of the 
first form and si- (perfective-durative) of the other two have no tones of 
their own. In ji/'i the high tone of 5-!- " my " is due to the contrastive 
influence of the low tone of the radical element (examples of other tones 
for 51- " my " are sr(dha " my grease ", s\la " my brother [woman speak- 
ing] ", contrast ji/^' " my hand "). In sit^i the high tone of perfective- 
durative si- is due to the fact that the subject of the verb is a second per- 
son singular. The old personal prefix *-«- (Ath. *se-n-^ Sar. si-), which 
has been lost in Sarcee, seems regularly to have required a preceding- high 
or, in other forms, relatively hig-h, say middle, tone, although the syno- 
nymous prefix *ne- (> Sar. m-), which is used in certain forms, is itself 
variable in tone (e. g. amid " you do ", nits\' " you are crying"). In 
sit^i the si- is just low enough to bring oiit the inherent high tone of the 
stem -t'i. Just what tone a variable syllable takes cannot be foretold on 
the basis of simple rules ; the various syllables of this type do not all fol- 
low identical patterns. The tones so far illustrated belong, then, to three 
categ^ories : inherent or fixed tones, tones controlled by definite grammatical 
requirements, and tones controlled by contrastive requirements. It is 
naturally impossible to draw a hard and fast line between the last two 

All, or nearly all, primary syllables have a level tone. This applies to 
syllables with a long vowel as well as to short-voweled syllables; e. g., 
in forms like n'lts^i' " you are crying " or gdisyc:{ " they have run olT " 
OT wiictc^o' " leg-big " or yi'an " it has lain " the long vowel of -ts^c, 
-)'i';(, ~tc'd\ -a'n must be pronounced on a fixed register — middle, high, 
or low — without the slightest upward or dow^nward inflection. The 
numerous inflected tones ^ of the language arise from the contraction of 
two, rarely three, vowels of different registers. The theory of these 
inflected tones is simple enough. The vowels melt into a single long 
(over-long) vowel without losing their tones, which tlius combine into 
inflections whose tendency and range is determined by their component 
elements. Hence, a -\- a becomes a' (half-falling from high level), a -\- 

{. Yet far less nnmerous Ihan in Chinese or in certain other American lang-nages, 
e. g. Takelma, in which we have primary rising and falling inflections. The quick 
change from level lo level, with comparatively few slurs, is highly characteristic of 
the general How of Sarcee speech. 

Six: Athabaskan and Na-Dene Languages 179 


a becomes d' (fuU-fallingrfrom high level), a -\- a becomes a' ( half- falling- 
from middle level), a -\- a becomes a' (half-rising from middle level), 
a -\- d becomes a' (full-rising from lo^v level, a -\- a becomes a' (half- 
rising from low level) '. If the contracting vowels have the same tone 
there results a long (over-long) vowel of level tone ((i", a' ", or a"); 
long level-toned vowels of the same phonetic type result also from vowel 
-|- /. Only a few examples are needed to illustrate these tonal rules : 

/Wi isfscx, blanket I-wear >> t'sidi'sfsix^ 

k^d adi moccasin without, barefoot>» k^a'di 

nd-\-c-atc again-it (indef.)-I-keep-handling-one-object, I handle a 
certain object time and again > nd'cqlc 

Si-l^una my sickness>> sd':(i'ma 

da'" Ictc'o here I-seize-it>> da'ctc^o 

gu-d-s-tid thus-I-do> giva'stid 

\-nl-i-n'.-i it (indef.) -he-admires-the, the one v^ho admires^ Inl'nl: 

k^u dstid fire(wood) I-do, I make a rire>> k'wa'stid (note disappear- 
ance of -' in crasis) 

n'.sgdywlsl' dnd downward he-becomes, he stoops^ nisgd^(iu\sa"nd 

mit'uwa iss'd his-water I shall-heat > mit'-iava-ss'd 

t'd:(d-d-yi-s-\ grief-thus-him-treats, he is in mourning>- f'^i^a'^yu't 

t's4-i-s-sl noise-he-is-sounding, crow>> fsd'ssi 

t^u-^{w\-\-a water-in-it (indef.) -he-handles-one-object, he gets water 

The only certain example of an inflected tone in Sarcee that I know of 
which cannot be analyzed into two simple tones is the low falling tone, 
-a'-, of the first person plural subjective, e. g. gumca'Cd'j " we suckle 
them", nxsa'io' " we get moist ". Full historical knowledge would proba- 
bly indicate that this element is contracted from two simpler elements 
of middle tone and low tone respectively. 

It goes without saying that the high, middle, and low levels have no 
absolute fixation. Nor are they definitely fixed relatively to each other. 
The interval between high and middle or between middle and low or 
between high and low varies widely according to the care or emphasis of 
articulation. What under some circumstances constitutes a movement from 
high to low may under others be equivalent to no more than a half fall. 

1. These formulae do nol refer to vocalic quality. The quality of the contracted 
vowel is dependent on a number of factors, which do not concern us here. 

2. As several of these examples show, the rules of contraction apply not only within 
the " word " (a somewhat elusive concept in Athabaskan) but in sandhi between 
words (" crasis "). 

180 VI American Indian Languages 2 


Each level and each inflection must be judged or intuitively felt in the 
context of the preceding and following tones. Now one of the most cha- 
racteristic and also one of the most puzzling things about the Sarcee tone 
systtMU in practice is the tendency, within a phonetic phrase or breath 
group, to gradually lower the absolute register. This means that a high 
tone following on a low or middle tone that is itself preceded by a high 
tone tends to be a little lower than the first high tone and to be perceived 
by the ear as a middle tone. Many of the ''middle " tones that I have 
recorded in running text or in specimen sentences or even within the 
single word are really pseudo-middle tones — " dropped high tones ", 
as we might call them. I feel that the}' are not normally identical with 
true middle tones but keep above the normal middle register. The form 
dn\ld " you do " already quoted was often recorded as dnda, with a 
secondary middle tone, instead of the theoretically required high tone, 
on the last syllable, yet, unless my ear deceives me, the tone cadence of 
this word is not identical with that of words like dit'ibdja' " it is flabby ", 
diszuuga' '' it is velvety ", nUc^it'ca' " it is tiny ", in which the diminut- 
ive -a' falls regularly on the middle tone (cf. , with another tone cadence, 
nidot'ia' " it is light in weight ", S'.sowa' " it is a little sour "). In other 
words, the " dropped high " really falls between the high and the true 
middle level. Where there is an immediately adjacent middle tone to 
serve as a standard of comparison, this difference can frequently be direct- 
ly perceived. In the phonetic group ayd t^unihi ^' the one who wrings 
out water " the cadence of the second word is not adequately suggested 
by the orthography. In making the record I noted that the group t^umhi 
was really t'linihi on a high register (i. e., not equivalent in cadence to 
forms like dnlla with true low -n\). In fiimhi, in other w^ords, the -m is a 
true middle and the -hi a " dropped high ". The analysis is not difficult 
here, because we have an immediately preceding -|*a with a clearly mark- 
ed low tone to serve as gauge ; moreover, we have the analogy of other 
related forms in middle-tone -m and of hundreds of analogous relative (or 
participial) forms in -L 

A good example of the progressive fall in register is the sentence 
k'iyldji. ytca '* coat in-I-go, I put on a coat ", in which i indicates a lower 
tone than the normal low tone fixed by the preceding wnrd. The theore- 
tically correct form of the sentence is k\yidji y\cd, but actually no two 
tones are quite alike in it. Neither of the " middle " tones as the sen- 
tence was recorded {-dji and -c'(i) is a true middle, but merely a " dropp- 
ed high ", the second a trifle lower than the first. Thus, the three high 
tones of the sentence are graduated downward, and the two low tones 
are also graduated downward in a parallel curve. Functional register 

Six: Athabaskan and Na-Dene Languages 181 


chang-es are compounded, in other words, with a rhetorical register change. 
Just as there is a " dropped high " tone, there ia also not rarely a " dropp- 
ed middle " (or secondary low) tone, further a " raised middle " or 
secondary high tone. These niceties can hardly be properly studied with 
the unaided ear. They require the assistance of mechanical devices^. In 
any event, how^ever, it Avould be a great mistake to lay too much stress 
on variability of tone in Sarcee. It is probably no more disturbing a 
factor in the understanding of the basic tone theory of the language than 
in Chinese or any other tone language, each of which has its special laws 
or tendencies of tone sandhi. 

Tone differences in otherwise identical words. 

We shall devote the rest of this paper to a brief analysis of a number 
of cases of forms which dilfer only in tonal respects. In this way we 
shall gain a cumulative idea of the importance of tone in the study of 
Sarcee structure. In many cases the forms in question originally dilTered 
also in other respects and fell together completely in the consonantal and 
vocalic framework because of the operation of various phonetic laws. 
Thus tone was left as the last reflex of outward distinction between such 
forms. One must beware of makinsr the error, which is so commonlv 
made with Chinese and other tone languages that possess numerous near- 
homonyms, of supposing that the tone differences arose for the purpose oi 
keeping apart words which would otherwise be indistinguishable. As a 
matter of fact, perfectly analogous tone distinctions prevail in groups of 
non-homonymous forms, w^here tone is not " necessary " as a gran^mati- 
cal device. F'urthermore, there are many examples in Sarcee of true 
homonyms, of two or more w^ords which belong to distinct form cate- 
gories but are identical in all phonetic respects, tone included. 

1 . a. -^a'^dd'l he has eaten the berries 
b. -^a'dd'l we have eaten the berries 

Form a. goes back to *^alddl, in Athabaskan terms *ys.-l-di'l, '' in 

1. I hope to secu e phonograpliic samples of ottier Athabaskan dialects, so as to 
be able to work out some of the more elusive tone problems -with the help of strictly 
objective material. [Since this article was written such material has been secured 
from a Kulchin informant]. 

2. These Athabaskan reconstructions, it is hardly necessary to say, do not imply 
that we are certain that precisely such forms were in early use, but merely that, 
if they existed, tiiey probably had such a phonetic form. The reconstructions are 
practically Athabaskan formulae for the dialectic forms. 

182 y^f American Indian Languages 2 


which *)'£- is contracted from *yc-ye'. Form b. g-oes back to *'(a'ldd'l, in 
Ath. terms *-(-€'-l-di'l (perhaps contracted from *'^e-\-l-di'l). -l-dd'l is 
perfective of imperfective -l-dal " to eat berries". In form a. *^a- <i*yt- 
consists of third person objective *ye- (Sar. y\.-) and perfective prefix *--^e- 
(Sar. -yi-) ; as usual, forms without specific subjective pronominal prefix 
are used as definite 3 d personal forms. In form b. -^-a'- consists of per- 
fective y(^)- and first person plural subjective -rt"-. Transitive verbs with 
first or second person or indefinite person as subject have no expressed, 
but clearly implied, third personal definite object in many forms, but in 
verbs with definite third person as subject third personal definite objects 
must be expressed. All that remains in Sarcee of the marked phonetic 
difference between Ath. *ye-(eldi'l and *'(eildi'l is the fact that in the latter 
form the voice sinks by a slig^ht interval, perhaps no more than a semi- 
tone, before leaping to the high- toned verb stem. 

2. a. k^u'sdal I have many (e. g. buffalo) coming in 

b. k'ii'sdai he has many coming in 

c. k'u'^sdai you have many coming in 

Form a. is contracted from k'ulsdai, Ath. *k'iuen-'e-c-i-dki, literally : in 
(k^wen-) l{c-) cause {i-) a certain one [or certain ones, 'e-, indefinite object) 
to go in a group [-dki). -dai is imperfective, -dd'l perfec1,ive ; cf. no. 1. 
Form b. reconstructs to Ath. *k^iuen-e-i-dzi. In Old Sarcee, the older form 
of the language still spoken by the oldest men and women of the tribe, 
forms a. and b. are still distinguished SiS k^u'sddi and k'H'idai (contracted 
form k'uddal) respectively. In the former of these forms the element -i- 
dropped out after the sibilant, as in so many Ath. dialects (e. g. Navaho 
c- for theoretical c-i-\ Ghasta Costa has c-i- but -I- may have been restor- 
ed because of the analogy of the other forms). In New^ Sarcee, as spoken 
by my chief informant and most of the other Indians, syllabically final 
-I- has shifted to -s- ; hence in form a. -s- represents older Sarcee -s- 
<i-c- <C-c-i-, but in form b. -s- represents older -/-. Form c. goes back to 
Ath. *k'wen-e-n-l-dii. 

Syllabically final -n- (here "thou") has disappeared after a short vowel 
(there was doubtless nasalization of the preceding vowel, oldest Sarcee 
*-i-, as in Chipewyan, before all trace of the old -n- was lost) and, as 
in a. and b., m + i or older u + ^has contracted to an over-long M-vowel. 
k^u- has fixed high tone. Objective -i- is variable in tone; in c. it takes 
the high tone because of the second person singular subject (cf. p. 194). 

3. a. m-^i he will grow up 

b. ni-^i. you will grow up 

c. niYa he has grown up 

Six: Athabaskan and Na-Dene Languages 1 83 


Forms a. and b. correspond to Ath. *ne-yi and *ne-n-yi respectively ; 
cf. nic^(7. '^ I shall grow up ". Form c. is a perfective in -i-, cf. nl^iY^^ 
" you have grown up, " nlcu:vi (-^ic- assimilated from 5 u:- ') " 1 have 
grown up ". Moreover, -vi of c. has lost a final -n (contrast ni-^in-i 
"the one w^ho has grown up" with n'.-yoL-hl " the one who will grow up"). 
Hence c. reconstructs to *ne-z-yin. Perfective *-yin, however, is modified 
from the normal Ath. *-ydn by the analogy of imperfective *-yi (cf. Chi- 
pewyan imperfective -yt : perfective -yan) ; the older *-ydn is preserved in 
Sarcee in the causative : -c-ca <^*-i-ydn " to cause to grow up, to raise "'. 
Note that imperfective -yi, perfective -yi (n-) belongs to the type of verb 
stem with fixed tone in all aspects ; cf. also continuative —/di. The prefix 
ni-, however, of variable tone, is only one step lower than the stem in 
the imperfective forms, but two steps lower in the perfective forms 
(contrast nic^fi " I shall grow up " with nlcic^^i " I have grown up '"l. In 
other words, even where the stem is invariable in tone, other elements 
in the verb complex are not necessarily treated in a uniform tonal manner 
in the varying aspects. It is quite probable that originally the difference 
between m- of a. and m- of c. was purely mechanical. The aspect prefix 
{*se-*ye-) tends to adopt the middle tone and, if the stem is high-toned, 
a prefix of variable tone which precedes the aspect prefix generally 
takes the low^ tone (the cadence nlcic-^fi is very typical). It is likely, there- 
fore, that form c. is reduced from an old trisyllabic Ath. form of type 
*ne-se-ydn, while a. goes back to a disyllabic form *ne-yi. Now, in Sarcee, 
where forms a. and c. have come to have the same syllabic structure, 
the difference of tonal approach to the stem reflects the old difference in 
phonetic build. 

4. a. gof^ddlmsts'C I shall punch a hole (.with an awl) 
b. goy'^'ddlnists^C I have punched a hole 

5. a. goy'^ddini''ts^i you will punch a hole 

b. goy'^ddim''ts\'' you have punched a hole 

6. a. goy'^'ddl"ts^C he will punch a hole 
b. goy'^'ddl'ts'i' he has punched a hole 

In these forms the momentaneous aspective prefix m- is used in both 
perfective and imperfective forms, so that the only dilference between the 

1. In perfective forms in which the pronominal subject " I " is directly follow- 
ed by the stem or by the element -/-, Ath. * -c- is regularly replaced by *-i-, reflexes 
of which occur in most of the dialects, but in certain dialects, such as Sarcee and 
Chasla Costa, the * -c- is used by analogy with the imperfective forms. 

184 VI American Indian Languages 2 

two aspects, for each of the three persons, is in the tonal treatment of 
the syllable or syllables following the first three syllables {-yd- " hole, 
through '', labialized to -y'^d- because of preceding o, has a fixed high 
tone, to which the variable elements go- " it, indefinite locality " and ^t- 
accommodate themselves). In 5 and 6 the i" is due to the loss of an old 
/ after the i, in 4 the / dropped out after s without leaving a trace. The 
element -w. <C *-ne- became reduced to -«- in 6 and eventually disappeared. 
Aside from the first two syllables, these forms reconstruct to Ath. : 

4 . a . *de-ne-c-l-ts*ed 
b. *de-nt-c-l-is''ed 

5 . a . *de-ne-n-l-ts'ed 
b . *de-ne-n-l-ts^ed 

6. a. *de-n-l-ts'ed 
b. *de-n-l-ts'ed 

In 4 the prefix -ne keeps one step below the stem, hence middle tone 
in a., low tone in b. In 5 the -ne- needs the high tone because of the 
second person singular subject. The verb stem, which means " to 
handle a pointed object (like an awl) », is *-ts^ed in the imperfective, 
*-ts^ed in the perfective (the final -d, reduced to -' in Sarcee, reappears 
in certain forms, e. g. goy'^'dd'c'ts'idi. " the one who will punch a hole " 
g(yf"ddi''ts''idi " the one who has punched a hole '). It is probably a 
denominative in -d from the noun stem *ts'i (Sarcee ts'd) " stone '". 
There is evidence to show^ that the imperfective aspect uses the more 
primary form of the verb stem and that the stem forms appearing in 
other aspects are secondary formations. It is therefore no accident, in all 
probability, that imperfective *-ts"ed is identical in tone with *ts'i. It 
should be remarked that a large number of the Athabaskan verb stems 
are either identical with noun stems or clearly derived from them. 

7. a. iditi'jl I shall hide myself 
b. idinis'i he hid himself 

Form a. uses the imperfective form of the verb stem, -l, " to hide ; 
form b. , the perfective form, -'i ; -/- has dropped out after -s- in both a and 
b. The reconstructed forms are *ede-ne-c-l-\n and *ede-ne-s-l-'hi. The third 
person corresponding to a. is idin'/'\ <C *ede-ne-l-ln ; the first person 
corresponding to b. is idimsts'i <C *ede-ne-se-c-l-' iji . Hence -s- of a. is a 
pronominal element, while -s- of b. is an aspective prefix. With the 
reflexive prefix [di- as well as with the perfective stem - { contrasts the 

Six: Athabaskan and Na-Dene Languages 185 


low-toned -nl- of b. In a. the -m- takes a middle tone as a transition 
between the high idi- and the low-toned imperfective -'i. 

8. a. :i^is^(d I shall kill him 
b. :(is^(a kill him ! 

These are imperfective forms based on --rd " to kill one person or ani- 
mal) ". The middle tone in b. is organically a high tone which has been 
depressed to accentuate the high tone of :{{- (cf. also s\.-{i5^(a " kill me ! ") 
used in verbs of killing and probably referring to " death. " It is variable 
in tone. In a. it is one step lower than the stem, in b. it is raised to the 
high tone because of the second person singular subject. In perfective 
forms, based on-^i <; Ath. *-'(in, the :(i- sinks to a low tone (e. g. yi;(U^t 
" he has killed him "). The Ath. prototypes of a, and b. are : 

a . *:(e-c-i-^i 

b. ■{e-n-i-'{i (or -^£ ?) 

9. a. yi/'ii*;^ you have worn it 
-b. yi/'5i';( he is wearing it 

Form a. is based on the perfective form of the stem, -/Vi*:( (cf. yisfs\'7 
" 1 have worn it ",yVyi/Vf;( " he has worn it "). 

Form b. is based on the continuative form of the stem, -t'si':(, with 
middle tone. The imperfective is -fsist\ e. g. Ist'sist^ " I shall wear it ", 
yifsist' he will wear it ". In a. yi- is an aspective prefix ; the old 
pronominal -«- has dropped out (yi- << *-^e-n-), but the yi- is not raised to 
the high tone, as it is already high relatively to the stem. In b. y- is the 
definite third personal object, Ath. *ye-. Its high tone, as contrasted 
with the low yl- of equivalent forms in other aspects (ylyj'' si :( " he has 
worn it ", ylt'sist', " he will wear it ") is characteristic of certain conti- 
nuative forms, which demand a high tone for the pronominal object ; 
analogous to form a., as compared with imperfective yU'sist\ is sIs^ols " he 
is throwing (bags) at me " : s\s'(xs " he will throw (a bag) at me ", 
yisk'a' " he keeps hitting him (with a cloth) " : ylsJid' " he will hit him '', 
yigx " he is poking him " : y\ga. ' " he will poke him '\ yi:(i " he is 
calling him " : y'-i: " he will call him" (perfective yi«i:{i " he has called 
him "). Observe that a. and b. are identical in cadence but dill'er in 

10. a. nayi:(idi the one who stood 

b. nayi:(id'. he will stand later on 
Form a. is a relative (or participial) in -t formed from the perfective 

185 ^^ American Indian Languages 2 


nayiii {-nd-)- The stem, -iid-, means " to become ", the idea of " stand- 
ing " being contributed by the low-toned prefix na- " erect, on the 
ground ". The imperfective form of this important verb stem is high- 
toned, -71 {-Z'-d-), e. g. na:(C " he stands ". y.- is perfective and takes the 
hio-h tone by contrast with the preceding prefix and the following stem. 
Form b. is a " delayed future ", characterized by aspective yt- and the 
middle-toned suffix-i (analogous forms are dylstldb'. " I shall do it after a 
while ", yisdam " I shall drink it later on ", Vu-('^l''yis'ak " I shall get 
water later on ^'). Besides the relative suffix -i and the delayed future -i, 
there is a low-toned -\ which makes gerundives. It is not properly a 
suffix, however, but an enclitic particle attached to the pause-form of the 
preceding verb, e. g. nayi^t' -|- i, whence, with contraction, ndyb^j,' " he 
having stood. 

11. a. *n7.y\sk\' he threw it (a cloth-like object) away 

b. n6Lyisk\' I threw it down 

c . nxyisk'i' you threw it down 

All three forms are perfectives, based on the perfective form, -■ ,k\ ot 
the stem ; its imperfective is -^', continuative -k'a'. The Ath. proto- 
types are : 

a. *nd-ye-s-i-k\'y 

b. nd-^e-c-i-k\'y 

c. tia-^e-n-i-kU'y 

Form a. is really of a very different pattern from b. and c. It has a 
high-toned prefix nd- " away " that requires *-s(/)- as a following per- 
fective element, whereas the na- of b. and c. is a prefix of variable tone, 
generally low, na- " down ", which is followed in perfective forms by 
*-v(^)-. The -yl- of a. is the definite third personal object, contrasting in 
tone with the high-toned prefix and the middle-toned stem. In the other 
persons the -yi- must be replaced by an objective -1-, which contracts 
with nd- to nd'-, e. g. nd'slskc " I threwit away", nd'sisk\' " you threw 
it away " (Ath.* nd-'e-se-c-i-k'e'y^*' and nd-'e-se-n-i-k\'y). Contrast noLyiyisk" i 
'' he threw it down " (Ath. *na-ye-ye-i-k't'y). Note that the only difference 
in tone between b. and c. is that the latter form begins a step lower. This 
has the effect of accentuating the height of the following syllable, 
dependent, as usual, upon the second person singular subject. 

12. a. xdy\sdl^' I pulled it out quickly 

b . xayLsdlg,' you pulled it out quickly 

c. xdyisdla' he will pull it out quickly 

Six: Athabaskan and Na-Dene Languages 187 


The verb stem is-dla {dlad- ) in the imperfective, -dig.' {-dlgd-) in the 
perfective, xa- " out '" has a fixed low tone. Form b., having a second 
person singular subject, has a higher tone on the perfective prefix -yt- 
than form a., in which -yi- takes the middle tone, transitional between 
the low-toned prefix and the high-toned stem; cf., fora parallel cadence, 
form 4 a (last 3 syllables). Form c. assigns ^ middle tone to the variable 
syllable -yi- (definite third personal object, Ath. *-ye-) by contrast to 
the fixed low tones which precede and follow. 

13. a. nxdiT^i you will disappear 

b . nadlif he will disappear 

c. nidlX} he disappeared 

d. nodi:(C he will appear 
(e. nadi:(C you will appear) 

I have put form e. in parentheses because I have no actual record of it 
under my hand at the moment, but the general analogy of numerous 
other forms makes it a safe enough form to accept. The stem, " to be- 
come ", is high-toned {-:(i) in the imperfective forms (b. , d., e.), middle- 
toned {-:(}') in the perfective (c.) ; cf. no. 10. Forma, is remarkable. It is 
an imperfective parallel to b. , yet it seems to have a middle tone on the 
verb stem, as in form c, where a middle tone is justified. The reason is 
simple. The form nid<.:(iis a secondary development oi n6di:(l.\ with fixed 
high tones on na- " away " (cf. no. 11) and -:(C and with a high tone on 
the prefix -di- because of the second person singular subject. Certain ele- 
ments with a fixed high tone, like our «a-, depress the following syllables 
to a slightly lower register, so that an immediately following high 
tone takes what seems to be a middle position. (A similar effect is often 
produced by the common demonstrative prefix i-, e. g. dsfa " I am thus " 
«< theoretical dsfd). I strongly suspect that the -di:(i of a. is not in a true 
middle position but is midway between high and middle (cf. the " dropp- 
ed high" tones, discussed on p. 196, which arise in another way) ; 
while I often noted tones that fell between high and middle or low and 
middle, I did not fully realize in the field the theoretical importance of 
these secondary " dropped" and " raised " tones. Thus, the syllable -;(i' 
of these five forms probably occurs on three registers — high, " dropped 
high ", and middle, hut not low. The element noc- of forms d. and e. is 
a totally distinct prefix from the ni- of a. , b., and c. In b^ and c. variable 
-di- has contrastive tone, in d. it has transitional tone (cadence as in 
12a.), in a. and e. it has functional tone. Forme, is one of the curious third 
personal perfective forms in -s- in which this element has disappeared 

VI American Indian Languages 2 


(cf. 3 c) ; analogous forms are found in other Athabaskan dialects. The 
reconstructed Ath, forms are : 

a . *nd-dt-n-7^d 

b . "nd-de-^ed 

c . *nd-de-ied 

d . *na-de-x^d 

e. *na-dt-n-/^M 

It must not be imagined that the examples which we have selected 
are in any way exceptional . The number of such homonymous sets — 
homonymous except for tone — is legion in Sarcee and our difficulty has 
been to reject rather than to find examples in the recorded material. But 
enough has been given to indicate the nature of the tone problem in Sar- 
cee. Tone is not a matter entirely of the inherent pitch of a given word 
or element nor is it entirely a matter of grammatical symbolism. Both 
types of tone function are inextricably interwoven in Sarcee in a system 
of considerable complexity, from a morphological standpoint. Phonetic- 
ally the complexity is only moderate. 

What is the fundamental nature of this tone system ? It is much too 
early to speak with assurance, as comparative material bearing on tone 
has still to be gathered from other Athabaskan dialects. It is already 
reasonably certain that the inflected tones of Sarcee are of secondary 
origin and cannot be credtied to the original Athabaskan tongue. In other 
words, Athabaskan is in all probability a tone language of varying regis- 
ters (high, middle, and low; or, possibly, fundamentally high and low), 
not of inflections. The ultimate theory of Athabaskan tone would have 
to take account of three important problems : 1 . How are the syllables 
with variable tone to be explained ? Did these syllables originally have 
fixed tone and did they later, having sunk to the status of relatively weak 
proclitic or enclitic elements, lose their tone and take on new tones 
according to certain rules of tone sandhi? 2. How did tone come to have 
grammatical function (e. g. the high tone, or relatively high tone, so often 
found on the syllable preceding the stem in second person singular forms ; 
the high tone on the objective pronominal elements in certain conti- 
niiative paradigms) ? Are these functional uses the consequence of merely 
mechanical tone principles and not directly symbolic in origin ? 3. How 
explain the characteristic changes of tone in the verbal aspects? Why do 
certain verbs allow of no tone changes in the stem, while others have 
one in the imperfective, another in the perfective or continuative ? May 
we suppose that originally a verb stem had only one inherent tone, 
preserved say in the imperfective stem, and that the other aspects deve- 
loped secondary tones on certain mechanical principles, say the amalga- 

Six: Athabaskan and Na-Dene Languages 



mationofthe stem with other elements that have disappeared as such but 
have left their trace behind in tone alterations ? 

It is not likely that all of these questions can be answered from the 
standpoint of Athabaskan alone. The same questions, paiticulary 1 . and 3., 
arise in Tlingit as well, whence it follows that the ultimate explanation 
of Athabaskan tone will be given by the comparative study of Atha- 
baskan and Tlingit, possibly also of Haida. Should it appear that Tlin- 
git-Athabaskan tone originally depended entirely on the inherent 
tone value of independently significant syllables, we should still have to 
ask what were the determinants of these tone values. But we are very 
far from having arrived at the point where such a question is profitable. 

Editorial Note 

Originally published in Journal, Societe des Americanistes de Paris 17, 
185-205 (1925). Reprinted by permission of the Societe des Americanistes. 

The Sarcee tone system is more complex than that of most tonal Athabaskan 
languages, which normally have only two tones (high and low). The Sarcee mid 
tone apparently results from the partial lowering of an original high tone in 
certain phonological circumstances, a fact that may be historically connected 
with a similar lowering (from high to low) in Navajo and the other Apachean 
languages (Krauss and Golla 1981: 85-85). The most recent work on Sarcee is 
that of Eung-Do Cook, who has carried out extensive field work. He has pub- 
lished a reanalysis of the Sarcee tonal system (Cook 1971, 1984), but does not 
challenge the essential correctness of Sapir's description. 

Sapir's ms. corrigenda in his copy are as follows: 




p. 190, lib 
p. 190, lib 
p. 191, 1. 22 

[second z] 



p. 195, 1. 1 

a + a becomes a 

a + a becomes a- 

p. 195, 1. 15 



p. 195, 1. 19 



p. 197, 1. 2 



p. 202, 1. 14 
p. 202, 1. 17 



p. 202, 1. 18 



p. 202, 1. 20 


na- . . . 

p. 202, 1. 31 
p. 202, 1. 31 
p. 204, 1. 22 

[1st form] 

[delete final superscripts 



The Similarity of Chinese and Indian Languages 

New light has been thrown on the ancestry of the American Indian by Dr. 
Edward Sapir, the Canadian anthropologist now on the faculty of the Univer- 
sity of Chicago. Dr. Sapir said that his research work on Indian linguistics has 
convinced him of the identity of the language of certain Indian tribes with that 
of the primitive Chinese. 

The similarity of the two tongues and the linguistic distribution of tribes scat- 
tered at random over the Americas have convinced Dr. Sapir that these groups 
must have entered this continent as a wedge from Asia. By a close comparison 
of the primitive Chinese, Siamese and Tibetian, all in the same language cate- 
gory, with the language of the "Nadine group" of North America, Dr. Sapir has 
found the same peculiarities of phonetics, vocabulary and grammatical struc- 
ture on both sides of the Pacific Ocean. 

The American Indian groups speaking the language of the Nadine group are 
found in all parts of the North American continent from northern Mexico to the 
southern boundary of Alaska, widely distributed among other Indian tribes 
whose language and customs are entirely different. 

With minor changes the Navajo of New Mexico speaks the language of the 
Sarcee in Alberta, and the linguistic stock of the Tlingit, just south of the 
Eskimos in Alaska, is much the same as that of the Hupa in California. 

It is probable, according to Dr. Sapir, that the migration of Asiatics speaking 
primitive Chinese or Tibetian took place at some time in the past, and that 
these immigrants settled or moved over the mountains and plains, some 
remaining in northwestern Canada to become the Tlingits, and others moving 
out to the Queen Charlotte Islands off the west coast to form the Haida group, 
and still others penetrating to the deserts of the Southwest. 

From the modern Chinese, which in academic circles is considered relatively 
simple, students of linguistics can reconstruct primitive Chinese which is far 
more complex than any of the dialects known to the Mongolian layman of 
today. Dr. Sapir has discovered not only that the Indians of the Nadine groups 
speak with a tonal accent, raising or lowering of the voice to give certain mean- 
ing to words, in a manner similar to the tonal peculiarities of the early Chinese, 
but also that the meanings of certain words arc identical. Further, he has dis- 
closed the fact that the Indians have retained certain prefixes and suffixes that 
long ago have disappeared from the Chinese speech, but which are clearly dis- 
cernible in the early forms. 

192 VI American Indian Languages 2 

Editorial Note 

Originally published in Science 62 (1607), supplement of 16 October, xii 
(1925). Reprinted by permission; copyright by the American Association for 
the Advancement of Science. 

This anonymous, and somewhat shoddy, piece of scientific journalism seems 
to have been based on a brief interview with Sapir shortly after his move to 
Chicago in the fall of 1925. Certain phrases sound like Sapir s, but most of the 
text must have been reconstituted from hasty notes (hence "Nadine," "Tibe- 
tian," and other gaffes). The article appeared in the separately paginated 
Science Supplement section of the journal, indicating that it was not intended as 
a formal announcement of scientific results, but rather as news of work in 

Review of 
Berard Haile: A Manual of Navaho Grammar 

A Manual of Navaho Grammar. Arranged by Father Berard Haile, O.F.M., 
of the Cincinnati Province of St. John the Baptist. Santa Fe, New Mexico: Santa 
Fe PubHshing Co., 1926. Pp. xi + 324. $6.00. 

This is an invaluable work, useful both to the practical student of Navaho and 
to the scientist interested in comparative Athabaskan linguistics, emanating 
from the enthusiastic band of Franciscan missionaries to whom we owe the 
excellent Ethnological Dictionary of Navaho and Navaho Vocabulary. It is a 
cruelly difficult language that Father Berard treats in this "manual," and if a 
number of statements or formulations seem doubtfully correct from the stand- 
point of the comparative linguist, it should always be remembered that the 
author's point of view is distinctly dialectic and descriptive, and not at all com- 
parative or historical. It is very interesting to observe that Father Berard 
explicitly recognizes the necessity of noting the pitch of the various syllables of 
the Navaho word, though the record in the book is extremely fragmentary on 
this point. This statement corroborates Father Legoffs remarks on pitch in 
Chipewyan, and E. Sapir's field data on Sarcee and Kutchin pitch. 

Editorial Note 

Originally published in American Journal of Sociology 32, 511 (1926). 
Reprinted by permission of the University of Chicago Press. 

Sapir's references are to the Franciscan Fathers (1910, 1912) and to Legoff 
(1889). Father Haile later published a considerably more sophisticated intro- 
ductory grammar of Navajo (Haile 1941-48) and a stem dictionary (Haile 
1950-51), both reflecting Sapir's influence. 

A Summary Report of Field Work among the Hupa, 

Summer of 1927 

Dr. E. Sapir spent the end of June, July, August, and the beginning of Sep- 
tember, 1927, in a linguistic investigation of the Hupa Indians, who occupy a 
reservation in the valley of the Trinity river in northwestern California. This 
work was done under the auspices of the [360] Department of Sociology and 
Anthropology of the University of Chicago. A number of good informants was 
secured, chief among them being Sam Brown, who served as the most impor- 
tant source of material and as interpreter of everything obtained from him and 
others. The information gathered was chiefly linguistic in character but a large 
body of ethnological material was also obtained, partly in the form of texts and 
partly as notes directly communicated in English. 

The linguistic material was chiefly obtained as part of a programme for the 
comparative study and reconstruction of the Athabaskan languages. A careful 
study was made of the grammatical structure of Hupa, which offers many diffi- 
culties, and a reasonably complete vocabulary was obtained. The texts, which 
number about seventy-five, were so chosen as to duplicate as little as possible 
the valuable material already published by Dr. P. E. Goddard. They are pre- 
vailingly ethnological in content. 

Some of the main results secured from the study of the Hupa language may be 
mentioned here. The sound system, as might have been expected, proved to be 
much more complex than hitherto represented and more in accordance with 
the typical Athabaskan patterns worked out for Sarcee, Kutchin, and Navaho. 
The old Athabaskan k-series (intermediate g, aspirated k', and glottalized 'k) is 
represented in Hupa by a velar series (g, x, and 'q), but the old prepalatal series 
(g, k', 'k) is preserved as such. Curiously enough, Hupa has also developed a 
new series of k-sounds which are neither velar nor prepalatal but mid-palatal 
(g, k', 'k). These do not represent the old Athabaskan series of k-sounds but are 
the diminutive form of the prepalatal set. The sounds s, ts', 'ts represent not 
only the old Athabaskan s-sibilants but also the diminutive form of the 
Athabaskan c-series (in Hupa terms: voiceless w, tew, "Ic). Vocalic quantity 
proved to be of fundamental importance for the understanding of Hupa mor- 
phology. This is true also of the use of the glottal stop and of final aspirations. 
Many final consonants arc glottalized and there is a characteristic and probably 
archaic difference between non-syllabic final consonants and consonants which 
have half-syllabic value. This difference is responsible for the parallelism of 
"light" and "heavy" syllables, a distinction which had already been worked out 
for other dialects. 

The old Athabaskan tone system, which can be reconstructed in large part 
from striking resemblances in the tone patterning of Sarcee, Kutchin, Navaho, 

196 VI American Indian Languages 2 

and no doubt by many other northern and [361] southern dialects as well, no 
longer appears in Hupa. Mr. Li s researches prove the same loss of tone for 
Mattole and Wailaki. Later investigation may disclose the fact that the absence 
of tone is one of the distinguishing characteristics of the Pacific (or perhaps only 
southern Pacific) group of Athabaskan dialects. There are, however, interesting 
tone cadences in the relative forms of Hupa verbs which are most easily 
explained as survivals of older forms with a high tone on the final syllable, such 
as we actually find in Sarcee and Navaho. 

The morphology was completely overhauled. Owing to a faulty phonology 
the details and some of the fundamentals of Hupa structure had not been fully 
grasped in former works on Hupa. The relative forms of the verb, which are as 
important here as elsewhere in Athabaskan, had not been properly kept apart 
from the non-relative forms. Owing to this fact the whole verbal system needs 
to be presented in a new light. The so-called "past definite" is merely a special 
use of the relative form of the perfective (Goddard s "present definite") and is 
paralleled by analogous relative forms based on the imperfective ("present 
indefinite") and on the continuative. The aspect system of Hupa needs to be 
revised also in other respects. The formation of the continuative (as part prefix 
and stem form) is entirely analogous to its formation in other dialects. A dis- 
tinct permissive paradigm must be recognized not only for the third person but 
also for the first. A distinct potential mode was discovered. There are also 
special modal forms for the verbal abstract and for the prohibitive. In none of 
these cases is reference merely had to the use of certain prefixed or suffixed 
particles. As in Navaho, a number of verbs distinguish durative and momen- 
taneous stem forms of the imperfective. 

Since this report was first written, P. E. Goddard has published a paper 
entitled. Pitch Accent in Hupa (Univ. of Calif. Publ. Amer. Arch. Ethn., 
23:333-338, 1928), in which it is shown that a study of Hupa tracings fails to show 
that Hupa syllables have inherently high or low tones. This is entirely in accord 
with Sapir's auditory record for Hupa and quite opposed to his auditory record 
for certain other languages of the Athabaskan groups, particularly Sarcee, 
Kutchin, and Navaho. A small amount of independent material obtained on 
Anvik (Chapman's "Ten'a," also known as Ingalik) indicates that here too tone 
is lacking. 

Editorial Note 

Originally published in American Anthropologist 30, 359-361 (1928). 
Reprinted by permission of the American Anthropological Association. 

References to Goddard's earlier work are mainly to his Morphology of the 
Hupa Language (1905). 

Six: Athabaskan and Na-Dene Languages 197 

Sapir's Hupa notebooks and lexical files are now in the Library of the Ameri- 
can Philosophical Society (manuscript 497.3 B63c Na20a.4; cf. Kendall 1982: 
56, item 4369). The texts, with linguistic and ethnographic notes, have been 
edited for publication in Volume XIV. The circumstances of Sapir's and Li's 
1927 "expedition" to northwestern California are described informally by Sapir 
(1927b, Volume IV). 

Golla's Hupa grammar (1970) was based on Sapir's materials, supplemented 
by Golla's own field work. Sapir's lexical data are incorporated into a compre- 
hensive Hupa dictionary now being prepared by the Hupa tribe. 

Excerpt from 

The Concept of Phonetic Law as Tested 

in Primitive Languages by Leonard Bloomfield 

Bloomfield's experience with the Central Algonkian dialects is entirely par- 
allel to my own with the Athabaskan languages. These constitute an important 
linguistic stock which is irregularly distributed in North America. The north- 
ern group occupies a vast territory stretching all the way from near the west 
coast of Hudson Bay west into the interior of Alaska. To it belong such lan- 
guages as Anvik (in Alaska), Carrier (in British Columbia), Chipewyan, Hare, 
Loucheux, Kutchin, Beaver, and Sarcee. We shall take Chipewyan and Sarcee 
as representatives of this group. The geographically isolated Pacific division of 
Athabaskan consists of a number of languages in southwestern Oregon and 
northwestern California. We shall take Hupa as representative. The southern 
division of Athabaskan is in New Mexico and Arizona and adjoining regions, 
and is represented by Navaho, Apache, and Lipan. We shall take Navaho as 
representative of the group. In spite of the tremendous geographical distances 
that separate the Athabaskan languages from each other, it is perfectly possible 
to set up definite phonetic laws which connect them according to consistent 
phonetic patterns. Navaho, Hupa, and Chipewyan are spoken by Indians who 
belong to entirely distinct culture horizons, yet the languages themselves are as 
easily derivable from a common source on the basis of regular phonetic law as 
are German, Dutch, and Swedish. 









1. s 





2. z 





3. dz 





4. ts 





5. ts' 





1. c 





2. j 





3. dj 





4. tc 





5. tc' 





1. X 





2. y 





3. gy 





4. ky 





5. ky' ky' tc' ts' tc' 

200 ^f American Indian Languages 2 

Table VI shows the distribution in Hupa, Chipewyan, Navaho, and Sarcee of 
three initial consonantal sets, each of which consists of five consonants. In other 
words, the table summarizes the developments of fifteen originally distinct 
Athabaskan initial consonants in four selected dialects. Each of the entries 
must be considered as a summary statement applying to a whole class of exam- 
ples.^ The table merits study because of its many implications. It will be 
observed that no one dialect exactly reproduces the reconstructed Athabaskan 
forms given in the first column. Series I is preserved intact in Navaho and 
Sarcee and very nearly so in Hupa, but has been shifted to another series in 
Chipewyan. Series II is preserved intact in Navaho, but has been shifted in 
Sarcee to identity with the series that corresponds to original I, while Hupa has 
introduced several peculiar dialectic developments and Chipewyan has shifted 
it to the original form of I. Series III is nowhere kept entirely intact but nearly 
so in Hupa, while in Chipewyan and in Sarcee it has moved to the original form 
of Series II, in Navaho to a form which is identical with the original and the 
Navaho form of Series Litis clear from the table that a Sarcee s is ambiguous as 
to origin, for it may go back either to Athabaskan s or Athabaskan c. On the 
other hand, a Sarcee s which is supported by either Navaho or Hupa s must be 
representative of an original Athabaskan s. Sarcee tc is, in the main, unam- 
biguous as to origin, for it corresponds to the original Athabaskan ky. It is 
curious and instructive to note that, of the four languages given in the table, 
Hupa and Chipewyan are the two that most nearly correspond as to pattern but 
never as to actual sound except in the one instance of y (III, 2). 

Let us take a practical example of prediction on the basis of the table. If we 
have a Sarcee form with tc, a corresponding Navaho form with ts, and a 
Chipewyan form with tc, what ought to be the Hupa correspondent? According 
to the table it ought to be ky. 


Ath. Hupa Chipewyan* Navaho* Sarcee* 

*kyan ... tcq n-t-tsq tea 

'rain' 'there s a rainfall' 

*q represents nasalized a. as in French an. Sarcee a is a peculiar a with velar resonance, regu- 
larly developed from Athabaskan a. 

Table VII shows the distribution in three dialects of the Athabaskan sound ky 
(III, 4) in the word for "rain'. When I first constructed the Athabaskan pro- 
totype, I assumed an initial ky, in spite of the absence of the test form in Hupa, 
on the basis of the dialectic correspondences. Neither an original ts nor tc could 
be assumed in spite of the fact that these sounds were actually illustrated in 

4. The apostrophe symbolizes a peculiar type of consonantal articulation, characterized by simul- 
taneous closure of the glottis and point of contact in the mouth, with glottal release preceding oral 
release. 7 is the French/ of yowr, djh they of English yu^/.x is the c/j of German /c/i; Wis approximately 
the wh of English what; e is the th of English thick; 8 is the th of English then. 

Six: Athabaskan and Na-Dene Languages 201 

known dialects, whereas ky was not. The Hupa column had to remain empty 
because the cognate word, if still preserved, was not available in the material 
that had been recorded by P. E. Goddard. 

In the summer of 1927, however, I carried on independent researches on 
Hupa and secured the form kyax^-kyoh,^ meaning 'hailstorm'. The second ele- 
ment of the compound means 'big' and the first is obviously the missing Hupa 
term corresponding to the old Athabaskan word for 'rain'. In other words, an 
old compound meaning 'rain-big' has taken on the special meaning of 'hail- 
storm' in Hupa. The Hupa form of the old word for 'rain' is exactly what it 
should be according to the correspondences that had been worked out, and the 
reconstruction of the primitive Athabaskan form on the basis of the existing 
forms was therefore justified by the event. 

Dialectic Forms for 'Rain' 

Anvik (Alaska) rcoA^* 

Carrier (B.C.) tcan 

Chipewyan tcq 

Hare tcq 

Loucheux tcien 

Kutchin tscin 

Beaver tcq 

Sarcee tcq 

Navaho n-t-tsq 

*D represents open o, as in German voll; q is nasalized o. N is voiceless n. 

Table VIII gives the chief dialectic forms that were available for the recon- 
struction of the Athabaskan word for 'rain'. Observe that not one of these has 
the original sound ky which must be assumed as the initial of the word. This is 
due to the fact that the old Athabaskan ky and related sounds shifted in most 
dialects to sibilants but were preserved in Hupa and a small number of other 
dialects, some of which are spoken at a great remove from Hupa. In other 
words, in working out linguistic reconstructions we must be guided not merely 
by the overt statistical evidence but by the way in which the available material is 

Editorial Note 

Originally published in Rice, Stuart A. (ed.). Methods in Social Science: A 
Case Book, pp. 297-306. Chicago: University of Chicago Press (1931). 
Reprinted by permission of the University of Chicago Press. Only pp. 302-306 
are reprinted here; footnote and table numbers are as in the original. 

5. r) is the ng of English sing. 

202 y^ American Indian Languages 2 

This discussion of comparative Athabaskan phonology forms part of a longer 
paper on comparative linguistic methodology (printed in full in Volume I). 
Although Sapir s methods are irreproachable, his reconstruction of the Proto- 
Athabaskan affricates is almost certainly incorrect, as Krauss has shown (1964). 
Sapir gave little weight to forms in his own Kutchin and Anvik notes that are 
now generally interpreted as retaining the Proto-Athabaskan contrast between 
retroflexed (or flattened) and non-retroflexed palatal affricates (*ts'^ : *ts). 
Sapir apparently believed that the retroflex/non-retroflex contrasts in Kutchin 
and Anvik (and — had he known — several other Alaskan languages) were of 
secondary origin. The retention of the contrast in Tsetsaut with quite different 
phonetic realizations, and a corresponding contrast in Eyak, make this an 
unlikely explanation. The Tsetsaut data were available to Sapir (Boas and God- 
dard 1924), and Krauss has commented that "why or in what way Sapir ignored 
that important material remains a dark mystery" (1986: 162). 

Two Navaho Puns 

It is a well known fact, often stressed by Boas, that the American Indians do 
not go in for riddles and proverbs. What few exceptions have been found serve 
only to emphasize the rule. If to riddles and proverbs we add puns, as it seems 
we have a right to do, it begins to appear likely that the American Indian has a 
generalized lack of interest in light verbal fancy. There is plenty of metaphor in 
his rituals, there is considerable etymologizing ad hoc in his legends, and his 
oratory is famous, but the zest in quick, irresponsible reinterpretation of famil- 
iar words or phrases which lies at the bottom of the pleasure that we experience 
in the telling of riddles, proverbs, and puns seems strangely un-Indian, what- 
ever may be the reason. 

Yet puns are not entirely absent, as the two following Navaho examples show. 
They were collected in the summer of 1929 at Crystal, New Mexico. 

In a council held some time ago the leader of one of the two contending 
parties said,^ xdct'e ndhdnd^'d-' 'You people decide on one thing!' A cripple 
who was present whispered to one near by, ndcidi-'d-' 'Pick me up!' The latter, 
catching on at once, picked up the cripple and, holding him in his arms, asked, 
xd-dic ndhdnc'd-' 'Where am I to put him down?' Everybody laughed. This is 
said to be a favorite anecdote among the Navaho and depends for its point on 
the double meaning of the verb nd-hd-ni- . . .- '«•', which may mean either 'to 
decide on the matter' or 'to put him down'. 

A close analysis of this pun shows that it is more subtle than appears on the 
surface and that to enter fully into its humor requires sensitiveness to no less 
than three changes of linguistic front. The fundamental pun is simple and 
would hardly be enough by itself to raise a laugh, one suspects. This is the use of 
the verb stem -a' 'to handle the "round" object' in the transferred sense of 'to 
handle the affair, words, plan, date, decision'. The secondary use oi-'d' (-'d) in 
an abstract sense is very common in Navaho, e.g. tc'6-hd- ni. . .-'«•' 'to tell'; nd- 
hd-ni- . . .-'d- 'to make the decision'; nd-hd-. . .-'a 'to make plans', - d-nd-hd- . . . 
-t'd-' ( = -d-'d-') 'a date is being set for one'. The tingling moment in the anec- 
dote comes with the cripple's whispered [218] request; for, in addition to the 
very general transfer of meaning already noted, there is the added point that 
one does not normally use the verb nd-di- . . .- '«•' 'to pick it up' (and its cor- 
relative ni-ni- . . .- '«•' 'to put it down') of an animate being, but only of such 
inanimate 'round' objects as a potato, or apple, or watch, or rock. In other 
words, the wily cripple, turning his helplessness to humorous account, classifies 
his hunched up body as a 'round object', ndcidi- a' substitutes for ndcidi-ftxi-' 
and xd-dic ndhohc'd-' substitutes for xddic ndhdhctxi' {-t-txi-' 'to handle the 
animate being'). Had a little boy of normal physical health made the request, 

1. Grave accent (a) represents low tone, acute (a) high tone, circumflex (a) falling tone. 

204 yf American Indian Languages 2 

the pun would have seemed a bit far-fetched, for he could not easily be thought 
of as lifted up and put down like an inanimate object. The quick understanding 
by the second punster of the cripple s use of -'a-' socializes the pun and kindles it 
into something like satire of the ponderous doings of important people. Finally, 
the climax of the pun, 'Where am I to put him down?', reinterprets the -hd-, 
w hich in the first usage has the meaning of something like 'the affair, circum- 
stance', while in the second it is a personal pronoun referring to 'this one'. 
These two uses of -hd- are historically distinct in Athabaskan. Briefly, then, the 
first element of the pun is impersonally contributed, as it were, by the language 
itself; the second is the creation of a masochistic cripple; the third is the echoing 
understanding of his friend, who equates 'the great business at hand' with 'this 
poor chap'. 

The second pun is much simpler. It is told as a joke rather than as an anec- 
dote. 'So and so has gone over there', one says. 'What for?' xwdjdd-txal, 
apparently 'he is going to give one a kick' (future of semelfactive verb yictxat 
'I give it a kick'), actually 'he (the medicine-man) will perform a ritual "chant"' 
{{uiwTt oi xatx a- 1 '\ perform a "chant"', a denominative verb based on jco/xfl/ 'a 
ritual "chant"'). Here again the xwd- (= -hd- above) changes from an imper- 
sonal to a personal application. It adds to the flavor of the joke if one remem- 
bers that Navaho medicine-men are often hired to come from great distances in 
order to direct curing rituals and that the spaces between the scattered Navaho 
hogans are wide indeed. 

The great number of homonymous elements in Navaho, due largely to the 
leveling influence of phonetic laws, and its peculiarly intricate structure, which 
derives quite definite meanings from the assembling of elements that are gener- 
alized and colorless in themselves, combine to make Navaho a peculiarly 
tempting language for the punster. 

Editorial Note 

Originally published in Language 8, 217-218 (1932). Reprinted by permission 
of the Linguistic Society of America. 

Problems in Athapaskan Linguistics 

The Athapaskan languages, in spite of their essential uniformity in phonetic 
type and morphology, are richly ramified and we need many more dialectic 
studies than we have yet had in order to understand the group as a whole and to 
reconstruct it adequately. The dialects of the Southern division are com- 
paratively well known, though less has been published about these dialects than 
about the chief languages of the other divisions. Even here, however, much 
important dialectic work remains to be done. Manuscript materials on Navaho, 
Mescalero Apache, Chiricahua Apache, Jicarilla and Lipan are probably abun- 
dant and accurate enough to be called satisfactory. Further work remains to be 
done on Kiowa Apache and on San Carlos and other Western Apache dialects. 

In the Pacific area the manuscript and published record is probably adequate 
for Hupa, Mattole and Wailaki. The languages in this area which need careful 
study are Kato, Lassik, Chasta Costa, Galice Creek, and Upper Umpqua. 
Kwalhioqua and Tlatskanai, which seem to have belonged to the Pacific group, 
are probably extinct. It is particularly important to try to determine if there are 
any survivals of the early Athapaskan tone system in the languages of the Pacific 
group. That there was such an early archaic system follows from a comparison 
of the Southern languages with most of the languages of the Northern group. So 
far, however, no Pacific dialect has been found with a tonemic system, though 
there is an interesting survival of such a system in certain grammatical forms of 

The greatest complexity prevails in the north. Carrier, Chipewyan, Hare and 
Kutchin are relatively well known though only a part of the material which has 
been obtained is in print. Kutchin, however, probably needs a careful overhaul- 
ing in the field because of its great complexity and the necessity of getting fuller 
grammatical material. The Kutchin group of dialects is one of the most spe- 
cialized and needs careful attention. Carrier may be considered reasonably 
well known because of Father Morice's monumental work but the tonal pecu- 
liarities of the Carrier group of languages still remain to be worked out. We 
have some published material on Beaver and Ingalik but it is altogether 
unsatisfactory and needs to be completely overhauled. The Athapaskan 
dialects which are either not known at all or only fragmentarily include: 
Chilcotin, Tahltan, Tanaina, Tanana, Han, Copper River, Koyukukhotana, 
Slave and Dogrib. 

Owing to the many intricate problems of phonology set by the task of recon- 
structing Athapaskan, we cannot have too much dialectic evidence. The treat- 
ment of original final consonants is particularly intricate. Another problem is 
the adequate reconstruction of the tone system. Certain languages, like most of 
the Pacific languages and Ingalik, seem to have lost the old tone system; others 

206 ^f American Indian Languages 2 

like Chipewyan and Navaho possess a two-register system; still others, like 
Sarcee, have a three-register system, which may either be archaic or, more 
likely, a secondary development of an earlier two-register system. In certain 
Athapaskan languages, such as Hare and Kutchin, there is a tonal complication 
due to the tact that the tonemic structure of a word differs according to whether 
it is in final or absolute position on the one hand, or protected by an immedi- 
ately following word. All these differentiations suggest the importance of get- 
ting as much dialectic material as possible. 

An important task in comparative Athapaskan is the recovery of the entire 
vocabulary which is basic to the stock as a whole. Because of differentiations of 
meaning in different dialects and the partial survival of many words in scattered 
dialects, we cannot afford to do without the testimony of all or most of the 
recoverable dialects. 

Editorial Note 

From a typescript with Sapir s ms. revisions, now in the American Philosoph- 
ical Society Library, manuscript 497.3 B63c Na.l. Published by permission of 
the American Philosophical Society. 

The original of this short statement is a short typescript, edited and signed by 
Sapir but bearing no date or other indication of its origin. It was perhaps pre- 
pared for the American Council of Learned Societies or a similar source of 
research funding. Morris Swadesh donated the typescript to the American 
Philosophical Society Library in 1946. 

Review of 
A. G. Morice, The Carrier Language 

The Carrier Language (Dene Family), a Grammar and Dictionary Combined. 
A. G. Morice. (2 vols., I: xxxv, 660 pp.; II: 691 pp. RM 80. Anthropos Lin- 
guistische Bibliothek, IX u. X Band. Modling bei Wien: Verlag der Interna- 
tionalen Zeitschrift "Anthropos," 1932.) 

This magnificent work, as its sub-title indicates, is not an ordinary grammar 
but also a dictionary, the lexical materials being skilfully disposed under appro- 
priate grammatical rubrics. The "Vocabulary" at the end of the second volume, 
with its page references, helps the student to find his way in these materials. 
There is no confusion, only a mutually fertilizing treatment of the complex 
grammatical forms of the language and its lexical content. Not only is such an 
interweaving of grammar and dictionary allowable for Carrier and its cognate 
languages, it is in many respects necessary, at least if the grammatical survey is 
to be complete and definite. In all the Athapaskan languages many compli- 
cated grammatical rules apply only to single "words" or to small sets of words. 
One cannot, therefore, give as adequate a notion of the more intimate structure 
of Carrier or Kutchin or Navaho with a schematic statement of processes and 
categories as is possible in such languages as Yokuts or Arabic or Jabo (in Lib- 
eria), in which grammatical principles, once mastered, [501] can be applied 
with a high degree of confidence to the words or word elements. It is not, of 
course, a question of the relative complexity of the grammar as of the relation of 
the grammar to the vocabulary. Father Morice's "Carrier Language" deserves 
to rank as a real contribution to linguistic method. So far as I know, the problem 
of handling grammar and vocabulary as a formal unity has never before been 
attacked in so original a manner nor with so sure an instinct. Incidentally, this 
single attack on a dual problem seems not unsuited to the genius of English, 
different as its grammatical contours are from those of Carrier. 

The body of the work consists of a preliminary treatment of Phonetics; Part 
First, "The normally non-verbal Parts of Speech" (subdivided into "The 
Noun," "The Adjectives," "The Pronouns," "The Postpositions, Conjunctions 
and Interjections," and "The Adverbs"); Part II, "The Verb morphologically 
considered" (subdivided into "The Verbal Stems," "The Verbal Prefixes," and 
"The Incorporating Verbs"); Part III, "The Verb grammatically considered" 
(subdivided into "General Notions," "The chief irregularities of the Verbs," 
"Grammatical Divisions," "Personal Divisions," "Modal Divisions," "Mor- 
phological Divisions," "Temporally incomplete Verbs," "Divisions based on 
Endings," and "Verbs with number-indicating Endings"); Part IV, "Syntax and 
Linguistic Peculiarities" (subdivided into "Syntactic Notes" and "Linguistic 
Peculiarities"); Part V, "Texts" (five texts with interlinear and free translations); 

208 VI American Indian Languages 2 

and Part VI, "Vocabulary" (an English check list). This list of the main headings 
will serve to give some idea of the scope of the work. Details are hardly in place 

It should be remembered that the strength of Father Morice's grammar lies in 
its wealth of descriptive detail, not so much in its ultimate configurative analy- 
sis. A comparative student of Athapaskan may want to reassemble much of the 
detail and redefine some of the fundamental outlines but he will always be 
grateful to Father Morice for the facts, of which he has an obvious mastery. One 
weakness of the work — which is, however, not an important drawback in prac- 
tice — is that the orthography is not sufficiently systematized from the 
phonemic point of view. Thus, both t and d are used for what is obviously a 
single phoneme, a lenis stop, while the corresponding aspirated fortis stop is 
written th. One might write these phonemes, with advantage to economy and 
phonemic accuracy, either d:t (defining d asa voiceless lenis, varying to the ear 
between French d and t, and t as a strongly aspirated stop) or t:t\ Either method 
is justifiable and adequate, whereas d:t:th corresponds only vaguely to the true 
facts; would be better but uneconomical. It is time that linguists realized 
that one of their first and most fundamental problems is to interpret the mass of 
purely phonetic data in terms of a phonemic symbolism that is as simple, as 
economical, and as powerful as possible. This ideal is rarely attained, yet its 
realization is implicit in the phonetic facts themselves. 

There are many ethnological remarks and implications scattered in the book. 
The anthropologist's attention should be called to the list of ethnological refer- 
ences on pages xxix-xxxv of the first volume. 

Editorial Note 

Originally published in American Anthropologist 37, 500-501 (1935). 
Reprinted by permission of the American Anthropological Association. 

Internal Linguistic Evidence Suggestive of the Northern 

Origin of the Navaho 

TNTERNAL linguistic evidence for inferences as to cultural antecedents 
-*■ is not in much favor among cultural anthropologists at the present time, 
and this for two reasons. Such linguistic evidence is often, if not generally, 
tricky as to what of a factual nature can be gathered from it, for words may 
change their meanings radically and, furthermore, it is often difficult to 
tell whether community of nomenclature rests on early linguistic relation- 
ship or on linguistic borrowing attending cultural diffusion. In the second 
place linguistic evidence is difficult to handle, full of phonologic pitfalls, 
requiring a closeness of knowledge that is often out of proportion to what 
little can be obtained from it for tangible cultural inference. Nevertheless, 
at its best linguistic evidence, properly controlled, may throw an un- 
expected light on remote cultural perspectives. There is reason to think 
that as our descriptive and comparative knowledge of unwritten languages 
increases, their value for cultural reconstructions and other kinds of in- 
ference — not least among which is elimination of theoretically conceivable 
possibilities — will grow in importance. It is natural that in the American- 
istic field linguistic evidence has as yet yielded but a scanty return to the 
historian of culture, but this need not continue to be the case indefinitely. 
I shall try to show that there is tangible evidence in Navaho itself for 
the secondary origin of apparently fundamental elements of Navaho cul- 
ture, such as agriculture, and that such evidence seems to point to an early 
association of the culture of these people with a more northern environ- 
ment than their present one. It may be said — and with justice — that the 
distribution of the Athapaskan languages is such as to make this historical 
theory as good as certain, but dialectic distribution is external, rather than 
internal, linguistic evidence. It is conceivable, if not plausible, that the 
Athapaskan-speaking tribes were originally massed in the Southwest and 
gradually rayed out to the north in successive waves of migration. One 
might argue that the Navaho and, to a greater degree, the various Apache 
tribes present the non-Pueblo aspect they do, not because of their relative 
recency in the area of Pueblo cultural development but because, like the 
Walapai and other Yuman tribes of Arizona, they represent a simpler and 
more archaic Southwestern culture, which proved impervious, aside from 
a late Pueblo veneer, to the influence of the more elaborate cultures in their 
neighborhood. It is true that the linguistic homogeneity of the Southern 
Athapaskan dialects is such and the dialectic cleavages in the nothern 


210 VI American Indian Languages 2 


Athapaskan area are so profound that the suggested theory fails to carry 
conviction either to the linguist or to the ethnologist, but here again we 
are dealing with external linguistic evidence. This external evidence is far 
more compelling than can be any evidence derived from details of dialectic 
structure or vocabulary, for it is more direct and sweeping. None the less, 
the more elusive internal linguistic evidence has its place in giving con- 
firmation to a hypothesis based on linguistic distributions. 

There is undoubtedly a large amount of relevant cultural evidence 
packed away in the vocabularies of Navaho and Apache. For the present 
I must content myself with considerations based on the study of four 
words or groups of words. 

1. The Navaho word for "gourd" is 'ade'.^ The word is used both for 
the plant and for the "gourd dipper, ladle. "^ The "gourd rattle," on the 
other hand, is otherwise named ('a7al).' But 'ade-' means not only "gourd 
ladle" but "dipper, ladle, spoon" in general, the gourd ladle being the ladle 
or spoon par excellence. Hence we find the earthen spoon called "mud 'ade'" 
or "earth 'ade'," while the modern tablespoon is called "metal (< flint) 
'ade'."^ Now the term 'ade' (in form a possessed noun -de*' with indefinite 
possessive prefix 'a- "somebody's" or "something's") means not only 
"gourd," "gourd ladle," and "ladle, spoon" in general, a natural family of 
words, but also "horn" or rather "somebody's, some animal's horn" (d6 
"horn" as absolute; 'a-de*' "an animal's horn," parallel to bi-de' "his 
[animal's] horn"). In no other Athapaskan dialect does 'ade' or its dialectic 
equivalent mean "gourd" or "gourd ladle," while, so far as I can discover, 
it is only in Apache that it means not only "horn" but also "ladle" in 
general. In Chiricahua Apache* we have possessed -de' "horn (of animal)" 
and 'ide' "cup, dish, dipper"; in Mescalero Apache -de" "horn (of animal)" 
and 'ide- "cup, dish, dipper." In both Navaho and Apache 'ade*', 'ide', 
'ide', in its meaning of "gourd ladle" or "dipper," keeps its indefinite 
possessive prefix 'a-, 'i-, when itself possessed, e.g., Nav. be-'ede^' (as- 
similated from *bi-'ade') "his gourd ladle," Chiricahua Apache bi-'ide' 
"his dipper," Mescalero Apache bi-'ide*. This does not in the least prove 

' See, e.g., Franciscan Fathers, A Vocabulary of the Navaho Language (2 vols., St 
Michaels, Arizona, 1912), Vol. l,p.99, sub "gourd;" Vol. 2, p. 13, sub ade', where it is defined 
as Cucurhita. 

' See Franciscan Fathers, An Ethnologic Dictionary of the Navaho Language (St 
Michaels, Arizona, 1910). 

'/fc/rf, p. 401. 

* Vocabulary, Vol. 1, p. 186, sub "spoon." 

' My Chiricahua and Mescalero Apache forms are quoted from manuscript material 
kindly put at my disposal by Dr Harry Hoijer. 

Six: Athabaskan and Na-Dene Languages 211 


that Navaho 'a-de' "one's horn" and 'ade' "gourd ladle" are unrelated 
words, for we have other examples in Southern Athapaskan of double 
possessives of type "his-one's . . ."; e.g., Navaho bi-t'a' "his (i.e., bird's) 
feather," 'a-t'a' "a (bird's) feather," but be-'et'a' "his-one's-feather," i.e., 
"his (secondarily owned) feather, his plume (used in hair decoration)." 
All this suggests that Navaho 'ade*' "gourd ladle" originally meant "ladle" 
in general and that this word in turn originally meant "an animal's horn," 
reinterpreted as "horn spoon," very much as our musical instrument, the 
"horn," originally a "ram's horn" used for blowing, is now a brass instru- 
ment with no obvious relation to an animal's horn. The semantic history 
of 'ade'' would, then, be: (1) an animal's horn; (2) ladle made of horn; 
(3) any ladle; (4) gourd ladle; (5) the gourd, Cucurhita, of which ladles are 
made. Stage 1 would be pro to- Athapaskan; 2, a dialectic Northern and 
Pacific, and presumably early Southern, development based on the wide- 
spread use of horn for spoons; 3, a Southern Athapaskan transfer of mean- 
ing due to the fact that spoons were no longer made of horn; and 4 and 5, 
a specific Navaho (in part perhaps also Apache) development. Inasmuch 
as stage 2 no longer has validity in Navaho, the meanings of the word 
group into two disconnected "sets (1; 3-5), so that 'ade'' is now felt to be 
two distinct and unrelated words, the more so as it is tabooed among the 
Navaho to use the horn of the deer for the making of spoons. My inter- 
preter Albert Sandoval once volunteered surprise that identically the same 
Navaho word meant both "a horn" and "gourd, gourd ladle." 

If we turn to other than Southern Athapaskan dialects, we find that 
the absolute *de "horn," the possessed *-de"' "horn of . . . ," and the form 
with indefinite possessive prefix *^e-de'' "an animal's horn" are found in 
both of the two other Athapaskan areas. Corresponding to Navaho de we 
have, e.g., Kutchin |i "horn" and Hupa -de- (in compounds); correspond- 
ing to Navaho -de' we have Carrier -de, Chipewyan -de (Li) (Chipewyan 
high tone = Athapaskan [Navaho, Apache, Sarsi, Kutchin] low tone). Hare 
-de, Loucheux -|i, Kutchin |i', Beaver -de', Sarsi -da', Hupa -de', Kato 
-de', and Mattole -de'; while Navaho 'a-de' has an exact correspondent 
in Chipewyan 'i-dk, Hare e-de, Loucheux e-|i, Kutchin c'i-^i', Batard 
Loucheux e-d^e, and Hupa ^i-de', all meaning "an animal's horn." The 
early use of horn for spoons, which can only be inferred for Southern 
Athapaskan, is linguistically reflected in Hupa ^ide-^in', literally "a horn's 
handle," whence "spoon," and in Hare ede-k*a "cuiller en come" (Petitot: 
k^a "plate, bowl") and Sheep Indian (esbatahot'ine) ede-ka "corne aplatie" 
(Petitot), whence "spoon." Obviously, to the Navaho mind 'ade' in its 
meaning of "gourd" must be referred to the beginning of things, for the 

212 VI American Indian Languages 2 


term is used in ritual and mythology, for example in the compound term 
Gourd Children,® but the feeling of the Navaho is of no more importance 
in the historic problem than their conviction that ly' always meant "horse" 
(though we can easily prove from comparative evidence that its original 
meaning was "dog") and that their ancestors became acquainted with the 
horse not too long after the Emergence, as indicated by the origin legend 
for the creation of the horse in the four cardinal points out of the four 
ritualistically proper materials. 

Our linguistic analysis, in short, points unmistakably to two things of 
historical interest: that the gourd is not an original element of Southern 
Athapaskan culture; and that horn spoons, not directly given by present- 
day Navaho culture, must be assumed to have been known to the remoter 
Athapaskan-speaking ancestors of the Navaho or, at the least, to early 
Southern Athapaskan culture. These inferences go well with a theory of 
immigration of the Navaho and Apache from the north (or east) into the 
Southwest. Even if one goes no further than to infer the absence of the 
gourd and the presence of horn spoons in an early phase of the culture of 
the Navaho-Apache tribes, the illumination brought by a close analysis 
of Navaho 'ade"' and its Apache cognates is useful for the reconstruction 
of the period antedating the massive influence of the Pueblos on the Navaho 
and the Apache tribes. 

2. The Navaho verb for "seed lies" is -sas, a perfective neuter, e.g., 
sisas "the seed lies," nsas "the seed lies in a row." The original meaning 
of these forms is obviously not specifically "the seed lies" but, more gen- 
erally, "the mass of finely divided particles (e.g., grain, sand) lies." A cor- 
responding active verb, nasas, means, for instance, not merely "I scatter 
the seed,"^ but also "I let the mass (of grain, sand) spill (e.g., out of a 
bag); I sprinkle it (e.g., sand, water)." I can find no cognate for these 
verbs in the material available to me from other Athapaskan dialects, and 
. the inference — as so often in analogous Athapaskan cases of apparently 
isolated verbs — is that we probably have here a dialectic denominative 
formation, i.e., a secondary set of verbs based on a noun. 

Now it is perfectly clear from Navaho phonology that all verb stems 
beginning in s (after vowels) are contracted products of a "classifier" -1- 
and either z or y; in other words, -sas must go back to either -1-zas or 
-t-yas. The perfective neuter *si-}-zas or *si-l-yas is analogous in form to 
such a perfective neuter verb as silcoz* "the fabric lies." But what is the 

« An Ethnologic Dictionary, pp. 351, 353. 

' See, e.g.. Vocabulary, sub "broadcast (in sowing)." 

•c = ts. 

Six: Athabaskan and Na-Dene Languages 213 


underlying zas or yas? Quite obviously, "snow," Navaho zas, yas (these 
two forms, of which the latter is the more archaic, constitute one of the 
few cases of dialectic difference within Navaho). Hence the verb sisas 
must have meant, originally, "it lies like (flakes of) snow," whence "the 
seed lies"; the derived active verb nasas originally meant "I scatter it 
about (so that it lies) like snow," whence "I sow the seed broadcast." 

As in the preceding case, while the present cultural term is not wide- 
spread in Athapaskan but is confined to Navaho (or Southern Athapaskan), 
it is not difl&cult to establish a close connection with a universal Atha- 
paskan term of differing cultural connotation. Athapaskan *yaxs' "snow" 
is found in the majority of Athapaskan dialects: Ingalik yi5, Babine yis 
(Jenness), Carrier y3§, Chipewyan ya0 (Li), Slave zah, Hare, Dogrib z^ah, 
Loucheux ?iow, Kutchin zah, Kaska zds (Jenness), Beaver yas, Sarsi zas, 
Kwalhioqua yaxs, Hupa yahs,^° Kato yas, Mattole yas, Jicarilla Apache 
zas, Mescalero, Chiricahua zas. The original meaning of the Athapaskan 
word is not "snow" in general but specifically "snow lying on the ground;"^^ 
another common Athapaskan term, represented by Navaho cil, means 
"snowdrift" or "falling snow." This restricted meaning, "snow lying on 
the ground," is clearly the prototype of the present Navaho term for "the 
seed lies." To summarize, a non-agricultural term ("snow lying on the 
ground") takes on a transferred and more general meaning in a classifica- 
tory verb ("the finely divided particles lie [snow-like] on the ground") 
and, in a secondary, agricultural environment, advances to the technical 
meaning of "the seed lies." No other sequence of meanings fits the linguistic 

3. The Navaho word for "corn" is nad4'. The second element, -d4', 
occurs in a number of compound nouns referring to plants in which it tends 
to be translated "corn" by Navaho interpreters, e.g., hasce'd4'' "box- 
thorn," approximately "god-corn" (hasce' is a familiar Navaho god name 
and, in slightly abbreviated form, is the first element in the native term 
for the Talking God); ci'd4' "buckthorn," perhaps "bitter corn" (cf. 
dici' "it is bitter"); ma'id4' "cedar-berries," literally "coyote-corn;" 

» My reason for reconstructing to high-toned *y4xs rather than low-toned *yixs is too 
technical to give here. 

" The Hupa word means not "snow" but "white frost (on trees)." 
" See, e.g., Fang-Kuei Li, A List of Chipewyan Stems (International Journal of American 
Linguistics, Vol. 7, 1933), p. 146: "y^ snow on the ground;" similarly, for Chipewyan, L. Le 
GofiF, Dictionnaire Franfais-Montagnais (Lyons, Marseilles, and Rome, 1916), sub "neige 
tomb^e;" for Carrier see A. G. Morice, The Carrier Language (2 vols., Modling bei Wien, 
1932), Vol. 1, p. 25, where ye§ (our yas) "snow" contrasts with eel (our Sal) "snow (heavy and 
not yet settled)." 

214 VI American Indian Languages 2 


gahcohdi' "winterfat," literally "jackrabbit-corn." The reason why, in 
compounds such as these, -di' is translated "corn" rather than "food," 
which is obviously more logical in such terms as "coyote-food" and "jack- 
rabbit-food," is probably the use of the abbreviated -d4' for "corn" in 
possessed forms (e.g., sid4' "my corn") instead of the fuller na-d4'of the 
absolutive. But it is quite easy to prove that -d4' is not, in any true sense, 
abbreviated from nad4' "corn" but, on the contrary, is an old term for 
"food" which lingers, somewhat disguised, in such compounds as have been 
quoted and in possessed forms for "corn" (sidi'' "my food" par excellence, 
whence "my corn"). This interpretation, not clear to the Navaho himself 
because the word in actual use for "food" is ci-yan and he therefore feels 
that the primary meaning of -d4' is, or should be, "corn," is at once made 
plausible from within Navaho when we compare -d4' with the medio- 
passive imperfective neuter verb -d4 "to be eatable" (e.g., yid4 "it is 
eatable"), itself closely related to the durative transitive verb -y4 "to eat 
it" (from which ciyan above is independently derived). It looks, there- 
fore, as if -d4' originally meant "what is eatable," i.e., "food," secondarily 
"corn" in possessed forms. 

The na- of nad4' is quite obscure to the Navaho. It seems to follow 
no obvious analogy and cannot be equated with the common na.-- "about, 
here and there" of continuative verbs. One might venture nad4', originally 
"corn is here and there," whence "planted corn, standing corn," finally 
generalized to "corn." This is to be taken no more seriously, however, than 
an attempt to see our common word tide in the -tide of eventide, whereas 
every historical student of English knows that this compounded -tide is 
a survival of an old word tide synonymous with time and cognate with 
Danish tid and German Zeit. Our problem cannot be considered com- 
pletely solved until we have done more than plausibly surmise that -d4'' 
originally meant "food" and have found a linguistically unforced explana- 
tion of na-. The former requirement is met by a consideration of Atha- 
paskan cognates, which reconstruct to *dan-6 (itself reduced from *dg-han-6 
"that which is eaten, food," relative form in *-g of *d5-han, whence *-dan, 
"to be eaten, to be eatable"), possessed form *-dan-e', *-dan-g' "food 
of . . .:" Chipewyan d^ne (Li), possessed -d^ne, Sarsi dani, Mescalero 
Apache dan, possessed -dan and -d4'', Chiricahua Apache dan, possessed 
-dan (also -d4' in nad4' "corn," perhaps borrowed from Navaho), Hupa 
possessed -dan' in -dan' sa'an "food of ... is lying" = ". . . is saving with 
food," Mattole possessed -dane' "possession, property" (presumably a 
meaning enlarged from "food"). These forms enable us to understand the 
exact status of Navaho -d4'. It is not the reflex of the primary *dan6 

Six: Athabaskan and Na-Dene Languages 215 


"food" but of its possessed form *-dan6' "food of . . ."; the former (exem- 
plified by Chipewyan d^n^, Sarsi dam, and Apache dan [read dan, for an 
old Southern Athapaskan *dan, monosyllabic, would have yielded Navaho, 
Apache *dan, while an old Athapaskan *dan would have given Navaho, 
Apache *d4]) would have resulted in Navaho *dani, *dan. The Apache 
possessed forms in -dan are merely generalized from the absolutive dan, 
the variant Mescalero -d4-' "food of , . ." being the true reflex of Atha- 
paskan -*dan-6' and an exact cognate of Navaho -d4' "corn." We see, 
therefore, that on strictly linguistic grounds such Navaho forms as ma'rdi"' 
mean, not "coyote-food," but "coyote's food." This makes it doubly im- 
possible to interpret nadi' as "corn here and there," which form, if it 
ever existed, would have had to yield *na-dan in Navaho. We are driven 
to infer that nad4' originally meant "food of na,--," whatever na*- may be. 
Once we see that na,'- must have referred to certain beings, human or 
animal, whose food was corn, we advance rapidly to a satisfactory linguistic 
solution. Many Athapaskan dialects have reflexes of an old word for 
"enemy, aliens," occurring in two forms (*na', na,-- in compounds, and, 
with indefinite possessive prefix, *]|e-(dg-)na-', *^e-na- in compounds). 
These words are frequently used to refer to specific neighboring tribes. 
Examples of *^e-na-', *^e-d6-na-' (*de- is collective) and compounded 
*^e-na-- are: Carrier 9-d-na, Chipewyan '^-na (Li) "enemy, Cree Indian," 
Slave e-na-kie "Eskimo," Hare e-h-da "enemy," e-na-ke "Eskimo," Dog- 
rib e-h-da, Loucheux 9-ne "enemy, Eskimo," Batard Loucheux a-ra-ke 
"Eskimo," Kutchin ce--k''6i (contracted from *ca-ne--) "Eskimo," Hupa 
]^i-na' "Yurok Indian," Navaho 'a-na-', Mescalero, Chiricahua Apache 
'i-nda\ (The -kie, -ke, -k^oi of some of these forms, analogous to Navaho 
-ke, is a plural animate sufl5x.) The old compounded form without in- 
definite prefix, *na-, is illustrated in Chipewyan na-t'j-i "enemy" (Petitot) 
(literally, "the one who acts as an enemy") and, presumably, in Kato 
na-cal "orphan" (from "alien" + "child, little"). In Navaho this na,-- is 
found in compound nouns, particularly such as refer to foreign peoples, e.g., 
na-lan "Comanche Indians" (from "enemy-many- the"), na'-sfezi "Zuni 
Indians" (contracted from na-yist!ezi "enemy"+"the ones who are black- 
ened"), nasgali (apparently made over, in accordance with the Navaho 
tribal name pattern, from masgali "Mescalero"),^^ natoho "Laguna In- 
dians" (apparently also "Isleta Indians"?). The last of these tribal names is 
interpreted as "enemies at the water" by the Franciscan Fathers^^ but a 

" See Vocabulary, Vol. 1, p. 127, sub "Mescalero Apache." 

" Ibid., Vol. 2, p. 135, sub natqoho; better "at the river," for this name is based on 
to h, possessed -td -h, "river which does not dry up," rather than on t6 "water." 

216 VI American Indian Languages 2 


more natural interpretation is to take the name as a relative in -i (assimi- 
lated to -6) from natoh "enemy-river," presumably an old name for the 
San Jose (and Rio Grande?), in contrast to the two normal interpretations 
of toh, namely San Juan River and Little Colorado River, the two rivers 
in or near the old Navaho habitat which never completely dry up. This is 
confirmed by the place name na*t6'(h) siiai' given by the Franciscan Fathers 
for Grant, New Mexico (natqo s3,kal)," literally "the enemy river has 
its legs distended," "(where) the San Jose turns crotch-wise." The point is 
of some importance linguistically as indicating that Navaho compounds in 
na"- "enemy" not only mean "... enemies" but also "enemy ..." In 
other words, both natoh and nad4' are archaic Navaho words which 
qualify basic nouns ("river" and "food") by referring them to the enemy, 
in this case the Pueblo Indians. 

The Navaho word for "corn," nad4', in summary, which can be an- 
alyzed with great probability into an older "food of the enemy," "Pueblo 
food," implies that there was a time when the Navaho, an agricultural 
people in historic times, were still thinking of corn as an alien food. Later on, 
when they had adopted corn as a staple and had built so much of their myth 
and ritual around it that it was inconceivable to them that there could be 
anything alien about it, they could not possibly feel the na- of their word 
for "corn" as akin to the -na' of 'ana*' "enemy" and the na- of tribal 
names. The sentiments clustering about the two terms had become irre- 

4. There is a curious verb stem in Navaho which seems to be used only 
in certain quite specialized verbs; this stem has the forms: imperfective 
-keh (probably error for -keh), perfective -k\, progressive and future -ket, 
usitative and iterative -keh, optative -kel. It is used in an idiomatic verb 
referring to sleeplessness, e.g., iterative bit sicanake'h "sleeplessness always 
bothers me," perfective bil sicankj "I have been sleepless." The form of the 
verb is such (bit "sleep" is subject; -ca- "away from" is preceded by the 
indirect pronominal object) as to suggest that the verb stem refers to a 
specific type of movement. My interpreter, Albert Sandoval, had no notion 
what the underlying metaphor was but said he felt, somehow, that there 
was a reference to gliding movement in it: "sleep glides (slips) away from 
me." There is no linguistic support for this feeling, which is hardly more 
than an ad hoc interpretation to fit the linguistic form. This obscure verb, 
as Sandoval pointed out, must have the same stem, in its progressive form 
(-kel), as the sacred name of the owl, cahalxel yil nakeli "darkness with- 
it the-one-who-comes-gliding(?)-back, the one who comes gliding (?) back 

** Ibid., Vol. 1, p. 226. 

Six: Athabaskan and Na-Dene Languages 



with darkness." The image of gliding is not so apposite here. The simple 
progressive form, which would be *yiket "it glides (?) along," is not in use 
in Navaho. There is nothing to be done with these isolated forms except to 
see in them survivals of an old set of verbs of movement which perhaps still 
occur in other Athapaskan dialects. 

Turning to Chiricahua Apache, we find the verb stem: imperfective 
momentaneous -ke" (continuative -ke), perfective -kj, progressive and future 
-kel, usitative and iterative -ke, optative (evidently transferred from im- 
perfective) momentaneous -ke* (continuative -ke). Its meaning is given as 
"several run, trot," which is by no means easy to reconcile with the hypo- 
thetical "glide" of the Navaho words. If the Navaho and Chiricahua 
Apache words are historically related, as is indicated by their strict formal 
parallelism, it must be because each dialect has developed specialized mean- 
ings that diverge from a third term. Now the distribution of the meanings of 
the Northern and Pacific Athapaskan verb stems which are demonstrably 
cognate to the Southern Athapaskan stems is such as to leave little doubt 
of what this third term must have been. The following table of stem forms 
gives a summary of dialectic meanings and of phonetic equivalents for four 
selected stems of the set: 


1. Ingalik "to travel by canoe' 

2. Loucheux (ditto) 

3. Kutchin (ditto) 

4. Carrier (ditto) 

5. Beaver (ditto) 

6. Hare (ditto) 

7. Chipewyan (ditto) 

8. Sarsi "to travel by canoe; 

to go for trade" 

9. Ts'ets'aut "to travel by canoe' 

10. Chasta Costa (ditto) 

11. Hupa "to travel by canoe; 

several objects float" 

































(read -keh?) 










kin- before 

' -ki- 


-XI g 



218 yi American Indian Languages 2 



Mattole "to travel by canoe' 

' -kxix 






Kato "several bathe" 



from cont. pf., 

it., and opt.?) 


Chiricahua Apache "several 

run, trot" 






Navaho (only as survival in 

obscure forms) 

(read -keh?) 



The history of the meaning of these verb stems is now reasonably clear. 
The primary meaning of the Athapaskan verb stems may have been "sev- 
eral objects (or persons) move in the water, float" (see Hupa and Kato 
above), whence "the group travels on the water, to travel by canoe (as one 
of a canoe-party)." Both meanings are preserved in Hupa. The latter mean- 
ing, however, may well have been the primary one. The specific meaning of 
a group traveling by water seems, under changed environmental conditions, 
to have taken on a new meaning in Chiricahua Apache ("several run, trot"), 
though the old plural or collective implication is still preserved. In Sarsi 
the meaning of "to travel by canoe" is now felt to be rather archaic and to 
belong to myth and story. The natural meaning today is "to go to trade, 
to go (by foot or horse-back) in order to shop;" this is developed from "to 
go by boat (or canoe) in order to trade at a Hudson's Bay Co. trading post," 
itself specialized from the common Athapaskan meaning "to travel by 
canoe." Here too the gradual passage to a typical Plains life, with little or 
no use of water craft, has brought about a redefinition of a familiar set of 
words. The Navaho words seem to stem from an old meaning "to travel by 
canoe," naturally entirely eflfaced from tribal memory. A generalized mean- 
ing "to float," applying to singular as well as plural subject, cannot be 
assumed for early Navaho because there is no evidence anywhere in Atha- 
paskan for a reflex of *-ke"x in the sense of "one person floats" and because 
all Athapaskan dialects are peculiarly sensitive to the difference between 
singular and plural forms of verb stems referring to characteristic types of 
movement. "I become sleepy," in other words, seems originally to have 
meant "Sleep paddles away from me;" the Owl was ritualistically described 
as "he who brings Darkness back in his canoe." Such locutions seem to stem 
from a cultural setting in which travel by canoe was so much a matter of 
course that it could be transferred to the supernatural world. 

Six: Athabaskan and Na-Dene Languages 219 


The Navaho nakei "he comes 'gliding' home" (of which nakeli in the 
sacred name of the owl is the relative form) is contracted from an old 
Athapaskan progressive *na-7g-(d6-)ke"l, of which there are exact reflexes 
in many of the other dialects, e.g., Sarsi na-71-kal "he's coming back on a 
boat, he's returning from shopping;" Beaver na-7a-kil (read -kel?) "he is 
paddling back;"'^ Carrier na-s-ket (contracted from *na-7e-s-ke"l) "I am 
again navigating, I am returning by boat;'"® Ingalik n3-79-d3-kal "he 
paddles again. '"^ 

The evidence collected in this paper may now be summarized. (1) It is 
assumed that there is important external linguistic evidence, distributional 
in character, to provide a prima facie probability of the northern origin of 
the Navaho and Apache. All the Southern Athapaskan dialects (Navaho, 
Western Apache, Mescalero and Chiricahua Apache, Jicarilla Apache, 
Lipan, and Kiowa Apache) obviously form a close-knit dialectic unity 
which contrasts with the more complex dialectic ramifications of Pacific 
and Northern Athapaskan. The geographical center of gravity of these 
languages, in short, lies in the north. (2) If we could find internal linguistic 
evidence in Navaho, of cultural implications, tending, as it were, to free 
Navaho and Navaho culture from their present Southwestern environment, 
the initial probability of a northern provenience would be strengthened. 
Such supplementary strengthening of an inherently probable hypothesis 
is suggested by the linguistic analysis of four Navaho words having cultural 
connotations. The cultural inferences that may be derived from this analy- 
sis are: that the gourd was not originally an element of Southern Athapas- 
kan culture; that spoons in this culture were originally made of horn; that 
broadcast sowing of seed was foreign to the culture; that maize, a staple in 
historic times, was at one time felt to be an alien food — in other words, that 
the Southwestern agricultural complex was originally lacking; and that a 
glimpse, faint but not to be lightly argued away, may be had of a time when 
the Navaho, or Southern Athapaskans collectively, made use of canoes. 
(3) All of these inferences deepen, in a historical sense, the cultural gap be- 
tween the Navaho and the Pueblos. This gap is already given, in a descrip- 
tive sense, though in lesser degree, by the modern ethnologic evidence. The 
first four of the cultural inferences we have listed are theoretically compati- 
ble with a non-Pueblo Southwestern cultural setting and, equally, with a 

" Pliny Earle Goddard, Beaver Dialect (Anthropological Papers, American Museum of 
Natural History, Vol. 10, Pt. 6, 1917), p. 506. 

" Morice, The Carrier Language, Vol. 1, p. 279. 

" John W. Chapman, Ten 'a Texts and Tales from Anvik, Alaska (Publications, American 
Ethnological Society, Vol. 6, 1914), p. 158, 1.1. 

220 ^f American Indian Languages 2 


more northern setting. The last of these inferences, if valid, points more 
positively to a northern setting. 

"Northern origin" does not in the least imply a direct line of movement 
from north to south across the Great Basin. Such a line of migration is 
most improbable. It is far more likely that the movement of these peoples 
proceeded via the western plains. If this is correct, an analysis of Southern 
Athapaskan culture would aim to reveal four strata: a fundamental north- 
ern layer, comparable to the culture of the tribes of the Mackenzie basin; 
an early western Plains adaptation, more archaic in its outlines than the 
specialized culture of the Plains as now defined by ethnologists; a first 
Southwestern influence, tending to assimilate these tribes to the relatively 
simple non-Pueblo culture of the Southwest; and a second, distinctively 
Pueblo, Southwestern influence. To these must, naturally, be added a good 
deal of Navaho specialization on the basis of the Pueblo influence. The 
disentangling of these various layers is work for the future and, in any 
event, is hardly likely to be ever more than fragmentary. Meanwhile, the 
geographical sequence: Chipewyan, Sarsi, Kiowa Apache, Jicarilla Apache, 
Navaho, may stand as a suggestion of the reality of the historical problem, 
though, no doubt, the Plains character of Sarsi and Kiowa Apache culture 
is in each instance of a much later type than the hypothetical Plains in- 
fluence to be worked out for Navaho cultural antecedents. 

Yale University 
New Haven, Connecticut 

Editorial Note 

Originally published in American Anthropologist 38, 224-235 (1936). 

A. L. Kroeber called this paper a "matter of virtuosity. . .Sapir just took out of 
his head or from scattered literature the particular little bits that he needed and 
wove them into this masterpiece" (Kroeber 1984: 133). 

Contribution to 
Cornelius Osgood, The Distribution 
of the Northern Athapaskan Indians 

Linguistic Classification within the Northern Athapaskan Area 

Linguistic classification within the Northern Athapaskan area is still for the 
most part extremely uncertain. I am indebted to Dr. Edward Sapir, whose 
authority is preeminent in this field, for the few statements and tentative sug- 
gestions which may be presented at this time. First, it appears certain that there 
are major linguistic divisions within the Northern Athapaskan area, which are 
individually equal in comparative weight to the whole Southern, or Pacific, 
Athapaskan division taken collectively. Linguistically, the difference between 
Ingalik and Chipewyan, or between Carrier and Sarsi, is as great a contrast as 
that between Chipewyan and Navaho. Whereas there is apparently a true lin- 
guistic unity in both the Southern and the Pacific Athapaskan divisions, the 
Northern Athapaskan area, as such, does not form a linguistic unit. Secondly, 
two languages of the Northern Athapaskan group, Kutchin and Tsetsaut, stand 
out from any further internal [22] alignment as individually distinct, the former 
being probably the most specialized of all Athapaskan languages. Finally, it 
seems probable that there are relationships among fifteen other groups which 
resolve themselves into six divisions beside the two given above. That these 
eight divisions given below are exactly comparable linguistic equivalents is not 
to be assumed: 

1. Kutchin 

2. Tsetsaut 

3. Tanaina — Ingalik 

4. Carrier — Chilcotin 

5. Tahltan — Kaska 

6. Sekani — Beaver — Sarsi 

7. Chipewyan — Slave — Yellowknife 

8. Dogrib — Bear Lake — Hare 

Of the remaining eight groups, Koyukon, Tanana, Nabesna, Ahtena, Han, 
Tutchone, Mountain, and Nicola, nothing is certain concerning their classifica- 
tion. Ahtena may prove to be a distinct division by itself, and it is likely that 
most of the others are to be consolidated with the divisions numbered above as 
3, 5, and 7. 

It is greatly hoped that both linguistic and cultural research may be under- 
taken during the coming decade to bring to more definite conclusions, while 
still possible, some of the problems which have been mentioned. Such conclu- 
sions will have an important and far-reaching effect on the study of the whole of 
American culture. 

222 Vf American Indian Languages 2 

Editorial Note 

Excerpt from Osgood, Cornelius, The Distribution of the Northern 
Athapaskan Indians, Yale University Publications in Anthropology 7, 21-22 
(1936). Reprinted by permission of Human Relations Area Files. 

Cornelius Osgood (1905-1985), an anthropologist who did extensive eth- 
nographic work among Athabaskan groups of Canada and Alaska during the 
1930s, was a junior colleague of Sapir's at Yale during the 1930s and had earlier 
studied with Sapir at Chicago. 

Section Seven: 
Penutian Languages 


As originally proposed by Dixon and Kroeber (1913), the Penutian lin- 
guistic stock encompassed five California groups that had hitherto been 
separately classified: Wintun, Maidu, Yokuts, Miwok, and Costanoan. In 
modern discussions of North American linguistic classification Dixon and 
Kroeber's grouping is referred to as California Penutian, whose status as a 
unit within a larger Penutian relationship is still an unresolved question 
(Whistler 1977; Shipley 1980; Berman 1983). The recognition of a wider 
affiliation of California Penutian, and the extension of the term "Penutian" 
to the larger relationship, must largely be credited to Sapir. Between 1915 
and 1925 Sapir proposed a series of additions to the California kernel, 
beginning with an Oregon group — limited at first to Takelma, Coos, and 
Lower Umpqua, but quickly expanded to include Alsea and Kalapuya — then 
Chinookan, followed by Tsimshian, a Plateau group (Klamath-Modoc, 
Sahaptian, and Molale-Cayuse), and finally by a cluster of Mexican and 
Central American outliers. Subsequently, others have used the term Penutian 
in even broader senses, but Sapir's proposals have remained the point of 
departure in any discussion of the "Penutian Hypothesis" (as for example in 
Silverstein 1979 or Greenberg 1987). 

Compared with his achievements in the comparative linguistics of Hokan 
(see Volume V), Sapir's published record in comparative Penutian is a thin 
one. He wrote only one major paper on Penutian ("A Characteristic Penutian 
Form of Stem", 1921b), which included no etymological sets or reconstruc- 
tions. It is fair to say that, for all his interest in historical questions, Sapir's 
major contributions to Penutian linguistics were descriptive rather than 
comparative. His Takelma texts and grammar (1909c and 1912h) and his 
Wishram Chinook texts (1909d) and grammatical synopses (1907c and 191 Ig 
in this volume) are the foundational works in the descriptive study not only 
of these two languages but of Penutian languages generally. Moreover, Sapir's 
perspective on the diachronic relationship of the Penutian languages was 
heavily influenced by his detailed knowledge of the synchronic facts of 
Takelma and Wishram. Any general assessment of Sapir's Penutian work 
must, then, largely be concerned with his work on these two languages of 

It was in fact Wishram on which Sapir, as a 21 -year-old novice field 
worker, cut his analytic teeth in the summer of 1905. He had been sent by 
his teacher, Franz Boas, on funds provided by the Bureau of American 
Ethnology, to gather data on the upriver dialects of Chinook. "Preliminary 
Report on the Language and Mythology of the Upper Chinook" (1907c) is 

226 ^f American Indian Languages 2 

the published version of a paper which Sapir originally delivered shortly after 
returning from the field, at the December, 1905, meeting of the American 
Anthropological Association. It is a series of comparisons between the Upper 
Chinookan (Wasco-Wishram) material that Sapir had collected and the 
Lower Chinookan data gathered earlier by Boas. For a student's first pub- 
lication of his own field data it is strikingly self-assured, showing many of 
the qualities (e.g., meticulous attention to phonetic detail) and concerns (e.g., 
the relation of linguistic to socio-psychological phenomena, as in consonantal 
symbolism) of Sapir's mature work. Of particular interest are some historical 
suggestions that adumbrate the Penutian hypothesis: that a comparison of 
Upper and Lower Chinookan phonetic systems indicates an original Chi- 
nookan homeland east of the Coast Range (p. 535), or that the "puzzling" 
occurrence in Wishram of locative postpositions reminiscent of case suffixes 
can be taken to indicate a connection of some sort with Sahaptian and 
Klamath (pp. 541-542). 

In the course of his 1905 work on Wishram Sapir collected an extensive 
vocabulary, rich grammatical data, and some 30 narrative texts. He published 
the texts in 1909, together with 25 Wasco texts collected earlier for the Bureau 
of American Ethnology by Jeremiah Curtin (1909d, in Volume VII). The 
extent of his material would certainly have allowed him to prepare a full 
grammatical description had Boas encouraged him to do so. Instead, Sapir 
was asked only to contribute several illustrative sections of Upper Chinook 
to the grammar of Lower Chinook that Boas prepared for the first volume 
of his Handbook of American Indian Languages (1911). These sections are 
reprinted here. Two are expansions of briefer treatments of the same topics 
in Sapir's 1907 report: "Diminutive and Augmentative Consonantism in 
Wishram" (cf 1907c: 537-538) and "Post-positions in Wishram" (cf 1907c: 
541—542). In addition to these relatively self-contained sections, Sapir's 
contributions to Boas's Chinook grammar also include Section 17, "Modal 
Elements: 2. ni-, 4. ga-, 5. ga-'' (pp. 578 — 579); Section 44, "Demonstrative 
Pronouns and Adverbs in Wishram" (pp. 625 — 626); and Section 45, "In- 
dependent Personal Pronouns" (the last five paragraphs of which are Sapir's, 
paraphrased by Boas; pp. 626 — 627). 

Sapir's second field trip, in the summer of 1906, was to the Siletz Reser- 
vation in western Oregon for the purpose of gathering linguistic data on 
Takelma, a moribund language of unknown affiliation, along with whatever 
incidental ethnographic information could be obtained. Sapir wrote a full 
grammar of Takelma for his 1907-09 doctoral dissertation at Columbia 
University (published as 1912h) and shortly afterwards prepared a volume 
of Takelma texts for publicaUon (1909c). Both of these works are reprinted 
in Volume VIII. He also published two papers on Takelma ethnography 
(1907b, I907d), reprinted in Volume IV. 

Seven: Penutian Languages 227 

Takelma exhibits many of the phonological and morphosyntactic features 
that have come to be recognized as characteristically Penutian. As early as 
1907, when Kroeber's study of Yokuts appeared (Kroeber 1907), Sapir had 
begun noting structural resemblances between Takelma and certain California 
languages. It was not, however, until well after Dixon and Kroeber first 
proposed the Penutian and Hokan relationships in 1913 that Sapir began 
active exploration of possible genetic links for Takelma. Apparently reading 
Leo J. Frachtenberg's grammar of Coos (Frachtenberg 1914a) first stimulated 
Sapir's interest in this work. Sapir's correspondence with A. L. Kroeber 
reveals that by mid- April, 1915, he had concluded that Coos and Takelma, 
and possibly some other Oregon languages, formed a "North Penutian" 
group that had been "cut loose" from the California ("South Penutian") 
group by the "northern intrusion of Hokan languages" (GoUa 1984: 182). 
He compiled a list of 145 potential cognates among Coos, Takelma, and 
California Penutian (see Sapir and Swadesh 1953), although his most con- 
vincing evidence was grammatical and typological. It was at this time, in all 
hkelihood, that Sapir drafted "A Characteristic Penutian Form of Stem" 
(1921b), the paper that was to be his only major contribution to comparative 
Penutian linguistics. 

This paper, however, was not completed until 1919, and its publication 
even further delayed until 1921, so that in part it reflects later developments. 
Chief among these was the broad expansion of the network of Penutian 
connections. By late 1915 Sapir had come to believe that "North Penutian" 
should also include Siuslaw and Chinookan, as well as (more tentatively) 
Alsea, Kalapuya, and Tsimshian (see GoUa 1984: 201 —203). Leo Frachten- 
berg, meanwhile, was independently pursuing a similar line of investigation, 
and in 1918 he pubhshed evidence linking Takelma, Kalapuya, and Chi- 
nookan, and suggesting (but not claiming) a connection between this group 
and the California Penutian languages (Frachtenberg 1918). Sapir recast "A 
Characteristic Penutian Form of Stem" as a reply to Frachtenberg, offering 
confirmatory evidence of a morphological nature and placing Frachtenberg's 
comparisons in the context of his own speculations. 

Of particular interest to Sapir was the possibility of a relationship between 
Chinookan and Penutian, not least because it seemed so improbable on the 
basis of morphological structure. Chinookan is a thoroughly polysynthetic, 
head-marking language (Nichols 1986) differing radically in this respect from 
most other Penutian languages, which tend to be dependent-marking, with 
nominal cases and a rather Indo-European-like distinction between nouns 
and verbs. For Chinookan to be Penutian it would either have to be assumed 
that all Penutian languages had originally been head-marking, or that Chi- 
nookan had originally been dependent-marking. Sapir saw the latter as the 
more probable development and believed that certain grammatical differences 

228 Vf American Indian Languages 2 

between Upper and Lower Chinookan (some of which he had noted in 1907c) 
indicated a general drift toward polysynthesis. In "A Chinookan Phonetic 
Law" (1926a), pubUshed in 1926 but probably written much earlier, Sapir 
explores the processes through which a typically Penutian-like clause syntax 
of pre-Chinookan might gradually have "petrified" into the phonologically 
close-knit polysynthetic words of attested Chinookan. In this paper — a tour 
de force of historical argumentation — Sapir musters internal evidence to 
show that the Chinookan ergative prefixes g- (3 sg. masc.) and c- (3 sg. fem.) 
plausibly derive from the post-pronominal element -g- that occurs with the 
ergative prefixes for other persons. This element, in turn, he sees arising out 
of an accentuated demonstrative particle in the original clause syntax. 

Sapir's proposal that Tsimshian, spoken in northwestern British Columbia, 
belongs in the Penutian family was a bold step that remains controversial. 
The idea of such a connection seems to have occurred to Sapir not long after 
he had collected a small amount of Tsimshian data from a speaker of the 
Nass River (Nisqa) dialect in February, 1915 (see Sapir's letter to Kroeber 
of December 9, 1915 in Golla 1984: 201 —202). The nature of Sapir's evidence 
remains uncertain, although marginal notes in Sapir's copies (now in the 
possession of Professor John Dunn of the University of Oklahoma, who 
provided copies to the editors) of Boas's Tsimshian grammar and texts show 
that he had noted a number of lexical resemblances (see "Sapir's Comparative 
Penutian Glosses" in this volume). In a 1917 letter to Kroeber Sapir cited 
several possible kin-term cognates between Tsimshian and Yokuts (Golla 
1984: 242 — 244), and he again alluded to this evidence (without presenting 
it) in his paper on "Nass River Terms of Relationship" (1920c: 365, note 1). 

Sapir began to explore Mesoamerican linkages to Penutian in the mid- 
1920s, at the same fime that he was noting the possible Hokan affiliation of 
other languages of that area (see 1925b). In 1924 Kroeber suggested to Sapir 
that Mixe and closely related Zoque, languages of Oaxaca, were possibly 
Hokan (Golla 1984: 409). At about the same time Roland B. Dixon inde- 
pendently suggested that Zoque was Penutian (Freeland 1930: 28, note 1). 
Sapir evidently found Dixon's evidence more convincing, for he placed Mixe- 
Zoque and Huave (whose relationship to Mixe had been postulated by Radin 
in 1919) in a Mexican Penutian group in the classificatory table of his 
Encyclopaedia Britamiica article on American Indian languages (1929a: 139). 
In the same place Sapir also suggested that Xinca (in Guatemala) and Lenca 
(in Honduras and El Salvador), as well as perhaps Paya and Jicaque (in 
Honduras), "may be remote southern outliers" of Penutian. Here he followed 
Lchmann (1920: 767), who had grouped Xinca and Lenca with Mixe-Zoque, 
although Sapir rejected Lehmann's further comparisons with Coahuiltecan 
and Chumash (i.e., Hokan). In the mid-1920s Jaime de Angulo collected 
more field data on Mixe. His wife, Lucy S. Freeland, used these data to 

Seven: Penutian Languages 229 

prepare a paper on "The Relationship of Mixe to the Penutian Family," a 
draft of which she sent to Sapir for "criticism and suggestions." The paper 
as published (Freeland 1930) incorporates many of Sapir's comments and 
suggestions and for all practical purposes is a coauthored work. In addition 
to ten signed footnotes, Sapir is almost certainly responsible for many of the 
lexical sets and for the suggested reconstructions. This paper may be consid- 
ered Sapir's last public statement on comparative Penutian. 

During the 1950s Sapir's student Morris Swadesh was the most active 
scholar in the Penutian field. In addition to his own work, which included a 
field survey of all surviving Oregon Penutian languages (Swadesh 1954), 
Swadesh also edited and published, as "Coos-Takelma-Penutian Compari- 
sons" (Sapir and Swadesh 1953), a Hst of 152 comparative Penutian lexical 
sets compiled by Sapir around 1915. This is apparently the list mentioned 
by Sapir in a letter to Kroeber on April 21, 1915 (see Golla 1984: 184-186) 
and also the data Sapir promised to "present in extenso in the future" in "A 
Characteristic Penutian Form of Stem" (1921b: 59). In his letter to Kroeber, 
Sapir calls particular attention to the sets involving grammatical elem.ents. 

"Sapir's Comparative Penutian Glosses" is a collection of Penutian com- 
parisons that Sapir jotted down in the margins of works on other Penutian 
languages, and which apparently he never thoroughly collated. A few are 
contained in the 1915 list (Sapir and Swadesh 1953), but most were probably 
made subsequently. The original publication of these marginal glosses (Swa- 
desh 1964) contains numerous typographical errors, and the material has 
been completely re-edited for publication here. Integrated with these glosses 
are the Penutian marginalia in Sapir's copy of Boas's Tsimshian texts (1912). 


In the summer of 1905 I was commissioned by the Bureau of 
American Ethnology to continue the study of Chinookan Hnguis- 
tics and, incidentally, mythology, which had been begun some ten 
years ago by Professor Boas, and the results of which, so far as 
published, have appeared in "Chinook Texts" and " Kathlamet 
texts," both bulletins of the Bureau, and in Dr Swanton's " Mor- 
phology of the Chinook Verb " and Professor Boas' " Notes on the 
Chinook Vocabulary," both of which articles appear in the Ameri- 
can Anthropologist} This published material deals with the dialects 
of the Chinookan family spoken at or near the mouth of Columbia 
river. It was therefore desirable, in order to gain a somewhat more 
comprehensive idea of the peculiarities of Chinookan grammar, to 
devote study to the extreme eastern dialects. 

The dialect or language to which the following notes refer is 
that spoken by the Indians formerly living on the northern shore 
of Columbia river, roughly speaking, from White Salmon river to 
the Long Narrows. These Indians, who are now on the Yakima 
reservation, Washington, called themselves iidxlnit, the ist per. 
sing, of which {itcxluit, ' I am an Uaxluit ') is in all probability 
the " Echeloot " of Lewis and Clark. They are known by their 
Yakima and Klikitat neighbors (tribes of the Sahaptian stock) as 
Wticxam, which, in its anglicized form of Wishram, or Wishham, is 
their common apellation to-day. The language spoken by them is, 
to all intents and purposes, the same as that of the Wasco on the 
other side of the river and of the White Salmon and Hood (or Dog) 

' Read before the American Anthropological Association at Ithaca, New York, De- 
cember, 1905. Published by permission of the Bureau of American Ethnology. 

^ Bureau of American Ethnology, Bulletins 20 and 26. American Anthropologist, 
N. S., II, 1900, pp. 199-237, and VI, 1904, pp. 1 18-147. i he phonetic system followed 
in these works is used in this article. See either Bulletin for the key. 

AM. ANTH , N. S., 9-35. 533 

232 VI American Indian Languages 2 


River Indians farther down the stream. More prominent dialectic 
differences appear when we get as far down as the cascades ; the 
dialect of this locality may be considered transitional between the 
Wishram and the Clackamas of the Willamette region. 

Viewing the Chinookan dialects as a whole, we find that the 
same general morphological characteristics apply to both Upper and 
Lower Chinook. In both groups we have the concept of the word 
as distinct from the.sentence clearly developed.' Pronominal incor- 
poration of subject, object, and indirect object in the verb ; a some- 
what elaborate apparatus of pronominal elements and pronouns 
(including the dual and an inclusive and exclusive in the first 
person dual and plural) ; a peculiar method of expressing the pos- 
sessive pronouns (these are prefixed elements related to the pro- 
noun subjects of transitive verbs) ; a characteristic use in many cases 
of invariable particles accompanied by auxiliary verbs instead of the 
use of verb-stems to express the main idea (as though one were to 
say in English: "He made the bell ding-dong " instead of "he 
rang the bell ") ; a general tendency toward onomatopoesis ; the 
extraordinary phonetic weakness of many of the verb-stems (often 
consisting of but a consonant or cluster of consonants) ; local or 
adverbial prefixes and local and quasi-modal suffixes in the verb ; 
and a thoroughly developed system of grammatical sex-gender 
(masculine, feminine, and neuter), both in the noun and in the 
verb — all of these Jeatures are shared by both the upper and the 
lower dialects. 

The first important difference between the Wishram and Lower 
Chinook is found to be in the phonetic systems of the two. Whereas 
the lower dialects affect on the whole a surd articulation (with pre- 

* Such a word, for instance, as the Wishram gatctcxcgam,. ' he took them away from 
the two (women) ' {,ga- := tense sign indicating remote past ; -tc-^ 3d sing. masc. subj. 
of trans, verb; -/- = 3d pi. obj. of undefined gender; -<:-:= 3d dual indirect object of 
undefined gender ; -x- = reflexive element indicating that object, -t- is possessed by per- 
sons referred to by -c-, here most easily rendered by ' from ' ; -eg- = verb-stem or " root " 
meaning ' take ' ; -am ^= verbal suffix generally denoting ' arriving, coming or going to do 
something,' but not quite transparent in its application to this verb) must be conceived of 
as an indivisible unity in the same sense in which a Latin form like conscripsi is an or- 
ganic unit (not merely con -(- icrib -\- s -\- l •&% agglutinated elements intelligible per se^ ; 
none of the elements in the given verb-form has any sort of meaning outside of its par- 
ticular place in such form. In other words, the word and sentence do not flow into one 
another in Chinookan. 

Seven: Penutian Languages 233 


dominance of/, /, and velar surd q over b, d, and velar sonant g), the 
Wishram is prevailingly sonant in its use of stops. Thus, where the 
Lower Chinook has o'/a, ' yellow-jacket,' and an/oVena, ' I killed 
them,' the Wishram has wa^^a and in^/u^ina. Moreover, the short u 
and i of Wishram are generally represented in Lower Chinook by long 
o and e, as seen in the latter example cited. The peculiar voiceless 
palatalized / (written l) of the Pacific coast appears in Wishram with- 
out the characteristic stop quality of the Lower Chinook ; thus we 
have Lower Chinook Lon ' three,' but Wishram lun. These phonetic 
differences, together with a number of local phonetic changes that it 
is not necessary to go into here,' would suffice to give the two 
groups of dialects a marked acoustic difference. From internal 
evidence I am very strongly inclined to believe that the phonetics of 
Wishram represents better than that of the lower dialects the orig- 
inal condition of Chinookan. Inasmuch as the phonetics of Lower 
Chinook is closely allied to that of the neighboring Coast Salish 
(such as the Tillamook and Chehalis), the interesting possibility pre- 
sents itself that the Chinookan tribes were formerly all located east 
of the Coast range and that some of them, proceeding down the 
river in their well-built canoes, came to the Pacific coast and there 
assimilated the phonetic system of their new neighbors. This, 
however, is confessedly mere speculation, and needs confirmatory 

Leaving aside these phonetic differences, perhaps the most strik- 
ing morphologic difference is in the treatment of the demonstrative 
pronouns. Both the upper and lower dialects possess different 
forms for the various relations of near the speaker, near the person 

' Characteristic Coast features found in Lower Chinook but not in Wishram are be- 
sides : the presence of the voiceless palatal spirant x- as in German ich (Wishram employs 
instead a voiceless palatal spirant x pronounced far forward, yet quite distinct acoustically 
from X', which to Wishram ears sounds like c) ; and the difficulty of distinguishing be- 
tween m and b and also n and d, a characteristic Coast Salish phonetic feature. 

^ Such a movement of the Chinook down the river would satisfactorily explain also the 
severed position of the Salish Tillamook, in Oregon, who are separated from the linguistic- 
ally related Chehalis only by Chinookan tribes. Even though all the Salish tribes be of 
interior provenience, as generally believed, their occupancy of the Pacific coast, including 
the region directly north and south of the Columbia, may have long antedated the coming 
to the coast of the Chinook. See A. B. Lewis, "Tribes of the Columbia Valley and the 
Coast of Washington and Oregon," Memoirs of the American Anthropological Associa- 
tion, 1906, I, pt. 2, p. 198. 

234 VI American Indian Languages 2 


addressed, and near the person spoken of, and both distinguish the 
three numbers and the three genders of the singular in the demon- 
strative. Whereas, however, the Lower Chinook further distin- 
guishes between visibiHty and invisibility of the person or object, no 
such difference could be observed in the use of the demonstratives 
in Wishram. Moreover, the principle of formation of the demon- 
stratives is, in detail, quite dissimilar in the two groups. In Lower 
Chinook the demonstrative is built up of three exceedingly weak 
phonetic elements : a consonant expressing visibility or invisibility, 
a vowel or consonant denoting the number and gender of the person 
or object referred to, and a consonant or two vowels defining the 
demonstrative relation. In Wishram the principle of formation is 
simpler; each demonstrative form is built up of two agglutinated 
syllables, one of which is the short form of a 3d pers. pronoun 
(defining both gender and number), and the other a characteristic 
element indicating the demonstrative relation. Speaking generally, 
the demonstratives in W^ishram seem to stand in much closer rela- 
tion to the personal pronouns than they do in the lower dialects.' 

Reference was made above to the general tendency toward 
onomatopoesis in the Chinookan dialects. The impression which 
Professor Boas had obtained of such a character in his study of the 
lower dialects was in every way confirmed by my own study of the 
Wishram. The frequent rhetorical lengthening or shortening of 
vowels and consonants, the duplication or quintuplication of imi- 
tative elements, and the frequent use of ononiatopoetic particles in 

* For convenience of comparison the demonstratives of both Lower Chinook and 
Wishram are tabulated below. Those in parentheses are the Chinook correspondents of 
the Wishram forms immediately above : 

Near ist Person f ddnya 
hU \{xi'k) 

Near 2d Person f ydxdan 
iste \ (x'iait) 

Near 3d Person j ydxia 
ilk \ yx-ix-) 

Shortened Pro- 
nouns in Wish- yni^x^ 

The Lower Chinook forms here given are those implying visibility. The correspond- 
ing demonstratives used to refer to invisible objects are obtained by changing the initial 
X- -Xo q-. 








dduda \ 

{x-uik) r 







ddxdau 1 
[xila) \ 



{^x tcta ) 




ddxia "1 
{xota) \ 






cda ( JT ) 

da {x) 

Seven: Penutian Languages 235 


lieu of verb-stems are not the only phenomena which illustrate this 
onomatopoetic tendency. Most characteristic of Wishram, and prob- 
ably of the other Chinookan dialects also, is the employment of a 
series of changes in the manner and, to some extent, in the place of 
articulation of the various consonants, in order to express diminution 
and augmentation. This singular rhetorico-grammatical process 
works in such a way that all surd and sonant stopped consonants be- 
come exploded consonants (better known as " fortis ") to express the 
diminutive idea (i. e. b and p become />/, d and / become t!, g and 
k become k.'\ while all surd and exploded consonants become 
sonant to express the augmentative (i. e. p and pf become b, t and 
/.'' become d, k and k! become g, q and q! become ^); in the 
case of the velar consonants a possible change to the "fortis" to 
denote the diminutive is attended also by a more forward, i. e. 
palatal, articulation (i. e. g and q become, not q!, but k!). More- 
over, the sibilant consonants c, tc, and tcf on the one hand, and s, 
ts, and tsf on the other, are related to each other as augmentative 
and diminutive consonants, while dj may sometimes, though rarely, 
be employed as the augmentative grade of tc and ts (e. g., idjik-) 
djik ' big wagon,' from itsiktsik ' wagon.' The guttural spirant 
X becomes x in the diminutive form. Subjoined are a few illus- 
trations for the purpose of making the process more easily under- 
stood. The normal word in Wishram for ' hip-joints ' is ck.'dlkal. 
The diminutive of this word is skfdlkal, in which, it will be noticed, 
the c of the first word has been changed to s in consonance with 
our rule. The word skldlkal would be appropriately used to 
designate a baby's hip-joints, for instance. On the other hand the 
augmentative would require a change of the fortis k! to a sonant g 
— hence cgdlkal is used to denote ' big hip-joints,' as of a giant. 
Similarly, while a^.^^a*/ with velar fortis (^/) is the normal word 
for 'knee,' akliixt with palatal fortis (yk!) and guttural spirant 
pronounced farther front i^x) is the diminutive, and aodxt^\\\\ sonant 
velar (^) the augmentative. Not infrequently there is a slight 
change of meaning accompanying the phonetic change. Thus, 
while itdVndn (masc.) denotes ' eagle,' itts.'V non (neut.) with 
diminutive consonantism means ' bird '; ik.'dlaniat denotes ' stone,' 
but igdlainat with augmentative consonantism means ' rock.' It 

236 Vf American Indian Languages 2 


must not be supposed that this characteristic consonant-gradation 
is confined to the noun ; it is found just as well in every other part 
of speech. An example of its use in the verb will serve to give an 
idea of its rhetorical possibilities. InigEltcim is the normal word 
for ' I struck him with it.' If the verb-stem -tcim appears, with 
diminutive consonantism, as -ishn, it implies that the person struck 
is small ; if the verbal prefix -gEl-, which implies in this case 
intent to hit, is pronounced -k/sl-, the implication is that the 
missile used is a small one. Hence we have four forms : inigAlicim 
' I hit him with it,' inigHltsivi ' I hit him (a child perhaps) with it,' 
inik.'Eltcim ' I hit him with it (something small),' and inik.'kltsim 
' I hit him (a child) with it (something small).' It would seem then 
necessary, so far as Chinookan grammar is concerned, to allow as a 
regular grammatical process, alongside of reduplication, vowel 
change or "ablaut," and pre-, in-, and suffixation, a fourth process 
— consonant-gradation or " ablaut." 

Turning again to morphology, there was one feature which was 
well calculated to arouse a certain degree of surprise. The work 
which had been done on Lower Chinook disclosed a paucity of 
tenses that is, on the whole, quite in accordance with the general 
morphologic character of many American linguistic stocks. In 
VVishram, however, I found that it was necessary to distinguish 
carefully six tenses : ist, a tense characterized by the prefix ga- 
(before consonants) or ga/- (before vowels) in certain cases option- 
ally by the prefixed consonant w-), which refers to time long past, 
say more than one year ago, and which is used regularly in the re- 
cital of myths ; 2d, a tense characterized by the prefix ;/z- (before 
consonants) or fiig- (before vowels), used to refer somewhat indefi- 
nitely to time past and which is used in speaking of events that 
happened say less than a year ago, yet more than a couple of days ; 
3d, a tense characterized by prefixed nn- (before consonants) or tia/- 
(before vowels) and sufifixed -a, which seems to refer to recent time 
exclusive of to-day, more specifically to yesterday ; 4th, a tense 
characterized by prefixed t- (before consonants) or ig- (before 
vowels), which refers to an action already performed to-day ; 5th, 
a tense characterized normally by suffixed -/", referring to an action 
now going on but, as it seems, with the implication of its soon being 

Seven: Penutian Languages 237 


completed ; and 6th, a future tense, normally characterized by pre- 
fixed a- (before consonants) or al- (before vowels) and suffixed -a} 
Besides this series of six positively characterized tenses, I should 
not omit to mention that some verbs, when referring to present 
time, are morphologically tenseless, and seem to form their immedi- 
ate past tense by a verbal prefix -/- which ordinarily denotes action 
toward the speaker.^ 

In this connection I may also mention a group of verb-forms 
which are characterized by the consonant / (assimilated in nasal 
surroundings to «) suffixed or infixed to the verb-stem, sometimes 
by -lal (or -7ian) suffixed to the verb-stem. These forms denote 
frequentative or continuative action and, as a rule, do not allow the 
verb to be further characterized by a tense element. They may 
then, from a certain point of view, be considered as forming a 
seventh tense — the present tense with no implication of comple- 
tion.^ The most interesting point about these /-frequentatives is 

^Examples — 





f gay^y^ 

J 'he went ' 

' «/yuya 

f wayuya 

1 galaya. 
\_ 'she went ' 





' he saw him' 




' he became' 

'they two did to me' 




Pres. /-Form 

FuT. rt-FoRM 

i iyiiya. 


f ayuya 
\ a/uya 





'he looks at him 



^Thus uxt (=a -\-tt -{- xt) means 'she is seated,' but 'she was sitting' is rendered 
by dtxt, in which the prefix -«- has been changed to -t-. Cf. , for this interchange, 
^wgwat 'they fly (away from me) ' and <Vgat 'they fly toward (me).' 
^ Such frequentative forms are : 

With Tensb-sign Frequentative 

^atksanbnaj'x tksanbnawi'x 

' they jumped in the water ' ' they keep jumping in the water ' 

(verb-stem bna-) 


VI American Indian Languages 2 


that certain verb-stems apparently infix the / or ;/. If our EngHsh 
word ' look,' e. g., were also a Wishram verb-stem, ' he looked at 
it ' would be itciiilook, but ' he keeps looking at it ' would probably 
be tcmloolk} I pass over many other verbal peculiarities, such as 
the distributive suffix -yit (alxk/wdya ' we shall go home,' but 
alxklivdviiiva ' we shall go each to his own home ') or the passive 
suffix -ix {itcudxuui ' he ate it up,' hut yulxu mix 'it is eaten up') to 
mention the considerable difficulty experienced in analyzing the noun, 
apart from its syntactic elements which are transparent enough. 

The pronominal elements prefixed to the noun (every noun is 
either masculine singular, feminine singular, neuter singular, dual, 
or plural) are in Lower Chinook identical with the pronominal 
object elements incorporated into the transitive verb, except for the 
feminine singular, which in the noun shows o- (from original wa-) 
as compared with -a- in the verb. In Wishram, however, the noun 
has prefixed a pronominal element differing from the corresponding 
element in the verb by an initial zv-{masc. and fem.)or z- (neuter, dual, 
and plural). The following table shows the corresponding elements : 

Noun in Wishram 

Noun in Chinook 

Obj. in 

Wishram Verb 


wi-, i- 




wa-, a- 








ic-, {is-) 

^-, {S-) 

-C-, {-S-) 







*he jumped ' 

'he keeps jumping ' 


tccinq ! w6'^k°na«p^ 

* he grasped him wi 

th his 

claws ' 

' he scratches him ' 



' she gathered driftwood ' 

' she keeps gathering drift 




'he flew' 

' he flies about ' 

' Some examples of th 

lis phenomenon are : 

With Tensk-sign 




' he hid himself 

'he is hiding himself 



' he used it ' 

' he keeps using it ' 



' she looks at me ' 

' she keeps looking at me 

(verb-stem /k- 

Seven: Penutian Languages 239 


The choice between wi- tva- and i- a- in Wishram is dependent 
chiefly upon considerations of syllabic length : wilx ' land ' (cf. 
Chinook z'le'e), but ?^nuk ' beaver '; icdtrnX ' marrow ' (cf. Chinook 
d'mala), but ^^agilak ' woman.' 

It had been hoped that some light would be thrown on the 
derivative elements of the noun, but it cannot be claimed that all 
desirable success was attained in this direction. Perhaps the most 
transparent derivative elements that were found are the suffixes -/// 
and -mat. The former of these seems to denote a group, particu- 
larly a grove of trees. Thus the word ilibuni ' apple ' (borrowed, 
of course, from the French /a pomrne) forms the derivative noun 
UibumElit ' orchard.' The suffix -?;iai is perhaps best defined as 
denoting 'something used for so and so.' For instance, isqxiis 
denotes ' the eyes,' isqxusniat means ' something for the eyes,' i. e. 
spectacles.' An interesting group of nouns is formed by descriptive 
verb phrases, such nouns being in effect pure verb forms. The 
loon, e. g., is described as 'he shouts along the river' {tci-iliiviat 
tviviat)y and ' telescope ' is rendered by ' people keep looking 
through it ' {(jexgElgElini). 

The most puzzling linguistic phenomenon found in Wishram, 
because it is at complete variance with what we have in the lower 
dialects, is the use of a certain number of loosely tagged on postposi- 
tions, in some cases optionally prepositions. We have a suffix -ba 
denoting ' in ' or ' at,' a suffix -iarnt meaning either ' towards,' or 
' from,' a suffixed or prefixed element bdvia meaning ' for,' the post- 
er pre-positions dniKiii and knEgi meaning ' with ' or ' made out of,' 
and an element -^i"/, meaning ' when,' suffixed to verb forms. ^ The 

^ Further examples of this suffix are : igictxmat 'load' (from verb-stem -ctx ' to 
carry on one's back') ; itkHcitmat 'tools' (verb-stem -cit 'to use') ; aklixivacamat 
' plane ' ( verb-stem -xwac ' to plane ' ) . 

^ The following examples illustrate the use of these elements with nouns, pronouns, 
and verbs : 

ba : wimatba 'in the river; ddiiyaba wilX 'in this country' (lit. ' this-in 
country ' ) ; galcigElkElba ' where he saw him ' (gatcigklkEl ' he saw him ' ). 

iamt : ivimajiamt ' to or from the river ' ; imigdii naikdyaint ' you are bigger than I ' 
(lit. 'your bigness [is] me-from, compared with me ') ; dtpXiamd agdiax ' to where she 
goes out towards [us] {atpX ' she goes out towards'), the sun,' i. e. 'east.' 

bdma : cdn bama ' for whom ?' ; Miilmul bama ' from, belonging to Fort Simcoe.' 

dniEni : igdbEnac aniEni ' made out of young oak.' 

EnEgi{ngi) : aqti'wiqxi ngi ' with a knife.' 

bEf: gayuyabEt 'when he went'; nkldckacbEt ' when I was a child.' In length- 
need form bd^^ it means ' as soon as' : gayuyabd't ' as soon as he went.' 

240 VI American Indian Languages 2 


extent of pronominal incorporation of indirect objects and the use 
of local or relational prefixes in the verb are such in the Chinookan 
dialects that the employment of these local tags (one might be 
inclined to call them " cases," if they had less individuality) seems 
quite unnecessary. It is of considerable theoretic importance, there- 
fore, to note that the neighboring Sahaptian dialects, quite similarly 
to the Klamath, make an extended use of such case-suffixes. We 
would then have here a good example of the graniniatic, not merely 
lexical, influence that dialects of one linguistic stock may exert on 
geographically contiguous dialects of a fundamentally distinct stock.^ 
In conclusion a few words may be devoted to the mythology of 
the Upper Chinook. I have not as yet enough texts of myths to 
present a really complete description of the mythologic concepts 
and elements present in the tales of the Wishram, but some of the 
main points seem patent enough. As in other Indian mythologies 
it is believed that there was a time antedating the present one when 
animals walked about as men, though having approximately the 
same mental and, to a large extent, physical characteristics as now. 
At that time, when there were no Indians, properly speaking, in the 
country, but only anthropomorphic animals, many things were not 
as they should be, and, in order to make the country fit for habita- 
tion by the Indians destined to hold it, it was necessary for a culture- 
hero or transformer to rectify the weak points in creation. This 
transformer is, as in the plateau regions to the east, the Coyote. 
There is a cycle of myths made up of local tales telling how Coyote 
traveled all the way up the Columbia river, transforming monsters 
and instructing the people in the various arts of life. This string of 
local tales is, if I am not mistaken, continued in unbroken succes- 
sion by the Sahaptian tribes living farther up the river, so that we 
have here a series of myths, belonging together yet distributed over 
a large number of different tribes. Some of the things that Coyote 
does are : to stock the Columbia with fish that had been withheld 
from the rest of the world by two women ; to transform two women, 
who entice wayfarers, into birds ; to provide the people of the 

* Of the postpositive elements given above, three, bdma, EnEgi, and dmEni, are 
certainly of Sahaptin origin, probably also -ba (cf. Yakima -pa ' in ' ). This explains their 
entire absence in Lower Chinook. 

Seven: Penutian Languages 241 


Cascades country with moutlis that had formerly been lacking ; to 
instruct men in the art of catching white salmon in basket traps and 
of spearing and steaming salmon ; to put an end to the atrocities of 
the merman who swallows canoes with men and all, and of the dread 
woman, At!at!aha, who steals children- and roasts them on an 
island still pointed out at the Long Narrows; and so on. In all 
this Coyote is distinctly the benefactor of mankind, but at the same 
time he is, as often elsewhere, conceived of as cunning, deceitful, 
and gluttonous. In some stories, particularly in such as do not 
belong to the cycle of Coyote as Transformer, he is an insufferable 
marplot, as when he, contrary to Eagle's injunction, opens a box 
containing the souls of his and Eagle's wife and son, thus bringing 
death into the world. At the same time he is indescribably obscene ; 
some of the deeds of this kind performed by the culture-hero of the 
Tillamook, as communicated by Professor Boas, are also told by 
the Wishram of him. Although Coyote is the main transformer^ I 
think it would be incorrect to speak of him as the hero of the 
Wishram. This point comes out clearly when Coyote himself, in 
one of the transformation myths, admits that he is no chief, that 
title being reserved, among the animals, for the Eagle and the 
Salmon. These two may, indeed, be considered the true heroes of 
Wishram myth, their deeds being narrated with considerable sym- 
pathy and admiration. The Salmon, in particular, may be described 
as the local hero of the Chinookan tribes, an elaborate salmon myth 
being common to both the Lower Chinook and the Wishram. I 
cannot say definitely whether Bluejay, who figures so prominently 
as buffoon among the coast tribes, such as the Kathlamet and Qui- 
naiult, occupies a corresponding position among the Wishram. So 
far as the material collected is concerned, he is quite a subordinate 
character, and I suspect that he is almost entirely superseded by 
Coyote. The mischievous and spiteful elements of his character, as 
of the Mink of more westerly and northerly regions, are embodied 
also in the Weasel. 

Besides the main type of myth — i. e. the Transformer or Cul- 
ture-hero myth, one can discern also a species of nature myth that 
is somewhat different in character. This type is represented, e. g., 
by the tale of the contest between the East Wind and the West 

242 VI American Indian Languages 2 


Wind, in which the latter proves successful. Another example of 
this type is the struggle of the five East Wind brothers with the five 
Thunder brothers, resulting in the death of all but one of the latter, 
which exception accounts for the existence of a certain amount of 
thunder to-da}\ 

The single myth motives of Wishram mythology are many, 
probably most of them, found distributed over considerable areas 
elsewhere. Such well-known incidents as the magic increase of a 
small amount of food, the blundering imitation of the host, the life 
and death contest at gambling bones, the unsuccessful attempt to 
destroy strangers in an overheated sweat-house, the abandonment 
and later enrichment of a poor boy while his maltreaters are 
starving — all these and many others are common property of the 
Northwest Pacific coast and regions to the east and south, though 
the setting in which they occur may vary indefinitely. On the 
whole, the chief interest of Wishram mythology seems to lie in its 
transitional character between the mythologies of the coast and of 
the plateau. Although it shares, as we have seen, a local and 
specifically Chinookan salmon myth with the Lower Chinook, many 
of the myth motives are not duplicated farther down the river, 
but are found in other regions, such as the plateaus. Here again 
we observe that linguistic and cultural, more specifically mythologic, 
distribution areas are by no means necessarily congruent. 

Editorial Note 

Originally published in American Anthropologist n.s. 9, 533 — 544 (1907). 
Reprinted by permission of the American Anthropological Association. 


Contributions to Franz Boas, Chinook 

Diminutive and Augmentative Consonantism in Wishram 

Very characteristic of Wishram, as also without doubt of ail other Chinookan 
dialects, is aseriesof changes in the manner, and to some extent in the place, of 
articulation of many of the consonants, in order to express diminutive and aug- 
mentative ideas in the words affected. This peculiar process of "consonantal 
ablaut," though perhaps most abundantly illustrated in the case of the noun, is 
exemplified in all parts of speech, so that it has almost as much of a rhetorical as 
of a purely grammatical character. Of the two series of consonantic changes 
referred to, that bringing about the addition to the meaning of the word of a 
diminutive idea is by far the more common, an actual change to augmentative 
consonantism hardly being found outside of the noun. The main facts of conso- 
nantic change may be briefly stated thus: To express the diminutive, non-fortis 
stopped consonants become fortis, the velars at the same time becoming back- 
palatals (the treatment of velar stops, however, seems to be somewhat irreg- 
ular); c and its affricative developments tc and tc! become s, ts, and ts! (s seems 
sometimes to be still further "diminutivized" to ts, ts to ts!, so that c, s, ts, ts! 
may be considered as representing a scale of diminishing values) ; x becomes x, 
in analogy to the change of velar stops to back-palatal stops just noted; other 
consonants remain unmodified. To express the augmentative, fortis consonants 
become non-fortis (generally sonant) stops, no change taking place of back- 
palatal to velar; s, ts, and ts! become respectively c, re, and tc! (in some few 
cases ts and tc fricatives become dj, pronounced as in English judge, this sound 
not being otherwise known to occur in Wishram); other consonants remain 
unmodified. [639] The following table of consonantic changes will best make 
the matter clear: 

Normal Diminutive Augmentative 

b, P p! (b) 

d, t t! (d) 

g> k k! (g) 

^, q k!, (g, k) (^) 

qx kx (^) 

q! k!, (kx) (qx) 

p! (p!) h 

t! (t!) d 

k! (k!) g 

c s, ts (c) 

tc ts (tc), (?)dj 

tc! ts! (tc!), dj 

s (s) c 


VI American Indian Languages 2 





(ts), ts! 




tc, dj 
td, (?)dj 



On the whole, there is a distinct tendency to have all the consonants of a word 
bear a consistent diminutive or augmentative coloring, though absolute con- 
cord in this regard is by no means always observed. In general it may be said 
that c and s sounds are most easily varied in accordance with our rule. Final 
non-affricative stops seem incapable of change. It often happens that the nor- 
mal form of a word is itself partly diminutive in form owing to its meaning; in 
such cases the form may be still further "diminutivized" if it is desired to give 
the word a more than ordinarily diminutive force. Thus -k!ac- in il-k!a'c-kac 
'child' is evidently a semi-diminutive form of the stem-syllable -kac; 'little 
child, baby' appears in more pronouncedly diminutive form as itk!a'skas 
(Wishram Texts 176.3). 

The following table of body-part nouns will serve as a set of examples of 
diminutive and augmentative forms. The diminutives would naturally refer to 
the body-parts of a tiny child, the augmentatives to those of an abnormally large 
being, as a giant. 




i-p!a'qxa 'flat-headedness' 



i-gE'tc 'nose' (aug.) 


i'i-pc 'foot' 



i-qxwi't 'leg' 


a-q!o'xl 'knee' 



a-ms'luqtan 'cheek' 


i-mElExtk!u'lamat 'tongue' 


i-mi'ct 'lips' 


i-k^cxa't 'mouth' 


wa'-kcEn 'finger' 


is-qxu's 'eyes (dim.)' 


id-mE'qco 'face-hair' 


i-k!wa'yat 'crown of head' 


a-tckE'n 'shoulder' 


wa-qxa'tc 'breast' 


i-kxa'tc 'tooth' 


i-q!a'qctaq 'head' 


ic-k!a'lkal 'hip-joints' 



is-q!wd'gwdst 'jaws (dim.)' 


a-mu'q!w>al 'paunch, stomach' 


1. Cf. wa-q!a'lc 'thorn', dim. Wa-kxa'ts (Wishram Texts, p. 261). 

Seven: Penutian Languages 


Examples of other than body- 


it-q!u'tcu 'bones' 

i-tcfi'au 'snake' 
i-tsi'ktsik 'wagon (dim.)' 

i-cgi'lukc 'wolf (aug.)' 

da-ga'c 'yellow' 
i-cga'n 'cedar board' 

i-kfa'lamat 'stone' 
a-k!a'munaq 'fir' 
il-k!a'ckac'c\\M (dim.)' 
a-t!u'-gagilak 'good, strong 


part nouns are: 

it-q!utsie'lxlEm 'dog' (literally, 

eater of small bones) 
is-ts!i'ktsik 'buggy' 

il-skli'luks 'new-born wolf 

cub (Wishram Texts 56.30) 
a-qx-k!i'c 'gold' 
wa-ska'n 'box' 
wa'-tsk!un 'cup' 


'heavy truck' 



'strapping big 

In these lists, (dim.) and (aug.) mean that the words so designated are wholly 
or partly diminutive or augmentative in consonantism owing to their primary 
significance. In i-pfa'qxa, for instance, the diminutive notion implied by p! is 
easily understood if we remember that head-flattening is associated with 
infancy. In some cases a consonant change involves or is accompanied by a 
vocalic change; it seems that the change of a to m or £ has in itself more or less 
diminutive force (cf. wa'-tsk!un from wa-ska'n with ila-k!d'its 'very little' [Wish- 
ram Texts 176.3] ordinarily -kfaits 'small'). The case of i-cga'n as compared with 
wa-ska'n and wa'-tsk!un illustrates the fact that the diminutive form of a noun 
often has a specialized meaning of its own. A few more examples are: 


i-tc!i'ndn 'eagle' 
i-tc!i'laq 'cricket' 
i-q!apca'lwac 'turtle' 
a-tca'la 'grindstone' 


il-tsfi'non 'bird' 
i-ts!i'laq 'grasshopper' 
is-k!a'psalwas 'lock (of door)' 
a-tsa'la 'file' 

It will be observed that several nouns on becoming diminutive in form at the 
same time change to a more suitable gender, masculines often becoming femi- 
nines (e.g., wa-ska'n), neuters (e.g., ii-sk.'i'luks), or diminutive duals in is- 
(e.g., is-k!a'psalwas). Most examples of diminutives and augmentatives hith- 
erto given have been formed from nouns that in themselves have no necessary 
diminutive or augmentative force. Other examples than those already given of 

246 ^f American Indian Languages 2 

words with inherent diminutive force, and hence with at least partial dimin- 
utive consonantism, follow: 

a-k!u'ksk!iiks 'ankle' 

a-p!u'xp!ux 'elbow-joint' 

i-pfu'xc 'cotton-tailed rabbit' 

a-t!antsa 'crow' (contrast i-cka'lax 'raven') 

i-skfu'lya 'coyote' (? cf. i-cgilukc 'wolf') 

a-gu'sgus 'chipmunk' 

a-p!una'tsEktsEk 'mosquito' (? cf. - bEna 'jump') 

is-ga'k!aps 'hat' 

i-k!a'its 'smallness' (contrast -^ail 'bigness') 

i-k!a'stila 'crab' 

H-xan '(somebody's) child' 

i-sklwo'latsintsin 'swallow' 

wa-tsk!E'nLX 'nit' 

Particularly instructive as indicating a live feeling for diminutive consonan- 
tism are such words as a-lik!u'k 'chicken' and a-lap!u's 'cat' borrowed from 
Chinook jargon (p in -pus would not be consistent [642] with diminutive s). It is 
perhaps not too far-fetched to recognize augmentative consonantism in the fol- 
lowing nouns: 

i-ga'nuk 'beaver' 

i-gii'nat 'Chinook salmon' (contrast wa-tsu'iha 'blue-black salmon') 

i-CE'lqcElq 'porcupine' 

ic-ga'kwal 'eel' 

i-du'iha 'buffalo' 

ic-kcku'ct 'testicles (contrast is-qxu's 'eyes') 

i-gu'cax 'sky' 

ic-gwo'lala 'gun' 

wa'-itc 'tail of mammal' 

Ic-li'ct 'fish-tail' (contrast is-p!i'ost 'tail of bulb, dried fish') 

It sometimes happens that a change to diminutive consonantism implies not 
so much the diminutiveness of the object referred to as a sense of endearment. 
This seems particularly true in the case of certain terms of relationship: 

N on- diminutive Diminutive 

-qcE-n 1 'man's son's -/:.'fl'c-w-c 'paternal grandfather' 

ga'c-u (vocative) j child' 

-gak-an 1 'man's daughter's -ga'/c.'-w-c 'maternal grandfather' 

ga'g-u (vocative) /child' 

-gi-an 'woman's son's child' -k!i-c 'paternal grandmother' 

Seven: Penutian Languages 


Interesting as examples of augmentative consonantism are the names of 
Coyote's four sons, all of which are derived from words denoting body-parts of 
the salmon. The augmentative consonantism implies the lubberliness of 
Coyotes sons. 

Body-parts of salmon 
i-k!la'tcin 'salmon-head gristle' 

i-ksa'lk!uts 'backbone of fish' 

i-qfwi'nan 'fin' 

a-k!a'tk'^tgwax 'adipose fin' 
(? better -qfa'tk'^tgwax) 

Names of Coyote's sons 

Sipa'-glatsin 'Big Gristle' (Wishram 

Texts 66.5) 
Sipa'-ksalguts 'Big Backbone' 

(Wishram Texts 66.6) 
Sapa'-gwinan 'Big Fin' (Wishram 

Texts 66.7) 
Sapag-a'thHgwax 'Big Adipose Fin' 
(Wishram Texts 66.8) 

As has already been remarked, the noun is not the only part of speech that 
illustrates the consonantal play here discussed. Adverbs and particle verbs of 
appropriate meaning sometimes show diminutive consonantism: tsfii'nus 'a lit- 
tle'; sdk! 'to whistle'; sa'u sau 'to whisper' (contrast Lower Chinook cdu)\ 
Lower Chinook k!a 'and' may be diminutive to ka. The diminutive form of a 
particle verb denotes a less intense state of being or activity than its correlative 
form. Sometimes its meaning is considerably specialized: [643] 

N on- diminutive 
tcic 'cold' 

ma'ca 'to spoil' 

gut 'to break up (earth) by digging' 

Possibly also: 
wax 'to pour out' 
Iq.'up 'to cut' 


(ts/u'nus) a- its d's 'just (a little) cool' 

(Wishram Texts 190.15) 
ma'sa 'to be ashamed' 
kfu'tkfut 'to pluck' 

wax 'to set on fire; to bloom' 
fkfup 'to shoot' 

The dual in is- is not the only example of a diminutive form of a purely gram- 
matical element. The diminutive stem -qlwa'lasup 'fast running' occurs with 
possessive prefixes showing diminutive consonantism. Thus the normal ele- 
ments -tea- 'her' and -cda- 'of them two' appear as -tsa- and -st!a- in i-tsa-qfwa'l- 
asup 'she runs fast' (Wishram Texts 66.9) and i-st!a-q!wa'lasup 'they two run 
fast' (Wishram Texts 66.13). Similarly, in a song (Wishram Texts 94.23), where 
the reference is to is-p!i'ast 'tail of bulb', a noun of diminutive form, the pro- 
nominal element cd- and the post-positive local element -ha appear as -st 
(? better st!-) and -p!a. Thus: 

staimapid' giskip'.i'ast 'it-alone-at the-my-tail' 

Finally the verb may show diminutive consonantism, partly in the stem itself, 
partly in its local and adverbial prefixes and suffixes, partly and most frequently 

248 VI American Indian Languages 2 

in its pronominal prefixes. Examples of verb stems in distinctly diminutive form 
are not exactly common, but certain cases seem clear enough. Thus gaqiulatla'- 
ulx 'he was tossed up' (Wishram Texts 84.26) and gatciutat!a'mElq 'he swal- 
lowed him by sucking him in' evidently contain a diminutive form of the verb 
stem -lada- 'to throw away'; ^//w'^^wajc 'it trembles' (Wishram Texts 116.10) and 
gasi'ximk!na-uk"atsk 'he looked around' (Wishram Texts 30.6) show dimin- 
utive consonantism both in their stems {-skw- and -k!na-u-) and in their first 
incorporated pronominal objects (dual s-), the latter verb also in its adverbial 
suffix -tsk, doubtless the diminutive form of -tck 'up from position of rest'; 
gats(s)altsgi'ma 'he laid her belly up' (Wishram Texts 56.27) shows diminutive 
consonantism in both stem (-tsgi) and incorporated pronominal subject (-ts-) 
and first object (dual -s-). 

We have already given -tsk as an example of a derivative suffix with dimin- 
utive consonantism. Other such suffixes are -p!a 'slightly out (of position)' 
(from -ba 'out') in ayulapfa'tcguxwida 'it will tilt up', literally 'it will spon- 
taneously move out up from its sitting [644] position' (Wishram Texts 184.10) 
and tsu (from -tcu 'down') in ililu'stsu '(water) moved down into the (hollow 
place)'. As examples of diminutive forms of local prefixes may be given -klEl- 
(from -gEl- 'directed toward') in ga-tssi'k!Elutk 'he looked at him' and its reflex- 
ive correlative -xeI- (from -xeI) in gasi'xElutk 'he looked'; -sklEm- 'under' in 
iniasklEmla'datcu 'I threw it down under her' is doubtless diminutive to -gEtn- 
'next to' (cf. -tcu and -s-tsu above). 

The only examples of diminutive consonantism in the pronominal prefixes of 
verb forms occur in the case of ts (for tc, third person masculine subject tran- 
sitive) and s (for c, third person dual subject intransitive and transitive and 
object transitive). Whenever the object of the transitive verb (or the apparent 
subject, really first object, of the "half-transitive" verb) is diminutive in form, 
the pronominal prefixes tc and c appear as ts and s; the ts by no means implies 
the diminutive character of the transitive subject. Examples are: I'wi gatssu'x 
isie'nqxoq 'he looked at his fish-line' (Wishram Texts 140.28), where the incor- 
porated pronominal dual element -s- of gatssu'x refers to the diminutive dual 
object is-ie'-nqxoq 'his fish-line', while the pronominal subject -ts- 'he' agrees 
with the object in diminutive consonantism; gatksu'klam {-Iks- always appears 
for -skt-) 'the two (women) came home with the (baby)' (Wishram Texts 2.12), 
the diminutive dual -s- referring to the grown-up women, not to the baby ; ga^f- 
ngatklagwa'x gas ktenaklwd 'st 'it-waves-freely-over-me-my-feathered-cloak' 
(Wishram Texts 142.5), where the first object -s- of the half-transitive verb 
refers to the diminutive dual noun s-tenakfwd'st '(small) feathered cloak'. Par- 
ticularly noteworthy in this connection is the idiomatic use of a diminutive dual 
object -.s- referring to an implied, unexpressed noun of diminutive significance; 
there need not even exist such a diminutive dual noun to which reference, if 
desired, could be explicitly made. A good example is: gaksi'lutk 'she cradled 
him', literally, 'she put the-two-small (objects) down to him', where 'the two 
small (objects)' refer to an implied word for 'cradle', though the word for 'cra- 
dle' in actual use is a masculine (i'-lkau). Similarly, verbs of jumping and somer- 
saulting have an incorporated diminutive dual object -s- referring to 'the two 

Seven: Penutian Languages 249 

small (feet)', though the actual word for 'feet' is plural (i't-pc). Examples are: 
gaksu'bEna 'she jumped'; gasixmi' Lgwa 'he turned a somersault' (Wishram 
Texts 82.18); and gats(s)altsgi'ma 'he laid her, belly up'. The [645] most trans- 
parent example of the use of an incorporated diminutive dual object to refer to 
an unexpressed but existing noun is afforded by certain verbs of looking, in 
which the -s- has reference to is-qxu's 'the two eyes'. A frequently occurring 
example of such a verb is gatssi'klElutk 'he looked at him', literally, 'he put the 
two small (eyes) down toward him', the -tc- and -gEl- appearing in their dimin- 
utive forms -ts- and -k!El- to agree with the object -s-; gasiximk!na'-uk"atsk 'he 
looked around' is another such verb. 

As a rule, it will have been observed, a verb form tends to be consistently 
diminutive or non-diminutive in its consonantism. It is at least possible, how- 
ever, to limit the application of the diminutive idea to some specific element of 
the action by "diminutivizing" only some corresponding element of the verb 
form. An example already published elsewhere will again do service here. The 
normal word for 'I struck him with it' is inigE'ltcim. If the verb stem -tcim 
appears, with diminutive consonantism, as -tsim, the implication is that the 
missile used is a small one. Hence we have four forms: inigE'ltcim 'I hit him with 
it'; inigE'tsim 'I hit him (a child perhaps) with it'; iniklE'ltcim 'I hit him with it 
(something small)', and iniklE'ltsim 'I hit him (a child) with it (something 
small)'. To be sure, such examples are very uncommon and the one just given is 
perhaps little more than a linguistic tour de force. Nevertheless, it shows very 
clearly how thoroughly alive is the feeling for the significance of consonantal 

Post-positions in Wishram 


Wishram, differing markedly in this respect from Lower Chinook, makes 
rather considerable use of a series of post-positive particles [651] defining mate- 
rial case relations (chiefly local and instrumental). As most such relations can 
be expressed by means of local and adverbial prefixes and suffixes in the verb, 
the denominating parts of speech being in apposition to incorporated pro- 
nominal elements, this use of postpositions must be considered as un- 
Chinookan in origin; the fact that some of the postpositive particles are 
phonetically identical with corresponding Sahaptin case suffixes proves the 
whole process to be borrowed from the neighboring Sahaptin linguistic stock. 
As a rule, such postpositive particles are used with denominating parts of 
speech (nouns, pronouns, adjectives), but some of them may also be suffixed to 
predicating words (verbs, particle verbs); in the latter case the predicate is to be 
considered as substantivized syntactically, though not morphologically, and is 
used subordinately to another predicate. Wishram thus utilizes its postposi- 
tions to some extent in the building up of subordinate clauses. Where a noun or 

250 ^' American Indian Languages 2 

other denominating part of speech has been already represented in the verb by 
an incorporated pronominal element, its relation to the verb and to other nouns 
in the sentence is necessarily already defined, so that no postposition is neces- 
sary; even here, however, it not infrequently happens that a postposition is 
pleonastically used (compare such English possibilities as "He entered into the 
house"). If a noun is modified by a preceding attributive word (demonstrative 
pronoun, numeral, noun, or adjective), the postposition is used with the modi- 
fying word. The postpositions, with examples illustrating their uses, are listed 
in the following paragraphs: 

1. -ba (-pa) 'in, at\ With this element should be compared Yakima -pa 'in'. 
Examples illustrating its use with nouns and pronouns occur with very 
great frequency, so that only a few need here be given. 
cikxa'-imdt ci'tfix yak"cxa'tpa 'half of it lies in his mouth' 4.3^ 
gaklakxa'-ima ilkfa'ckac akni'mba 'she put the child in the canoe' 2.11 
atgadi'mama da'uyaba wi'lx 'they will come in this land' 6.17 
gayu'yam ixtpd' wilx 'he arrived at one land' 6.28 
itcqxE'mEm axqxatcpa 'I am sick in my breast' 12.27 
gatci'upmt it.'o'xwatckpa 'he hid it in the bushes' 18.25 

gatu'ya yaxka'ba 'he went up to him' 20.10 (one can also say gaiiglu'ya 'he 
went to him' with local prefix -gEl-) 
gadiq'.Eltxi'uba icia'gitcba ya'k"cxatpa wamtfu'xiba 'they went out through 

him at his nostrils, at his mouth, and at his ears' 28.24 
gahi'xuni yaga'iipa wi'mal 'it floated in the great river' 48.7 
alxu'ya wa'tcktib' itga'qpuks 'let us go on the tops of the grass' 70.26 (literally, 
'the-grass-at its-tops') 
Observe that the first two examples illustrate its pleonastic use; the nouns 
yak'^cxa't and akni'm have been respectively anticipated in the verb by the 
pronominal elements -/- and -a-, while their local relation to the verb is 
defined by the prefix -k- 'on' following these elements, -ba is also used with 
demonstrative stems to form adverbs of place where: da'ba 'here'; kwo'ba 
'there'; ia'xiba 'yonder'. 
As subordinating element, -ba denotes 'where'; less frequently it indicates 
cause. It is suffixed either to the verb itself, or, similarly to the case of the 
modified noun, to an adverb or particle preceding the verb. Examples are: 
cta'xya I'nadix qla'tsEuba gatccgE'lgElx 'across yonder (were) the two where 

he had first seen them' 8.10 (literally, 'first-at he-saw-them') 
galiktcYptck gatcc^ElkE'lxpa 'he came to land where he had seen them' 8.5 
e'wi gali'xdx gayaxa'limaixpa 'he looked back to where he had thrown himself 

into the water' 8.6 
ma'sa gali'xox qfu'mba gagi'iix 'he was ashamed because she had disturbed 
him in his sleep' 58.26 (literally, 'disturb-in-sleep at she-made-him') 

2. References are to Wishram Texts. 

Seven: Penutian Languages 251 

2. -/amr (often with palatalized a as -ffl>?i/, -iemt) 'to, from'. This suffix is proba- 

bly Chinookan in origin; it may be plausibly analyzed as verb stem -/- 'go' 
+ verb suffix -am 'arriving' + tense suffix -t. This analysis would explain 
its two apparently contradictory meanings. It tends to draw the accent to 
itself. Examples are: 
ickte'lgwiptck wimalia'mt 'they collected (driftwood) from the river' 2.2 
nigElga'ba iciagitcia'mt 'it flew out of his nostrils' 80.29 (literally, 'out of him 

from his nostrils') 
gacx^klwa'x txtdqtia'mt 'the two returned to their house' 2.12 
gayukfwi'xa ilaxni'miemt 'he swam to the person's canoe' 18.23 
mxa'tcktcam wimatia'mt 'go to the river and wash yourself 22.18 (literally, 

'go-and-wash-yourself to-the-river') 
gatciu'k"{ itq^Udmt ittcqoa 'he took the water to the house' 28.8 
As subordinating element it may be translated as 'to where'. An example of its 

use after verbs is: 
asEmxElu'tka a'tpxiamd aga'tax 'you shall look towards the east' 188.21 (liter- 
ally, 'she-comes-out to-where the-sun') 

3. ba'ma 'for, belonging to'. This is evidently the Yakima suffix -pama 'for'. 

Examples of its use with denominating words are: 
na'ikabam' amtklni' dama itqagi'lak 'for my sake you two will go and get me 

the woman' 62.25 
ya'xtau laxka'bama IgiubttcEma 'that (fish) he obtains for himself 186.4 
gaqxd'gwigax itsH'ndnks wi'lxpama 'animals were taken belonging to the 

country' 16.13 
ctmo'kct gactu'ix ntca'ikabama 'two of our men (literally, us-for) went on' 

da'nbama qxe'dau mxu'lal 'what for do you speak thus?' 132.24 
tga'tqwdm luwa'n qa'xbabama 'he has come I know not where from' 128.17 

(literally, 'what-in belonging-to') 
k.'a'ya kwo'babama idE'lxam tcduxt 'he had not made people belonging to 

there' 44.23 
gi'gwalbam' itk'.i'tit 'underclothes' (literally, 'below-for clothes') 
Less commonly bama may precede. An example is: 
bam ' iiJ'wulx aktugwi" ilk'.a'lamat 'he carries rocks for (i.e. , in order to gain) 

strength' 186.17 (cf iLxe'wulx bama 188.2) 
When used at the beginning of a predication, bama gives it the meaning of a 

clause of purpose. Examples are: 
ba'ma ta'-itcka a'lEm' atcludi'na 'in order that he might kill them' 54.2 (liter- 
ally, 'for them will he-will-kill-them') 
bama capca'p qiuxu'nnit ika'ba 'for chopping up the ice' 188. 19 (literally, 'for 

chop-up it-is-always-made the-ice') 
When accented {bama'), it is used after predicates to mean 'ever since'. An 

example is: 

252 ^f American Indian Languages 2 

nk'.a'ckacbama' k/d'ya qxantcix itctcgE'ruEin 'ever since I was a child I have 
never been sick' 190.9 

4. (E)nEgi 'with, by means of, less frequently 'made out of. It seems to be the 

Yakima genitive case ending -ngi. Examples are: 
axk' E'riEgi amcgiu'xa Iqfo'p 'with it you will cut it off 12.4 
Lqfo'p gaigi'ux aqE'tiEkc E'riEgi 'they cut it off with the stone knife' 18.5 
gatkld'qi' aiakcE'n EtiEgi 'he counted them with his finger' 18.19 
itia'ma ngi gayu'ya 'he went by means of a round-pointed canoe' 38.21 
iga'hEnac E'riEgi gatdu'x 'he made them out of young oak' 4.13 


Less frequently ngi may precede. Examples are — 
xa'u xau galxu'x ng' ilkcE'n 'they combed themselves with the hand' 78.10 
ayakla'lamat ngi wa'nux 'his pipe (was) made out of a stomach' 94.9 

5. a'mEni 'made out of, less frequently 'with'. It is perhaps the Yakima -nmi. 

Examples are — 
sd'q" itk.'a'lamat a'mEni akitxax 'it is entirely out of stones' 82.13 
isklu'ly' amEni isga'klaps aqsu'xwa 'a hat is made out of coyote' 182.7 
alklwa'dit amEni aqiu'xwa 'it is made of tule' 182.9 
itq/u'tc' a'mEni tsE'tsEx gaqtu'x itkfa'munaq 'they split trees by means of 

antlers' 182.14 

6. -pt 'up to' is used to form adverbs out of demonstrative stems: dapt 'up to 

here'; kwopt 'up to there, then, enough'; ya'xpt 'up to yonder'. Probably 

etymologically identical with this element is -bEt, frequently added to 

verbs or other words in the predicate to form temporal clauses. Examples 

are — 
gatclE'mquit Iqa'wulqt gagiula' dabit 'he spit blood when she threw him down' 

galikta'tckpEt pla'la igi'xox 'when he had come up out of the water, he 

stopped' 22.18 
lE'p(b)Et atxu'xwa anigElgd 'ya 'when he dives, I shall take hold of it' 18.20 
nk'.a'ckacbEt 'when I was a boy' 188.8 
aga'lax alaxu'xwa yaxtadi'wi gali'xux gaixo'qbEt 'the weather will be as it was 

when they came together' 130.27 

When rhetorically lengthened to -bd't, this post-position has a general 

cumulative significance; with verbs it is best translated 'as many as'. 

Examples are — 
^wE'nEmabd'd Hgwo'mEx antk.'wa'lalaqwida 'I shall be absent as much as five 

days' 122.12 
kwo'pt natcdup^Enayabd't 'that many (ropes) as he had apportioned' 188.6 
qxa'ntcipt alkixa'tgway' atciulxamabd't 'he piles up as many as he tells him to' 



Seven: Penutian Languages 


diwi (emphatic dd'wi) 'like'. This element is very likely of demonstrative 
origin, and so does not perhaps belong here. It is freely used, however, as a 
post-position, and so may be included. Examples are — 
ickla'li diwi datdi'p 'striped like a basket' 166.2 
iya'tqx ilgwa'Ulx diwi 'his body (was) like a person's' 166.17 
naika dd'wi itcE'lgulit 'exactly like my appearance' 104.10 

Demonstrative Pronouns and Adverbs in Wishram 


Near 1st person 

Near 2nd person 

Near 3rd person 
Near 3rd person 

(formed from ya'xdau) 


















Near 1st person 

Near 2nd person 

Near 3d person 



f cda'xdau 


Near 3d person 

(formed from ya'xdau) cdakd'xdau 






Plural, persons 

, da'(u)a-itc 

Note. — It is somewhat doubtful whether ya'xdau should be so read or as 
ya'xdau. (x) in personal and demonstrative pronouns is deictic in value. 

-ka may be added to demonstratives in -/7c. 

Elements -tfa and -tHkc are perhaps "diminutive" forms of demonstrative 
pronominal stem da 'this' and personal plural dike. 

254 VI American Indian Languages 2 

Following is a list of the demonstrative adverbs of the Wishram dialect: 


up to 

towards, on . . . side 

Stem da(u) 

da'ha 'here' 


dabd't 'little ways further 


Stem kwo 

kwo'ba 'there' 


iyax da'ba ^^A^y 

Stem iaxi 

ia'xi 'away, off 


iaxd't 'further on' 

Stem di 

di'ka 'here' 

(pt also in 

di'gat (18.17) 

(dika data' 92. \\) 

qa'n tcipt 
'how long?') 


Stem gi 

g/>/ (18.17) 

{i'wat 'to you (place)' 
i'wa 'thus, there' (106.22) 

[iwa'tka (158.24) 
Note. — Compounded with gi are also da'ngi 'something'; qa'tgi 'somehow'; 
qxa'matgi 'somewhere' (96.11). 

Related to di'ka and di'gad is perhaps digu'tcix 'perhaps' (96.17); also di'wi 

In -xi we have, besides ya'xi, also (aga) du'xi 'oh, well'! (60.4). 
Note. — Ya'xa 'indeed ' (also in quct i'axa 'as it turned out'); au (perhaps = 
aw', a'wa, and related to Chinook _vflwfl) in da'n au ayamlu'da 'what, pray, shall I 
give you?' (154.6); yaxa'wa 'however'. 

Note also kwo'bixix 'right there, not very far'. 

-a'dix forms: a'ngadix 'long ago'; ixtka'dix (192.2); ina'tkadix (192.5). 

With stem dau: kwo'dau 'and'; da'ukwa 'just as before'; qxi'dau 'thus'. 

Independent Personal Pronoun 

These correspond to Wishram forms recorded by Sapir: 

na-ima 'I alone' 

Uiinuidikc, dci-iniadikc , dimadikc 'they alone' 

mci-ima "thou alone' 

Ixa-imadikc 'we (incl.) alone' 


Besides these, Doctor Sapir has recorded in Wishram the following: 

Shortest form: 

nafx) T ya(x) "he da'-ifc 'they' 

la'-itc 'they' (Wishram Texts 48.4) 
a'-itc 'they' 
nai't'.a 'I too' ya'xt'.a 'he too' la'-it'.ikc 'they too' 
Ixai't'.ikc 'we too' da'-it'.ikc 'they too' 
a'-it!ikc 'they too' 

3. References are to Wishram Texts. 

Seven: Penutian Languages 255 

He remarks that the demonstratives of the third person (ya'xia) seem mor- 
phologically parallel to first and second personal emphatic pronouns (na'ya); 
that the demonstrative element -/- is characteristic of the first and second per- 
sons, -X- of the third; as in 

na-i-ka T ya-x-ka 'he' 

na'-i-tfa 'I too' ya-x-t!a 'he too' 

na'-(i)-ya T ya'-x-ia 'he' 

These elements -/- and -x- are probably identical with Chinook -/"- and -x-, -x 
in x-i 'lu and x-ix-, xax. 

Modal Elements 


2. ni-. This prefix is confined to the dialects east of the Kathlamet. It takes the 
form nig- before vowels, like the preceding a-, prefix indicating a change 
from one state into another, translated 'to become' in intransitives and as an 
aorist in transitives. It occurs in transitive and intransitive verbs. It expresses 
a somewhat indefinite time past, and is used in speaking of events that hap- 
pened less than a year or so ago, yet more than a couple of days ago. 

ni-y-u'ya 'he went' {ni- 'past'; -y- 'he'; -uya 'to go') 

nig-u'ya 'she went' (the same before vocalic element; -a- 'she', being con- 
tracted with -u- into -u) 

ni-tc-i-gil-kErhe saw him' {ni- 'past'; -tc- 'he'; -/- 'him'; -g// verbal prefix; 
-knl 'to see') 

3. a-. [Prefix indicating the future in intransitive verbs.] In the dialects east of the 
Kathlamet it is used also with transitive verbs. 

a-tc-i-gE'l-kEl-a 'he will see her' {a- 'future'; -re- 'he'; -/- 'him'; -^eI- verbal 
prefix; -kEl 'to see'; -a 'future') 

4. ga-, before vowels gal-. This prefix is confined to the dialects cast of the 
Kathlamet. It expresses time long past, and is always used in the recital of 

ga-y-u'ya 'he went' (see analysis under 2) 
gal-u'ya 'she went' (see analysis under 2) 
ga-tc-igE'l-kEl 'he saw him' (see analysis under 2) 
n- may be used in place of this prefix. 

5. na-, before vowels nal-. This prefix is confined to the dialects east of the 
Kathlamet. It refers to recent time exclusive of today, more specifically to 
yesterday. Its use is analogous to that of the preceding. 

256 VI American Indian Languages 2 

By Edward Sapir 
Coyote and Itc!e'xyan 

673 Aga' kwo'pt^ gayu'ya* isklu'lya'^ wi'tlax." Na'2wit' gayu'yam;* 

Now tlu'ii he went Coyote again. Straightway he arrived 


galixE'ltcrna(i® isklu'lya gwa'nisini'*' qtiilatla'iiiElqt" idE'lxam" 

he heard Coyote always they findef.) are always the people 

Mwallowing them down 

674 itclE'xyan.'^ Qxa'damt" gayu'y"^ ikni'm** na'wit gatcigE'lga" 

Merman. Whither it went the canoe straightway he got hold of it 

itclE'xyan; gatciutatla'mElq^'* ka'nawi'® dan.^° "NaitI'" a'g'" 

Merman; he always swallowed it down every thing. "Me too now 

atcnuiatla'niElEqEma,"" isklu'lya galixtu'xwa-it." Aga kwo'pt 

he will swallow me down," Coyote he thought. Now then 

gayu'y' isklu'lja; gatcigE'lga yag'ail" ikla'munaq.^^ Aga kwo'pt 

he went Coyote; he got hold of it its the tree. Now then 

675 La'x" gali'xox.'* GatcigE'lga itclE'xyan, gaqiulat!a'raElEq." 

insight he made him- He got hold of merman, they (indef.) swallowed 

self. him him down. 

Na'wit ittcqo'ba^" gi'gwaP' isklu'lya galixi'maxitam^'' wi'lxba.*' 

straightway in the water below Coyote he arrived falling on the ground 

Aga kwo'pt gatcugi'kfil^* Igabla'd'* idE'lxam; Jgabla'd'^ akni'ni" 

Now then he saw them their multitude the people; their multitude the canoes 

axu'xf* kwo'ba^® gi'gwal iltcqo'ba. Aga kwo'pt gatcigE'lkEl*" 

they are there below in the water. Now then he saw it 

piled together 

isklu'lya itclE'xyan yago'mEnii^' qxwoL^'' iki'ax." Aga kwo'pt 

Coyote Merman his heart hanging it is. Now then 

676 gaqiu'lxam"* isklu'lya: "Ya'xdau"'' itclE'x3'an yago'mEnii.'' Aga 

they (indef.) told Coyote: "That Merman his heart." Now 


kwo'pt Lq!6'p"'= gatci'ux;"^ Lqlo'p"*^ gali'xox*^^ itclE'xyan yago'niEnii. 

then cut he made it; cut it made itself Merman his heart. 

Aga kwo'pt ka'nawi gatkxEni'yutck" sJi'q"''^ akni'm kwo'dau" 

Now then all they each floated up out entirely the canoes and 

of water 

idE'lxam kwo'dau isklu'lya. 

the people and Coyote. 

Aga kwo'pt gali'kim" isklu'lya: "Lga^^ pu" qii'ma^" ma'ima" 

Now then he said Coyote: " Perchance would how you alone 

itclE'xyan qxi'dau" amdu'xwa" idE'lxam? Da'uya^^ wi'gwa" aga 

Merman thus you will do to them the people? This . day now 

677 kwo'pt^® qxi'dau amdu'xwa idE'lxam. Na'ika" isklu'lya yamu'lxam.^* 

that far thus you will do to the people. I Coyote I have told you. 


Kwa'ic*" da'uyaba®" wi'lx atgadi'mama" idE'lxam. Kwo'pt 

Soon in this land they will arrive coming the people. Then 

alugwagi'ma,®^ 'Qxi'dau ^ex«^ gatci'ux^* isklu'lya itclE'xyan.' 

they will say, 'Thus exercising he did to him Coyote Merman.' 


Kwo'pt a'ga itclE'xyan pla'l'^^ amxu'xwa."" 

Then now Merman being quiet you will make yourself." 

Seven: Penutian Languages 257 

1 A connected English translation of this text will he found in Sapir's \\ ishram Texts, Puhlicatlons of 673 

the American Ethnological Society, ii, 4 1 , 43. The Indian text as here given has been very slightly normal- 
ized from its form as there published (pp. 40, 42). 

« Used partly with weak temporal force, partly as mere connective in narrative. It is frequently prac- 
tically untranslatable into English. 

3 kw6pt, THEN, AT THAT TIME, is regularly used with preceding aga to mark new step In narrative. It 
can be analyzed into demonstrative stem kwo- (or ktoa-) that (= Chinook go there) and local suffix 
-pt UP TO (so AND so) FAR. Neither of these elements occurs freely, kwo- is not used to form demon- 
strative pronouns, only occurring stereotyped in several adverbs; l)esides kwopt we have kwo'ba there 
(note 39), and kwo'dau and (note 46). -pt also hardly seems to occur except stereotyped in adverbs; 
cf. dapt AS FAR AS THIS (related to da'ba, this-in=here, as kwopt is to kwo'bc), and yaxpt, as far A3 
THAT YONDER, from ya'zi off yonder). See also note 56. 

* ga- (gal- before vowels) = tense prefix denoting remote past, regularly used In myth narrative. - y-^ 
3d per. masc. subj. intr., referring to isk.'u'lya. before consonants it would appear as -i-, while gal- would 
then appear as tense prefix (ga-y- = gal-i-: see notes 9, 28, 32, 47). -u- = directive prefix away from 
SPEAKER, -ya = verb stem to go. 

* i- = masc. noun prefix with which -y- in gayu'ya is in agreement, -sk.'u'lya = noun stem coyote, 
apparently not capable of analysis; perhaps loan-word from Klickitat spi'lya. Chinook has another stem, 

« Composed of wi'tfa again and deictic particle -i: cf. da'uya (note 54) and da'uyaz this. wi'Ha 
is most plausibly explained as stereotyped adverb from wi-, masc. noun prefix (originally independent 
masc. pronoun? See notes 19 and 33), and -t!a, emphatic particle added to pronouns, too, also (see note 
21). According to this analysis wi't!a(x) was originally formed from *wi as ya'xtfa(i) he too from ya-X' 
HE. Originally it must have meant that (masc.) too, but was later generalized in meaning. 

' Rhetorically lengthened form of nu'it immediately, right away. When thus lengthened to na'wit, 
it seems to imply direct, unswerving motion without interference of other action; it may then be rendered 


8 As in note 4, except that instead of verb stem -ya we have its shorter form -y. -i- (as in yu'it he goes; 

cf. also note 61). To this is suffixed verb suffix -am arrive while — ing, go (or come) to do . Several 

verb stems have two forms,— one in -a, and one without this -a (e. g., -pa and -p to go out; cf. gali'pa 


9 gal- = tense prefix ga- before vowels, -i- = 3d per. masc. obj. before reflexive element (reflexive verbs 
have, morphologically speaking, no subject). -xEl- = indirect reflexive composed of reflexive element -r- 
and local verb prefix -I- to, into, -tcmaq = verb stem to hear. galixE'ltcmaq means literally to him- 
self HEARD. TO HEAR SOME ONE is expressed by -x-tcmaq with prefixed transitive subject and object 
pronominal elements. 

10 Adverb not capable of analysis. 

" q- = indefinite transitive subj. -t- = 3d per. pi. obj. tr., referring to idE'lxam. -u- =directlve prefix 
(very many verbs have this "directive" -u- even when no definite idea of direction away from speaker 
seems to be implied). -tatfaniElq- is exampleof rarely occurring compound verbs. -Za^'a- is "diminutive" 
form of verb stem -lada- to throw down, away (in this case its meaning seems to correspond somewhat 
more closely to that of its Chinook cognate -Lata to pull back); -mElq- is best explained as verb stem 
-mEq. (or -mq-)TO vomit with infixed -I- of frequentative or continuative significance (that -Ms not really 
part of stem is shown by form itciulatfa'maq he swallowed him down); pull back -i- vomit may 
be construed as meaning vomit backward, draw to one's self and swallow, -t = tpnse suffix of 
present time. Observe peculiar sequence of tense, he heard . . . they sn- allow them down. Verbs 
that are dependent on other verbs, chiefly of saying or perception, are always present in tense, no matter 
what tense is logically implied; cf. below gatcigE'lkEl . . . iki'az (note 43) he Syi if it ... it /5. 

'2 id- = 3d per. pi. noun prefix, in concord with -t- in preceding verb, -ham (-E- is inorganic) = noun 
stem village (wi'liam village is formally masc. sing, of idE'lxam people); -liam is evidently 
related to -li (see note 33). 

" j- as in note 5. -tc.'Eiyan = noun stem merman, protector of fishermen (see Wishram Texts, p. 40, 
note 2; p. 42, note 2; p. 256, note 2); no etymology suggests itself. Syntactically itc.'E'iyan is subject 
Implied, but not grammatically referred to, by q- of preceding verb. This clause can hardly be considered 
as quite correct; properly speaking, itc.'E'xyan should go with tclulat.'a'mElqt. 

258 Vf American Indian Languages 2 


•< From interrogative stem qia- (or qa-), seen also in qa'xba what-in? = where? qa'xia of what 
KIND? and qa'ngi what-with? = how? -damt = local suffix toward found suffixed to several 
adverbs (cf. ca'ialadamt toward above, gigwa'ladamt toward below). This -damt is evidently 
related to local noun sufTix -iamt to, from. Qxa'damt here introduces indirect question, and may best 
be translated as no matter where. 

li = gayu'ya. Final vowels are regiilarly elided when following word begins with vowel. For analysis 
of form, see note 4. 

•« i- as in note 5. -knim = noun stem canoe. This stem can be only secondarily monosyllabic, for 
otherwise we should have * wiknim (see note 3.3); its Chinook cognate -kanlm shows original dissyllabic 
form. See also note 37. 

1' ga- = tense prefix as in note 4. -tc-.= 3d per. masc. tr. subj., referring to following itc.'E'xyan as sub- 
ject, -i- = 3d per. masc. tr. obj., referring to ikni'm of preceding clause as object. -gEl- = verb prefix of 
adverbial force, toward (with purpose, intent to reach); it here replaces directive -u- of most transitive 
verbs, -ga = verlj stem TO get hold of, seize; it is possibly to be identified with verb stem -ga stick to, 
its particular active significance being gained by use of transitive pronominal prefixes and verb prefix -gE 1-. 

18 ga-tc-i- as in note 17, -i- here referring to following dan. -u-lat.'a'-mElq as in note 11. 

" ka'nawi all, every is most probably compounded ofkana- all together (found in such niuneral 
fonns as ka'nactmokct all-the-two = both and, with unexplained -m-, in kanEmlu'nikc all three 
PEOPLE) and old 3d per. masc. demonstrative pronoun *wi (cf. note 6) now no longer preserved as such 
(except in such petrified words as wi't.'a and ka'nawi), but specialized, like its corresponding fem. wa-, 
as 3d per. noun prefix (see note 33). These old pronouns *wi and *wa are best explained as substantivized 
from pronominal elements-!- (masc.) and -a- (fem.) by means of demonstrative element w- (or «-); this 
latter element is probably identical witli -u- in demonstrative stem da'u- this (found also as da-; see 
note 54), and with ( hinook -o- in demonstratives near 3d per. {I'OLa, locta, iota), ka'nawi must origi- 
nally have meant something like all (of) that (masc), but, like wi't.'a, was later generahzed in signifi- 
cance, ka'nawi is here, as often, rhetorically lengthened to ka'nawi to emphasize its meaning of totality. 

""> Interrogative and indefinite pronoun referring to things, what, anything, something. Though not 
provided with any sign of gender, it is always construed as masculine, hence -i- in gatciulat.'a'mElq. Its 
correlative can (Kathlamet ban) referring to persons, who, anybody, somebody, is always neuter in 
gender; he swallowed everybody down would he gatciulat.'a'mElq ka'nawi can. 

21 Elided from na'it.'a (see note 15). Composed of 1st per. sing, pronominal stem nai- (seen also in na'- 
ika I) and emphatic suffix -t.'a too, also (see note 6). All independent pronouns in -fco can be changed 
to emphatic pronouns by merely replacing -ka by -t.'a (e. g., ya'ika he becomes ya'it.'a he too). 
Syntactically na'it.'a here anticipates -n- in following verb (see note 23) as 1st per. sing. obj. 

M = a'ga (see note 15). This particle is very frequently used before future verb forms in conversation. 

23 a- = tense prefix of future time, -tc- = 3d per. masc. tr. subj. -n- = 1st per. sing. tr. obj. -u-lat.'a'- 
mElEq- as in note 1 1 {-E- before -q- is inorganic). -Em- = connective before future suffix -a: verbs that are 
continuative or frequentative in form regularly use this connective -Em- before certain suffixes (such as 
future -a, cessative -tck, usitative -nil), -a = tense suflix of future time; in \V ishram verbs regularly form 
their future by prefixing a- or al- (before vowels) and suffixing -a. It is somewhat difficult to see why this 
form should be frequentative; one would rather except atcnulat.'a'mEgwa. 

2< gal-i- as in note 9. -x- = reflexive element; literal translation of verb would be (to) himself thought. 
•lux^w)- = verb stem to think, -a-it = verb suffix of rather uncertain significance here; it is found in all 
tenses of verb but present, where it is replaced by -an {iilu'iwan he thlnks). 

26 ya- = i-ya-. i- = masc. noun prefix, determining gender of noun stem -gail. -ya- = 3d per. masc. 
possessive pronominal prefix, referring to masculine noun ik.'a'munaq. -gail = abstract noun stem big- 
ness, yagailik.'a'munaq the tree's bigness may, like all other possessive constructions, be construed 
either attributively (the big tree) or predicatively (the tree is big). Its attributive character is here 
determined by presence of true verb (gatcigE'lga) as predicate. 

•6 J- as in note 5. -k.'a'munaq = noun stem tree, stick, wood. This word is difficult of etymologic 
analysis, yet can be no simple stem; -k.'a- is undoubtedly to be regarded as noun prefix (cf. ik.'a'lamat 
ROCK, perhaps from verb stem -la to move), -k.'a- is most plausibly considered as "diminutive" form 
of verb stem -ga- to fly '- r in air (as first elenient in comi)ound verbs); cf. itciuk!wa'la he whetted it 
with itci'ula HE filed it, and miugwala' da-ulx i threw it up on top (of something) with iniula'da-uLf 
I threw it up. 

2' Particle verb. Though verbal in force, it is purely adverbial morphologically, having no grammatical 
form of its own. In regard to tense and person it is defined by following verb, which serves as its form- 
giving auxiliary. 

» gal-i- as in note 9. -i- = reflexive clement. -6- (modified from -«- because of preceding and following 
velar consonants) = directive prefix; ordinarily reflexive -i- replaces directive -u-, but there are several 

Seven: Penutian Languages 259 


verbs that retain it even when reflexive in form, -i = verb stem to do, make, -x-u-x to do to one's 
SELF, MAKE ONE'S SELF, is regularly used to mean become. For other forms of verb stem -i see notes 
43, 53, 64, and 66. 

^ ga- as in note 4. -q- = indefinite tr. subj. -i- = 3d per. masc. tr. obj. -u-lat/a'-mElEq as in note U. 
FOTms with indefinite -q- subject are very commonly used in Wishram in lieu of passives. 

'• il- = 3d per. neut. noun prefi.x. -t- = inorganic consonant, serving as glide between I and c. <q6- 
(= <qa-; a is velarized to 6 by preceding q) = noun stem water ; its shorter form <q- is seen in kta'cq 
THE WATER OF THE TWO (Wishram Texts 190.14). -ba = local noun suffix in, at (see also notes 33, 
39, and 60). 

31 Adverb; -al is probably not part of stem, for it is found also in correlative ca'i-al above. 

32 gal- as in note 9. -i- = 3d per. masc. intr. subj., referring to preceding i-sk.'u'lya. -lima- = verb stem 
to put down, put on ground, lay down (as tr.); lie down (as intr.); probably composed of -x- on 
GROUND(?) and -ima- put (cf. ga-ya-i-a'l-ima-lx HE put himself into the water [Wishram Texts 2.5)); 
whenever indirect object with -k- on is introduced, -x-ima- becomes -la-ima- (e. g., ga-k-l-a-k-ia'-ima 
she laid it down on it [Wishram Texts 2.11]). -lit = quasi-passive suffix; -x-ima-xit- = be laid down, 

LAY one's .SELF DOWN, FALL DOWN TO GROUND, -am = verb SUffix ARRIVE ING (cf. note 8). 

33 wi- = 3d per. masc. noun prefix; masc. noun stems that are non-syllabic or monosyllabic require wi- 
(cf. note 55); those that have more than one syllable have i- (see notes 5, 13, 16, 26); for probable origin of 
wi- see note 19. In Chinook wi- has entirely given way to i-, except as archaism in some place-names 
and in songs, -li- = noun stem land; seen also in wi'lxam village, idE'lxam people (see note 12); 
probably also in wa'lxi FisHittG station and icE'lxlx staging for fishing, -ba as in note 30. 

3< ga- as in note 4. -tc- = 3d per. masc. tr. subj. -u- = 3d per. pi. obj., referring to following idE'lxam 
(before verb prefix -gEl-M per. plural obj. -Ms replaced by-w-, -gEl-ih^n becoming -^(w)'-; in other words, 
-t- before gEl- is treated analogously to when it comes before -gel-), -gi- = plural form of -gEl-{see note, 
40)^ouT FROM enclosed SPACE (cf. gi-l-a-g E' l-ba it flowed out of her [Wishram Texts 94.4]); analo- 
gously to -</ir/-( see note 17) directive -u- is here replaced hy -gEl-. -kEl= verb stem to know (cf. l-k-d-u'- 
kulHE knows them [Wishram Texts 176.10)); -gEl-ksl = to know from out one's (eyes), hence to 

SEE, GET sight OF. 

34 ^ = 3d per. neut. noun prefix, defining gender of abstract noun stem -bla4. -ga-= 3d per. pi. pos- 
sessive pronominal prefix, referring to idE'lxam. -blad = noun stem multitude, great number. Igabla'd 
idE'lxam is construed like ya'gnil ikfa'munaq (see note 25). 

36 As in note 3.'>, except that -ga- = .3d per. fem. possessive pron. prefix (merely homonymous with -ga- 
of note 35), referring to akni'm. 

iT a- = 3d per. fem. noun prefix; though many fem. dissyllabic stems have wa- (e. g., wala'la pond), 
it is here replaced by analogy of iknt'm (see note 16), as in related nouns i- and a-, wi- and wa- generally 
pair off respectively, -knim as in note 16. logically akni'm canoes is plural, morphologically it is 
fem., being so referred to in aiu'xt, (note 38); another example of fem. as plural is wa'mwa maggots, masc. 
wi'mwa maggot. 

38 a- = 3d per. fem. intr. subj., referring to akni'm. -i- = verb prefix on ground, on bottom (?) -u- = 
directive prefix, -xt = verb stem to lie, sit, be placed, corresponding in use to Chinook -c. This verb 
stem allows of no formal modification by means of tense affixes. 

3" Composed of denionstrativestem kwo- (see note 3) and local suffix-6a Csee note 33): that-in = there. 

♦0 As in note 34, except that incorporated obj. is -i- = 3d per. masc, referring to yago'mEnil, and that 
-gEl- is unmodified. 

<i ya- = t-j/a-asinnote25, i- defining heart as masc. in gender, while -ya- refers to itc.'E'iyan. -g6mEnil 
heart seems to be verbal in form, -Enil being usitative suffix; yagO'mEnil may also be used predicatively 
to mean he is alive. 

« Particle verb, for which iki'ax serves as auxiliary. 

<3i-= 3d per. masc. intr. subj., referring to yagd'mEnil. -kiax to be is another tenseless verb (cf. note 38). 
It is best, though somewhat doubtfully, explained as composed of verb prefijc -ki-, which shows lack of 
object of ordinarily trans, verb, and verb stem -i to do (cf. Kng. he oo£S well, i. e., gol^ along well)- 
-a- would then have to be explained as inorganic glide vowel (cf. Chinook i-ke'-i he is and Wishram 
i-ki'-i-ai he is, has become). For syntactic construction, as subordinated to gatcigs'lkEl, see note 11. 

«B ga- as in note 4. -q- = indef. tr. subj. (cf. note 29). -i- = 3d per. masc. tr. obj., referring to Lik.'u'lya. 
•u- = directive prefix, -liam = verb stem to say to with personal object. This verb form is logically 

260 VI American Indian Languages 2 


43b Donionstrative pronoun, showing location near 2d person, composed of simple form of independent 
3d personal pronoun + demonstrative element -z- (of. also ordinary forms of independent 3d personal 
pronoun ya'x-ka and similarly for other genders) + demonstrative stem -dau (= -da + -u), for vifhich see 
note 54. Syntactically ya'zdau, here used suijstantively, agrees in gender with yago'mEnil, to which it 
refers. There is no expressed predicate in this sentence, yago'menil (it is) his heart being so used. 

«<: Particle verb, to which following verbs gatci'ux and gali'iox, both from verb stem .-i to do, serve 
as auxiliaries. Lqfdp doubtless has onomatopoetic force. 

«d See note 64. 

<*» As in note 28. cut it-m.\de-itself = it became cut. 

<< ga- as in note 4. -t- = 3d per. pi. intr. subj., referring to akni'm, idE'lxam, and isk.'u'lya as combined 
plural subject, -k- = regular replacement of directive -u- whenever intr. subj. -t- would theoretically be 
expected to stand before it. -xEni- (or -luni-) = verb stem to float, drift, -j/m- = distributive suffix 
EACH separately (gatkxEni'tck would mean they floated up in one body), -tck = local verb suffix 
UP to surface, up from position of rest (cf. also gal-i-x-lE'-tck he moved himself up from sitting 
POSITION, he arose [Wishram Texts 4.6]; gal-i'-kta-tck he rose (sticking his head) out of water 
[op. cit., 10.5]); combined with -6a out of interior, -tck appears as -ptck from water out to land 
{gatkxEni'yuptck they each floated on to land; for change of -ba to-pcf. galags'lba with lagE'lpt 
[Wishram Texts 94.7]). This -tck should be distinguished from -tck of cossative significance, whose 
function it is to deprive verbs that are continuative or frequentative in form of their continuative 
force (e. g., yuwi'lal he is dancing, gayuwi'lalEmtck he was dancing (but is no longer doing so). 

« Adverbial in force. Logically sa'g"- (rhetorically lengthened to sd'qu to emphasize idea of totality) 
often seems to be used attributively with nouns (translated as all), but grammatically it is best con- 
sidered as adverbial, even when there is no expressed predicate. 

« Composed of demonstrative stems kw6- (see note 3) and dau- (see note 54). Its original significance 
was evidently that (which precedes) and this (which follows). 

<7 gal-i- as in note 32. -kim = verb stem -to say (without personal object; cf. note 58). 

« Adverb of modal significance, serving to give doubtful coloring to verb. 

« Adverb of potential and conditional significance; in formal conditions introduced by cma'nii if, it 
often has contrary-to-fact implication. This use of modal particles in lieu of verb modes is characteristic 
of Chinookan. 

60 Evidently contains interrogative stem qa- what, seen also in qxa'damt (note 14). -ma can not be 
explained. This word has been found only in such passages as here, and is very likely felt to be archaic. 
Iga pu qa'ma occurs as stereotyped myth-phrase in transfonner incidents (cf. Wishram Texts 6.13, 
38.6, for similar passages). 

51 Forms in -aima alone may be formed from simplest forms of personal pronoims (subject intr. 
incorporated); e. g., na'ima i alone, ma'ima you alone, ya'ima he alone. It is doubtful, however, 
whether these forms should be considered as intransitive verbs from verb stem -aima. Since personal 
plurals in -dike (e. g., la'imadikc they alone) occur, it seems preferable to consider them as formed by 
suffixed -ma alone? (cf. qa'ma note 50) from independent pronoun stems in -ai- (as in na'ika, note 57, 
and na'it.'a, note 21); this -ai- is in these forms found also in 3d persons (e. g., la'ima it alone, as con- 
trasted with la'xka and la'xt.'a). Chinook nd'mka i alone, analyzed by Boas as intr. subj. pronoun + 
verb stem -nmka, Is probably best explained as simple independent pronoun in -a- {na, ma, and corre- 
sponding fonns for other persons occur not rarely in Wishram) + -m{a) + -ka just, only (cf. lu'nka just 

'"- Adverb composed of.relative particle qxi- (cf. qzi as relative pronoim in Wishram Texts, 188.1) and 
demonstrative stem dau- this (cf. note 54). qii'dau thus means literally as, like this. 

"a-= tense prefix of future time, -m- = 2d per. sing. tr. subj. -d- = 3d per. pi. tr. ohj., referring 
to idE'lxam. -u- = directive prefix, -x- = verb stem to do (to), -w- = inorganic consonant induced 
by -u- preceding k- sound, -a = future suffix. 

M Demonstrative pronoun, showing location near 1st person, composed of demonstrative stem daw 
(= da-, as in da'6a here + -u-, see note 19) and simple form of 3d per. independent pronoun in -a (masc. 
ya, fem. a, neut. la, du. cda, pi. da). Forms without -u- (e. g., da'ya) occur, though much less frequently; 
deictic -i may be added without material change in meaning (e. g., da'uyax or da'yax). -dau also occurs 
as second element in demonstrative pronouns showing location near 2d person (e.g., ya'idau that masc., 
note 43b). da'uya is here masculine because in agreement with masc. noun wj'j/u'a. Chinook seems to 

Seven: Penutian Languages 261 

preserve da- only in isolated adverbs like <a'fc£ then (= da'ka just this or that [cf. Wishram da'uka 

JUST so]). 

is wi- = masc. noun prefix, with w- because noun stem is monosyllabic, -gwa = noun stem day. 
da'uya wi'gwa this day is regularly used as stereotyped phrase for to-day; dau' aga'lax this sun is 
also so used. 

56 Analysis given in note 3. Here kwo'pt, with well-marked stress accent, preserves its literal meaning 
of that far, thus much, aga kwo'pt being regularly used, outside of narrative, to mean enough. Chi- 
nook kapc't enough is doubtless related, but fco- can not be directly equated with kw6-, which corre- 
sponds rather to Chinook go (see note 3). 

5' Ordinary form of independent personal pronoun, composed of stems in -ai- (for 1st and 2d persons) or 
-a-x- (for 3d persons) and suffi.xed particle -ka just, only, found also suffixed to numerals, na'ika is here 
grammatically unnecessary, but is used to emphasize subject of following verb form. 

58 = iyamu'lxam. i- = tense prefix of immediate past time, -yam- = combination of 1st per. sing. subj. 
and 2d per. sing. obj. -u- = directive prefix, -ham = verb stem to say to with expressed personal object. 

59 Temporal adverb referring to action just past or about to occur, either just now, recently, or soon. 
Seems to be Klickitat loan-word. 

6" da'tij/a as in note 54; masc. because in agreement with masc. noun will, -ba = local noun suffix in 
regularly suffixed to demonstrative pronoim preceding noun instead of to noun itself. 

81 a- as in note 53. -t- = 3d per. pi. Intr. subj., referring to idE'lxam. -ga- = element regularly Intro- 
duced after 3d per. pi. Intr. -t- before -d-i- to come and, before verb stems beginning with k- sounds, 
after 3d per. pi. intr. -u- (cf. note 62). -d-i- to come consists of -d- = directive prefix hither, toward 
SPEAKER, correlative to directive -«-, and -i- = verb stem to go. -mam- = form of -am- (see notes 8 and 
32) used after vowels, -a as In note 53. 

62 al- = tense prefix of future time employed before vowels (al- and a- used analogously to gal- and ga-). 
-u- = 3d per. pi. Intr. subj.. used, Instead of -t-, before verb stems beginning with k- sounds (as here 
-gim-). -gwa- = -ga- as In note 61, -w- being Inorganic, due to Infiuence of -u- preceding k- sound (cf. 
note 53). -gim- = verb stem to say; -kim (as in note 47) is used when accent immediately precedes, -gim- 
when suffix (here -a) is added and accent Is pushed forward, -a as In note 53. In Chinook -ugwa- appears 
as -ogo- (gwa regularly becomes go); alugwagi'ma is paralleled in Chinook by ogogoe'ma. 

63 Particle verb to use supernatural power, transform, to which following gatci'ux serves as aux- 
iliary. It is one of those very few Wishram words in which glottal catch Is found (other words are -tci^ 
OR, H'c^ic BLUEJAY, daWa'x perhaps). 

6* ga- as In note 4. -<c-= subj., referring to jsfc/u'^ya. -i- = obj., referring 
to UdE'xyan. Observe that subject noun regularly precedes object noun, their order being thus analogous 
to that of Incorporated pronouns with which they stand In apposition, -u- = directive prefix, -x = verb 
stem to do (TO). 

* = p.'a'la. Particle verb, with which following amxu'xwa is used as auxiliary. pfa'V amiu'xwa QUiET 
You-wiLL-BECOME (I. e., you will stop, desist). 

6«a- as In note 53. -m- = 2d per. sing. obj. with following reflexive element (see -i- in notes 9 and 28). 
-z- as in note 28. -u-x-w-a as in note 53. 

262 ^f American Indian Languages 2 

Editorial Note 

Excerpts originally published in Franz Boas, "Chinook," in Handbook of 
American Indian Languages, Part 1 (Bureau of American Ethnology, Bulletin 
40), 578-579, 625-627, 638-645, 650-654, 673-677. Washington, D.C.: Smith- 
sonian Institution (1911). 

Sapir s extensive grammatical files on Wishram were later used by Walter 
Dyk, a student of Sapir s at Yale, in preparing a Wishram grammar for his 1933 
doctoral dissertation. This grammar, which remains unpublished, also drew on 
new material collected both in the field and at Yale, where Philip Kahclamat, 
a young Wishram speaker, was in residence for a few months in 1932. Sapir's 
and Dyk's manuscript materials on Wishram are now in the Boas Collection, 
Library of the American Philosophical Society. 

Sapir s description of Wishram diminutive-augmentative consonant "grada- 
tion," here and in the previous paper, is the first detailed treatment of this 
common feature of western North American languages. Sapir's exhaustive 
description of the elaborate phonetic symbolism of Nootka (1915a) is also 
reprinted in this volume. For a thorough survey of the literature on American 
Indian consonant symbolism see Nichols (1971). See also Haas (1970). 


In spite of its somewhat top-heavy intro- 
duction Dr. Leo J. Frachtenberg's " Com- 
parative Studies in Takehnan, Kalapuyan and 
Chinookan Lexicography " (International 
Journal of American Linguistics, vol. I, n° 2, 
pp. 175-1S2) is a decidedly valuable and wel- 
come contribution to our gradually progressing 
knowledge of the relations between the seve- 
ral linguistic " stocks " of the Pacific Coast. He 
does not go so far as to deduce genetic rela- 
tionship between Takelma, Kalapuya, and 
Chinook on the basis of the lexical evidence 
presented in his paper, but he does__" make 
bold to predict that additional data will be 
produced in the near future " tending to 
confirm such a conclusion. Elsewhere in the 
paper he states : " I have collected a mass of 
material establishing a probable conmion ori- 
gin for the Kusan, Siuslawan, Yakonan, and 
(perhaps) Kalapuyan languages which will be 
presented in the near future. " Though he 
chides Dixon and Kroeber for their hasty 
annouqcement of the Hokan and Penutian 
groupings, he remarks, somewhat unexpec- 
tedly : " The absence of conclusive evidence 
concerning Penutian and Hokan is the more 
unfortunate, as there exist strong reasons to 
believe that the Takelman, Kusan, Siuslawan, 
Yakonan, Kalapuyan, and (perhaps) Chinookan 
languages spoken in Oregon may be proven to 
be Penutian sister-tongues. 

All this is very interesting to me, as it 
chimes with conclusions or hypotheses I had 
arrived at independently. On the appearance 

of Frachtenberg's Coos grammar ' it soon became 
clear to me that the morphological and lexical 
resemblances between Takelma and Coos were 
too numerous and fundamental to he explain- 
ed away by accident or plausibly accounted 
for by borrowing. This in spite of the very 
great differences of phonetics and structure 
that separate the two languages. The appea- 
rance of Frachtenberg's" Siuslaw material ^ has 
only tended to confirm this impression, further, 
to make it perfectly obvious that Coos and 
Siuslaw, as Frachtenberg announces, are 
divergent representatives of a single linguistic 
stock. Meanwhile comparison of Takelma, Coos, 
and Siuslaw with Dixon and Kroeber's Penutian 
group of California (Costanoan, Miwok, 
Yokuts, Wintun, and Maidu) disclosed an 
astonishing number of both lexical and mor- 
phological correspondences, correspondences 
which were first dimly brought to my con- 
sciousness years ago by certain morphological 
resemblances between Takelma and Yokuts, 
later and more vividly by the decidedly Penu- 
tian " feel " of Coos grammar. In spite of our 
slight knowledge of most of the Californian 
languages involved, I succeeded, in getting to- 
gether what I believe to be a quite respectable 
mass of evidence tending to unite the southern 

1. Coos, an IlUisirative Skctcli (Handbook of Ame- 
rican Indian Languages, BBAE 40, part 2, 1914, 

pp. 297-429)- 

2. Lower Umpqua Texts (CU, 4, 1914) , Siuslawan 
(Lower Umpqua), an Illustrative Sketch (Handbook of 
American Indian Languages, BBAE 40, part 2, 191 7, 
pp. 431-629). 


VI American Indian Languages 2 


languages with those of Oregon into a large 
and highly differentiated " stock ". 

Unfamiliarity with Alsea (Yakonan) and 
Kalapuya made it impossible for me to follow 
the chain of evidence geographically. Neverthe- 
less, Chinookan points of contact soon mani- 
fested themselves too persistently to be brushed 
aside. After hesitating for a long while to take 
up seriously the possibility of affiliating 
Chinook, one of the most isolated and mor- 
phologically specialized languages in America, 
with the Penutian languages of Oregon, I now 
find myself forced by the evidence to admit 
such an affiliation as not only possible bat 
decidedly probable. In view of the clear points 
of lexical contact and of the phonetic shifts 
that Frachtenberg has established, even if only 
in a preliminary way, between Takelma and 
Kalapuya, his further hypothesis of a funda- 
mental connection between Kalapuya and 
Chinook was, forme, to be looked for a priori. 
I believe it only fair to add that the manuscript 
evidence that I possess of the relation between 
Chinook and various languages to the south is 
much stronger than the comparatively scanty 
lexical data presented by Frachtenberg '. 

1. Even of the nineteen cognates given by Frachten- 
berg no less than nine or ten, it seems to me, are almost 
certainly borrowings. This is particularly obvious in the 
case of Kalapuya U skan cup and u'lxayu seal. Both of 
these nouns have taken over bodily the Lower Chinook 
feminine prefix u- of u'-cgan cup and u'-lxaiu seal. 
Moreover, the Kalapuya, as an inland people, could 
hardly be expected to possess a native term for the seal. 
As for u'-cgan, that is merely a secondary feminine for- 
mation from Chinook i'-cgan (masc.) CEDAR. This 
alone would prove Kalapuya u'skan cup to be a borrow- 
ing from Chinook. 

The fifty-three Takelma-Kalapuya correspondences are 
of a very different sort and bear all the ear-marks ot 
genuineness. I take this opportunity of correcting a few 
errors. For Takelma Ja*'- to finish (p. i8o)read t\ap'g-; 
for t'pdW SNAIL (pp. i8o, i8i) Ttzd t'billt' ; {or dip' 
CAMASS (p. i8o) read dip' ; for t]ewex flea (p. i8i)read 
tUwix. Frachtenberg's statement that « Takelma ts 
remains unchanged >i in Kalapuya (p. i8o) is incorrect, as 

However, further evidence serving to hnk 
Chinook both with Kalapuya and with Coos- 
Siuslaw-Alsea will no doubt have been discov- 
ered by Frachtenberg by the time this paper is 

The greatest surprise was still awaiting me. 
Tsimshian occupies a peculiarly isolated posi- 
tion. In its morphological aspects it offers 
hardly any specific points of resemblance to 
the neighboring Nadene languages (Haida, 
Tlingit, Athabaskan). With the Wakashan- 
Salish-Chimakuan group to the south it shares, 
e. g., initial reduplication to express distribu- 
tive and plural ideas and a series of classifying 
suffixes appended to numeral stems. In its 
deeper morphological features, however, 
Tsimshian stands quite apart from these lan- 
guages. Lexically too, aside from a number of 
more or less patent loan-words, there seem to 
be no stems that Tsimshian shares with any of 
the neighboring languages. A tentative com- 
parison with the Penutian (or, if one prefers, 
hypothetically Penutian) languages of Oregon 
revealed a considerable number of correspon- 
dences both in the lexical material and in some 
of the more intimate and fundamental features 
of the morphology. Should it be possible to 
demonstrate (and I am fairly sanguine that it 
can be demonstrated) thai Tsimshian is a de- 
tached northern offshoot of Penutian, we would 
be compelled to face a most interesting fact in 
linguistic differentiation and in the distribu- 
tion of American tribes. 

The data for the various assertions I have 
made in this paper I expect to present /;/ 
extenso in the future. The fundamental type of 
Penutian language seems to be a predomi- 
nantly inflective one, as exemplified say in 
Miwok, Yokuts and, in more specialized form, 
in Takelma. The tremendous morphological 

Takelma possesses no ts ; al-ts-M red (pp. i8o, i8i) 
should be corrected to al-ts-\il. 

Seven: Penutian Languages 



divergences exhibited by the various languages 
here grouped together are due, it v»'ould seem, 
to two main types of change : the gradual 
breaking down of inflective forms and the 
consequent growth of more analytic forms (e. 
g., in Costanoan and, in a less extreme way, in 
Coos) ; and the evolution of special synthetic 
developments, as in Chinook, often, no doubt, 
under the influence of alien languages, as in 
Maidu and Tsimshian. It is, of course, entirely 
premature to group the Penutian languages 
genetically, but the following provisional 
scheme may be of some s'ight value as an aid 
to convenient visualization. 

A. Californian group 

[ Miwok 
\ Costanoan 

2. Yokuts 

3. Wintun 

4. Maidu 

B. Oregon group 

1 . Takelma 

2. Oregon Coast group 


3. Kalapuya 

C. Chi nook an 

D. Tsimshian 

It may well be that Frachtenberg's as yet 
unpublished data on Kalapuya and Yakonan 
will force a readjustment of the languages 
under B ; as for group A, possibly Wintun and 
Maidu form a closer unit, analogous to that 
undoubtedly comprised by Miwok and Costa- 
noan '. 

I. See A. L. Kroeber, The Chumash and Costanoan 
Languages, UCP 9: 237-271, particularly 259-263(1910); 
also The Languages of the Coast of California north of 
San Francisco (/oc. cit., 191 1, pp. 273-455), pp. 306-508. 

One of the most striking of the more inti- 
mate structural peculiarities of Takelma is the 
presence of a large number of stems of type 
cv, cv(c2)-, i.e. of stems with repeated vowel. 
In the verb, this method of stem formation has, 
to a large extent, taken over a grammatical 
significance, in some cases characterizing the 
aorist stem, in others, less numerous, the verb- 
stem proper. It is, however, frequently illus- 
trated in the noun^ adjective, and adverb as 
well ^. Examples are : 

yana ' acorn 

viahai large 

waga- TO BRING (non-aorist) 

baxam- to come (aorist) 

herein- KIDNEY 


yeuue^ perhaps 

sehe- TO roast (non-aorist) 

lehei- to drift dead to land (aorist) 

yiwin speech 

-xiivi- TO ROT (non-aorist) 

ginig- TO GO SOMEWHERE (aotist) 

nwt'op' SEED-BEATER 

odo- TO HUNT FOR (nou-aorist) 
Jop'od- TO RAIN (aorist) 

Silli'V'k' CRICKET 

ilyftfs- TO LAUGH (aorist) 

Turning to Coos, we find that this same type 
of stem formation, though apparently not pro- 
vided with formal significance, is very common. 
It is so very characteristic of the noun, indeed, 
that Frachtenberg's failure to mention it in his 
Coos grammar must be set down as a curious 
oversight. By actual count I find in his Coos 

2. For details see The Takelma Language of South- 
western Oregon (BBAE 40, part 2, 191 2, pp. 1-296), 
§§40, 86, 109. 

3. lam omitting all stress and pitch accents in this 
paper, also normalizing all orthographies to correspond 
to the' recommendations of the Phonetic Committee of 
the American Anthropological Association. 


VI American Indian Languages 2 


vocabulary ' no less than 140 examples of such 
nouns (and adjectives), 35 of verb stems, and 
9 of adverbs. These figures show conclusively 
that we are dealing with a real formative 
feature of the language, not with an arbitra- 
rily selected congeries of words. A few of the 
examples are : 

yabas maggots 
yalaq gray (of hair) 
fiianat crowd 

Cola't'iS, CRACK 

(jaiva CHEEK 

maha- to scare 

ak'auak- to stick out 

kayal- to decay 

qala"- to hurt 

//rt/a'^'- to bark 

hakwai as if 

ehe GONE 

luebel stomach 

tsehes alive 

^el'eths corner 

tlpene wings 

en'ek- to stick out 

eqe to die (plur.) 

xeyetl to twist to one side 

lek shouts 

luekx- TO stoop 

kweek to live 

ehenlc far away 

Cccd MAT 

kzvims feather 
W'.t'.n blood 
f'./"u: razor-clam 
hkiuKt feathers 


h'.tl- TO CARRY 


^l//-^- TO OVERTAKE 

IV'.xkm- TO STEAL 

I. See pp. 191-215 of Coos Texts, CU i, 1913. 

ydki' LONG AGO 

nwytis anus 
tso'no' BOTH ways 

It does not seem plausible that these and 
numerous analogous examples are merely 
secondary phonetic developments due to pro- 
gressive or regressive vocalic harmony ^. To at 
least some extent Coos possesses forms with 
repeated vowel that are of grammatical sig- 
nificance, notably certain irregular plurals of 
nouns and adjectives, e.g. : 

tSi'y-UX^' SMALL 

plur. tsi'yt'-ne 

The very irregularity of such forms points to 
their archaic character. They are obviously 
analogous to such Takelma adjectival plurals 
as : 

-t'gem BLACK 
-t'gcy-ap-x ROUND 
ba'Hs i,ONG 

plur. -fgeme-t'it' 

Analogous forms seem to be found in Lower 
Umpqua (Siuslaw), though apparently not so 
characteristically. I do not feel as clear about 
them as yet as about the corresponding Coos 
formations. Examples are : 

yafqa-a"' hole 
hayci'tit WIDOW 
halca-t LONG 
pa'iiqa medicine-man 
taha-mk quiver 
ca'ya penis 
t'saxan stomach 

k\iyak EAGLE 

qa'xal fish-net 

XlVil'ka HEAD 

iazvat- to gamble 
2. Sec Frachtenberg's Coos, BBAE 40, 2 : § 7. 

Seven: Penutian Languages 



amha- to be willing 
atas ONLY 

prs<.p FISH-HAWK 
misi'-a" ELDER SISTER 
t\hmS CHILD 
tSimd'z' MUSKRAT 
tsi'kwi LEGGINGS 

hmilta MATCHES 

pmd-t SHARP 

aiorhin- to lose 
zui'tsm- TO CAMP 

tlxmu'hv BOW-STRING 

Many more forms of this type, though not as 
many as in Coos, could be cited, but, on the 
whole, one gets the feeling that it had ceased 
to be a live and productive method of stem 
formation. An apparent tendency to loss of 
vowels and consequent formation of consonan- 
tal clusters would seem to have counteracted 
the persistence of the type. Comparison with 
Coos seems to confirm this, e.g. : 

L. Umpqua 


nia''q"li crow 

tipa'^'n, ipz't'v wings 



In both Coos and Siuslaw, it should be noted, 
forms with repeated vowels are sometimes 
related, as in Takelma, to monovocalic forms, : 

Coos manat crowd 

t'ipenc wings, feathers 

xalaw-is heat 
L.U. valqa-a" ■ hole 

piurcja medicine-man 

nianta-^- to keep compan\- 
tlpe wings 
xahu-is hot 
yafq-' i-lq- to dig 
p:iiq- to dance a mtdicinc 

Whether the type of stem with repeated 
vowel is found also in Alsea (Yakonan) must 

wait for an answer until Frachtenberg's mate- 
rial is available. This applies also to Kalapuya. 
The Kalapuya forms that he gives in his com- 
parative study number several of the type, 
e.g. : 

thanaq fly 

qalavi silverside salmon 

qauwan Chinook salmon 

palam drunk (Chinook Jargon) 

qoloq swan (Chinook) 

diigulhu' owl 

niu'Iukwa cow 

but, suggestive as such examples are, not much 
can be done with them when torn apart from 
their grammatical and statistical context. In 
Chinook and Tsimshian the type cannot be 
demonstrated to exist. It is evident from inter- 
nal evidence and from its numerous and often 
complex clusters of consonants that vowels 
have dropped out very frequently in Chinook, 
so that an originally present type of stem for- 
mation (c)vc,v(c2)- would be expected, under 
the destructive influence of phonetic processes, 
to have disappeared as such. Possibly, however, 
compar,ative evidence and closer study of 
Chinookan material may eventually reveal the 
former existence of the type. Such forms as 
-kala MAN (cf. also Wishram -gi-kal husband), 
-ka-holf EEL < *-ka'kwale' (cf. Wishram 
-gakwal\ -saJa fresh-water clam, /av.'//; six (cf. 
Wishram tx.^ni), and ca'y.m grizzly bear (cf. 
plural caya'm-ukc) are, for the present, sugges- 
tive but too isolated to constitute convincing 

As soon as we turn to the Penutian languages 
of California, we are at once struck by the 
large number of nouns, adjectives, and verb 
stems thatareformed on the pattern(c)vc,v(c2)-. 
Examples from Wintun ' are : 

I. Taken from S. A. Barrett's vocabulary, pp. 81-87 
of The Ethno-oeography of the Povio and jicighhoring 
Indians (University of California Publications in Ameri- 


VI American Indian Languages 2 


apti- mother's brother 

tahal tongue 

daka belly 

anak knee 

kama bed 

taral mush-paddle 

kada deer-snare 

aka bitter 

k'afta to sleep 

hara to go (D) ' 

elct infant 

k'me rib 

keicel house 

wene medicine 

ii'nik nose (Southerly dialect) 

iniil milk 

tkvi-l rattlesnake 

hi'ii' shell-beads 

t'iki sand 

luini to see (D) 

ko'so'I lungs 

poto intestines 

mrko' arrow 

no'Io' smoke 

do'ko' knife 

tco'nr fish-net 

cotok pestle 

SH'iio' nose (Xortherly dialect) 

pi fill' heart 

irdii'i rabbit-skin robe 

tii-nfl magnesite beads 

tii'nii'k cradle 

swrii't burden net 

yu'rw rain 

ku'pirm fingers 

ti'tcw- younger sister 

lu'iiiw dead 

iinrhii' to sin" 

can Archaeology and Ethnology, voL 6, 1908, pp. i- 

I. Forms marked (D) are quoted from R. B. Dixon's 
Outlines of Wintun Grammar (Putnam Anniversary 
Volume 1909, pp. 461-476). 

It is evident at a glance that the type is very 
much in evidence in Wimun. By actual count 
I find that considerably more than a third of 
the nouns and adjectives listed by Barrett for the 
« Southerly » dialect belong to it. 

Equally evident is the type in Maidu. From 
Dixon's grammar ^ I may quote : 

luaka' meat 

tsaka- pitch 

tsa'zva teeth 

paka sinew 

ma' tun hip 

yaha' good 

ivasa' bad 

-tala- to crush 

heme gopher 

sede blood 

benek tomorrow 

/;/;// eye 

hini net 

ki'ivi' back 

;///;//■ nipples 

yiini arm 

pili' dung 

si mi month 

tsibi' nails 

-piuili- to roll 

-tibil- to wind around 

U'dhol heavy 

koJo- to rotate 

orw- head 

ko'zuo' armpit 

yoso' field-mouse 

kono' baby 

nolo' pack-basket 

bono' ear 

poko' sun 

koyo' valley 

WO' no'- to die 

2. Maidu, an illustrative Sketch (Handbook of Ameri- 
can Indian Languages, BBAE : 40, part i, 19 10, pp. 679- 

Seven: Penutian Languages 



buhl tail 
butu fur 
suhu' smoke 
ku'lw evening 
viu'su' face 

It is clear that the type is as much alive in 
Maidu as in Wintun, and, as in Wintun, it 
seems, further, to be rather more characteristic 
of the noun than of the verb. This may be 
only apparent, however, as our knowledge of 
the verb morphology of these languages is far 
from complete. 

Turning to Yokuts, we find the type of stem 
formation still very much in evidence. I select 
from Kroeber's lists '. Examples of nouns 
are : 

V'apnd shoulder 
ma ,ad tears 
kabad wings 
tabak deer-rick 
capan coal 
wadak head-net 
k'twz't back 
tthzty head louse 
tczhvr, fog 
bidik faeces 
dimik prairie falcon 
tyipin sky 
idik water 
cikid arrow 
ip'i7i ground acorns 
P'ir,ik nose 
mi'kJc throat 
lir,it earth 
wdho'cit panther 
iirta youth 
oty^ hair 
p:>tJ penis 

I. See The Yokuts Lau^uas^e of South Central Califoinin, 
UCP 2 : 165-577(1907). For this paper I use only 
Yaudanchi forms (see pp. 240-254 for vocabularv). 

comot lungs 
■rpho grizzly bear 
coyod antelope 
topo■r^ buckeye 
tcoxJtc soaproot 
so'hn tobacco 
cohd hole 
odJt ball 
luocok belt 
putuc acorn 

The list might be greatly extended. A large 
number of verbs also belong to this type, e.g. : 

awaty to dislike 
tawac to be thirsty 
paxat to mourn for 
tcadax to turn 
dixid to make a basket 
pinity to ask 
pitid to tell 
pitciiu to catch 
icitid to pull hair 
luinis to be ready 
xityiw n be angry 
tdxitc ■ be sick 
i'dik to sing 
iv:)d:> to dodge 
od:^y to be on 
ciitnx to skin 
dukud to bury 
dumuk to sweat 
k'uyuk to scratch 
nuhuk to kneel 
hupnc to select 

Before certain suffixes these stems, like other 
disyllabic stems, lose their second vowel, e.g. : 

diikttil (imperative) ) 
dxkud-ji (past) ) 

( duhd-un (present-future) 
( duhd-ut (passive) 

According to Kroeber this stem alternation is 
purely phonetic, not morphologic, in charac- 
ter, so that direct comparison with the stri- 


VI American Indian Languages 2 


kingly similar stem alternations of Takelma 
(^e-g. zonst p'eleg- TO go to war : non-aorist 
p'elg-) is not feasible. However, this matter 
certainly needs further investigation, particu- 
larly as not all of Kroeber's forms seem to 
conform to his phonetic rule (loss of second 
vowel before a suffix beginning with a vowel). 
It may well be, on the other hand, that what 
was in remote origin a purely phonetic alter- 
nation in Takelma gradually assumed a mor- 
phological function. There is nothing to in- 
dicate this at present, however, for Takelma. 
In Miwok (Moquelumnan) the presence of 
the type (c)vc,v(c),- is as easy to demonstrate 
as in Wintun, Maidu, and Yokuts. For examples 
of nouns and adjectives I quote from Barrett's 
vocabularies of Coast Miwok ' : 

nawa old man 

ata elder brother 

saka cradle 

panak red-headed woodpecker 

patca poison 

luala Indian potatoe 

heleke neck 

ewe milk 

wene medicine 

killi antler 

ki'ti'lak butterfly 

pi'tci' nails 

ko'no' bow 

loko tule 

olok ocean 

posol lungs 

mo'M condor 

koto grasshopper 

tso'to'i short 

pu'tw infant 

pu'lu'k belly 

ku'lu'm bone 

pu'tcw wild onion 

I . See The Etbno-geography of the Porno and neighboring 
Indians, UCP 6 : 68-80. 

u'kw hand 
tirnu'k cradle 
su'ku'i sand 
mu'lu'-ta black 
n'lu'-ta red 
hu'hf stinking 

From Kroeber's material ^ the following verb 
stems are cited : 

yar^a to sleep 

kata to shut 

ete to see 

etepo to lie on one's stomach 

kelpe to swallow 

nete to count 

nipito to sit with folded leg 

hili to pinch 

oiuo to eat 

koyok to see 

toloye to hear 

lutsu to ascend 

uhii to drink 

iinii to come 

ynnu to kill 

kusH to sit with stretched leg 

nuxii to undress 

uku to enter 

iitcii to stay 

yiiku to swing 

Costanoan, finally, is perfectly analogous in 
this respect, as in so many others, to Miwok. 
Examples of nouns and adjectives are again 
quoted from Kroeber > : 

ama person 
wara body 
patcan blood 

2. See his Miwok sketch (pp. 278-3 I9)in The Languages 
of the Coast of California north of San Francisco, UCP 9 : 
273-435 (191 1). The Northern Sierra dialect is quoted. 

3. See vocabularies in The Chiimash and Costanoan 
Languages (UCP 9: 237-271 [1910]). 

Seven: Penutian Languages 



lawan bow 

wasar wind 

lanai neck 

hatac foot 

tcara sky 

wahan snow 

hesexem old woman 

£)'Ci beard 

/c^jyw jackrabbit 

t^d;'(?u rabbit 

wetel large 

j/;n' child 

kinir fish 

^/7m belly 

wznzf heart 

liti bow 

tt/fx/ fish 

worox hair 

/co/^o boy 

o/co ear 

koro foot, leg 

rono^ bow 

moto cloud 

yokovi ice 

/o/oi knee 

co/o^ moccasin 

to/o fire 

locko-, laskci- white 

^o/^o- black 

xntii belly 

kunuc pipe 

«;/n<x wolf 

pMJM/ small 

It is worth noting that disyllabic stems ot this 
type are in some dialects, particularly in that 
of Monterey, sometimes reduced to mono- 
syllables by the loss of the second vowel, e.g. : 

xurks neck 
tols knee 
tuis jackrabbit 

S. Juan Bautista 




Some examples, furthermore, of vocalic repeti 
tion may be only the secondary result of as- 
similation, e.g. : 

S. Clara oroc bear ores (four other dialects) 
S. Cruz lasa tongue lase- (four other dialects) 
tcipi knife S. J. B. Pipe 

S. Francisco tipa 

For examples of Costanoan verb stems of type 
(c)wc'y{c2y I quote from De laCuesta's Mutsun 
vocabulary ' ; 

atiia to eat 

ara to give 

axa to comb 

ata to examine 

icara to cut 

xaiva to call 

saya to shout 

ene to write 

ele to raise 

luexe to shield 

nietc to hide 

sepe to cut hair 

tere to cut hair 

pele to stick 

ipili to lie down 

witi to fall 

wixi to fish 

niini to strike 

xiri to make dried meat 

olo to become blind 

yoko to become ashes 

vioko to be born 

roto to drown 

xopo to give water 

tcorok to become sad 

iipu to pay 

utu to guard 

Inxun to stick in mud 

I. See J. A. Mason, The Mutsun Dialect of Costanoan 
based on the Vocabulury of De la Cuesta (in UCP 1 1 : 
399-472 [1916]). 


VI American Indian Languages 2 


rusu to spit 
siimti to decay 
ttipun to finish 
tculti to jump 

This list might be vastly extended. That the 
second vowel is felt as an integral part of the 
stem is shown by its persistence after infixed 
frequentative -s-, e.g. : 

ele-pu to go 
epe to pass 

else -pti many go 
epse many pass 

The final impression left on one's mind is 
that the stem type(c)vc,v(c2)-is quite unmista- 
kably, not to say exuberantly, represented in 
all the Penutian languages of California. It is, 
further, just as unmistakably in evidence in 
Takelma and Coos. Further north it appears 
to die out, in all probibility owing to the ope- 
ration of destructive phonetic laws. Its persis- 
tence to a minor extent in Lower Umpqua 
(Siuslaw), however, is highly probable, while 
further data may eventually indicate its 
presence also in Kalapuya. In other words, we 
are not far from having demonstrated its charac- 
teristic presence in the Penutian languages as 
a whole. 

Whether any conceptual significance origi- 
nally attached to this type of stem formation 
it is quite impossible to say at present ; quite 
probably it will always remain impossible. It 
will be interesting to determine later whether 

the Takelma alternation betwen the aorist verb 
stem with repeated vowel and non-aorist verb 
stem with single vowel is an archaic Penutian 
feature or a specialized development peculiar 
to Takelma. It is quite likely, indeed, to prove 
related to the alternation in Coos and Siuslaw 
between the short form and the « amplified » 
form of stem (generally with a-vowel), e.g. : 

Coos tkwi'tl- to follow tliwi'yali 

tC'.Jiti- to reach tanatl 

Siuslaw tku'nt- to close tkwani- 

hi'ts- to put on h'yats- 

i'iq- to dig yaiq- 

hau'- to be ready haiua- 

wi'in'- to agree lui'hca- 

The primary function of the amplified stem is 
the indication, apparently, of durative activity. 
Assuming the « amplified » stem of these lan- 
guages to be related to the aorist (present-past) 
stem of Takelma, it remains to be determined 
whether their amplifying rt-vowel is genera- 
lized from stems with radical a {e.g. Siuslaw 
hail'-) or, on the contrary, the repeating vowel 
of Takelma is a secondary assimilation from an 
older -a- (e.g. p'eleg from *p'elag-'). More pene- 
trating study of the Californian Penutian lan- 
guages mav help to solve this and other simi- 
lar problems. 

Seven: Penutian Languages 273 

Editorial Note 

Originally published in International Journal of American Linguistics 2, 
58-67 (1921). 

Sapir's suggestion of the "characteristic presence in the Penutian languages 
as a whole" of disyllabic stems with a repeated vowel should be compared to his 
identification of a similarly diagnostic stem-type (with deletable initial vowel) 
for Hokan (see 1917e, 1920b, 1921o, and 1925b). 

There is evidence that Sapir had second thoughts about Penutian stem shapes 
after the completion of this paper in early 1919. In November of that year, fol- 
lowing the publication of Dixon and Kroeber s full lexical evidence for Califor- 
nia Penutian (Dixon and Kroeber 1919: 56-61), Sapir wrote to Kroeber, "I feel 
the typical Penutian stem is long, often 3 or 4 syllables in original form," sug- 
gesting such reconstructions as *ilapita 'fish' and *yikati 'one' (Golla 1984: 315). 
A characteristic stem-type is not included by Sapir as a diagnostic trait of Penu- 
tian languages in his final classificatory statement (1929a: 140). Later work in 
comparative Penutian morphology has tended to emphasize the centrality of 
stem vowel ablaut, frequently coupled with suffixation as a "morphological pro- 
cess," rather than focusing on stem shape per ^e (Hymes 1957; Silverstein 1979: 


It is the purpose of this paper to show how 
the operation of a phonetic law, hitherto unno- 
ticed, brought about a number of irregularities 
in the use of pronominal elements in Chinoo- 
kan. Certain incidental inferences on more 
fundamental points of Chinookan linguistic 
history also suggest themselves. These will be 
briefly referred to at the end of the paper. 

If we examine the Chinookan system of 
transitive and intransitive pronominal prefixes 
of the verb and corresponding possessive pre- 
fixes of the noun ', we shall note three appa- 
rently unrelated irregular features which 
involve an alternation of g (which may be 
modified to k or kx ^) and the palatal sibilant 
aff'ricate tc. These are as follows : 

I. The possessive prefix for the third person 
singular feminine (" her ") is -ga- when the 
noun itself is feminine, neuter, dual, or plural, 
i.e. is preceded by the gender-number pre- 
fixes : 

sing. fem. 

Lower Chinook Wishram ' 

6- (w) a- 

L- ii- 

C-, s- ic-, is- 

t- id- (it-) 

but \s-tca- when the noun itself is masculine, 

1. See F. Boas, Chinook (in Handbook of American 
Indian Languages, Bureau of American Hthnology, Bul- 
letin 40, part I, igii.pp. 559-677), particularly pages 
580, 581, and 585. 

2. For consistency's sake I am preserving Boas' Chin- 
ook and my own Wishram orthography without modi- 

3. An Upper Chinookan dialect. I quote from my MS 

i.e. is preceded by the gender-number prefix : 
masc. sing. e-, i- (jio)i- 

Examples (Wishram dialect) are : 

Absolute Possessive : " her ' 

wa-ska'n cup a-ga'-skan 

ii-tcqiua' water il-ga'-cq 

is-qxu's eyes is-ga'-xus 

i't-qHi house it-ga'-qH 



wi'-lxam village i-tca'-lxatn 

1. The possessive prefix for the first person 
singular (" my ") is -^e- (Wishram -^-, -k- ; 
-.y- before /e-stops) when the noun is femin- 
ine, neuter, dual, or plural, but -/^e-, -ki- 
(Wishram -tc-') when the noun is masculine. 
Lower Chinook (C.) and Wishram (W.) 
examples are : 

Absolute Possessive : " my " 

C. o-pO.ike bow d-gu'-pL\ike {-gu- labra- 

lized from -^e- because 
of preceding 0-) 

W, a-kni'm canoes a-x-hii'm 

C. L-gE'-qacqac my grand- 


W. is-qxu's eyes is-k-xu's 

W. i'f-pc feet i't-k-pc 

but : 

C. i-tslENiEnd wooden 

W. wi'-iq body 



VI American Indian Languages 2 


3. Aside from certain secondary irregularities 
in the third person dual and third person plur- 
al which do not concern us here, the pronom- 
inal subject of the transitive verb differs from 
the pronominal subject of the intransitive verb 
(and pronominal object of the transitive verb) 
only in the case of the third person singular 
masculine and third person singular feminine, 
the difference between the two sets of forrns 
being for the most part indicated by position 
(the subjective pronominal prefix preceding the 
objective pronominal prefix) and, in part, by 
the use of a " postpronominal " particle -g- 
which indicates that the preceding pronominal 
element is used as the subject of a transitive 
verb. For " he " (and " him ") and " she " 
(and " her "), however, the following distinc- 
tive forms are used : 

Intransitive Transitive 

3d person singular, masc. tc- he z'-he, him 
fem. ^- she a- she, her 

The forms will be better undestood from the 
following Wishram examples : 
i-tc-i'-uwaq he killed him (/- is temporal ; tc- 

" he " ; -i- " him ") 
i'-tpx he comes out of the house (/- " he") 
i-g-i'-uwaq she killed him (-^- *' she ") 
a'-lpx she comes out of the house (a- " she ") 

Contrast : 

i-m-i'-uwaq you (sing.) killed him (-///- 
" thou ") 

a-ni-tbay-a you will come out of the house 
(a- is temporal) 

i-tc-m-u'woq he killed you, 
in which -m- " thou, thee " is used both as 
transitive subject and as intransitive subject and 
transitive object. 

How are wc to explain these irregularities ? 
The distribution of the forms in question is 
such as to make it probable that we are dealing 
with a phonetic factor rather than a morphol- 

ogical one in the first instance. Cases i and 2 
are parallel phonetically : 

1. a-ga- " her " (fem. noun) : i-lca- " her" 
(masc. noun) 

2. a-g- " my " (fem. noun) : i-tc- " my " 
(masc. noun) 

and suggest at once that the masculine prefix 
/- palatalized the older -ga- " her ", -^(e)- 
" my ", to -tea-, -/^(e)-, perhaps via palatali- 
zed k- sounds (*-§a-, *-^^-). But how account 
for the forms in case 3 (fem. g- : masc. tc-\ 
and why should only these forms be exclusive- 
ly characteristic of the subjective transitive 
verbal paradigm ? If we venture to reconstruct 
them in accordance with cases i and 2, we get : 

*ag- " she" (transitive subject) 
*itc- " he " (transitive subject) 

< V 

The phonetic parallelism would then be per- 
fect in the three cases. If we compare the theo- 
retical forms *ag- " she " and *itc- " he " with 
the remaining subjective forms of the transi- 
tive verb, we obtain at once a perfectly regu- 
lar and intelligible set of forms. Including the 
" post-pronominal " -g-, the system is as fol- 
lows : 

1st pers. sing: n- 

exclusive dual nt-g- 
exclusive plural nc-g- (also heard as 

ntc-g-, with f-glide) 
inclusive dual /a:-^- (simplified in Wish- 
ram to t-g-^ 
inclusive plural Ix-g (simplified in Wish- 

hram to /-f-) 
2nd pers. sing, m- 

dual mt-g- 
plural mc-g- 
3d pers. sing. masc. *i-tc- <i*i-g- 
sing. fem. *a-g- 
sing. neut. i-g- 
dual. c-g- 

plural t-g- 

Seven: Penutian Languages 



Compare these pronominal prefixes with 
the corresponding intransitive subjects (and 
transitive objects) : 

1st pers. sing. n- 

exclusive dual nt- (nd-) 

exclusive plural nc- (nlc-^ 

exclusive dual tx- 

exclusive plural Ix- 

2nd pers. sing. tn- 

dual mt- (^md-) 

plural mc- 

3rd pers. sing. masc. i- (y- before vowels) 

sing. fern. n- 

sing. neuter i- 

dual c-(in certain cases c/-, erf-) 

plural t- (in certain cases u- 
gwa-, Lower C. o-go-) 

Aside from the irregular intransitive subjec- 
tive (not objective) forms in the third person 
dual and plural (ct- ; ti-gwa-), whose use is 
limited to certain cases, the transitive par- 
adigm obviously derives from the intransitive 
by the addition of a transitivizing particle -g- 
to the pronominal element, except in the 
first person singular («-) and second person 
singular (w-), in which cases position alone 
differentiates the transitive and intransitive 
subjective uses. If our analysis is correct, the 
actual transitive subjects for the third person 
singular, masc. Ic- and fem. g-, are not true 
pronominal forms in origin; the older pronomi- 
nal elements, i- and a- respectively, still in 
evidence in the intransitive praradigm, have 
disappeared as such but have left their tVace in 
the different treatment of the old transitivizing 
-f-, which now appears in twofold form and with 
transferred function as tc- " he " and o^- " she. '' 

We may therefore reasonably infer that in 
all three cases what now appears as a peculiar 
morphological alternation of^^: tc is really a 
survival of an old phonetic law, according to 
which g (k) was palataUzed by immediately 

preceding i to anterior palatal ^ Q), which in 
turn shifted at an early period of Chinook his- 
tory to tc. Presumably an old *ik\ shifted to 
itc\, but I have no evidence of this. The law 
is no longer operative as such. It had run its 
course long before Chinookan split up into its 
present dialects, its consequences are now of 
a strictly functional character, and its operation 
was probably checked at an early period by 
analogical leveling. There may at one time 
have been such alternations as *i-tca'la " man " : 
it-ka'lukc " men " or *n-i-tcim " he said " : 
n-a'-kim "she said", which were then leveled 
out to the forms i-ka'la : it-ka'lukc, n-i'-kim : 
n-a'-kim that we now possess. But there is 
nothing to prove this and it is more probable 
that the phonetic law had ceased to operate 
before the welding of noun and verb stems 
with pronominal and with gender-number 
class prefixes. It is not at all unlikely that such 
elements as /-, wi- of masculine nouns and 
n-i-, n-a- of verb forms were independent elem- 
ents or assemblages of elements (e.g. *w-i 
" he " ; *n-i " then-he ", *n-a " then-she "), 
which became attached to noun and verb stems 
at a comparatively recent date. On the other 
hand, we must assume that such assemblages 
as *i-tca < *i-ga " hers " (masc), *i-tc3 < 
*i-gd " mine " (masc), and *i-tc3 < *i-g3 " he " 
(transitive subject), *a-gd *' she " (transitive 
subject) and *c-g3 " they two " (transitive sub- 
ject) formed firm units at a much earlier date. 
Within such units the phonetic law could 
operate but not outside of them. 

Comparative evidence, making use of data 
outside of Chinookan, may some day succeed 
in confirming our phonetic law by showing 
that certain cases of tc in stems go back to g 
(/^) after ;■ (say -i/c, fem., "tail of quadruped "), 
but at present we cannot do this. So far I 
know of only one other case of tc which may 
be presumed, with some plausibility, to derive 
from palatalized k. This is Upper Chinook 


VI American Indian Languages 2 

1 08 

-i-tc{i), which forms personal plurals of demons- 
trative and personal pronouns, e.g. Kathlamet 
La-i-tci " those " (indef.), ta-i-tci " those '' 
(def.), based on pronominal stems La- " it ", 
ta- " they " ; Wishram da'-i-tc, fa'-i-tc, a'-i-tc 
" they ", based on pronominal stems da- 
" they ", in- " it ", a- " she"; Wishram 
da'uda-i-tc " these people ", h'xia-i-tc " yon 
(indef.) people ", based on demonstrative 
da'iida, la'xia '.This suffix corresponds morphol- 
ogically to -kc-, -di-kc, -i-kc of other forms 
(e.g. nominal plural -^^in cases like W. it-ka'- 
lu-kc " men " ; W. ia'it\i-kc '' they too ", cf. 
fa'itla " it too " ; da'ima-di-kc " they alone ", 
cf. da'ima ; Kathlamet tata-ihc " these people ", 
cf. tala-x " these "). Presumably Upper Chin- 
ook -i-kc is umlautcd from older *-a-kc -C 
*-a-ki-c (cf. ia'ii\a : fa'it\i-kc ; -di-kc < *da-ki-c, 
personal plural in *-ki-c oi da- " they "). In 
Lower Chinook plural forms in-kc,-i}ic,nnd -tike 
occur plentifully with nouns, both animate and 
in animate (e.g. L-q\eLxa'pu-kc " coats ", /-w'- 
gala-i-kc " his firrs ", id'wux-ti-kc "his younger 
brothers "), but not with demonstrative or 
personal pronouns. The Lower Chinook suffix 
for personal plurals in the demonstrative is -c 
(t.o.x'ltac " those people", visible, qota-c "yon 
people", invis.), probably the same element as 
the -c- of prefixed nc- " we " (excl.) and mc- 
" ye " (cf. n- " I ", m- " thou "). The best 
way to explain the various plural suffixes in 
Chinookan seems to be to assume an old ele- 
ment *ki, preserved in palatalized form in Upper 
Chinook -i-tc{i) ; a plural element -c \ and a 
double plural -kc <C *-ki-c. 

Another survival of the old *-ki plural may 
be Lower C. L-a-tcl " mothers " •< *i-a-ki-t : 
L-aa " mother ". 

The data presented in this paper suggest a 
number of further problems, which we can 
hardly do more than touch upon. 

I. The disappearance of */- and *a- in the old 

I. See F. Boas, op. cit., pp. 623, 62 5, '^627. 

iorms*itc(py "he"and*a^^- "she" is merely an ' 
early phase of a phenomenon that seems to have 
been characteristic of Chinookan at all times, 
the loss of short unaccented vowels. The accent 
of Chinook is a strongly expiratory one, seems 
to have been regulated by morphological consid- 
erations (contrast W. gahi'pa " she went out of 
the house " whith aluba'ya " she will go out of 
the house " ; future -a, as shown also by Lower 
Chinook evidence, shifts the accent forward), 
and has left in its train a number of phonetic 
consequences, both early and dialectic, the 
chief of which are the disappearance of short 
vowels (cf. vowelless stems like -iq " body ", 
-tcktc " to wash"; alternations like ti- " I", 
as verb prefix, with independent na'-, W. -0^- 
" my " with Lower C. -^e'- ; loss of final 
vowels which reappear in protected forms, e.g. 
Lower C. i-saniEF " lid " <; *-sa'm^lga : L-ia- 
SEniElqa-ks '' their lids " -< *-savid'lqa-), the 
shortening of unaccented and the lengthening 
of accented vowels (e.g. Lower C. *i-cayiin 
" grizzly bear " : plur. L-caya'm-u-kc <C early 
Chinook *cayam : *caya'm-)., and the weaken- 
ing of consonants after unaccented vowels 
(cf. above examples of W.'/): /»'; Lower C. q: 'g'). 

2. As regards the old unaccented */- and *a- 
which we must suppose to have disappeared 
before transitivizing *-gy-, we may note that 
it is characteristic of all transitive forms that 1 
the pronominal subject is unaccented, while in 
many instances the intransitive subject receives 
the accent. Lower Chinook seems to preserve | 
the old accentual conditions better than the 
upper 'dialects, which have undergone further 
shifts of accent with resulting loss of reduced || 
vowels (e.g. Lower C. atcE'tax " he made 
them " : W. gatclux^. A good ex^ample of such 
accentual alternation is Lower C. le'- " it " 
(intr. subj.) : L- " it " (tr. subj.) in aLE'nkatka 
" it comes flying above me " : aLgigE'ltcxEm 
" it sings for him ' ". This is not the place to 

I. Op. cit., p. 588. 

Seven: Penutian Languages 



pursue the matter further, but we may at least 
point out that the transitivizing *-gi- is in all 
probability identical with the -gE- of the 
** adverbial prefixes " gE-l- "■ for, on account 
of" and -gE-tn " with, near ". It is remarkable 
that the pronominal element (indirect object) 
to which -gEl- and -ghn- are suffixed never 
receives the accent, which either strikes the -^e- 
or some syllable following it. In other words, 
the treatment of pronominal element + tran- 
sitivizing *-^<?-and thiU of pronominal element 
-f- indirective -^e/-, -gEm- are parallel. But 
note that the pronominal element (indirect 
object) is frequently, even typically, accented 
before the " adverbial prefix " (really postpo- 
sition) -/- " to, for ", with which -gEl-'\s com- 
pounded; e.g. W. inia'lut *' I gave it (masc.) 
to her, " Lower C. Ld'Joc " it was to her ". 
If -/- and -gEl- were strictly parallel elem- 
ents, it should be possible to have such paral- 
lel forms as a'-l- " to her " and a'-gEl- " for 
her ", whereas we consistently have a'-/- " to 
her" but a-gE'l- or a-gEl-' " for her ". This 
can only mean that -/- and -gvJ- are not mor- 
phologically parallel, but that -gE- is an element 
which somehow displaces the pronoun and 
draws the accent to itself. Its power to take the 
accent away from preceding elements is further 
indicated by the fact that it regularly occurs 
with voiced 0-, not with voiceless k or affricative 
kx (cf, remarks in i.). 

What is this old element *-^^'-, which now 
appears as transitivizing -^(e)-, as third person 
singular masculine and feminine transitive 
subject (/c[eJ- " he ", g^E)- " she '"), and as 
first component of the verb prefixes -gEl- 
" for" and -gEm " with " ? It seems likely that 
it is an old demonstrative or deictic stem which 
is either predicatively related to the preceding 
pronominal element or which serves to emphas- 
ize or displace the pronominal element and 
to which the postposition (-/-, *-ni-) is at 
tachedas an enclitic. A cluster of elements like 

i-gEl- " for him " originally meant i-gE'-l- 
" him -that one-to, for him ". Similarly, at a 
far earlier stage, a transitive cluster like t-gE-n- 
(W. t-g-n-') " they (subj.) -me (obj.) " or 
*i-gE'-m- > *i-tcE-m- >> tcE-m- (W. tc-m-') 
" he (subj.) -thee (obj.) " really indicated 
" ihey-that (it is) -me ", " he -that (it is) 
-thee ". It this is correct, the original difference 
between the intransitive and transitive phrase 
must have been one of sentence idiom. " He 
goes " was expressed as " he goes ", but *' he 
kills her " as " he that (is) (who) kills her "'. 

This deictic or demonstrative *o-p'- can only 
be a reduced form of post-accentual -ka, which 
occurs freely in Chinookan numerals, pro- 
nouns, and adverbs as deictic element (" only, 
just "). Examples are W. i'xt-ka " just one ", 
na'i-ka " I ", a'x-ka " she ", Lower C. nd'm- 
hi " I alone ", e'-ka " thus ", kaiua't-ka " soon ", 
nau'it-ka " indeed ", W. iiua't-ka " to yon 
(place) ". This deictic '-ka, in turn, is obvious- 
ly merely an enclitic use of an old demonstra- 
tive stem ka " that " which is no longer in 
free Chinookan use but which survives in Lower 
C. ka, c-ka " and " and as petrified temporal 
ka- " that (time) " in Lower C. ka-iua't-ka 
" soon " (< " to just that [time] ", parallel 
in form to W. i-wa'l-ka "to just yon [place] ") 
and ka-wi'x' " early " {c{. luiix't '' tomorrow ", 
W. wax "dawning") ; cf. also W. tense pre- 
fix ga-, ga-l- of remote past time. All these 
Chinookan elements (ka, ka-, -ka ; ga-; -gE-), 
finally, are reflexes of a wide-spread demon- 
strative stem *ka " that ", often used as a gener- 
al term of reference, found in other Penutian 
languages (e.g. Coos -kd in xd-ka " he " ; 
Takelma j^^ " that"; Yokuts/fert " that "[vis.] ; 
Miwok i-ka " that " ; Tsimshian -_^e absent 
connective, -ga absent demonstrative added to 
final noun in sentence). 

3. It is fairly clear that the two fundamental 
factors in the development of the somewhat 
irregular morphology of Chinookan were a 


VI American Indian Languages 2 

I 10 

strong and movable stress accent and, as a 
result of this, the tendency for vowels to drop 
out and for originally independent elements to 
melt together into complex assemblages. Thus 
the old sentence, which seems to have been 
constructed on rather simple, analytical, lines, 
tended more and more to petrify into a highly 
synthetic sentence-word. We have already hint- 
ed at the probability that the phonetic change 
of g to tc antedated the inclusion of certain 
elements in the verb. Internal evidence makes 
it practically certain that at least the tense pre- 
fixes were late in coming into the verb com- 
plex. In the first place, the tense prefixes of 
Lower Chinook differ considerably from those 
of the upper dialects ', so that it looks at 
though an old set of temporal particles or 
adverbs (Lower C. a ; «-; W. a, a-l; i, i-g-^ 
na, na-l ; ni, ni-g ; ga, ga-l \ ;/-) had coalesced 
with the following pronominal prefixes of the 
verb in the independent life of the various dia- 
lects. Moreover, these elements do not behave 
as though they had ever coalesced into a 
phonetic group with the early Chinookan 
forms of the transitive forms for " he" (^iicc-) 
and " she " (^ags-'). Thus, in Wishram we 
have forms like i-g-i'-ux " she made him ", in 
which the tense prefix /- does not palatalize 
the following -g- to -Ic-, no doubt because it 
did not enter into the verb complex until long 
after the palatalizing effect of an /- had spent 
its force. If the /- had been prefixed at the time 
that the pronominal element^^- " she "still exis- 
ted in the fuller form *agj-, it would have re- 
quired an intervocalic -g~ and the form *ig-ag3'- 
would have arisen (cf. modern forms like ig-a- 
-tpa "■ she came out of the house "). Similarly, 
a form like Lower Chinook atcE tax " he made 
them "evidently arose before the tense prefix a- 
was part of the verb complex, for a- could not 
have palatalized an original *-_^?- to*- to- while the 
I. Op. cit., pp. 577-79. 

older pronominal form */to- would have requi- 
red as tense prefix the prc-vocalic «-, hence 

4. It is a well known linguistic phenomenon 
that similar or identical sounds, groupings of 
sounds into phonetic patterns, or phonetic pro- 
cesses may characterize a number of indepen- 
dent languages or even linguistic stocks within 
a continuous area. Such examples are sugges- 
tive of phonetic interfluences between distinct 
languages presumably through the medium 
of bilinguals. The change oi g or k to dj 
or tc because of the palataUzing influence of a 
preceding or following front vowel (/ or e) is 
perhaps too general a process to warrant our 
attaching much importance to its occurrence in 
a number of contiguous languages. Neverthe- 
less it is of some interest, and, it may be, of 
historical significance, to point out that the 
change of ^- sounds or of palatalized k- sounds 
to Ic- sounds is found in a continuous or nearly 
continuous area from a northern point on the 
west coast of Vancouver Island south to the 
mouth of the Columbia. All the Nootka dia- 
lects, both NootKa proper and Nitinat-Makah, 
have altered the original Wakashan anterior 
palatal Ic- sounds, preserved in Kwakiutl, to 
corresponding palatal sibilant affricatives; Kwa- 
kiutlj g' (^'•)and t (^)appear as Nootka tc; t \ 
(^') as tc' ; and x (.v) as c. A large number of 
Salish dialects, furthermore, have altered the 
original unlabialized /.'-sounds to /^-sounds. I 
am quoting Dr. Boas' personal statement on 
this point and am unable to give the geograph- 
ical distribution of the Salish /^-dialects. I 
should perhaps add that the Lower Chinook k'- 
Q-) sounds, which correspond to ordinary k- 
sounds in Wishram, are a comparatively recent 
dialectic development before /-vowels and that 
they have nothing whatever to do with the 
old, general Chinookan, change of 0^ to tc after 
/-vowels which is the subject of this paper. 

Seven: Penutian Languages 


Editorial Note 

Originally published in International Journal of American Linguistics 4, 

Silverstein (1977), working with richer comparative data, has reaffirmed 
Sapir s analysis of the Chinookan ergative prefixes, although he posits a basic 
zero marker, rather than a-, for the third person masculine subject/agent pre- 
fix. Like Sapir, Silverstein sees "heavy categorial influence from the languages 
surrounding [Chinookan] on the coast," resulting in a rich elaboration of 
ergative inflectional morphology out of an earlier "nominative-accusative phra- 
sal-enclitic syntax" (1977: 154). 

Sapir's ms. corrigenda on his copy are as follows: 



p. 106, 



first table 

p. 107, 



last para. 


p. 108, 



1. 11 


p. 108, 



1. 22 

in animate 

p. 108, 



fn. 1 


p. 108, 



1. 1 


p. 108, 



1. 3 


p. 108, 



1. 8 


p. 108, 



1. 26 


p. 109, 


1. 4 


p. 109, 


1. 5 


p. 109, 


1. 18 

to her," 

p. 109, 


last 1. 


p. 110, 


1. 25 


p. 110, 



1. 29 

Kwakiutl, g- 


Reverse headings "intran- 
sitive" and "transitive" 




to her" 

tached as 


Kwakiutl g- 


By L. S. Freeland. 


THE structure of Mixe is rather bare and 
scanty, at least when compared with the 
rich and fairly intricate grammatical patterns 
found in some of the Penutian tongues of 
California and Oregon. One gets the im- 
pression that in Mixe the morphology has 
worn thin. 2 There is a strong tendency to 
fall back on word-order to express relational 
ideas. New morphemas appear which are 
but one step removed from concrete seman- 
temas. In some cases these "empty words" 
still retain their full concrete meaning in 
other connections, as for instance the word 
mid which is the verb "to have". We find it 
in Mixe used to express the concept "with, 
of" very much in the same manner as the 
"belong" of Beche-de-mer. In Miwok or in 
Maidu this would be expressed by a comita- 
tive suffix. 

^ The Mixe material for this study was obtained 
by J . de Angulo in Oaxaca during the course of a 
linguistic survey of that region for the Department 
of Anthropology of Mexico under the Direction of 
Dr. Manuel Gamio. 

After writing this article I sent it to Dr. Edward 
Sapir for criticism and suggestions. Dr. Sapir was kind 
enough to send me a Ust of further cognates from 
Takelma, Coos, and other languages from the northern 
Penutian group. Dr. Sapir was also kind enough to 
write some marginal notes which I have taken the 
liberty to reproduce as footnotes with the initials K. S. 
In his communication to me Dr. Sapir says that Prof. 
Roland B. Dixon had already been struck by the Penu- 
tian character of Zoque, and had sent him a list of 
Penutian cognates in this language. Zoque is very 
close to Mixe. The credit for the discovery of a Penu- 
tian language in southern Mexico belongs therefore to 
Dr. Dixon, by priority. 

^ Cf. analytic wearing down in Mukne and Coos as 
contrasted with Miwok and Takelma. — E. S. 

One somehow gets the impression that the 
Penutian morphology having reached too 
great a degree of intricacy in its usual 
methods for expressing the relational, has 
thrown the whole baggage overboard and 
started on a new tack. This may be due in 
some measure to the influence of other 
Central American tongues which seems to 
be in the direction of sparseness of linguistic 
expression. Nevertheless it is true that a 
very similar drift is at work in some places in 
California, for instance in the western dialects 
of Miwok as compared with the eastern 

Whatever may be the interpretation, the 
fact of the morphological sparseness of Mixe 
remains, and naturally precludes extensive 
morphological comparisons. The evidence 
for classing Mixe in the Penutian family must 
therefore necessarily^ be largely lexical. But 
in spite of its limited grammatical apparatus 
Mixe possesses morphological traits that 
have a strong Penutian flavour. We will 
consider these briefly: 


As in Maidu, in Miwok and in Yokuts, so 
in Mixe the radical often appears in several 
forms so closely allied that it is sometimes 
difficult to say which is the primary one. 

^ This is just what happened in Chinookan, where 
older Penutian features are only survivals and a new 
autonomous and rather complex morphology was built 
on top of the older system by means of coalescences of 
sequences. — E. S. 


VI American Indian Languages 2 


In Maidu these symbolic changes express 
derivational concepts, in Miwok and Yokuts 
temporal aspects, in Mixe differences of 

Examples (a hyphen indicates the omission 
of irrelevant affixes) : 

Maidu : 

a) tot to drag ; tttt to touch ; tat to stroke ; tet to reach 

b) tea to eat grass; tco to eat meat. 

c) yopom-ia to throw; yapam to slap. 

d) witcap to tear; witcep to tear a little; witcohit to 
wring; witcup to tear off. 

e) ivitcot to skin a hare or small animal; iviicut to 
scrape hides; witcat to split with the hand; 
ivitcet to split a small thuig. 

Miwok: 2 

to play 
to die 
to smoke 
to run 

Yokuts : 

cadik, cadak to awake; tcadxin tcadax to turn; 
ipe, epi to get water; hiwet, hcut to walk, go, move; 
tawidc, taudj to overcome ; waid, waadi to break up ; 
diikdu, dukud to bury; tui, toy to shoot. 


TiME Removed 


FROM Present. 















to get up 

to fall 

to like 

to sleep 
































Yokuts and Mukne (Costanoan) are appar- 
ently lacking in incorporated pronouns. 
Southern Maidu uses independent forms in 
the ordinary statement, but possesses v^hat 
seems to be a depleted series of suffixed 
pronouns (possibly old objective forms), used 

Examples : 

in subordinated expressions of time and in 
hortatory and optative modes. Miwok is 
rich in pronominal series, specialized in their 
use for present, past, future, subjunctive 
and the Hke. Although the}'' are much less 
numerous the Mixe forms come closest to 
those of Miwok. 3 




Past and 



-5, -se 






-ne, -no, -men 










y-, IV 


-es, -nes, -hese 







-maw, -mem 










y-, w 

Both Miwok and Mixe show a considerable 
amount of fusion and obliteration of elements 

^ Cf. internal vocalic changes of verb themes in 
Takelma, Coos, Lower Umpqua. — E. S. 

- These alternations are amazingly like the Takelma 

in putting together their verb forms. This 
gives often quite a similar impression: 

ones, e. g. ydan- "to go": fut. yana-. — E. S. 
^ Takelma like Miwok and Mixe. — E. S. 

Seven: Penutian Languages 





I have eaten 

i'wicak I 


hi' kcebes 








i' w'icakmas 



i'wicak tos 






Rather a large number of the Mixe verbs 
in the list are bi-syllabic, and many can be 
analyzed into two elements. They suggest 
very strongly the Maidu verbs with prefixes 
or first-position-stems such as he- action with 
the foot, ka- action with the hand, and the 
like. Mixe probably has nothing like the 
number of these prefixes that there are in 
Maidu, but three or four are quite clearly 
discernible and very suggestive: 

Maidu: heye creep; bedokoi run; bedoi 
kick; helom step over; hewet hurry, kadokok 
push away; kapadum roll up; katut touch; 
katcaduk grab; pai foot, trail; -pat after. 

Mixe: petc-h to climb; pegig to run; pedig 
to rise; pasl-h to go out; pa- along, by; 
pa-otsm to go along; pa-witsm to creep; 
pa-nas to pass by {nas "the ground") ; pa-boy 
to chase; pa-son to follow; pa-wep to kick 
[wep to strike) . kuhath to pay ; kudigx to put ; 
kutuk to command; kuyat to play; kunas to 
fall {nas ground); ni-pan to cover; ni-kes-b 
to be on top of something high [kos, kes 
above, on, upon); ni-wits to close; neksm to 
go, ni-neksm to go for the benefit of some 
one else; kutuk to order, ni-kutuk to order 
through the agency of another; kes-wits to 
place; kes-ets to place to boil {ets-b to boil); 
niwis-dud to open {niwits to close) ; nipan-dud 
to uncover [nipan to cover) ; rswits to take 
off [kes-wits to place, ni-wits to close) ; 
rspit-n a skirt [pit to make thread) [rspit-m 
also means "to tangle"); rsmats to let go 
[mats-b to take, pats-b to find) ; rsmats 
trotisers ; 

II. Semasioi,ogicax^ 


1. sun: si Cf. Maid, sa fire; Miw. hi, hit sun, day; 
Muk. icnien, hismen sun; Wint. sun sun. 

2. moon: po Cf. Maid, pombok moon (also poko 
sun, poho night, poketcuk star); Yok. opodo, 
upic moon; Wint. po fire; Tak. be svm; Kal. 
pyd-'n sun, day. 

3. shine : tuks Cf . Miw. lulep- shine ; Muk. tuxts day. 

4. night: su, us Cf. Miw. sien- be night; Wint. 
sinol night; Tak. xii'n night (Tak. x regularly 
comes from 5). 

5. cloud: yots Cf. Maid, ya; Tak. hat; L. Um. hiai ; 
Pen. * h-yai. 

6. sky: tsap Cf. Muk. tcarak. 

7. star: musa Cf. Muk. mur night. 

8. fire: hen Cf. Miw. wik-, wik-, wul-, wel-; Yok. 
licit, oxit; Muk. xii, xihi, xute go for a fire, 
light a fire ; Tak. -'/cm/- to set fire to, to catch 
fire; Coos cutl to set fire, Hcil- to burn, 'tcwel- 
fire; Pen. * swilihe, *'lswil- to bum (cf. "to 
burn" next semantema).^ 

9. bum isai Cf. Maid, tco; Miw. tsup. 

10. salt kan Cf. Muk. akes, aks. 

11. dry tes Cf. Miw. tsutul. 

12. stone: tsa Cf. Miw. cawa; Muk. isin. 

13. earthquake: pumimb Cf. Wint. pomoko. 

14. hole: Aw/ Cf. Maid, tuke; Muk. kutui. 

1 In making the following list of semantemas I 
found it necessary to alter somewhat the orthography 
of the different sources in order to obtain a common 
basis of comparison. The system of transcription we 
have followed is that of the American Anthropological 
Association. It may be well to point out that in Mixe, 
s and c are interchangeable ; i' and f'are the "unrounded" 
forms of u and o; in Mixe the dynamic stress accent 
is invariably on the semantema, and this in turn is so 
seldom of more than one syllable that a di-syllabic stem 
may be looked upon confidently as resolvable under 
further scrutiny. 

For Miwok, Wintun and Maidu, I have made use of 
my own unpublished material and of that of Mr. J. do 
Angulo, also of the grammar of Maidu by R. B. Dixon 
in the first volume of the Handbook of American Indian 

For Pomo I have made use of my own and J. de 
Angulo's unpublished material, also of The Ethno- 
geography of the Pomo Indians by S. A. Barrett. For 
Mukne I have used the Mutsun Dialect of Costanoan 
by J. A. Mason. 

For Yokuts I have used The Yokuts Language of 
Southern California by A. L. Kroeber. 

Many of the kinship terms are to be found in E. W. 
Gif ford's exhaustive work "California Kinship Termi- 

* Pen. *swil- "fire" ? : Pen. *'tswil- "to burn" ? — 
E. S. 


VI American Indian Languages 2 

15. water, river; ni Cf. Maid, dim wet; Miw. nuni- 
to be submerged; Wint. hm boat. 

16. wet: 50^ Cf. Miw. cumek to soak up. 

17. rain: tub, tsuy Cf. Miw. upa; Yok. ciwek drizzle. 
iS. wash: iL'ibux Cf. Maid, dp to dive; Miw. upux 

bathe, swim; Yok. ep-, ip- swim; Muk. upxi 
to sip; Tak. 'pag- to bathe. 

19. wind: poh Cf. Maid, biye; Miw. pwc to blow; 
Yok. piuate whistle; Muk. put blow, pusninyis 
whirlwind; Wint. pul whistle; Tak. -p'ouw- 
to blow; Chinook po to blow; Kal. ptU to blow; 
Coos pjix- to spout; Tsim. bmi to blow. 

20. red: (saps Cf. Muk. lasas. 

21. green: stisA Cf. Maid, kotcis; Muk. icutsu. 

22. then: we/ Cf. Miw. ne this; Muk. «e this. 


23. Temporal affixes: Past te- Cf. Miw. -/it; Muk. 
-te, -kie. "essential" : -b, -n Cf. Muk. -n infinitive; 
Tak. -n noun suffix, sometimes used to make 
verbal nouns, exactly as in Mixe; Maid, -n 
infinitive ending, "contmuative" -its, -ats 
(forms of the verb to be) ; Cf . Miw. -ntcu to live, 
also a continuative ; Mukne tcira always, 

24. big: mi, mik, mix (V. also old) Cf. Yok. nioxodo 
old; Muk. muk adult; Tak. mahai. 

25. little: mutsk. Cf. Muk. pusnt. 

26. round: pik Cf. Miw. pol; Yok. huk-\ Wint. 

27. Numbers:' 









wite, suii 
























usit, utit 




mawik [ma 

hand) macoka 






























watsu, paki 








pampa-sem (two hands) 






28. more: maak Cf. Miw. manik. 

29. no, not: ka-, kedi. Cf. Miw. ke, ken, ket; Yok. 
k'a>nu; Muk. ekwe, kwe. 

30. yes, positive, good: oy Cf. Maid, he; Miw. hi; 
Yok. houu; Muk, he; Wint. 0. 

31. thus: sa Cf. Miw. saka like, as; Muk. sata. 

32. if: pen. Cf. Muk. pini. 

(2) animals: 

33. dog: uk. Cf. Maid, suku; Miw. tcuku; Yok. 
tce-jej; Muk. hutce; Wint. cukut; Tak. 'tsixi. 

34. bird: hon Cf. Maid, hu to fly, horn nest; Miw. 
hotju egg; Yok. hoy egg. 

35. eagle: wits Cf. Miw. wipayak; Yok. witisnl, 
eagle, witc, condor. 

36. snake: tsan Cf. Muk. lisana. 

37. frog: ink Cf. Miw. wataksay, kotola; Wint, 

38. fish; ak Cf. Maid. mako. 

39. flea; picg Cf. Muk. wipsur. 

^ Note to "two" Chin, mokct., Mixe matsk cannot be 
directly compared to Mixe us- in us-tug "7" (= 2 -|- 5) 
and to Miwok otto and Mukne utxi-n but may be related 
if we assume m- as archaic durative intran.sitive prefix, 
such as we have far more transparently in Hokan and 
Muskogian. — E. S. 

40. louse: ag Cf. Muk. rax. 

41. bedbug: /// Cf. Maid, dt^ louse. 

(3) plants: 

42. wood, tree: kep, kip Cf. Miw. kapiim, bark; 
Yok. yapkin ; Muk. xipur, tree. 

43. brush, bush, forest: yuk Cf. Miw. yomyum, 
brush; Yok. yawud; Maid, yo, flower. 

44.. tobacco: xuyge Cf. Miw. hutia. 

45. moss: sarjk Cf. Muk. sasuk. 

46. fruit: uts Cf. Muk. owos. 

47. flower; pux Cf. Maid, pu, to blossom. 

(4) man: 

48. person, people: hay Cf. Maid, maidik; Miw. 
miw; Muk. muwe; Coos md; Yok. mai. 

49. man: yayek, yadyek Cf. Miw. naya-yak, Yok. 
muk-yamk. (also Coos -iyag plural suffix in 
terms of relationship). 

50. woman ; tosdyek Cf . Miw. occayak, Yok. 

51. old man: mix (also "big") Cf. Maid, muk, big; 
Yok. moxodo; Muk. muk, adult. 

52. foe, enemy: was Cf. Mukne wayas. 

Seven: Penutian Languages 



53. youth, boy: nakn Cf. Yokuts notco. 

54. father: tyedj, tat Cf. Maid. te\ Yok. -atet; Muk. 
ete, grandfather; Win. tata. 

35. mother: tagh, nana Cf. Maid, na, ne; Miw. ita; 
Yok. -ajaj; Mukne ti/a, parents: Tak. ni-, -hin; 
Kal. 7ii; Low. Chinook -wafl; Coos e'n-dtc {-ate 
kinship ending); Coos nik/a "mother!" (voc); 
Tsim. (Nass R. dialect) nc-^i "my mother" 
(-H "my"), nd'd (voc.). 

56. grandfather: ap Cf. Maid, opa, pa; Miw. apa, 
father; Mukne apa, father; Win. apa; Yok. 
bapa, bap', father's mother; Tsim. (Nass 
R. dial.) no-beb-i', my mother's brother, bip' 

57. elder brother: ats Cf. Miw. ata, tatci; Mukne 

58. yoimger brother: its Cf. Maid, tu, t'i; Miw. 
tcale; Mukne tare; Miw. iDi'' (Sapir MS.). 

59. younger sister: vis Cf. Wint. utcii; Miw. Wa'* 
(Sapir MS.). 

60. uncle: haym Cf. Maid, yam, kam; Miw. kaw\ 
Yok. kawa, komoy. 

61. elder sister: isyo Cf. Maid, cti; Miw. De'De* 
(Sapir). 1 

62. aunt: tsugu Cf. Yok. guiha, nusus. 

63. child: tt»;g Cf. Miw. aijst, son; Mukne rwis, son. 

64. son: wza»; Cf. Mukne mos. 

65. daughter: ^e/, wyi's. 

06. diminutive or endearing suffix (suggested by 
Sapir) -s, -c, -j (e. g. wi'-y, daughter; Cf. ur)-g, 
chUd. Pen. stem perhaps *ariu, child; dimin. 
*a'yiu-si > Mixe tji-s, Miw. ajj-si, son, Mukne 
ini-s, son). With stems ending in -t, -d this old 
diminutive combines in Mixe to form -ts, -ic, 
-dj, hence: tyed-j, father (probably dim. form 
of tat) ; at-s, elder brother : Miw. ata ; it-s, 
younger brother: Maid, ti, Miw. iDi'^; ut-s, 
yoimger sister: Miw. Wa'\ This dimin. 
*-s! survives also in other Pen. dialects: Miw. 
ay-si, son; S. Coast Miwok (Barrett) taiyi-s, 
man; kuleyi-s, woman; hena-s, boy; oyi-s, old 
man; potci-s, old woman( ?) ; Mukne ini-s, mo-s, 
son; Wintun utcu, younger sister (diminuti- 
vized < *utu < *utaP); Tak. -xi, -x < *-si, 
*-s (in hap-xi, child, haap-x, one's children, 
cf . haap'-, child in other combinations) ; 
Chinookan (Wishram dialect) -c in -k!acu-c, 
paternal grandfather, -gak ! u-c, maternal grand- 
father, -k/i-c, paternal grandmother (these 
stems are etymologically related to respective 
reciprocals: -qcE-n, -gaka-n, -gia-n), further in 
wi-n-am-c, my father (stem -am) and wa-n-aq-c, 
my mother (stem -aq-) ; Coos -ca, endearing suffix 
(e. g. huumik-ca "dear old woman"; umd-ca- 

^ c- of Maidu cti and -sy- of Mixe tsyo may be 

tc, uma-c "grandmother": vocative uma. This 
Penutian *-si diminutive is characteristic, it 
would seem, as contrasted with its vmdoubtedly 
cognate Hokan *-tsi {*-'tst). — (E. 8.)^ 

67. I: es, -s, -es Cf. Maid. -5, -mus; Tak. -xi "me", 
from < *-si. 

68. I: n- Cf. Miw. kan; Yok. na; Muk. kan; Wint. 
nanu "my"; Coos n; I/Ow. Ump. na "1", -n 
"I, me"; Tak. -'n, -n (trans, subj.); Chin, -n- 
"I, me" (Wishram dial, also indep, «a.) 

69. thou: migs, m- Cf. Maid, mi, -mam, -mem; 
Miw. mi; Yok. ma; Mukne me; Wint. mi; Tak. 
ma; Chin, -m-, -mi-; Tsim. — n (thee); m-. 

70. he: yt Cf. Maid, -i; Miw. /. 

71. we: -m,, -em, -ma Cf. Miw. -m, -me; Yok. mai; 
Mixkne makse, fnak; Tak. -aw; Coos -ami; 
Tsim. -Em. 

72. ye: -t, -te, -ta Cf. Miw. -tos, -ton, -tok, -tc. 

73. which: wud Cf. Mukne watt "someone"; Coos 
wit; Low. Ump. watc. 


74. eye: win, is "to see" Cf. Maid, hin; Miw. c'in-ti, 
eye, c'iy, to see ; Yok. oil, to see ; Mukne xin ; 
Wint. wini, to see, ca, eye; Tak. al-xi-g-, -xi-k-, 
(Tak. -xi- < *-si-), also -xanw- ("to look") 
< *-san-w-.- 

75. mouth: aux Cf. Miw. awo; Yok. ata, ca; Mukne 

76. tongue: _va»e Cf. Maid, eni; Miw. He/??/, leiip; 
Yok. tadxat, palat "project tongue"; ^lukne 
lase; Wint. tahal; Tak. ela; Coos Ae//a ; Low. 
Ump. i^al. 

77. to drink: ug Cf. Miw. wcczt; Yok. uk; Mukne 
ukis; Tak. iigw-. 

78. to speak: mo^ Cf. Maid, mo-, ma-, action with 
mouth; Miw. mo-, answer; Mukne }>!ons. 

79. to answer: azoy Cf. Miw. has-, ask; Mukne 

80. word: kops Cf. Maid, ka, say; Miw. /fa-, say; 
Mukne ko, say. 

81. to call: mo Cf. Maid, wo; Miw. «'o; Low. Ump. 
waa, say, speak, tell; Chin, wawa, talk; Tak. 
yaway-, talk (dissimilated from *waway- ?). 

82. nose : hikb Cf. Maid, hikii; Miw. A«/j ; Mukne hus. 

^ Nominalizmg suffix -c = Mukne -s verbalizing 
suffix (e. g. yitc-s "language" <; rite "to speak"; 
= Tak. -{a)x infinitive suffix of intransitive verbs 
< *-(a)s; = Low. Ump. -s verbalizing suffix (waas 
"language" << waa "to speak"). 

^ In other words we have two old Penutian stems: 
*(i)sa or *(i)say (> Miw. ciy-, Wint. ca, Mixe is, Tak. 
-xi-) and *{i)san, *(i)sat > Miw. cln-ti, Yok. cil, Mukne 
xin. Maid, hin, Tak. -xan-w-, Tsim, saal "to notice", 
perhaps also Coos xil-, "to look around". — E. S. 


VI American Indian Languages 2 


83. ear: tatsk Cf. Miw. tokoc; Yok. luk; Mukne 

64. hand: ki Cf. Maid. ka-. with the hand; Miw. 

eku. uka « *esu. Miw. k = Mukne s. regularly. 

— E. S.); Yok. ko, to hit with hand; Muk. 

isu; Win. kupum. 

85. left hand: ava Cf. Mukne an-si. 

86. to take: mats Cf. Mukne nm;tr, get; Low. Urap. 
mate-, to lay; Tak. mats lag-, put, place; (Pen. 
base *mats- > Ta^. mas-g-); Chin. (Wishr. 
dial.) -A--n»a, to lay, -a-l-ima-lx, to put into 
water; Tsim. mag-, to put dowTi. ; Maidu me, to 

87. to give: mo, to have: mid Cf. Maid, me; Miw. 
a>«-; Muk. ;trMm, cumi. 

88. to do, make, work: dun Cf. Miw. taiiha; Yok. 
/nc; Muk. /w« finish. 

89. to put, place: -wits Cf. Maid. wi-. pull, press; 
Miw. wtk-; Tak. -ttyhA, to spread out. 

go. foot, leg: t'ik Cf. Maid. /^A, jump; Miw. tek, 
kick, Aa/f, foot, tirji, hip; Mukne /^A^, walk on 

01. knee: koc, Cf. Tak. gouk; Low. Ump. qoq; 
Chin- -q!oxl. 

92. to go: ois Cf. Maid, u-; Miw. «-. 

93. to walk: wag Cf. Miw. iwi-n, weeta; Yok. wadix, 
pass by; Mukne twa//, come, ze/a/^, go; Tak. wi-, 
to go about. 

94. to go: y- Cf. Maid, i-ye; Miw. yoc-; Chin, -ya, 
-i-\ Tsim. yaa; Tak. yaw-; (Cf. Hokan *iya, 
to go). 

95. action with the foot, walk, trail, etc.: pa- 
(pason, to follow, paboy, to chase, pesamb, to 
go out) Cf. Maid, be-, with the foot, puiyi, hunt. 

track; Miw. puy, move, depart; Mukne paya, 
run, payta, himt. 

96. to stay : wi (also in : tti^ic, a bed)^ Cf. Tak. teiay-, 
to sleep ; Kal. ivai, to sleep ; Tsim. woq, to sleep ; 
Coos, haya-ti, to lie down (Coos A often comes 
from iv) ; Yok. woi-, to sleep. 

97. to sit: uy Cf. Maid, in-; Yok. uy, to lean; 
Mukne one. 

98. bone: pak Cf. Maid, -/'o^, -puk; Wint. /jaA. 

99. skin: po Cf. Maid, ^tt, posa. 

100. neck; yok-yi Cf. Yok. ogun; Tak. gwen; Coos 

kwin-ts; Low. Ump. kwininicu. 
loi. rib: /la/c Cf. Maid, /cz ; Miw. Atf/c, bone; Mukne 

/ca/c, bone. 

102. phallus: tsii-k Cf. Maid, utcu, piss; Yok. 
icwyo ; Mukne tcoxo, pudenda. 

103. testicles: istca Cf. Miw. caii-tal; Mukne tcatia. 

104. anus: pxut Cf. Miw. poii, navel; Yok. poiodo, 
intestines; Mukne puttus, belly; Wint. pot. 

105. to urinate: tatsp Cf. Miw. ota, otso; Mukne 
tcala (Cf. Wint. tcunu-s; Tak. xan. — E. S.). 

106. heart, belly, thought: hot Maid. ho7i; Yok. hut, 
hon, to know, horj, heart; Mukne xtitu, belly; 
Coos haw-, to imagine, think; Low. Ump. hai, 
opinion; Tak. heivhaw-, to think. ^ 

107. fear : tsok Cf . Miw. ceki ; Yok. dote : Mukne susu. 

108. love, like: tsok Cf. Maid, -sak-; Yok. icik; 
Mukne otciko. 

^ < "Sleeping instrument". See -i nominalizing 
suffix. — E. S. 

' Maidu ho-n and Yokuts ho-r] probably nouns in -n, 
cf. "infinitive" and noun-forming -n, from base *ho- 
< Pen. *haw-, "hau-; Mixe ho-t probably factitives in 
-t from same base. — E. vS 

Editorial Note 

Originally published in International Journal of American Linguistics 6, 
28-33 (1930). 

Ancestral Mixe-Zoque is now supposed by most authorities to have been the 
language of the Olmecs, the oldest of the archeologically known high cultures 
of Mesoamerica (Campbell and Kaufman 1976). 

Sapir's "Mexican Pcnutian" (Mixe-Zoque and Huave) was expanded by 
Whorf ( 1935) to include Mayan, Totonac, and Uto-Aztecan. Greenberg (1956) 
accepted the addition of Mayan and Totonac but excluded Uto-Aztecan. 

Seven: Penutian Languages 289 

Greenberg has recently (1987) reiterated his version of Mexican Penutian as 
part of a hemisphere-wide classificatory scheme bolstered by extensive lexical 
comparisons. Neither Whorfs nor Greenberg's claims have met with favorable 
reaction from specialists in Mesoamerican languages (see Campbell 1979: 964). 
The connection of Huave to Mixe-Zoque, proposed by Radin (1916) and 
accepted uncritically by Sapir, is now considered improbable. The most likely 
affiliation of Huave is with Otomanguean (Campbell 1979: 964). 

Edward SAPiRf and Morris Swadesh 

1. Introduction 

1.1. Date of manuscript 

1.2. Description 

1.3. Phonology 

1.4. Why withheld 

1.5. Symbols and arrangement 

2. Sapir's comparisons 

1. The Coos-Takelma-Penutian compari- 
sons presented here were made by Edward 
Sapir. Morris Swadesh has prepared them 
for pubhcation and has provided the intro- 
ductory information and discussion. 

1.1. Sapir's manuscript, which now forms 
part of the Franz Boas Collection of the 
American Philosophical Society, evidently 
dates back to about 1914, the publication 
year of Leo J. Frachtenberg's Coos, An 
Illustrative Sketch, Extract from Handbook 
of American Indian Languages (BAE-B 
40.2.297-429). It surely represents an im- 
portant part, though far from all, of the 
lexical evidence Sapir refers to in A Charac- 
teristic Penutian Form of Stem (UAL 
2.58-67), pubHshed in 192L In this article 
he concurs with Frachtenberg's suggestion 
of Penutian affinities of certain Oregon 
languages, stating: 

All this is very interesting to me, as it chimes 
with conclusions or hypotheses I had arrived at 
independently. On the appearance of Frachten- 
berg's Coos grammar it soon became clear to me 
that the morphological and lexical resemblances 
between Takelma and Coos were too numerous and 
fundamental to be explained away by accident 
or plausibly accounted for by borrowing. The 
appearance of Frachtenberg's Siuslaw material 
has only tended to confirm this impression, 
further, to make it perfectly obvious that Coos 
and Siuslaw, as Frachtenberg announces, are 
divergent representatives of a single linguistic 
stock. Meanwhile comparisons of Takelma, Coos, 
and Siuslaw with Dixon and Kroeber's Penutian 
group of California (Costanoan, Miwok, Yokuts, 
Wintun, and Maidu disclosed an astonishing num- 
ber of both lexical and morphological correspond- 

ences. ... In spite of our slight knowledge of 
most of the Californian languages involved, I 
succeeded in getting together what I believe to 
be a quite respectable mass of evidence tending to 
unite the southern languages with those of Oregon 
into a large and highly differentiated "stock". 

In the same article Sapir refers to evidence 
for still further Penutian relationships. It is 
possible that other manuscripts will yet be 
discovered containing this material. Much of 
it may have been in the form of marginal 
annotations made by Sapir in books and re- 
prints, like those compiled in Comparative 
Penutian Glosses by Edward Sapir, in the 
Boas Collection. 

1.2. The present manuscript consists of 
13 pages 8x11, handwritten in Sapir's usual 
small writing. There are 152 numbered 
groups of cognates, with columns for Coos, 
Takelma, Penutian. There is no other title. 
At least two shades of ink are distinguishable, 
a dark blue for all but set 152, and a lighter 
blue for the latter and for additions to the 
first 151 sets. Since the additions include 
items in the Coos column marked "Mil." 
(that is Miluk dialect of Coos) and four 
annotations of 'see L.U.' (that is, Lower 
Umpqua or Siuslaw), one may infer that the 
main list was made before and the added 
notes after the appearance of Frachtenberg's 
Coos and Siuslawan sketches. The original 
Coos items must then have been based on 
Frachtenberg's Coos Texts, which appeared 
in 1913. This assumption is corroborated by 
the fact that Sapir's copy of Coos Texts has 
marginal glosses of Takelma comparisons in 
what seems to be the same ink as the manu- 
script list, and there are 'L.U.' annotations 
in the lighter ink. All the foregoing observ^a- 
tions on the ink used in the manuscript were 
made by inspection only. The identifications 
seem for the most part quite obvious, 
except for those parts of the manuscript 
which have been most affected by light. 



VI American Indian Languages 2 


Besides the entry numbers the manuscript 
has a set of check marks varying with a 
diagonal criss-cross at a few points. Sapir 
used such marks to register a later critical 
reexamination of his work. That is, at some 
time after he had made the list, he must 
have gone through it again, with constant 
reference to the sources, checking the entries 
which seemed plausible and putting the 
criss-cross by the ones that on reflection 
seem improbable. During this process he 
would often insert additions to the original 
material. In the present manuscript, we find 
that the criss-cross sometimes confirms and 
sometimes overrides an earlier hesitancy re- 
flected by parenthesized question mark or 
annotations (e.g. in 57 after the Yokuts 
form: "perhaps better to Tak. -ha-n, 
no. 55"). Three asterisks between entries 55 
and 56 may be a pause mark, showing where 
Sapir had temporarily stopped in his 
check-up. A short horizontal line between 86 
and 87 is evidently another pause mark. 

1.3. The manuscript does not include any 
full analysis of phonological relationships. 
However, one finds phonological notes at 
various points. Some of these merit general 
comment. The occurrence of Yokuts d for 
earlier 1 is indicated or implied in a number 
of places (8, 9, 19, 50, 54, 71, 72, 74, 78, 79, 
83, 84, 87, 94, 95); Newman's work' has 
since shown that this is a regular dialectal 
development in Wikchamni, while the other 
dialects preserve the lateral. In two cases (10, 
23) we have Wintun X, Takelma s recon- 
structed to t; item 7, disregarding the first 
Coos form (Hanis dialect), suggests that 
I or X might be a better reconstruction. The 
development of Takelma x from earlier s is 
posited in 11, 35, 41, 61, 75, 99, 121, 123. The 
origin of Takelma s from earlier c is suggested 
in 129. This fits in with the fact that Takelma 
has no c even though it has s and the com- 
bination s?, the latter being pronounced as a 
glottalized affricate. Sapir's idea was evi- 
dently that pre-Takelma s changed to x, 

' See Stanley S. Newman, Yokuts, Viking Fund 
Publications in Anthropology 2 (1944), p. 16. 

coinciding with original x, and that new 
instances of s developed out of older c and ch. 
The evidence in general seems to bear out 
Sapir's theory, but the proof is complicated 
by the presence of more than one s-sound in 
Yokuts, Coos and perhaps other Penutian 
languages, and by the existence of two 
x-sounds in Coos and Siuslaw. Robert Shafer 
has attempted a clarification of Penutian k 
and x-sounds,^ but he does not fully suc- 
ceed; his material is insufficient in quantity 
and does not properly distinguish between 
probable cognates and similarities which 
might have resulted from borrowing. How- 
ever, the problem may be expected to yield 
to patient research in the future. 

Where Sapir gives two or more spellings 
for a word, they may be variant phonetic 
recordings of one and the same form, dialect 
variants (as in Yokuts), or functional vari- 
ants (e.g. verb stem and aorist stem in 
Takelma) . The first kind of variation is here 
eliminated to the extent possible by using 
phonemic spellings. The other two types 
have been retained because of their value in 
tracing phonological relationships. 

For Yokuts it was desirable and easily 
possible, by consulting Newman's mono- 
graph and with additional data kindly pro- 
vided by Newman, to specify the dialect of 
most of the cited forms. Where Sapir gives a 
single form and it coincides with the Yawel- 
mani dialect, it has been labeled Yawelmani 
(Yy). Where Sapir's form is from another 
dialect or includes more than one dialect, the 
Yawelmani has been added after Sapir's 
Yokuts entry. 

1.4. Why did Sapir not publish his Coos- 
Takelma-Penutian evidence?' The answer is 
perhaps suggested by the following remarks 
(UAL 2.59, 1921): 

Unfamiliarity with Alsea (Yakonan) and 
Kalapuya made it impossible for me to follow 

* Penutian, UAL, 13.205-19 (1947). 

' Some of Sapir's comparisons and structural 
observations are included in L. S. Freelander, 
The Relationship of Mixe to the Penutian Family, 
UAL 6.28-33 (1930). 

Seven: Penutian Languages 



the chain of evidence geographically. Neverthe- 
less, Chinookan points of contact soon manifested 
themselves too persistently to be brushed aside. 
After hesitating for a long while to take up seri- 
ously the possibility of affiliating Chinook, one 
of the most isolated and morphologically 
specialized languages in America, with the 
Penutian languages of Oregon, I now find myself 
forced by the evidence to admit such an affiliation 
as not only possible but decidedly probable. . . 

The greatest surprise was still awaiting me. 
Tsimshian . . . Should it be possible to demon- 
strate (and I am fairly sanguine that it can be 
demonstrated) that Tsimshian is a detached 
northern offshoot of Penutian, we would be com- 
pelled to face a most interesting fact in linguistic 
diflferentiation and in the distribution of American 

These comments show that Sapir con- 
sidered the Coos-Takelma-Penutian rela- 
tionship to be part of a larger grouping which 
he felt could be demonstrated. Evidently he 
was awaiting the appearance of Kalapuya 
data and was hoping to work out in detail 
the relationship of Chinookan and Tsim- 
shian. He felt sufficiently sure of himself to 
publicly state his theory, but perhaps he 
thought it well to wait with the specific 
lexical data until he could present the evi- 
dence for the fuller picture. The case would 
then be comparable to that of Sapir's theory 
relating Na-dene and Sino-Tibetan.'* When 
students asked Sapir why he did not publish 
his evidence for this connection, he said it 
might be better to hold off until he had 
worked out and published the full details of 
Athapaskan phonology. However, in retro- 
spect, one can only regret that this material 
was not published long ago. 

1.5. Sapir's 1914 orthography is here 
modernized and spellings are given in 
phonemicized form to the extent possible. In 
the case of Takelma, we take the aspirates 
to be clusters of stop with h and the glot- 
talized to be clusters with glottal stop; 
Takelma accent is here reduced, with reason- 
able justification inferred from Sapir's de- 

* See discussion (by Morris Swadesh) of 
Athapaskan and Sino-Tibetan by Robert Shafer, 
UAL 18.178-81 (1952). 

scription,^ to two types, a normal accent (') 
with tonal variations according to position 
in the word, and a rising or sustained long 
accent (') occurring on long vowels or on 
vowel-sonant groups in contrast to falling 
tone for the normal accent in this position. 

Points of phonetic usage: x is for back 
velar spirant, c is for sibilant affricate (ts), 
H for shibilant affricate, double letters are 
used for long vowels, r is for voiced velar 
spirant (in Coos). The languages are abbrevi- 
ated by using the first letter for Takelma, 
Coos, Siuslawan, Yokuts, Wintun, the first 
two letters for Chinook, Miwok, Maidu, 
Mukne (Costanoan — Freelander's term). Yy 
is used for the Yawelmani dialect of Yokuts. 

For convenience in printing, Sapir's 
columnar arrangement of the manuscript 
has been dropped. Sapir's diagonal cross for 
improbable comparisons is retained, his 
check marks are omitted. His parenthetic 
question marks, also indicating doubt, are 
retained but are placed after rather than 
before the entry. Other than in such matters 
of arrangement, punctuation and phonetic 
symbols, the materials are given as fisted by 

2. 1. T kuu-x-m/e (?) (see S); Mi kule, 
kulei wife, woman. Ma kiila woman. 2. C 
k'^aaxaX how; T gaP bow; Mi kono, W kul-sak 
how. 3. C walwal knife; T wiih-A;m/e; Mi 
hulaya knife. 4. tuu good (aor. tuu\\^iu-k- to 
be good); Mi towi-s good. ^5. C asoo again, 
asoow-is second; Mi os(s)a, otta two. 6. C 
e'^n-ee^ mother, niika mother!; T hin-, ni- 
mother; W nen-^m, nenin. Ma ne, Mu ana 
mother. 7. C huuI nose, Miluk fin-nuuq; T 
sin-, sinii-x- nose; W Xinik, suno, Y tiiDiik 
nose, Yy tinik. 8. C helta tongue; T ela 
tongue; Ma eni, W tahal. Mi le-mtep, le-tip, 
Mu las, lase, lasa, Y talxat, '''aladis, m-ada-t 
tongue, Yy talxat tongue, "^aalat- to lick. 
9. C kxla leg; T k«el- leg; W koli leg 
(<*k*ili ?), Mu kolo, koro foot. Mi kollo 

* Takelma, Handbook of American Indian Lan- 
guages, 40.2.17-18 (1922). 


VI American Indian Languages 2 


foot, Y kada-Sa lower leg, Yy kalassa'i*. 10. C 
Miluk Hil-li leg; T sal- foot (<*tal-?); W 
Xel-ma (<*tel-?). 

11. T xuu''*n night (for x:s, cf. xi water 
below); W sinol night. 12. T b§e sun, day. 
poo now, today; Ma poko sun, W po day, 

Y 'i'opo-do sun, "i^op-di day, Yy "^op moon. 
13. T n6ox rain-, W luha rain. 14. T p'^ii^re; 
W po fire. 15. T s'i'aaw lake; W nahi lake. 
16. T som mountain (<*tom?); W toX, ho\ 
mountain. 17. T tan rock; W Son rock. 18. C 
Sximl black bear; T x§,mk grizzly bear; W 
Silai bear, grizzly bear. 19. T -s'i'il red; W 
tulu-ka, tede-kit red. 20. C bni ^ood; W 
laiyok good. 

21. T ma thou; W mi thou, [see 47]. 22. T 
pam- up, pam-ls sky; W panti-be up, panti 
on, upon. 23. tqanX- (?) [to hit, to strike with 
instrument] ; T saansan-, sana- to fight with, 
spear; W Mia to shoot (<*tana?). 24. T -t 
present participle, adjective suffix; W -u-t 
future participle, -t adjective suffix with sub- 
jective noun, Ma -do- present participle. 

25. T -khi?, -ki? if; W -kila conditional. 

26. C -uu interrogative suffix; W -wi inter- 
rogative. 27. T -de?, the? / (intr.), future 
-tee, -thee; W -da /, Mi -t, -te /, me, my. 

28. C iin not; T ?anii? not; W eleu 
(<*ene-u?), Y am, Yy ?ohom, Mi (E.) ela. 

29. T yaan-, yana- to go;W hara to go. 30. T 
p^nx hunger; W bira to be hungry. 

31. C tk"*- to kick; T tkuuntkan-, tkuun- 
to kick; W qow(a) to kick. 32. -eenii reciprocal 
verb suffix, -inii (relatives) to one another; T 
-an- reciprocal verb suffix; W -pu-ra re- 
ciprocal verb suffix (?). 33. T ?al- to (?), 
ka-nan in; W el in, into. 34. T pay- out of 
house (vb. prefix); W pat out. 35. C yipsn 
three (?) (see S [Siinx]); T xipini three; Ma 
sapwi three, Yy soopin, Mu kapan, kaphan, 
kapxan three. 36. C SuuX- to set fire to, m\- to 
burn, Hweel^re; T -s^iiluii-, -s^iil- to set fire 
to, to catch fire; Ma sa, Y ?oSit, Yy ?osit, 
Mu Soto fire. 37. C mee human being; Mu 
ama person, Ma mai-dii Indian, mai-ki boy, 

Y may person. 38. T yap?a person; W 
yapaitu person. Ma yepi man, husband. 39. T 

tak-ax- head; Mu taxa-S head. 40. C x'^iHux'^ ^ 
head (?); T ?ulu-k?-i- hair of head; Mu ut, ^| 
uri, uli hair, Yy ?otow head. Ma ono head. 

41. C '))[ to drink; T xl water; Mu si water. 
Ma sewi river. Mi kik water. 42. C Xipi^ 
arrow; Mu tep-s, temo-x, nemo arrow. 43. C 
Iqawe to die; T lohoy-, loho- to die; Mu 
laku- dead. 44. C cimsimt to sleep (plur., 
redupl. <*tim-t or reformed from *ir)tim > 
*t-ir)tim); Mu eten, enen, etin-i to sleep. Mi 
eM, Y ?er)tim, Yy ?entam a sleep-inducing 
substance. (Is Esselen acin- after all merely 
borrowed from Costanoan?). 45. T ti inter- 
rogative enchtic; Mi mau-ti whof hi-ti 
whatf Ma -de interrogative, Y ti enclitic, 
Mu -sa, -s (< *-ti). 46. C n- /; T -?n I—him, 
-n (fut.) I— him; Ma ni /, W ni, Yy na?. 
47. T ma thou; Mu me, men, mene. Mi mi, 
Yy ma?, W mi, Ma mi thou. 48. C Xtaa 
earth (?) (see S [Xa?ay]); T tkaa earth; Ma 
kau ground. 49. Ch -k-, -x- my; T ki /, -te-k, 
-t-k my, koom we; Mu ka, kan, kana /, Mi 
kanni 7, -ka, -ka-n my. 50. C -en numeral 
suffix so and so many times; T -n, -(t)an j 
adverbial numeral suffix; Mu -na adverbial ' 
numeral suffix, Ma -nini, Y -d, -1, Yy -? . . . il'. 

51. C puux"^- to spout; T phoophaw-, 
phoow- to blow; Ma bo to blow. 52. T mena 
bear; Y moloy bear. Ma mode brown bear. 
53. T maan- to count; Ma -mak- to know, 
count, measure. 54. C -enii to do, to make 
something; T -(a)n(a)- causative verb suffix; 
Mi -ne causative, Y -la, -da causative (for 
1 < n, cf. no. 50), Yy -aala. 55. C -n, -ni-, 
-eene plural element found in few nouns (e.g. 
mee-n human beings, ceey-eene stnall pi., 
qe-ni-xet short pi.) — its irregularity and 
rarity show it to be survival from remote 
past; T -ha-n, -an noun plural (not very 
frequent); Y -n, -in plural of pronouns, -in 
numeral collective, W -1 dual of pronouns, 
-li-t dual obj., -le-t plur. obj., -li-n dual pos., 
-le-n plur. pos., Ma -nono many (?). 

56. C -e, -ii petrified noun plural (e.g. 
tummeeX-e old men, timil-ii men, keneeyes-e | 
hunchbacks); -a petrified noun plural (e.g. f 
aXimaq-a big pi.); Y -i, -a noun plural (for 

Seven: Penutian Languages 



Y ablaut, e.g. 'i'onmid: ?onemad-i, cf. C 
toomiX: tummeeX-e, knes: keneeyes-e), Yy 
-i, -a subject case with plural stems. ^57, C 
hinna so and so many each; Y -hin collective 
(e.g. yapkan-hin many trees) (perhaps better 
to T -ha-n, no. 55). 58. T ^uuk'^- to drink; 
Yy "i^ugun- to drink, Mu ukis, Mi uhu. 59. T 
ka that (indef.); Y ka that (vis.), Mi i-ka that. 
60. C t that there, ta so, such, te- article with 
pos. pronoun; T ii-ta (ka) that yonder; Y ta 
that (inv.), Yy ta that. 

61. C -is noun suffix (often suffixed to 
verb stems: Xeey-is language Xee-c he spoke), 
-s, -t-s, -en-is nouns of quaUty; T -(a)x 
infinitive (often used as verbal noun: S 
?ipn-ax speech); Mu -s noun-forming from 
verb stems (e.g. ni^i-s language < nm to 
speak), W -s verbal noun suf., Mi -s noun 
ending (?), Y -0$ noun suf. 62. C -ne it is (e.g. 
n-ne it is I); T -(a)n-, -in- petrified intransi- 
tive suffix; Y -in- intrans. (?), Mi -ne, -fie in- 
transitive verbifying. 63. C ii-ta emphatic 
particle; ii-ta(ka) that; Mi i-, i-ni, i-sa, i-mo, 
i-ka, i-ti that. 64. T -n noun ending (e.g. 
yiwi-n speech < yiw to speak); Ma -n in- 
finitive ending. 65. T -a'i'n person, people of; 
Yy -? . . . inin people of. 66. C axaax maternal 
uncle.; T has- maternal uncle; Y 'i*aga$ 
mother's brother, Yy "i^aagas. 67. C ^inl^in 
eyebrow; Y nimejid eyebrows and lashes. 
68. C cneex beard; Y jamo^ beard, Yy 
daamut-. 69. C k'*'in-c throat, neck, k'^'in- to 
swallow (Coos kw: T k^, cf. k^'ees-is wind, T 
k^alt); T k*en- neck; Yy ?oogun, Ma kuyi 
neck. ^70. Xpene wings, feathers; Y ^a'i'ada 
fur feathers (?). 

71. T telk-an buttocks; Y teda anus, Yy 
iooto*^. 72. T nihwlk'' black bear; Y du?uxun 
bear (<*lu?uxu-). 73. teec lice (see S); T 
fi^eld louse (?); Y tehet head louse, Yy ti'i^it. 

74. T laap leaf; Y dapdap leaf (<*lap-). 

75. T bix-dl moon (?) (x < 5); Y 'i'upiS moon 
(cf . sun), Yy "^op. ^76. C t(^een soot; Y neheo 
fog. 77. T tk^'d thunder; Y taka'i'a thunder. 

78. C lahX-is earth; Y diioiit earth (< *luout). 

79. T hok--ai, -h6k-h-al hole (<*-Sok-al); 

Y Sogod hole (<*Sogol), Yy sogol. 80. T 
tkism- green; Y nii'i'imat green. 

81. T -tkSm black; Y qimgutan black. 
82. T mii now, then (particle) ; Yy 'i'ama'i* and 
then, then. 83. C liixUi to pass out, to pass by, 
landing place; Y day (< *lay), lay to step, to 
kick. 84. C al-$ toy; T loo-l-, loo- to play, 
loo-sf toy; Y do to play (<*lo). 85. C tooh-, 
tous- to hit, strike; T t^omom-, toom- to hit, 
kill; Y do battle, (?), do5 to beat, overcome 
(?), Ma -tul- to break flat thing (?), tup-, tus- 
to break (cf. 117). 86. C Xou- to eat; Y duy to 
eat (<*\uy). 87. T -(a)l-, -Iha- continuative, 
frequentative; Y -ad (<*-al) continuative, 
-li frequentative. 88. C maux- to chew (up) 
(cf. Ch); Y mok to swallow, Yy meeki-. 
^89. C nix(t)- to touch; Yy niti- to squeeze 
with hand. 90. C tou- to fall; Y ?ot- to fall. 

^91. C t^- to come in; Y tax- to come. 92. C 
$anx- to shake (see S); T s'i^elel-, s?eel- to 
rattle (?) ; Y ta^it to shake. 93. C ^au- to come 
apart, to pull apart; Yy tatay- to break, Y taw 
to overcome. 94. T helel-, heel- to sing; Y 
'i'uduk-, Yy ^ilik- to sing. 95. C welex- to 
stoop, to lie down; Y wodo to dodge ( < *wolo), 
Yy wolooyee-. 96. C w^kin- to touch with 
stick; Yy woto- to hit with a stick. 97. C 
yak"-, yak'^-t- to pick, to gather (see S); Y 
yam- to gather seeds, yitw- to gather. 98. T 
-koyo-k*^-, -koy-k*!*- to touch, nudge; Yy koyo- 
to butt. 99. C xil- to look around (?); T 
-xanaw- to look (out) (?); Y Sil'i- to see, Yy 
sil'i-. ^100. C xint- to go fast, to run (?); Y 
Silit- to jump, Yy siUt-. 

101. T wayaan-, way- to sleep; Y woy- to 
sleep, Yy woo'^uy-. 102. T mahay large; Y 
met, mat, may- large, Yy mayaahay. 103. T 
khay- woman, girl; Yy gaa'i'ina woman, girl. 
^104. T tola hollow tree; Y ton, toga-S digger 
pine. 105. C k6s shell used for ornament; Y 
gu^i^ beads. 106. T s^fxi dog; Y neSe^, ^exa, 
Ma sii, Mu huhu dog. 107. T hSiiy cloud; 
Ma yaa cloud (?). 108. T nf, nii- teats; Ma 
mini nipples (dissimilated < *nini?). 109. T 
-k?otk'i*at-, -k'i*oot- to break in two; Ma -kot-, 
-kut- to divide. 1 10. T kele-k, kel-k- to drill for 
fire; Ma -kel- to perforate. 

111. C qal- to dig (?); T k'^olol-, kool- to 
dig; Ma -kol- to bore (?). 112. T kala-p-, 
kal-p- to twist (thread) by rolling ; Ma -kol- to 


VI American Indian Languages 2 


roll. 113. T k'^ataay-, k'^'aat- to pick, pluck; 
Ma -ket- to graze. 114. T lopap-, loop- to 
pound (acorns, seeds); Ma -lop- to move with 
friction (?). ^115. C six- to shake off; Ma -sil- 
to shake (?) (see no. 92 [in different ink, 
showing that comparison with 92 was prob- 
ably made after the original entry and after 
the entry had been marked with criss- 

^116. T tama-k?-, tam-k*^- to choke; Ma 
-tap- to squeeze (?). 117. T -tk^eltk^al-, 
tk'^eel- to break in two; Ma -tala- to crush (?). 
118. T t?alal-, taal- to crack; Ma -tala- to 
crush (?). 119. C -aami /, we 2, or we-thee, 
you 2, or you; T -(i)k-am we (fut. subj. intr.), 
-(a)nak-am we (fut. subj. trans.), -am us, 
-t-am our; Mi -ma(si) our, -m, -me us, me 
(subj. 1 [sic]). 120. T -nk he (fut. trans, 
subj.); Mi -k, -ko him, he (subj. 1, including 
future), W -k he (past). 

121 . T -xi me, I (subj . of passive) ( < *-si) ; 
Ma -s(i) /, Mi mu-su, mo-s, mu-$u, $i-ma 
I — thee. 122. C -aa'^is thou, ye 2, or ye — me, 
u^ 2, or us; W -s-ka, -s-ke-n thou, ye 2, ye 
(-ka, -ke- as in Ma -nka-no thou, ye 2, ye, 
-nka-s we 2, -nke-s we?). Mi -s thou, -tok-su 
ye (subj. 3: present and perfect). 123. T -x 
non-agentive (quasi-passive or reflexive), 
-al-x intransitive (-x < *-s) ; Ma -us reflexive, 
Mu -s-tap, -s-tapse impersonal, passive (?), 
Y -wi-5, -wi-s reflexive. Mi -si passive. 124. C 
C xeen-is sick; T xil-am sick (< *xin-an) ; Mi 
hali sick. 125. T -molo'i'mal-, mol^mal- to 
turn things over, .stir food in basket-bucket; Mi 
mole to spill. 

126. T phili-phal-, philphal- to squash 
(insects), to whip (children); Mi pilapa to 
pinch (?). 127. T -hit, -thit plural of ad- 
jectives; Mi -ti plural of adjectives and 
verbs, W -te subjective pronoun plur. 128. T 
-k inferential past; Mi -ke, -ka past. 129. T 
-sii, -saa agentive (-s- < *-c-); Y -i^, -ic 
agent, Yy -ic. 130. C -a-'^eiwat causative 
frequentative (-eiwat frequentative), -iya-t 
causative, -e-'^et (-a-'i*at) causative passive 
(-et, -at passive); Y -i, -u, -a causative. 

131. T -(a)pa'^ let us'.; Ma -p6, -pe, -pe'^'e 
let us I ^132. C -qm, -xm to be in a condition, 
to be in the act of; Mi -imi continuative (?). 
133. C -H instrumental noun suffix; Mu -cu 
comitative (?). 134. C -et (-at) past passive; 
Y -t passive, -n-it future passive. 135. -iiyas 
plur. of nouns of relationship, -iiye, -eeye 
adjective plur.; Mi -ya noun plur. 136. C han 
about to, hanX shall, will; Y hi future particle. 
137. C k^'a it seems, as if, hak'*'a-i as if, kind 
of; Y akam perhaps, it seems. 138. C hi'i 
particle indicating surprise; T his almost, 
trying to but in vain; Y hiqa perhaps. 139. C 
hei emphasizing particle, hi enclitic em- 
phasizer indeed; Ma -hehe only, just. 140. C 
qa-1 down, below, under (<qa -f- at, qa- is 
local adverbial prefix), qa-lin from under 
(<qa- -f- ahn); Y 'i'adid down, low, below 

141. C yuu very, very much; Yy yow ami, 
also, again. 142. T -ta*^ subordinating suffix 
of intransitive aorist; Ma -we-te after having 
(for -we- cf. -wea and -weu temporal 
suffixes), -ce-te when, while, -ya-tan past 
participle (for -ya- cf. -yak temporal suffix). 
143. T -na*^ subordinating suffix of inferen- 
tial past and transitive aorist; Ma -wono 
past participle (-wo- assimilated to -no, 
<-we- as above). 144. T -mai* subordinating 
suffix of aorist passives, -tat from, to, in — 
direction, -t locative (occurs in very few iso- 
lated adverbs: k'^cn-t in back, behiml < 
k^en- neck, nape, back), -(a)ta local adverbial 
suffix; Ma -di in, on, at (?), Mi -ta for, Mu 
-ta, -ta-k locative, -ta-5 in, on, at. 146. T 
-k'a- comitative verb suffix; M -ko having. 
^147. C witin blood; Mi weteti, wetete red. 
^148. C -ume where . . . is; Mi -m, -mo loca- 
tive, Mu -me with, at the house of, -m, -mo. 
149. C -a imperative with object of third 
person; Mu -i imperative with object of 
third person. 150. C -is, -inis place of; Y -lis 
habitual place of (-1- frequentative). 

151. T -(a)n aorist passive; Y -han pas- 
sive in dependent clauses. 152. See S [-uus 
locative]; Y -w locative. 

Seven: Penutian Languages 297 

Editorial Note 

Originally published in InternationalJournal of American Linguistics 19, 
132-137 (1953). 

Comparative Penutian Glosses 

Edited by Victor Golla 

Based on "Comparative Penutian Glosses of Sapir," 
edited by Morris Swadesh (1964) 


It was Sapir's custom to enter linguistic notes, both comparative and descrip- 
tive, in the margins of his books and offprints, especially dictionaries and gram- 
mars of American Indian languages. In this way, he frequently began various 
researches which he pursued throughout his lifetime. One of the most impor- 
tant of these was the gathering of comparative data substantiating his theory of 
a Penutian stock enlarged to include languages of Oregon, the Plateau, British 
Columbia, and Mexico. The broad outline of this theory was first presented in 
the introductory paragraphs of "A Characteristic Penutian Form of Stem" 
(Sapir 1921b, this volume), but with no comparative lexical data. That Sapir had 
what was to him a satisfying amount of such data is clear from a few published 
allusions (see especially Sapir 1920c: 265, note 1) and miscellaneous citations in 
his letters to Kroeber (see Golla 1984: 201-204, 242-244, 314-315) and other 
colleagues. Only a few fragments of Sapir's Penutian data reached print during 
his lifetime, the most substantial being the comparisons incorporated into L. S. 
Freeland's "The Relationship of Mixe to the Penutian Family" (Freeland 1930, 
this volume). Unfortunately, most of Sapir's comparative Penutian working 
files seem not to have survived. What these may have consisted of is hinted at in 
a passage in a letter from Sapir to Robert Lowie, dated July 6, 1918: 

Just at the moment I am carding some of my Penutian-Takelma-Coos-Siuslaw-Chinookan- 
Tsimshian correspondences. It is technical work, of course, but quite interesting, as many 
lines of historical linguistic research are opened up. Yes, my boy, Tsimshian. Not a bit 
isolated. Very specialized in development, but showing clear threads, in my humble and 
heterodox opinion, binding it to Oregonian "stocks" (Lowie 1965: 27). 

No collation of the sort Sapir refers to in this passage is known to exist among his 
surviving papers. A 14-page handwritten list containing 315 sets, later published by 
Morris Swadesh as ''Coos-Takelma-Penutian Comparisons" (Sapir and Swadesh 
1953, this volume), seems to have been compiled quite early in Sapirs Penutian 
research (probably in 1915); it includes no Chinookan, fsimshian. Plateau Penu- 
tian, or Mexican Penutian material. Under these circumstances, it is useful to 
publish as many as can be retrieved of the marginal jottings that were Sapir's raw 
material. Unsystematic and abbreviated as they arc, they nevertheless indicate the 
breadth of Sapir's knowledge of the data and the general dimensions of his theory 
of Penutian comparative phonology and grammar. 

300 VI American Indian Languages 2 

The comparisons in sections 1 through 5 below were culled by Morris Swadesh 
from several books left to him by Sapir. The work was done in 1953, and Swadesh s 
original manuscript is in the Boas Collection of the American Philosophical Soci- 
ety (ms. 2U[P1.4]). A decade later, Swadesh published these lists under the title 
"Comparative Penutian Glosses of Sapir" (Swadesh 1964). For the present work, 
Swadesh s lists were thoroughly checked against the publications annotated 
(although the copies with Sapir s notes could not be found), and against the sources 
of most of Sapir s comparative glosses. A number of errors in transcription have 
been silently corrected. Translations of the glosses — for the most part omitted by 
Sapir (or Swadesh) — have been supplied where they are not obvious from the 
context. In a few (but far from all) instances where Sapir cited a language without 
giving a specific form, it has been possible to identify the form Sapir had in mind. 
These and all other editorial interpolations appear in brackets. 

Swadesh, both in his 1953 manuscript and in his 1964 publication, cited all forms 
in a normalized orthography. To minimize future philological difficulties, the origi- 
nal orthographies have been restored wherever possible. This includes all of the 
forms in Wintu, Tsimshian, Coos, and Lower Umpqua in the publications that 
Sapir annotated. Lacking access to Sapirs original marginalia, I have retained 
Swadesh's orthography for these, but in many instances his re transcriptions have 
been supplemented (in brackets) by the orthography of Sapir's probable source 
(most frequently Sapirs own published work on Takelma). This, unfortunately, 
could not be done with Sapir s citations of his still unpublished Chinookan field 
data. Swadesh described his normalization conventions as follows (1964: 182-183): 

Back velar voiceless spirant is shown by /X/ as against /x/ for the front type , /G/ is a velar voiced 
stop, /y is a velar voiced fricative; /i/ is central vowel. /«/ is front-rounded; double letters are 
used for long vowels, but length is disregarded in Chinook, where it is non-distinctive. The Ta- 
kelma stops are given in the system adopted for Coos-Takelma-Penutian comparisons (Sapir 
and Swadesh, 1953): /p t/ and so forth for Sapir s/bd/ and so forth, and also for his p /orp' t' in 
final or preconsonantal position. Aspiration, in positions where it is distinctive, is repre- 
sented by /h/; glottalization by a separate P/ after the character; /s^/ is phonetic glottalized 
(ts). The Takelma tonal accents are interpreted as normal and rising, the latter found only 
on long vowels and on vowel-sonant sequences; it is assumed, on the basis of Sapirs treat- 
ment of tone, that the sustained (on sequences of "inorganic" vowel plus sonant) and the 
rising may be phonemically equivalent. The acute sign is used for the normal accent in its 
different forms, and the inverted circumflex is used for rising or sustained long accent. 
Comparable phonemicization has not been attempted for the other languages. 

The comparisons in the last section (6) are marginalia from Sapirs copy of Franz 
Boas's Tsimshian Texts, New Series (1912), which was apparently not available to 
Swadesh in 1953. It is now owned by Professor John Dunn of the University of 
Oklahoma, who graciously provided the editor with a photocopy of the book 
and a list of all annotations in Sapir's hand. This material is presented in the 
same format as the comparisons in sections 1-5. 

Abbreviations for the languages cited by Sapir, and the sources from which 
most of his data were obtained, are as follows: 

Tak. = Takelma (Sapir 1909c, 1912h; section references to the latter are given 
where appropriate). 

Seven: Penutian Languages 301 

Coos = Hanis and Miluk Coos (Frachtenberg 1913b, 1914a). 

L.U. = Lower Umpqua-Siuslaw (Frachtenberg 1914b, 1917). 

Ts. = Tsimshian (Boas 1911, 1912). 

Chin. = Chinook (Boas 1911; Lower Chinook unless otherwise indicated). 

Wish. = Wishram Chinook (Sapir's fieldnotes). 

Yok. = Yokuts (Kroeber 1907). 

Miw. = Miwok (Barrett 1908). 

L Wintun 

The following are comparative glosses found in Sapir's copy of Roland B. 
Dixon's "Outline of Wintun Grammar" (1909). Dixon's frequently inaccurate 
forms are here supplemented by phonemic forms from Alice Schlichter's Wintu 
Dictionary (1981) and Harvey Pitkin's Wintu Grammar (1984). Section refer- 
ences to the latter are given in parentheses. Swadesh's normalizations of the 
Miwok and Yokuts forms cited by Sapir have not been checked against the 
original sources (Barrett 1908 and Kroeber 1907). 

-ibi, -be, present [-be-, visual evidential (262.11)]: Miw. -b«s, bi-s, 

-liba, -libo, future [-le-, inevitable future (241.61) + -ba-, -boh, durative auxil- 
iary (262.12)]: Miw. -bo, fut. imperative, 
-kila, conditional [-kele, hearsay evidential (243.12)]: cf. Tak. -khi^, -ki^ [- 

k'i^, -gi* conditional suffix, §71]. 
-wu, -u, imperative [-U-, imperative stem formant (222)]: Miw. -pa (?). 
-ut, future particle [= ?]: cf. Tak. -tu [?]. 
-s, forms nouns from verbal stems [-s, general aspect (322)]: cf. Tak. agentive -"^s 

[-^s §80] (?). 
-wi, interrogative [-(w)u-, 1st person interrogative suffix (241.75)]: Coos -uu [-u 

interrogative, Frachtenberg 1914a: 372]. 
-t, used with adjectives, where the noun is in the subjective case [-t, particular 

aspect (320)]: cf. Tak. -t [-f adjectival suffix, §108.1]. 
-da, first person [-da. 1st person subject (243.21)]: Tak. -the^, -te^ [-t'e^, -de^ 

eleu-ibida tconmina T am not dancing' [^elew, negative auxiliary (262.422)]: cf. 

Tak. anii^ [an! negative particle, §113.3]. 
ni hara-kila 'if I go' [har- 'to go, move away from speaker']: cf. Tiik. yana- ('to 

-s, verbal noun [-s, generic aspect (322)]: cf. Coos -is [-is nominalizing suffix, 

Frachtenberg 1914a: 365]. 
bira'ibe-wi 'is he hungry' [bira-- 'to be hungry']: Tak. panx [banx 'hunger'], 
tcona-bem 'he kicks' [c'on-, c'una- 'to kick']: cf. Tak. tkuuntkan- [t'gu"nt'gan- 

'to kick off']. 

302 VI American Indian Languages 2 

tcon-pura-ibi-da 'we kick each other' [-p'ure-, reciprocal (241.41)]: Tak. -an- 

[reciprocal, §55, being compared here with -ra-]. 
ya'paitO 'person' [ya-paytu- 'person, white man, guardian spirit']: Tak. yap^a 

[yap!a' 'person, people'], 
el 'in, into' [^el]: cf. Tak. al- ['face, with eye, to, at', §36.15]. 
pat 'out' [pat]: Tak. p^ay- [p!a-i- 'down', §37.13]. 
xan 'away from' [xan]: Tak. hee- [he*^^- 'off, away', §37.3]. 
panti 'on, upon' [pan 'on top' + -ti, locative suffix (500)]: Tak. pam- 'up' [bam- 

'up into air' §37.15]. 
"Sometimes the locative particle is repeated after the verb, and in some cases two 

particles may be used together": cf. Tak. for use before verbs and in postposi- 
tion with nouns [§37.93]. 
kiilun' 'edge's' [kulu 'edge (of basket)' + -n, genitive case]: Tak.-Coos-Maidu- 

Yok. type of stem [i.e.,disyllabic, with repeated vowel, the "characteristic Penu- 

tian type of stem" discussed in Sapir 1921b]. 

2. Tsimshian (Boas 1911) 

The following are comparative glosses found in Sapir's copy of Franz Boas's 
Tsimshian (1911). The page on which a gloss is found is given in brackets. Forms 
not marked "Tsimshian" are in the Nass River dialect. 

bax- 'up along the ground' [300]: Tak. paa- [ba''- 'up', §37]. 

y!a§a- 'down along the ground' [300]: Coos qa-yaa-c [qaya^tc 'down-stream']. 

ts'Elem- 'into, from the side' [301]: cf. L.U. q-cii [q-tsi 'inside'] (?). 

na- 'out of the woods in rear of the houses to the houses' [303]: cf. Tak. noo 

[no" 'downriver'], hinaw [hina'" 'upriver']; Coos nooc [?], L.U. tnu 'outside, 

lagauk- 'from the side of the house to the fire' [303]: for-k-cf. L.U. adverbs in 

spl- 'out of water' [304]: Tak. pay- [ba-i- 'out (of house, water)'] (?). 
logol- 'under' [305]: Coos qat [qal 'below, under'], 
lax- 'to and fro, at both ends' (Nass), lagax- (Tsimshian) [309]: cf. Chin. laX 

sa- 'off [309]: Tak. xam- [xam- 'into the water'] (?). 

his- 'to do apparently, to pretend to' (Nass), sis- (Tsimshian) [316]: Tak. his 
[his, hl's 'nearly, almost, trying', §114.7] (?). 

wadi- 'like' (Tsimshian) [317]: cf. Wish. tiwi. 

xpl- 'partly' [326]: Coos x- [? discriminative prefix, Frachtenberg 1914a: 324]. 

hwin- 'innermost part' (Nass), wun- (Tsimshian) [329]: Tak., Chin. [ = ?]. 

qa- 'location' (Nass), g-i- (Tsimshian) [331]: cf. Coos, L.U. [ = ?]. 

an-, used to transform verbs into nouns, and express abstract terms, local 
terms, and even instruments (Nass), n-, nE- (Tsimshian) [333]: cf. Coos n- 
[n- 'in, at, to, on, with'], prob. = a-n- 'it-on', cf. Chin. -n-. 

Seven: Penutian Languages 303 

yu-..-k" 'one who has' [334]: cf. Tak. comitative -k"^- (-(a)gw- §46]. 

ha- instrument' [334]: cf. Tak. wa- [wa- instrumental, §38] (?). 

iagai 'already, however, rather' (Nass), y!a^ai (Tsimshian) [339]: cf. Coos, 

tak. [ = ?]. 
-En, causative [344]: Tak. -an- [-an- §45]. 
-sk", expresses primarily the elimination of the object of the transitive verb 

(Nass), -sk (Tsimshian) [344]: Tak. -xa-, -x- [-xa- intransitivizing suffix, 

-A, instead of -sk with words ending in p, t, s, ts, q, x, L, and sometimes in 1 

(Tsimshian) [346]: Not at all. Evidently suffix has different meaning from - 

k^. Cf. Coos -ani (?). 
-s, used in Nass and in Tsimshian in place of -k and -tk after k, x, k^, q, and x 

[346]: cf. Tak. -x. 
-ES, a few times after p in place (?) of -sk, denotes back of object (Tsimshian) 

[347]: Tak. -xa-. 
-d, indicative of many transitive verbs [347]: cf. Coos [-t, transitive, cf. 

Frachtenberg 1914a: 328]. 
ga" 'to take' (Tsimshian) [348]: Chin. [ = ?]. 
dzak 'dead', dzak-d-u 'I kill' (Tsimshian) [348]: cf. Coos ck'^i-t-s [tskwits 'he 

speared him'], 
reduplication of words beginning in hw (Nass), w (Tsimshian) [372]: cf. Coos 

(see vocabulary under w-) [Frachtenberg 1913b: 194-195]. 
smax- 'meat' (Nass), cf. Tsimshian sami 'bear' [377]: Tak. xamk [xam'k' 

'grizzly bear'], 
ie' 'to go (sg.)' [381]: cf. Tak. [?yana- 'to go']. Chin. [ = ?]. 
L6 'to go (pi.)' [381]: cf. Coos [ = ?]. 
gaksk 'to wake up (sg.)' (Tsimshian) [381]: Tak. [ = ?] (?). 
-Em 1st pi. objective (Tsimshian) [381, 384]: cf. Tak. [-am, §62]. 
awa'" 'proximity', awa'"t 'near him (his proximity)' (Tsimshian) [393]: Tak. 

waa-ta [wa'^-(da) 'to, at-him']. 
gwa" 'that' (Tsimshian) [394]: cf. Chin. [ = ?]. 

-1, demonstrative element (Tsimshian) [395, 406]: cf. Tak., Chin., Coos, 
nda 'where' [406]: Chin, ta-n, sa-n, Tak. ne'k-ti [nek'di 'who?', §105]. 
-L, imperative of transitive, indef. obj. (Nass), -\ (Tsimshian) [407]: cf. 

Coos 1. 
dzE, tsE, weakens statements [408]: Wish, -ci^, Coos na-cii [natsi, Frachten- 
berg 1914a: 387], ci [tsi, restrictive particle, Frachtenberg 1914a: 394]. 
optsE 'else, lest' [408]: Wish. pu. 
ami 'if (event assumed as not likely to happen) (Tsimshian) [409]: Tak. 

mii^wa [ml'^wa 'probably, perhaps']. 

3. Coos (Frachtenberg Texts) 

The following are comparative glosses found in the vocabulary section of 
Sapir's copy of Leo Frachtenberg's Coos Texts (1913b). 

304 VI American Indian Languages 2 

a'lEC 'toy': Tak. lou-si [lo"-s-i]. 

a'lqas *fear": Tak. hin^x [hin^x] (?). 

alts- 'to be in the wrong place, to be mistaken': Tak. iils^-ak^ [i'lts!-ak'^ 'bad']. 

ai^- 'to kill (pi. object), to take away': Tak. ey-yi- [ei[y]-i-]. 

a"q- 'to take off: L.U. aaq- [aq-]. 

e« 'thou': Tak. [see Section 4 below]. 

e^natc 'mother': Tak. ni- [ni-]. 

is 'we two (inclusive)': L.U. -ns [-ns]. 

ix- 'canoe': Tak. ey [ei]. 

ilx- 'to look': Tak. liw- [liw-]. 

ilxa'ka 'they': Tak. [see Section 4 below]. 

itE, emphatic particle: Tak. [cf. hi (§114)]. 

in, negation: Tak. [see Section 4 below]. 

I'n-ta 'not so, bad': Tak. [cf. hit (hiit) 'no' (§113)]. 

I'l- 'to tell, to say, to send': Tak. hiim-t- [hi'm-d- 'to talk to'] (?). 

yEq- 'to go away': Tak. yew- [yeu- 'to go back, return'] (?). 

yat- 'to coax, to persuade': Tak. yamat- [yamat- 'to ask (tr.)'] (?). 

yi'pSEn 'three': Tak. xipini [xib'ni]. 

yi'qa 'nevertheless, right away': Tak. [ = ?]. 

yux"- 'to rub': Tak. yuluyal- [yuluyal-]. 

yuwil- 'to divide': Tak. yowo-, yows*^- [yowo-, yo"^s- 'to start (when startled)' 

yut- 'to tear off: Tak. yonk^- [yonk!- 'to pull forcibly'] (?). 
waha'^tcas 'sickness': Tak. hawax [hawax 'rottenness, pus, foul odor'] (?). 
wa'lwal 'knife': Tak. wiil-ii [wi'li' '(stone) knife'], waya [waya 'knife'], 
wa'wa 'little girl': Tak. waiwii [waiwi' 'girl, female'], 
weste'n 'so many times': Tak. [ = ?]. 
wint 'forehead': Tak. «l«k^-i- [ii'liikl-i- 'head-hair'] (?). 
wispa'ya 'arrow': Tak. wilaw [wilau] (?). 

wif 'to look far, to search': Tak. wow-lt- [wo"-ld- 'to go far, go to get']. 
weL-, wIL- 'to twist': Tak. wiik*^- [wi'k!- 'to put around (neck, head)'], 
wl'n- 'to cheat': Tak. wiim-at- [wi'm-ad- 'to exercise supernatural power upon'] 
ha'yati 'to lie down (pi.)': Tak. wayaan-, way- [waya^n-, wai- 'to sleep'], 
ha^- 'to think, to imagine': Tak. hewe-haw [hewe-haw-]. 
han, temporal particle: Tak. [ = ?]. 
han- 'to wrestle': Tak. hemem- [hemem-]. 

hant- 'to pick out, to choose': Tak. henee-t- [hene'^-d- 'to wait for'] (?). 
hak- 'to crawl': Tak. waka-xa- [wage-xa- 'to climb up']. 

hala'qi:s,hali'qas 'relatives of husband': Tak. wayaw- [wayau- 'daughter-in-law'] 
halt! 'now': Chin. alta. 

halk"- 'to take off: Tak. weet-k- [we^t'-g- 'to take away from, deprive of] (?). 
halq-, helaq 'to climb up': Tak. hilw- [hilw-]. 
halq 'tir-tree': Tak. x6 [x6] (?). 
he, temporal particle: Tak. [see Section 4 below]. 

Seven: Penutian Languages 305 

heilta 'tongue': Tak. ela- [ela-]. 

hen, particle denoting hearsay: Tak. [ = ?]. 

henl- 'a while, a long time': Tak. [see Section 4 below]. 

he'lmi 'tomorrow': Tak. [see Section 4 below]. 

helq-, he'laq 'to arrive': Tak. wowk- [wo"g-]. 

hitc, particle denoting surprise: Tak. [see Section 4 below]. 

hrk!am 'monthly courses': Tak. wtHilh-am- [wii^lk-am- 'to have first menstrual 

courses'] (?). 
hu"'mik- 'old woman': Tak. wuun- [wu"n- 'to be, grow old'] (?). 
pEnL- 'to tear off, to come off: Tak. potpat- [bot'bad- 'to pull out (somebody's) 

pa^w- 'to smoke (a pipe)': Tak. phoy-amt- [p'oy-amd- 'to smoke out (wasps)'] 

pa^- 'to fill': Tak. ptwk^- [bii'^k!-]. Chin. paal-. 

pils- 'to tear, to smash to pieces': Tak. philiphal- [p'ilip'al- 'to squash, whip']. 

pilx"- 'to break, to crush': Tak. pilw- [bilw- 'to jump (at), fight with'] (?). 

pidj- 'to come up (from water), to appear': Tak. piis- [bi's- 'to look up, lift up 
one's head' (used only in myths)]. 

pux"- 'to spout': Tak. phowphaw-i- [p'oyp'aw-i- 'to blow']. 

mEanl'yas 'parents, grown-up people': Tak. ma- [ma- 'father']. 

maha- 'to watch, to look after': cf. Hokan. 

ma'qaL 'crow': Tak. mel [mel]. 

ma'luk" '(Indian) red paint': Tak. maanx [manx 'white paint'] (?). 

ma^L! 'flood': cf. Chin. maX-ni. 

ma'x" 'vulva': Tak. min- [min- 'vagina'] (?). 

mitsl'le 'pregnant': Tak. mix^ii- [mAhwI-]. 

mi^la 'liver, waist': Tak. p^an [p!an 'liver']. 

mi'iaq 'arrow': Tak. smelaw'^x [smela"^x]. 

ml'naqas 'pole': Tak. mal [mal 'salmon-spear shaft']. 

da'mil 'strong, male being, husband': Tak. tomxaw [domxau 'big crooked-nosed 

drl 'something': Tak. al-til [al-dil 'all']. 

tE 'that there': Tak. [ = ?]. 

tEW- 'to be in upright position': Tak. t^epe-, teep- [t!ebe-,de<-"b- 'to get up']. 

ta, conjunction: Tak. [ = ?]. 

tat'n- 'to fix, to divide': Tak. t'-^ilii-, tiln- [t!ili-, dlln- 'to distribute to']. 

te'lex- 'crosspiece': Tak. tele-p- [dele-b-] 'to stick into', taia-k- [dala-g- to pierce 
nose, ears']. 

ti'k- 'to stand': Tak. tiiktak- [di'k'dag- 'to erect, cause to stand up'j. 

ti'lpl 'gopher': Tak. thiis |t'i's]. 

tlyet- 'to store up food': t««lt^al- [dult!al- 'to stuff (basket) with']. 

tl^- 'to coil': Tak. t^uukuuy-, ttmk^- [tlugui-, dii^gw- 'to wear (garment)']. 

ti'ntc 'remnants of meal': Tak. t'-^ayay-, taay- [t!ayai-, dai- 'to go to get some- 
thing to eat'] (?). 

toh- 'to hit, to strike': Tak. toy-k^ [dui-k!- 'to push'] (?). 

306 VI American Indian Languages 2 

to"s- "to hit, to strike': Tak. t^omom-, toom- [tlomom-, do"m- 'to kill'], 
tkwii'tuk" 'shoulder', t^kwa'tkwis 'elbow': Tak. tk^iintk^- [t'gwi'nt'gw- 'upper 

tEk-elm- 'to dive, to sink': Tak. tkeP [t'gel^ 'drop, fall']. 
tqanL- 'to hit, to strike (with instrument)': Tak. tk^el-tk^al- [t'gwelt'gwal- 'to 

break in two']. 
tqa'L- 'to put on a belt': Tak. tkenc*?- [t'gents!- 'to put about one's middle']. 
tEq!e''^n 'soot': Tak. sin [s-In 'wood-coals'] (?). 
t!c- 'to shove, to push': Tak. t^osot^as [t!os-ot!as- 'to walk around']. 
t!kw- 'to kick': Tak. tkuuntkan- [t'gu"nt'gan-]. 
t!kwa 'roof: Tak. tkuup- (t'gu^b- 'to cover over, put lid on'], 
nl'kla 'mother (vocative)': Tak. ni- ['mother']. 

s^aL! 'pitch': Tak. seel [se'^l 'black paint, writing'], k^^al [klwal' 'pitch'] (?). 
sin! 'hair on penis': Tak. seen- [sen- 'hair'], 
sil- 'to drop (of liquids)': Tak. tk^il-i- [t'gwili-] (?). 
so^t- 'to trade, to exchange': Tak. -xoot- [-xo"d-] (?). 
sto"q- 'to stand, to be in an upright position': cf. Tak. [ = ?] (?). 
sq- 'to seize, to take': Chin, -sk- (?). 

sla '(male) friend, cousin': Tak. snaa [sna 'momma!' (vocative)] (?). 
ci- 'to drink': cf. Yana si-, 
canx- 'to shake': Tak. s'^^el- [ts!el- 'to rattle']. 

cima- 'to draw by means of breath': Tak. s^usm- [ts-!usm- 'to chirp'], 
cl'tla 'pet': Tak. c^ixi [ts-!u'xi] 'dog'. 
cuL- 'to set fire to': Tak. s^«l-k^- [ts-!ul-k!-] (?). 
tSE7i- 'to crush': Tak. c'^atat-, saat- [ts-!adad-, sa^d- 'to mash'] (?). 
tSEL- 'to stand, to lie side by side': Tak. xoxok^- [-xoxogw-'to string salmon'] (?). 
tsa^ntc 'on the back' (borrowed from the Siuslaw tca^'n- 'to lean backwards, to 

recline'): Tak. xaah-am- [xa-^h-am- 'back, waist'] (?). 
tsn- 'to stretch': Tak. tin-k^- [din-k!-] (?). 
tsk"- 'to hit against, to strike against, to spear': Tak. saak^- [sa^gw- 'to shoot 

(arrow), paddle (canoe)'], 
tsx- 'to take': Chin, -sk- (?). 

tsqal- 'to defecate': Tak. xalaxam- [xalaxam- 'to urinate'] (?). 
tshm 'summer': Tak. sama- [sama-]. 
tci'ntsan 'friend': Tak. k^hinax- [k'winax- 'kinsman, relative']. Upper Tak. 

k^htHtnakst [k'ii'^naks-t' 'his relatives'], 
tcils 'penis': Tak. khal [k'al]. 
tco"-, tco^" 'to jump': soo-'^k- [s-6"^k'-]. 
tcut 'nose': Tak. sin- [s-in-]. 
tcH 'raft': Tak. sil [sil 'canoe' (in verbs)]. 
tc!a- 'to walk': Tak. s'^aak- [ts!a^g- 'to step']. 
tc!ici'lis 'sweet', tc!ical- 'to be sweet': Chin. [ = ?]. 
tc!il- 'to burn': Tak. -s^«l-k^- [-ts!ul-k!- 'to set fire to']. 
tc!iltc! 'hammer': Tak. s'^al-s'^al- [ts!alts!al- 'to chew']. 
tc!6"- 'to lie down, to go to bed': Tak. s'^oot- [ts!6"d- 'to touch, strike against']. 

Seven: Penutian Languages 307 

kwE'li7Es 'intestines, a person created from the intestines': Tak. k^aas [gwas 

kwe'ik- 'girl', kwe's 'young woman, girl': Tak. khay^- [k'ai^-]. 

kwil-, k!"hil- 'to burn': Tak. k«l-k^- [giil-k!- 'to blaze, glow'], of. Tak. k'winax 
: k'ii'^nax [for phonological detail]. 

kwit- 'to leave': Tak. k*?otot-, koot- [k!odod-, go"d- 'to bury']. 

kwiL- 'to dig (clams)': Tak. k^olol-, kool- [k!olol-, go"l- 'to dig']. 

kxla 'foot, leg': Tak. sal- [s-al- 'foot'], k^el- [gwel- 'leg'] (?). 

k!a 'rope': Tak. k^ank^an- [k!ank!an- 'to twist (hazel switch)']. 

k!a' 'to listen, to keep quiet': Tak. skek^ii- [sgek!i'- 'to listen']. 

k!walxa'ya 'butter-ball' [duck sp.]: Wish. [ = ?]. 

kiwa'sis 'wind': Tak. k^alt [gwaPt']. 

k!wehe 'willow': Tak. k^^aay [k!wal 'grass']. 

k!wints 'throat, neck', k!win- 'to swallow': Tak. k^en- [gwen- 'neck, nape']. 

g-a'we 'sea otter': Tak. tkam [t'gam 'elk (hide, hide armor)'] (?). 

k-i'nwis 'lazy', k-i'ria" 'tired': Tak. kenaw- [genaw- 'lie curled up dog- 

k-ow- 'to munch, to pick and eat': Tak. k^owoo-, kuuw- [k!owo"-, gu"w- 'to 
throw a mass of small objects']. 

k-!em- 'to practice': Tak. k^emn- [k!emn- 'to make, treat as']. 

k-!si'lis 'green': Tak. tkism- [t'gis-m-] (?). 

k-ele 'shouts': Tak. skelw- [sgelw- 'to shout']. 

qano'tc 'outside': Tak. noo^s [n6"^s 'next door']. 

qal- 'to dig': Tak. k^olol-, kool- [k!olol-, go"l- 'to dig'], -holohal- [-holohal- 'to 
dig into (ashes)'] (?). 

qal- 'to cry (sing.)': cf. Chin. [ = ?]. 

qa'ya 'breath': Tak. hekehak- [hegehag- 'to breathe'] (?). 

qa'yis 'sky, day, world': Tak. haay [hai 'cloud'] (?). 

qai'na 'cold': Tak. tkuun-p- [t'gu"n-p'-] (?). 

qa"'wa 'evening, night', qa"'m'(tc) 'evening': Tak. hoo-xa [ho"xa 'yesterday', 

qEto" 'to be in suspended position, to hang': Tak. xataxat-na- [xadaxat'-na- 'to 
hang up in a row']. 

q!al- 'to take out': Tak. k^^alk^- [kiwalgw- 'to let alone']. 

q!a'na 'young': Tak. k^^al-thaa [k!wal-t'a'' 'youngest (of two or more)']. 

q!e}e 'pitchwood': Tak. k^^al [k!war]. 

q!m- 'to eat, to cook': Tak. k^uum-an- [k!u^'m-an- 'to fix, prepare'] (?). 

xa'yusLatc 'relatives by marriage after death of person causing that 
kinship': Tak. ximni- [ximni-]. 

xan- 'to be sick, to be sorry', xii'nis 'sick': Tak. xilam [xil-am 'sick, dead per- 
son, ghost']. 

xa'ka 'he, she, it': Tak. haa^ka [ha'^^ga 'that yonder', §104] [see Section 4 

xo'xwel 'frog': cf. L. Chin. [ = ?]. 

x"kwi'natc 'maternal aunt': Tak. xaka- [xaga-]. 

308 VI American Indian Languages 2 

xwelap 'lungs': cf. L. Chin. [ = ?]. 

xwi'nLis 'snot': Tak. xin [xin]. 

xwi'lux" 'head': Tak. xuul-i- [xu"l-i- 'brains'] (?). [See different comparison in 

Section 4 below.] 
xqas 'white': Tak. tk«y^s [-t'gu'^s-]. 
xlls 'slime': Tak. xla^px [xle^p-x 'roundish dough-like cake of deer-fat or 

X il- 'to look around': Tak. xanan- [xanan- 'to look out (pi.)'] (?). 
Eqa"'we 'to die (sing.)': Tak. loho- [loho- 'to die']. 
E'71 'good, nice': Tak. tuu [dii] (?). 

aa- 'to put around': Tak. laat- [la^d- 'to put (belt) around waist'], 
atclya- 'to call by name': Tak. laalaw- [la^law- 'to name, call'], 
aix"- 'to jab': Tak. laaw-t-an- [lau-d-an- 'to hurt'], 
ewi 'it is, that is': Tak. laaHi- [la-^'l-i'- 'to become'], 
em- 'to be in upright position, to stand (of inanimate objects)': Tak. lemk^- 

[lemk!- '(people) move, go, to take along (pi. obj.)'] [see lim- below], 
e'xalx 'string': Tak. laaw- [la^w- 'to twine (basket)'], 
e'xum 'buzzard': Tak. moxo [moxo]. 
ic- 'to shake, move (intransitive)': Chin. -la. 
icla7- 'to swing': Tak. lewe^law- [lewe^iaw- 'to swing (shells) in one's ear'], 

smilismal- [- smilismal- 'to swing'] (?). 
i'kwit 'feathers': Tak. xliwi [xliwi 'feathers worn in war-dance'] (?). 
^x-il- 'to like, to love': Tak. miil- [mlli'-d- (?). 
, abbreviated form of dl4 ['something']: Chin. t-. 
a- 'to go': Tak. yaan- [ya'^n-] (?). 
im- 'to put inside (pi. object)': Tak. lemek^- [lemek!- '(people) move, go, to 

take along (pi. object)'] (?) [see lem- above], 
^yuwil- 'to move, to wiggle': cf. Ts. [ = ?]. 

h- 'to get weir, the- (the-) 'to rest': Tak. likii-n- [ligi'-n- 'to rest'], 
n'nas 'name': cf. L.U. hin [tin], 
k"- 'to sew': Tak. lep-t- [lep'-d-] (?). 
kwilt'red': Tak. al-s^iil [al-ts!il] (?). 
qalk"- 'to bite': Tak. lek^ei- [legwel- 'to suck']. 
Lala^- 'to bark, to growl, to shout, to wail': Tak. khewekhaw-al- [k'ewet'aw-al- 

'to bark at'] (?). 
L'^va 'son (vocative)': Tak. khapa- [k'aba- 'son'] (?). 
Loc 'clam': Tak. kos [gos- 'clamshell'] (?). 
Lkwa^- 'to cut off: Tak. skoot- [sg6"d- 'to cut'] (?). 
Lx-- 'to drift (away)': Tak. thiyii- [t'iyi'- 'to float']. 
L!ta 'land, earth, country, ground, place': Tak. tkaa [t'ga]. 
L!ka- 'to string': Tak. tpaak- [t'ba^g- 'to tie up (hair, sinew)']. 
L!kw- 'to cover up': Tak. tpook-tpak- [t'bo"k't'bag- 'to put away']. 

Seven: Penutian Languages 309 

L!k-- 'to pour, to spilK: Tak. tkiiy-al-x- (t'gi'y-al-x- 'tears roil down ones face']. 
L!le- 'to come out (from water)': Tak. theek- [t'e'^g-]. 

4. Coos (Frachtenberg Grammar) 

The following are the comparative glosses found in Sapir's copy of Leo 
Frachtenberg's Coos (1914a), omitting comparisons duplicating those in his 
copy of Coos Texts (Section 3, above). The page on which a gloss is found is 
given in brackets. 

he'niye 'a while', he'nihen 'many times' [313]: Tak. he^ne [he'^ne' 'then', 

§112.3], hem-ti [hemdi 'when', §113.2]. 
In, negative particle [314]: Tak. anii^ [anP §113.3]. 
"Hiatus .... Broadly speaking, it may be said that the coming-together of two 

vowels of like quantities and qualities is avoided by means of infixing a weak h 

between them" [314]: cf. Tak. ["inorganic"] -h- [§24]. 
helmi'his 'next day' [314]: Tak. te-wenxa [dewenxa 'tomorrow', §112.2]. 
xwi'lux" 'head' [316]: Tak. «l«k^i- [ii'liikl-i- 'head-hair'] [see different com- 
parison in Section 3]. 
n-, 1st person [321]: Tak. -a^n [-a^n 1st person sg. transitive subject (aorist)], 

-an [-an (future), §63]. 
e*-, 2nd person [321]: Tak. -te'^, -'^t [-de^, -'^t', 2nd person sg. possessive, §90]. 
nkwa'xLa 'with bows' [322]: Tak. kaP [gal^ 'bow, gun']. 
mitsT'ltfye 'become pregnant' [323]: Tak. max^ii^ [mAhwi^] ? = mix^ii^ [see 

also Section 3]. 
ntcla'ha '[equipped] with walkers [i.e., legs, feet]' [323]: Tak. s^aaks^ak- 

[ts!a^k'ts!ag- 'to step'], 
-t, transitive [327]: Tak. -s-, from -tx- [indirective, §47]. 
-ts, transitive [327, 329]: Tak. -s-, from -t-s. 
-I, e', neutral [ = mediopassive] suffix [327, 334]: Tak. -ii [-1, suffix forming verbs 

of position, §57]. 
-s, general nominal suffix [328]: Tak. -s [§87.8] (?). 
-aai, intransitive [332]: cf. Tak. -xa- [§53]. 

yoyo^waai 'is stopping' [332, 364]: cf. Tak. yo-, yowo- ['be sitting'] (?). 
-me", reciprocal [332]: cf. Tak. -an- [§55]. 

-ts-XEm, -t-XEm, reflexive [333]: cf. Tak. -k^i-, -k^i- [-gwi-, -k'wi- §54]. 
laatsxEm 'put herself in' [333]: cf. Tak. las^ak-, lask- [lats!ag-, lasg- 'to touch']. 
-e'wa(t), -o"wa(t), frequentative [336]: cf. Tak. -eeha, -iiha [-eha, -Iha §43.6]. 
nhaml'yat 'I brought it out' [340]: Tak. hemek- [hemeg- 'to take out, off'], 
-ani, transitive [341]: Tak. -an- [causative, §45]. 

310 VI American Indian Languages 2 

tsak"kw-, -tskw- 'to spear' [341]: Tak. saak^- [sa^gw- 'to shoot (arrow), paddle 

(canoe)'] [see Section 3]. 
tqa'nL- 'to strike' [341]: Tak. saansan- [sa^nsan- 'to fight'] [see Section 3 for dif- 
ferent comparison]. 
\\ tqanLLii'nl 'they mutually strike one another' [341]: Tak. sana-x-in-iyan- 

wil- 'to fight' [342]: Tak. te-wiliw-alt- [de-wiliw-ald- 'to fight with, "go" for' 

(< wiliw- 'go, proceed, run')] (?). 
-ayu, -e'yu, -lyu, (past) passive [327, 344]: Tak. -yaw- [-iau-, impersonal, §58]. 
-aya", past participle [347]: Tak. -lya [§83] (?). 
-a, in formation of past participle [347]: from-ii, [-1] cf. L.U. and umlauting force 

in Coos, 
k-ela 'hand' [348]: Chin, -ksi, -ksn (?). 
te*L 'this' [348]: Chin, ti-ka. 
-It, suffixed to verbs that are transitivized by means of the transitive suffix -aya, 3 

pers. obj. [348]: cf. Tak. [ = ?]. 
-enl, verbalizing suffix, 'to do, make something' [349]: Tak. causative -an- [§45]. 
-a, transitive verbs [354]: cf. Tak. [§44-51] (?). 

-is, nominalizing suffix [360]: Tak. -(a)x, [intransitive] infinitive [§74]. 
-onis, -si, verbal noun formants [363]: cf. Tak. passive [(aorist)] -an [§66]. 
L!x-T'nt 'he examined it' [364]: Tak. xiik- [xi'g 'to see'] (?). 
e^niitc 'mother', ni'kla 'mother!', vocative [366]: Tak. ni- [ni-(xa) 'his mother'], - 

hin [(wi-)hln '(my) mother'], 
teka'^tsi 'granddaughter!', vocative [366]: cf. Chin. [ = ?]. 
a'tatc 'paternal aunt' [366]: Tak. that- [t'ad-]. 
mi'nkatc 'son-in-law' [366]: Tak. mot-, moo- [mot', mo"-] (?). 
-ex, -lyEX, -lyetEX 'pertaining to . . .' [367]: Tak. -^iixi [-^r'xi 'belonging to', 

yiqantcimex mii 'the last [ = previous] generation' (< yiqantc 'behind') 

[367]: Tak. -imik^i [-imik!i, suffix forming adjectives from temporal adverbs, 

slaatc 'cousin' [371]: cf. Tak. snaa [s-na 'momma!' vocative, §91] (?). 
-en, multiphcative suffix, 'times' (in numerals) [373]: Tak. -an, -tan [-an, -dan 

hl''me 'children' [374]: Tak. haap- [ha'-'p'- 'small child'], 
ma 'human being', men, plural [374]: Tak. -(h)an [§99]. 
L!ta'yas 'village', derived from L!ta 'earth, ground, country' [375]: ? from 

X'taay- [L!tay-] (cf. L.U. X'ay [L!a'»'] 'place, world, people', X'ayuus [L!ayu's], 

locative form) + -as. [See Section 3 for Takelma comparison], 
hu^mi'k-ca 'dear old woman' [375]: cf. Chin, wi-na-m-s wa-n-aq-s. 
-I, found suffixed to the article, expressing instrumentality [377]: cf. Ts. -ii (-1) 

final reduplication [380]: cf. Tak. [§30]. 
tco'xtcox 'rabbit' [381]: Tak. hoow [hou] (?). 
he 'usually, frequently, habitually' [384-5]: Tak. -wi^ [-wi'^ 'every, §101] (?). 

Seven: Penutian Languages 311 

c*^, expresses slight surprise at a state of affairs that has come into existence con- 
trary to one's expectations [389]: cf. Chin. s-. 

hitc, indicates surprise [391]: Tak. his [his 'nearly, almost, trying', §114.7]. 

L 'must, necessarily' (exhortative particle) [392]: Chin. XX (?). 

hamiL, mIL, iL 'let me, I should like to, better, you may, please, a while' (exhor- 
tative particle) [392]: Ts. amii [ami], Tak. mii^wa [ml'^wa 'probably, perhaps', 

n'ne, 1st person (independent pronoun) [396]: L.U. -n. 

e^ne, 2nd person [396]: L.U. -nX [-nx]. 

i'sne, dual inclusive [396]: L.U. -ns. 

i'cne, 2nd person dual [396]: L.U. -c [-ts]. 

xwin'ne, dual exclusive [396]: L.U. -Xun [-xiin]. 

lin'ne, 1st person plural [396]: L.U. -nl, inclusive, -nXan [-nxan], exclusive. 

cin'ne, 2nd person plural [396]: L.U. -ci [-tci]. 

itxa, 3rd person plural [396]: cf. Chin. 1- (?). 

ntet, 1st person singular reflexive (cf. tet 'body') [400]: Tak. -ta^x [-da^x 
restrictive suffix, "denoting the isolation of the person," e.g. 'only I', §103]. 

la", ha" 'he, it is' [401]: Tak. haa^- [ha'^^ demonstrative stem, §104]. 

lewi 'it is' [402]: Tak. laalii- [la^l-i'- 'to become']. 

-tc, adverbial suffix of modality [405]: Tak. -xa, -x [adverbial suffix, §112.2]. 

i 'when, as, since, while' [409]: Tak. isi^ [T's-i^ 'despite, although, even if, 

a'watu 'whether or not' [411]: Wish, awa-ci^ 'or'. 

nouns as qualifiers [412]: cf. Tak. [§88]. 

k!al- 'to shout' [413, 416]: Tak. skelw-, skelew- [sgelw-, sgelew-]. 

cx-Iml 'bear' [415]: Tak. xamk [xam'k' 'grizzly bear']. 

5. Siuslaw (Frachtenberg Grammar) 

The following are comparative glosses found in Sapir s copy of Leo Frachten- 
berg's Siuslawan (Lower Umpqua) (1917). The page on which a gloss is found is 
given in brackets. 

nx-l-k becomes nak [446]: cf. Wish. -tX-k becomes -t-k. 

winx- 'to be afraid' [446]: cf. Tak. hin^x [hin^x 'fear']. 

-Imts, indirect object of first person imperative [501]: cf. Tak. -xi [?]. 

Jan- 'to call by name' [514]: Tak. laalaw- [la''law-]. 

wihc- 'to send' [520]: cf. Coos will- [wil- 'to look for, to search'] (?). 

-t, present tense [527]: cf. Wish. [ = ?]. 

mate- 'to be in a horizontal position', ma-'tc- 'to lay' [536]: cf. Tak. mas'^ak-, 
mask- [mats!ag-, masg- 'to put']. 

-a, nominal object of an action, also the local idea of rest [ = accusative case-end- 
ing] [541]: cf. Yokuts -a. 

312 yi American Indian Languages 2 

-a, modal adverbs [557]: cf. Tsim. adv. -a. 

qa'wl 'blood' [560]: cf. Chin. [ = ?]. 

-a^mu, nominalizing suffix indicating place [563]: cf. Coos -ame. 

-t, adjectival suffix [564]: cf. Tak. -t [-(i)t' §108.1]. 

tsami'tsEm 'chin' [565]: Coos cneex [ts'na'x] 'beard'. 

-wi, suffix found in a small number of nouns [565]: cf. Coos -eyeewe [-eyawe, 

noun of agency, Frachtenberg 1914a: 364]. 
tsxan- 'to comb one's hair' [565]: cf. Chin. [ = ?]. 
duplication of final consonants [567]: cf. Tak. [§30]. 

tcaq- 'to spear' [568]: Tak. saak^- [sa^gw- 'to shoot (arrow), paddle (canoe)'], 
hlq" 'wildcat' [570]: Tak. yaak^ [yak'^]. 

stem amplification [573]: equivalent to Tak. aorist verb stem [§39-40]. 
tqut- 'to shout' [575]: Tak. skelel- [sgelew- 'shout', sgelel- 'keep shouting'] (?). 
xa'ts!un 'four' [586]: = 2 x (2), -n cf. Tak. -an, -tan [- an, -dan multiplicative 

suffix, §111]. 
qo'x"m 'offshore, out in the water' [589]: Tak. xam- ['in river, into water', §37.16] 

stim, stimk 'there' [589]: cf. Ts. [ = ?]. 

aldu 'Hkewise' [592]: Tak. altii [aldl 'all']. 

"I 'and' [592]: Coos il [il 'surely, indeed'], Tak. altii-1, -tiil [aldI-1, -dil 'all, 

aL 'now' [593]: L. Chin, alta, alqa. 

hanhan 'indeed, to be sure' [594]: Coos han [temporal particle], 
na, particle of interrogation [599]: L. Chin. na. 

6. Tsimshian (Boas Texts) 

The following are comparative glosses found in the vocabulary section (pp. 
256-284) of Sapir's copy of Franz Boas's Tsimshian Texts, New Series (1912). 

aya- 'successfully': Chin. aiaq. 

awa' 'proximity': Tak. wa- ['to, together with']. 

ap 'bee': Wish, wa-ba. 

am- 'only': Wish, nai-ma, L. Ch. nam-ka. 

ami 'if: cf. Tak. mT''wa ['probably, perhaps'] (?). 

amuks 'to listen': (cf. mu 'ear'?). 

al 'but': L.U. "1 ['then, so, and', introductory particle] (?), Coos il ['surely, 

a'Mks 'servant': cf. Kwakiutl Elk", 
at 'not' (in interrogative sentences): cf. Na-dene. 
emx 'beard': Chin, -miqcu. 

Seven: Penutian Languages 313 

ya° 'to go, sing.': Chin, -ya, Tak. yana-. 

y!an 'excrement': Tak. xa^n- 'urine' (?). 

yaltk, pi. yilyaltk 'to return': Tak. yew-. 

yei, pi. yikyei 'fat': cf. L.U. [= ?]. 

y!u'«ta 'man': L.U. hit-c (?). 

wak- 'brother', waik- 'elder brother': Tak. wa^-, Chin. awi. 

walp, pi. huwalp 'house': Tak. wih. 

wa 'to invite': Tak. wo"- ['to go for, go to get'] (?). 

woq 'to sleep': Tak. wai-, Yok. woi. 

wun- 'innermost part', wun-^aus 'brain' (cf. gaus 'head'): Chin, -wan, Tak. 

ha-^win-i-(de) ['inside of (me)'], 
wul 'being': Tak. wala' ['really, come to find out'] (?). 
wuir^l, pi. wulwuir^l 'to rub': Tak. yul-, final red[upIication]. 
hau 'to say': ? < *waw-. 

haowi'n 'before': Tak. ha'wi- ['still, yet', §113.2] (?). 
hats! 'hardly, now': demon, ha- (cf. Coos [he 'the', ha 'his, her, its']), 
ha^x 'goose': cf. Tak. [ha^k'a^]. 
haldEm- 'upward': cf. lak. hilw- ['to climb']. 
hi- 'beginning': L.U. hlq!- ['to begin'], 
ha 'to shout': L.U. hal-. 

han 'salmon' . . . SE-wI-ha'n 'twins ( = making plentiful)': cf. Nootka. 
hum 'to smell something': cf. Chin. 

hu^t 'to escape, pi.': Tak. ho"gw- ['to run (without expressed goal of motion)']. 
hu°tk, pi. hukhu'ot'^ 'to call, summon': L.U. hant'- . 
huk- 'expert' (par.): Tak. yok'y- ['to know'] (?). 
bEn, pi. §a-bE'n 'belly': Chin, -wan (?). 
bEla'x 'moss': Tak. blls. 

balax, pi. bilba'lax 'ghost': Tak. bilam ['having nothing, empty'] (?). 
pe 'liver': cf. Tak. p!an. 

bu", pi. gabu'o 'to blow': cf. Tak., Coos, Chin, 
pha'r 'to relate': Tak. p!al-g- ['to tell a myth to'] (?), L.U. pulk-na- ['to speak'] 

p!El-mu' 'ear-ornament': cf. p!al 'button'. 

p!ia'n 'smoke', x-p!ia'n 'to smoke ( = to enjoy smoke)': cf. Southern Paiute. 

m 'thou (transitive subject)': cf. Chin., Tak. 

mES- 'reddish' . . . mEsi'^n 'copper': cf. Hokan, Algonquian. 

mEsx 'breast': Chin, -mxtc 'heart'. 

m!an- 'up through the air': cf. Hokan 'up, sky'. 

mag 'to put down one object', mak-sk 'to put down several objects': cf. Tak. 

[masg- 'to put']. 
ma'Mk 'to burn': cf. Tak. [me^l- 'to blaze'], 
mat, pi. mElmal 'to tell': cf. Tak. [malg- 'to tell, speak to'], 
mu 'ear': Hokan? 

du'la, pi. ga-du'la 'tongue': cf. Wintun. 
tgi- 'down through the air': Tak. t'gi'y- ['tears roll down one's face']. 

314 Vf American Indian Languages 2 

tgin 'to drill fire, [t]gu- 'around': Tak. t'gei- ['to roll, put around']. 

txa 'skin': Tak. t'gwa^x-an- ['to tattoo'] (?). 

txal- 'against': Tak. =t'gwa"l- ['to run about, whirl past'] (?). 

t!a« 'to slap': cf. L. Chin. 

t!alp 'to cover with hot ashes or fire': Coos t!al 'roast'. 

t!ii" 'to sweep, tran.', t!ii°sk, intrans.: Tak. -xa- [intransitivizing suffix]. 

t!u"s 'to push, to beat with fist': cf. Coos t6"s ['to hit, to strike']. 

SEksu' 'urine': cf. Tak. xa^n (?). 

sa°l 'to notice', only in negative form: Tak. -xanaw ['to look out'] (?). 

su"nt 'summer': Tak. sama- (?). 

sta 'half, one side of a long thing': Chin. cit!-ix, Tak. -xdil'- ['notch']. 

stslal 'beaver': Coos tEtci'na (?). 

sg-an 'gum, pitch': Coos s'^aL! (?). 

dzab, pi. dzEbdzab 'to make': Tak. xe^b- ['to do (so)'] (?). 

dzai, pi. dzEldzat 'to eat up, to beat in game': Tak. -xeml- ['to desire to eat'] (?). 

tslE- 'inside of a thing' (only in the form ts!E-wa'lb 'inside of house'): L.U. [qtsi 

ts!ats!a 'hail': cf. Tak. [ts!ele-m- 'to hail']. 
ts!al, pi. ^-ts!Elts!al 'face, eye': Tak. tc!elei- ['eye']. 
ts!i"p 'to tie, to close eyes': Tak. -ts-!ibib- ['to shut (doorway, hole)'], 
n-, UE-, nominal prefix: Chin, [illegible]. 

na'ya 'mother!' (said by girl): Tak. [s--na 'momma!' (vocative)]. 
ne'tSEks 'fish-tail': Chin. -lict. 
n!axn!6 'to hear': Uto-Aztecan. 
gao 'to take sing, obj.': Chin. -ga. 
gu'p!El 'two round objects': Tak. ga"m ['two']. 
gun- 'to cause' (par.): Haida (?). 
gwa" 'that': cf. Chin. 

gwanks 'cooked, done', SE-gwanar 'to cook': disyllabic stem, 
ksax 'to go out, pi.', ksEr 'to go out, sing.': ? = *ksaxwax"-. 
gP-tc 'to swell, to flood', g-I"n 'to cause to swell': cf. Tak. [-(a)n-, causative], 
qa 'rabbit' (?): Athabaskan. 

gal-, qal- (par.) 'empty': cf. Tak. k!wal- ['let alone'] (?). 
gala'r 'cedar': Chin, -cgan (?). 
ga"d 'heart, mind': Tak. giixw- [ = ?] (?): 
gox, qox 'to peck', q!ax, pi. q!alx 'to pluck': note infix -1-. 
la- 'not quite in the right manner' (par.): Chin, la'ax. 
lEt 'excrement' (?): Tak. la'. 

lamdzEX 'to enter, pi.': Coos tim- ['to put inside (pi. object)']. 
laq 'to take name' (?): Tak. la^law- ['to name, call'], 
lo'e 'neck-ring of cedar-bark': Chin. -Ixwa (?). 
lu- 'in': Chin. -1-. 
lu"p-k 'to sew': Tak. [lep'd-, lebe-]. 
tEb 'smooth', Ie'Iep 'to smoothen': infixed -I-. 
lantk, pl. hitantk 'to move', la^'l, pl. h:»a '^'1 'to move something': Chin. -la. 

Seven: Penutian Languages 315 

ta" 'to slide': L.U. sLox" ['to descend, slide down']. 

latk 'to move': Chin. 

toga 'to go, pi.': Coos la- ['to go']. 

Ika'ak 'sister' (said by man): Chin, -ixt (?). 

Section Eight: 
Wakashan and Salishan Languages 


Within a few weeks of assuming his duties as Chief of the Anthropological 
Division, Geological Survey of Canada, in September 1910, Sapir set out for 
British Columbia for linguistic and ethnographic field work on the Nootka of 
Vancouver Island. A genetic connection between Nootka and Kwakiutl had 
been proposed by Boas (1891:678-679), and although this was accepted by 
Powell (1891), who provided the family with the name "Wakashan," the details 
of the relationship remained to be worked out. This was the sort of task Sapir 
relished, and his first report on his Nootka work (1911e) was a preliminary sur- 
vey of Nootka-Kwakiutl comparative linguistics. He apparently continued to 
work on the project intermittently for several years. The notes he accumulated, 
now in the Franz Boas Collection in the Library of the American Philosophical 
Society (manuscripts 497.3 B63c Wla.2 and W1.3; cf. Freeman 1966: 216, item 
1954, 380, item 3836), formed the basis of an extensive "Wakashan Com- 
parative Vocabulary" (American Philosophical Society Library, manuscripts 
497.3 B63c Wla.26 and Wl.l; cf. Freeman 1966: 380, items 3837-3838) pre- 
pared by Morris Swadesh in 1950-51, which is published for the first time in 
Volume XII of The Collected Works. 

Sapir visited the Nootka twice, from September to December, 1910, and from 
September, 1913, through February, 1914. His principal informant was Tom 
(Sayach'apis), an elderly blind man (19181, 1921e, and 1922y). From Sayach'apis 
and others Sapir obtained an extraordinarily detailed record of Nootka lan- 
guage and culture, including 71 texts, several of considerable length. During 
the second trip Sapir taught two of his young interpreters, Frank Williams and 
Alex Thomas (the latter a grandson of Sayach'apis), to take phonetic dictation ; 
he arranged for them to continue to transcribe texts and collect other data after 
his departure (1915j: 172). This collaboration was explicitly modeled on the 
relationship Franz Boas had developed with his Kwakiutl interpreter, George 
Hunt, and proved just as successful, particularly in the case of Thomas (Golla 
1984: 133-134). In addition to providing documentation of many aspects of 
Nootka culture from a native" point of view, Thomas's contributions eventually 
more than doubled Sapir's collection of Nootka texts. Although Sapir worked 
on these materials fairly steadily during his Ottawa years, he was able to see 
only a small fraction of it into print until very late in his career. The three papers 
(1911e, 1924g, and 1929e), the short monograph (1915a), and the set of annota- 
tions (1916c) that are reprinted here represent nearly all of Sapir's published 
work on Nootka linguistics before 1939 (for his publications on Nootka eth- 
nography see 1911e, 1912a, 1913b, 1914b, 1915h, 1919e, 1921e, 1922y, 1922aa, and 
1925c, all in Volume IV). 

320 VI American Indian Languages 2 

"Some Aspects of Nootka Language and Culture" (1911e), subtitled "The 
Linguistic Relationship of Kwakiutl and Nootka," is one of two studies pub- 
lished together as the first fruits of Sapir s 1910 field work. (The other part, "The 
Nootka Wolf Ritual," appears in the full article reprinted in Volume IV.) 
Sapir s aim here is to provide Wakashan comparative linguistics with a firm 
structural basis. He outlines the basic sound correspondences between Nootka 
and Kwakiutl; he sketches the parallels in morphology and provides a number 
of cognate affixes; he discusses related morphological processes, particularly 
reduplication; and he broaches comparative syntax in a treatment of pro- 
nominal systems and conjunctions. He concludes with a list of 40 apparently 
cognate stems and affixes. 

The short monograph on "Abnormal Types of Speech in Nootka" (1915a) 
belongs together with Sapir s description of consonant symbolism in Wishram 
(in 1911g), his paper on male and female speech in Yana (1929d), and his experi- 
mental study of the psychology of sound symbolism ( 1929m) as an expression of 
his enduring interest in the points of contact between formal linguistic struc- 
tures, social categories, and psychological universals. This paper is sometimes 
cited in the sociolinguistic literature as a pioneering study of variation corre- 
lated with such factors as age, sex, class, and ethnicity. In fact, Sapir's subject 
here is not sociolinguistic variation as such, but a set of literary-rhetorical 
devices used in both formal and informal Nootka discourse to index the social 
categorization of the topic (having thus a clear link to Sapir's discussion of 
"recitative" in Southern Paiute myths, 1910d). Typically, these are regular pat- 
terns of phonetic alteration used in traditional narratives to identify the speech 
or actions of salient mythic characters or of socially stereotyped groups such as 
foreigners, deformed individuals, or outcasts. At least some of these devices 
are regularly used in conversational interaction for a variety of interpersonal 
ends, including affection, mockery, and social exclusiveness. 

"The Rival Whalers, A Nitinat Story" (1924g), a short text obtained from 
Sayach'apis in 1913, is provided with a word-by-word analysis in order to pres- 
ent a "serviceable" introduction to Nootka linguistic structure. Publishing an 
analyzed text in lieu of a full grammar was a device Sapir had earlier resorted to 
for Wishram ( 191 Ig: 673-677) and for Yana ( 1923m). In Sapir's hands what could 
easily be a mere pastiche of disconnected information is crafted into a tightly 
organized descriptive statement. Most major inflectional and derivational pro- 
cesses are explained (with several paradigms given in full), a wide range of 
morphophonemic alternations are exemplified, and the complex morphosyntax 
of Nootka words is thoroughly explored. Although superseded as a formal pre- 
sentation by the grammatical sketch in Nootka Texts (Sapir and Swadesh 1939) 
— the orthography of which is also fully phonemic — "The Rival Whalers" 
remains the best pedagogical introduction to Nootka linguistic structure. 

"Nootka Baby Words" (1929e) is a brief report on the Nootka baby-talk regis- 
ter — a small set of lexical substitutions, phonological alterations, special mor- 
phological devices, and other special features used by or in speaking to small 
children — and should be read in conjunction with the paragraph on the use of 

Eight: Wakashan and Salishan Languages 32 1 

the diminutive in baby talk in "Abnormal Types of Speech in Nootka" (1915a: 

During the 1930s two of Sapir's students, Morris Swadesh and Mary Haas, 
carried out further Nootkan field work under Sapir's direction (cf. Swadesh and 
Swadesh 1933). In the summer of 1934 arrangements were made for Alex 
Thomas to come to Yale for intensive work with Swadesh, with a view to prepar- 
ing some Nootka materials for publication. The eventual result, published in 
the year of Sapir's death, was a major volume on Nootka linguistics containing a 
selection of Sapir's (and Thomas's) texts, a grammatical sketch, and a list of 
lexical elements (Sapir and Swadesh 1939, Volume XI). A second volume of 
texts, largely Alex Thomas's, was published by Swadesh after Sapir's death 
(Sapir and Swadesh 1955). Even so, over half of the texts, many of considerable 
ethnographic interest, have remained unpublished, as have virtually all of 
Sapir's Nootka ethnographic notes. (See Volumes XI-XII.) 

Besides those specifically concerned with Nootka or Wakashan linguistics, 
the papers reprinted in this section include two that deal wholly or in part with 
Salishan languages. "Noun Reduplication in Comox" (1915f), a Straits Salish 
language, is based on data that Sapir collected during his first visit to the 
Nootka in 1910 in typically incidental fashion, from a Nootka man whose 
mother was Comox. The short list of "Salishan-Wakashan Comparisons," pub- 
lished posthumously (as part of Swadesh 1949), was apparently compiled by 
Sapir at the time of his Comox work or shortly afterward (see Golla 1984: 108). 
It is all that survives of Sapir's own lexical evidence for the "Mosan" hypothesis, 
the genetic linkage of Salishan, Wakashan, and Chimakuan originally pro- 
posed by Frachtenberg (1920: 295; see also Swadesh 1953: 26-28) and incorpo- 
rated by Sapir into his general classification of North American languages 
(1921a and 1929a). 

Sapir also prepared a phonetic retranscription of, and notes to, the anony- 
mous 18th century Nootka vocabulary published by Boas in "Vocabularies from 
the Northwest Coast of America" (1916c); this contribution, which appears in 
Volume XII, is a minor work, but not without interest. The vocabulary was 
taken at Nootka Sound, in the northern part of the Nootka dialect chain, 
whereas the dialect with which Sapir was familiar, that of Barkley Sound and 
Alberni Canal, was near the southern end of the chain. Sapir's notes include 
numerous observations on the differences between the two dialects and may be 
read as a preliminary essay on Nootkan internal diversity. 




Fast-Runner and Throw-away are chiefs of the Kwakiutl, and 
close friends. The latter makes up his mind to give a feast of salmon- 
berries, and sends his four attendants to invite his own clan and his friend 
Fast-Runner. After Fast-Runner has been assigned his seat and the 
feasting-songs have been sung, four canoes that Throw-away has been 
careless enough not to have cleaned out are brought in, filled with berries 
and boxes of grease, and put before Fast-Runner and his clan and three 
other chiefs. Fast-runner, instead of proceeding to eat of the food, lies 
down on his back and covers his face with a blanket, while his fellow- 
clansmen, observing that he is displeased, follow his example. After a 
long silence, Fast-Runner's attendant arises and tells his fellowmen of the 
dirty treatment accorded their chief, in that Throw-away has not washed 
out the canoes before using them as food-receptacles. Throw-away 
rebukes his friend for his haughtiness ; Fast-Runner expresses unwilling- 
ness to eat of the dirty food, and, to show his superiority in wealth, sends 
for his copper "Sea-Monster" and puts it into the fire. Throw-away 
retaliates by putting in his own copper, " Looked -at-askance," in 
order to " keep the fire burning." Fast-Runner meets this by sending 
for a second copper, " Crane," and putting this also into the fire so as to 
"smother it." Throw-away, not possessing another copper, tries to 
borrow one from his fellow-clansmen, but is unable to do so. Hence he 
virtually confesses himself beaten by his rival, and his fire " has died 
out." The guests thereupon leave. 

The following day Fast-Runner returns the feast, and sends his 
attendants to invite his friend Throw-away. The latter is assigned his 
seat, and the feasting-songs are sung. Then four feasting-dishes are filled 

' [The text as originally written down by Mr. Hunt has been kindly revised 
by Mr. Edward Sapir, whom I have to thank for having changed the system 
of spelling to that now adopted by Professor Boas in his Kwakiutl Texts of 
the Jesup Expedition Publications. For the explanation of letters and sounds 
see Vol. Ill of that series, p. 5. — Editor.] 


324 VI American Indian Languages 2 


with crab -apples, wild cherries, and grease, and put before Throw-away 
and his fellow-clansmen. Throw-away does not eat, but returns his 
friend's insult by saying that he will not taste the dirty food offered, and 
then sends his four attendants home. They soon come back, however, 
with the copper "Day-Face," which Throw-away puts into his rival's 
fire. Fast-Runner arises and says that his fire has been " extinguished." 
He then puts on the recklessness of the Fool-Dancer, and, with his 
father-in-law's leave, sends his attendants to the latter's house to break 
four of his new canoes. They return with the fragments. Fast-Runner 
puts these on the fire so as to " build it up " again, and wishes by the 
heat to drive away his friend, who is lying on the ground near the fire. 
The intense heat causes Throw-away great physical pain, but he does not 
flinch and holds his ground. After the blaze has begun to die out, he 
gets up and eats of the crab-apples, thus showing that his rival's deeds 
have not in the least affected him. Then the guests of Fast-Runner are 
directed by his attendants to leave, and they do so. 

Some time later Throw-away secretly informs his fellow-clansmen of 
his intention to give a winter-dance in order to outdo his rival. Fast- 
Runner, however, hears of this, and determines to do likewise. As pre- 
paratory to the winter ceremonial. Throw-away has his son and daughter 
"disappear," whereupon Fast-Runner, not to be behindhand, causes his 
two sons and two daughters also to ' ' disappear." At the approach of the 
ceremonial period, Throw-away has a Sea-Monster mask carved out for 
his daughter, who is to be a war-dancer, and a Grisly-Bear mask for his 
son. The "disappeared" children of Throw-away are "caught," songs 
are sung, the Sea-Monster and Grisly Bear perform the proper cere- 
monies, and a canoe is given to Fast-Runner. 

Fast-Runner then begins with his winter-dance and begs his people 
to stand by him in his attempt to shame Throw-away. The following 
day all are invited into his house to eat, whence they repair to the singing- 
house. Here they sing and dance continuously through the night. The 
next day the " disappeared " Grisly Bears and warriors of Fast-Runner are 
sent for and brought in four canoes. A slave is scalped, pursued, and 
butchered by the Fool-Dancers, Cannibals, and Grisly Bears, and his 
body eaten up by the Cannibals. Fast-Runner's recklessness in sacri- 
ficing a slave shows him to be a richer man than Throw-away, who is now 
clearly outdone. Fast-Runner gives Throw-away his slave's scalp. In 
the evening the Grisly Bears are ' ' tamed ' ' and the two war-dancers sing 
their secret songs. The attendant asks one of the war-dancers what she 
wishes done to her, and she requests that they be put into the fire. So 
the war- dancers are tied down on boards, a high wall of fire -wood is put 

Eight: Wakashan and Salishan Languages 



about the fire, and the daughters of Fast-Runner are ostensibly to be put 
inside to be burned. Two slaves are adorned just like the true war- 
dancers and tied down on boards. These are to be burned as substitutes 
for the war-dancers. They are told not to scream when put into the fire, 
as they will be struck dead if they do so, but otherwise will come back 
to life after four days. The slaves acquiesce and are burned to death, 
while the real war-dancers are hidden, being supposed to have been 
put into the fire. After four days they pretend to revive, the ashes of the 
two slaves having been preserved in boxes. Now Throw-away is com- 
pletely worsted. 

Throw-away goes off to fight against the Nootka ; but he and all his 
men are slain by them, only one man surviving to tell the tale. " This," 
says Mr. George Hunt, " is the true story of the two chiefs who were true 
friends in the beginning, and turned out to be worst enemies at the end." 

Go'kula^laeda g-a'lasa Kwa'g-ule lax Qa'logwise. Wa, 

They lived, it is first of Kwakiutl, at Crooked-Beach. Well, 

said, the the 

laE'm^awIse ^na'mok-aleda g-i'gama^yasa SE'riLlEmexa 

now it is said they were friends to each chief of the Sun tribe 


named Fast-Runner 
(long ago) 

other, the 

g-i gama'yasa 

the chief of the 

Lo'yalalawaxa Ts!ex- 

Hair-turned-up-in- named 
Front tribe 

^i'dEX'Lula. Wa, 

Throw-away Well, 

(long ago). 


then it is said 



then it is said 


they watched (knew) 
each other's 

la ga'la la ^na'mok-ala, 

now long time 

they were friends 
to each other. 




they loved each 





'na'xwaEm lae 
all, it is said, 

Wa, laE'm^lawiseda ga'lasa Kwa'gule xu'lsaxa ^riE'mxsa 

Well, then it is said the first of the Kwakiutl were downcast one 

^na'la. Wa, laE'm'^lawise TslEx^i'de ha'nak-axes ^nEmo'kwe, 

day. Well, then it is said Throw-away asked leave of his friend 

lax TsEx^wl'de. Wa, 
of Fast- Runner. Well, 

laE'm ne' 

then he told of his 


going to be a place of 
giving a feast with 

326 VI American Indian Languages 2 


qle'nEme qIa'mdzEkvva la'xa ga'la Kwa'kug-ula. Wa, 

many salmon-berries to the first Kwakiutl (clans). Well, 

he'x-^ldaEm^la'wTse TsEx^wi'de ^ne'ka : " ^ya, qast, la^mo'x 
right away, it is said. Fast- Runner said, " O friend ! Now that 

e'k'os wa'ldEmaqos qa e'k'lex^IdesEns g'o'kulotex qaxs 

good your word, so that may be happy tribe, for 

your our 

xE'nLElaex xu'Isa," ^ne'x-^lae. Wa, he'x-^idaEm^la'wise 

very they are he said, Well, right away, it is said, 

downcast," it is said. 

5 TslEx^i'de axk-a'laxes a^yi'lkwe qa e'x%itsE^weses g'o'kwe. 

Throw-away asked his attendants that they sweep out his house. 

Wa, g i'FEm^la'wise gwal e'kwasE^wa g'o'kwaxs lae'da 

Well, as soon as, it is said, finished was swept house when then the 

out the 

a^yi'lkwe mo'kwa yae'ltsEmtsa dEnE'me. Wa, la'xda^x"- 

attendants four hung about their cedar-bark Well, then they, 

waists the (belts). 

^lae Le'^lalaxa Kwa'kug-ule qaes gi'gama^e TslEx^i'de. Wa, 

it is said, to invite the Kwakiutl for their chief Throw-away. Well, 


g i'l^mese gax %e'laeLexs la'e axk'la^le TslEx^i'daxes 

as soon as came all in, when then asked Throw-away his 


lO a^yi'lkwe qa las e'tse^stasE%es ^nEmo'kwe TsEx%i'de. Wa, 

attendants that they go call again his friend Fast-Runner. Well, 

he'x-^ida^mesa mo'kwe a^yi'lkwa e'tse^staq. Wa, g-i'Fmese g-a'xe 

right away the four attendants again called Well, as soon as came 


TsEx^wi'de g'a'xeLa, la'e TslEx^i'de axkMa'Iaq qa las 

Fast-Runner came into then Throw-away asked him that he go 


k!wa'ga^lil la'xa ^nEqe'walelases g'o'kwe. Wa, g-i'l^mese 

sit down at the rear of his house. Well, as soon as 

in the house 

k!wa'g-a^ll}Exs la'asa kiwe'le klwe'^'lala dE'nx^ida, yi'sa 

he sat down in the then the feasters sang feasting- began to sing with 

house, when songs 

15 mo'sgEme klwe'la^layo q!a'mq!EmdEma. Wa, g-i'l^mese gwal 
four means of feasting songs. Well, as soon as they finished 

Eight: Wakashan and Salishan Languages 327 


klwe'^laleda k!we'laxs la'e LEle'LEmeda mo'ts!aqe xwa'xu- 

to sing the feasters, then were brought four (long) small 

when they into house 

xwaguma. Wa, laE'm hawe'xaEm ts!o'xwEgEntsE^waxs la'e 

canoes. Well, then never they were washed inside then 

when they 

qupIa'laxsElayowa qle'nEme q!a'mdzEk" laq. Wa, la- 

were used for pouring in many salmon-berries into Well, 


^me'se ^na'xwaEm ^naE'ngoyaleda mo'tslaqe xwa'xuxwagumaxa 

then all half full the four (long) little canoes 

qIa'mdzEkwe. Wa, la'^lae ax^e'tsE%a mo'sgEme dE'ngwats!e 5 

salmon-berries. Well, then it is were taken four (round) grease-boxes 

L!e"^na qa^s k!u'nq!EgEmaeda ^na'FnEmsgEme la'xa 

olachen-oil so that were put in the each one (box) into 

^na'l^nEmtsIaqe xwa'xwaguma. Wa, g-i'PEm^la'wise gwal 

each (long) small canoe. Well, as soon as, it is finished 

said, they 

k!u'nqasoxs la'e LE'lLElbEntsE^a xwa'xwagume qa^s le 

were put in, then it was lifted up at small canoe, so that went 

when each end the they 

ha'nx-d/.amolelEm lax TsEx^wi'de LE%e's ^nE^me'mote. Wa, 

was put down at Fast-Runner and his clan. Well, 

in front 

la'^laeda ^nE'mtslaqe ha'nxdzamolelEm lax O'dze^stalise 10 

it is said one (canoe) was put down in front at Wrong-around- 

the World 

LE%e's %E^me'mota Maa'mtagila, yixs SE'nLlEmae ^nE^me'- 

and his clan Maa'mtag-ila, while Sun tribe the 

motas TsEx^wlMe. Wa, la'^laeda ^nE'mtslaqe xwa'xwagum 

clan of Fast-Runner. Well, it is said one (long) small canoe 


ha'nx-dzamolelEm la'xa Ge'xsEme lax gi'gama^yase Ki'm- 

was put down in front at the Chiefs (clan) to their chief Met-by- 

k-EqEwede. Wa, la'^'laeda E'lxLa^e ha'nx-dzamolelEm lax 
(Chiefs). Well, it is said last one was put down in front at 


Lla'qwagila yixs gi'gama^yasa La'alaxsEnt!ayo. Wa, 15 

Copper-Maker when he was chief of the Breakers-(of-all- Tribes). Well, 

328 VI American Indian Languages 2 


g-i'l^Emla'^wise ^wi'lga^lllExs la'e TsEx^wi'de a'Em 

as soon as, it is said, all had been put then Fast-Runner only 

on floor, 

tIe'xbEtalel qa^s ^naxu'mdeses ^nEx^una'^e Lla'^ya. Wa, 

lay down on his and he covered his face blanket (of) black Well, 

back in his seat over with his bear. 

laE'm'^lawise TslEx^i'de axk'la'laxes a^yi'lkwe qa wa'xesexa 

then it is said Throw-away asked his attendants that they tell to go 

ahead the 

k!we'ie qa wa'g*es yo's^itses k'a'k'EtslEnaqe, qaxs 

feasters that they should eat with spoons spoons, for 

with their 

5 ^na'xwamaeda kiwe'ie qe'qEplEpElaxes k'a'k'EtslEnaqe. Wa, 

all the feasters carried in the folds of spoons. Well, 

their blankets their 

a'Em'^la'wisa ^na'xwa klvvei do'qwalax TsEx^wi'de, qaxs 

only, it is said, all the feasters kept looking Fast-Runner, for they 


^ma'lt!allmaaqexs ya'x-sE^mae na'qa^e la'xes k!wa^e'lase. 

recognized that bad the heart at his place of sitting 

on floor. 

Wa, a'Em^lawisa klvve'ie la ^na'xwaEm la tle'x-alila o'gwaqa. 

Well, only, it is said, feasters then all now lay down on also, 

the their backs 

Wa, he'lat!a^la gae'iExs la'e La'xulile a^yi'lkwas 

Well, that, however, long on floor, then stood up the attendant of 

it is said, 

lO TsEx^vvi'de. Wa, la'^lae ya'q!eg-a^}a. Wa, laE'm*la'wise 

Fast-Runner. Well, it is said he spoke. Well, then it is said he 

^ne'k-a, la'xes klwe'l^wEote : " V^, g'o'lg'Ekulot, we'g-a 
said to his fellow-feasters, "O fellow-tribesmen! goon 

do'qulaxg-in g-i'gamek-, la'xg-a gvvae'lasg-as la'xox 

see this our chief here at this his manner at this 

on floor 

g-o'kwasos ^nEmo'kwox TslEx^i'dex, yixs k-!e'sek- ^nex- 

house of his friend here. Throw-away, for he not says 

qa^s ha'^mx-^idexwa ^mo'^mxselayoxsa ^mo'xulax. Wa'lax'de 

that he eat these dirty things of this dirty (chief). Could not he 

15 la'Lox tslo'xulExsaxes le'loqlulela xwa'xuxwagumaxs 

have gone he washed out his feasting-dishes small canoes when 

Eight: Wakashan and Salishan Languages 



k-!e's^mex-dex gu'xalESElasa qla'mdzEkwe la'qwaq, qaxs 

not he began to pour the salmon-berries into these, for 

he'mEnala^ma'axEl t!e'lats!exa t.'e'tlElema (he'Em gwo^yo'seda 

they were all the time soaking-recep- what was (that is what he 

tacles for that 


meant the 

p!E'Lasde LE'^vva ma'leqasde LE'^wa mo'qwasde). Wa, yo'^mets 

1 hali- and the dried hali- and the (?) ). Well, this is 




gwo^yo's qa yo'sasoso'xda q.'wa'iobEse la'xg-exs la'xos 

wish that be eaten with spoons the soot now inside in 

lelo'qulelaqos TsIex^iM," ^ne'x-^lae. Wa, la'^lae he'x-^lda^me 5 

your feasting-dishes. Throw-away," he said, it is Well, it is said immediately 




La' xu Ilia. 

got up in the 




then, it is 



wa'ldEma la'xes a^yi'lkwe. Wa, la'^lae ^ne'k-a : 

word to his attendants. Well, it is said he said. 


he said 

^ya, qast, 
"O friend! 

gwa'ldzas xe'iileI LE'mlEmqIa'loL. Ha'aqos gwex-s q!e'q!ade 

don't you very talk proudly. That is you as if you much having 

qa'es wa'ldEmos. Wa'g-adza a'Em laxs gwa'exsdaasaos," 

on account word. Well, go on just to your wish," 
of your 

^ne'x'^lae. Wa, he'x-^IdaEm^la'wise TsEx^wi'de La'xuliia. 

he said, it is Well, immediately, it is said, Fast-Runner got up in the 
said. house. 

la'^lae ^ne'ka 

it is said he said, 

^ya, qast, Ts!Ex*^i'd, k-!e'sEn ^ne'x- 

' O friend Throw-away ! not I say 

Wa, 10 



that I 


eat the 


dirty man ! 

^ma'^moxselaqwe ha^mgi^la'yo g'a'xEn, 
dirty things given for food to me. 


Wa, la^me'sEn ^ma'^moxwalaLOL. A'la^meg'in 

Well, and now I shall meet your dirty Truly I am 



much having among you," 

a^yi'lkwe qa 



that they go 


he said, as 

then he 


sent his 

ax^e'dEx Lla'qwase Tsle'gese. Wa, 1 5 

get his copper Sea-Monster. Well, 


VI American Indian Languages 2 


g-i'l^Em^la'wise g'ax ho'gwcLcda mo'kwe a^yi'lx"sexs la'e 

as soon as, it is said, came went into house four attendants, then 


ts!as lax TsEx%i'de. 
gave it to Fast- Runner. 




la'^lae TsEx^wi'de da'x'^idxes 

then it 
is said 


took his 




then it is 
said he 


walked in 

la La'yabots lax 

and he went pushed it under to the 


lEqa'%alilas g'o'kwases ^nEmo'kwe. Wa, laE'm k'li'lxas la'xa 

fire in middle of house of his friend. Well, then he put out to 

the floor of the fire with it 

5 kiwe'lase. Wa, he'x-^idaEm^la'wise TslEX^i'de o'gwaqa ax^e'dxes 

his feast. Well, immediately, it is said. Throw-away also took his 

Lla'qwe LlEsaxEla'yo. Wa, laE'm^laxaa'wise La'yabots la'xes 
copper Looked-at-askance. Well, then it is said also he pushed it under to his 











in order 

laE'm^lae x'a'x"iq!as 

then it is 
said he 


it should 

tried to burn it 

die out. 


in his 






^nEma'x-Es l6^' lEqwi'las lEqwa', yi'ses Lla'qwax'de. 
just as if with he made fuel, with his past copper. 



then it 
is said 





^ya'laqases mo'kwe 

sent his four 

a^yi'lkwa qa las ax^e'd ^nE'msgEme Lla'qwase AdE'mgule. 

attendants that they go take one (other) his copper Crane, 
and (round) 

Wa, g i'l^Em^laxaa'wise g'a'xda^x" ho'gweLElaxs la'e tslas 


as soon as again, it is 
said, they 

went into house, 



la'xes gl'gama^c TsEx%i'de. Wa, a'Em^laxaa'wise TsEx^wi'de 

to their chief Fast- Runner. Well, only, it is said, Fast- Runner 


la La'yabolesas lax kIwe'lasdEma lEgwe'iases %Em6'kwe. 

went pushed it under to the feast-place fire of his friend. 

Wa, laE'mxaa kli'lxas lax lEgwe'lases ^nEmo'kwe. 

Well, then again he put out to the fire of his friend, 

with it 

Eight: Wakashan and Salishan Languages 



Wa, laE'm %Ema'x-Es lo^ dzn'mas lelao'xwasa Lla'LlEqwa 



just as if 

with he covered 

cost of the 


lax klwe'lasdEtna lEgwe'lts go'kvvases ^nEmo'kwe, qa 
to the feast-place fire of the house of his friend, so that 

k-!eo'ses xi'xsE^watsa lEgwe'le. Wa, he'^mes la'g-ilas 

it be nothing being burned iire. Well, that is reason of 

by the 

Le'gadEs kli'lxaxa lEgwe'lasa k!we'lase. Wa, la'La TslEx^i'de 

having name putting out fire of the feast. Well, but then Throw-away 

of the 

^nEma'xEs lo^ mo'xLalases Lla'qwa. Wa, laE'm ^mEx°stE%e'sa 5 

just as if 





his house. 


with he lighted his 
fire with his 

LE'^wa lEqwa'. 

and the fuel. 

copper. Well, then it is in place of the 

La'giias o'gwaqa LasLa'lases 

That is also pushing on 

reason of the fire his 

qa k-!e'ses k-!i'lx^ede k.'we'lasdEmas lEgwe'las 

for its not dying out the feast-place fire of 




then it is 

g-i'^Em^lawise TsEx^wi'de La'sLEntses 
Fast-Runner pushed on fire his 

as soon as, it 
is said, 



dak -la' lax 

asked for a 



to his 

*nE^me'mota Lo'yalalawa. 

clan Hair-turned-up- 


Wa, laE'l k'leo's Lla'qwas. Wa, 10 


it is there was no his copper. Well, 

la^me' ya'k-awe Ts.'Ex^i'de la'xeq. Wa, la^me' k'li'lxEkwe 
then he was beaten Throw-away in that way. Well, then was extinguished 

k!we'lasdEmax-de lEgwe'lts g-o'kwas. Wa, a'Em^la'wise la 
the past feasting-place fire of his house. Well, just, it is said, now 

ho'qawElseda kiwel la'xsde. Wa, la^me' hawe'xa yo's^Id la'xa 

went out the feasters the passed. Well, then they never ate wiih at the 


kIwe'ladzEm qla'mdzEkwa. Wa, la'^lae TsEx^wi'de ^ncx- qa^s 

given in feast salmon-berries. Well, it is said, Fast- Runner wished that he 

q!a'FaLElex na'qa^yases ^nEmo'kwe TslEx^i'de. Wa, laE'm'la'wise ic 
find out the thought of his friend Throw away. Well, then it is said he 

332 VI American Indian Languages 2 


axk!a'laxes a^yi'lkwe qa las Le'^lalax Le'lanEmx-dases 

asked his attendants that they go invite the ones that had been 

and invited by his 

^nEmo'kwaxa lE'nsde. Wa, laE'm^la'wise a^yi'lkwas Le'^'lalaq. 

friend the past Well, then it is said his attendants invited 

day. them. 

Wa, he'x-^idaEm^la'wise g-a'xeda Le'lanEme ho'gweLa. Wa, 

Well, immediately, it is said, came the invited ones came into Well, 


le'x'aEtn^la'wise TslEx^i'de k'!es g'a'xa. Wa, laE'm^^lawise 

only, it is said, Throw-away not came. Well, then it is said 

5 T.sEx'vvi'de '^ya'laqaxcs a^yi'lkwe qa las e'tse^staq. Wa, k!e's- 

Fast- Runner sent out his attendants to go and call him Well, not, 


^lat!a ge'x"^ideda a^yi'lkwaxs g'a'xae la'k'Elax TslEx^i'de. Wa, 

however, were long attendants when came following Throw-away. Well, 

it is said, the 

he'x'^idaEm^la'wise TsEx^l'de q!a'x"sldzeq. Wa, laE'm^lae 

immediately, it is said, Fast-Runner led his feet. Well, then it is 

said he 

%ex- qa las klwa'g-a'^lila Lla'salilases ^nE^me'mote. Wa, 

wished that he go sit down outside of his clan. Well, 


laE'm ma^e'ma^lokweda a^yi'lkvvas lax wax'sano'LEma^yas la'xa 

then there were two on his attendants at both sides of him at the 

each side the 

10 nExwa'Lala la'xa lEgwe'lases g'o'kwe. Wa, g-i'l'-Em^la'wise 

neighborhood to the fire of his house. Well, as soon as, it is 

said, he 

klwag-a^ll'laxs, la'as k!we'lg"a^'leda kiwe'lasa klwe'laya^layu 

sat down, then they began to sing feast-giver there were sung 

feast-songs the 

mo'sgEm q!a'mq!EmdEma. Wa, g i'FEm^la'wise 

four songs. Well, as soon as, it is 


qlwe'l^ededa k!we'lala dE'nxElaxs la'e ax'^e'tsE^a tse- 

stopped the feasters singing then were taken out crab- 

tsE'lwats!e mo'sgEm k!ek!i'myaxLa, qa^s ga'xe mExa'lelEm 

apple boxes four (round) boxes, and they came were put down 

on ground 

15 lax ma^sta'yasa a^we'LElasa klvve'layats! g'o'kvva. Wa, la'^lae 

at the near by to the inside of door feasting- house. Well, it is 

of the receptacle said 

Eight: Wakashan and Salishan Languages 333 


e'tled ax^e'tsE^vveda mo'sgEme dE'ngwatsle Lle'^na. Wa, la'^lae 

again were taken out four (round) grease-boxes (of) olachen- Well, it is said 


ax^e'tsE%eda ma'^lEx-La si'siul lo'quleia. Wa, he'Em^la'vvisa 

were taken out two on it double-headed feasting- Well, that, it is said, 

serpent dishes. 

^nEme'xxa na'na lo'qulela. Wa, he'Em^la'wisa ^nEme'x'La 

one on it grisly- feasting-dish. Well, also, it is said, one 


na'ne lo'qulela. Wa, he'Em^la'wisa ^nEme'x-La 

grisly- feasting Well, also, it is said, one 

bear dish. 

a'LanEm lo'qulela. HEm%ex-Lago la'e lo'Elqulelas 5 

wolf feasting-dish. That is all then the feasting-dishes of 

TsEx%l'de. Wa, laE'm^lawise guxtsla'laso^sa t.'E'lse. Wa, 

Fast-Runner. Well, then it is said were poured into wild Well, 

they with cherries. 

laE'm^lawise naE'nguya'leda lo'Elquleiaxa tsElx", 

then it is said were half full the feasting-dishes of crab-apples, 

la'e k!u'nq!Eqas6^sa Lle'^na. Wa, laE'm q6'qut!as. 

then were poured into olachen- Well, then they were full 

they with grease. with it. 

Wa, laE'm^laxaa'wise he'Em g"il k'a'x-^itso'^ses ^nEmo'kwe 

Well, then also, it is said, that first it was set before his friend 

TslEx^l'de. Wa, g-i'l^Em^a'wise ^wi'^la k-a'x"^Itsa 10 

Throw-away. Well, as soon as, it is said, all were set before 


io'Elquleiaxs la'e a^yi'lkwas wa'xaxa k.'we'le qa 

feasting-dishes, when then his attendants urged the feasters that 

wa'g'es yo's^eda. Wa, he'x'^idaEm^la'wise TslEx'^i'de La'xullia. 

they go eat with Well, immediately, it is said, Throw-away stood up in 

ahead spoons. the house. 

Wa, la'^lae ya'q!eg-a^la. Wa, la'^lae %e'k-a : "^ya, qast, 

Well, it is said spoke. Well, it is said said, "O friend! 

he he 

k-!e'seg-in he g-a'xile qEn plax^aLE'le la'xos kIwe'ladzEmaqos, 

not I that reason of that I should taste of your your feasling-objects, 


yO'Laxs a'laaqos mo'xula bEgwa'nEma, qfist, qaxs hawe'xaaqos 15 
you that really are dirty man, friend, since you never 


VI American Indian Languages 2 




washed inside my 

ha^maa'tslex, yi'xos lelo'qulelaqos," 

food-receptacle your feasting-dishes," 

^ne'x'^laexs la'e ^ya'laqases mo'kwe a^yi'lkwa qa las dadag'ill'lEla 

he said, it is then sent out his four attendants that they take out (some- 

said, when 


lax g-o'kwas. Wa, 
in his house. Well, 


not, it is said, 


long when 


immediately, it is said, 

g'a'xae e't!ed 
came again 

go and thing) 

la'x-da^xwa. Wa, 
went out. Well, 


they went into 

5 laE'm^lae dag-i'lqElaxa Lla'qwa '^nE'lgEmala. 

then it is carried in hand copper Day-Face, 

said they the 


^idaEm^la'wise TslEx^l'de da'x-^idxa 

mediately, it is said. Throw-away took the copper 

boleses lax lEgwe'lasa klwe'lase. Wa, 

it under to the fire of the feast-giver. Well, 

house of the 

k'li'lxax lEgwe'lasa gi'gama^e TsEx^wi'de 

put out the fire of the chief Fast-Runner. 




and he 



he'x - 





6 gwaqa 

also he 

Wa, la'^laeda 

Well, it is said the 

g i'gama^e 


house of the 


Fast- Runner 


got up in 
the house 


ya'qlega^e. Wa, 

and he spoke. Well, 

10 la'^lae ^ne'ka : "^ya, qast, la^mo'x k-!i'lx^edEn kIwe'lasdEmaqEn 

it is said 


"O friend! 

now extinguished my 
this is 

this my 

lEgwe'la. We'g-il la e'tla^lilEL k!wa'g-a^lllLOL, qEn do'qwale 

fire in house. Go on now will again will vou sit down so that I look 

in house 

will you sit down so that I 
in house 

qEn wa'ldEm o'gwaqa,' 

for my word also," 

^ne'x"^laexs la'e 

he said, it is said, as he 

xwa'sa la'xes 

was excited in his 

nulEmalaena^e. Wa, la'^lae ^ya'laqases mo'kwe a^yi'lkwa qa 
fool-dance. Well, it is said he sent his four attendants that 

las o'gwaqa d6'x%edEx axe'ilaxa lax go'kwases nEgu'mpe 

they go also look for what they might in the house of his father-in-law 


Wa, k-!e's^Em^lawise ho'qawElseda mo'kwe 

Well, not then, it is said, went out of four 

house the 




Eight: Wakashan and Salishan Languages 




attendants when 

lase la'xes 

place for he 
in house 




stood up in 

Moving- Lx)ad 


had also come 

as feaster. 


at his 

it is said he 






son-in-law ! 





they go 


be taken 

mo'tslaqa t!e't!Egu'na 

four (long) flat-bowed canoes 



lEgwe'los," ^ne'x'^lae. 

your fire 
in house," 

he said, it is 



he'x-^ldaEm^la'wise TsEx^wl'de ^ya'laqases mo'kwe a^yi'lkwa 5 
immediately, it is said, Fast-Runner sent out his 

Lo^me's ha^ya'Fa qa les 

young men that they 

and also 



flat-bowed canoes. 




break to pieces the 


not, it is said, 




ga'laxs . 

long when 




^we'g-iLEla'yoweda tso'gukwe t!e't!Egu'na. Wa, 

carried into house on broken flat-bowed canoes. Well, 

shoulders the 



then it is 
said they 

^mo'x''Lalayo la'xa kIwe'lasdEma lEgwe'la g-o'kwas Tsex- 

were piled up at the feasting-place fire in house the house of Fast- 


Wa, la^me' 


la x'ix^e'de 




Well, then 


now was burning 

fire in the 
house with it. 

Wi la^ 

me' TsEx%I 

'de wa'lac 

elax TslEx^i' 

de he'ltsasa 

Well, then Fast-Runner wished that Throw-away might run 

away from the 


Wa, la 



qo k!esL 


Well, then 

he thought it 
good that 

he be roasted 

if not will 






he will run 
away from it 


just, it is said, 


lay down on 
his back 



Wa, a' 



at his 


Well, just, it is said, 

his blanket 

in house. 


la tlE'mg 


Wa hewa 



(of) black 

now became scorched. 

Well, then never, it is said, 


VI American Indian Languages 2 


qlwe'nalElile TslEx^i'daxs wa'x-'^mae la pEne'^nakule Lle'sas 
moved in house Throw-away although became covered with skin of 

blisters the 


his knee 


at his 

a ene me 
just being 


face being covered 
with his 


past bear-skin 



Wa, g-i'FEm^la'wise k'li'lx-a^nakuleda lEgwe'laxs 

blanket. Well, as soon as, it began to die out the fire when 

is said, 

la'e TslEx^l'de klwa'g-a^lll qa^s yo's^ldexes lo'quleda tsElx". 

then Throw-away sat down and ate with spoon dish of crab- 

he out of his apples. 





^nEma'x'Es l6^ ne'laxs 

this is just if 

he showed 

k-!e'sae tslE'x-^aLEle 

not he became sick, 

gwe'x-^ldaasaq wa'x'^mae la lEgu'la. G i'Pmese gwal yo's^ededa 

what had been although he now he was As soon as finished eating with 

done him burnt. spoons the 

kiwe'laxa tsE'lxwe la'e tslE'lwaqaso^s a'^yi'lkwas TsEx%l'de. 

feasters the crab-apples then were praised by attendants of Fast- Runner. 

they the 

Wa, g'i'Fmese qlwe'Fideda a^yi'lkwaxs la'e ho'qawElseda k!we'le. 

Well, as soon as stopped speaking attendants then went out of feasters. 
the house the 

Wa, ga'laEm^a'wise k-leo's e'tled gwe'g'ilasa. Wa, laEm^la'- 

Well, it was long, it is said, nothing again way of doing Well, then it is 

lo Wise 




invited his 


^nE^me'moteda ^ Lo'yalalawaxa 

clan (were) the Hair-turnedup- 

in-Front at 

ga'nuLe. Wa, laEm^la'wise ne'lases ya'^wix-ilaexsda^e. 



then it is said he told that he desired to give a winter- 

immediately, it is said. 

^nE^me'motas ^mo'^las 

his clan 

were grate- 
ful for 


his word. 






then it is 
said he 




to find out 

that he 

la'g-iias ^nex* qa'^s he gwe'x-'^lde. 

for that said that he thus did. 

reason he 

1 5 qla'^le TsEx' 





his word. 

yak-a'masex TsEx%i'de ; 

might beat Fast- Runner ; 

Wa, he'x'^idaEm^la'wise 
Well, immediately, it is said, 

Wa, laE^m^Iaxaa'wise 
Well, then it is said, also he 

Eight: Wakashan and Salishan Languages 



o'gwaqa Le'^lalaxes ^nE^me'motaxa ga'nuLe. Wa, laEm'^la'wise 

also invited his clans at night. Well, then it is said he 

ne'iaxes ^'nE^me'motaxs lE^ma'e o'gwaqal ya'wix-ilalxa tslawu'nxe. 

told his clan that he will also will give winter- winter. 

dance in the 

Wa, laEm^la'wise ^nex* qa^s a'lagawa^eses hayo'te TslEx^l'de. 

Well, then it is said he wished that he be always equal rival Throw-away. 

to his 




laEm^a'wise gwa'ie wa'ldEmasexs la'e ho'qawElsa. 

then it is said were their words when then they went out of 




not, it is said, 





disappeared the 

ba'bagume 5 
boy the 

xuno'x"s TslEx^i'de LE^we's ts!a'ts!EdagEme xuno'x"s. 

child of Throw-away and his girl his child. 

Wa, he'x-'^'idaEm^laxaa'wise x-is^i'de bEgwa'nEme xuno'x"s 

Well, immediately, it is said, also disappeared man the child of 


ma^o'kwe ts!eMaqe sa'sEmaxa la'e 

two women children at then 


TsEx^wl'de LE^we's 

Fast-Runner and his 









it is said 




it was night when 

x'is^l'deda ^nEmo'kwe bEgwa'nEm xuno'x°s. 

disappeared one man 


then it is 

then it is said 

g-i'ta qae'. 

carve for him. 


danced winter- 
dance the 





hired the 

his child. 

Kwakiutl clans. 









he go 

Wa, laE'm^ae tsle'gese g-ita"-'yasa gitle'noxwe 


then it is 
said a 

sea monster carving of the 

qa Lo'gweltsa ts!a'ts!EdagEme xuno'x"s 

that it be the supernatural girl the child of 

treasure in house of the 


was to be war- 


if she 

should come 


should show 





when she 


then it 




VI American Indian Languages 2 








is said 

grisly bear 
was to be the 



he should show 



'wise TsEx^w 

i'de ho'Lelax 

wa'ldEmas. Wa, 


only, it is 

said, Fast-Runner was listen- 
ing to 

their words. Well, 




lax k I'm^'yaEnxLaq. 

Wa, laE'm^'la- 

it is said 


to the time of being about 
to meet them. 


then it 







is said 

sat down the 

Kwakiutl clans 

at the 



5 laEm^la'wise TsEx^wi'de k!wa'gElasEq, qaxs ^a'xwa^mae la'da 
then it is said Fast-Runner sat with them for all they went the 

be'bEgwanEme la'xa q!a'mdase. Wa, laE'm^lawise gwa'lExs 

men to the singing-place. Well, then it is said finished when 

la'e na'^nakwa. Wa, la'^lae ga'nuFlda, la'e k-ik-i'lnala ^wi'^leda 

then went home. Well, it is said night came, then tried to all the 

they bring back 

(the novice) 

q'.e'nEmula le'lqwalaLa^ya. Wa, laE'm^lawise ^na'x'^ldxa gaa'laxs 

many past tribes. Well, then it is said daylight came early mom- 

in the ing when 

la'e gwa'la. Wa, he'x-^idaEm4a'wise la da'da ga'la Kwa'g'ulxa 
then finished. Well, immediately, it is said, went took first Kwakiutl the j 
they the 

10 ma^o'kwe x'isa'ias sa'sEmas TslEx^i'de. 

two disappeared children of Throw-away, 


ga'xalesa da'xdaxa x'isa'laxde la'xa 

came on the who had taken having disap- on the 

Wa, g'i'l^Em^Ia'wise 

Well, as soon as, it is 


beach when 



beach those 


had her head cut 
off the 


the peared ones 

to'x^wide. Wa, la'^laeda 

war-dancer. Well, 


LE qases 

it is said the grisly bear struck with 



ge'ts!Eme la'xa Sva'lase tiEgu'na. Wa, g i'FEm^la'wise 

claws at the large flat-bowed Well, as soon as, it is said, 


gwa'lExs la'e g-e'x-^edayoweda t!Egu'ne lax TsEx^i'de. 

finished when then was given the flat-bowed canoe to Fast-Runner. 

Wa, laE'm^lawise gwal q!a'mtasE%a to'x%idaxs la'e 

Well, then it is said finished was sung for the war-dancer then 

when they 

Eight: Wakashan and Salishan Languages 339 


ho'x^wEsdesa qa^s la ho'gweL la'xa lo'bEkvve. Wa, la'^lae 

went up beach and went went inside to the (emptied one) Well, it is said 
they winter-dance 


ga'nuFldExs la'e nana'qamasE%eda da'nEme. Wa, laEm^la'wise 

night came when were brought to their caught ones. Well, then, it is said, 
senses by singing, the went the 

tsle'gese Lo'gwa^yasa to'x^wlde. Wa, laE'm^lae g-a'xustalileda 

sea-monster the supernatural war-dancer. Well, then it is said came up from floor 
treasure of the of house the 

tsle'gese la'xa ogwe'waliiasa g-o'kwe. Wa, he'Em^la'wise 

sea-monster at the rear of the house. Well, that, it is said, 

wa'lox%ede. Wa, laE'm gwa'le TslEx^i'de ya'vvixila. 5 

was all that Well, then finished Throw-away giving winter- 

was done. dance. 

Wa, la'4ae TsEx^wi'de Lla'yogulsa. 

Well, it is said, Fast-Runner changed with him 

on the ground. 

Wa, laE'mxa^e TsEx^wiMe ^ya'laqases a^yi'Ikwe qa 

Well, then also Fast- Runner sent his attendants that 

las Le'^lala Hvu^na'laxa la ga'ia ga'nuLa la'xes 

they go invite secretly in the now long night to his 


^nE^me'mote. Wa, gi'FEm^la'wise g*ax %!'^laeLExs 

clans. Well, as soon as, it is said, they came all into the 


la'e LEne'x'^ItsE^we t!ex-i'lasa ts!a'gats!e g'6x"s 10 

then was barred the door of the winter-ceremonial the house 

receptacle of 

TsEx^wi'de. Wa, laE'm4awise ne'le E'lkwas TsEx^wi'daxes 

Fast-Runner. Well, then it is said told the attendant Fast- Runner his 


^nE^me'motaxs lE^ma'e lal k!wa'iaixa hi'La dza'qwaltsa la'La 

clan that they go sit in the future evening of the future 

e'tledEl ^na'x'^IdEL. Wa, he'^mesexs lE^ma'e ^ne'k'es gi'gama^e, 

again morning. Well, that is when he said his chief 

yix TsEx%i'de qa^s we'g'iL e't!edEi mo'mas^idElxes ^nEmo'kwe 

that Fast-Runner that he will go will again will do harm to his friend 


TslEx^i'de qo lal ne'l^idELes x-e'xisala. " Wa, laE'm^la'wisEns 15 

Throw-away if will go will show disappeared " Well, then we are told to 

his ones. 


VI American Indian Languages 2 


gwa'laial qEns g-o'xHvedeLEnsaqEk- qa^go ha^ala'lasoLo," 

be ready that we shall help hitn if he will be faced by 

his rival," 

^ne'x-^ae. ^na'xwaEm^'la'wise ^nE^me'motas ^ne'kExs lE^ma'e %i'^la 

All, it is said, his clan said they all 

he said, it is 


gwa'laia. Wa, laE'm^lawise gwa'le wa'ldEmasexs la'e ho'qavvElsa. 

are ready. Well, then it is said finished their words they went out of 

when the house. 

Wa, laE'm^awise ^na'xwa ga'x'staeiaxs ga'lae ho'qawElsa. 

Well, then it is said all went to bed first they went out of 

when (as soon as) house. 

5 Wa, laE'm^awise a'l^Em tslEx-^i'dxa 

Well, then it is said they soon awoke at 

Wa, he'x-^idaEm^a'wise 

Well, immediately, it is said, 

la lE'nsa. 

now next day. 



Wa, g-i'l^Em^a'wise gwa'lExs la'e Le'^lale TsExVi'daxa 


and their 


a la 

ga'la '^na'laxa 



long day on the 
[late in) 



la kwa'seda 

went washed them- 
selves the 


g i'ng inanEme. 

and the 


Well, as soon as, it is said. 




Fast- Runner the 

. gwe'gudza LE'%a tsle'daqe Lo'^ma 

winter-dancers and the women and the 

10 las he'yasEla la'xa ts!a'gats!as 

they go breakfast at the winter-ceremonial 

and house of 








g'i'FEm^la'wise gwal ha^ma'pExs lae'da a^yi'lkwe o'paiaxa 

as soon as, it is finished eating when then the attendants whispered to 

said, they 

*na' xwa 

all the 

gwe'gudza qa la.s la'xa 

winter-dancers that they go to the 

qia'mdase la'xa a'Lle. Wa, 
singing-place in the inland. Well, 

bEgwanEme ^wi'^la q!wa'g"a4ii 



immediately, it is said, the 

le a'Le^sta. 

15 g-i'FEm^la'wise %i'^la 

as soon as, it is said, all 



and they went around 

la klusE'lsExs la'eda 

went to sit down on then the 

the ground, when 

to the 






Eight: Wakashan and Salishan Languages 341 

gade dE'nx^Itsa ma^ltsE'me q!a'mq!EmdEmasa to'x^wlde. 

masters began to sing the two songs of the war-dancers. 

Wa, he'^mesa ma%sE'me qIa'mqiEmdEmasa na'ne. Wa, 

Well, also the two songs of the grisly bears. Well, 

g'l'PEm^la'wise gwa'la ne^na'gade dE'nxElaxs la'e 

as soon as, it is finished song-masters singing when then 

said, the 

TsEx%i'de ya'qleg'a^a. La'^lae ^ne'k'a: " '^ya, ^ne'^nEmokwa', 

Fast- Runner began to speak. Then it is said, "O friends! 

said he 

laE'ms %i'^lal qlwa'lax'alxwa ga'nuLex, qas e'k-!eqEleL6s 5 

now you will all will dress this night, that may be happy 


qaE'n wa'idEmLex," 





on account word," 

he said, 


it is said 

all the 

of my 

it is said. 

gwe'gudza na'^nax^meq. Wa, la'^lae ^ne'k-a : " He'LEnu^x" 

winter-dancers answered him. Well, it is said said, " This we shall 


gwe'laLe " ^ne'x-^lae. Wa, la'^lae he'x-^ida^ma mo'kwe a^yi'lk" 

shall do they said. Well, it is said, immediately the four attendants 

thus," it is said. 

qa's^id qa^s la xa'set!alaxa le'ladEnokwe. Wa, g-i'FEm^la'wise 

walked and went asked to wash those who had Well, as soon as, it is 

they their bodies dances. said, 

la'da mo'kwe a^yi'lkwaxs la'e ho'x^wult'.eda qIa'mtlEsde, 10 

went four attendants then went out of the song-experts, 

the woods the 

Wa, he'x-^idaEm^a'wise xwa'nal^ededa , %a'xwa gwe'gudza 

Well, immediately, it is said, got ready all the winter-dancers 

LE%e's ts!e'daqe LE%a' ^'na'xwa g-i'nginanEme. Wa, 
and their women and all the children. Well, 

laE'm^lae k-iki'lnala. Wa, a'FEm^la'wisc gwa'lExs la'e 

then it is said they tried to bring Well, just, it is finished when 

they them back. said, they 

^na"^nakulaxa gaa'la. Wa, he'x'^idaEm^la'wise la'da mo'kwe 

daylight came early Well, immediately, it is said, went the four 

in the morning. 

a'^yi'lk" ax^e'dxa mo'tslaqe awa' tIe't.'Eguna qa LE'nkwesesa 15 

attendants took the four (long) great flat-bowed and tied them together 

canoes with the 


VI American Indian Languages 2 


mo'ts.'aqe g-i'lsgilt!a dzo'xuma. 

four (long) long poles. 

sao'kwc. Wa, g"i'FEm^la'wisc 

boards. Well, as soon as, it is said, 


^na'xwa gwe'gudza laq 

all the winter-dancers 





and the 



Lo^ '^na'xweda le'laenenokwe 

and all those having dances 

5 LaE'm^lae a'Em kludzl'l la'xa lo'bEkwe. Wa, 

Then it is said just sat in house in the winter-dance 



long the 


point of 

Wa, la'^lae pak-EE'ntso*^sa 

Well, it is said were covered over 
they with 

gwa'lExs la'e ho'x^aLExseda 

were finished then went into canoes 

laE'm^lae k-!es la'da nena'ne 

not went the grisly 

la'xa da. 

to the taking. 


not, however, 
it is said, 

then it is 




house (emptied 

Eyo'x^ed lax 
back at the 


Crooked- Beach. 

LlEma'esasa lo'bEkwe. 

beach of the winter-dance house 
(emptied one). 






came, it is said, 


then it is said 


opposite northern 
end of point 

ha'ng-a^les lax 

being in front at the 
of beach 

^nE'me na'ne 
one grisly bear 

hax"ts!a'gvvewe lax 

lay down inside in the 

a gwiwa^yasa 

bow of 

of bov 


one (long) 



10 Wa, la'^'laxaeda ^nE'me na'ne he'xat! gwa'la lax a'gwlwa'^yasa 

Well, also, it is said, one grisly that also thus in the bow of the 

(other) bear 


one (long) 





Wa, laE'm^lae k'!es ya'wix'a'lag"iiExsa 

Well, then it is said not moved about in canoe, 

(mother canoe). they 

laE'm'lawise q!a'mt!etsE^weda ma^lo'kwe teto'x^wlda. 

then it is said were sung for the two war-dancers. 

g-i'l'Em^la'wise gvval q!a'mtasoxs la'e 

as soon as, it is said, finished were sung for then 

they when 

^walatsa'yokwe q!a'k*6 bEgwa'nEma qa^s le q!o'dEg-ivvaletsa 

stout slave man that he 'go keep bow with a 

15 dzo'xume la'xa ^wa'lega^yasa t!et!Egu'ne. Wa, g i'PEm^la'wise 

pole in the large one among flat-bowed Well, as soon as, it is 

was sent a 


said, he 

Eight: Wakashan and Salishan Languages 343 







stood up in 
the canoe, 

then a 



Frowned -upon 

haxsE'mlExsaq. Wa, la'^lae ^nEmo'kwase '^na'xulalese, ax^e'dxes 

fell uf>on him Well, it is said his friend Warrior-all-round- took his 

in canoe. the-World 

k'le'LEnxe qa^s k-!e'ltse^stales mak'inxE'ndalaxox awE'nxa^yaxsEns 

sharp-edged and cut around with it close to this edge of our 

(knife) he 

SE^ya'x. Wa, g i'FEm^la'wise Ifi'lgEwe k!e'La^yasexs 

hair. Well, as soon as, it is said, met where it was cut 


la'e ne'xodEx Lle'tsEma^yas xo'msdasa q!a'k-owe. Wa, 5 

then pulled off the scalp of the head of the slave. Well, 


la'^lae LaxSvuito'dxa qla'k'o. Wa, g'i'FEm^la'wise la^sta'xs 

it is said they pushed out of slave. Well, as soon as, it is said, he went into 

canoe the he the water, 


ga'xaeda ha'Emats! LE%a' ne^na'ne LE^wa' no'EniEmala ho'qawEls 

came the cannibals and the grisly bears and the fool-dancers went out 

la'xa lo'bEkwe qa^s la qa'qayaxa qia'kaxs la'e dzExwae'sEla 
from the winter-dance and went pursued the slave then ran along 

house they when he 

lax LlEma'esas Qa'logwise. Wa, he'^mes la qaqaya'atsa 

at the beach of Crooked- Beach. Well, and that is then place of pur- 

suing of the 

ne^na'naq LE^vva' no'EnlEmala qa^s sa'k'apleq. Wa, la'^'lae 10 

grisly bears and the fool-dancers that they tried to Well, it is said, 

him spear him. 

na'papleda wao'kwaq. Wa, la'^laeda ne^na'ne Ui'qaplEses 

tried to throw several at him. Well, it is said the grisly bears tried to strike 

stones with their 

xexElya'yowe laq. Wa, we'gaa^latla lax ^na'lanegwesas 

claws at him. Well, not he reached, to the beach on south side 

however, of 

Qa'IogwIsaxs la'e ya'x%alesa. Wii, laE'm 1e4 la'xeq. Wa, 

Crooked-Beach when he fell down dead Well, then dead there. Well, 

on beach. he is 

he'x'^ldaEm^'la'wisa no'EnlEmala sEsox"sE'ndxa q!a'kox"de, 

immediately, it is said, the fool-dancers cut up in pieces the former slave. 

Wa, g-i'l^Em^'la'wise gwal sa'kwasoxs la'eda ha'Emats!a jr 

Well, as soon as, it is said, he finished was cut up when then the cannibals 


VI American Indian Languages 2 



ho'qunts!es la'xa LlEma'ese qa^s la ha^mx-^l'dEq. Wa, 

went down to to the beach that they go eat him. Well, 


la'^lae o'gwaqa'^ma ne^na'ne LE%a' no'EnlEmala ha^mek'la'la 

it is said, also the grisly bears and the fool-dancers asked to eat 

of it. 

Wa, wel?.x"dze'^lae ge'x-^idExs la'e '■"wi'^laq qaxs 

Well not it took, it is said, long when 


(ate) him 
all up 


ma^itso'gugiya^eda bEgwa'nEme ha'^ma'pxa q!a'k'oxMe. Wa, 

120 were the men eating the former slave. Well, 

la'^laeda ne^na'ne, yi'xa 

it is said the grisly bears, that is the 




(taken ones). 



le'nEmax SE^ya'x-das 

took away his past hair 

ha^mg'i'^layo laq. Wa, 
was given to Well, 

food them. 

qa^s le'nEmapleq. 

and took it from 

they each other. 






and that is 




was killed by 

TsEx%l'de qae's hayo'te TslEx^i'de. Wa, la^me' ya'k-aweda 

Fast- Runner on account rival Throw-away. Well, then was beaten 

of his 



g-i'gama^e Ts!ex^ 

chief Throw- 

"I'de la'xes ^nEmo'kwe 

away by his friend 


Fast- Runner. 

Wa, g-i'l^Em^la'wise gwal a'mleda ma'^le' 

Well, as soon as, it is said, finished playing the two 


grisly bears 
with the 

sa'bEkwe SE^ya'sa 

skinned hair of the 

q!a'k-oxMaxs la'e TsEx%l'de 

former slave, then Fast- Runner 

took the 

sa'bEkwe SE^^ya' 
skinned hair 

qa^s la ts!as lax 

and he went gave it to 



Wa, la'^lae 
Well, it is said he 

^ne'k-a : " Wa, qa'sta, 
said, " Well, friend. 


now this 

qo'sL," ^ne'x-^lae. 

will be he said, 
yours," it is said. 

Wa, g-i'P'mese gwa'lExs la'e 

Well, as soon as he finished then 


went out of 

15 ^wl"leda k i'm'yaxdaxa x i'sa'lax-de. Wa, g-i'FEm^a'wise 

all those who had surrounded former disap- Well, as soon as, it 

the peared ones. is said. 

Eight: Wakashan and Salishan Languages 



dza'qwaxs la'e a^'yi'lkwas TsExHvi'de qa's^Ida. Wa, k-!e's'lat!a 

evening came, then attendants of Fast-Runner walked Well, not, however, 

the (to call). it is said, 

g-a'laxs g-a'xae %i'^laeLeda ga'la Kwa'kugula. Wa, 

long when came all into house the first Kwakiutl clans. Well, 

he'x'^idaEm^la'wise ya'lasE^weda ma^e' 

immediately, it is said, were tamed the two 




g-i'FEm^a'wise gwa'iExs, 

as soon as, it is said, finished 

they with it, 


then it 
is said 


sang their sacred 
songs the 

grisly bears. 


teto'x^wlda. Wa, g-i'pEm^la'wise qlwe'FedExs la'e dE'nx^ededa 5 
war-dancers. Well, as soon as, it is stopped speaking, then began to sing 
said, they the 

ne^na'gade. Wa, g-a'xeda ma^lo'kwe teto'x^wid ^yEx%uit!a'- 




came out and 
of house, they 

Well, came the 


went around 
in house 



at the 

war-dancers dancing as they 


fire in middle 
of the 






then it is said 



singing with 


one (round) 






immediately, it is said, 


ya'yaq!antemiie la lax 
speaker of the house went to the 


of the 





then it is said 


wuLa'sE%eda ^nEmo'kwe to'x^wid la'xes axe'xsdESE%a. Wa, 

was asked one war-dancer for her desired (thing). Well, 

he'x-^ldaEm^lawise to'x^wide %e'ka: " Wa'laqeleg'anu^x" 

immediately, it is said, 



qEnu^x" lEqwi'lasE^e la'xwa lEgwe'lex," 

that we 

be made fire 
[put into fire] 

into that house-fire," 

la'^Iaeda ya'yaq!antemile e'talas la'xcs 

it is said speaker of the house repeated it to his 


" We desire 


she said, it 
is said. 






wa'ldEmasa teto'x^wide. Wa, he'x-^idaEm^^lawisa mo'kwe a^^yi'l.x^'.s 15 
word of the war-dancers. Well, immediately, it is said, the four attendants of 


VI American Indian Languages 2 


TsExSvi'de ya'q!eg-a^ia. Wa, laE'm4awise ^ne'k-a: "^ya, pepExala', 

Fast-Runner spoke. Well, then it is said they said, "O shamans 


wc'gadzaxins ^nE'mplEna na'nageg-exg-axg-a wa'ldEmg-asg ins 

let us for one time obey this desire of our 

friends here. 




go on 

ax^e'dEx ma^Exsa' ts!a'ts!ax"- 

take two flat short roof- 

SEtna qEns we'gl yiLEdzoda'lasEk- laq," ^ne'x-^lae. Wa, 

boards that we go on tie them on to them," they said. Well, 

it is said. 

5 laE'm^lawise Le'^IalasE%a ma^lo'kwe ba'bEbaklwa, yix 

then it is said were invited two warriors, that is 

QEnx%ida'yowe LE^e's ^nEmo'kwe 
Frowned-upon and his friend 

^na'xulalese qa 

Warrior-all-round- that 

ga'xes yiLEdzo'ts la'xa ts!a'ts!a6x"sEme. Wa, 

they come 

tie them on 

to the 

ately, it is said, 










g"a xe 



were taken 


put them down 

short roof-boards. 


two flat 





short roof-boards 

la'xa ogwewali'iasa g-o'kwe. 
in the rear of the house. 


then it is said 

were taken the 





naEriLEdzoda'yowe la'xa ^na'FnEmxsa tsIa'tslaox^sEma. 

were laid on their backs on the 

each flat 

then it is said a 

g-i'Itla dEnE'm 

long cedar-bark rope 

yaLa'4ayosa ts'.e'daqe to'xw^ld 

instrument of women war-dancers 
tying of the 

Wa, laE'm^laxaa'wise he' Em 

Well, then, it is said, also, that 


and the 

short roof-board. 

ax^etsE^wa qa 

was taken for 


short roof-boards. 


thus was done to the 

1 5 Wa, laE'm^lawise gwa'lExs la'e 

Well, then it is said they finished when then 

were taken 


other one. 


Eight: Wakashan and Salishan Languages 347 


tE'mg-ik° lEqwa' qa^s qElxasusta'lasE'^we lax awi'^s- 
blocks fire-wood and they were piled up at the around 

tasa lEgwi'le. Wa, a'^mese la nExts!E%e'da 

the fire. Well, and only then was in centre the 

lEgwl'le laq. Wa, laE'm^lawise ha'lsElaEm^la heltsla'pElcda 

fire in it. Well, then, it is said, hardly could look over a 

g'i'ltlExsde bEgwa'nEm lax xixtsla'laq. Wa, laE'm gwa'lala 

tall man in the putting head Well, then it is ready 

out into it. 

qae'da ma^lo'kwe teto'x^ld qo lal latslo'^yoLo laq. 5 

on account two war-dancers if will will be put in into it. 

of the then middle 

Wa, he'^maaxs la'e ^ne'k-a teto'x^wide qa^s laxLa'nowe 

Well, that when then they said the war-dancers that should be put 

they on top 

la'xa lEgwi'le. Wa, la ax^e'tsE%a ma^lExsa' ts!a'ts!aox"sEme 

on the fire. Well, then were taken two flat short roof-boards 

la ^na'^nEmax iyaak" LE%a' la nELEdza'yaatsa teto'x^wlde, Wa, 

made just like also the now place of lying on war-dancers. Well, 

back of the 

la'^lae ax'^e'tsE^a q!wa'xe, Wa, la'^laxae ^na'^nEmax-iyaak" 

it is said were taken hemlock- Well, then also, it were made just like 

branches. is said, they 

LO^ qeqExIma^yasa a'laklala to'x^wlda. Wa, laE'm- 10 

also hemlock head-rings true war-dancer. Well, then 

the of the 

^lawise qEximda'yo la'xa q!a'q!Ek*o ma^lo'k" tsle'daqa. 
it is said were tied around to the slaves two women. 


Wa, laE'm^lawise la ne'nELEdzodayo la'xa ts!a'ts!Ets!aox"sEme. 

Well, then it is said now were laid down on on the short roof-boards, 

they their backs 

Wa, laE'm^lawise yii^eda'yoweda gi'ltle dEnE'm laq lax 

Well, then it is said were tied the long cedar-bark to in 

ropes them the 

gwa'laasasa ma^lo'kwe teto'x'^wida. Wa, laE'm^lawiseda 

same manner as two war-dancers. Well, then it is said 

the the 

ma^o'kwe ba'bEbaklwa ^ne'x'xa ma^lo'kwe ts!e'daq q!a'q!Eka : 15 
two warriors said to the two women slaves. 


Vf American Indian Languages 2 


" ^ya, sa'sEm, gu'no gwaLlEXLa'lalaxo qa'so lal lax'La'not 

"O children! do not scream if you will you will be 


put on top 






on this 



that when 

you not 

gwaLlEXLa'laLOL, wa, la'LEs mo'plEnxwa^s^Eml k"!es g'ax 

you will scream, well, then you four (times) day not come 


q!ula'x'^idELOL. Wa, g-i'pEml^its gwaLlEXLa'laLoL, wa, 

you will come to life. Well, as soon as you will you will scream, well, 

5 la^me'sEnu^x" kwe'xaplELOL qa^s lelE'^'laos. Wa, laE'ms XEk'la'l 

then we shall strike the so that you die. Well, then you will stay 

nape of your necks away 

IcIe'^U la'xaq," ^ne'x-^lae. Wa, laE'm^lawise ^na'xwa ya'qleg'a^e- 

will be by this," they said, it Well, then it is said all spoke 

dead is said. 

da ^we'^woselaga ts!e'daq q!a'q!Ek-a. Wa, la'^lae ^ne'k-a: 

poor women women slaves. Well, it is said said, 


"We'g-a a'Em ha'^lilalax laE'mx- iak!we'masg-anu^x" ne^na'qek* 

" Go on only do it quickly this we are strong in our hearts 

qEnu'^x" k*!e'se gwa'LlEXLa'la qEnu'^x" k-!e'sel ga'ial qEnu'^x" 

for we not scream so that we shall not shall be that we 


10 ga'xei e'tledEl q!ula'x-^ida," ^ne'x'da^x^^Iae. Wa, laE'm^lae 

shall come shall again come to life," they said, it is said. Well, then it is 

said they 

gwa'llla. Wa, la'^laeda no'Enlamala Le'^lalaso^s TsEx^wi'de 

were ready. Well, it is said the fool-dancers were invited by Fast- Runner 







La'g ililaxa la ya'gudzayaatsa teto'xSvlde, qa*s 

war-dancers that 


carry up the now place of lying tied on 
board of the 

Lax"LE'ndEs la'xa lEgwi'le. 
put it on top on the fire. 

ga'xeda no'EnlEmalaxs 

came the fool-dancers when 

1 5 gwe'gudza. Wa, 

winter-dancers. Well, 







as soon as, it is said, 


stood up in 
house the 

xo'lexulila. Wa, 

then it is said they were all confused Well, 

(running about). 

Eight: Wakashan and Salishan Languages 




it is said the 



fool -dancers 


took up in 




in the 

rear of house 



la'xa go'kwe. 

in the house. 



it is said 

turned around 

at the 



went around in 
house with them 


door of the 






it is said 





went to rear 


it is said 




put down the 

heyak ili'lEla la'xa helk*!ote'waliie. Wa, 

went back to at the right-hand side. Well, 

inner room 

a'lak-!ala teto'x^id qa^s q!ula'Fedeq. Wa, la'^lae La'g-a^lllaxa 

true war-dancers and hid them. Well, it is said took up the 

they they 

ma^lo'kwe q!a'q!Ek*o qa^s le laltla'lii la'xa gEmxote'walile. 







it is said 

slaves and went 



went around with 

went out at the 

left-hand side. 






it is said 






was stopped 
^nEmo'kwe he^stali'lElaya 





at the 


and the 

was taken around 




to the 





rear of house 


door when 






at same time 








were on each side 
of the 






la"^lae La^na'kulamatsE^wa 

it is said were placed upright 

one after another the 

ma%'x"de qIa'qiEkoxs la'e la'xxana. Wa, laE'm'^lae havve'xa 

two former slaves when then were put on Well, then it is never 

they top (of pile). said they 

gwa'LlEXLalaxs la'e we'qumaxa. Wa, laE'm^lae lelE'^la. Wa, 

screamed when were shoved Well, then it is were dead. Weil, 

they down. said they 



350 Vf American Indian Languages 2 


gi'FEm^la'wise q!u'lx-^idExs la'e ax^e'tsE%a ma^tsE'me 

as soon as, it is said, burned to ashes then were taken two 


xa'xExatsEma qa^s g-a'xe ha'nEmg-a^lllEm la'xa ogwewali'le. 

small boxes and came were put down on floor at the rear of house, 


Wa, he'^latla le'da ^nE'msgEmeda o^sta'lile. Wa, la'^lae 

Well, that, however, was one at the door. Well, it is said 

it is said, 

ax'-e'tsE^a gi'lt!a klipLa'lae qa^s k!ip!e'tsE^we xa'lxEqasa 

were taken long tongs and they were picked up the bones of 

5 ^naFnEmo'kwe q!a'k-o qa^s le k-!ipts!a'layo la'xa ^nal^nE'msgEme 

each (person) slave and then were put in into the one to each (round) 

they with tongs 

xa'xatsEma. Wa, g-i'FEm^la'wise ^wl'^lts!axs la'e yikuyE'ntsE^wa 
box. Well, as soon as, it is were all in then they were covered 

said, they 

yi'ses ye'yik"ya^e. Wa, la'^ae ha'nga^lllEm la'xa ogwewali'lasa 

with their covers. Well, it is said were placed on at the rear of the 

they floor 

ts!a'gats!e. Wa, la'^lae mo'plEnxwa^se ^na'las he gwae'lExs 

winter-dance Well, it is said, four times day day of that being thus in 

house. house 

la'e ya'laqweda teto'x^wide. Wa, laE'm qlula'x-^ldbola. Wa, 

then sang their war-dancers. Well, then they pretended to Well, 

sacred songs the become alive. 

10 laE'm^Iae a'lax-^ld la ^ya'k'owe TslEx^i'de la'xeq. Wa, la'^lae 

then it is truly then was beaten Throw-away after that. Well, it is said 


TslEx^i'de ^nex- qa'^s le wl'naxa Mo'tsladxwe g*a'sa la'xa 

Throw-away said that he would make war Nootka going at the 

go upon the through 

dzE^la'lasa ^nE'mgese. Wa, la'^lae te'nox%ed la'xa wa. Wa, 

lake of the ^nE''mges. Well, it is said poled canoes on the river. Well, 


g'i'FEm^la'wise la'gaa la'xa t!ex i'las TiE'se la'e yu'dux"sEnd 

as soon as, it is went to the road of Nootka then in three pieces 

said, they Inlet they 

LE'mx'^idxes ^ya'^yats!e qa^s la ^we'k'ilk-ilaqexs la'e e'k'!e^sta 

split their canoe and went carried it on their they went up 

they shoulders when 

15 la'xa nEg-a', Wa, g i'FEm^la'wise la'g'aa lax was TIe'sc 

to the mountain. Well, as soon as, it is came to the river of Nootka 

said, they Inlet 

Eight: Wakashan and Salishan Languages 






tlE'mx-^ldxes ^a'^yatsle qa^s yo'Ix-^ide la'xa 
sewed together their canoe and they drifted down at the 


la'x'SE^yod la'xa vvaxs 

as soon as, it is arrived at at the river 

said, they mouth 

sex"sale'sEla la'xa tlo'kwaxs la'e a'Em 

paddled through at a narrow passage then only 

when they 

la'EnaLlEme. Wa laE'm^lae 

arrows. Well, 







were startled 




were shot at by 


then it is was 

said killed 




was alive 


and his 











it is said 

that is the 

^nEmo'kwa 5 



reason of know- 
ing of the 







were killed. 





^ya'k-owe Ts!Ex*l'de la'xeq. Wa, 
was beaten Throw-away after that. Well, 

laE'm la'ba. 

then end. 

Editorial Note 

Originally published in Boas Anniversary Volume. New York: G. E. 
Stechert, 108-136(1906). 

This paper, a contribution to the Festschrift honoring Boas on the 25th anni- 
versary of his doctorate, was Sapir's first published work. Student records at 
Columbia University (Murray and Dynes 1986) show that Sapir enrolled in 
Boas's introductory course on "American Languages" in 1903-04 while he was 
an undergraduate in Columbia College, and that he continued with a second 
course in 1904-05 during the time he was enrolled as a Master s candidate in 
Germanic Philology. According to the Columbia catalogue for those years, the 
"translation and grammatical interpretation of Indian myths" formed part of 
the work of both courses, and it is entirely possible that Sapir edited "The Rival 
Chiefs" for this purpose. The text was one of those transcribed by George Hunt. 
Boas's native Kwakiutl collaborator, and the original manuscript is preserved 
among Hunt's Kwakiutl materials in the Columbia University Library. 

352 VI American Indian Languages 2 

George Hunt, a native speaker of Kwakiutl, gathered and transcribed large 
quantities of ethnographic and Hnguistic material for Boas during a 40-year 
association. Boas first met Hunt during his initial field trip to the Kwakiutl in 
1886. In 1893 Boas arranged for Hunt to visit the World Columbian Exposition 
in Chicago as part of a delegation of Kwakiutls, and their close collaboration 
dated from that time. To facilitate their work Boas taught Hunt to write 
Kwakiutl phonetically, and Hunt began collecting texts for transmittal to Boas 
for editing and publication, with two of the resulting monographs explicitly co- 
authored (Boas and Hunt 1902-05, 1906). Boas described Hunt s work in some 
detail, including his orthographic practices, in his introduction to The Religion 
of the Kwakiutl Indians (1930: ix-xviii). See also Helen Codere, "George Hunt 
and Boas" (in Boas 1966: xxviii-xxxi). 

Sapir, impressed by the productiveness of the Boas-Hunt collaboration, 
established similar working relationships with several native speakers, includ- 
ing Pete McGuff (Wishram), Tony Tillohash (Southern Paiute), Albert San- 
doval (Navajo), and Alex Thomas (Nootka). The orthographic practices of 
these men later formed an important part of Sapir s famous discussion of the 
psychological reality of phonemes (1933c). 

Some Aspects of Nootka Language and Culture 


The Linguistic Relationship of Kwakiuti and Nootka 

The Wakashan linguistic stock is divided into two main branches, the 
Kwakiuti and the Nootka or Aht; the former embraces Kwakiuti proper, 
Xaisla, and He'ttsa^q", the latter Northern Nootka (from about Cape Beale north 
to Cape Cook on the west coast of Vancouver Island) and Southern Nootka or 
Nitinat (south of Cape Beale to Cape Flattery). By careful comparison of the two 
Wakashan branches one can in part reconstruct a Wakashan "Ursprache," but the 
actual differences between Kwakiuti and Nootka are in fact very great; they differ 
perhaps as much as Slavic and Latin. 

[16] As regards phonetics, Kwakiuti and Nootka, while both showing charac- 
teristic Northwest Coast features, differ rather considerably. The sonant or inter- 
mediate stop series of Kwakiuti is absent in Nootka, Kwakiuti p and b for instance 
being replaced by Nootka p. Besides the ^-series, which Kwakiuti and Nootka 
possess in common, Nootka has a c-series, which is doubtless derived from the 
Kwakiuti and Wakashan /:--series, which in turn Nootka lacks; thus Kwakiuti g- 
and k- are cognate with Nootka tc, k-f with tcf, and x- with c. There is no / in 
Nootka, n corresponding to both Kwakiuti / and n. The velars q! and x, while 
somewhat infrequently found in Nootka, are not the regular Nootka represent- 
atives of Kwakiuti q! and x; q! has developed into a peculiarly harsh and choky 
glottal stop, which I write ^, x into a strangulated-sounding h which I write //, 
these two consonants respectively resembling Arabic 'ain and ha; ordinary ^ 
and h are also frequently found in Nootka. As regards phonetic processes, 
Kwakiuti and Nootka agree in allowing no initial consonant clusters in words; 
initial Kwakiuti and Nootka ^m, ^'n, ^ w, ^'y, and Kwakiuti ^/ are undoubtedly 
related to ordinary Kwakiuti and Nootka m, n, w, y, and Kwakiuti / as are 
Kwakiuti and Nootkap.^ t!, kf, iJ, tsf, qf, Kwakiuti k-!, and Nootka tcHo non- 
fortis Kwakiuti and Nootka/?, t, k, l, ts, q, Kwakiuti k-, and Nootka tc. In both 
Kwakiuti and Nootka certain derivative suffixes "harden" the final consonant of 
the stem; thusp, q, and i, become Kwakiuti /7.^ qf, and ^7, Nootka /;.^ ^\ and ^'y. 
The "softening" of Kwakiuti seems to be represented in Nootka by but a few 
stray phonetic processes. Syllabically final glottal stops and glottally affected 
consonants — such as -^/ and -p! — which are common in Kwakiuti, are entirely 
absent in Nootka. Medial and final consonant clusters are not as freely allowed 
in Nootka as in Kwakiuti, / often serving in Nootka to lighten them (cf. Nootka 
-qEmit, 'round thing', with Kwakiuti -^Einl 'mask'). All final vowels and 

354 VI American Indian Languages 2 

Stopped consonants in Nootka are aspirated. Peculiar to Kwakiutl is the change 
of A:-stops to spirants (x, x", x-) before consonants, whereas in Nootka they 
remain; in this point Nootka seems more archaic than Kwakiutl. 

In general morphology Kwakiutl and Nootka are quite similar, [17] despite 
numerous differences of detail. In both the stem is, as far as its meaning allows, 
indifferently verbal or nominal and one or more suffixes are required to give 
rise to definitely verbal or nominal complexes; in Nootka a suffixed-'/' is often 
used to substantivize a verb form. Both Kwakiutl and Nootka are absolutely 
devoid of prefixes, most of the elaborate grammatical mechanism being carried 
on by means of suffixes, to a lesser extent by means of initial reduplication, and, 
in Nootka, consonantal changes. The suffixes of Nootka and Kwakiutl express 
similar ideas and are used in more or less parallel fashion, though the number 
of suffixes that are etymologically related form but a small percentage of those 
found in either; so far about ninety Nootka suffixes have been discovered that 
are entirely or in part cognate to Kwakiutl suffixes. Examples of local suffixes 
shared by Kwakiutl and Nootka are: Kwakiutl -o-^yo 'in the middle', Nootka 
-^'win'; Kwakiutl-A7e<7"'in the corner', Nootka -nikw-; Kwakiutl -atCls 'down 
river', Nootka -atis; Kwakiutl -tsfo 'in', Nootka -ts!o'; Kwakiutl -k-E 'top of a 
box', Nootka -tci' 'full'; Kwakiutl -!a^ 'on the rocks', Nootka -!a^a'\ Kwakiutl 
-es'on the beach', Nootka -is; Kwakiutl -/7'in the house', Nootka-//; Kwakiutl 
-xs 'in a canoe', Nootka -qs, -lans. A few examples of body-part suffixes are: 
Kwakiutl -Ids 'cheek', Nootka -as; Kwakiutl -xo 'neck', Nootka -as-Haul 
'chest'; Kwakiutl -dp! 'neck', Nootka -dpfaf 'back'. Important temporal ele- 
ments held in common are: Kwakiutl -l 'future', Nootka -^dq-L, -^il; Kwakiutl 
- x-^ld 'inceptive', Nootka -ci-L. There are some striking agreements in verbify- 
ing derivative suffixes, as: Kwakiutl -lexst 'to desire', Nootka -Hh" 'to try to 
get', -st!iH^ 'to have as goal'; Kwakiutl -!a 'to go in order to', Nootka -fas; 
Kwakiutl -k-!dla 'to make a noise', Nootka -^en' (= Wakashan *-q!Ela); 
Kwakiutl -g-a^i 'beginning of a noise', Nootka -^aL (= Wakashan *-q!a^l)\ 
Kwakiutl -qlES 'to eat', Nootka -lis; Kwakiutl -nuh' 'to have', Nootka -nak\ 
Examples of nominal suffixes are: Kwakiutl -aano 'rope', Nootka -d^nul 'long'; 
Kwakiutl -gas 'woman', Nootka -^aqs; Kwakiutl -asde 'meat', Nootka -act' 
'dried meat'; Kwakiutl -mis 'useless', Nootka -mis 'mass'; Kwakiutl -p!e-q [18] 
'stick, tree', Nootka -p.'U 'long board-like object', -q- 'tree'; Kwakiutl -(x)^'Enx 
'year, season', Nootka -q'^itcH" 'year', -fitcH" 'season'. On the whole it seems 
that Nootka has a rather larger number of derivative suffixes than Kwakiutl, 
many quite special ideas being expressed by means of suffixes where there seem 
to be no Kwakiutl equivalents. A few examples are -a/ 'blanket'; -'m//' 'son';-^^ 
'daughter'; -Htut 'to dream of; -!d^il 'to ask for as a gift in a girl's puberty 
ceremony'; - tfota'' 'to give a potlatch for'; -yaqn'^ 'to sing a song'; -Hl 'to begin 
to sing a song'; -lini 'to give a feast of; -Hd' 'to buy'. 

Both Kwakiutl and Nootka make use of two kinds of reduplication, one in 
which the first consonant, first vowel, and second consonant of the stem are 
repeated, and one in which only the first consonant and vowel are repeated; the 

1. ! denotes a "strengthening" of the preceding consonant. 

Eight: Wakashan and Sallshan Languages 355 

former type is employed in forming iteratives, the second in forming plurals or 
distributives and with certain suffixes (such as Kwakiutl -fa, Nootka -fas 'to 
endeavor, to go in order to'; Kwakiutl -^ydla 'to go to look for'; Nootka -////" 'to 
try to get'; Nootka -kfok'" 'to look like'). In Nootka the repeated vowel is in all 
cases the same as that of the stem, in Kwakiutl the second type of reduplication 
has a definite vocalism {e in some cases, a in others) in the reduplicating sylla- 
ble. In Kwakiutl verb stems ending in vowels inserts- after the first, k- after the 
second syllable of the iterative, while Nootka iteratives of like form insert l and 
y; Nootka sd- 'to crawl' forms iterative sd' Lsdtc, -tc being probably identical 
with Kwakiutl -k-. One other striking resemblance of detail between Kwakiutl 
and Nootka may be noted: both Kwakiutl diminutives in -Etn and Nootka nouns 
in -kwin' 'toy' require reduplication of the stem. 

In regard to pronominal development there is considerable difference 
between Kwakiutl and Nootka. While there is, practically speaking, but one 
series of personal pronominal suffixes in Kwakiutl, there are three in Nootka 
(represented, for second person singular, by -e^its, -k\ and -sok'), of which the 
second and third are etymologically related: the first Nootka series is used in 
indicative forms of verbs, the second in subordinate clauses, interrogatives, 
and possessive forms of nouns, while the third seems to be confined to [19] 
certain modal forms. Kwakiutl has distinct forms for first person plural 
inclusive and exclusive, while Nootka has only one form for both. Pronominal 
objects are, to at least a considerable extent, incorporated in Kwakiutl; in 
Nootka, however, only in the case of the first person (second series) of the 
imperative. A great degree of complexity in pronominal forms is brought about 
in Kwakiutl by the combination of the pronominal affixes with syntactic (sub- 
jective, objective, and instrumental) and demonstrative elements. Nootka has 
none of this syntactic and demonstrative complexity of the pronoun, but a 
series of forms is found built up of the second pronominal series and an element 
-tc implying that the statement is not made on the authority of the speaker. 

Almost all Nootka and Kwakiutl words are noun or verb forms, there being 
almost no particles properly speaking. Such apparent Nootka conjunctive and 
case particles as ^onoi. 'because', ^oyi' 'when, if, and ^okwil 'to' are mor- 
phologically verb forms built up of a stem^o- 'a certain one, thing' and deriva- 
tive verbifying suffixes. There is, however, in Nootka a syntactically important 
conjunctive element ^ani' 'that' to which may be appended pronominal affixes 
of the second series and which may perhaps be considered a particle in the 
proper sense of the word. The "empty stem," Nootka ^6-, is cognate with 
Kwakiutl 6- 'something,' which, however, is used primarily in noun forms. 
Other Wakashan "empty stems" are: Nootka ^ap-, ^am-, Kwakiutl dps-, used 
chiefly in forming nouns of body-parts that occur in pairs, and Nootka hit-, hi- 
'to be at', Kwakiutl he- 'that'; peculiar to Nootka is hin, hit- (before "harden- 
ing" suffixes) 'to be or do (as indicated by derivative suffix)'. 

In regard to vocabulary Kwakiutl and Nootka differ greatly. Considering the 
very striking morphological agreements between them it is somewhat disap- 
pointing to find comparatively few stems held in common. It is highly impor- 
tant, however, to note that many of these are rather colorless in content and 

356 Vf American Indian Languages 2 

thus hardly to be suspected of having been borrowed in post-Wakashan times. 
Such are Kwakiutl ^nd-, Nootka ^nds 'daylight'; Kwakiutl g-dl-, Nootka tcdn- 
'to be first'; Kwakiutl ax- 'to do, be', Nootka 'o//- 'to be'; [20] Kwakiutl we-, 
Nootka wi-, wik" 'not'; Kwakiutl g<?, Nootka qe' 'a long time'; Kwakiutl ^riEm-, 
Nootka ^'nup- 'one'; Kwakiutl gwe- 'thus', Nootka qwi- 'to be or do thus'; 
Kwakiutl .s^o-, Nootka 50- 'you'; Kwakiutl ek-!-, Nootka '77c/- 'above'. Thus Dr. 
Boas' first announcement in 1890 of the close relationship between Kwakiutl 
and Nootka has been confirmed in every way by new evidence. 

Editorial Note 

Excerpt from "Some Aspects of Nootka Language and Culture", originally 
published in the American Anthropologist 13, 15-28 (1911); the article appears 
in full in Volume IV. Reprinted by permission of the American Anthropologi- 
cal Association. Sapir's ms. corrections have been incorporated. 

For a much fuller presentation of the data on which this paper is based, see 
the "Wakashan Comparative Vocabulary," based on Morris Swadesh's rework- 
ing of unpublished materials of Sapir (now in the Boas Collection, American 
Philosophical Society Library), published for the first time in Volume XIL 
Sapir also returned to comparative Wakashan in a late paper on "Glottalized 
Continuants in Navaho, Nootka, and Kwakiutl, with a Note on Indo-Euro- 
pean" (1938b, Volume II), where he showed that the glottalized sonorant con- 
sonants (v, vv, /, and the nasals) of Wakashan have arisen from the coalescence 
of ^ and h with following or preceding plain sonorants. Recent work on com- 
parative Wakashan, which has largely been confined to studies within the 
Kwakiutlan and Nootkan branches, is surveyed in Jacobsen (1979). 

Abnormal Types of Speech 
in Nootka 

An interesting linguistic and cultural problem is the use 
in speech of various devices implying something in regard to 
the status, sex, age, or other characteristics of the speaker, 
person addressed, or person spoken of, without any direct 
statement as to such characteristics. When we say "big dog 
make bow-wow" instead of "the dog barks," it is a fair inference 
that we are talking to a baby, not to a serious-minded man of 
experience. Further, when we hear one use "thee" where 
most would say "you," we suspect that we are listening to an 
orthodox Quaker. In neither of these cases is there an explicit 
reference to a baby as person addressed or to a Quaker as person 
speaking. Such implications are common in all languages and 
are most often effected by means of the use of special words or 
specific locutions. Thus, in Nootka there are special words 
used in speaking of obscene matters to or in the presence of 
women; a number of "baby- words" also exist. Generally it 
is the speaker or person addressed that is thus signalized, but 
it is quite possible, though less frequent, to thus imply some- 
thing also in regard to the third person. A more specialized 
type of these person-implications is comprised by all cases in 
which the reference is brought about not by the use of special 
words or locutions, that is, by lexical, stylistic, or syntactic 
means, but by the employment of special grammatical elements, 
consonant or vocalic changes, or addition of meaningless 
sounds, that is, by morphologic or phonetic means. 

358 Vf American Indian Languages 2 

To enumerate all the possible types of person-implication 
expressed in language, from the point of view of resulting 
classifications of human beings, would lead one far afield. 
Two types, however, seem to stand out most prominently — 
those referring to sex-discrimination and to rank-discrimination. 
Several languages make a distinction between words or forms 
used by males and such as are restricted to females. Such a 
distinction, for instance, is made by certain Eskimo dialects, 
in which, at least in earlier times, according to Boas,^ final p, 
t, k, and q^ were pronounced by the women as the correspond- 
ing nasals m, n, V, and rj. In Yana, an isolated linguistic stock 
of northern California, the forms used by the women, whether 
in speaking to one another or to males, differ from the fuller 
forms used by the latter in the unvoicing of final vowels; final 
-na {-hi in Southern Yana), a common noun ending, is replaced 
by aspiration in the speech of the women, who further lengthen 
final vowels to express the interrogative, while the males suffix 
an element -n. Most languages that make such sex distinctions 
differentiate the sexes as speakers. In Yana, however, a further 
discriminating factor is the sex of the person spoken to, in so 
far as the men in speaking to the women use the forms charac- 
teristic of the latter. 

More widespread in language seems to be a discrimination of 
forms according to the rank or social status of the person speak- 
ing, addressed, or spoken of. Here belong the etiquette forms 
characteristic of several East Asiatic and Indonesian languages, 
by which the social grading of the speakers as inferiors or 
superiors in reference to one another is clearly reflected in their 
speech. An analogous American instance is the use in Nahuatl 
of reverential forms to imply respect to the person addressed or 
spoken of. These are morphologically nothing but indirectives 
or causatives in -lia, -Ha, or -Ilia with reflexive pronominal 
prefixes; "he sleeps" is thus more politely expressed as "he 
causes himself to sleep." Here belongs also the use in so many 
European languages (French, German, Russian, and others) 
of second or third person plurals, instead of the more logical 
second person singulars, in speaking to people with whom one 

* Handbook of American Indian Languages, Bulletin 40 of Bureau of American Ethnology, 
1911. p. 79. 

' See Phonetic Key at end of this paper. 

Eight: Wakashan and Salishan Languages 359 

is not on the most intimate terms. This usage has its parallel 
in Yana, where brothers and sisters address each other in the 
pluraP; other Californian examples of a similar nature have 
been given by Goddard^ and Kroeber.^ 

These preliminary remarks are intended merely to indicate 
the general class of linguistic phenomena to which belong the 
more specialized Nootka examples to be given presently. At 
the same time they will serve to render these latter less glaringly 
bizarre by providing them with parallels of a more general 
character. The data here presented were chiefly obtained in 
November, 1910, in the course of ethnologic and linguistic re- 
search for the Geological Survey of Canada among the Nootka 
Indians of Alberni canal, Vancouver island; the informant was 
Dan Watts, the young chief of the Hopdtdas'atH" tribe. Further 
data on this subject were obtained in the winter of 1913-14 from 
Alex Thomas, a young Indian of the TsUcd'atH" tribe of the 
same region. 

It is possible and often customary in Nootka to imply in speech 
some physical characteristic of the person addressed or spoken 
of, partly by means of suflSxed elements, partly by means of 
"consonantal play." Consonantal play consists either in alter- 
ing certain consonants of a word, in this case sibilants, to other 
consonants that are phonetically related to them, or in inserting 
meaningless consonants or consonant clusters in the body of 
the word. The physical classes indicated by these methods 
are children, unusually fat or heavy people, unusually short 
adults, those suffering from some defect of the eye, hunchbacks, 
those that are lame, left-handed persons, and circumcised males. 

In speaking to or about a child it is customary to add the 
regular diminutive suflfix -'is to verb or other forms, even 
though the word so affected connotes nothing intrinsically 
diminutive; affection may also be denoted by it. The -'is 
comes before temporal, modal, and pronominal suffixes. Thus, 
the normal qwlstci" "do so!" {qwis- "to do thus;" -tci' second 
person singular imperative, "go and . . , !") is changed to 
qwis'istci" "do so, little one!" when speaking to a child. 

' Sapir, Yana Texts, University of California Publications in Arncrican Archaeology and 
Ethnology, vol. 9, 1910, p. 95, footnote 139; p. 101, footnote 150. 
» Goddard, Kalo Texts, ibid., 1909, vol. 5, p. 143, footnote 185. 

' Kroeber, The Languages of the Coast of California north of San Francisco, ibid., 1911, 
vol. 9, p. 321 (Porno). 

360 ^f American Indian Languages 2 

Similarly, qwisma^ "he does so" {-ma'^ third person present 
indicative) is changed to qurls'ismd' when one is speaking about 
a child. In speaking about oneself or others when addressing a 
child, it does not seem to bo customary to use the diminutive 
suffix except to show affection at the same time. Thus, the 
word waidLaH "I am going home" {wal- "to return home;" 
-ciL- inceptive; -an "I") may be changed to walciL'tsan "I 
am going home, little one" when addressed to a child for whom 
one wants to show love, but this form would not be used in 
speaking to a child that is a stranger. As might be expected, 
diminutive verbal and other forms occur in lullabies, in some 
of which the child is represented as speaking about itself. 
Thus, in a lullaby supposed to be sung by a whale mother to 
its child, occur the words 'oH'^'eso¥ '.emiti' ("my) little name 
is" ('o//"- "to be;" -'is- diminutive; -ok^ "of, belonging to;" 
lemiW "name"). Some people were said by Dan to have the 
habit of using the diminutive suffix in order to belittle others, as 
though the persons addressed or referred to were of no more 
importance than children as compared to themselves. If a 
chief does this to too great an extent, he is set down as haughty. 

In talking to or about fat people or people of unusual size, 
the suffixed element -aq' is used in a manner analogous to the 
diminutive -'is. Thus, the normal hinV cihwe' in^ "he comes, 
it is said" {hin- "empty" verb stem "to be, do;" -V-, shortened 
form of -tw* "to come;" -a"L- inceptive; -we'in^ quotative) 
becomes hinVciLaq'we'in^; 'otsdtciLma' "he goes to it" {'o- 
"empty" noun stem meaning "something;" -tsa- "to start for, 
go to;" -tciL- inceptive, used after vowels; -ma' third person 
present indicative) becomes 'otsatciLdq'ma\ Other examples 
are: ha'gkwaq'ma' "he, clumsy one, eats;" (ha'w- "to eat;" 
-okw- intransitive verbal suffix); and ha'okwdqiVnak'^ "did you 
eat, fatty?" {-iV tense suffix denoting past time; -na- inter- 
rogative; -/c' second person singular). 

People who are abnormally small are spoken of in forms 
with the diminutive suffix; moreover, in such cases, all sibilant 
consonants (s, is, is!; c, tc, id) become palatalized c- sounds 
(s, th, t^!; compare, for s, Polish s and Sanskrit f ; for ts, com- 
pare Polish c), which sound acoustically midway between s- 
and c- sounds; the diminutive -'is itself becomes -'is. Thus, 

Eight: Wakashan and Salishan Languages 361 

hinVciLweHn* "he comes, they say" is changed to hinVkiL- 
'Uwe'in* "he, Httle man, comes, they say." These §- forms are 
also used to refer to small birds, such as sparrows and wrens. 
Sometimes a meaningless ^ is added to the word, as in wiMh"^ 
t6Hauk' from wikdn" toHauk^ "I am not afraid" (wik- verb stem 
"to be not;" -d//" first person singular present indicative; ton- 
verb stem "to be afraid;" -uk\ diphthongized to -auk' because 
of preceding a- timbred h, intransitive suffix). We shall meet 
this consonantal change again further on in another connexion. 

Quite analogously to dwarfs, are addressed or spoken of those 
suffering from some defect of the eye. Under this category 
are included cross-eyed people, those who squint, and such as 
have one eye run out, but not the blind. Here again the 
diminutive suffix is used, with the added feature that all s- 
sounds and c- sounds are converted into the corresponding 
voiceless lateral stops or spirants (s and c become I; ts and tc 
become l; ts ! and tc ! become lI); the diminutive -'is itself 
becomes -'il. This style of speech is termed daddtckHn'- "to 
talk in sore-eyed fashion" (cf. Lladdtck'sul "one-eyed per- 
son"). Thus, qimsma' "he does so" is changed to qwll- 
'ilma'. Similarly, tditciLma' "he cuts" {tdi- "to cut;" -tciL- 
inceptive; -ma' third person present indicative) becomes 
LliLJL'ilma'. A full-grown Indian named Sammy (or Se'mi 
as pronounced in Nootka), who is cross-eyed, is referred to 
as ie'mi'il "little cross-eyed Sammy." Another Indian of the 
same tribe, To'mic, who has only one good eye, is, in parallel 
fashion, referred to as To'miVil "little one-eyed To'mic." It 
should be remarked that such people, particularly when adult, 
are apt to become offended if addressed in this fashion, and that 
one would not use such forms in their presence unless with the 
express purpose of showing contempt or of teasing. As will be 
seen again later on, L!aL!dtck!in^ forms are used also in referring 
to the deer^ and mink. Thus, the mythological Mink, tc/dsti- 
mits'miC "Mink-son," is generally referred to as Lldltimifinit'. 

Hunchbacks (k.'wdpi') are also addressed or spoken of in 
forms provided with the diminutive suffix, a further peculiarity 
in these being the change of ordinary s- sounds and c- sounds 

* Deer is associated with sore eyes also in other Indian mythologies. An Ojibwa einmplo 
may be found in P. Radin, Some Mylks and Tales of the Ojibwa of Southeaatern Ontario, Gooloji- 
cal Survey of Canada, Memoir 48 (No. 2, Anthropological Series], p. 3 (episode d). 

362 VI American Indian Languages 2 

to peculiar thickish c- sounds, pronounced with the lower jaw 
held in front of the upper; the diminutive -is appears as -He. 
We may represent these c- sounds by c. In this hunchback 
talk qivisma'^ becomes qwicHcma\ Other examples are: ydtcuk'- 
'i^ma' "he is walking" (yats- "to walk;" -uk'- intransitive verb 
suffix); tc.'dtck'^miniH'^'icma'^ "all of them are" {tddick"- "to 
be all;" -miniH'^- plural); and tcldxciLHcma' "he spears" 
{ts.'ax- "to spear"; -oil- inceptive). Here again these distinc- 
tive forms are generally avoided when in the presence of hump- 
backed people, for fear of giving offence. However, a hump- 
backed child who is well known to the speaker would hardly 
take offence and would be addressed as described. Or, if an 
old humpbacked woman is good-natured, c- forms may well be 
used when she is about, as though to show that she is happy 
and not easily ruffled. Here the notions of contempt and 
affection commingle. 

In speaking of lame people the diminutive suffix is again 
used, this time in its normal form. Besides this, the meaning- 
less element lc or lci is inserted in the body of the word some- 
where before the diminutive suffix, its exact position apparently 
depending on the whim of the speaker. Thus, hinini'aLma" 
"he comes now" {hin- "empty" verb stem; -ini- "to come;" 
-ai- determinative suffix marking point of time, "now"; -ma' 
third person present indicative) becomes hininiLcVitsfaLmd' 
(diminutive -'is and -'ol regularly combine to form -itslab) 
or hiLcnini'its.'aLma" "the lame chap is coming." Similarly, 
the verb tdltd'aLma' "he cuts now" (inceptive -tciL and -'gl 
combine into -tci'aL) is changed to tditciLditsIaLma' when a lame 
person is spoken of. The word tfa'ne'is'i" "the child" {t.'a'na- 
"child, son, daughter;" -is diminutive suffix, i causing preceding 
a to become umlauted to e; -i" nominalizing element, about 
equivalent to our definite article) becomes tlaLcne'is'i' "the 
young lame fellow," which may be used in speaking to children. 

In speaking of or to left-handed people the diminutive suffix 
is used in its normal form, besides which the meaningless element 
few" is inserted after the first syllable of the word. Thus, ydV- 
aLma' "there now he is" (ydl- "to be there;" -'ai- and -ma' as 
above) becomes ydltcn'^'itslaLma' (-'is and -aL combine to 
form -iislaL) "there now he is, poor little left-handed chap!" 

Eight: Wakashan and Salishan Languages 363 

Similarly, from sukwi'ttLma' "now he takes it" {su- verb stem 
"to take;" -kwiL inceptive suffix, changed to -kwi- before -ai) 
is formed sutcH^kwiL^itslaLma'. The diminutive suffix may also 
be omitted. Examples are: hitcH'^nin^ from hinln^ "to come"; 
and tHtcHHciLan from tlitcxLaH "I throw it down" {Hi- "to throw;" 
-tcib inceptive suffix; -an first person singular indicative). Such 
a form as the last might be appropriately used in speaking to a 
left-handed person that one is well acquainted with and who will 
not take offence at being thus twitted. It is customary, particu- 
larly for jokers, to use these left-hand forms also in talking 
about bears, who are supposed to be left-handed.^ 

In speaking of or to circumcised males, forms known as 
H'icVklin'' "to make cC- sounds" are used. In these the mean- 
ingless element cV is inserted after the first syllable of the word. 
One of the TsUcd^atn'^ Indians, named T!6xmis "Slaying-while- 
moving-from-beach-to-beach," is often humorously referred to as 
Tfdctxmis because of his having been born circumcised. Other 
examples of this class of forms are: hict'ninnna' from hininima' 
"he comes;" and /idcf'pA;'" from hd'ok'"" "to eat." 

Similar phonetic changes are made in forms used to refer to 
one or two classes of individuals characterized by some mental 
quality. Thus, greedy people are addressed or referred to in 
forms having a meaningless tcx inserted after the first syllable of 
the word. Thus, from 'ou'^sdmaH "I hunger for it" ('o- 
"empty" stem which may be rendered by "something" or "so 
and so;" -u^sd- verbifying suffix "to desire to eat;" -maH first 
person singular present indicative, used after vowels) is formed 
'utcxHsdmaH. Similarly, hinini'aLma" "now he comes" becomes 
hitcxnini'aLma' "now he comes, greedy fellow that he is." These 
tcx- forms are also used to refer to ravens, regularly to the mytho- 
logical Raven, a character noted for his gluttony. 

Cowards may be satirized by "making one's voice small" in 
referring to or addressing them, in other words by speaking in a 
thin piping voice that suggests timidity. 

It is interesting to notice that in several of the above usages, 
the notions of mere smallness, of contempt, and of affection are 
found side by side, and doubtless the precise nuance of feeling 
expressed depends much on the relations subsisting between 

' According to Dr. Paul Radio, the Winnebago also consider the bear to be left-handed. !■ 
the bear clan feast of these Indians the guests eat with a spoon in their left hand. 

364 ^J American Indian Languages 2 


the speaker and the person addressed or spoken of. What is 
meant in the spirit of pitying affection for a poor lame or hump- 
backed child or for a good-natured squinting old grandpa, 
might be intended to convey contempt when addressed to a 
young man and would be promptly resented as an insult. It 
is significant that the various types of abnormal forms of speech 
that we have reviewed are used with little or no reserve when 
speaking of the persons referred to or when addressing children, 
but are, on the whole, avoided when within ear-shot of adults 
so referred to. It seems further significant that the traits 
satirized are chiefly such as are inherent in a person, not merely 
acquired in the accidental course of events, whereby he is set 
apart by nature as falling short in some respect of the normal 
type of individual and is to that extent stamped as inferior. 
This may explain why blindness, which is more often acquired 
rather late in life than congenital, is not made the subject of 
speech-mockery. Added to this may be the feeling that blind- 
ness is too grave an affliction to be treated light-heartedly, an 
explanation which gains weight when the well-known sensitive- 
ness of the Indian is considered. 

Outside of the normal use of the diminutive in addressing or 
referring to children, the peculiar forms of speech that we have 
seen to obtain in Nootka are not easily paralleled in America. 
For diminutive verbal forms of the Nootka type Uto-Aztekan 
affords a close parallel. In Southern Paiute the regular diminu- 
tive suffix -tsi-, which is employed to form diminutive nouns 
and adverbs of all sorts, is also used as a verb suffix when speak- 
ing to or of a child. Cognate with this element is the diminutive 
suffix -tzin{tli) of Nahuatl. Derived from this is the verb 
suffix -tzinoa, "which," according to Remi Sim6on,^ "serves to 
denote respect or love;" it is generally, like reverentials of the 
type already referred to, employed with reflexive prefixes. 
Examples given by Remi Simeon are: otechmo-chiuilitzino in 
Totecuyo "our Lord created us" (o preterit prefix; tech- first 
person plural objective prefix; wo- third person reflexive 
prefix; chiui-, from chiua, because of following -li-, verb stem 
"to make;" -li dative suffix, mo- . . . .-li "for himself;" 
-tzino reverential, final -a being dropped because of preterit 

' Dictionnaire de la Langue Nahuatl ou Meiicaine, s.v. tzinoa. 

Eight: Wakashan and Salishan Languages 365 

tense; in definite article, "the"; to- first person plural posses- 
sive prefix; tecuyo noun stem "lord"); and timo-qauhtzinoa 
(quoted from Olmos) "you fast" {ti- second person singular 
gubject; mo- reflexive;^ gauh-, from gaua verb stem "to fast;" 
-tzinoa reverential). These forms may be rendered in some 
such fashion as: "our Lord has created us for himself, revered 
one," and "you fast, honoured sir." 

Strikingly similar psychologically to the cases of consonantal 
play in Nootka just considered are the peculiar consonant 
changes characteristic of Chinookan, employed to convey 
diminutive and augmentative notions respectively in all parts 
of speech.^ The change here of c- consonants to s- consonants 
to express the idea of diminution further illustrates the tendency 
of sibilants in America to be subject to consonantal play. In 
Yana the phenomenon of diminutive consonantism is illustrated 
in the change of I to n. This process takes place regularly in 
forming diminutive nouns in -p.'o; thus, ntnimaupla "little 
nose," from ltlimau{na) "nose." The l-n type of consonantal 
play is another one of some currency in America, and seems to 
obtain also in Sahaptin. This matter of consonantal play to 
express modalities of attitude is doubtless a fruitful field for 
investigation in American linguistics and should receive more 
attention than has hitherto been accorded it. It may be 
expected to turn up particularly in connexion with notions of 
smallness, largeness, contempt, affection, respect, and sex- 

Such consonant changes and increments as have been con- 
sidered are evidently of a rhetorical or stylistic as much as of a 
purely grammatical sort. This is borne out by the fact that 
quite analogous processes are found employed as literary 
devices in American myths and songs. I have already drawn 
attention to the fact,^ that in American mythology certain 
beings are apt to be definitely characterized by speech peculiari- 
ties. The employment of consonantal play or of similar devices 
in such cases seems always to have a decidedly humorous effect. 

• This verb is intrinsically reflexive. 

* See Sapir, Preliminary Report on the Language and Mythology of the Upper Chinook, 
American Anthropologist, N.S., 9, 1907, pp. 537, 538; and, in greater detail, Sapir, section 
on "Diminutive and Augmentative Consonantism in Wishram," in Boas, Handbook of American 
Indian Languages, pp. 638-645. 

• Sapir, Song Recitative in Paiute Mythology, Journal of American Folk-Lore, XXIII, 
1910, pp. 455-472. Takelma, Ute, Chinookan, and Nootka examples are there given, p. 471. 

366 VI American Indian Languages 2 


The culture-hero KwdtiydV of Nootka mythology is in the habit 
of inserting a meaningless x after the first vowel of a word; 
thus, the normal form htnuse'i" ''come up out of the water!" 
{hln- empty stem "to do, be;" -use-, umlauted from -usa- 
because of following i, "to move up out of the water;" -'z' 
imperative singular) becomes, at the same time, inasmuch as it 
occurs in a song, with song-vocalism, hlxnusa'e. In the speech 
of the Deer and Mink all sibilants, whether of the s or c series, 
are transformed into the corresponding laterals (s and c to I, ts 
and tc to l, tsf and tc! to l/). Thus, the Deer says Limil for tdmis 
"black bear;" U.dpaL for tddpats "canoe." The Nootka Deer 
and Mink style of talking is of particular interest for two rea- 
sons. In the first place, it will have been noticed that the 
consonantal changes are identical with those employed in speech 
about or addressed to those that have some defect of the eye, 
the latter type of forms, of course, being further characterized 
by the use of the diminutive suffix -'il (from -'is). Here we 
see at once the intimate connexion between the two types of 
consonant play. In the second place, the speech of the 
Nootka Deer and Mink offers an interesting parallel, 
or rather contrast, to that of the Kwakiutl Mink. This 
latter character regularly transforms all laterals to cor- 
responding s- sounds {I, L, I, and lI become respectively s, is, 
dz, and tsl), the exact reverse of the Nootka process. From the 
• point of view of the psychology of phonetics, it is significant to 
observe that both Nootka and Kwakiutl have a feeling for the 
interchangeability of the sibilant and lateral series of consonants. 
But the Mink of the Kwakiutl is not content with this. He 
also regularly transforms all anterior palatals to corresponding 
sibilants (x'. A;', g , and Ml become respectively s, is, dz, and ts!). 
There are still other phonetic changes to be found in Boas' 
Mink texts, but they seem less regular in character than these 
two; the changes at times of I and 'I to y and 'y may be instanced 
as one of these (thus ss'yt for h'lt "dead").^ Now it is per- 
haps significant that the change in Kwakiutl of anterior palatals 
to sibilants is curiously like the change of original Wakashan 
(Kwakiutl-Nootka) anterior palatals, as preserved in Kwakiutl, 

• For data on Mink's peculiarities of speech, see F. Boas and G. Hunt, Kwakiutl Texts — 
Second Series, Publications of the Jesup North Pacific Expedition, vol. X, 1906, footnotes o 
pages 82 to 154: and Boas, Kwakiutl Tales, Columbia University Contributions to Anthr 
pology, volume II, 1910, footnotes on pp. 126-154. 

Eight: Wakashan and Salishan Languages 367 


to c- consonants in Nootka.^ Thus, a Mink form nsdzt in 
Kwakiutl for normal UEg't "mountain" is strikingly similar to 
the regular Nootka cognate nutci\ Suggestive also, a propos of 
the use by Mink of sonant palatal spirants {y and 'y) 
for normal sonant laterals {I and 'Z), is the fact that in 
Nootka so-called "hardening" suffixes change immediately 
preceding I to 'y, corresponding in such cases to Kwakiutl 'Z.^ 
The bearing of these facts on mythological consonant play in 
Kwakiutl is not easy to determine; a possibility will be 
suggested farther on. 

Consonant play as a device in mythology is not confined to 
America. In reading some recently published Bushman litera- 
ture the writer came across striking parallels. The Bushman 
Mantis, who, like the Kwakiutl Mink, is a trickster, consistently 
changes all the cerebral clicks of normal speech into lateral 
clicks.^ Similarly, the Baboon transforms all the clicks of 
ordinary speech into a compound click, consisting of cerebral 
followed by dental click.^ Evidently a comic effect is aimed at 
in both these cases. 

The phenomenon of consonant and vocalic play is also well 
illustrated in Indian songs. Song diction is an extremely 
important, though rather neglected, field of primitive lore, and 
only one phase of it can be touched on here. Song texts often 
represent a "mutilated" form of the language, but study of the 
peculiarities of song forms generally shows that the normal forms 
of speech are modified according to definite stylistic conventions, 
which may vary for different types of songs. Sometimes sounds 
are found in songs which do not otherwise occur in the language. 
Where the texts of a type of songs are in the language of another 
tribe, as happens so often in America, such an abnormal sound 
may be simply borrowed from the foreign language, as is the 
case with the mourning songs of the Southern Paiute, which, 
sung to supposedly Mohave texts, contain many examples of I, 
a sound otherwise unknown in Paiute. On the other hand, 
new sounds may be developed spontaneously or in imitation of 
foreign sounds. The former is probably the case in the frequent 

1 See Sapir, Some Aspects of Nootka Language and Culture, American Anthropologist, 
N.S., 13, 1911, p. 16. 

^ See Boas, Handbook of American Indian Languages, pp. 430, 435; Sapir, loc. cit. 

' Bleek and Lloyd, Specimens of Bushman Folklore, 1911, footnotes on pp. 6 and 8. 

* Ibid. Footnotes on pp. 18 and 22. At least this is indicjited by Block's orthography, though 
possibly the compound sign is meant to indicate a special click not otherwise found. 

368 VI American Indian Languages 2 


Nootka use of 77, a sound quite foreign to normal Nootka speech, 
in certain classes of songs; the latter explanation is more 
plausible in the case of the regular Nootka change of n to Z in 
many songs. This n-l interchange, again, is significant in so far 
as Kwakiutl, doubtless agreeing in this respect with primitive 
Wakashan, has both n and I, while Nootka, when cognate words 
are compared, is seen to have only n to correspond to both. 
Of particular interest in this connexion is the fact that such 
special song-sounds (Paiute I; Nootka I and rj) are, at least so 
it would seem, pronounced with difficulty by Indians under 
ordinary circumstances, as in the handling of English words 
that contain them. The obvious inference is that one may 
react quite differently to the same speech-sound entering into 
dissimilar associations. This fact, has, of course, a much wider 
psychological significance.^ Conventional consonant changes 
in songs are no more restricted to America than, as we have 
seen, are parallel changes in mythology. An example that 
happens to have come to the writer's attention lately is 
the change of voiceless stops to corresponding nasals plus 
voiced stops in the songs of the Karesau-Papua of German New 
Guinea. Thus, the normal apil becomes ambil in songs.' 

In seeking some comparatively simple basic phenomenon, 
from which, as a starting point, the various types of consonant 
play we have illustrated from Nootka could have originated, 
• one easily thinks of the vocalic changes or consonant substitu- 
tions that take place in the speech of those who have some 
specific speech defect. The most familiar case of this sort in 
English is lisping, which simply means that the ordinary 
alveolar sibilants (sometimes also stops) are changed to the 
corresponding dental sibilants or even interdental fricatives 
(and sometimes correspondingly for stops). Information was 
obtained of five types of speech defects found among the Nootka. 
The first of these is called ninikHn^ {nini- reduplicated stem; 
-k!in} "to make a sound of") and consists of the involuntary 

' Sounds falling outside the regular phonetic system of the language may be spontane- 
ously developed also by the operation of other systems of consonantal (or vocalic) play than 
are found in song diction. Thus, in Wishram (Upper Chinookan), the analogy of certain 
consonant changes of augmentative value (as of p to 6, ( to d, k to g) brought about the 
creation of dj, a sound otherwise unknown in Chinookan, as the augmentative correlate of 
tc or Is sounds. See Handbook of AmeTxran Irnlian Languages, pp. 638, 639, 640. 

' See Father W. Schmidt, abstract of Ober Musik und Gesange der Karesau-Papuas , 
Deulsch Neu-Guinea, Bericht iiber den III. Kongress der Internationalen MusikgesalLschaft, 
1909, p 297. 

Eight: Wakashan and Salishan Languages 369 


nasalizing of all vowels and continuants. Thus, the normal 
hayd'akaH "I do not know" (-an first person singular present 
indicative) is pronounced by people who have this defect 
hqijq 'qkqii. The father-in law of Dan Watts, who is a Ucluelet 
Indian that came to visit his son-in-law, was observed by the 
writer and definitely stated by Dan to have this "nasal twang," 
which is due to an inability, muscular or nervous, to raise the 
velum so as to shut off the passage of the outgoing breath 
through the nose. In speaking of the elk, ninik/in* forms are 

A second type of defective articulation is termed hahdCklin^ 
or hahdt/in^ {hahaC- reduplicated stem; -k!in^ "to make a 
sound of"), and is supposed to be due to a hole in the palate. 
I have no clear idea as to just what the organic basis of the 
faulty articulation is, but, judging from the examples given of 
it, it seems evident that those subject to it have difficulty in 
articulating against the hard palate. Perhaps the speech defect 
is due to cleft palate. All ^s and tc affricatives (presumably also 
lateral aflfricatives) become simple t- sounds (dental), while s, c, 
and I become interdental fricatives (6). The acoustic effect is 
that of an exaggerated lisp. Thus, tclotck" "all" becomes t!6Vk'; 
'otsVyukwaH "I go to it" ('o- empty noun stem "some- 
thing;" -tsi'yukw- "to go to;" -an "I") becomes ' otV yukwan ; and 
tc.'op'tdop'cinil "stretch around the neck; sweater" (tc.'op'- 
tdop'c- reduplicated stem; -inil "at the neck") becomes 
t/op^t/op'dinid. This latter rests on the authority of Dan Watts; 
Alex Thomas, starting from a form tcfop'tcfop'cimil for 
''sweater," gave tlop'tlop'timil as its hahdtUn^ correspondent. 
Those who are hahdt'kfin^ thus confound three distinct series of 
consonants in a single dental or interdental series. Such per- 
sons are imitated when addressed. The outward resemblance 
with the phenomena of consonant play is quite striking here. 

This resemblance becomes even stronger in the case of the 
third Nootka speech defect of which information was obtained, 
that known as tstska' (tsisk- verb stem; -a' verb suffix of con- 
tinuative significance) or istskaq'sul (tsisk- verb stem; -aq'sul, 
perhaps misheard for -ak'sul "at the lips"). Such as are 
subject to it are supposed always to keep their teeth open and 
to be saying ts-]-. As a matter of fact, those who are tstska 

370 VI American Indian Languages 2 


change all s and c- sounds to palatalized sibilants (s). Thus, 
'otsl'yukwaH "1 go to it" becomes 'ot^Vyukwan; si'ydsaH "it is 
mine" (si'yds- "to be mine," from independent pronoun si'ya' 
"I;" -aH first person singular present indicative) becomes 
si'y&saH. It will be remembered that these consonant changes 
are characteristic of the forms used in addressing or speaking 
about abnormally small adults, except that such discourse is 
further characterized by the use of the diminutive suffix -i^ 
(from -'is). Here there is a tangible connexion between the 
involuntary consonant changes brought about by a speech 
defect and the consonant play used to symbolize a body defect, 
though it is far from obvious in this particular case what asso- 
ciation there can be between a kind of lisp and a dwarfed condi- 
tion of the body. A further point of interest is that those who 
are tsiskd' are generally imitated when spoken of. The signi- 
ficance of this in the argument is obvious. 

Somewhat similar to the hahdtfin* speech defect, yet not to be 
confused with it, is that known as kakaVwin"^ "to talk as one 
with missing teeth" (cf. kdlxwak" sul ''to have teeth missing in 
one's mouth"). Such persons speak with a decided lisp, sub- 
stituting B for s and c, id for ts, id! for ts! and td, but, it would 
seem, t for ic. Examples are: 'e'pinid from 'e'pinis "apples;" 
'd'yintad from '6'yintcas "oranges;" timid from tdmis "bear;" 
Id/otk' from tclotck' "all;" tdfdpatB for tcfdpats "canoe" 
(contrast the corresponding hahdtfin^ form: tfdpaV). Here 
again, one who is afflicted with this speech defect is imitated 
when addressed; thus, Alex Thomas, before he had caps put on 
his vestiges of teeth, used to be mocked A;aA;dr'tt;m'-fashion. 

A fifth, not uncommon, speech defect among the Nootka is 
stuttering. Stutterers, like all other persons who have some- 
thing abnormal about their speech, are derided by being imitated. 

The West Greenland speech defect known as kutdt'oq^ is 
particularly instructive in that an individual speech-peculi- 
arity, which, however, seems to be a common one in the Eskimo 
settlements along the coast, has become one of the dialectic 
peculiarities of the northern settlements of the Upernavik 
district. The kuidt'oq habit consists in substituting ordinary 
gutturals (fc- sounds) for velars {q- sounds), and is evidently due 

' See W. Thalbitier, A Phonetical Study of the Eikimo Language, Meddelelser om Gron- 
Und. XXXI, 1904, pp. 178-180. 

Eight: Wakashan and Salishan Languages 371 


to the greater difl&culty of bringing about a contact between the 
root of the tongue and the velum than farther front in the mouth. 
This defect, it should be noted, brings with it the confusion of 
two etymologically distinct series of consonants with resulting 
grammatical or lexical ambiguities, at least theoretically. In 
this respect kutdt'oq forms are parallel to the forms resulting 
in Nootka from speech defects or the use of consonantal play. 
Children are particularly apt to be kutdt'oq, but generally lose 
the habit as they grow older. However, certain adults, particu- 
larly women, always remain kutdt'oq, whether because of the 
mere force of habit or because of a physiological or anatomical 
impediment. As for the Upernavik peculiarity, it seems clear 
that the kutdt'oq habit can hardly be due to the individual 
disability or carelessness of all the members of the district, 
but that what was originally a speech defect has become social- 
ized into a dialectic peculiarity. The analogy with the forms 
employed in Nootka in speaking of or addressing certain classes 
of people that are ill-favoured by nature is striking. 

The explanation and genesis of the various types of speech 
mutilation in Nootka can hardly be more than guessed at, yet 
certain probabilities, in part already suggested, seem to stand 
out. In the first place, the use of definite morphological elements 
to indicate some characteristic of the person spoken to or of 
(Nootka -is and -aq*; Paiute -tsi-; Nahuatl -tzinoa) needs no 
particular comment, at least from the purely linguistic point of 
view. Further, definite points of contact have been estab- 
lished between speech defects and "mocking-forms," with 
consonantal play, on the one hand, and between the latter and 
myth-character forms with consonantal play, on the other. 
I am inclined to believe that the observation of consonant 
substitutions such as take place, with involuntarily humorous 
effect, in the speech of those that articulate incorrectly, has set 
the pace for the consciously humorous use of the same or similar 
substitutions in both mocking and, directly or indirectly, myth- 
character forms. The Nootka mocking-forms, with their use 
of the diminutive affix and of consonant play, represent a 
combination, both linguistically and psychologically, of the 
pity and affection symbolized by the use of the diminutive 
element and of the contempt or jesting attitude implied by the 

372 VI American Indian Languages 2 


imitation of a speech defect. A myth character whom it is 
desired to treat humorously may, among other possibilities, 
be relegated either to the class of poor talkers or to that of nat- 
ure's step-children. Hence the consonant play of such characters 
is in part traceable either to speech defects or to mocking-forms. 
In passing it may be observed that the "enfant terrible" motive 
is fairly clear in the treatment of many humorous characters of 
American mythology, and that consonant play may in some 
cases be taken to symbolize this attitude. The socializing of 
the kutdt'oq habit among certain of the Eskimo forcibly suggests 
the influence of the speech of children as a contributing factor 
in the creation of myth-character forms. The Kwakiutl Mink 
is a very likely example of the "enfant terrible," both in action 
and speech. The possibility should not be lost sight of, of the 
use of myth-character forms to apply to a class of people or to 
an individual in ordinary life. This would be an extension of 
the well-known American Indian habit of comparing one that is 
marked by some peculiarity of temper or habit with a favourite 
mythological character.^ 

There is, however, another factor which has undoubtedly 
exercised a great influence both on the forms of speech used 
by myth-characters and on the forms peculiar to songs. This 
is the comic or novel effect produced by the imitation of the 
speech of foreigners, particularly of such as speak a dialect 
divergent enough from the home-dialect to be funny or impres- 
sive, yet not so different as to be unintelligible and, therefore, 
lacking in interest. Hence we often find mythological characters 
in America making use of a neighbouring dialect of the language, 
as in the case of the Nass River TxdmsEm and other characters, 
who talk in the dialect of the Tsimshian proper of Skeena 
river.^ Examples of songs whose texts are in a divergent 
dialect, not to speak of the common use of a totally distinct 
language, are frequently met with in and out of America. A 
well-known instance is the use by Melanesian tribes, according 
to Codrington, of the dialect of some neighbouring tribe for their 
own song diction; thus, the Melanesians of Mota (Norfolk 

' A few interesting examplea are given by A. Skinner, Notes on the Eastern Cree and 
Northern SauUeaux, Anthropological Papers of the American Museum of Natural History, 
vol. IX, 1912. p. 82. 

' See Boas, Tsimshian Texts, Bulletin 27 of Bureau of American Ethnology, 1902, pp. 8, 
18. 20. 30. 35, 46. 61-64, 78, 171. 

Eight: Wakashan and Salishan Languages yiZ 


island of Banks islands) use for their songs the dialect of Saddle 
island. Also in the clownish episodes of rituals, which are so 
characteristic of America, the impersonation and imitation of 
the speech peculiarities of foreigners are often resorted to and 
never fail to arouse a hearty laugh. In all these cases, it is 
rather important to observe, real accuracy of imitation is not 
generallj' attained or even aimed at, so that the foreign style 
often tends to reduce itself to a number of conventional vocalic 
and consonantal displacements. In dealing above with the 
change of anterior palatal k- counds to ts- sounds in the 
language of the Kwakiutl Mink, I pointed out that a similar 
change was involved in the passage of original Wakashan 
anterior palatal k- sounds to Nootka tc- sounds. It is 
just possible that the Mink ts- sounds are in such cases due 
to an imitation of the speech of the northern Nootka tribes. 
The difficulty with this interpretation is that Nootka and 
Kwakiutl are altogether too divergent to afford more than a 
quite inconsiderable number of illustrative cases of the k- tc 
change, and of these but few would strike the naive mind. It 
seems more plausible, on the whole, to assume that both the Mink 
and Nootka consonant changes rest on a common Kwakiutl- 
Nootka tendency, perhaps a tendency on the part of children to 
pronounce anterior palatals as sibilants. Data on the speech 
peculiarities of Kwakiutl children would be valuable here. 

The Nootka Indians of one tribe frequently imitate the real 
or supposed speech peculiarities of those belonging to other 
Nootka tribes, the stress being primarily laid not so much on 
peculiarities of vocabulary and grammatical form as on general 
traits of intonation or sound articulation (cf. our New England 
"nasal twang" and Southern "drawl"). For the purposes of 
this paper the Nootka now spoken by the TsUcd'atH" and 
HopdtcIas'atH" of Barkley sound and the head of Alberni canal 
may be taken as the normal form of Nootka speech; this is, of 
course, purely arbitrary, but so would any other point of de- 
parture be. It is instructive to note that one or two of these 
tribal speech peculiarities coincide with individual speech defects. 

According to the TsUcd'atH" Indians, the Houtcuq' Lis'atH" tribe 
of Uchucklesit harbour, a western inlet of Alberni canal, speak or 
spoke (for there are few of them left now) in a rumbling fashion 

374 VI American Indian Languages 2 


(JoL!o:en^) ; they are said to use their throat more than the other 
tribes. The peculiarity referred to seems to be a more than 
ordinary use of velar resonance, due to a tightening of the passage 
between the root of the tongue and the velum or perhaps the 

The Hoidi'atH'^ Indians of Sarita river and the southern 
shore of Barkley sound are said to speak LfdLfatdin'^, a splutter- 
ing effect being apparently referred to. As far as can be made 
out, their speech peculiarity consists in a more liberal use of tc 
sounds than ordinarily. Thus, according to Alex Thomas, the 
Ho'.di'atH" say 'ndtcciL instead of 'ndcciL "to look at" (as a 
matter of fact, this usage is probably etymologically justified, as 
'nac- and, in other forms, 'natc- are both used as verb stems in 
Ts.'ud'atH" itself); instead of pronouncing tdayi'is "give me 
water" (tc.'a- noun stem "water;" -yi- verbifying suffix "to 
give;" -is second person singular imperative with first person 
singular object) they say something like tdatcyVis, though Alex 
maintained that it was not a full clear-cut tc that was inserted. 
At any rate, the TsUcd'atH'^ have seized upon the tc- insert as a 
convenient means of poking fun at their Hdidi'atn'^ kinsmen, 
using it in ways that are certainly not, nor meant to be, accurate 
renderings of the tribal peculiarity. Thus, the tribe itself is 
humorously referred to as HotcidVatH'^; Numdqemiyis, the main 
inlet of their country, is similarly termed Nutcmdqemiyis. 
Evidently, we have here an example of a mocking usage, based 
on a tribal peculiarity, that is in form perfectly analogous to 
certain myth character and cripple-mocking usages (cf. inserted 
X for Kwatiyat and inserted tcu" for left-handed people.) 

The northern Nootka tribes, beginning with the La'okwi'atH'^ 
of Clayoquot sound and proceeding north, are said to speak 
tdHtdua', which refers to a drawling or long drawn out manner of 
talking. Apparently the peculiarity, which is often imitated in 
jest, consists not so much in lengthening out vowels as in a some- 
what exaggerated rise in pitch towards the end of a sentence, 
which gives the flow of speech a sliding cadence. The most 
northern Nootka tribe, the Tc.'vq^Lis'atn'^, are said to be all 
stutterers and are accordingly imitated in jest. 

In imitating the Nitinats {Nlttna' atn") , a group of Nootka 
tribes to the south of Barkley sound that speak a very divergent 

Eight: Wakashan and Salishan Languages 375 


dialect, the meaningless syllable -aq' is always added to the 
word, as this syllable is supposed to be a very common one in 
Nitinat. This device is strikingly similar to the use of sufiixed 
-ag' for large persons. 

The real old Hopdtdas'atH'* Indians, whose earliest homes were 
in the interior of the island along Somass river and about Sproat 
and Great Central lakes, were said to talk tstska', that is, to 
confound s and c sounds. As we have seen, this is also a well- 
recognized individual speech defect among the Nootka. In the 
case of the Hopdtdas'atH'', the tstska' habit was simply due to the 
fact that they carried over into Nootka speech a linguistic 
peculiarity found in the Salish dialect which they originally spoke 
(a dialect apparently identical with or closely related to Boas' 
PEnLatc; recognized as Pindd'atc by Tyee Bob, the leading man 
among the Hopdtdas'atH'* to-day and whose father is still re- 
membered to have spoken tsiska^). 

As for the TsUcd^atH" themselves, they are said by the other 
tribes to talk very fast. If one anywhere among the Nootka 
Indians talks too fast, the proverbial saying is that he is a 

It will, as we have seen, have to be admitted, that mocking 
forms for various classes of people are connected not only with 
speech defects and mythological devices, but, to a large extent, 
also with tribal speech peculiarities. 

Finally, the possibility of a direct psychological relation 
between the consonant change and the type of individual or 
attitude it symbolizes should not be summarily ruled out of 
court. That such an association once established by historical 
causes will be felt as a direct and simple psychological associa- 
tion is quite obvious, also that it may become productive, by 
analogy, of further associations of a related sort. I would, 
however, even be inclined to suppose, though proof may be 
difficult or impossible, that certain associations of sound and 
character or form arose more or less spontaneously, or, to put it 
more correctly, by virtue of the inherent associative value of 
the otherwise unconnected phenomena in the mind of a par- 
ticular individual or group of individuals. Such an individual 
association, if given outward expression, can become socialized 
in the same way in which any individual idea becomes socialized. 

376 VI American Indian Languages 2 


The type of association here thought of is quite parallel to the 
sound-colour associations familiar enough in psychology. It 
may be not uninteresting as a psychological datum to note that 
the writer himself feels, or thinks he feels, the intrinsically 
diminutive or augmentative value of certain consonant changes 
in Wishram. Moreover, the association of c- consonants with 
humpbackedness in Nootka seems not so far-fetched after all. 
The thickish quality of these consonants, together with the pro- 
trusion of the lower jaw in pronouncing them, suggests to me 
the same squat clumsiness as the image of a hunchback. All 
this may, of course, be merely auto-suggestion ad hoc. 

To summarize, evidence has been presented of the historical 
connexion between various linguistic and stylistic processes 
involving the symbolic use of sounds. These are diminutive 
and augmentative forms of speech, mocking-forms, myth- 
character and animal forms, and song forms. Moreover, further 
evidence has been presented to show the historical connexion of 
these quite specialized tricks of language with the far simpler 
phenomena of speech defects, children's language, and imitation 
of the phonetic peculiarities of foreigners. The direct association 
of some of the former with the types they symbolize, after the 
manner of primary association between data of distinct sense, 
has also been suggested as a possibility. 

The data brought forward in this paper as to the associations 
obtaining in Nootka between various classes of persons, mytho- 
logical beings and animals, linguistic devices designed to satirize 
or characterize, speech defects, and tribal speech peculiarities, 
may be most conveniently grouped in tabular form. The arrange- 
ment in the table is intended to emphasize the purely linguistic 

Phonetic Key. 

a, short as in German Mann; e, short and open as in English 
met; i, short and open as in English it; o, short and open as in 
German voll; u, short and open as in English put; e, short and 
close as in French ete) i, short and close as in French jini; o, 
short and close as in French chaud. 

a, long as in German Bahn; e, long and close as in German 
See; i, long and close as in German Sie; o, long and close as in 

Eight: Wakashan and Salishan Languages 371 


German ro/i; e, long and open as in French /e/e; d, long and open 
as in English saw, yet with back of tongue not so low. 

E (Kwakiutl), short obscure vowel like e of German Rose; 
I (Nootka), short open i-vowel of rather unclear quality; * 
(Nootka), occurring as syllabic final after n and m, barely 
articulated or murmured (yet not voiceless or whispered) /; 
" (Nootka), denotes a-timbre of preceding h (see below). 

c, like sh in English ship: tc. corresponding voiceless affrica- 
tive, ch of English church (in Nahuatl ch is used for tc); dj, 
corresponding voiced affricative, j of English joy; s and is, as 
in English sit and hats (in Nahuatl z and tz are respectively 
used instead); s and ts, palatal voiceless sibilant and affricative, 
acoustically midway between s-c and ts-tc respectively; c and 
tc, c and tc pronounced with lower teeth in front of upper; 6, 
interdental voiceless spirant, like th in English thin. 

q, voiceless velar stop like Semitic qof; qw, labialized form of 
same; x, voiceless spirant of g-position; x, voiceless spirant of 
fc-position, not pronounced as far back as German ch of Bach; 
k • and g • (Kwakiutl), anterior palatal stops (palatalized k-stops), 
approximately ky and gy; x- (Kwakiutl), voiceless spirant of 
fc- -position, ch of German ich; 17, voiced nasal of k- position, ng 
of English sing; rj (Eskimo), voiced nasal of q- position. 

I, voiceless lateral spirant; l, corresponding voiceless lateral 
affricative (written tl in Nahuatl); i (Kwakiutl), corresponding 
voiced affricative. 

' , glottal stop; ; (Nootka), strangulated-sounding laryngeal 
stop, similar in resonance to Arabic 'ain; h (Nootka), strangu- 
lated-sounding laryngeal spirant, Arabic ha; \ aspiration or 
breath-release of preceding vowel or consonant (p", V, k\ and 
g' are aspirated voiceless stops); .' denotes glottalized stops 
and aflfricatives {p!, t!, kf, q!, l!, ts!, tc!, ts!, tc!, k-!), that is, such 
as are pronounced with simultaneous closure of glottis, but with 
oral release prior to that of glottal release. All other con- 
sonants as in English. 

' , stress accent; *, denotes preceding long consonant (except 
in Kwakiutl k'- sounds); ,, denotes nasalization of vowel under 
which it is placed; +, denotes excessive length of preceding 
vowel or consonant. 


VI American Indian Languages 2 



-0 «> 



C b 




s •■ 









•*j 1 














■0 1 












V.S *^ 






rt u S 





':; i 



















•2 §"* 


•1 § 








73 •»' 















1 ■-T3 


3 3 

.9 " 

3 "„, *• 


3 fl 

3 d _, 









'3 * 





■ — aj 3 

a " 

3 v> 












^ fl 




S 3'S 

9> .. " 

2 S'S' 






•a 4) 









S fl { 


















.2 fl 

















01 .= 


_ "J 












.2£ fc> c 

U) c 

^ 'c 





— V 









$ *" 












3 — 

it c « 
0-5 a 








g-3 2 





























Eight: Wakashan and Salishan Languages 379 

Editorial Note 

Originally published as Memoir 62, Anthropological Series 5, Geological 
Survey, Department of Mines, Canada. Ottawa (1915). 

Although the formal patterns of consonantal symbolism, particularly dimin- 
utive/augmentative devices, have been much discussed in Americanist work 
(for an excellent survey see Nichols 1971), the rhetorical and literary uses of 
these have been noted only sporadically. Some attention has been given 
recently to the phonetic stereotyping of the Coyote and similar figures in myth 
(Aoki 1975: 190; Toelken 1969). Hymes's study of the expressive resources of 
Takelma traditional literature (1979; revised in Hymes 1981: 65-76) is closest to 
the spirit of Sapir's work. 

Noun Reduplication in Comox, 

a Salish Language of 

Vancouver Island 


One of the most characteristic grammatical processes of a 
group of Northwest Pacific Coast languages, embracing the 
Tsimshian, Kwakiutl-Nootka, Salish, and Chemakum linguistic 
stocks, is initial reduplication, employed in both noun and verb 
forms to indicate a variety of grammatical concepts, chiefly 
those of plurality, distribution, and iteration. The Salish 
languages in particular are known to make exuberant use of 
reduplication for grammatical purposes, but the subject, which 
seems to bristle with irregularities and intricacies of detail, has 
never been adequately treated for any of the numerous dialects 
of the stock. Indeed, a thorough grammatical study, at the same 
time phonetically adequate, of a Salish language, is still one of the 
desiderata of American linguistics. 

During the autumn of 1910, while prosecuting ethnologic and 
linguistic research for the Geological Survey of Canada 
among the Nootka Indians now living in two reserves near 
Alberni, B.C., opportunity was incidentally found to gather 
some linguistic data on Comox, a Salish language spoken on 
the east coast of Vancouver island near the present town of 
Comox. The dialect represented in these notes seems to be 

382 ^f American Indian Languages 2 

Comox proper (Q!6mgx"s), with which Uohos, spoken on the 
mainland of British Columbia, was stated to be identical. 
SdlolV'' was stated to be a northern dialect of the same language. 
This term is evidently identical with Boas' Qatloltq, which he 
uses to apply to the most northern group of Coast Salish tribes, 
excluding Bella Coola, inhabiting "Discovery Passage, Valdes 
Island, Bute and Mala-^pina Inlets."^ Boas adds, "The 
Catl61tq are called K'omoks by the Lekwiltok'" (southernmost 
Kwakiutl tribe.) 

The informant was Tommy Bill, an Indian of mixed blood, 
whose father belongs to the Ts'.ictiatH" tribe of Nootka Indians, 
while his mother was a Comox, he himself living with and being 
to all intents and purposes a member of the Hopdtdas'atH" 
tribe of Nootkas. His knowledge of Comox was obtained in 
his earher years, when living among his mother's people, whom 
he visits from time to time; it is only fair to add that he speaks 
mainly Nootka and English nowadays and does not claim to 
have a perfect command of Comox. However, the rather 
elementary character of the data obtained, together with 
convincing internal evidence derived from their study, leaves 
no room for doubt as to the essential accuracy of the material 
here presented. Most of the time spent on Comox was taken 
up with securing material pertinent to the problem of reduplica- 
tion in nouns. For most of the nouns obtained, plural, diminu- 
tive, and diminutive plural forms were secured, all of which 
involve various types of reduplication. Our linguistic material 
thus naturally divides itself into three heads, not to speak of a 
small number of nouns that are always used in reduplicated 
form. A few introductory remarks on Comox phonetics and 
some supplementary data are also added. 


Vowels. The short vowels found in the Comox material 
secured are: a (as in German Mann); a (as in English bat); 
e (short and open as in English 7net); e (short and close as in 
French ete); i (short and open as in English hit); i (short and 
close as in French fini); o (short and open as in German dort); 

' See F. Boas, First General Report on the Indians of British Columbia, Report B. A. A. S., 
1889, 5th Report on North-Weatern Tribes of Canada, p. 10. 

Eight: Wakashan and Salishan Languages 383 

(short and close as in French beau); and u (short and open as 
in English 'put). Of these vowels, e, i, and i are etymologically 
one sound, which is modified by phonetic surroundings; simi- 
larly, o and u. Velar consonants tend to lower preceding or 
following i to e (possibly sometimes e), while certain consonants 
(particularly s and I) tend to palatalize i to i. e and o, which 
latter does not occur often, are doubtless etymologically related 
to e and o respectively, but seem in every case to be clearly 
kept distinct from these, a is not common. 

Corresponding to each of the short vowels is a long vowel 
(long d, however, has not been found). These are indicated as: 
d (as in German Bahn); e (long and open as in French mere, or 
as in English hear, but without "r-vanish"); e (long and close 
as in German See); i (long and open as in English beer, but 
without "r-vanish"); i (long and close as in English see); 6 
(long and close as in English roll, or as in German Sohn) ; 6 
(long and open as in English born, but without "r-vanish"); 
u (long and close as in English rule) ; and u (long and open as in 
English poor, but without "r-vanish"). Similarly to the 
corresponding short vowels, and under parallel phonetic circum- 
stances, e, %, and I are variants of one sound, etymologically 
speaking, though i is often to be interpreted as lengthened form 
of inorganic vowels, in which case it does not seem to vary with 
e and l; 6, u, and u are likewise representatives of what is 
etymologically a single sound. 6 does not often occur; it is 
probably etymologically related to 6. e occurs often and cannot 
be considered a mere variant of e. 

As not infrequently happens in American Indian languages, 
the long vowels are not always held out with even stress, but 
end with short rearticulations which give the whole vowel in 
each case a quasi-diphthongal effect. Such vowels have been 
noted by the writer in Takelma, Southern Paiute, and, at least 
to a moderate extent, in Nootka; Boas has noted them in 
Tsimshian. While they occur to a considerable extent in 
Comox, they cannot as in Takelma be considered the normal 
forms of the long vowels; sometimes the short rearticulations 
seem to serve as glides to following consonants, particularly 
velars. The quasi-diphthongal long vowels are here indicated 
by long vowels followed by superior short vowels, the vocalic 

384 VI American Indian Languages 2 

quality of the latter being indicated as in normal short vowels. 
There are found: a"; e%' e^; e' (occurs before anterior palatal 
consonants) ; 1'; i'; i^ (occurs before velar consonants) ; 
0° and o"; and w". A number of cases also occur of short 
vowels followed by weak rearticulating vowels; such are e", 
q^, and i' (here the * is a glide to the following velar consonant). 
Some of these may well represent secondarily shortened long 
vowels. Differing from such long or short vowels with quasi- 
diphthongal character are vowels that are secondarily diph- 
thongized by a vocalic glide whose timbre depends wholly on 
the following consonant; such is i" in /cwpw"mi"a;" "hill," in which 
the second " is a glide due to the u- timbre of the final consonant. 

Short vowels of somewhat obscure quality are also found, 
either representing dulled forms of normal short vowels or being 
of inorganic origin and meant to lighten consonant clusters or 
serve as glides. Such vowels are: a (as in English but, yet 
sometimes less clearly marked in quality), which is sometimes 
inorganic, sometimes dulled from a; e (obscure vowel with e- 
quality) ; and / (very short rather unclear i). 

At times short vowels are so weakly articulated as to be 
barely audible; these are rather "murmured" short vowels 
of etymological significance than merely glides, timbre-echos 
of preceding consonants, or voiceless vowels. Examples are: 
? in Idl^bo'm' "small clam" (-^?6- reduced from ld'"b- in I6"°bnm' 
"clam"; yet in this case ? can just as well be morphologically 
dispensed with and phonetically explained as a timbre-echo of 
-ol-); ^ in qt'w^x "steel-head salmon" (that ^ is organic, 
despite its dull quality and extreme brevity, and reduced from a, 
is indicated by Nootka qe'wan "steel-head salmon," with which 
Comox qt^w^x is evidently identical; borrowing has doubtless 
taken place); "^ and " in hew^qen" "swan" and its diminutive 

Another class of "murmured" vowels (German 'Murmel- 
vokale") is formed by weakly articulated, yet not voiceless, 
vowels occurring in syllabically final position after glottal 
stops ('). Such vowels are only in part "murmured echoes," 
i.e., reduced repetitions of immediately preceding fully voiced 
vowels (such are a'", e'*, V\ ai'\ o'?, d'"; vowel breakings of this 
type occur often in American languages) ; in some cases we have 

Eight: Wakashan and Salishan Languages 385 

also murmured vowels after glottal stops that are of different 
quality and etymologically distinct from immediately preceding 
vowels (such are a'' and a'O- 

Some consonants, notably glottalized ("fortis") consonants, 
are apt to be followed by timbre-echoes dependent in quality 
on the preceding vowel. This simply means that the oral 
resonance chamber characteristic of a vowel may, failing to be 
materially disturbed by the following consonant position, 
linger on and thus become acoustically noticeable as a voiceless 
(sometimes aspirated) vocalic echo; if the consonant is a 
spirant, the vocalic timbre may be audible during its production. 
Examples of such unaspirated timbre-echoes after glottalized 
consonants are: " in pld'alats!'^ "skunk" and ° in k!6°ddt!° 
"porpoise." In Id^g^et!" "herring" the t! was heard with definite 
a-timbre despite preceding e. After u (o)-vowels syllabically 
final A;-sounds are regularly followed by echoes (aspirations 
when consonant is not glottalized) with w-timbre. Hence A;'", 
A;.'", X", q'", q!", and x" (see below for orthography of fc-sounds). 
These sounds, however, are also very frequent after unrounded 
vowels, as in Id'^dak'"' "skin;" in such cases they represent 
original labialized A;-sounds (see below). Aspiration with 
definite w-timbre is also found after t, as in sdW" "woman." 

Excluding such inorganic diphthongs as are formed by vowels 
and following glides (e.g., i"), there have been found as true 
short diphthongs ai, au (also au), di, ei, and long diphthongs 
di, du. Vowels normally forming diphthongs that do not so 
unite, each preserving its full value, are separated by . (thus, 
a.i as distinct from true diphthong ai). Stress accent is indi- 
cated by ' over vowels. 

Consonants. The consonant system of Comox is fairly 
full, including, as it does, eleven distinct series that differ 
according to place of articulation. As regards manner of articu- 
lation, six distinct series are to be recognized (voiceless stops, 
glottalized or "fortis" stops, voiced stops, voiced nasals, voiceless 
spirants, and voiced spirants), though by no means all of these 
are represented for all places of articulation. The voiceless 
stop and glottalized stop series are complete, the voiceless 
spirants nearly so, while the others are quite defective. All 
these consonants may be represented in the form of a table : — 


VI American Indian Languages 2 

Laryngeal glottal) 


Imbialiekd velar 


Labialized guttural 

Pbe-quttural (anterior pala 


Dorsal lateral 

Palatal sibilant 

Alteolar aiBILANr 





















c is pronounced like sh of English ship; x^ like ch of German 
ich. tc, td, dj (like j of PJnglish jam), ts, and ts! are affricatives 
(stop plus corresponding spirant; no simple stops correspond 
to <c-series). l and l! are also affricatives, but with lateral 
(voiceless spirant /) release. 

h and d are phonetic variants of m and n; h and d were often, 
though not consistently, heard between vowels, m and n rather 
consistently as initials, while m and n were more often heard as 
syllabic finals than h and d. These h-m and d-n sounds have 
been at various times analysed by Boas as "semi-nasalized" 
consonants. "The nasal opening," he v/rites, "may differ in 
width, and the stricture of the upper nares may produce semi- 
nasalized consonants. "2) Again, in speaking more definitely 
of Coast Salish, "... the b sound ... is produced with 
half-closed nose by the Indians of the Strait of Fuca, in the State 
of Washington. . . . The characteristic trait of the sound is a 
scmiclosure of the nose, similiar to the effect produced by a cold 
ill the head."' These remarks doubtless apply to Comox as 

Eight: Wakashan and Salishan Languages 387 

well as to more southern Coast Salish languages, yet it seems 
likely to the writer that under certain phonetic conditions these 
semi-nasals become true nasals. No attempt will here be made 
to normalize orthography on this point, a faithful record of 
what was heard, or thought to be heard, being presented. 

Eliminating h and d as of secondary origin {g^ and dj, it should 
be carefully noted, are true sonant stops, not "intermediates"), 
all the other consonants listed in the table are etymologically 
distinct, that is, none of them are mere variants, {k, k!, and 
X, however, may prove to be merely secondary forms of kw, 
k!w, and xw.) This gives us no less than thirty-six (or thirty- 
three) organically distinct consonants to operate with. A 
secondary series of aspirated surds (voiceless stops followed by 
aspiration) arises when voiceless stops occur as syllabic finals 
(written p\ t\ k\ fc^', q\ tc^); kw and qw become fc'" and 5'", 
that is, their aspiration-release has w-timbre; similarly, k!w 
and q!w in this position become fc.'" and g/". q, it may be noted, 
is often released into a weak spirant glide x (written ^) before 
the following vowel is attacked (thus, q^'a for qa). Final vowels 
and m and n are also often followed by aspiration (-a' and 
similarly for other vowels, m' or less often h\ n' or less often d"), 
though this was not consistently heard. Final m and n are 
etymologically distinct from final glottally affected m and n, 
which are written wi' and n' (sometimes breath release is heard 
after glottal release, when they are written m" and n"). Long 
consonants (indicated by ' after consonant) were noted, but 
seem to be of no etymological significance (examples are q', d'). 

Sound Changes. Lengthening and reduction of vowels are 
important phonological processes in Comox, also, though to 
less extent, changes of vowel quality. As these, however, are 
generally of grammatical significance, they are best taken up in 
their proper place under types of reduplication. As more 
strictly phonetic pure and simple in character is to be considered 
the palatalizing of a to i in the neighbourhood of g", also the change 
of A to w and i in appropriate phonetic circumstances. These 
changes also, however, are most clearly brought out in connex- 
ion with morphological processes. 

Many cases of g", perhaps all, are undoubtedly due to original 
w. It seems that w, when it came to stand between vowels 

388 VI American Indian Languages 2 


(not, it would seem, including cases of preceding vowel plus 
glottal stop), also initially in many cases, regularly passed into 
g^. Thus, as diminutive of xAucin* "bone" is found xt^xig^icin' 
<*xtxAwidirC {-Ag^- becomes -ig^-, as noted above). Similarly, 
from qt'w^x "steel-head salmon" is formed qeqeg^e^x "little steel- 
head salmon" and qeqAuqd'''g"e^x "little steel-head salmon 
(plur.)." This phonetic law explains a class of plurals, formed 
by reduplicating with o- vowel, derived from stems in internal 
-g"-. Thus, from t!eg"em {<*t!ewem) "sun, moon" is formed 
plur. UoHleg^em {<*t!AWt!ewem); other examples will be 
given in their proper place. So also is explained suffix -dg^il 
"canoe" in such forms as tcddd^g^il "three canoes," seyatsd^g^il 
"five canoes", as compared with -dul in mdsdui "four canoes;" 
-dgHl is evidently from *-dwil (cf. KwantlEn, of Cowichan 
group of Coast Salish, -agitl "canoe" in numerals,'^ i.e., -axwil; 
perhaps cf. Comox n6xwU "canoe"). An interesting test case 
is qPgyas "deer," doubtless a loanword from Kwakiutl (cf. 
Kwakiutl gtwas "deer"^). Another such test case is afforded 
by Comox tigH^x" "nine" < *tAWAX'' or *tAwux'' (cf. KwantlEn 
tuq "nine,"^ i.e., tux or tux, contracted from *tuwux). Compare 
also Comox ht'g^os "chief" with PEntlatc and Siciatl htwus*. 
On the other hand a number of words have been found with w 
between vowels. Such are ts.'atsldwicin^ "hail," xwd'awVC 
"fire," and 'dwdk'" "tobacco." It is not clear how this -w- is 
. related to -w- > -g''-. 

Just as g^ and w are related, so there is reason to believe that 
dj and y are related, though there is perhaps not quite as con- 
vincing internal evidence at hand. See Type VIII of plural 
formations for such evidence. Moreover, with Comox djidis 
"tooth" compare KwantlEn yenis "tooth;"^ with Comox djicin* 
"foot" compare Siciatl yicin.^ 

* See C. Hill-Tout, Ethnological Studies of the Mainland IlalkomilBm, a division of the 
Salish of British Columbia, Report of British Association for the Advancement of Science, 
1902, Ethnological Survey of Canada, p. 65. 

' See F. Boas, Kwakiutl, Handbook of American Indian Languages, Bulletin 40, Bureau 
of American Ethnology, 1911, p. 447. 
' C. Hill-Tout, ibid., p. 64. 

* F. Boas, Comparative Vocabulary of Eighteen Languages spoken in British Columbia, 
Report of British Association for the Advancement of Science, 1890, 6th Report on the North- 
western Tribes of Canada, p. 148. 

' C. Hill-Tout, Ethnological Studies of the Mainland HalkomilBm, a division of the Salish 
of British Columbia, Report of British Association for the Advancement of Science, 1902, 
Ethnological Survey of Canada, p. 86. 

' F. Boas, Comparative Vocabulary of Eighteen Languages spoken in British Columbia, 
Report of British Association for the Advancement of Science, 1890, 6th Report on the North- 
western Tribes of Canada, p. 14' 

Eight: Wakashan and Salishan Languages 389 


A considerable number of Comox nouns always appear in 
reduplicated form, reduplication in these cases being of no 
grammatical significance, but belonging to the noun as such. 
Many of them are animal names, and of these some are quite 
evidently onomatopoetic. Ten fairly distinct types of redupli- 
cation seem to be illustrated in the rather limited material 
available. Very likely others exist. 

Type I. Completely Reduplicating. 

hd^mho^m blue grouse xop'xop' humming-bird 

k'^dck'^ac bluejay tsU'xHsHx" fish-hawk 

pok'^'pok"'' liver gH'gH'- panther 

qe'n'qen"^ duck qwVqwi' sea-gull 

"Duck" and "sea-gull" have both syllables with vowels alike 
in quality but with short vowel in the second. 

Type II. Completely Reduplicating with e. 

te'ltoV small butter-ball duck hd'ihei' arrow 
"Arrow" belongs perhaps rather with Type I. Both of these 
nouns lose a glottal stop in the reduplicating syllable. 

Type III. Reduplicating Syllable: cvci.' 

titctitcV c owl kiva'kwd'^djo' grey squirrel 

t.'Aq't.'Aqdi dog-wood 

Type IV. Reduplicating Syllable: ce. 

mf'mau cat k^It'k^Idk^! crow 

td'' salt-water hunter 
In "salt-water hunter" reduplicating icl- is broken into tcl'i-. 

' In these formulae c represents first consonant of stem, v first vowel, ci accond consonant 
of stem, vi second vowel, and so on. V represents any long vowel, <■ any shortened vowel. 

390 VI American Indian Languages 2 


Type V. Reduplicating Syllable: ci. 

Only one or two certain examples have been found of this type. 
They differ from the preceding in that the vowel of the redupli- 
cating syllable is short. 

qwi'^qwdH! .dd'^^k' butterfly we'wdlos young man 

(form probably diminutive in). 
Possibly also: — 

e'ddjam' young woman 

Type VI. Reduplicating Syllable: cd or ca. 

LdLdpx pocket-knife qwdqumi^s marten 

xdxe'* nit mdmstco^m mink 

Type VII. Reduplicating Syllable: cv. 

ts!ats!dwicin' hail tdatddH'.dn" mouse 

xwdxwadjo'm fly (word probably diminutive in form).'td'amas game with wooden ball^ 
qoqgwVm^ down (of bird) 

Type VIII. Reduplicating Syllable: cv'. 
qd''^qa'^ rush mat djd'''dja' tree 

Type IX. Reduplicating Syllable: eo. 

Only one example has been found of this type: — 
tofxHal necklace 

Type X. Reduplicating Syllable: cec. 

Of this very peculiar type (doubly reduplicating consonant, 
otherwise like Type IV) also only one example has been found: — 
qliqlq.'d'adje'uk''^ butter-ball duck 

' Formed from q'td'ahas "wooden ball covered with spruce-roots." There were two 
sides in the game, with the same number on each. Each side had a goal consisting of a little 
pit, which was guarded by one man. All but the two guards gathered in the centre. One 
man threw up the wooden ball and everyone tried to catch it, run with it to the goal of the 
opponents, and put it into the pit. Those of the other side tried to take the ball away from 
the one that had it. The side that 6rst made ten goals won the game. After four goals had 
been made, the game was suspended for a while and a general free-for-all fight took place. 

Eight: Wakashan and Salishan Languages 



Here may also be given: — 
qldq.'tux'* big fire (form is augmentative?) : 
scattered around. 

cf. q!6tix'' fires 


By far the larger number of Comox nouns form their plural by 
reduplication, in a few cases different stems are used for singular 
and plural, while still other nouns seem to form no plural. The 
most persistent type of plural reduplication is that in which 
both first and second consonants of stem are repeated, though 
'ess numerously represented types also occur. 

Type I. Reduplicating Syllable: cvci 

lAkoni"^ beaver 
kuvidqin^ sea-lion 
qwAdPs humpbacked whale 
qwASAm woolly grouse 
xop'xdp' humming-bird 
ts!oxd"° codfish 
LlAxwd'"' dog salmon 
sd^arC cohoe salmon 
qlwAt'lHcin" humpback salmon 
xd'd big clam 
Lp'Am' cockle 
xApd'^ red cedar 
qo'^'a''^ hemlock 
q!dp!xwai oak 

p'.tHxdi alder 

t.'t'ibdi wild cherry bush 

'dwdk''' tobacco 

qfwA'ix wood 
xd'a.idatc stump 

plural tlAk^HlAkom"^ 


qwxd'qw Adit's 








Lp'Lp'Am'^ (type viii?) 



q!ap!q!dp!exwai (with 
lengthening of first 
stem-vowel; -e- is in- 

p.'e'p.'t'ixdi (type viii?) 

HeH.'t'ibdi (type viii?) 

'au'dwdk'^ many bun- 
ches of tobacco 




VI American Indian Languages 2 


mAqsm nose 
djiciti^ foot 
djidis tooth 
dikuiriKS heart 
xAucirC bone 
k^it! little finger 
tslAmdla" index finger 
qlwdt'Am river 
pdxai' creek 
Lfdqe'nac spring 
kupu^'mi^x'' hill 

Lldxai' old man 

qAl'q! warrior 

L.'Ams house 

ajiSAm box 

kwd'am coiled storage basket 

L.'pdtjl basket bag 

q!dk'" board 

k^Iik^dyu oar 

SAq'Ak'" war-club 

lAq!" bow 

tcHCqdmin knife 

sipfAvitn" shinny stick 

lAq'.AS mountain-goat blanket 

L.'pi'ts.'d''^ yellow-cedar 

q.'As'Addi buckskin shirt 

Lldq.'acin" moccasin 

pdq'dos white-eyed 

tdxdos red-eyed 

plural niAqimAqsin' 
ts! Amis! AmaW 
kup'kupumt"x" (with 

shortening of second 

l! Ap' l! Apdtil 
tcHVtdit' qdmin" 
q! Asq! As' addi 

Type II. Reduplicating Syllable: cac. 

This type differs from the preceding in that, while both first 
and second stem-consonants are reduplicated, the stem vowel 
between these consonants is not, but is replaced by an inorganic 
A-vowel. If the vowel is followed or broken by a glottal stop, 
or if there are two successive vowels, the second consonant is 

Eight: Wakashan and Salishan Languages 



repeated just the same, the glottal stop being neglected in the 
reduplicating syllable. Thus, tc'.e'dd- and L.'d'al- reduplicate as 
tdin- and l!aI- respectively. Several nouns with stem-A and 
reduplicating-A, listed under Type I, should perhaps belong 
here. Three sub-types are to be recognized, according to 
whether a remains as such (sub-type a), is palatalized by s, tc, 
td, k^, I, or y to i (/) (sub-type 6), or is labialized by xw to u 
(sub-type c). 

Sub-type II a. 
mi'xdl bear 
L/d^ard'm' wolf 
qld^L.' land otter 
qld^sa" sea otter 
xd'^'wa fur seal 
dsx" hair seal 
k!6°ddt!° porpoise 

pIdqiAddtc goose 
qen^qerC duck 
hew^qen' swan 
qt'w'^x steel-head salmon 
td^qfwa' devil-fish 
mdt'.di horse clam 

sd^^ba" mussel 

mdHdirC louse 

dsd'i huckleberry bush 

xwdsAbdi soapberry bush 

t'.e'^^de'qwai salmon-berry bush 

t/d'abuxwdi gooseberry bush 

qex" ring finger 
bld^qlwdi fish-gill 
sopAdatc tail 
tsldmuql cloud 

plural mAxmi'xdl 






klwAd'k.'wd^dot!'? (with 
shortening of second 
vowel of stem) 


qAd' qeri' qerC 




mAt.'mdHldi (with length- 
ening of first vowel 
of stem) 



^ As'dsdH 



t! Amt! Amuxwdi (with re- 
duction of d'a of stem 
to a) 






VI American Indian Languages 2 


tld'^qfaV mountain 

st'qeV dug hole, well 

td'mic man 

xd''p! baby basket 

t.'o'mV paddle 

waxdHsU pipe 

toVxHal necklace 

q'td'ahas wooden ball used in 

mitdli beaver-tooth die 

plural t!Ag!t!d''q!aC 

mAfmiHdli (with length- 
ening of first stem 



qd'^qa rush mat 

Idq.'wdinop cedar-bark mat 

L.'dxe oldest dAXhldxe 

Lldlsdmi strong LfAlLldlsdmi 

An irregular example of this sub-type is: — 

sdts'.Am tyee salmon SAmsdHs! Am 

Here the first and third, instead of first and second, consonants 

are reduplicated. 

Sub-type II 6. 
tde'ddg dog 
k^dck^dc bluejay 
l6"''honC small clam 
tsfdtc.'ilbai spruce 

sdsin'^ mouth 

sdpdxos horn 
ko^sAd' star 
ydxai^^ pack-basket 




taHtc'ts.'dtcHlbai {tsUtc"- 
instead of tsUtc!-) 

sissosirC (with shorten- 
ing of second stem- 




Irregular examples of this sub-type are: 
idatddH.'dn" mouse 

Itx^sal tongue 

tclUHddH.'dn" (for tdiC- 
instead of tdit.'-see 
"spruce" above) 


Eight: Wakashan and Salishan Languages 



In the first of these the plural is built not on the already redupli- 
cated simplex (as e.g., in "bluejay" above), but on a simpler 
unreduplicated stem abstracted from it. In the second example 
the first and third, unstead of the first and second consonants, 
are reduplicated (cf. "tyee salmon" above). 

Sub-type II c. Only one example is available: — 

xwdtoqo'm "falls" plural xuCxwdtoqo^m 

Type III. Reduplicating Syllable: co or co. 

Nearly all of these nouns have g" as their second consonant, 
representing, as we have already seen, original w. These nouns 
could be considered a sub-type of Type II, were it not that they 
form their reduplicating syllable not in -au, as might perhaps 
be expected (cf. xAuxAUciii'^ under Type I), but in -o- (-w- after 
dj- and g^-) or -o- (probably due to contraction of original -yiw-). 
Two sub-types can be recognized, according to whether the 
reduplicating vowel is short (sub-type a) or long (sub-type 6). 

Sub-type III a. 
td^ag^ax"^ fern 
td'agHn salmon spear 

Sub-type III b. 
Id^g^et!" herring 
pleg^di halibut 
gT/figvp panther 

t.'cg^em sun, moon 
Mg^os chief 
djigHn' song 
Id^dak'" skin 



*g^ug''Vg"l^ (not obtained 
as such, but implied 
in diminutive plural 
gyVg^ugH'gH' "pan- 
ther cubs") 

tldH.'eg^em sun and moon 




It is not clear why "skin" should reduplicate with o-vowel. 

396 ^I American Indian Languages 2 


Type IV. Reduplicating Syllable: cv; Syncope of First Stem 


Only one example has been found of this type. As it begins 
with g"-, the stem -g^ of the plural, coming immediately before 
another consonant, reverts to it;, uniting with preceding a to 
form au. 

g^dq'dhas married woman plural g"duq'dhas 

That *wdq'dhas is to be presupposed is corroborated by com- 
parison with KwantlEn s-wd-wskus "married woman. "^ 

Type V. Reduplicating Syllable: cvc. 

Nouns belonging to this group have long stem-vowels and 
differ from Type I in that the reduplicated vowel is shortened, 
though it keeps its quality. 

xdug"as grizzly bear xduxdug^as 

qd'um' eye qduqd'um' 

qd^'mai" snow on ground qiirnqo"' mai' 

}6kd°min bailer luk'loko'^min 

Type VI. Reduplicating Syllable: caci. 

ti'hd''ddn' chief's wife tdhtihd''ddn' 

heq'sd^min pole for poling canoe hdq'heq'sd°min' 

6lqai'' snake 'dVolqai'' (with shorten- 

ing of first stem- 
aL leggings 'dt^aL 

"Leggings" may, of course, just as well belong to Type I. 

Type VII. Reduplicating Syllable: cv. 

q!6a'dda ear q!oq!oa'dda 

tt (L.'.-ims) big (house) titl (LfAvis) big (houses) 

dx" snow-flake a'dx" falling snow 

' C. Hill-Tout, Ethnological Studies of the Mainland HalkomilBm, a division of the Salish 
of British Columbia, Report of British Association for the Advancement of Science, 1902, 
Ethnological Survey of Canada, p. 89. 

Eight: Wakashan and Salishan Languages 


Type VIII. Reduplicating Syllable: ce. 

According to varying phonetic circumstances we have either 
I or e, the latter occurring after q, g! and x. The examples of 
this type obtained are: — 

qldik''' eagle 
kwudjak''^ trout 
ft'x" yellow cedar 

djd^'dja' tree 

sd'idJA^ leaf 
tcdyac hand 
sayd^ada neck 
qd^ya' water 
s&'yal lake 
xd'adjaic stone 
tcu"i child 

k.'oyokobPn (or -mVd) fisherman 
sidjdqo'p' basket hat 
IdidatdAn woman's cedar-bark 

plural q!t'q!dik'" 


tltPxwai' (may belong 
also to type vii; note 
-ai' in plural) 

*djidjd'''dja" (not ob- 
tained as such, but 
implied by diminu- 
tive plural djedjidjd- 








kfwtk! oy okomi 'n 



Eliminating "yellow cedar," which, as was pointed out, may 
just as well be reckoned as belonging to Type VII (there is 
reason, however, to believe that il'x" goes back to *tiyix"; see 
diminutive type I b and diminutive plural type II f), all these 
plurals may be plausibly explained as cases of Type II, redupli- 
cating -I- or -e- being the contracted result of -Ay-. It will be 
observed that the stems of these nouns contain either i- diph- 
thongs, including broken groups (-di-, -a'l-, -u'i-), vowel plus y 
{-ay-, -ay-, -a'y-, -oy-), or vowel plus dj {-udj-, -d'^dj-, -d^adj-, 
-idj-) ; dj, as we saw above, is probably a resultant of original y. 

398 VI American Indian Languages 2 


Type IX. Reduplicating Syllable: cd (or ca). 

Sub-type IX a (with d). 

tdd rain plural tcldtclel 
qd'"qwai speaker qiuaq6'"qwai 

ylp'i'x" hole ydyipl'x'* 

Sub-type IX b {with a). 

tct'^ salt-water hunter* 

Type X. First Stem-vowel Changed to e. 

These nouns are reduplicated to begin with, and substitute for 
plural reduplication a change of the first stem-vowel to e (long 
and open). 

The few examples are: — 

we'wdlos young man we"wdlgs 

e'ddja7n' young woman e"ddjam' 

ky.'t'ky.'dky! crow k«!e''k«!dk«! 

Type XI. Reduplicating Syllables: cdCAC. 

Only two examples have been found of this doubly reduplicat- 
ing type of plural formation. In the first, the a, coming after 
g", is palatalized to i; in the second, the reduplicating -Ay- 
becomes -I- (see Type VIII). 

g^d^di'm slave gyagndg^ddPrn 

tdyac killer-whale idtltdyac 

Irregular Plurals. 

Several plurals listed above are somewhat irregular, but there 
has been no difficulty in assigning them to definite types. The 
two that follow are quite irregular. The second shows not only 
reduplication but breaking of -a- to d'a-. 

djd'°dja' tree djadj'id'm 

inAVq" fawn mamd' aliq' " 

Eight: Wakashan and Salishan Languages 399 


A few nouns change the stem entirely in passing from singular 
to plural. Such are: — 

sQ.lV'^ woman plural nig'-^dp'tai 

sa'as?r" girl (diminutive of ssiir") ninig^ap'lai (diminutive 

of nig^dp'tai) 
Involving this same change of stem is: — 
sdHux'^ married man nig^dpHahai' 

Rather different, presumably, is: — 

tatfndtcap^ leg tcuklu'ndtcap' 

which keeps the same suffix in the plural, while changing the 

Nouns without Plurals. 

Quite a number of nouns were secured which form no plural. 

Some of these are reduplicated to begin with, and there is clearly 

a feeling, though one by no means consistently applied, against 

re-reduplication in forming plurals. Others, however, are such 

as might easily be reduplicated, were it usage to do so. It is 

possible that reduplicated plurals might have been given for 

some of these by other informants. Reduplicated nouns that 

form no plural are: — 

gypgyp panther qivdqumi^s marten 

titditcVc owl mdmstco'm mink 

is'.i' xHs'.ix'' fish-hawk qwVqwV sea-gull 

qwi^qwdH! Ald^'k" butterfly xivdxwadjo'77i' fly 

(probably diminutive: dim. plur. 

is found) 

pok^^pok'"^ liver xdxe'i nit 

LaLdpx"^ knife hdiihei' arrow 

Non-reduplicated nouns for which my informant would give 

no plurals are: — 

mdyos raccoon pVk! ground-hog 

qlt'etc elk p.'d'alats!" skunk 

v!dxd'° \ ^dmaxHdjo'" ant 

^ ■■ Vraven ,. • •" 

p. 'ah ) qeix salmon-egg 

tdeq" robin mo'os head 

For "robin," tdAq'tcfeq', which might well enough l)e expected 

as plural, was explicitly denied. If necessary to express plurality 

in these nouns, qax or q.ix "many" can be juxtaposed before 

any of thorn. 

400 VI American Indian Languages 2 



Diminutives in Comox, as in other Salish languages, are formed 
by means of reduplication. Reduplicated diminutive forms, 
however, differ from reduplicated plurals in that the reduplicat- 
ing syllable rei:>eats the first consonant of the stem, never also 
the second. Moreover, the vowel of the reduplicating syllable 
is formed according to different rules from that of the redupli- 
cating syllable of plural forms. Further complications result 
from the internal changes to which the stem is often subjected, 
so that altogether a large number of more or less distinct types 
of diminutive formations may be recognized. It will be advan- 
tageous to list in a purely analytical way the various features 
that are found in diminutives, so that ready reference may be 
made to them when discussing the types as^such. 

Diminutivizing characteristics are: — 
{I. ) Reduplication of initial consonant of stem, followed by 

a. Short e (i or i). Two types of c- reduplication may be 
recognized, according to whether e is or is not accented. 
Thus, mimosas from yno^os "head"; qeqn^ya' from qd^ya^ 

b. Long e (i or i), always accented. Thus LllLlAXWd'* 
from lIaxwo''- "dog-salmon." 

c. e, always accented. Thus qfe'^q'.e^L! from q.'d'^L!" land- 

d. V, which may or may not be accented. Thus, lolko°min 
from loko^min "bailer." 

e. V, which is regularly accented. Thus, k!6k!gddf!° from 
k!6"ddt!'^ "porpoise." 

f. Short a, accented or not. Thus, LdLV'tm'^ from lV' attC 

g. Long a. Thus, djddjd''gHn' from djig^in' "song." 
h. Long d'a. Thus, sd^aslV^ from sd/f" "woman." 

i. Short 0. Thus, LloLld'ami's from dAms "house." 
(2.) Glottal stop inserted in stem. This may occur as 

a. Breaking of (non-final) yowel or diphthong. Thus, 
tcitcd^'^yac from tcdyac "hand." 

b. Glottalizing of final consonant (generally m or n); this 
should probably include breaking of vowel when final. 
Thus. Idtbo'm' from 16'^^hom' "small clam." 

Eight: Wakashan and Salishan Languages 401 


(3.) Quantitative vocalic changes (increments). These include 

a. Lengthening of (last) stem voweL Thus, tdtig'^dx" from 
td'agyax"" "fern." 

b. Change to wd or wa of u of stem. Thus, diminutive 
plural kwikumkwd'''mdqin' from plural kumkumdqin" 

c. Lengthening of inorganic a (or i. e) to i. Thus, xexsim' 
from XASAm "box." Less often, full a is changed to i 
(cf. 4b), as in kwekwi'im' from kwd'am "coiled storage 

d. Insertion of i. This is probably but another form of 3c, 
inorganic a and absence of vowel being perhaps con- 
sidered as phonologically equivalent. Thus, qeqAli'q! 
from qAl'q! "warrior." 

e. Insertion of short vowel (a, i) before syllable with length- 
ened vowel. Thus, xt^xigHctn^ (note second i) from 
xAucin^ "bone." 

f. Lengthening of a or a (non-final) to d. Thns, qlwdq.'wd''^- 
djix from q'.wAHx "wood." 

(4.) Qualitative vocalic changes. These include 

a. Umlaut of a to short e (i). Thus, xexd'adjeHc from 
xd'adjaic "stone." 

b. Umlaut of a {or d), rarely o, to long e {i, i). Thus, 
qle^qle^k'^ from qfak^"- "board." 

c. Change of stem vowel to a'a. Thus, tot&'amic from td'mic 

(5.) Vocalic reduction. Under this head may be grouped 

a. Shortening of stem vowel before syllable with lengthened 
vowel (or inserted i). This shortening before lengthening 
is doubtless due to quantitative rhythm. Thus, qicrl''- 
qwi'qwV* (note second i) from gmV/iw' "sea-gull. Such 
shortened syllables regularly lose their glottal stop, if 
there is one present, as in qeqawem" from qd'um' "eye." 

b. Syncope of stem vowel after reduplicating syllable with 
accented vowel. Long vowels may thus fall out quite as 
readily as short ones. Thus, se''sp'xos from sdpdxos 

These twenty-two diminutivizing features occur in various 
combinations, so that a large number of possible types of 

402 VI American Indian Languages 2 


diminutive formation maj' result. A considerable number of 
such types can be constructed from the available material, but 
this need not exemplify all that actually occur. As to which of 
the features listed are fundamental to Salish and which merely 
secondary in Comox or several Coast Salish languages, it is 
useless to speculate. Adequate comparative data are necessary. 
A few points of a comparative nature will be brought forward at 
the end of the paper. The various diminutive types will now 
be taken up in order, the main stress being laid on the form of 
the reduplicating syllable. 

Type I. Reduplicating Syllable: ce. 

Various sub-types occur, according to whether the stem 
vowels remain unmodified or are subjected to comparatively 
slight changes. 

Sub-type I a. Diminutive feature la only: — 
mo' OS head diminutive mimo'os 

qtaabas wooden ball qeq'td'abas 

q.'ga'dda ear q!weq!oa'dda 

'dwdk^'' tobacco 'e'awdk'" 

qwAdi^s whale qweqwAdi^s 

qo'^'a'^ hemlock qioiqd'^'a'' 

sidjdqo'p' basket cap sisidjdqo'p' 

qeix salmon-egg qeqeyix {-eyi- prob- 

ably merely variant 
of -ei-) 

Sub-type I b. Diminutive features la, 3c (or d) : — 

mAqsirC nose mimAqsi^n' 

djidis tooth djldjidVs 

^I'x" yellow cedar {<i*tiyix") titiylx'^ 

qAl'q! warrior qeqAWq! 

Sub-type I c. Diminutive features la, 3a, 5a: — 

/■ey^os chief heheg'-'d"s 

Eight: Wakashan and Salishan Languages 403 


Type II. Reduplicating Syllable: ce; stem: feature 2a. 

In these diminutives the first vowel of the stem is broken, the 
broken vowel taking the form v'v. If the final vowel is long, it 
seems to be shortened (-'d becomes -'"). 

pi'k! ground-hog diminutive pipi'ik! 

qfs^etc elk gleq'/e'e'etc 

xd'd big clam xexA'a'" 

Though the last diminutive seems to correspond exactly in 
form and rhythm to the second, the final -a'" may perhaps here 
be better explained as breaking of the last vowel (-d) of the stem 
(feature 2b). 

Type III. Reduplicating Syllable: ce: stem: features 3a 
or d, 5a, 2b. 

qd.'unC eye qeqaweni" {-e- doubt- 

less merely variant 
of -1-) 

qwPqui" sea-gull qv/i^qwi'qwi"^ 

Type IV. Reduplicating Syllable: ce; stem: features 4a, 3c, 26. 
kwd'am coiled storage basket kwekwi'im' 

Type V . Reduplicating Syllable: ce; stem: feature 5b. 

hHkuinAS heart LlidkuinAS 

xApd^^ red cedar xexpd''' 

Type VI . Reduplicating Syllable: ce; stem: features 5b, 


qwAsAm wooll}^ grouse qweq'"se''m-ol 

pdxai' creek pfp'xe'* 

Type VII. Reduplicating Syllable: ce; stem: features 5b, dc, 2a. 
XASAVi box xexsim' 

404 VI American Indian Languages 2 

Type VIII. Reduplicating Syllable: ce. 

In this type the redupHcating e is unaccented. According to 
whether or not the stem is modified in regard to vocalic length or 
quality, various sub-types may be recognized. 

Sub-type VIII a. Diminutive feature la only: — 

tsldtdilbai spruce diminutive tsUts.'dtcHlbai 

qd'ya' water qeqd'ya' 

qd'^qwai speaker qwiq6'''qwai 

Here probably also belongs qwi'qwdH! Ala^'k" "butterfly." 

Sub-type VIII b. Diminutive features la, 5a (accent on third 

syllable of diminutive) : — 

sayd'ada neck dsiyd'ada (sa-short- 

ened to sa-, which, 
coming before y, 
has to be palatal- 
ized to si-) 

xa'aidaic stump xexA'd.idatc (-a.i- 

probably equiva- 
lent to -di-) 

Sub-type VIII c. Diminutive features la, 3a (or c) : — 
sd'yal lake sisd'ydl 

mdHdirC louse mimdHdin' 

Sub-type VIII d. Diminutive features la, 4b: — 
pdh'^'pok'"' liver pipuk'^pVk'*' 

Type IX. Reduplicating Syllable: ce; stem: feature 2a. 

Here again the reduplicating vowel is an unaccented e. The 
stem, however, is characterized by the breaking of one of its 
vowels. According to whether or not umlaut also takes place, 
two sub-types are to be recognized. 

Sub-type IX a. Diminutive features la, 2a: — 
tcdyac hand tcitcd'''yac 

Eight: Wakashan and Salishan Languages 405 


Sub-type IX b. Diminutive features la, 2a, 4a: — 
xd'adjaic stone diminutive xexd'adje'ic 

As irregular representative of this type may perhaps be con- 
sidered : — 

qe'n'qen' duck qeqA'dd-ol (built on 

unreduplicated sim- 

Type X. Reduplicating Syllable: ct. 

Various sub-types are to be recognized, according to whether 
or not the stem vowels are quantitatively modified. 
Sub-type X a. Diminutive feature lb only: — 

tclel rain tcHtdel 

pleg^di halibut plVpHg^di 

tsfoxd"" codfish ts!i'ts!ox6'° 

LlAxwd'^ dog-salmon LltiJAXwd'^ 

x&^p! baby-basket xt'xd^p! 

p!oxo^° raven pH'p.'oxd'' 

yip'Vx" hole ytyipVx'^ 

L.'pi'ts.'d'"- yellow-cedar bark l!i' dpttsfa'^ 


titctitcl'c little owl tititctitcl^c 

knacky dc bluejay kH'^k^dc 

qoqgwVm' small breast feathers qivi'^quwihn' 

In the last two examples the diminutive is formed, not from 
the already reduplicated simplex, but from the unreduplicated 
form abstracted from it. 
Sub-type X b. Diminutive features lb, 3c: — 

Llpdtil bag hlt'Llpdlil 

tcleq'' robin tdiHcH^q' 

kumdqin' sea-lion kwi'kumdqip'^ 

djicin' foot dji'djichC 

Uq!'' bow li'Hi'q!'* 

Sub-type X c. Diminutive features lb, 5a, 3c (or 3a) : — 

qlwAt'lHcin" humpback salmon qlwtq'.utVtcin" 

t.'d'abuxwdi gooseberry bush tUt! Amuxwdi 

pld'alats!" skunk pltp'.A^ldts! (mis- 

heard for -p!aI-?) 

406 VI American Indian Languages 2 


Sub-type X d. Diminutive features lb, 3e, 3c:— 

xAucin' howe diminutive arg'xrg^/dn' {-igH-< 


It should be noted that this type of diminutive formation, 
while externally similar to Type VIII of plural formation (of., e.g., 
xt'Td^p! "little basket" with qlt'q.'dik'" "eagles"), is in reality 
quite distinct in origin, the latter, as we have seen, tracing its 
reduplicating -e- to -Ay- and being limited to nouns with z'-diph- 

Type XI. Reduplicating Syllable: ct; steyn: feature 26. horse clam me'mAtfa''- (me'- per- 

haps misheard for 
hdihei {hdihei'7) arrow htheihei'^ 

qlAs^Addi buckskin shirt q!t''q!as'add'i 

Type XII. Reduplicating Syllable: ct; stem: 46. 

g.'dfc'" board qlt'qle'k"'' 

aL leggings t'eL 

Type XIII. Reduplicating Syllable: ct; stem: feature 5b. 

There are two sub-types, according to whether or not the stem 
vowel is modified. 

Sub-type XIII a. Diminutive features 1 b, 5 b: — 

q'.dpfxwai oak qliq.'plxivai 

pleixdi alder p'.i'plxdi 

L!d"q!wdi fish-gill Lli'^Uqlwdi 

kup-il''nn"x'' hill kivik'"p-VV 

Idqlwdinop' cedar-bark mat H'lq.'wdinqp'^ 

V.eHbdi wild-cherry bush t'.i'V.bdi 

i'.Aq't'.Aqdi dog-wood t'.et'.qdi 

In the last example the diminutive is built up on the unre- 
duplicatod stem abstracted from the already reduplicated 
simplex. The broken stem vowels -ei- of "alder" and "wild- 

Eight: Wakashan and Salishan Languages 407 


cherry bush" disappear in the diminutive apparently without 

trace of ',. but this may in part be due to following q! and p! , 

which imply '. With these contrast: — 

sd'an' cohoe salmon diminutive sts'ad-ol 

Here the -a'a- is treated, not as a broken vowel, but as two 

vowels with intervening consonant. 

Sub-type XIII b. Diminutive features lb, 5b, 3c: — 
lAq'.AS mountain-goat blanket li'l'qUs 

LlAqlacin"^ moccasins Llt^Uqlaccn'' (mis- 

heard for -en"?) 

Type XIV. Reduplicating Syllable: eg; stem: features 5a, 

Sc, 2b. 

tUg^evi sun, moon tH'tHgHm^ 

-i- is for -A-, because of following g^. 

Type XV . Reduplicating Syllable: ce; stem: features 5b, 4a. 
SAg'iA;'" war-club sVsqek'^ 

Type XVI. Reduplicating Syllable: ce'. 

qd'^qa' rush mat qe'^qa''^ 

toVxHal necklace te'H'x"lal 

The diminutive of "necklace," as often happens with nouns 
reduplicated to begin with, is built uf) on the implied unredupli- 
cated stem. The same applies to the diminutive of "rush mat," 
except that here it is the reduplicating syllable of the simplex, 
which doubtless more nearly represents the simple stem, that is 
taken as the base of the diminutive form. 

Type XVII. Reduplicating Syllable: ce'; stem: feature 4a orb. 

Two sub-types are found, according to whether or not there 
areat the same time quantitative changes in the stem. 

408 VI American Indian Languages 2 


Sub-type XVII a. Diminutive features Ic, 4b: — 
qfa^t! land-otter diminutive qle'^qle'L.' 

qld^sa" sea-otter q!e''q!e^s (note loss of 

Sub-type XVII b. Diminutive features, Ic, 5a, 4a (or b) :— 

qt'W^x steel-head salmon qe'qeg^e'x 

-gy- is from original -w-. It is not clear whether -qeg^e^x repre- 
sents *-qewe'x or *-qewex. 

Type XVIII. Reduplicating Syllable: ce; stem: features Sc, 2b . 

q.'wdt'Am river q!we'q!wat'im' {qlwe'- 

not equivalent to 
q!we^-; see diminu- 
tive plural type iv) 

Type XIX. Reduplicating Syllable: ce'; stem: feature 6b. 
There are two sub-types, the latter with modified stem vowel. 

Sub-type XIX a. Diminutive features Ic, 5b: — 
sdpaxos horn se''sp'xos 

heg'sd^min" pole for poling canoe he'^hq^sd^min" 

Sub-type XIX b. Diminutive features Ic, 5b, 3c: — 
tlAkom" beaver (-/;p- doubt- Ue'Hikwim' 

less for -kwA-) 

Type XX. Reduplicating Syllable: cv. 

Here again there are two sub-types, the latter with vocalic 

Sub-type XX a. Diminutive feature Id: — 

xdug^as grizzly bear xdxdug^as 

Here probably belongs also xwdxadjo'm' "fly." 

Sub-type XX b. Diminutive features Id, 5a, 3c or d: — 

Id^gyet!" herring Idlig^et!" {-i- < -a-) 

IdidatctAu woman's cedar-bark laltdatctin {-I- <-Ai-) 

t!6'mV paddle tlgUAbiH' 

Eight: Wakashan and Salishan Languages 409 

Type XXI. Reduplicating Syllable: cv; stem: feature 5b. 

There are three sub-types, based on diflferences in the further 
treatment of the stem. 
Sub-type XXI a. Diminutive features Id, 5b: — 

yaxai'* pack-basket diminutive yd.ixai'* 

Sub-type XXI b. Diminutive features Id, 5b, 3a: — 
waxdHsU pipe wauxdHsli* 

Sub-type XXI c. Diminutive features Id, 5a, 5b: — 
I6kd°mtn bailer l6lko"m%n 

Type XXII. Reduplicating Syllable c'^; stem: features 3a or c, 

and 26. 

There are two sub-types, depending on whether or not the 
first vowel of the stem is reduced. 

Sub-type XXII a. Diminutive features Id, 3c, 2b: — 
sdtslAin tyee salmon sas'dHs'.VnC 

Sub-type XXII b. Diminutive features Id, 5a, 3a, 2b: — 

sd'idJA^ leaf sasidjd'" (-1- reduced 

from -a'i-) 

Type XXIII. Reduplicating Syllable: cv; stem: features 5a, 

Sa or c. 

kfd'^ddt!'? porpoise k!6k!od6th 

mdyos raccoon mdmiyo°s (-?'- palat- 

alized from -A-, re- 
duced from -a-) 
td'ag''ax'' fern tdtig'-'dx"" {-i- palatal- 

ized from -A-, re- 
duced from -d'a-) 
td'ag^in salmon-spear tdtigHn (dit.) 

y^d^di^m slave g^dg^idVm (-i- pala- 

talized from -.1-, re- 
duced from -d"-) 

410 VI American Indian Languages 2 


tn^dak'" skin diminutive idlidd'^k''' (dit.) 

(Vf.r" hair-seal 'a'asix" 

ts.'amuql cloud ts!dts!imAgwil (-/- 

palatalized from-x-, 
reduced from -a-; 
-wiA- merely vari- 
ant of -mu-) 
olqai'"^ snake ^d^olqai'^ 

'dmax^idjd'" ant 'd^imax^idjo'" 

In the last two examples the final vowel is considered quanti- 
tatively long and hence cannot be further lengthened. Quite 
irregular is: — 

tdyac killer-whale tdtlyac 

The long -i- and the short -d- of the stem are the exact reverse 
of what would be expected (*tdtiydc, cf. tdtig^dx'' above). 

Type XXIV. Reduplicating Syllable: cv; stem: features 

5a, 46. 
tdatddH.'dn" mouse tddtcHt.'in" (-i- pal- 

atalized form of -A-, 
reduced from -d"-) 
The diminutive, as often, is based on the unreduplicated 
stem abstracted from the already reduplicated simplex. 

Type XXV. Reduplicating Syllable: cv; stem: features 5a, 2b 

Two sub-types are to be recognized, depending on the treat" 
ment of the last vowel of the stem. 

Sub-type XXV a. Diminutive features la, 5a, 2b: — 
l6"°bom' small clam lol'^bo'm" 

kloyokobVn fisherman kfokloyokobVn" 

Sub-type XXV b. Diminutive features le, 5a, 3a, 2b: — 
xd^'wa fur seal xdxAwd'a 

sd'^^ba' mussel sdsAbd'" 

tc.'e'ddo dog tde'dtdidd'" {-i- pala- 

talized from -.4-, re- 
duced from -e'd-) 
In the last example -e'd is treated as a reduplicating long vowel. 

Eight: Wakashan and Salishan Languages 411 


Type XXVI. Reduplicating Syllable: cv; stem: feature 5b. 

Three sub-types are to be recognized, according to whether 
the stem undergoes no further change or is further modified. 

Sub-type XXVI a. Diminutive features le, 5b: — 

sOstn' mouth diminutive sossin" 

pldqlAddtc goose p.'dp.'qiAddtc 

ti'hd'^ddn' chief's wife tiVhdddn' 

sopAdatc tail so^'spAdatc 

xwdsAbdi soapberry bush xwdx"sabdi 

tix'^sal tongue tlHx'^sal 

osd'i huckleberry bush '6'ASd'i {-'os- cannot 

be further reduced 
than -'as-) 

mi'xdl bear mt'mExdl {-e- is mere- 

ly glide) 

siplAmirC shinny stick sPsp! Ami^n" 

mitdli beaver-tooth die mi'm{i)tdli i-i- is 

merely glide) 

ky'.ikydyu oar k^!t'k«!k^dyu 

si'^qeV dug hole, well si'^sqeV 

"Bear," "shinny stick," "beaver-tooth die," and "oar," 
which have short stem-vowels, are perhaps better listed with 
type X. 

Sub-type XXVI b. Diminutive features le, 5b, 3c: — 

kd'^SAd' star kok'sid' 

Sub-type XXVI c. Diminutive features le, 5b, 5a, 3a: — 

V.e'^'de^qwai salmon-berry bush V.eV.dAqwdi 

Type XXVII. Reduplicating Syllable: cv; stem: features 

5 b, 4 b. 
tld^qlaV mountain t!dt!q!eH" 

Type XXVIII. Reduplicating Syllable: cv; stem: features 5b 

(or a), 3a, 2a. 
td'^qlwa' devil-fish tdCqlwd^" 

djd'^dja' tree djddjidjd^" {-i- pal- 

atalized from -.4- 
reduced from -a"-) 

412 ^f American Indian Languages 2 


Type XXIX. Reduplicating Syllable: ca. 

Two sub-types have been found illustrated, each represented 
by but one example in the material obtained. 
Sub-type XXIX a. Diminutive features If, 3c, 2b: — 

Lp'Am' cockle diminutive LdLV'im'' 

Sub-type XXIX b. Diminutive features If, 3b, 2a: — 

kwudjdk'" trout kwakwd'^djak''' 

Type XXX. Reduplicating Syllable: ca; stem: feature Sf. 

Two sub-types may be recognized, the second with further 
modification of the stem. 
Sub-type XXX a. Diminutive features Ig, 3f : — 

qlwA'ix wood qlwdqlwcC^djix {-dj- 

<*-y-, glide be- 
tween -a'"- and -i-). 
Sub-type XXX b. Diminutive features Ig, 3f, 3c: — 
djig^in" song {<*dJAWAn') djddjd'gHn" 

Type XXXI. Reduplicating Syllable: cv'v; stem: feature 5b. 

sdlV"* woman sd'aslV'' girl 

Lfd'ard'm" wolf Ud'aLfrd'm' 

Type XXXII. Reduplicating Syllable: co; stem: feature J^c. 

Two sub-types, each represented by one example, are found, 
the second involving a further change of stem. 
Sub-type XXXII a. Diminutive features li (perhaps rather 1 
d), 4c:- 

td'mic man totd'amic boy 

Sub-type XXXII b. Diminutive features li, 4c, 3d: — 

L'.Ams house Lfoild'amPs 

Diminutive in -ol, -ol. 

Besides forming diminutives by means of reduplication and 
internal stem change, Comox can also make diminutives of 
animal nouns by means of a suffix -dl{t'") or -oi(r"). Some of 
the diminutives in -o?(r") or -ol{V") are nouns whose simplex is 

Eight: Wakashan and Salishan Languages 



already reduplicated (cf. reduplicated nouns which form no 
redupHcated plural), yet not all. Of those formed from unre- 
duplicated nouns, somo have diminutive reduplication at the 
same time, others not. By an interesting phonetic law of 
rhythmic balance -o?(r") is suffixed to stems whose last vowel 
is short, -pZ(r") to those whose last vowel is long. The examples 
obtained of the suffix are: — 
1. -dl{V-) 

hew^qen' swan diminutive Jiew^qAdol 

mi'^mau cat 
q.'dik'" eagle 

qe'n'qen' duck 
tsfitsqU'^nas chicken hawk 
tdeq' robin 

sd'arC cohoe salmon 

xdp'xop'^ humming bird 
gvfigvii panther 


little eagles 
ts!itfiq!e'^nasdlV " 
tditdeqHdeqolV" little 
[sisosg^dddl plur. 

The last two seem irregular as regards rhythmic balance; 
perhaps they were respectively misheard for *x6p'xopdlt''' and 
*gH^gHyul. -ol has also been found in mim'ini'dl A;'" mdmstco'm 
"little mink." 
2. -p^ (r") 

hdhnhd'm blue grouse hd^mho^mol 

qwdqumi^s marten qwdqunii^solV" 

qwAsAm woolly grouse qweq'"se^mol 

tsU'x^tsHx^ fish-hawk tsU'xHsUxwol 

kwa'kwd'^djo' grey-squirrel kwa'kwd'^djgl 

414 Vf American Indian Languages 2 




The plurals of diminutives are, as a rule, doubly reduplicated, 
the first reduplicating syllable expressing the diminutive idea, 
the second that of plurality; the first reduplicating syllable is 
almost invariably of diminutive type, the second of plural type. 
Hence diminutive plurals are morphologically, and psycholo 
gically, diminutivized plurals, not pluralized diminutives. 
While they may be said, on the whole, to be formed from the 
plural of the simplex, the diminutive singular has often influence 
on the form of the diminutive plural, both as regards the inner 
stem changes and the vowel of the reduplicating syllable. Thus 
diminutive plurals may be said to combine, roughly speaking, 
the characteristics of both the plural and diminutive of the 
simplex. In order better to understand the formation of the 
diminutive plural and to assist in cross-referencing, the types to 
which the non-diminutive plural and the diminutive singular 
belong will be indicated in the following lists. 

Type I. Reduplicating Syllable: ce; followed by plural of simplex. 

The reduplicating syllable is analogous to that of diminutive 
types X, XI, XII, XIII, XIV, and XV. According to whether 
or not the remaining part of the word is somewhat modified 
from the plural of the simplex, sub-types may be recognized. 

Sub-type I a. Plural of simplex unchanged: — 

LllkuiriAS heart 



dim. V. 

dim. plur. lU'lIeJc"- 

tsloxo"" codfish 


x a 


lIAxwo''' dog-salmon 




Lfpt'tsld'" yellow 


bark basket 




tc.'eq'' robin 

no plur. 



I implied 


in dim. plur.) 

aL leggings 


[. (or VI.) 



kupu''mi"x" hill 



a. kwlkup'kup-lH' 
(with -IH' as in 
diminutive singu- 

Eight: Wakashan and Salishan Languages 


k"!ik''dyu oar 


plur. dit. dim. 

sip.'Amtn' shinny-stick dit. 

qwdqumi's marten no plur. 

(type I, based 
on stem-form of 
simplex, im- 
plied in dim . 

xd^p! baby basket ii a. 

t.'d'abuxwdi goose- 
berry bush dit. 

Idqfwdinop' cedar- 
bark mat dit. 

dd'qlwdi fish-gill dit. 

k.'o^dot!'? porpoise dit. 

tsldmuql cloud dit. 

mitdli beaver-tooth 

XXVI a. dim. plur. k^Hk^li- 
dit. sisip!sip!amVn' 
-olV'* qwtqumqwdqumVs 

X a. xexAptxd'^p! 

X c. t!it! Amt! Abuxwdi 

XIII a. lilAqfHdqfwdinop'^ 

dit. lUl! Aq!" ild^qlwdi 

XXIII, k!wik!wAd'k!6°doi!^ 

dit. tsUts.'Amtsldmuql 



XXVI a. 


mi^xdl bear 

II a. 

XXVI a. 

{-E- is glide) 

st^eV well 




Dsd'i huckleberry 





sdpAdatc tail 




td'mic man 




g^d'di'm slave 

XI. (i 





g"d- belongs 

(based on -gHdg^d 

to II 


dVm of plur.) 

ydxai'^ pack-basket 


XXI a. 


tdatc.'dHIdn" mouse 




kd^'SAd' star 


XXVI b. 


pleg^di halibut 

Ill b. 

X a. 


gypgvii panther 

no pi 





Ill b 



in dim 



olqai'* snake 





VI American Indian Languages 2 


sd'idJA' leaf plur. viii. dim. xxii b. dim. plur. smsd'trfj^' 

djd'^dja' tree irregular xxviii. djedjidjd^''dja'' 

(built on plur. of 
type viii) 
yipi^x'* hole ix. x a. yeyip'yip'l'x'' 

(built on plur. of 
type i) 

Sub-type I b. 
feature 3a, c, or d 

djidis tooth 
L.'pdtil bag 
djicin' foot 
lAq.'AS mountain- 
goat blanket 
L.'Aqlacin^' mocca- 

tfAkom"^ beaver 
tfo'mV paddle 
waxdHs.'i pipe 
td^qlwa' devil-fish 

djig'^in' song 

Sub-type I c. 
feature 5a: — 

I6k6°min bailer 
ti'hd''ddn' chief's 

Plural of simplex modified by diminutive 




II a. 


Ill b. 

Plural of simplex modified by diminutive 




lUl! Ap^ l! Apdtil 



XIII b. 




(-tn' misheard for 

-tn" ?) 

XIX b. 

tat! Ak'H! Akwim' 

XX b. 


XXI b. 




(-'q' misheard for 

-q! ?) 

XXX b. 



XXI c. 

XXVI a. 


Sub-type I d. Plural of simplex modified by diminutive 
feature 2b: — 

W'^bonC small clam ii b. 

kloyokobVn fisher- 
man VIII. 

XXV a. lilimld"°bo^7rC 

dit. k!wtk!wlk!oyoko- 

Eight: Wakashan and Salishan Languages 



Sub-type I e. Plural of simplex modified by diminutive 
feature 4b: — 

tld'^qlaV mountain plur. ii a. dim. xxviii. 

dim. plur. t!et!Aq!t!d''q!eH' 

Sub-type I f. Plural of simplex modified by diminutive 
features 3b, 2a, 3c: — 

kumaqirC sea-lion i. x b. kwikurnkwd"' mdqirC 

A couple of aberrant diminutive plurals with ce- are given 
under type ii f. 

Type II. Reduplicating Syllable: ce; followed by plural of 


The reduplicating syllable is analogous to that of diminutive 
types I, II, III, IV, V, VI, VII, VIII, IX. Sub-types are 
to be recognized here as in type i. 

Sub-type II a. 

Plural of simplex 

unchanged : — 

qwAdi^s hump- 

backed whale 


I a. 

qweqwAd ' qwAdVs 

qo^'a'^ hemlock 




'dwdk''' tobacco 




xApd'^ red cedar 




q!dp!xwai oak 


XIII a. 


q'td'abas wooden ball 

used in game 

II a. 

I a. 


qe'n'qerC duck 


IX b. 

qeqAd'qen" (based 
on unreduplicated 

qd'^qa" rush mat 




toVx^lal necklace 



tftAxHdVxHal (re- 
duplicating sylla- 
ble for plurality 
based on unredu- 
plicated form of 


VI American Indian Languages 2 


qfa^L! land-otter plur. 

II a. dim. xvii a. dim 

. plur. qleqlALfqld^d 

qld'sa" sea-otter 



qfeqlAsqld^s (with 
loss of -a\ as in 
dim. sing.) 

xd^'wa fur seal 


XXV b. 


sd^'ba' mussel 




xwdsAbai soapberry 



XXVI a. 


p.'dqfAddtc goose 




L'.d'aVo'm" wolf 



lIcl! AIl! d' aV 6' rrC 

ts'.dtcHlhai spruce 


VIII a. 


k^dck^dc blue jay 


X a. 

k^dky'ickydc (based 
on unreduplicated 
-form of simplex) 

sdpdxos horn 


XIX a. 


t&'ag^ax'' fern 

Ill a. 



htg^os chief 

Ill b. 

I c. 


xdug^as grizzly bear 


XX a. 


heq'sd''min' pole for 

poling canoe 


XIX a. 

hehdq' heq' sd^min'^ 

qtod'dda ear 


I a. 


sidjdqo'p' basket 



I a. 


tdyac killer whale 




(plur. of type 


implied in 



Sub-type II b. Plural of simplex modified by diminutive 

feature 3c or d: — 

qAl'q! warrior 




Ug.'" bow 




q!wAt ■ iHcin' hump- 

back salmon 


X c. 


xAucin' bone 




mdHdin' louse 

II a. 

VIII c. 


<isx" hair seal 




td'agHn salmon spear 

Ill a. 



Eight: Wakashan and Salishan Languages 



Sub-type II c. Plural of simplex modified by diminutive 
features 3a or c, and 2b: — 

xAsAm box 

plur. I. 



dim. plur. 

qwAsAm woolly 





{-qus- probably 
merely variant 
of -qwAS-) 

qfAs'Addi buckskin 





LV'Am' cockle 

dit. (or VIII.) 

XXIX a. 

LiLVLdi'im" (with 

irregular lengthen- 

ing of -V- = -Al- to 


htw^qerC swan 

II a. 



mdV.di horse clam 




qd'um' eye 



qeqoqd'om"^ (-qo- 
heard for -qau-, or 
perhaps for -qAu- 
reduced hom-qau- 
— see type in) 

Sub-type II d. Plural of simplex modified by diminutive 
feature 4a or b : — 

pdxai' creek i. 

q!dk'" board dit. 

SAg'ifc'" war-club dit. 





Sub-type II e. Reduplicating syllable of plural of simplex 
changed to cau-: — 

xd'a big clam 

II. xexAuxd'A (note 
change of xa'd- to 
-xd'a, perhaps due 
to rhythmic anal- 
ogy of dim. sing. 

qfdik'" eagle 


-ol q!eqfAuq!dik-dl 


VI American Indian Languages 2 


These strange diminutive plurals can hardly be explained 
otherwise than as formed by analogy of such diminutive plurals 
as xexAUxd'wa "little fur seals," xexauxdug^as "little bears," 
and xexAUXAUcin' "little bones," where -xAU-{-xau-) is etymo- 
logically justified. The parallelism of xd'd "big clam" and 
xd^'wa' "fur seal" seems particularly plausible. 

Sub-type II /. Plural of simplex modified by diminutive 
feature 4c (for convenience of comparison one form with ce is 
included) : — 

V.t'ihdi wild cherry plur. i (or viii), dim. xiii a. dim. plur. 


qe'w^x steel-head 

ft*x" yellow cedar 



XVII b. 


(really belongs to 
type i; based on 
reduplicated plu- 
ral of type ii) 

{-g^e^x as in dim. 

titold'ayix'' {tVx* 
> Hiyix'', U- being 
modified to ta'a- ; 
-to-, cf . type II e, 
is peculiar and is 
probably due to 
analogy of titotd'- 
ag^ax'' "little 

Another diminutive plural with erratic -o- vowel (in both 
reduplicating syllable for plurality and stem) belonging to 
type I, is: — 

safari' cohoe sal- 

XIII a. sisoso'dd-ol 

The material at hand does not permit to see what analogies 
have operated here. 

Eight: Wakashan and Salishan Languages 421 


Type III. Reduplicating Syllable: ee; reduplicating vowel of 
plural of simplex shortened. 

A new feature is here introduced, the shortening of the long 
reduplicating vowel characteristic of the plural. Sub-types 
are here also to be recognized. 

Sub-type III a. Plural of simplex not otherwise modified : — 
xdp'xop' hum- plur. i. dim. -olV" dim. plur. 

ming bird xwexop'xo^p'^ (bas- 

ed on unredupli- 
cated simplex) 
^"^"6^'" herring iii b. xxb. lilola^g^et!" 

Id^dak'"* skin iii b. xxiii. lei old" dak"" 

qd'ya* water viii. viii a. qeqeqd'ya' 

xd'a.idatc stump i. viii b. xexexd' aidatc 

(type VIII im- 
plied in dim. 
sd'yal lake viii . viii c. sjsjsd'yal 

IdidatctAn woman's dit. xx b . lilildidatctAn 

cedar-bark skirt 

Sub-type III b. Plural of simplex modified by diminutive 
feature 2a: — 
tcdyac hand viii. ix a. tcitcUcd'yac 

Sub-type III c. Plural of simplex modified by diminutive 
features 2a, and 3b or f : — 

sayd^ada neck viii. viiib. sisfsd'ya'ada 

kwudjak'"^ trout dit. xxix b. kwikwikwd'^djak''' 

Sub-type III d. Plural of simplex modified by diminutive 
features 4a and 2a: — 

xd'adjatc stone viii. ixb. xexexd' adje'ic 


VI American Indian Languages 2 


Type IV. Reduplicating Syllable: ce; followed by simplex. 

It seems that a reduplicating syllable with e tends to be 
considered the morphological equivalent of double reduplication 
(see plural type X, diminutive type XVI), in this case of combined 
diminutive and plural reduplication. Various sub-types are to 
be recognized, according to whether the reduplicating syllable 
is followed by the unmodified (or modified) simplex, the modified 
form characteristic of the diminutive, or by a form still further 

Sub-type IV a. 

Simplex unc 

changed : — 

pWixdi alder 

plur. I. 

dim. XIII a. dim. plur. 

(or viii). 


L.'Ams house 


XXXII b. hle'^LlAms 

mo' OS hand 

no plur. 

I a. me' mo' OS (may 
also be considered 
as belonging to 
type IV b) 

Sub-type IV b. Simplex modified by diminutive feature 5a:- 
ic.'e'ddo dog ii b . xxvb . tcle'tdin'am" (ir- 

regular in that -o 
of stem is dropp- 
ed; with -dm' cf. 
perhaps -d'm of 
djddjid'm 'trees') 

Sub-type IV c. Reduplicating vowel of diminutive changed 
to e: — 

xwdxwadjo'm' fly 

(dim. in form) 
qwi^qwdH.'Ald'k* butterfly 

(dim. in form) 
mAqsirC nose i. 

kwd'am coiled 

storage basket dit. 

q!wdt' Am river dit. 

ttx'^sal tongue ii b . 



I b. me'mAqsi^n 

IV. kw^'^kwi'im' 
xviii . qlwe'^q'.wat ' tm' 
XXVI a. te'Hx'^sal 

Eight: Wakashan and Salishan Languages 



Sub-type IV d. Reduplicating vowel of diminutive changed 
to I; stem further modified by diminutive features 5a and 3c: — 

q'.wAHx wood plur. i. dim. xxxa. dim. plur. 


Sub-type IV e. Reduplicating vowel of diminutive changed 
to e; stem further modified by diminutive feature 5b: — 

sdtslAm tyee sal- 

II a. 

XXII a. 


Type V. Reduplicating Syllable: ce; followed by plural of 
simplex modified by diminutive features 5a and 3a: — 

tlc'^'de'qwai salmon- plur. ii a. dim. xxvi c. dim. plur. 
berry bush t'.e't'.AnV.An'qwai 

{-€'- is lost, cf. 
diminutive feature 
5 b) 

sdstn^ mouth lib. xxvi a. se'sossin' 



1. pd'a 

2. sd'a 

3. tcdlas 

4. mos 

5. siydtcis 

6. tfdxam (oT-ab) 

7. ts'.o'HcVs 




IL op'an haik'" pd'" 

20. simcyd'a 200. 

30. tcanaux'^cyd^a 300. 

40. mosalcyd'a 400. 

50. seyatsfalcyd'a 500. 

60. V.dxamalcyd'a 600. 

70. ts.'otci'alcyd'a 700. 

80. td'atdsalcyd'a 800. 

90. tigHxwalcyd'a 900. 

100. Vsd'ntc 1000. 

6p' an 
2000 is sdba Vsd'agHtc or sd'a Vsd'agHtc. 








tig^ixwd' ag^itc 



VI American Indian Languages 2 


Numerals with classifying suffixes, referring to class of objects 

counted, are: 






1. pipd'a 





2. sisd'a 





3. tcdldy 





4. mOsdvi 





5. s^yatsdyi 





6. tfdzamdyi 


7. ts!6tcisdvi 

8. ta'dtcisdyi 

9. llgt/'iTWdyi 

10. dpdndyi 

The series for "dollars" refers, properly speaking, to round 
objects, including such objects as heads and turnips. 

Body-part suffixes. Examples of body-part "substanti- 
vals," as they have been termed by Boas, which occur only in 
composition (better perhaps derivation), are: — 

head: ■pdq-e'^q'warC white-headed 

idx • e'q'warC red-headed 
(or -ad') 
hand: pdq'd'''dja' white-handed 

tcixo'"dja' red-handed 
eye: pdq-dos white-eyed 

pdq'paq'dos white-eyed (plur.; refers to several persons 
or to two eyes of one person) 

tcixdos red-eyed 

tdxtcixdos red-eyed (plur.) 
nose: ts!dts!e'miq'"^ red-nosed 

paqe'^q'"" white-nosed 

t!dt'ts!d'''miq'" nose bleeds 
foot: pdq'dn' white-footed 

pdq'paq'cin' white-footed (plur.) 

With these contrast independent use of "ear" in pdq^paq* 
qloa'dda "white ears." 

Eight: Wakashan and Salishan Languages 425 


Possessive and subjective pronouns. Only very frag- 
mentary data were secured on Comox pronouns. I do not 
consider them as particularly reliable. 

tAtsi md'os my head tAmsi md'gs our heads 

tAn mo^os your head Ia mo^osap' your 

(plur.) heads (vis- 
tA mo'gss his head (visible) ku mo'osap' your 

(plur.) heads (in- 

ku mo'oss his head (invisible) 

tA and ku are articles implying visibility and invisibility respect- 
ively. Possessive pronouns modifying verb subjects are: — 

'd tsi mo' OS my head is sore ('a' to be sore) 
'd' tAn mo' OS your head is sore 
'd' tA mo'oss his head is sore 

'd tA mo'oss tA 6-d?r" the woman has headache (literally, sore 
the her-head the woman) 

Possessive pronouns modifying verb objects are: — 

tc'kfudA wad tsi mo'os I see my head 
tc'kludAxwad das mo'os I see your head 
tc'kludAxwad dA mo'oss I see his head 
tc'kludAxwad das tciHcdyac I see your hands 
tc^k/udAxwad dA tdHcdyacs I see his hands 

Subjective pronominal suffixes are: — 

titc to'mic I am a big man (tl big) 
ttHc'"" to'mic you are a big man 
tV" to'mic he is a big man 
tt'^djan sdlV" I am a big woman 
tV^'djaux" sdW'' you are a big woman 


VI American Indian Languages 2 



This is not the place to enter into anything like a systematic 
comparative treatment of Salish reduplication, the more so as the 
phonetics of most of the material available for comparison are 
not such as to allow one to make definitive classifications of plural 
and diminutive types (this remark applies particularly to vocalic 
quantity and glottal stops, both of which, as we have seen, are 
important for our present purpose). Certain facts of a compara- 
tive nature, however, come out quite clearly and may be briefly 
noted here. 

Plural reduplication. It is evident that all Salish lang- 
uages make use, like Comox, of different types of plural reduplica- 
tion. Both types I and II are plentifully illustrated and are 
without doubt the fundamental Salish processes. Examples of 
type I are: — 

Bella Coola 

s-t7i tree 

plur. s-tntn^ {s-, as 
often in Salish, is 


s-kwomdi dog 


(Cowichan group) 


sk'dqa dog 


noqonuq woman 



s-k' EltEmtQ man 

s-k' Elk' EltEmtQ* 

Thompson River 




s-nukoa friend 


s-k6um crumpled 


Examples of type 

II are: — 


s-pdl raven 


s-tdlo river 


' F. Boas, The Salish Languayes of British Columbia, Report of British Association for 
the Advancement of Science, 1890, 6th Report on the Northwestern Tribes of Canada, p. 

' C. Hill-Tout, Report of British Association for the Advancement of Science, 1902, 
Report on the Ethnological Survey of Canada, p. 20. 

• F. Boas, ibid., p. 131. k' is here and in other forms equivalent to our q; q to our i; 
Q to our f, tl to our I (and l); IV to our l!. 

« Ibxd., p. 135. 

' F. Boas, Report of British Association for the Advancement of Science, 1898, 12th 
and Final Report on tlie Northwestern Tribes of Canada, p. 28. 

• F. Boss. Report B.A.A.S., 6th Report on N.W. Tribes, p. 129. 

Eight: Wakashan and Salishan Languages 



Thompson River 


s-kdpk'Eii head plur. s-k' Epkdpqsn^ (prob- 
ably misprint for 
-kspkapk' En) 

k'est bad ky'Eskest^ (probably 

misprint for -k^est) 

s-k'elg Indian 
cdsuQ stone 
s-pam camp fire 
s-nikidp coj'ote 


s-niknikidp^ (-t- is 
very open and short, 
-I- is close and equi- 
valent to our -t-; 
hence type iib) 
s-Quasit to walk s-QUsquasit^ {type II c) 

An interesting Thompson River example of type II is: — 

cirdp tree cipcirdp^ 

An example of type III (reduphcating -aw- contracted to -p- 
or -0-), but with retained -w- (Comox -g^-) is: — 

Okanagan iEiuwU boy totuit* (based on unre- 

duplicated form of 
simplex ; final vowel 
of stem apparently 

This example follows type III b. 
usage in the treatment of the same 
languages, compare with this: — 

Lower Lillooet tu'u''wut' boy 

As illustrating diversity of 
stem in different Salish 

tutu'u'^wuV^ (-U- is short 
and close) 

This follows type III a, besides which the stem itself seems to 
differ markedly in regard to vocalic quantity and rhythm from 
the cognate Okanagan stem. Shuswap agrees better with 
Okanagan: — 

tuwtui boy tutuwtut^ 

i Ibid., p. 131. 

* Ibid., p. 135. 

' F. Boas, Report B.A.A.S.. 12th Report on N.W. Tribes, p. 28. 

* F. Boas, Report B.A.A.S., 6th Report on N.W. Tribes, p. 135. 

' Some Lower Lillooet linguistic material was obtained iu January, 1912, from I nace 
Jacob (Indian name Yisp). 
•F. Boas. ibid., p. 131. 


VI American Indian Languages 2 


It would seem that type VII, which is only sporadically repre 
sented in Comox, is more typically developed in Interior Salish 
Examples are: — 

tsitQ house 
gitia old woman 
tcltQ house 
s-tsuk' picture 
s-k'dk'qa dog 
s-pEzuzo bird 

Thompson River 

s-kikElUQoa musk- 

plur. tsUsttQ^ 





s-pEpEzuzo^ (this form, 
however, may really 
be diminutive plural, 
s-pEzuzo being dimin- 
utive, with final re- 
duplication, of s-pEz6 
"animal," whose plu- 
ral is normally form- 
ed : s-pEzpEzo,^ type I) 



Lower Lillooet 

tctC'^x house 


gd'? water 


Note also : — 


k'unes whale 

k'okuinis* (probably mis- 

(i.e. qunes) 

print for -k'uinis) 

It is interesting to contrast with this plural (qogwinis in our 
orthography) Comox qwAd'qwAdi^s humpbacked whales ( <qwAn- 
qwAnis) of type I. Here again we see the tendency for different 
Salish languages to form the plural of the same stem according 
to different types. 

Type IX also is illustrated outside of Comox. Examples are : — 



lalEm house 
wuqas frog 

mela son 


hdiiweqas* (-u- presum- 
ably glide; haw- dis- 
similated from *wdw-?) 


' ibid., p. 131. 

« F. Boas, Report B.A.A.S.. 12 Report on N.W. Tribes, p. 28. 

' Obtained from Ignace Jacob. 

< F. Boas, Report B.A.A.S., 6th Report on N.W. Tribes, p. 129. 

»C. Hill-Tout. Report B.A.A.S.. 1902, Ethnological Survey of Canada, p. 20. 

Eight: Wakashan and Salishan Languages 429 


Type X is illustrated in : — 

Nanaimo k'dk'En post plur. k'&lak'En} (d is 

apparentl}' our e) 

The last example, with its inserted -la-, shows also another 
method of plural formation, one not found, at least as far as 
can be judged from available material, in Comox. Other 
examples of this inserted -lia)- are: — 

Nanaimo hd'pet deer hala'pst^ (type IX) 

tcitdek'an mink tciletciek'an^ (type VII) 

spdk'Em flower spdlak'Em} 

Tcil'Qeuk k''dmi maid k''dlami^ 

stEktyu horse stElEktyu^ 

ydsuk hat ydlsuk^ 

There seem to be still other types of plural formation in 
Salish that are not represented in the Comox material given in 
this paper. One of these is to prefix -a- (Boas and Hill-Tout 
write -E-), which may be palatalized to -j-, to the stem. Examples 
of this type are : — 

Nanaimo s-mtysg deer s-Emtyeg^ 

Tcil'Qeuk s-wtEka man s-iwtska^ {-a- palata- 

lized to -i-, -T- by S-?) 

This type is perhaps a reduced form of another one that occurs 
with some frequency, reduplication with ca-. Examples are: — 

Tcil'Qeuk IdlEm house lEldlEm^ 

s-mdlt stone s-mEvidW 

Shuswap la good IeW 

Nanaimo laldlEm "houses," as compared with Tcil'Qeuk 
IsldlEm, suggests, in turn, that CA-reduplication is reduced from 
ca-reduplication (type IX). Tcil'Qeuk yESidm "chiefs"^ from 
sldm may be dissimilated from *SEsld7n (or does y- reduplicate 
-I- of stem?). Vocalic changes {e to o and d) are illustrated in: — 

Tcil'Qeuk s-weEkdtl boy woEkdil^ 

' F. Boas, Report B.A.A.S., 6th Report on N.W. Tribes, p. 129. 

* ibid., p 128. 

» C. Hill-Tout, Report B.A.A.S., 1902, Ethnolocical Survey of Canada, p. 20. 
< F Boas, Report B.A.A.S , 6th Report on N.W. Tribe-, p. 128. 

* Ibid., p. 131. 

430 VI American Indian Languages 2 


s-wtwilus youth s-wAvrilus^ (this may 

be considered, how- 
ever, as formed from 
unreduplicated sim- 
plex according to 
Type IX) 

With the latter example compare Coraox we"wdlos "young 
men" from we^wdlos. 

To sum up, it is clear that there are a number of wide-spread 
Salish methods of forming the plural, which may, however, at 
last analysis turn out to be capable of reduction to Type I (of 
which Type II may be a reduced form). It is conceivable that 
sub-types, which have developed in particular cases from this 
by secondary phonetic processes (cf., e.g., Comox Types III and 
VIII), set the pace for new purely analogical, not etymologically 
justifiable, forms, so that now any one Salish language exhibits 
great irregularity. Certain of these secondary types seem to be 
favoured in one language, others in another, so that, as we have 
seen, the same stem is sometimes differently treated in different 
languages. To unravel the history of reduplicated (and other) 
plurals in Salish, however, requires a far more abundant body of 
material, for purposes of comparison, than has as yet been made 

Diminutive reduplication. The last remark applies even 
more forcibly to the study of Salish diminutive formations, for 
here there is a still greater variety of types represented. Avail- 
able comparative data are quite scanty, so that only a few 
points can here be referred to. The most consistently carried 
out difference between plural and diminutive reduplication in 
Salish is that in the former the first two consonants of the stem 
(though not infrequently only the first) are reduplicated, while 
in the latter only the first is reduplicated, never also the second. 
At the same time there is a marked tendency, as in so many 
Comox examples, for vocalic reduction of the stem. Redupli- 
cation with e- vowel seems also characteristic of many forms; 
also breaking of stem vowel and umlaut of a to e or e seem 
to be found. 

Some of the types represented, outside of Comox, are: — 

' C. Hill-Tout, Report B.A.A.S., 1902, Ethnological Survey of Canada, p. 20 

Eight: Wakashan and Salishan Languages 


Type X. 



IdlEm house 

diminutive lUEm^ 
(based on unredupli- 
cated simplex) 
HtnotEm little girl* 
(H = our x") 

Tyye XII . 


wuqas frog 


Type XIX a. 


k'dk'En post 


Type XXI a. 


pasitlkua lake 


Thompson River 

s-nukoa friend 


Type XXIII. 


s-tdlo river 



s-td'lo river 


Type XXVI a. 

Nanaimo s-pak'sm flower s-pdpk'sm^ 

Comparable perhaps to Comox Type XXX a is: — 
Thompson River s-pee'tc s-pdpaats' {-aa-^-d'a- 

black bear (e = our e) ?) 

Other diminutive types than those listed for Comox un- 
doubtedly exist in Salish. Among these is reduplication with 
CA- (cf. plural types above), as examples of which may be given: — 

Thompson River 

c-mtits deer 
IdlEm house 


Isldm? (based on unre- 
duplicated form of 
simplex; change of 
-E- to -a- is perhaps 
parallel to that of 
Comox -A- to -%-) 

' F. Boas, Report B.A.A.S.. 6th Report on N.W. Tribes, p. 129. 

' C. Hill-Tout, Report on the Ethnology of the Okandkin of British Columbia, Journal of 
th« Royal Anthropological Institute of Great Britain and Ireland, vol. xli, 1911, p. 143. 
' F. Boas, ibid. 

* Boas, ibid., p. 131. 

• C. Hill-Tout, Report B.A.A.S., 1902 Ethnological Survey of Canada, p. 20. 

• Boas, ibid., p. 129. 

' Boas, Report B.A.A.S., 12th Report on N.W. Tribes, p. 29. 
' Boas, ibid. 

* Hill-Tout, ibid. 

432 VI American Indian Languages 2 


Similar apparently to Comox type VII (but without diminutive 
feature 2a), except for its incomplete reduplication (loss of 
reduplicating consonant after s-, cf. plural types above), is: — 

Tcil'Qeuk s-mdlt stone diminutive s-eniElet^ 

Reduplicating with cv-, and with breaking of stem-vowel, is: — 

Thompson River kss bad kEkEES-f^ (? = -kA'As) 

This tj^pe may well exist in Comox, but not happen to be 
represented in the material collected. Such diminutive forms as 
Thompson River qEzuzum^ , with interior reduplication, from 
qzum "large," and Thompson River spEyiizu^, with change 
of -z- to -y-, from spEzuzu "bird," are evidently representatives 
of very specialized tj^pes. Neither of these, so far as known, 
has a Comox counterpart. 

Judging from the analogy of Comox and from a few Interior 
Salish forms obtained by the writer, it seems very likely that 
glottal stops are frequently employed in Salish as diminutivizing 
elements, though this is not apparent from most of the material 
that has been published. Examples are: — 

Upper Lillooet' s-mulcitc woman sE-m'Em'letc girl 

Thompson River* c-muldtc woman c-mu'ni'ldtc 

(type XXVIa) 
Comparative data on diminutive plurals are too scanty to 
enable us to gather much of interest. Some Interior Salish 
forms obtained by the writer seem to indicate quite clearly 
that in those languages the diminutive plural is not, as in Comox, 
a diminutivized plural, but a pluralized diminutive; in other 
words, of the tv/o reduplicating sj'Uables, the first contains the 
first two consonants of the stem (plural type), the second syllable 
the first consonant only (diminutive type). Examples are: — 
Upper Lillooet plur. s-muhnu'fdtc dim. sE-tnErn'letc girl 


dim. plural 

Thompson River c-muhnuldtc dim. c-mu'm'ldtc 

women dim. pi. c-mElmu'm'ldtc 

' Hill-Tout, ibid. 
' Boas, ibid. 

' L'ppcr Lillooet forms were obtained in January, 1912, from Chief Jim (Indian name 
Aide^i'q.'t). F has here been used to indicate very short obscure vowel of undeiined quality. 
* .Some Thompson River forms were obtained in January, 1912, from Chief Tetlenitsa. 

Eight: Wakashan and Salishan Languages 433 


This difference of treatment again indicates that in many 
respects each dialectic division of SaHsh has gone its own way 
in the use of morphologic features common to Salish generally. 

Editorial Note 

Originally published as Memoir 63, Anthropological Series 6, Geological 
Survey, Department of Mines, Canada. Ottawa (1915). 

Pluralizing and diminutive reduplication is nearly universal in Salishan 
(Kuipers 1978: 612). Haeberlin's survey of Salishan reduplication (1918), com- 
piled shortly after the appearance of Sapir's paper and undoubtedly stimulated 
by it, summarizes the Comox data (169-170) and cites similar patterns from a 
number of other languages, but does not go much beyond Sapir's own com- 
parative remarks. Paul Kroeber (1988) has recently published a study of incep- 
tive reduplication in Comox. The reduplicative morphology of several other 
Salishan languages has been described in detail; see in particular Hess's study of 
"chameleons" in the Snohomish dialect of Lushootseed (1966). 


(Nootka Text witli Translation and Grammatical Analysis). 

The following text was dictated to me in 
November, 191 5, by Tom (Sa'ya't'capis 
" Stands up high on the beach "), one of the 
oldest and best informed men of the T'sica'- 
'ath tribe of Nootka. The grammatical analysis 
should give a serviceable idea of Nootka 
structure, pending the appearance of a full 
grammar of the language. The phonetic system 
used in this paper is explained in " Phonetic 
Transcription of Indian Languages, Report of 
Committee of American Anthropological Asso- 
ciation " (Smithsonian Miscellaneous Collec- 
tions, vol. 66, no. 6, 19 16, particularly pp. 7- 
15) ; my u, however, is always open, as in 
English full, and varies freely with close 0. 
The tale is Nitinat (Nootka dialects south of 
Cape Beale, including Makah of Cape Flattery, 
Washington), but its linguistic form is Nootka 
proper (T'sica'^ath dialect), except for the 
names of the rivals, which are unmodified 


'o''si.m't'cat} ' ma""ak' - 

Now trained secretly for sue- California whale 
cess in so and so 

''.•'h-'to"p' ' k'waTs'.ts. •♦ wz"'.tcu"ati > 

humpbacked K'walisits. Now went to sleep, 

hawt"atl '' 'o''stm'tc. ' t'si't'k'pi'atl *^ 

now train secretly Now lay down in 

finished for success in the house on 

so and so. (l^is) back 

'uvk"t'.9 w-.'k'at}uk''° t'si't'k'" 

the fire. Now of (him) lie down in the house 
was not on (her) back 

lo''ts"ma''^ t'l'qwd'atL'' 'na''ts'a"ti''* 

woman, now sat in the Now was look- 

house ing at 

t'aci'''ak"i'5 'a'the'".'^ ya-'i" 

the door of (them) be night. There 

ka'tlh^'cttt'* t'aci-"i'"^ t'^'h't'sai'.'" 

appeared the door head . 

qwa^' qw:""tt'q'''^ k'a'yu"min'.^' 

Was in quality as is in panther, 


yai hi'Lat'^4 si't'a'^^ tb'hn's'.tat"-.''*^ 
There of (him) tail the head of (him), 
was at 


Was not 

qwa* ^' 
was in quality 

t'hi"ya-x.'" ya"l h-.stsa'q't'so'^' 

move quickly. There was provided at 

each end with 

tb'h't'sit'.'. su'kwitL'' tca'kopokw"t.'' 

head Took hold of the husband of 


lo-'ts-nu''.'.>^ tci''tC'.ti'> tlup'k'sa"'p'at}."'*' 
the woman. Pulled, now caused to 

wake up. 

tiu'p'k'citL'7 k'waTsas. yai 

Woke up K'walisits. There was 

t'o'tbh'tsaq't'so''" su'kw<.tl tS'/kuinn'- 

Head-at-each-end. Took hold of the iron 


VI American Indian Languages 2 

'ak''-.'''' k'wuTsits t'ci'tcah" 'a'h''^' 

of(Him) K'walisits, cut this, 

hi's'a'p'^^ 'ah*ko-^> 'a"ap'tsa'atcat"'.'.+'* 
caused to bleed this the thigh of (him). 

tuxtspa''-<5 k'waTst.ts. t'c'.'tc.tiiia'*^ 

Jumped over K'waHsits. Cut also 

kw'.sa"'s'at''». ■»^ tli'cth.n'.*^ hz's'c-.ti. *'^ 

the other side of (him) foot. Bled. 

kwr'spanoici.'atl^" lu's*i,m''yawi'atl> ' 

Now began successively now became blood- 
(to jump) from side to side, covered 

t'o't'-">h*tsaq't'so"'<.'.5' qah^c/'Eti^' t'o't':>h-"- 
the Head-at-each-end. Now died the Head- 

tsaq't'so"'-,'. ?'u'y'.tci"lc'.'ztl.>^ qw-.'sh^'atl^' 
at-each-end. Now began to Now was acting 
mal>e medicine. thus that 

'a'yMii'k'c'.'atl^'' ma-"ak'. 

now began to obtain many California whale, 
in hunting 

'napxta'"atlqo"W£''.n'5' t'saxci";tiqo' .>** 
Now would die imme- whenever now speared, 
diately, it is said, 

'ah'''a-"atKv£'uVi'' tta'o-"atltta'''" 

Now was thus, it is said, now another was also 

t'so"ach-'citl*" W£"it'cah's'atl'^' 

winter take place. Now was sleeping in (his) 


k'wa'l'S'.ts h'.'l 'anah*'i.s^' t'ca'pats*"* 

K'walisits, was at little canoe 

'athe''.'*' 'oyo"aratl^> kwi'stsatcith'ca'^^ 

be night. Now perceived the, as they say, " go 
so and so off elsewhither " 

'yh-^'atF'' tca"ats--.-b'^~* 'ukta'^? 
now the one was Cha'atssib' was named so 

and so 

.'£'m'"k'"" tla"o" 'o"o"tah\'' 

have as name, was another hunt such and such 


t'i'.'"wi.n''ap'at'''^ kwaTsas 'ani"> 

Was caused to be laughable K'walisits that 

qa"'ya*panatcqa'^ Wz"itcqa' wi'k'arattqa"* 
being drifting being sleeping, being now not 
aimlessly aware of 

kw'.'stsatcittqa''7 'na^'s."** hac'i"'tC'.ti''' 
being " going off else- day. Heard about 

whither " 

k'wal'S'.ts 'an-.'"' t"h.""win''ap'at'qa'.''^" 

K'walisits that being caused to be laughable. 

ya'k'ci"zt''*' h/'maqst'.'*'^ wi'k' 

Of (him) became sore heart, was not 

citlsti^'s**' hi."'sasa^^ tlo""ow's.*^ 

move inland, stayed right on Tlo'owis. 

the beach 

citistr's'atl'"*^' ni-ti-'na'ath-*^' 'utsatc' 'ii\^^ 
Now moved Nitinat now wentoff to so 

inland people, and so 

t'sa"akokw'i'.''' w-.-'napatl''" k'wa'l's-.ts 

the stream of Now remained K'walisits 


tto'"ow's hd'atl. W'.k'r't'^' 

Tlo'owis now was at. Became non-existent 

qo-"as^^ 'o-Sf.m'tc-i"atl9' 

person. Now began to train secretly for 

success in so and so 

kNva'l's'.ts ma'"ak'. 

K'walisits California whale. 

hr'ms'o"- j 
Beaan to 

'uk'c.ti'-*. wa-'lak'''i 

move up and down (like a Was bound for 
blowing whale). 

say£""i''^^ mo''tci'l''' lv.''niso"'uk' ''*' 

the far- was for four move up and down 
distant ; days whale-fashion 

mo- 9'* 'na-'s mo" 'a'thai' ""'. haia's-af'°' 
four daylight four be night. Arrived nt 

kaxi-'k'.s '"^ 'ukt£-"e '"' nis-'ma'"°^ 

Kahikis the so-and-so-namcd land. 

Eight: Wakashan and Salishan Languages 



t'-.'q'satt '°5 k'wa'l'stts. yai 

Now sat down on the beach K'walisits. There 

hi'tahnis '°^ t'ca-'pok'" '"7 t'sa"ak'i'. '°** 
came down-stream canoeman the river. 

tbwi-tci"£t' '°' h'.-'s'it'q' "° 

Now was approached where (he) was on the 


k'wa'l'3its na\va*'yis"' 

K'waUsits be seated on the beach looking 


hmii'ha "^ h'.'iyin''-.' '" t'ca"'pokw'r.'. 
Went out of the the one at the canoeman. 
canoe the bow 

si'ma'tsy.n'"^ mi'h'yi.' "5. 

(They) had sticking up in the bow spear. 

t't'q'stC'.sa'at' "^ k'waTsits 'o'h^'at' "" 
Was sat alongside of K'walisits was soed by 
on beach 

t'ca^'pokw''.'. qatcct"zt' "*'. 'o"'quni'him'a "^ 
the canoeman. Now was " It is fine 

nudged with weather, 
(his) elbow. 

takha'^'^ 'ah' 'na-'sT '^' wa-"at' '". 

is it not ? this the day," was said to (him). 

'moqwt'"yu*tl 'V k'wa'l'sas h£'s*iH*tcr,'a'h* '^+ 

Became speechless K'walisits, became unable 

in (his) throat in any way 

tS','q'c'.tl'^> qalrkwatC'."at' '^^ qa"yap't'at"t'. '^" 
speak, of (him) the limbs of 

completely died (him). 

tb'k-icitl "** qo'"as''.' ya-'tscai "^ 

Stood up the person, walked off 

ho'a'tsatcai '5" t'ca'patsukw't' '•'. haa'qsal "^ 
go off back the canoe of (him). Went into 

(his) canoe 

t'ca'patsukw'-.'. th'h'c.t} ''' ys-f-'" 

the canoe of (him). Paddled off yonder, 

ta'kh'tatcat ''^ '.^'h^'atl ^^ 

became far out at sea. Now was the one 

t'c'.-n'.-"ath't'ca' ''^ 'oya-'tl ti-'tcatc.'sii '^^ 
the, as they say. Now was at now come 
Ch'ini-person. such and such a to life 

th'h'C'.'str-.t'q' ■5'^ t*ca-'pokw''.' t'c',*n'.-"ath^'e' '■'°. 
when now the canoeman the Ch'ini- 
paddled off person. 

su'kwal'^ ''.•'nax'mak''r.'.'4' tr.mi"'s'ap' ^'>^. 
Took hold of the regalia of rubbed about 
(him), on the beach. 

hr's '■♦5 'ah-*'.-!" '■»■> mo"'tci"y'.s '■*>. 

Was there on the that be for four 

beach days on the beach. 

h'.'tats--">hnatltla ''^ qwa'yst's'.'k' ■•♦" 

Now came out of the woods also wolf. 

tiaw'.-tC'.";t'tla '*** k'wa'l's-.ts. tsusk'c'."£tuk'"'t^ 
Was approached K'walisits. Of (him) was 
again urinated on 

mu'tsmDhaq' '5° k'wa'l'sas. qwa""ak' '5' 

bearskin K'walisits. Of(him)wasin 


qwz""tt'q' '>^ kats"o"'m'.n' ''5. 'ah'''a""atl 
as is in quality hail. Now was thus 

haw'.""ati ho"atsatC'.tl k'waTs-.ts 

was now finished, went offback K'walisits 

walc'."£tl''4 maht'f"ak"'.' ■>> ha'wi-t}'>^ 

Now returned the house of Was finished 
home (him). 

t'so"'.tcly\ t'cukwe"ati ■'" ma""ak'. 

winter." Now began to run California whale. 

t'sa'xcal tca"ats"',"b'. qa'h-'ba'p' '>^ ma""ak"i 
Speared Cha'atssib'. Caused to die the Califor- 
nia whale 

ya'.a"'n'.t "'.t'q' "''' t'i'.'"wtn''ap'at' k'waTs'.ts. 
the one by whom caused to be K'walisits. 
had been — ed laughable 

t'sa'.xc'.tl 'yo"qwa" ""' k'wa'l's-.ts. t'sa'xcaitia '^' 
Speared likewise K'walisits. Speared again 


VI American Indian Languages 2 


tca"ats-'.-b''p'tb' '^'. t's.i'xca^ 'yo-qwa' 
Cha'atssib', again obtained. Speared likewise 

k'wal's'.ts h'.n-'.-'p' yo-'qvva-' mo-'y-.p' '^' 
K'walisits, obtained likewise. Obtained four 

tca"ats''.'b' ma""ak' 'an'a'^+ 

Cha'atssib' California whale, was only 

kwa'l'S'.ts c-.iVV.-r^'ifii 


sutC'.'p '"' ma'"ak'. 

obtain five California whale. 

t'tu'p'i'tch*C'.tt"^^ 'o'o'i''h^C'.ti'^' ''.-'h'to'p'. 
Became summer, began to be in whale, 
pursuit of so and so 

t'sa'xc'.thla tca"ats"'.'b' ''.''h^to'p'. 

Speared again Cha'atssib' whale. 

'o"\v'.'a'i'ap' ''"'^ t'sa' tv:a"ats''.'b'. 

Repeatedly caused so and spear Cha'atssib'. 
so to be first 

t'saxc'."£thla 'yo'qwa' k'waTsas ''.''h-^to'p "•.'. 
Now speared likewise K'walisits the whale. 

nio""yf.p'athla '^^ '-.•'h'to'p' tca"ats''."b', 
Nowobtained four again whale Cha'atssib', 

W'.ki'm'.'tC',"atl ■''' h'.n''."'p' tca"ats''."b'. 
now became unable obtain Cha'atssib'. 

sut'c'.'"y'.p'athla' '7' k'waTs'.ts ''."'h^to'p'- 
Now obtained five again K'walisits whale. 

su't'ci"'y'.p'tk nia""ak'. hayo'"y'.p'c'.t} ''^ 
Obtained five also California Had obtained ten 

k'wa'l'sits 'a'tiakwah.'p' '■' tca"ats"'.-b' 

K'walisits, obtained eight Cha'atssib' 

'oyo"ai'.t "f. ■■•» 'o'h* "'> kwi'stsatcitl. 

the one who 'had be the one " go oft else- 
perceived so and so wither". 

'a'ttakwal'.-p' 'o'h* 'ana k'waTs-.ts 

Obtained eight was the one, was only K'walisits 

hayu""y'.p' ''^. ha"okw'.'ctt '7' k'wa'l's-.ts 

obtain ten. Now took revenge K'walisits 

ya!a-'na"'.t'q' '''^ t'}'.""w'.n''ap'at'. 

the one by whom had been caused to be laugh- 
— ed able, 

sa'tckok't'cac ''* .'^'u'yi' '" tca"ats''.-b'. 

"Of (him) is sharp medicine Cha'atssib'. 

'a-'q£n'qha'^° 'a'tlakwal-.-p' tca"ats-'.-b' 

For what reason is obtain eight Cha'atssib'. 
he that 

'a'naqa '^' 'p'u'yiwatl ■'- hayo-"y'.p'ath'. '*' 
being alone have medicine fall Now I obtained 
to (his) lot? ten 

si"ya' '^* yaqe''s '^> w-.'k' .'^'u'yiwatl 

I I who am not have medicine fall 

to (my) lot, " 

wa-"att "*^ k'wa'l's'.ts ha"okw'.'£th 

now said K'walisits now take revenge. 


K'walisits was training in secret for success 
in hunting California whales and humpbacked 
whales. And he went to sleep, having complet- 
ed his ritual training. He lay down on his back 
near the fire, but his wife did not lie down on 
her back but sat up. She was looking at their 
door at night. There at the door appeared a 
head, looking just like a panther's. There was 
his head with a tail attached to it, that was 
what it was like. He did not move briskly. 
There he was with a head at each end. 

The woman took quick hold of her husband, 
and pulled at him, endeavouring to wake him. 
K'walisits woke up. There was the superna- 
tural being known as Head-at-each-end. 
K'walisits seized his iron knife and cut here, 
making bleed this thigh of his own. K'walisits 
jumped over him. He made a cut in his other 
leg. It bled. And so he continued, jumping 
from one side to the other, until the Head-at- 
each-end became all covered with blood and 

Eight: Wakashan and Salishan Languages 



died '. K'walisits proceeded to make medicine 
.of him. And this is how he began to capture 
many Cahfornia whales when he went out to 
sea. Whenever lie speared, the}^ say, they 
would die at once. 

And then there came another winter. K'walis- 
its was sleeping in his canoe, was in a little 
canoe through the night. Now there was one 
bearing the name of Cha'ats^ib', another 
whaler, who saw the thing they call " going 
off to another place " ^ 

K'walisits was laughed at because he was 
drifting about aimlessly, asleep, and because he 
was not aware that it was the season of 
" going off to another place ". K'walisits heard 
about how he was being laughed at. His heart 
grew sore and he did not move inland with his 
people for the drying of salmon but stayed 
right at Tlo'owis. The Nitinats moved inland, 
they went off to their river, but K'walisits 
remained behind there at Tlo'owis. All the 
people had gone. 

And then K'walisits began to train secretly 
for success in hunting California whales. He 
began to imitate the movements, up and down, 
of a blowing whale, while on his way to a far- 
distant place ; four days he made motions as 
of a whale, four spans of daylight and four of 

1. Tlic onlvjW.u' toliill this bciiig is to sprinkle luiman 
blood over iiim. 

2. The Nootka Indians believe that twice during tlie 
year, at unknown dates, a big liikev^'arni tide conies in at 
night and shifts everything about in the village, houses 
and all. After a short time everything is restored to its 
proper place. If one happens to be awake at such a time, 
he can train (Jo-S'-iii'lc) for anvthiug he pleases, such as 
wealth or success in whaling, and be sure of attaining 
his desire. Should he step into the water, however, he 
becomes paralyzed for life. He must step into a canoe or 
move back to higher ground. Signs of the approach of 
the mysterious " shift " are the birds starting in to sing 
and the mice running through the house. The people 
become very drowsy just before the " shift ", so that 
few are fortunate enough to be awake during the spell 
and make " medicine " of it. K'walisits too was caught 
napping, great wlialer though he was. His rival was 
more fortunate. 

night. He came to the land which is called 

K'walisits sat down upon the beach. Yonder 
on the river was a canoe-party coming down- 
stream. They came near to where K'walisits 
w^as, seated on the beach and looking around. 
The one of the canoe-party that was at the 
bow came out of the canoe. They had a spear 
sticking up in the bow. Someone sat down 
alongside of K'walisits — it was the canoeman, 
who nudged him with his elbow. " It is fine 
weather today, what do you think ? " he said 
to him, but K'walisits' voice stuck in his throat. 
He became unable to speak and his limbs 
became lifeless. The person stood up and walk- 
ed back to his canoe. He went into his canoe, 
paddled off way yonder, until he was far out 
at sea. Now this one was he whom they call 

It was when the canoeman, the Ch'ini-person, 
set off that K'walisits came to life. He took his 
bearskin robe and rubbed it about on the beach ' . 
For four days he stayed at that place. And then 
also a wolf came out of the woods. He came 
near to K'walisits and urinated upon his 
bearskin, and his urine was like hail. And 
then K'walisits was done and started back ; he 
returned to his home. 

When winter was over, the California 
whales began to run ^ Cha'atssib' threw his 
harpoon and killed the California whale, he 
who had laughed at K'walisits. K'walisits too 
speared a whale. Again Cha'atssib' threw his 
harpoon and again he got his quarry. K'walis- 

1. Wherever the (Ch'ini-person had been in direct con- 
tact with tlie ground was medicine. In most Nootka 
tales, legendary or modern, of the acquirement of power, 
the seeker carries away with him an object granted bv 
the supernatural being, some part or effluvium of his 
body, or some tangible evidence of direct contact with 
his bod\ . 

2. They migrate nortii with the coming of warm 
weather and touch ar various points on the west coast ot 
Vancouver Island at fairly regularly recurring dates. 


VI American Indian Languages 2 


its too speared a whale, he too got his quarry. 
Cha'atssib' secured four California whales, but 
it was K'walisits alone who got five of them. 

The summer came and they started to hunt 
humpbacked whales. Once more Cha'atssib' 
threw his harpoon at a humpbacked whale. 
K'walisits too speared a whale. Each time he 
allowed Cha'atssib' to be the first to throw his 
harpoon. And then, once more, K'walisits too 
speared his whale. This time too Cha'atssib 
secured four whales, but then he proved unable 
to get another. And K'walisits, once more, 
obtained five humpbacked whales. And he had 
obtained five California whales. 

K'walisits had ten ; Cha'atssib', the one who 
had seen the " going off to another place ", 
had eight. He was the one who got eight 
whales, but it was K'walisits alone who obtain- 
ed ten of them. Now K'walisits had his revenge 
on him by whom he had been laughed at. 
" I have found out that his medicine is sharp. 
Why has Cha'atssib' obtained but eight whales, 
seeing that he alone has had real medicine fall 
to his lot? And I have obtained ten, I who 
have had no medicine come to me ", said 
K'walisits. He had his revenge. 


I. 'o'-sun't'c-ati consists of radical element 
'o'-, derivative stem-suffix -son'tc-, and word- 
suffix -ati, 'o'-, lengthened from 'o- because of 
following Simple-, is exceedingly common in 
Nootkain both noun and verb forms; it appears 
as 'o-, 'o*-, and as reduplicated 'o'o-, 'o"'o-, 'o'o'-, 
and 'o''o'-, each of these forms being determined 
by the following element. It cannot be used as 
an independent element but needs always to 
be followed by a stem-suffix to specify its 
meaning. Its function is relational; it indicates 
the person, object, or activity required to 
limit the following element and frequently, as 

here, anticipates a word of specific content 
(here nta''ak' and 'clfto'p'). The nature of the 
relation between 'o- and the suffixed element is 
implicit in the latter ; thus, it may be con- 
strued objectively, as here, subjectively, geni- 
tively, causally, and in other ways. A form in 
'o- always implies that the psychological inte- 
rest centers in the person or object or idea 
with which the logically significant concept is 
connected, not in this concept itself. It may be 
translated a certain (pi-rson, thing), such 
AND SUCH A (person, thing), SO AND SO. Verbs 
in 'o- should be conceived of as answering 
questions of the type '' IFho is it that. . . ? '' 
or " What did he. . . ? " in contrast to ques- 
tions of the type " What did he do ? " Thus, 
'o'swi'tc ma''a¥ means not so much to train 


ELSE. California whales, not secret training, 
is the psychological predicate and is anticipated 
by a place-filling 'o- ; one can also say, more 
synthetically, ma'^ah'san'tc and \'h"lo-p's<.m'tc. 
There are practically as many verbs and nouns 
in 'o- as there are derivative suffixes and com- 
binations of derivative suffixes to append to it, 
in other words, many hundreds. If the signifi- 
cant content of an 'o- word, that is, the idea 
expressed by the derivative suffix, is the true 
center of interest, 'o- is replaced oy hin(a)-, 
h'.l(ay, or an entirely different word is used. 
A few examples of 'o- words, with parallel and 
contrastive forms, will make these remarks 
clearer : 'o-i's to eat so and so, like t^sis'.-"is 
to E.^T meat, contrast ha'w- to eat (as such) j 
'o'o'-/«/ (reduplicated) TO dream of so and so, 
WHAT one OF IS..., like fuiul'c-itul to 
DREAM OF A WOMAN, Contrast po'huds- TO 
DREAM (as such) ; 'j-ha-' to buy a certain 
THING, like futc-ha'' to buy a woman, to 
MARRY, contrast mahv- to buy ; 'o-so'tf the 
one who dies is..., like ivik'-so'ti nobody dies, 


Eight: Wakashan and Salishan Languages 



contrast ^Z^/;- to die ; 'o-'ycha' thk cause of 
one's dying is..., like t'sax-'yrha' to die from 


Hkc nio'-yc' TO give four things, contrast 
hi.n-K' TO give (as such); 'o-iiwt' so and so's 
SON, as verb the one of whom (he) is son..., 
contrast /'fl';/fl' child. 

-sun'tc, derivative verbifying suffix following 
stem form, to undergo secret ritual train- 
ing IN ORDER to gain success IN...; it lengthens 
stem vowel if short. (Derivative suffixes leave 
the stem vowel unaffected, lengthen it if short, 
shorten it if long, or reduplicate the stem 
according to varying quantitative patterns). 
Cf. further 'titc-s\m'lc to train for long 
life {trie TO BE alive) ; ha'ivii-sun'tc to train 
FOR WEALTH (ha lud CHIEF, ha\ud-mis wealth); 


SEA-LIONS (tok-o'k'" sea-lion) ; hi'tc-sr.m'fc to 


(hitc-ina' torch). There are probably several 
hundred such verbifying suffixed elements in 
Nootka, many of them very specific in content, 
which differ from primary verb-stems not 
only in their position but in that they are 
always construed, according to an implicit 
syntactic relation, with a preceding denomina- 
tive term (which may be a " noun " or a 
" verb "stem). Composition of primary stems 
is as good as unknown, 'o'-suii'tc is durative in 
aspect (all verbs have durative and momenta- 
neous, or inceptive, aspects, most have also at 
least one iterative aspect, and m.uiy have still 
other aspects). 

-all, word-sutlix (i. c., attached to complete 
word, not stem) of colorless content. It may 
be translated now, then, and then, so ; it 
seems to indicate state or activitv at a given 
moment of time and implies that the verb 
form is finite. Its use is not obligatory, however. 
The 'of -'<!// combines with precedingsiop (/>, /, 
k, hu, q, (}w,ts, tc, if) mto glottalized stop (p", 
', k\ k'lr, ', '., fs, t'cy I'i respectively ; origi- 

nal q' and q'lv have become .', a peculiar glot- 
tal stop of strangulated articulation and velar 
resonance) ; other consonants remain unaffect- 

'o-sivi'i'cati is absolute (of undetermined 
tense-mode) in form. Absolutes, withor with- 
out -'all, are freely used in narrative ; without 
-all they are frequently used as complemen- 
tary infinitives (see note 7). Absolutes with 
3d personal subject have no personal suffix 
(contrast 3d person indicative forms : present 
'o-sun'fcati-nia' ; preterital 'o'sun't'catf-U-a' ; 
future 'o'sitn't'c-a'q'li-ma'). The personal end- 
ings of absolute (or narrative) forms differ from 
those of indicative forms : 

Sing. I. 'o'smi'Uatf-S'. 

Absolute Present Indicative 


2. 'o'sun'l'cati-.uik' 'o'sun'l'calf-tits 

3. 'o'suii'fcali 'o'S'.nVfcati-ma' 
Plur. I. \rswi'fcat{-n\' 'o's<jn'l'cat{-.n' 

2. \r Sim' t' calf -so-' 'o-suii't'catt-i'Uso'' 

3. 'o'sun'l'catf(''at) 'o's<jit'l'catf-)iia{-'tif) 

A third set of personal endings, used in 
various modal and subordinate forms and in 
possessives of nouns, is clearly related to the 
absolute series : 

Interrogative Possessive 

Sing. i.\ys'.m't'catf-ha-s t'a'tia'k'-qa-s 

MY child 

2. \rs\ui'Ccalf-}M-k Caiurk'-d'-qa-k' 

3. 'o-S'.niifcalf-lm' Cana'lc^-'.') 
Plur. \.\)SKin'lcatf-hi-n' t'ana1;'-q\.-n' 

2. \rs<.)n'l\atf-f_m- 

so'' Ca'na'k' -^'.I'-q'-io' ' 

3. 'o'SKin't'i-alf-lui 

(-'rt/) t\ina'k\-\-af). 

2. Noun of uncertain analysis, -ak' is pro- 
bably durative intransitive {d. qah-ak' to be 
dead; 'nio-'ak to burn [intr.J) or, what at last 


VI American Indian Languages 2 


analysis amounts to the same thing, absolutive 
noun suffix (d. I'sa-'ak' to flow, hence 
STREAM ; t'ca-ak' water ; tca-'ak' island), but 
is now petrified. iiur~, which does not occur 
alone, may be an old verb-stem (cf. Kwakiutl 
7na' TO CRAWL, TO SWIM ?) that has become 
obsolete ; tna-'ak' originally to swim about, 


3. ''.-If-to-p', literally big thing, big variety 
(OE animal). \-h", durative intransitive big, 
TO BE BIG ; inceptive \'iv-a-tciti TO get big, 
GROW UP (Nootka h \ w <. Wakashan xw : w ; 
" of h" indicates voiceless rt-timbre of h, which 
colors and lowers following high vowels, e. g. 
his- < h"\s-, }ps- < }fus-). -to'p' (after conso- 
nants), -cto-p' (after vowels), shortened to 
-{c)liip under appropriate rhythmic circum- 
stances, noun forming suffix added to verb 

stems, THING, KIND, CLASS (c. g. Jlh'-tO'p 
diving kind, SEA-MAMMAL, kwiS'tO'p' DIFFERENT 

KIND, kiu\ktu>.s-tup'-sa-p' [red.] to cause various 

MER, -lo'p' < -to'in-, cf. kui'.sio'm-ah i am a 


ma-'ak' 'vh"krp' are object of 'o-S'.ni'Uati, 
which they follow, as regularly ; or, perhaps 
more accurately, they may be looked upon as 
merely appositional to 'o'-. Like all nouns not 
specifically distributive or plural in form, they 
are indeterminate in number. " And " is gene- 
rally omitted in Nootka ; \c, a conjunctive 
particle, may be placed between the two 

4. A Nitinat name. / does not occur in 
Nootka except in songs for n. K'lual'S'.ls is the 
subject oi'o'sun'tcati ; verb, object, subject — 
this is the most common Nootka order. 

5. ¥rom vji':tc-iiti, momentaneous form cor- 
responding to durative wiitc to sleep, -f -'ati. 
Momentaneous -nil is uncommon for primary 

verbs (cf. also durative -ap</ standing, up in 
THE AIR : momentaneous -apuli ; dur. -tcict' 
on the surface of the water : mom. -Icic- 
tiili ; dur. -'aq'li inside : mom. -aq'stuli^. 
Nearly all momentaneous or inceptive forms 
end in //. This -// disappears before -'all, pas- 
sive (or possessive) -'rt/', imperative -'i', finalis 
-a'-, irrealis -V///' (see note 124), future -<.k' ; 
e. g. -sa-ali (mom.) on the beach <C -sati -j- 
-'ali, -o'-at' (mom. passive) on the face 
<C -o'tl (cf. dur. -0'/) + - (it\ -'t'l' (mom. 
imperative) on the ground < -\ti (cf. dur. 
-'as) + -'i', -C'/a'-h" in order that i may... 
■< -Cili- (see note 18), -c.-ik-ah i shall... 
■<-C',//-. -'rt// (see note i). luiUcu'atl is narrative 
absolute, 3d personal subject, as in note i ; this 
is true of all verb forms commented on in 
this text, unless otherwise explained. 

6. From haioi'-tf, momentaneous form, to 
cease. -//, momentaneous suffix, drops before 
-ali (see note 5) ; comparatively few primary 
verbs add simple -// in their momentaneous 
form (cf. also 'a'k^o'-ti to borrow ; na'o'-ti to 
have for e.\ting). Certain verbs, like hazvc-ti, 
are basically momentaneous because of their 
radical significance ; they can form only a 
secondary quasi-durative by making an incep- 
tive, more properly graduative, out of the 
momentaneous by lengthening its stem vowel, 
e. g. ha'ivi-tf TO begin to cease, to be finish- 
ing (ct. momentaneous wai-citi to be gone 
home, to return home : graduative lua'i-citi 
to be going home, to be on the point of 
RETURN home; moui. tiawoi'ti to approach: 
graduative tia-uvi'lf to be approaching, cf. 
static durative llaiua'' to be near), -ati, see 
note I. 

7. See note 1. Depends as complementary 
infinitive upon preceding haw\'atl. This use 
of an absolute verb form to complete the 
meaning of a preceding finite form is exceed- 
ingly common in Nootka. Note that o"- does 
not need to be specified by a following noun. 


Eight: Wakashan and Salishan Languages 



8. From t'sit'k'-pdf, momentaneous form, 
-\--'ati ; -ti lost as in note 5. t'sit'k-, verb-stem 
TO LIE (in bed) on one's BACK (aspirated stops 
are of purely secondary origin in Nootka, 
being developed from un aspirated surds when 
they stand at the end of a syllable or immedia- 
tely before another consonant at the end of a 
syllable ; to the two Kwakiutl series d [inter- 
mediate] : /{aspirated surd] corresponds a single 
Nootka series/, /'). -/)•.//, momentaneous form of 
-//, locative suffix in the house. Examples of 
dur. -// : mom. -pdi are wt\tc-ii to sleep in 

BED : lUi:tc-pUl TO GO TO BED ; t\qw-\.l TO BE 
INSIDE THE house) : 'natC-p'.tl TO LOOK DOWN 

(when inside). Every local and body-part 
suffix has a durative and a momentaneous form 
in -//, further an iterative with lengthened stem- 
vowel, lengthened suffix vowel, and change of 
momentaneous -// to -/, e. g. t\'q'-pi'i to sit 

DOWN OM the floor SEVERAL TIMES, -'flf//, Sec 

note I. 

9. \n'k' , noun or durative intransitive, fire, 
TO BE BURNING (saidof fire, not burning object). 
From original *'an-aV ; -a¥ , durative intran- 
sitive suffix, see note 2. The common Nootka 
groups in' and iui\ in which the ' represents a 
murmured i-vowel, go back to fuller forms of 
type a (or i, //) + n or ;// + a (or ■., u), in 
which the second vowel is unaccented (e.g. 
-all momentaneous sound : -\in' continuous 
sound < -.'"r.«' < *-.ana, in Wakashan terms 
*-qa-ti : *-qa'-la ; 'ana'h'-is small : reduplicat- 
ed W'n*h'^-\s SEVERALLY SMALL < *'a'a'nah-). 
-\ (-' is merely a breath-release after all final 
vovv'els ; it is not heard if the word is pronoun- 
ced in close contact with the following word), 
suffixed definite article, often used as nomina- 
lizing element. Properly speaking, all " nouns" 
are indeterminately such, being formally iden- 
tical with durative intransitives (e. g. qo''as a 
PERSON, TO BE A person) Until noniinalized by 

-C or an equivalent element. Syntactically, 
\u'k'''.' is objectively related to the preceding 
verb ; the difference between a direct object 
and an indirect object or local phrase (at the 
fire) does not exist for Nootka, because the 
indirect or local relation is generally expressed 
by a suffixed element in the verb or is other- 
wise absorbed in the verb ; the fire here ampli- 
fies the more general local idea of in the house 
conveyed by -pi-(ti), no specific rendering of 
our AT or near being therefore necessary. 

10. u'lk-, durative intransitive, to be not ; 
wik' not, no ! is really verbal in form, w.k' 
consists of archaic stem lui- and durative intran- 
sitive -k', d. -ak', -ak' (notes 2, 9) ; most 
Nootka derivatives of to be not are based on 
Will- (e. g. luiFnit' to be not-stocked, to have 


there are also a number of more archaic forma- 
tions based on wi- (e. g. wi-ma-h'ti to be 
UNABLE ;'ti a man is angry), -all, 
see note. i. -uk' (after consonants), -ak' 
(after vowels ; often contracts with prece- 
ding vowel to -a'k', -ak'), possessive word- 
suffix indicating that the subject of the verb 
is the owner of the following alienable noun 
(here woman) : his (wife) did not, he had (a 

wife) WHO DIDNOT, W/d''fl//M^'ti;df/» nowmine isnot, 
NOW I HAVE WHAT IS NOT ; cf. nOtCS I 5, 24, 26. 

11. See note 8. Complementary infinitive 
dependent on preceding verb. Negatived state- 
ments are always expressed by treating the 
negative as the main verb and having the verb 
proper follow as an infinitive, e.g.ivikila himn' 


hiq'citl DO NOT (imperative) to-^speak! do not 

12. Absolute form of noun, assimilated from 
io'lc-sma. io-tc-, lengthened from stem iutc- 
WOMAN ; examples of derivatives are iut'c-i'h" to 
want, BE after a WOMAN, iiitc'-na'k' -citl to 



VI American Indian Languages 2 


tci' TO LIVE AT oxe's wife's HOME, fo'tc-azu'/qc 
TO CALL FOR A WOMAN, fufulc-atah (red.) TO GO 


-sma\ -s'ma\ absolutive stem-suffix lengthening 
stem vowel ; not freely used as derivative ele- 
ment, probably compounded of absolutive ele- 
ments -S-, -s'- (c{. absolutive suffixes -s-yi', 
■s-'yup' , -s-yin', -s'-fnnni') and -ma' (cf. li'tc'- 
ma' REDHEADED WOODPECKER), presumably 
reduced from older *-sa- as indicated by irregu- 
lar plural io'ts-sa-mch" women. Object of 
tuik'atfuk\ though logically subject of iv.k'atf 

13. t'<.qw-, verb stem to sit ; cannot be used 
without following local suffix, e. g. t'e'.-as (<i 
*t'\q\i'-as) TO SIT ON the ground, Ciqw-a's to 
BE seated on, i'e\-ih"ta to sit at the end, 
t\q'-s'ato--'asTo siton thegroundatthe door. 
-t/, durative local suffix on the floor, in the 
house ; for corresponding momentaneous form 
see note i. 

14. Assimilated from 'na'lc-sa'ti{c{. note 12). 
'invic-, lengthened from 'natc-, verb-stem to 
look ; other derivatives are 'natc-fso' to look 
into (a barrel), 'natc-i'itf to look into the 
house, 'natc-a'yii to be looking up in the air, 
'natc-mai-apC to look all around, 'nat'c- 
aq'ff-a'a' to be looking into the fire, 
'na'tc-uk' to look for, 'natc-n'af to see, 'tm- 
'na't'c-a'i {red.) to watch, 'na'ts-sa' to look 
at, watch, see is durative ; as momentaneous 
is used 'natc-ii'af (-n'at, -yiiai to get sight 
of, perceive), -sa' , durative suffix lengthening 
stem-vowel, not freely used, cf. durative -a' ; 
perhaps identical with stem-lengthening -sa' 
very, just, -most, too. -sa-il is contracted from 
-sa-'alf (^-'atf contracts with certain preceding 
vowels to -a'tf, -ati, according to rhytiimic cir- 
cumstances ; e. g. -ape standing + -'ati > 
-apati, -'alo' into the water -f -ati > -'alaii, 
durative -a' -f -aii > -a'ti, -ati). -'ati, see 
note I . 

13. t'aci-', noun trail, doorway ; absolutive 

in -/-'jStem t'ac- (e. g. t'ac-kuin' with a trail 
IN the center), -'ak' (after vowels ; often con- 
tracts with preceding vowel to -a'k\ -ak'), -uk' 
(after consonants), possessive word-suffix for 
alienable nouns ; for inalienable possession see 
note 26; for possessive paradigm, see note i. 
Their is ordinarily not distinguished from his, 
her. its. -'i', nominalizing element, see note 9 ; 
faci''\ik' alone would mean to be one's door. 
t\ici'"ak'\ is object oVna'tsiVtt. 

16. 'alh"-, verb-stem to be night, 'ntl/'-citi 
night comes, 'ath-e-' is durative in aspect ; verb 
stems endingin /'take -e'', -e' as durative suffix 
instead of normal -a'', -n' . Like other absolute 
durative forms, \ilhc'' can be used adverbially ; 
properly speaking, it is a complementary infini- 
tive following 'na'ts'a'li. 

17. Demonstrative pronoun and advert, that, 
there ; properly a verb to be there. It is bas- 
ed on simpler demonstrative 3'^', ya that; for 
-/, probably an old local suffix no longer freely 
employed, cf. /;// (note 24). 

18. katih"-, verb-stem to appear, come into 
view, -ff.//, momentaneous (or inceptive) suffix. 
Most verbs form their momentaneous aspec- 
by adding -c<.-li to the stem if it ends in a cont 
sonant, -ta-ti if it ends in a vowel, often -kw\.- 
ti (see note 32) if it ends in w, o' . Complemen- 
tary infinitive depending on ya'i : there-was 
to-appear ; ya'isi katilfcdi there i appeared, 
not ya-i katiJfatisC. 

19. Seenote 1 5. Object ofyfl"/^a////Vi//,/W''t' 
amplifying the local idea in ya'i ; cf. note 9. 

20. Absolutive. Noun-stem t'oh"- head (e.g. 
t'j}f-n<.' cod-head that drifts to shore, t'o''w- 
rs to eat a fish-head). -fsitC, absolutive suf- 
fix, not otherwise found. Subject of ya'i 

21. Properly qwa'', verb stem and durative 
absolute to be in quality, to be like. 

22. Umlauted from *qzva'-\t'q' ; a and a' 
immediately followed by i or r are umlauted 
to open z and z' (these vowels are felt as dis- 

Eight: Wakashan and Salishan Languages 



tinct from secondary e, i and e' , '' that are 
merely lowered from i, c because of preceding 
or following velar consonant), qwa'-, see note 
21. -\t'q' third person relative or subordinate, 
indicating various subordinating relations, such 
as comparison, time, place, relative clause (cf. 
notes no, 139, 159) ; the precise nature of 
the subordination depends on the verb. -'1/ q' 
may be considered as a nominalizedform, paral- 
rel to -'i' (see note 9), of the subordinate -qa- 
series (see notes i, 7^1). 

23. Absolutive form of noun. Stem k'ayntn-, 
k'ayiip'- (intervocalic -m-, -n- become stopped 
to -/)', -t' at the end of a syllable), e. g. 
k'ayup'-q-inakTOimT.\TE A panther in a dance. 
-in', absolutive noun suffix ; other examples are 

bttC->.n' SMALL CLAM, ho'p-m' SALMON TROUT, 

mxl-m' SALMON-EGG, Miv-in' GRISTLE, tca'skw- 
i«' BACKBONE, hd-f sn' t-'.n' SEA. Subjcct of subor- 
dinate clause qivi'\t'q' . 

24. Jhf, verb-stem and durative absolute to 

BE HERE, to BE THERE, TO BE AT ; probably 

consists of old demonstrative stem /;',- {d. bin-, 
h'.t-, notes 1,101 \h.s- to be at such and such 
A place) -f -/-, petrified local element (see 
note 17). -'at' (often contracts with preceding 
vowel to-rt*/', -al' ; affects preceding consonants 
like -'fl//, d. note i), possessive word-suffix 
referring to possession of inalienable noun, 
nearly always body-part (cf. note 10 for cor- 
responding alienable possessive suffix), hfat' 
is complementary infinitive depending on ya'i : 
there (he) was with his (tail) at, there he 
(was) having his at. 

25. Absolutive noun, tail (or mammal). 
Stem probably 5i/'- ; -a' absolutive noun suffix, 
identical with durative intransitive -a' (other 
nouns in -a' are faha' ghost, qama' trap, 
pa-liprftfa' substance for face paint, nat'ca' 
tail [of fish], kap't'a' pointed stick). Object 
of Ihfat', th :)ugh logically subject of /;-./ ; d- 
note 12. 

26. Contracted from t'oh^l'sdr. head (see 

note 20) -\- -at'-V. -'at', identical with verb- 
suffix -'a/' (see note 24), inalienable possessive 
suffix appended to nouns, chiefly bodv-part 
nouns ; for corresponding verbal and nominal 
alienable possessive suffix, see notes 10, 15. 
With pronominal elements -'at' combines 
exactly as does -uk' , -ak' (see note i) : 

Sing. i.t'ok't's\tat' my head Plur. i. t'ojft'sdat' 

-qa-5 -qi-n' 

2. t'jh"t's'.tat' 2. I'oh^i'sdat' 
-d'-qa-k' -'>.t'-q'-so'' 

3. t'oh^t'sdat' 3. t'oJfl'sKtai' 

t'jJft's'Jat"r.' is local object of /?>./ (syntax as in 
notes 9, 19), while s\t'a' is direct object of -Vr/' 
in hifat' to have. . . being at. . . 

27. See note 21. 

28. Demonstrative pronoun (this), adverb 
(thus), or verb (to be thus), used as general 
demonstrative ('al/'ko' ' is more specifically this, 
ya-' that); syntactically a complementary infi- 
nitive defining the preceding verb, which gives 
the nature of the relation, such as time, place, 
or manner, as here. 'ah"'a' is compounded of 
demonstrative stem 'aJf this, that, which may 
occur alone, and -a' , probably petrified demons- 
trative or local element. 

29. See note 10. 

30. Durative form of verb. No etymological 
analysis suggests itself. 

31. Assimilated from hc-tsaq't'so' ; cf. notes 
12, 14. h\c-, verb-stem to bf. all, to be both ; 
other derivatives are hic-im'f to be assembled, 
h'X-sa'tso' TO BE everywhere, -tsaq'fso' , local 
suffix AT THE END (see also note 38). Many 
verbs with local or body-part suffixes are to be 
interpreted as "bahuvrihi" compounds, i. e. 
the radical element expresses a concept which 
is possessed by the subject ; e. g. fwk'^-t'so' to 

NOGRAPH, 'ayaqs to have much (game) in the 
CANOE. The object of the verb is head. 


V/ American Indian Languages 2 


32. SU-, verb-stem to hold, get hold of ; 
its aspects are Jurative so' ' to hold (other mono- 
syllabic duratives with lengthened stem-vowel 
are /<:/•' TO pull, drag, t'fo'' to remember, qiun'' 
to be in Q.UAL1TY), momentaneous su-kwjf to 
GET HOLD OF, graduative scr-huUJ to begin to 


iterative so'tiso'ya' to hold time and again, 


-kwitf, momentaneous sufhx (cf. note 18), 
etymologically identical with postvocalic -tcuf 
(cf. -tci' AT : 'o-hvi to be at such and such a 
place) ; Nootka tc: hw (after u) goes back to 
Wakashan ^':/«c; (^d. Kwakiutl change of Oi;;- to 

33. tcakop' MALE, HUSBAND, irregular absolu- 
tive to which corresponds as stem Irapxw- (e.g. 
tcapx-nak'-f'.tf to marry a man), -o^w- (final 
form-o^'', -ok'"; ^sounds are labialized aftero), 
alienable possessive suffix after consonants (cf. 
notes 10, 15) ; -'i, see notes 9, 15. Object of 
suhviti. Note that -okvj-'i refers to possession 
by the subject ; if HER husband had referred to 
another woman ,than the subject, siikivitftcp' 
tcakop'^i took-hold-of-another's the-husband 
would have had to be used. 

34. Umlauted from *lo'ts'ma-\\ cf. note 22. 
See notes 12, 9. Subject^of sukwiti. 

35. tci'-, verb-stem to pull ; durative tci'', 
momentaneous tci'-tc.ti. -Icdi, post-vocalic 
form of -citi, momentaneous suffix, see 
note 18. 

36. tlup'k- or ti'.m'k-, verb-stem to be awake; 
see note 37. -sa'p' (alternates for rhythmical 
reasons with -sap'), causative of momentaneous 
-all, while causative -'yap' corresponds to -tcUi 
(other examples of 'Sa'p\ -sap' are qah"--sa'p' 
TO KILL : qah"~Cili to die ; 'utcq' -sa'p' to cause 

TO BE A FOG ; hh''-sa'p' TO MAKE A HOLE ; 

hamat'-sap' to cause to be known, to find 
out) ; alternates with -sa'm-, -sam- (e. g. tiup'k'- 
sa'in-ah i cause to wake. up). Every intran- 
sitive aspect has its corresponding causative in 

-p' , -in- ; momentaneous -tl always drops in 
causative forms (e. g. -pdi in the house, 
mom. : caus. -pitap' ; -as\ti on, mom. : caus. 
-asip'). -ati, see note i. him is understood as 
object ; third personal subjective and objective 
pronominal ideas are not specifically expressed 
in Nootka. 

37. Momentaneous intransitive, see notes 36, 
18 Aspects recorded of tlup'k- : dur. tfnp'k-a'' 
or tiim'k-a'' to be awake, mom. tlup'k' -ati to 
WAKE UP, graduative llo'p'k'-all or thm'k'-c<ji 
to be waking UP, durative-iterative //t>-/)7v''-n'-/ 
TO be waking up time and again, momenta- 
neous-iterative tlup'k' tliip'k-c or thm'k'lhm'k-c 
to keep waking up by fits and starts. 

38. Cot'olf-, reduplicated from f^h"- head, 
.see note 20 ; reduplication expresses distribu- 
tion, HEAD HERE AND THERE. -tSUq't'so', SeC 

note 31. This word is a "bahuvrihi" : having 

A HKAD AT EACH END, d. note 3 I. 

39. tsikwun', noun iron ; borrowed from 
Chinook Jargon, -'ak -'<.', see note 15. 

40. /V/-, verb-stem 10 cut (e. g. t'ci-maht' 

TO CUT A BODY, tC\~h"tiVk TO BE CUT APART, fci'- 

mcf.' mussel-shell knife) ; dur. t'c'rya to be 
cutting, mom. Cci-tcitl, cf. note 35. 

41. General demonstrative, object of /V/Zr-.// ; 
see note 28. 

42. = his-sa'p . his-, verb-stem to bleed and 
noun-stem blood (e.g. his-m\s blood, hihcs-sitl 
[red.] TO be bloody-eyed, hey-i's to drink 
one's blood), -sa'p', see note 36 ; to 
BLEED, mom. 

43. Demonstrative pronoun, object of /i^^v/'/)'. 
Compounded of \ih" (see notes 41 , 28) and -ko', 
not otherwise found. 

-14. Contracted from 'a'ap'-ts\.fata-al'-\', d. 
note 26. 'aap'-, reduplicated from 'ap'-, 'am-, 
noun-stem of general locality part, body-part 
(e. 'g. 'ap'-qe' summit, ^ap-p'.q' tl-il on a mat 


chest, \ip"-'win' WAIST, 'am-ak'th.' butt end); 
suffixes indicating body-parts occurring in pairs 


Eight: Wakashan and Salishan Languages 



generally reduplicate preceding stem even if only 
one of two is actually referred to (c. g. \{ap'- 
p\qa' KNEE, 'a'''ap-s"'iuin'i arm-pit, 'a am- as 
CHEEK, 'aa'ni-anid shin ; from other stems, 
e. g. yaya'k'-'nuk'" to be sore-handed, yaya'k'tv- 
wi'i TO be sore-eared, lotop'k'-atsoh'^ to be 

BLACK ON THE SOLES, 'nilts'}lO'q'"-Sui TO HAVE A 

BOIL ON THE eye). -ts-fatcC, body-part suffix 
THIGH ; compounded of -tS'.t- side, apparently 
not found uncompounded (other examples are 


THE head], hopai-tsu'ak'lii to be moon-si de- 

UPPER THIGH AND RUMP), and -a'tci, -atCi PRI- 
VATE PARTS, VULVA (c. g. t'h'.a'tc' TO SHOOT AT 

THE PRIVATE PARTS ; compounded, e. g., in 
'tiJq' "-C sa'.alcC to have a boil below the navel, 
yaya'k'-si'o'w-atcr to be sore on the inner 
PART OF THE thigh), lieuce properly on the 
SIDE OF THE PRIVATE PARTS. -at'-C see notc 26. 
45. tuxw-, verb-stem to jump (cf., further, 


mom. tux"-citi, iterative lo'x"to-xw-a' to jump 
UP AND DOWN, -tspa' , -tspa'\ local suffix over, 
PAST, durative aspect (other examples are fsax- 

tSpn'' A SPEAR GOES OVER, h\ta-tspa' TO GO OVER, 

TO PASS, kannt'q'-tspa' to run past) ; corres- 
ponding momentaneous aspect, -tspm'ii. 

/\G. See note 40. -tia, word-suffix (or enclitic 
particle) also, too, again. 

47. klUlS-, stem the other, DIFFtRENT (c. g. 

kw\s-i-'y-as to be atthe far end of the village, 
kwi.s-aq't'so' [a house] stands opposite, kwis-t- 
i'ya' to be at another time, Inw.s-to'p' to be 
of a different class, abnormal), -a's, -as, -s 
(after vowels), local suffix of durative aspect 
on, at (e. g. hin-a's to be on, fiqw-a's to be 
sitting on [a box], k'wa'i-as branches are on 
[the logs]), mom. -a'siti, -{a)siti ; kims-a's to 
be different, the other On,'s'al'i' iiiclh.n' 
the leg of him which is otherwise, elsewhere 
on, atfached to [him], i. e. his other leg. 

-'at-W see notes 26, 44; -'i' relates kiuisa' s' al ' 
to tiictii.n'. 

48. Absolutive form of noun, probably an 
irregular reduplication. Object 

49. Momentaneous intransitive. See note 42. 

50. kwcs-, see note 47; lengthened form ot 
stem because of iterative aspect, -pano'i, uera- 
tive form of -pa'' side end {kwis-pa'- to be 
on the other side, at the other end, often 
with -s- ON. : kwis-pa'-s-'); to durative -pa'" 
corresponds momentaneous -/>'.«'// from *-paimif, 
whence iterative -pano'i by change of-// to -/ 
and lengthening of u to o' (momentaneous 
forms in -in'ti correspond to iteratives in 
-ano'i or -m'i ; another example is dur. -misa' 
MOVING UP, mom. -tmsmHi, iter, -nnsano'f). 
-ci-'atf, from -ati -\ — 'ati, see notes 18, i ; for 
loss of-// in inceptive suffixes, see note 5. 
kivi'spano'icitf is iterative-inceptive in aspect : 


examples of this aspect are isu'tstsu'tsati to 


51. his-, see notes 42, 49. -an^'yaiui-, longer 
form of -\.tn''yo'-tf, -uni'yu-li, momentaneous 
form of dur. ->,;//'/ (after' vowels, -q-wi'f), used 
partly as classifying suffix (round object ; 
moon), partly as local suffix all over, cover- 
ing A ROUNDED OR BULKY SURFACE (e. g. f hs- 

iw'/to be WHITE ON THE OUTSIDE; ya'k-iini'f TO 


/;'.f-iw''j'()'//TO ASSEMBLE, COME together). When 
momentaneous -// drops, as before -'«// (see 
note 5), modifications often appear in the pre- 
ceding derivative suffix (e. g. -tmsin'li up, 
mom., but -tmsano'alf ; -o'ti on the face, 
mom., but -aw.- ati, causative -azvtip' ;. -so'li 
so AND so DIES, butsa'wCati, -sawi-'ati, causa- 
tive -sawup' ; such cascsof -(r-(-()-) : -au'i-,-au'u- 
suggcst that Nootka o' is sometimes contracted 
from older rfzi^, an (cf. (/o/ slave : qaqo'i slaves, 
reduplicated, from *qaqaiii). -ati see note 1. 


VI American Indian Languages 2 

52. See notes 38, 9. -t'so-- with long vowel, 
-t'so' with short ; long vowels are 
frequently heard shortened in final posi- 

53. qaJf-, verb-stem to be dead, to die ; 
dur. qah-ak' to be dead, mom. qalf-cdt to 
DIE. -a-\ti, see note 50 ; t phonetic variant of 
a, X, because of preceding i. 

54. \:>uyi' (.f'ji' ; o' and i' are often brok- 
en to Jii, ail, and ti, ai after h and .), noun 
medicine; probably derivative in -y\ (cf. mih- 
SPEAR, absolutive mds-'yi ; further, derivatives 
in -s'yC , e. g. 'mukw- stones lie : 'nuik-s'yi' 
STONE, ''.«'/r' fire, to burn : '^n'k-s'y.' wood, 
stick). -Ici'i (after a and -.), -kwi'i (after 0- 
vowels), -r/ (after consonants), derivative suf- 
fix attached to noun stems, to make (other 
examples are'o-fef/"/, to make so and so, iihai- 

i'f TO MAKE A CEDAR-BARK MAT). -C\.-\tl^ see 

note 53 ; -ati is here inceptive. 

55. qwis, durative verb to do thus, to act 
as described ; perhaps related to qwa'' (see 
note 21) and to qwK-, stem used in relative 
verb forms to make indirect questions (e. g. 
qu'iyifftaqak-i'tc WHAT, as they say, it is made 
OY , qwe-sa'h'^i-vjos-C why he would be..., qwe- 
y.-n'-yi's at what time i come) ; -s-, possibly 
identical with local -s on, see note 47. -//' (after 
all consonants but h), -qff (after vowels and h), 
before all other word-suffixes) and indicating 
that the activity or state predicated by its verb is 
accompanied by or in some way conditions the 
activity or state predicated by the following 
verb or an understood verb (e. g. h.t-ii-h-ah 
ho'ya'l i-am-in-the-house-while dancing, uut- 
qh-atl-ni now we say it while [thus occu- 
pied!) ; here it implies that the following verb 
(to obtain many in hunting) results from 
the activity (to do thus, i.e. to make medi- 
cine) of its own verb, -ati, see note i. 

56. 'a'y.m'k'y from *\rya-mik' , see note 9. 
'a'ya-, lengthened from 'a\a-, verb-stem to be 


MUCH, MANY (durative absolute '(iya) (e. g. 
'aya-qs to have much [game] in one's canoe, 
'a-y-ip TO secure much, 'ayir-p'd' to be many 
times), -niik', -mi'k\ derivative verb suffix 
lengthening preceding stem-vowel, to succeed 
IN hunting... (e. g. 'o'-ini'k' to get so and 
so IN hunting, to be a successful hunter). 
-a-<xtl, see note 50. 

57. 'napxta'', durative intransitive to die 


note I. -qo'-, conditional suffix, see note 58, 
frequently used in main clause as past usita- 
tive (e. g. mala'- ati -qo'-k' whenever, if you 


-wt\n' shows that-(/o'- cannot be here under- 
stood as subordinating, as -tc- is quotative in 
subordinate clauses (e. g. 'napxta'-ati-qo'-lc 

IF, AS IS said, he dies IMMEDIATELY). -lVt\.n\ 

quotative word-suffix it is said in main clau- 
ses ; replaces third personal indicative -ma\ e. 
g. qaipak' -wiai' he is dead, they say : qaipak'- 
ma' HE IS dead (ist pers. sing. -wi\-S'.' , 2nd 
per. sing, -u'z'oi-tsuk') ; << *-wa-\n', probably 
petrified nominal derivative (omwa'- {wa-) to 


58. t'saxw-, t'sax- (labializations regularly 
disappear in syllabically final position), verb- 
stem to spear (e. g. fsaxw-i''nak' to imitate 


OVER, t'sax-yak' spearing-ikstrument, speak); 
mom. t'sax-C'.ti, iter, t'sa'xt'sa'xzv-a' . -c.-eti, 
see note 50, 53. -qo'', conditional suffix; its 
paradigm is : 

Sing. I. -qo'-s 

2. -qo'-k' 

3. -qo-' 

Plur. I . -qu-ti' 

2. -qo'-so'' 

3. qo(-a{) 

With quotative -tc (cf. note 57) it forms : 

Sing. I. -qo'-ts-i Plur. i. -qo'-t'a-n' 

2. -qo'-tc-k' 2. -qo'-ts-so'' 

3. -qo'-tc }.-qo--tc(-'ai) 

Eight: Wakashan and Salishan Languages 



59. 'ah'^'a', see note 28. -'ati, see note i. 
-wz\n\ quotative, see note 57. 

60. ttao'' ANOTHER, as durative verb and 
verb-stem tla'o'- to be another (e. g. tia'o'-fsiq ' 


TO BECOME another). -'«//, see note i. ~iia\ 
see note 46. 

6 1 . f so- itch" to be winter ; mom. t'so\tch"citi. 
t'so-\lch" is explained by theNootka as washed- 
SEASON, i. e. season when everything is washed 
clean by rain and snow, cf. t'so-, verb-stem 
TO WASH, -itch", -'i'tclf, stem suffix season (e. 
g- 'mil'i-i'lflf rainy season, la'y-i'lcJf rotting 

SEASON, FALL; cf. ^\sO -(]'-\tch" YEAR, Q. g 

nio'-q'\tch" four years), -cilf, momentaneous 
suffix, see note 18. 

62.iuz\tc, see note 5. -'ah"s, dur. local suffix 


mom. -'al/'sitf, caus. -ah"-s\p' (cf. further W/- 


hayu.-ah"s lo have 200 (salmon) brought 
HOME IN A canoe) ; -alfs is one of those suf- 
fixes that " harden " preceding final consonants 
of stems, i. e. preceding p, t, k, hv, q, qiu, ts, 
ic, // become glottalized to p\ /', k\ k'u', ', !, 
t's, fc, /'/ respectively (cf. note i), s, c, / to 'y 
(-/-'- above is irregular), .v, xiu to \v, n to ';/, 
in to 'w, h to }/ (sometimes 'zu) ; see -^.tch", 
note 61, for another " hardening" sutfix. -ati, 
see note i. 

63. 'annh-, probably identical with \inah to 
BE (so AND so) IN SIZE ; -\s, diminutive suffix. 
'annh- ' is doubtless based on \via- (dur. abso- 
lute 'ana') only, see note 164 ;cf. further V/z/^t- 
(with interrogative -ha) how many ? \rna- 
TO last, to be IN EXTENT, 'ani-ts- as long as, 
'ana'-...-\s to be near. Diminutive -'a is 
freely used as word -suffix, less frequently stem- 
suffix, in both noun and verb forms, e. g. 
la' nz~ '.s cuiLD, ha'hua'ti-\s little girl, tand- 
q'tl-'nnkw-'is TO BE child-in-handed-little, 
to hold a child in the hand. In 'anah-'is 

little, -\s may be separated from 'anah-, e. g. 
'rt"jm/>(r;)//i''-'u to BE LITTLE-HEADED. Here used 
as adjective qualifying canoe. 

64. Absolutive noun t'capats canoe (irregu- 
lar plur. tcn'ya'pats), local object of hd. t'cap- 
(sometimes t'catn-), noun-stem canoe (e, g. 
t'ca'p-ok'i canoeman, see note 107, t'cam-e.'ali 
TO return in a canoe after going in some 
OTHER way) ; in most derivatives t'capais is 
treated as stem (e. g. fcapats-nak' to have a 
CANOE, t\apals-diua' canoe-place on the 
rocks), -ats, derivative noun-torming suffix 
receptacle (e. g. \oqw-ats urine-receptacle, 
bladder) ; cf. also -sats receptacle (e. g. 
k'o'f-sats vessel for eating a relish out 

65. V, see note i ; refers to following noun. 
-yoaf (after vowels), -o'ai (after consojiants), 
derivative suffix to get sight of, perceive 
(e. g. wawa'-yn'af to hear what one says, 
qo'ats-u\if to SEE A person), -atf, see note i. 

66. kw.stsatcili, momentaneous verb form to 
GO off elsewhither, nominalized by -t'ca\ 
kwis-, see notes 47, 50, -tsa-tcitf, mom. suffix 
TO GO off to, cf. note 35 (e. g. 'u-tsa-tcilf to 
GO TO so and so, ho a-tsa-tcili to turn back); 
corresponding caus. is -tsa-'ap' , e. g. 'n-isa-ap' 
TO take so AND SO UP To) ; as corresponding 
dur. intr. is used -lsuuk\ -tsc'nk' (e. g. 
lua-S'lsu'nk'-ha-k' where are you going ?). 
-tea' , suffixed article or nominalizing particle 
with quotntive color, the..., as they say (for 
related forms in -ic- see note 58); syntacti- 
cally equivalent to non-quotative -\' (see 
note 9). 

67. 'jh", durative verb to be the one, thh 
one who... is so AND SO, indicates that the 
noun following or understood is the one refer- 
red to in a preceding verb ; here : the one who 
(perceived the shift) was (cha'atssib'). This 
verb probably consists of '()-, see note i, and 
petrilied suffix //', possibly to be {d. perhaps 
-/.'" in 'aJf, note 28, and 'aiia-h", note 63). 


VI American Indian Languages 2 


68. NitiiKU name ; h and final glottal stop 
show at one that it cannot be Nootka proper. 

69. 'u-, variant of '0-, see note i. -kia(^y 
(after u- vowels), -/^/rtf()' (after other vowels), 
-iaCy ('"^f^ter consonants), dur. verbifying suf- 
fix TO BE NAMED... (c. g. 'ag'-tctu-ha' HOW IS HE 


70. lim'h.-, noun-stem name, absolutive 
.im'tc' ; cf. verb-stem Um'/- to sing the signi- 
ficant syllables OF A song. -«a"A'', -Mfl//, verb- 
ifying suffix TO have... (e. g. 'u-na'k' to 
HAVE so and so, t'atia-nak' to have a child, 
imy-tci'i-uak' to have for four days). [=i)i'li- 
na'k , complementary infinitive depending on 

'iikla' ; literally, (cha'atssib') is what he was 
called TO have a name (or in having a name). 
The idea conveyed by a derivative suffix is fre- 
quently supplemented by a following primary 
element conveying the same notion, e. g. 'o'cs 
ha ok'" to-consume-so-and-so to-eat, 'utshtn'tf 
hifcna'k'ati to-marrYtSO-and-so to-begin-to- 

71. 'o'o'-, reduplicated form of '0-, see note 
I. -laff, derivative suffix reduplicating stem 
and lengthening stem-vov/el to hunt (such 


J2.l'h-'u".n', laughing-stock, TO BE A CAUSE 

OF LAUGHTER, derivative of verb-stem t'ii'xiu- 
TO LAUGH (e. g. mom. fii'x-citi to laugh 
ONCE, dur. t'H-xzv-a' to be laughing, I'ii'x-p'i- 

tdf to laugh while ENGAGED IN SOMETHING 

else) ; -\u- is " hardened " from -xw-. -in', 
" hardening " derivative suffix, apparently 
m.ikes passive or resultative nouns out of 
verb-stems (e. g. t'coc- to suspect : t'coc-'<.n' 
ONE who suspected, haca'- to hear about : 
hady-oi' famous). -ap\ -am-, causative suffix 
(ci. note 36); t'ic'ivi.n''ap' to cause to to be 
laughable, TO treat as a laughing-stock, i. 
c. to laugh at. -'at' , passive suffix, see note 
I ; identical with inalienable possessive -'al', 
see note 2^. 

73. Relative particle introducing staiemen.t 

of cause or other attendant circumstance in 
following subordinate verb. Probably petrified 
relative in -i, as shown by its pronominal 
forms, which are those of relative forms in -i-, 
-.-, -3'..-, -_)V- : 

Sing. I. 'atn-s 

2. 'am-V 

3. 'am' 

Plur. I. 'aui-n' 

2. 'ajH-so' 

3. 'anC 

Cf. relative paradigm in note 185. 

74. qa'ya'- to drift (in a canoe) ; may 
contain -^--outat sea in a canoe (e. g. hin-a'- 
tc'.tf mom. TO GO out to sea in a canoe, ivik- 
a'-nak' to have none in a canoe), -panatc, 
derivative suffix generally lengthening stem 
vowel, to wander acout aimlessly (e. g. 
^o'tcq' -panatc to get lost in a fog, ya'is-panatc 
TO BE OUT FOR A WALK, hita'-q' t'l-as-panotc to 
SPEND one's time IN THE bush). -qa. Subordi- 
nating suffix indicating cause or other atten- 
dant circumstance, often, but not always, in- 
troduced by \ui<. ; it is etymologically identical 
with -qa- of possessive paradigm (see notes 
I, 26). Its pronominal forms are ; 

Plur. I. -qi-n' 

2. -\t'-q'-ifl'' 

3. -qa{-'af) 

Sing. I. -qa-s 

2. -it'-qa-k' 

3. -qa' 

'an<. qa'ya' panatcqa indicates cause of t'fi'- 
'u".?i''ap'at' ; more explicit causal statements 
are rendered by 'o-no'tf 'am... -qa' to-be-for- 
such-and-such-a-reason, that... 

75. See notes 5, 74. Here -qa' denotes at- 
tendant circumstance, while asleep ; iviilcqa' 
follows clo.sely on qa'ya' panatc and needs no 
formal 'am to introduce it. 

76, u".k'-'ai NOT TO BE AWARE OF, dur., con- 
sists of luik' (see note 10) and derivative suffix 
-ai (certain suffixed elements beginning with 
glottal stop do not glottally afl^ect preceding 
consonants) to be aware of, have news OF(e. 
g. 'o-'af TO HAVE news OF SO AND so), inceptive 
-'<:-tc<.ti TO become aware of (contracted from 

Eight: Wakashan and Salishan Languages 



older *-'a/-, « softened » form of -'d/). -'atf, see 
note I. -qa, see note 74. u'ik"afalfqa follows 
'aw. (note 73) and is parallel to qa'ya'panatcqa ; 
AND is not necessary (ci. note 3). 

77. See notes GG, 74. Here -qa marks a 
subordinate clause that is objectively related to 

78. Absolutive and stem-form : daylight, 

DAY, also PERIOD, SEASON, WEATHER ; as dur. 

verb, TO be day, mom. 'na's-atl day comes. 
Possibly 'na's is composed of simple 'na'- (cf. 
Kwakiutl'wrt'-/^ day) and petrified -s (on ?), 
cf. qwis, note 55 ; this seems to be confirmed 
by '^oolkd.^ na' - p' <.naq- the proper time comes. 
'na's is subject o{ kwistsatcitiqa : that day was 


79. Umlauted from haca'-tcili. haca'-, dur. 
stem TO HEAR ABOUT; dur. -a' (but not short 
-fl) is umlauted to -r- before momentaneous 
-tati {c{. 'tiiitii'-tcitf to begin to rain from 
'ni'.tia'' TO BE RAINING ; but 'i'wa-tcdi to get 
big), -tciti, see note 35. had' tati is momen- 
taneous; most forms in -r-/<:i// are definitely 

80. See notes 72, 74. Here, as frequently, 
am. . . -qa' mark an objective clause of indirect 

81. ya'kiu-, ya'k'- to be sore, used either 
with body-part suffixes (e. g. ya'k-o'i to be 
sore-faced, ya'liiv-in'i to be sore-necked, 
yaya'k'-niik'" to be sore-handed) or with ina- 
lienable possessive -at' (see note 24) and fol- 
lowing body-part noun (e. g. ya'k'-at-ah qasc' 
mine-is-sore eye, I have a sore on the eye; 
d. yaya'k-sui to be sore-eyed) ; dur. ya'k'-at', 
inceptive ya'k'-ci-'at\ -ci-tt', see notes 18, 5, 
24, 53 ; -at' is treated analogously to -'ati (see 
note i), and they combine into -'atlat' .