General and Amerindian Ethnolinguistics: In Remembrance of Stanley Newman 9783110862799, 9783110118223

Citation preview

General and Amerindian Ethnolinguistics

Contributions to the Sociology of Language

55

Editor Joshua A. Fishman

Mouton de Gruyter Berlin • New York

General and Amerindian Ethnolinguistics In Remembrance of Stanley Newman

edited by

Mary Ritchie Key Henry M. Hoenigswald

Mouton de Gruyter Berlin • New York

1989

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

Library of Congress Cataloging-in-Publication Data General and Amerindian ethnolinguistics : in remembrance of Stanley Newman / edited by Mary Ritchie Key, Henry M. Hoeningswald. p. cm. — (Contributions to the sociology of language : 55) Bibliography: p. Includes index. ISBN 0-89925-519-1 1. Indians — Languages. 2. Linguistics. 3. Newman, Stanley S. (Stanley Stewart), 1905-1984.1. Newman, Stanley S. (Stanley Stewart), 1905. II. Key, Mary Ritchie. III. Hoeningswald, Henry M. 1915. IV. Series. PM102.G46 1989 497—dc20 89-93331 CIP

Deutsche Bibliothek Cataloging-in-Publication Data General and Amerindian ethnolinguistics : in remembrance of Stanley Newman / ed. by Mary Ritchie Key ; Henry M. Hoenigswald. — Berlin ; New York : Mouton de Gruyter, 1989 (Contributions to the sociology of language ; 55) ISBN 3-11-011822-X NE: Key, Mary Ritchie [Hrsg.]; GT

Printed on acid free paper. © Copyright 1989 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 in any form — by photoprint, microfilm or any other means nor transmitted nor translated into a machine language without written permission from the publishers. Typesetting: Arthur Collignon, GmbH, Berlin. — Printing: Ratzlow-Druck, Berlin. — Binding: Luderitz & Bauer, Berlin. — Printed in Germany.

Preface

The life of Stanley Newman has a special place in the history of linguistics in this country, for he was one of the last linguistic scholars who personally knew and worked with three of the "greats" who are credited with the development of linguistics in this century: Franz Boas (1858 — 1942), Edward Sapir (1884-1939), and Leonard Bloomfield (1887-1949). The correspondence in Newman's personal papers speaks to the close and vital relationship with these scholars during his formative professional years. Franz Boas founded the journal International Journal of American Linguistics in 1917 and was the fourth president of the Linguistic Society of America in 1928; Sapir followed shortly afterward as president in 1933, and Bloomfield in 1935. These were the years when Newman was "in training"; and his close relationship to them resulted from their mutual interests in the American Indian languages. This volume reflects the prominence of the Indian languages in the development of linguistics in this country, as well as in Newman's career. Also of interest is the observation that he bridged the gap between the linguistics of the 1920s and 1930s and contemporary linguistics, when he was elected VicePresident of the Linguistic Society of America in 1962. Again in the 1980s he was Vice-President; this time with the additional title of PresidentElect. Stanley Newman was a leading figure in using the comparative method on Indian languages, which Sapir and Bloomfield had pioneered in the Americas. His first comparative work was a study (with Robert Weitlaner, 1950), of the Otomian languages in Mexico. He moved with ease from one language to another and became an authority on other language families as well. He was a loyal friend in all of his relationships, but he could disagree with the linguistic stance of a close friend, even as his correspondence shows throughout his long and respectful friendship with Morris ("Morrie/Morry") Swadesh. In this volume we bring together several items of interest for the history of linguistics in this century and for a portrait of our beloved mentor, colleague, and friend. His history begins in the essay "Prague", a little sketch of his childhood trip to the old country for a short period of schooling. His own autobiographical essay, "How I discovered linguistics," fills out his place among other scholars in the growing discipline. This piece was solicited by Konrad Koerner, who generously agreed to

vi

Preface

our including it in this volume. Koerner is preparing a sequel to First Person Singular (John Benjamins, Amsterdam, 1980), for which this autobiography was written. We have also included here three examples from his collection of poetry, which he wrote when a member of the poetry club, along with Edward Sapir, at the University of Chicago. Among his papers are a few photographs — very few of this modest man. With the help of John B. Carroll, Susan Ervin-Tripp, Herbert Landar, Howard Maclay, Sol Saporta, and Sally Sperling, we have identified other people in the photographs as best we can. We express our appreciation to Professor Newman's wife, Lucy, and to the Maxwell Museum of Anthropology, University of New Mexico, for supplying the pictures. In the future scholars will benefit from the materials left at the Maxwell Museum, and for this reason we have included an inventory of Newman's linguistic collection which Mrs. Newman turned over to the Museum. Franz Boas had left some of his Bella Coola notes with young Stanley Newman; and it is a poignant experience to leaf through the small notebooks and miniature size slips of paper that are over a century old. Scholars did not have a glut of paper at that time! It is especially appropriate that Michael Silverstein's Obituary of Stanley Newman can be included here; Silverstein fell heir in recent times to Sapir's old office, Social Science 222, at the University of Chicago. This piece is a modified version reprinted from Language 63.2:346 — 360 (1987), with permission of the Linguistic Society of America, to which we express our appreciation. We have also reprinted the Obituary written by Newman's close colleagues, Philip Bock and Harry Basehart, at the University of New Mexico, because it supplements and fills out the picture of this man and his place in history. We thank them for their perspective of a colleague with whom they worked very closely, and we thank the American Anthropological Association for permission to reprint this Obituary from American Anthropologist 88.1:151 — 153 (1986). Stanley Newman had prepared a list of his publications during his younger years. Barbara and Bruce Rigsby compiled his bibliography up to the early 1970s, and this was published in a special issue dedicated to him, in the International Journal of American Linguistics 41.4:399—405 (October 1975), edited by Bruce Rigsby. Michael Silverstein recently brought Newman's bibliography up to date. We have used these three sources for the complete bibliography of Stanley Newman included here. We express our appreciation to the many people thus mentioned who have made this collection possible. Finally, we thank Marian Rodee, Curator of Collections of the Maxwell Museum of Anthropology, who

Preface

vii

cleared off her own desk and provided excellent working conditions in order to review the personal linguistic papers archived there. We all have been enriched by Stanley Newman's career, the epitome of a scholar whose life was a model of decency and courage. The authors in this volume consistently expressed their appreciation of his earnestness, warmth, and wit — not to mention his linguistic insights. These offerings are brought together to honor Stanley Newman, who would have been the next president of the Linguistic Society of America, had he lived to hold the position he richly deserved. They were written by students, colleagues, and friends who did not want his name to be lost in history. December 1987

Henry M. Hoenigswald Mary Ritchie Key

Contents

Preface Henry M. Hoenigswald and Mary Ritchie Key List of Contributors

v xiu

A. The History of Linguistics and Stanley Newman's Six Decades Prague Stanley S. Newman How I discovered linguistics Stanley S. Newman Songs for a windy day: poems Stanley S. Newman

31

Bibliography of Stanley S. Newman

33

Inventory of Stanley Newman's linguistic materials Mary Ritchie Key

41

"Singularly like our ideal of a scientist" Michael Silverstein

45

Obituary: Stanley S. Newman (1905-1984) Philip Bock and Harry Basehart

65

Stanley Newman and the Sapir school of linguistics Regna Darnell

71

Sapir's panoramic view (1926) of recent advances in linguistics Yakov Malkiel

89

B. American Indian Studies Institutional language maintenance resources of American Indians in the early 1980s Joshua A. Fishman

107

Spanish loans in Wikchamni Geoffrey Gamble

123

Some principles of Alaskan Athabaskan toponymic knowledge James Kari

129

X

Contents

Deified mind among the Keresan Pueblos Jay Miller Creek curing in academe Willard Walker

151 157

C. Grammar and Discourse

On the unit of paragraph analysis in formal monologue discourse Betty Lou Dubois Object agreement in the Halkomelem Salish passive: a morphological explanation Donna B. Gerdts Argument obviation and switch-reference in Hopi LaVerne Masayesva Jeanne and Ken Hale Some agent hierarchies in Upper Chehalis M. Dale Kinkade Vowel ablaut and its functions in Yuman Margaret Langdon Aspect in Isthmus Zapotec Velma B. Pickett A later view of Gitksan syntax Bruce Rigsby The Kuna verb: a study in the interplay of grammar, discourse, and style Joel Sherzer

175

185 201 213 219 229 245

261

D. Word Formation

Navajo stem variation Herbert Landar Lexical morphemes in Bella Coola Ross Saunders and Philip W. Davis Lexical elaboration in Navajo Robert W. Young

275 289 303

E. Phonology

Loss of contrast between voiced and voiceless alveolar flapped stops in American English J. Donald Bowen f

323

Contents

xi

Some environments which may condition vowel length Eunice V. Pike

331

What is a 'register' language? Richard S. Pittman

337

F. Comparative

Studies

The Proto-Otopamean vowel system and the development of Matlatzinca Doris A. Bartholomew

345

A new look at Aztec-Tanoan Irvine Davis

365

Interpreting the past from the present: a Nahuat example Mary Ritchie Key

381

Prenasalized stops in Proto-Indo-European André Martinet

387

G. Oral Tradition

Tsimshian poetics John A. Dunn

395

Thoth and oral tradition Carleton T. Hodge

407

Some aspects of textual relations in Jawoyn, Northern Australia Francesca Merlan

417

H. Ethnological Studies

"Say 'Hello' to your (second) cousin Claude:" kinship terminology and recursive rules Philip K. Bock

447

Kumix: the Chorti hero John Fought

461

Visualizing the physical context of discourse in languages of the past Saul Levin

469

xii

Contents

Ethnographic notes and observations on the Big Man Complex among the Nacirema J. Anthony Paredes

483

Index

491

List of Contributors

DORIS BARTHOLOMEW

Instituto Lingüístico de Verano HARRY BASEHART

University of New Mexico PHILIP K . BOCK

University of New Mexico J . DONALD B o WEN |

University of California, Los Angeles REGNA DARNELL

University of Alberta IRVINE DAVIS

Summer Institute of Linguistics PHILIP W . DAVIS

Rice University JOHN A . D U N N

University of Oklahoma BETTY L O U DUBOIS

New Mexico State University JOSHUA A . FISHMAN

Ferkauf Graduate School of Psychology, Yeshiva University JOHN F O U G H T

University of Pennsylvania GEOFFREY GAMBLE

Washington State University D O N N A B . GERDTS

State University of New York at Buffalo KENNETH HALE

Massachusetts Institute of Technology CARLETON T . H O D G E

Indiana University

xiv

List of Contributors

HENRY M . HOENIGSWALD

University of Pennsylvania L AVERNE MASAYESVA JEANNE

Massachusetts Institute of Technology JAMES M . KARI

Alaska Native Language Center M A R Y RITCHIE KEY

University of California, Irvine M . D A L E KINKADE

University of British Columbia HERBERT LANDAR

California State University, Los Angeles MARGARET LANGDON

University of California, San Diego SAUL LEVIN

State University of New York at Binghamton YAKOV MALKIEL

University of California, Berkeley A N D R É MARTINET

Université René Descartes FRANCESCA MERLAN

University of Sydney JAY MILLER

The Newberry Library J . ANTHONY PAREDES

Florida State University VELMA B . PICKETT

Instituto Lingüístico de Verano EUNICE V. PIKE

Summer Institute of Linguistics RICHARD S . PITTMAN

Summer Institute of Linguistics

List of Contributors BRUCE RIGSBY

University of Queensland Ross SAUNDERS Simon Fraser University JOEL F . SHERZER

University of Texas at Austin M I C H A E L SILVERSTEIN

University of Chicago WILLARD WALKER

Wesleyan University ROBERT W . Y O U N G

University of New Mexico

XV

A. The History of Linguistics and Stanley Newman's Six Decades

Prague Stanley S. Newman

I When mother brought me here for the first time to look at the place, I was afraid of the long dark stairway. And I did not like Miss Greta Lederer who was to be my governess because she was stiff and foreign looking. Like all other foreigners, she said to mother I kiss the hand. In the extra room that was to be mine the feather bed was so high that it looked as if it were going to float. But on the bed was a cover with little black flowers. They went in straight lines up and down and sideways. And if you ran your fingers crossways they went in straight lines too. It was a pretty bed cover. I wanted to stay. But mother said you will be here for a long enough time. And then she bent over and kissed me. When I turned around there was a lady leaning against the wall near the door. She had a round fat face and her hair hung in wisps. She stood there staring at me. He name was Hedva and she was the maid. Father is sick. And mother is going with him to Karlsbad or Venice, she says. I am going to school here. I am in the zweite Grade. All the boys make fun of me. They say Amerikaner, Amerikaner, Amerikaner. I do not answer them. In the morning I walk around near school until I hear the bell ring. Then I come and form in line with the other boys. We march up the stairs in fours. Der Herr Professor waits on the first landing. He has a large shiny bell in his hand. When we pass him we say guten Morgen Herr Professor. But he is always angry. I have a seat near the window on the third bench. There is a big tree outside the window. The shadows of its branches wave on my desk in the morning. Sometimes I have my pencil box try to catch the shadows. But once I had to kneel on the bamboo rod because I was not paying attention. So I try to listen to Der Herr Professor. But then I watch his hair shiver when he talks and the bubbles come to the corners of his mouth. I wish I could pay attention. Mrs. Lederer is Greta's mother. She is always in the kitchen watching Hedva, the maid, and drinking coffee. She puts a piece of sugar in her mouth and drinks the coffee through it. She drinks so much coffee that her skin is brown and the whites of her eyes are brown. When she looks

4

Stanley S. Newman

at me she rolls her eyes. It is queer that Hermann is not afraid of her. Hermann is her son and he is a hunchback. He comes home in the evening and swears at her and calls her bad names. Then she lifts her apron over her face and cries. Greta says that it is not nice for Hermann to do that. Now she takes me to another part of the house when he comes home. One day as I was coming from school, I heard Mrs. Lederer's voice. It was loud and high. She was scolding someone. So I waited outside the door until she had finished. Hedva was standing in the hallway when I came in. I wanted to pass her quietly and slip into my room. But the hallway was very narrow. I had to squeeze past her. I could see that her eyes were red. Then I stepped on her foot because the hallway was so dark. I said entschuldigen Sie. Greta had told me that I must always say that when I bumped into anybody or stepped on anybody's foot. But Hedva laughed when I said it to her. A few days after, Greta said, why do you not go to play in the back court? The window in my room opened out on the court. I had often thought that it would be a lot of fun if I could play there. But I did not know how to get down. Greta said you must go down the back stairs and through the basement. So I found the back stairs and went into the basement. It was filled with steam. Hedva was washing clothes. She would bend over into the big white balls of steam that were puffed out of the wash tub. Then she would stand away and wipe the sweat from her face with her apron. But that was wet. And so were her slippers. Every time she took a step, water oozed from them. But she seemed to be having a lot of fun washing clothes, dipping in and out of the steam. All of a sudden she turned around and saw me. I raised my cap and said guten Nachmittag, because Greta had told me that I must always greet people who are older than I am. But Hedva did not say guten Nachmittag. She held a wet sheet in her hand and stared at me. You are the Amerikaner? she asked. I said I was. Do you like it in Amerika better as here? I said I guess so. She went back to her wash tub and disappeared into the steam. I sat down on a box to watch her. She had on white stockings like those I always wear on Sunday. Only her stockings had a lot of holes in them. Her clothes looked stuffy because they were thick like blankets. Then she came out of the steam again.

Prague

5

Amerika is very far away, nicht wahr? It is on the other side of the ocean, I said. You take a big boat to cross the ocean. It is very nice on the boat, I said. The steward brings you tea in a glass with lemon in it every afternoon. I am going to Amerika soon if Der Lieber Gott wills it, she said. After I get married. I am going there too, I said. When mother and father come for me. We shall all go together maybe. After she had said that she laughed a lot. I laughed too. Are all the Amerikaners rich? she asked. I said I guess so. It is nice to be rich, she said. When you are poor then people can treat you like a dog. Frau Lederer knows that I did not steal her stockings. But she calls me a thief because she knows that I can not do anything. She went to the tub. So I walked about in the basement. There were a lot of posts. If you put your hand on the other side of a post and run, you keep on going around. After a while you get dizzy. Then you must sit down. The floor rocks as if you were on a boat. And if you close your eyes and smell the steam, you can play that you are on the ocean. Hedva called to me. She said you ought to go up already. They will be looking for you. I went upstairs. But Greta was not looking for me. I was glad that she did not ask me where I had been. The next afternoon I went into the basement again. I found a lot of rods and sticks. So I duelled with the posts until I was out of breath. But Hedva was not there the next day or the next. It was a long time afterwards before I smelled the steam when I came down the back stairs and found Hedva washing the clothes. When I came in this time she called out guten Nachmittag. And how goes the little gentleman today? Very well thank you, I said. Is your mother and your father coming soon to take you back to Amerika? I said yes. When I get married, she said, then I am going to Amerika too. She stirred the water in the tub with a stick. Der Lieber Gott will be kind to me, she said. I am giving money to the church. Suddenly she came over and sat on a box next me. Do you think that Der Lieber Gott is always fair? She looked very angry.

6

Stanley S. Newman

Der Lieber Gott in Amerika, I said, is an old man with a long white beard. He sits on a throne, I said. We have his picture on the wall of our Sunday school in Amerika. I have a sister, she said, who is very good and very religious. Already in three years two of her children have died. Now what do you think of that? Would Der Lieber Gott take away her children if He ... She crossed herself and jumped up and went to her wash tub. But she came back again in a few seconds. She talked to me for a long, long time. I do not remember what she said. And she spoke so fast that I did not understand everything. But she told me that she was going to be married in about a month maybe. If Der Lieber Gott would be kind, He would let her husband take her to Amerika, where they do not treat you like a dog. There they could be rich and live like human beings. And she said that when you are born poor here then people want that you should stay poor. You did not get a chance to live like a human being. She said she would not stay here and be treated like a dog. She would go where she could live like a human being. After she had finished talking, we sat there for a while. It was getting dark in the basement. Hedva lit one of the gas jets. The steam was all melted. The basement was cold. You must tell no one what I have told to you, she said. Maybe it was wrong for me to put such ideas into your head. Her eyebrows were drawn up tight when she said this. And she whispered as some of the boys whispered in class when Der Herr Professor was not watching them. I ran upstairs. I saw Hedva upstairs every day. But of course she did not talk to me there, because it was different from talking in the basement. The spring came and then the summer when mother wrote that she would soon be ready to take me back to Amerika. I was hoping that Hedva would not be going back with us. But she did not say anything more about it. Then one day I did not see her. She had left. At supper Mrs. Lederer talked with Greta about getting a new servant maid. It will be hard to get another one like Hedva, said Greta. Yes I had thought about that, said Mrs. Lederer. But we had to send her away. We could not keep her here. What could we do with the baby when it came? Then Hermann put down his knife and fork and leaned back and laughed out loud.

Prague

7

Who would have thought that she would do anything like that and keep it from us so long. But I like her better for it. You ought to be ashamed of yourself, said Greta. That evening I asked Greta where Hedva was. She said that I would not understand. She said that Hedva had gone away and was not coming back. I asked if Hedva was not ever coming back ever. Greta said that it was as if Hedva had died. She was never coming back. But when grandfather died in Amerika, he left his trunk in the corner of his bedroom. And Hedva's trunk is not under her bed any more. She has taken it along. II Lieutenant Hergel travels a lot with the army. But when he is stationed in Prague, he comes to see Greta. He is tall and straight like a stick. And the way his uniform is pulled tight around his chest makes you think that there is a board underneath. He spent one Sunday with us just before he was going to Berlin. I was writing my penmanship exercises when he came. I could hear his sword clanking and a lot of little chains jingling in the hall. Then he knocked on the door and rushed into my room like a wind. He grabbed me under the arms and threw me up in the air and rubbed his bristly moustache on my cheek. That is a French kiss, he said. And a French kiss is the best of all. Greta had just come in behind him and she laughed. He went to the window and flung the curtains aside. A lot of yellow light poured in. Ein prachtvoller Tag! he cried. It is a sin to stay in the house on a day like this. Come! Have some lunch and then we shall go out walking. So we went into the kitchen and Mrs. Lederer gave us some sandwiches and coffee and kuchen. But she did not eat anything, because she said that she was not feeling well. She sat near the stove and listened to Lieutenant Hergel. He told us stories of the barracks, about how the new men they were training were soft and awkward, and how the food that the army sent them was getting worse. He laughed a lot and his strong white teeth showed. I told him that I bet his army could beat any army in the world. And he said that his army was a bad enough place when there was not a war, without trying to think how it would be if there was a war. After we had finished eating, Greta helped Mrs. Lederer put the dishes in the sink. But Mrs. Lederer said that she would stay home and wash

8

Stanley S. Newman

the dishes, so we went out. The sun glared so hard on the white pavement that I had to keep my eyes closed at first. I held on to Greta's hand and I had to run almost, because Lieutenant Hergel was taking such long steps. We walked to the broad promenade street around the museum. The promenade was crowded with people walking slow. A lot of men wore frock coats and high hats, and almost all the men swung canes, except the soldiers. You could see even far away their brass buttons and their swords that looked like fire when they reflected the blinding rays of the sun. A lot of bright green and yellow parasols bobbed up and down along the avenue. Whenever we passed a soldier with white stars on his collar, Lieutenant Hergel saluted easily and the soldier saluted stiffly. But when we passed a soldier with more silver stars or with gold stars, then Lieutenant Hergel saluted very stiffly. We turned off the promenade and went into the park. It was crowded there too. The benches were all taken by people who sat reading newspapers or were talking and laughing. Zuckerwerke, Zuckerwerk, cried a man with a tray of sticky-looking candy. I wonder who would want to eat candy on a hot day like this, said Greta. This is a day for drinking, said Lieutenant Hergel. I think we shall go to the beergarden and have something cool, nicht wahr? So we walked up a little hill on the side of the park. The path led around and around in a circle up the hill. We could hear the music of a band playing at the top, but it sounded soft and far away. There were benches set along the path. And whenever we passed soldiers sitting there, they would jump up with a jerk and salute. At the top of the hill we went through a gate and came into the broad flat beergarden. It was dotted with small, iron-legged tables covered with white tablecloths. There were a few trees scattered about. And flocks of birds swooped from tree to tree and chirped and sang and made a lot of noise. In the middle was a pavilion draped with the German red, white and black colors. Then a Kellner came and took us past a lot of tables, where people sat drinking and smoking or talking excitedly, to a table next the railing at the edge of the beergarden. Here we could see the sky all around us and it was very blue. There was a cool wind blowing, puffing the tablecloths and chasing leaves across the ground. But the sun was hot. And so you could not lean against the railing, because that was hot too. On a day like this, said Lieutenant Hergel, I want to stay in a beergarden all the time and not think about the army.

Prague

9

Then he laughed and leaned over to Greta and took her hand. But why do you not leave the army? asked Greta. It is already so long. It is not so bad in the army. Lieutenant Hergel looked worried. A few more years, maybe ... Then he laughed again and patted Greta's hand. But today is today. He turned around and motioned to a Kellner standing near us. The Kellner came and took their orders and then he bent down to me. He had a long bony nose and little bird eyes that blinked. And will the little gentleman have something today? A piece of cake or an ice? I said that I did not know. He will have a pineapple ice, said Greta. The Kellner flicked his serviette over his arm and walked away. Look, said Greta, pulling Lieutenant Hergel's sleeve. They are going to play. The Kapellmeister, with gold and silver braid on his uniform, was standing in the pavilion. He was holding his stick up, ready to start. The brass horns looked like gold in the sunlight. They were quiet, waiting. Then the stick came down and the drums and blaring horns began a sharp, jerky tune. The Kapellmeister flapped his elbows in time to the music. That is a song we sing at the barracks, said Lieutenant Hergel. He swung his head from one side to the other with the music and leaned over to Greta and sang the words quietly to her. That is not a very nice song, said Greta. It is a good soldier song, said Lieutenant Hergel. When the band stopped playing, he leaned back and whistled the tune. Then the Kellner came and brought them each a stein of beer and me my pineapple ice. How would you like to taste some Pilsner? said Lieutenant Hergel, holding the stein to my mouth. It will make him sick, said Greta. He must learn to be a man. All men drink Pilsner. You want to be a man, nicht wahr? he asked me. Yes, I said. And a soldier too. So I drank some of the beer. But the foam felt so much like soap and tasted so sour that I did not like it. We sat there for a long time. The band always played a song and then rested for a few minutes. Greta and Lieutenant Hergel talked together and laughed a lot. They did not notice me much. Greta was altogether

10

Stanley S. Newman

different here from the way she was at home. She was not like a governess here. But Lieutenant Hergel was as he always had been. He told more stories and jerked his arms when he talked. But there were getting to be less and less people in the beergarden. The sun was almost touching the line of trees in the park and it was becoming cold. Some Kellners were piling chairs on top of the tables in the corner of the beergarden. We ought to go home already, said Greta. So we left the beergarden. It has been a nice day, said Lieutenant Hergel. Because it was getting dark, we took a tramwagen home. When we got there, all the lights were out. Mrs. Lederer was lying on the couch, because she was not feeling well, and no one else was at home. We sat in the kitchen a while and had some coffee and kuchen. Then Lieutenant Herbel gave me a French kiss and kissed Greta and went out. We watched him from the window. We could see him like a shadow walking swiftly to the corner of the street. He stopped under a lamp and took off his hat and waved it at us. Then he disappeared around the corner. One day during the next winter I was duelling with Ernst in the hallway. When it got too dark for us to duel, we sat down and talked. I am going to get in the army while I am young yet, Ernst said. Because then I can have a better chance of getting gold stars on my collar. But the army is a bad enough place when there is not a war, I said. I bet Lieutenant Hergel told you that. Did you hear what happened to him? What? The typhoid, Ernst said. It took him just like that. And he snapped his fingers. But don't tell Greta that I told you. And don't say anything about it to her. We sat there in the dark for a long while without saying anything. Then Ernst said, The typhoid is pretty bad, I guess. It took him just like that. He snapped his fingers again and nodded his head.

How I discovered linguistics Stanley S. Newman

I was born in Chicago on July 18, 1905. Early in life I was exposed to foreign languages as ways in which people talked. English was the allpurpose language of the home. But my parents, who had come to this country in the 1880s from the Czech-speaking area of what was then Austria-Hungary, used both Czech and German as well as English. Like many middle-class families at the time, we had a live-in cook, a recent immigrant, a Czech girl who spoke no English. By the time I was three, I was able to speak the language on a child's level and understand the folk songs she sang and the stories she told me. Czech became the warm, intimate language of the kitchen. Neither my older sister nor I spoke German, but we were motivated to learn; our parents, we came to realize, used German when they talked about something they didn't want us to understand. When I was four or five, my parents took my sister and me to central Europe and, while they visited relatives, left us with some of my father's people on a farm near Prague; there, for several weeks, we spoke only Czech. In 1913 the family went to Europe again, this time for more than a year. My parents arranged for us children to live with a family in Prague, where we could go to school. We spoke Czech in the home, but German was the official language of the country and of the schools. After some sixty years I have retained vivid memories and images of that emotionally charged period of my life. In school the sexes were segregated, and discipline was strict in the class of boys where I was placed. Even before I understood much German, I learned that there were three levels of punishment. The gray-haired, bearded man who taught the class — we addressed him as Herr Professor — spent much of the day walking up and down the aisles between the benches where we sat, holding a bamboo switch in his hand; he kept a supply of these in a small closet. For such minor indiscretions as whispering or staring out the window when we were assigned to work on reading or writing, we might get a flick on the back or head with the switch. For more serious infractions, the offender was marched to the front of the room, where the Herr Professor administered a predetermined number of swats on the upturned palms, the number depending on the level of misde-

12

Stanley S. Newman

meanor. For the most reprehensible categories of classroom crimes, the culprit was made to kneel on a bamboo switch placed on a step of the dais where the Herr Professor's desk stood. I learned about all three types of punishment from first-hand experience. I also learned what it meant to be a foreigner. During the first few weeks of school, some of the boys in my class would gather outside the school building in the mornings to greet me with jeering rhymes in Czech. One of them went, Amerika, Afrika, tata bije Pepika ('Amerika, Africa, pa is giving Pete a licking'). The rhyme didn't make much sense, but its intent was unmistakable. From listening in on adult conversations, I gradually learned that Americans were unpredictable and ill-mannered; American children, in particular, were wild and undisciplined. I heard myself referred to as a 'Red Indian' in both German and Czech. About a month after the beginning of the school year, the young woman who was in charge of my sister and me came to my classroom on a parents' visiting day. The Herr Professor told her, while I was standing by his desk, that I was bright enough, 'aber er hat kein Sitzfleisch'. I took the remark literally to mean that I lacked sitting meat and understood it as another affront, this time a scornful reference to my behind. Only later did I discover that this was a common expression, a strange way of saying that a person was unable to sit still. As I look back on it now, I realize that I had to adjust to a new identity. In America I had been a relatively quiet, well-behaved boy, a youngster who tried to avoid the fights that always broke out during recess periods. But here I was treated as a boisterous, untamed foreign kid in a school where recess consisted of the class marching briskly, in two-by-two formation, around the small courtyard adjoining the school building, with the Herr Professor leading the company. When the school year ended, the family was preparing to leave for home in the summer of 1914, just as Austria-Hungary declared war on Serbia. In Hamburg we had booked passage on a small British ship bound for England. But, before we could leave, England declared war on Germany, which had joined Austria-Hungary, and the ship was held in harbor. A palpable nervous apprehension hung over Hamburg. People gathered in knots on the sidewalks. Police seemed to be everywhere. Soldiers marched through the streets, and sometimes at night we could hear the sound of their marching boots on the pavement. In all this excitement my father was arrested as a British spy. He had been heard talking English to an American friend while walking along the street, and some nervous citizens had alerted the police. It took several days to clear

How I discovered linguistics

13

papers for his release. After that we spoke nothing but German until we left for Rotterdam, where we eventually obtained passage for New York. Back in Chicago I resumed my American habits, but some unforeseen language problems had to be overcome. Having spoken very little English for a year, I had acquired some foreign speech habits and had forgotten how to pronounce some English words. The daily classroom recitations of reading aloud became ordeals. When I heard the class snicker or giggle at my funny pronunciations, I felt embarrassed and stupid. Mercifully, this period of speaking like a foreigner in my own country lasted only a few months. In highschool I selected Latin as my foreign language and continued with it for four years. I enjoyed the language, especially the translation assignments. After the family moved from Chicago to Los Angeles, I enrolled in the University of California branch, located within walking distance of our home. At that time, in 1923, UCLA was a small college, primarily a teacher training institution, within the city. My highschool interests in language and literature carried over into college. In English I found writing particularly stimulating, and I started a modern foreign language — French. The instructor, in contrast to my easy-going Latin teacher in highschool, was business-like and demanding. By the end of the year I had some notion, not always clearly defined, of the strange pronunciation, and I was able to read with some pleasure. By the end of the second semester, when I talked over future plans with my English and French teachers, both advised me to continue my education at another university, since I was not planning on a career in public-school teaching. I had heard that the university on Chicago's South Side, near my former home, was a pretty good school, so I moved back to Chicago. The University of Chicago offered a variety of new and exciting challenges. The English department gave a minor sequence in American literature, a rare specialty in those days. The courses, taught by Percy Boynton and Napier Wilt, attracted me more than the standard period courses. I expanded my interest in modern languages, continuing with French, taking some courses in Italian and Russian, and trying to keep German alive by reading novels. One of my extracurricular activities was writing reviews for the Literary Section of the Chicago Daily News. Another was the Poetry Club, which I joined after recovering from the surprise of winning $ 50 in a poetry contest sponsored by the University. The Club was a lively and vigorous group. At our biweekly meetings, members handed in the poems they had written, with no names attached, and the main duty of the Club

14

Stanley S. Newman

president was to read the poems aloud and keep the discussion from getting too heated. The Club published The Forge, a poetry journal, which was later enlarged to become a literary magazine; my first editorial experience came when Sterling North and I edited The Forge for about two years. This post-war period of the 1920s was a time of artistic and intellectual ferment, a time when a host of new writers burst on the scene. It was at the Poetry Club that I first met Edward Sapir. He was a member of the anthropology faculty, who did language research, according to our more knowledgeable members. We were flattered but puzzled by his appearance at our meetings. Although nearly all of our members were majors in English, none of the English faculty ever came to our meetings. From our perspective, furthermore, anthropologists were people who dug in the ground for buried treasure or studied apes, and language specialists were preoccupied with subjects like Old English and Beowulf, which seemed remote, even antagonistic, to an interest in writing poetry. We could recognize the poems that Sapir submitted, a polished classical type, more restrained than the frankly lyrical verse that most of us wrote. We were particularly impressed by his incisive comments on the poems read aloud and by his broad acquaintance with modern poetry; I doubt that any of us had heard of Gerard Manley Hopkins until Sapir mentioned his name and quoted a few striking lines of a Hopkins poem. We were impressed, too, by the elegant fluency of his speech. His spontaneous choice of words had the characteristics of carefully crafted prose. My curiosity about Sapir led me to take one of his courses in 1927, a wide-ranging course on the psychology of culture. It introduced me to a world of fresh ideas and new writers — Jespersen, Vendryes, Croce, Boas, Jung, Spengler, Veblen. But this was my senior year, and I had to complete my requirements for the bachelor's degree and to continue with graduate work in English. After receiving my master's degree, I accepted a job as instructor at the University of Texas in Austin. This turned out to be a searing experience. I came to realize that my training had not prepared me to teach composition and English literature. One of the two courses I had taken in the School of Education at Chicago dealt with the teaching of English in highschool. Its main message was that students must be taught 'good literature', which could be identified by ten characteristics, two of which, for example, were that it must 'reflect good morals' and 'be true to life'. That ended my search in the School of Education. Discussions among the twenty or so young fellows who came from various parts of the country to fill the English instructorships at Austin revealed that none

How I discovered linguistics

15

of them had any clearer notion than I on how to teach the courses assigned to us. The freshman and sophomore courses, I learned, were low prestige assignments, unavoidable burdens that the department had to bear, and the less said about them the better. After this humbling year of teaching, I returned to the University of Chicago with mixed feelings about continuing in English. I decided to try another course from Sapir, one in phonetics, that I could fit into my English schedule. The novel insights about language that I received and the vitality of his presentation persuaded me to change my major from English to anthropology. I recall one lively session in which a student challenged Sapir's statement that the t sounds in stop and top differed phonetically. Sapir, in response, smiled broadly and his eyes brightened as he came forward to face the student. There he pronounced the pair of words slowly and had the student repeat them. The student still insisted that there was no difference he could hear. The class, mainly composed of undergraduates, was delighted at the confrontation, and Sapir was equally delighted at the chance to elaborate the point. By pronouncing stop with an aspirated t and top with the unaspirated variety, then by placing them in sentences, he finally managed to convince the student of the difference. During the second quarter, I enrolled in one of Sapir's seminars. Here the major task of each student was to read a grammar, give an oral report on it, and write a précis of no more than three double-spaced pages, whose purpose was to present the dominant patterns, the forest rather than the trees. I chose Kroeber's Yokuts, which led to my fieldwork the following summer. The shift of majors brought with it a change in orientation. I lost contact with my former teachers and fellow-students and, rather abruptly, moved in a new circle. Because of my late start in anthropology, my program became loaded with required courses — archeology from FayCooper Cole, who took his classes on week-end digs; physical anthropology from Bill Krogman; ethnology from Robert Redfield; and statistics from Leon L. Thurstone, a recent acquisition in psychology. Among my new fellow-students were Morris Swadesh, Mary Haas, and Walter Dyk. I often met Bloomfield in Sapir's company; they were close friends as well as colleagues. In the late 1920s, it must be recalled, no departments of linguistics existed. The term 'linguistics' was rarely heard outside linguistic circles, and to most people 'linguist' meant a person who spoke many languages.

16

Stanley S. Newman

It became the task of linguists of that period to define their field of study and make it a more widely recognized area of research and applied work. In the summer of 1930 I had my first field trip to study the Yokuts language of California. Walter Dyk, who was to work with the Chinookan Wishram, joined me in buying an old Ford for $ 25, which we drove over the dirt and gravel roads of that era from Chicago to Berkeley, where Dyk had received his undergraduate training. Through him I met several of the Berkeley linguists — Paul Radin, Jaime de Angulo and his wife Nancy Freeland, and Alfred Kroeber, who generously let me use his Yokuts field notes. From Berkeley I went to Bakersfield to see the Commissioner of Indian Affairs for Southern California to obtain information on the location of Yokuts speakers. In my naivete I was surprised to find that the Commissioner had no acquaintance at all with anthropological literature. He had never heard of the Yokuts; all the Indians in his part of California, he told me, were 'Digger Indians'. I located the Yokuts from a map in one of Kroeber's monographs, showing the Tule River Reservation in the former territory of the Yokuts tribes. Most of the Indians on the Reservation, where I lived during the summer, were speakers of Yawelmani Yokuts, and this became the language of my dissertation. My first day of fieldwork was a memorable one. My informant and I were working at a table set up under a tree, and about fifteen men were gathered around to see what sort of strange business I was conducting. At one point, when I repeated the Yawelmani word for 'duck', there was a loud burst of laughter from the crowd. Sensing a minimal pair, I asked my informant what I had said, but he dismissed the question and looked so uncomfortable that I thought it best to let it go. A week or so later, when we were working alone, he told me that I had said a 'bad word', the word for a venereal disease. I had indeed stumbled on a minimal pair: instead of the long vowel in the first syllable of 'duck', I had pronounced a short vowel. The story of my blunder must have spread immediately over the entire Reservation; for several days afterward I was greeted with wide smiles by every Yokuts I met along the road. During the first few weeks of fieldwork, I became increasingly discouraged. I was unable to find any explanatory basis for the vowel changes that were the most conspicuous part of the grammar. The problems seemed endless and insurmountable. Added to the linguistic bafflements were the uncertainties of informant work and the niggling problems of day-to-day living in the California valley, where the temperature often reached 120°. From hindsight I realize that I was experiencing an attack

How I discovered linguistics

17

of anomie on this, my first field trip. Sapir apparently diagnosed my ailment from my letters. Besides sending me helpful suggestions, he contacted the girl who was to be my wife, as I learned later from her, and suggested that she try to buoy up my spirits in her letters. Eventually, of course, the symptoms subsided. The difficulties were endured or were finally resolved when I experienced the excitement and joy of discovery in solving the ablaut problem. The next year (1930 — 31) at the University of Chicago brought eventful changes. I married Lucy Price-Benedict, who was doing graduate work in English; we have shared our triumphs and disappointments for more than half a century. Later that year, Sapir called me to his office to tell me that he was accepting a position at Yale as chairman of a new Department of Anthropology. Yale had offered to award fellowships to some of the advanced students that he would bring with him, and I would be among them if I wished to make the change. I was delighted, of course, to join the others. During that year at Chicago, Sapir suggested that I continue the investigation that he had begun in "A Study in Phonetic Symbolism", published the previous year. I had been taking statistics courses from Thurstone, who had recently devised an experimental technique that could be adapted to the phonetic symbolism problem. I discussed the possibility with Sapir. I sensed that he seemed unenthusiastic, but he made no negative comments. Not being aware at the time of his antipathy toward stimulus-response psychology and statistical techniques, I went ahead with the experiment, using pairs of spoken nonsense words as stimuli and groups of student subjects whose responses were made on the basis of a large-versus-small symbolism. Thurstone suggested that a visual presentation of the words, using a tachistoscope, would provide more consistent stimuli. When I explained that I wanted the subjects to respond to speech sounds, not letters, he looked dubious. But he helped me with statistical advice and allowed me to use the statistics lab, where I pounded electrical calculators for several months in analyzing the raw data. I felt rewarded when my article describing the experiment was accepted by the American Journal of Psychology. The move to New Haven was made in 1931, and within the first year I was able to complete my doctorate. In that year, too, my first meeting with Franz Boas took place. During the course of a linguistics meeting, I remember entering a reception room with Sapir, who pointed him out to me. I had seen photographs of Boas, but in the flesh he appeared more formidable. He was a small, rather frail man with a large head, a

18

Stanley S. Newman

shock of white hair swept back, a long scar across his cheek, and piercing eyes. Sapir took me over to him and introduced me, remarking that I had worked with the Yokuts the past summer. As we shook hands, he fixed his eyes on me and said, "Well, Newman, did you get the whole language?" I probably stood for a moment with my mouth open at the unexpected remark. I understood the words but didn't get the message. Then I heard Sapir laughing, and a light dawned: Sapir had recently published a paper arguing for the 'holistic' approach, and Boas was indulging in a bit of kidding. After I recovered, I made a reply in kind, mentioning that I had an eighteen-inch-long storage box stuffed with four-by-six slips of notes, and there certainly couldn't be any more to the language than that. The incident gave me a preliminary insight into the relationship between the two men, which had some of the complexities and ambivalences of a father-son relationship. From later talks I had with Boas, I know that he regarded Sapir as his most gifted linguistics student but one who was apt to let his imagination carry him off into the realms of fancy. Sapir considered Boas an uncompromising empiricist with little insight, but he always behaved respectfully in the presence of his teacher and, in spite of his love of verbal play, never traded jibes with him. In private, however, he often referred to Boas as 'the Old Man'. He once showed me a letter he had just received, one of Boas's notoriously terse letters, which he wrote as chairman of the Committee on Research in Native American Languages. The letter merely listed for Sapir, as a member of the Committee, the names of those awarded Committee grants. This was apparently the first news Sapir had of any Committee action, for Boas did not believe in wasting precious time and money on such formalities as committee meetings and consultations. When I handed the letter back, Sapir, who was fond of puns, remarked, "This is another of the Old Man's Caesarian operations". During the first year at Yale I heard rumors that Sapir had applied for membership in the Faculty Club and had his application rejected, apparently because some members felt uncomfortable at the prospect of associating with a Jew. Not until many years later was I able to verify the painful incident and to learn of Sapir's unhappiness at Yale. At the time, the rumors assumed credibility when Sapir told me one day as we were walking on campus that he had been appointed master of one of the Yale colleges. Uncharacteristically, he took me to admire the handsome study assigned to him in the college dormitory. Such official gestures

How I discovered linguistics

19

would not ordinarily have impressed him and might have occasioned a deprecating remark. When I received the doctorate in 1932, the country was sinking rapidly toward the bottom of the Great Depression. There were no teaching jobs for linguists with degrees in anthropology. As an academic discipline, anthropology had begun to develop during the 1920s, following World War I. In a few major universities, it had emerged as a distinct department, in others as a minor specialty within the established Department of Sociology (or Department of Sociology and Anthropology). It was represented primarily, however, by archeology and ethnology; linguistics and physical anthropology, the 'weak sisters', could be included in only the large departments. The expansion of anthropology, however, came to a halt as the Depression spread like an epidemic throughout the economy. In the absence of teaching jobs, I remained at Yale for the next five years, subsisting on research fellowships. Morris Swadesh and I received a two-year grant from the American Council of Learned Societies to work, under Sapir's direction, on an English grammar project. I continued with the phonetic symbolism experiments, using pupils in the New Haven public schools as subjects. At Yale's Institute of Human Relations my office adjoined the observational mental hospital staffed by members of the Department of Psychiatry. I became acquainted with some of the psychiatrists, one of whom collaborated with me on a study of the speech of patients with affective disorders, which we published in the American Journal of Psychiatry. During these years I revised my dissertation on Yawelmani Yokuts, expanding it to include the additional Yokuts languages on which I had done fieldwork. The Chicago linguists were joined by new linguistic students, and I met some of them in Sapir's seminars, one on Navajo, another on comparative Athabaskan, and a memorable one in which the participants — Mary Haas, Fang Kuei Li, Morris Swadesh, Carl Voegelin, Ben Whorf, and others — gave papers on languages they had studied in the field. Jerzy Kurylowicz was one of the foreign students who had received fellowships for the seminar on 'Personality and Culture', organized by Sapir and John Dollard. We had a number of lively bull sessions with Kurylowicz, who impressed us with his keenly critical mind and his wide knowledge of current linguistics, especially European linguistics. We told him about our research projects, but he was strangely reticent about his own, in spite of our attempts to pump him. Swadesh and I mentioned this puzzling behavior to Sapir, who told us that questioning Kurylowicz about his on-going research would be considered bad form among European

20

Stanley S. Newman

scholars. Most European journals paid for articles accepted for publication, and authors, for fear of plagiarism, did not discuss their work until after their articles had been published. We learned about subcultures in the world of scholarship. By the mid 1930s the Depression had made research funds increasingly scarce. Sapir told me about the possibility of a project on the speech of mentally ill patients, organized by a friend of his, the psychiatrist Harry Stack Sullivan. I prepared a project plan, sent it to Sullivan, and went to see him at his home in New York City. He was a man of extraordinary charm and quick responsiveness. We talked a good deal about the problems of recording the speech of mental patients, with which I had had experience at Yale and in some state hospitals of Connecticut. He had considerable knowledge of the technicalities of recording devices and the acoustics of speech. He told me he liked the plan I had outlined and hoped it would be possible to raise funds for it. But I heard no more about the matter from either Sullivan or Sapir. One of the pleasant diversions from academic affairs was offered by the informal dinners given by the Sapirs for some of his students and their spouses. Often, as we approached the house, we could hear Sapir playing the piano. There was always much lively talk, and usually there was music. On one of these evenings, Ben Whorf and his brother Richard sang Irish songs to Sapir's accompaniment. As the Depression deepened, fellowships began to run out at Yale. In 1937 I left New Haven to join the research team of the Adolescent Study in New York City, a project funded by the General Education Board. Most of the people on the team were guidance workers or psychoanalysts; I was the only linguist. My job was to study the language behavior of highschool students and interpret the results for teachers and school administrators. The project ended after two years, and I was jobless for a few months. During this dry period Boas was able to find enough funds to pay me $ 50 a month to analyze my Bella Coola field data, collected in 1932, and write summaries of the phonology and some of the morphology. As I was to discover, Bella Coola was very close to Boas's heart because it was among the first American Indian languages on which he had done fieldwork. His secretary often phoned me to make appointments, and sometimes he sent me one of his characteristically laconic letters, one of which read, 'Dear Newman: What about the Bella Coola? yours sincerely, Franz Boas'. He was a stern taskmaster, but a kindly one. After Sapir's death in 1939, Boas asked to see the manuscript of my revised Yokuts

How I discovered

linguistics

21

grammar. He said he would try to find funds to publish it, but he was unable to raise the money. During the closing years of the 1930s, the ominous noises of the cold war threatened open conflict in Europe. Mortimer Graves, Executive Secretary of the American Council of Learned Societies, organized plans for research on little-known languages which were spoken by large populations in countries important in world affairs. If war came, the United States would need reliable data about such languages. I was assigned by the Intensive Language Program, set up under the ACLS, to work on Persian with the aim of preparing a series of lessons for teaching the spoken language. I began searching for Persian speakers by visiting oriental rug stores, only to learn that the Persian rug business was run by Armenians. Persians, I finally discovered, were to be found in importexport companies dealing in Near East products. I soon located a native speaker, Kazem Kaihani, a highly sophisticated young man who was taking some graduate courses in chemistry at Columbia. He quickly grasped the purpose of our linguistic sessions and became invaluable in helping to explore the grammatical and semantic structure of the language. After we had prepared a few lessons, we were notified that a group of army officers would be stationed in New York until they received travel orders for Teheran. Arrangements would be made with Columbia University for us to instruct them in modern spoken Persian. Among the formal arrangements, I had to obtain the approval of a well-known Indo-Iranian scholar on the faculty. He questioned me about my training and research. He was genuinely shocked to learn that I had no formal training in Sanskrit or Old Iranian and that I had not even worked with any of the classical Persian texts. He told me frankly that he could not understand how I would learn enough to teach the modern language merely by interviewing a native speaker, since I had no background at all in the history of the language. He said, however, that he would raise no objection because I had been highly recommended by Edward Sapir. When the officers arrived, we found they had no notion when their travel orders might come. Kaihani and I set up daily morning and afternoon meetings. There were only ten officers, an ideal size for an intensive course, and they were highly motivated. Because Persian grammar is relatively uncomplicated, we were able to make rapid progress, using the format of the Army Method. By the time their travel orders arrived, after about two months, they were able to speak the language

22

Stanley S. Newman

with some fluency within the limits, of course, of a rather restricted vocabulary and syntax. For a year or so after the United States entered the war, Kaihani and I taught beginning spoken Persian again and an advanced course at Columbia to local students. During this period, however, most of my effort went into teaching anthropology as a substitute in the Department of Sociology at Brooklyn College. I found the job stimulating, in spite of the fact that I taught fifteen hours a week: the students were young and eager, and I was able to reorient myself in general anthropology, for I taught ethnological and survey courses, and none in linguistics. In 1943 I joined the Language Section of the Army Service Forces at 165 Broadway. At that time the Language Section was headed by an army major whose qualifications consisted of his having managed a shoe factory he owned in civilian life. His lack of understanding of what was going on in the Language Section and his inept handling of both professional and nonprofessional personnel resulted in considerable grumbling and several confrontations. A few months later he unobtrusively disappeared, and Major Henry Lee (Haxie) Smith was put in charge of the office. Since he was one of us — a linguist, with a doctorate in English — his arrival provided a much-needed boost to our morale. He had a deep conviction that linguistics had an important role to play in the war, and his enthusiasm was communicated to the staff. He was a skillful administrator in dealing with office affairs and with personnel. In outside contacts he had an ability, rare among Ph. D.s, to talk to the military brass in their own language and convince them of the practical value of linguistics. For a few weeks I continued to work with Kaihani, who was now employed by the Language Section as part of a Persian team that was compiling a military dictionary. But I soon received orders to discontinue the Persian project and participate instead in more pressing tasks. For several months I worked with a team that prepared small, pocket-size Language Guides, designed for use by troops on their way to foreign countries. These booklets of fifty to a hundred pages had to be produced at a hectic pace. We usually had ten days to two weeks to locate native speakers, organize the material, and complete a typed manuscript. The finished product, published by the War Department as a TM (technical manual), contained cartoons to liven up the text and was accompanied by phonograph records on which a native speaker, serving as a model for imitation, pronounced the words and phrases in the manual.

How I discovered linguistics

23

After working on a number of Language Guides and one of the longer Phrase Books in Persian, I was assigned to take charge of the 'Reverse English' courses. Like the Basic Courses for teaching foreign languages to English speakers, the Reverse English courses followed the format of the Army Method but their purpose was to teach English to speakers of other languages. The motivation for these courses first arose when groups of Latin American mechanics were brought to the United States for technical training in war-related projects and for instruction in English. Three of us — a bilingual typist, a bilingual editor, and I — prepared a course of thirty units, titled El Inglés Hablado, for teaching English to these Latin Americans. It was a substantial undertaking, which came to over seven hundred pages in print. We next were ordered to organize a similar course for Italian prisoners of war who were being held in camps within the United States. A group of us visited one of the camps located on Staten Island. There an Italian officer, who had studied English in Italy but was unable to speak it, lectured on the English language to some twenty noncommissioned prisoners. We were told by the American officer in charge of the camp that many of the prisoners were eager to learn English and spent part of their small allowance on bilingual pocket dictionaries and self-teaching English booklets. Some had relatives in this country and hoped to remain here after the war. We formed another three-person team to produce L'Inglese Parlato for the Italians. After this was finished, another thirty-unit course, Englisch wie man's spricht, was prepared for German prisoners of war. As the war began to wind down in late 1944 and early 1945, the pace of work at 165 Broadway slackened, and some of the projects were brought to a close. I was transferred to the Institute of Social Anthropology, one of the organizations under the Smithsonian Institution. I was instructed to report to Washington, D. C. for inoculations, passports, and other papers, which would take about two weeks, before moving to the Mexico City station of the Institute. My job would be to conduct research projects and teach linguistics to Mexican students at the Escuela Nacional de Antropología e Historia. When I came to Washington with my wife and nine-year-old son in mid-1945, just after the war had ended, the city was in a state of bustle and confusion, with some people moving out and too many moving in as old projects ended and new ones began. It was difficult to find a place to live, but we finally located a tiny and none-too-clean apartment at an exorbitant rental. While waiting for my papers, I helped Julian Steward, who headed the Institute of Social Anthropology, in editing the Handbook

24

Stanley S. Newman

of South American Indians. The projected two-week holdover in Washington stretched out to four months as the wheels of the federal bureaucracy turned at their ponderous pace. The delays gave me additional time to prepare for the Escuela Nacional, where classes were conducted in Spanish. I had never studied the language, but my highschool Latin supplemented by university courses in French and Italian gave me a background of Romance structure and vocabulary. While I was still in New York, I had obtained a copy of the Basic Course for Spanish with the accompanying set of records, and this material gave my family and me an opportunity to practice the spoken language. Finally, however, all the paper work and other arrangements for leaving were completed, and we departed in a weatherbeaten station wagon that had to be towed from a government car lot and repaired in a government garage. In Mexico City the academic pace, I discovered, was much more casual than in the United States. Frequent holidays punctuated the academic year, and it was seldom clear in advance which of the federal holidays and saints' days were to be observed as school holidays. I found it best, when in doubt, to phone the Escuela office to see whether students were showing up for classes that day. Besides this uncertainty, many of my students were very irregular in attending class. I learned that a large proportion of the student population in Mexico lived in pinching poverty, and many suffered from the all too prevalent illnesses. Most of the Mexican teachers at the Escuela were hired to teach only one or two courses and had jobs at other college-level institutions in the city because their salary at any one was too meager to support an individual, much less a family. After each class hour I would often see a number of the faculty members rush out to catch a bus for their next post. A few of the teachers, however, belonged to the upper-class Mexican gentry and had private incomes. They held lucrative positions elsewhere and taught at the Escuela as an avocation. One of these, for example, was the vice president of a major Mexico City bank. Those of us from the Institute of Social Anthropology were envied by most of the Escuela faculty because our salaries were lordly in comparison to theirs, and we received our checks regularly. The salary checks of the Mexican government were periodically delayed, often by a month or two, particularly at the beginning of a fiscal year. Some of the teachers and service staff would have faced severe hardships had not Dr. de Borbolla, the dedicated director of the Escuela and the Museo Nacional, tided them over the starvation period with his personal funds.

How I discovered linguistics

25

Most of the students were older and more mature than college students in the United States. Perhaps, too, because there were no social events or athletics to serve as distractions, their efforts were focused on classwork. Several of my students became professional linguists and made useful contributions to Mexican Indian linguistics. Although I did some fieldwork on Nahuatl, my main research was directed at Otomi and comparative Otomian. In this I had the congenial collaboration of Robert Weitlaner, who had learned to speak Otomi. At a banquet I attended, he was introduced by Alfonso Caso, the dean of Mexican anthropology, as the only Otomi Indian who was born in Vienna. Weitlaner, who had been trained as an engineer, had come to Mexico as a young man. He was one of the growing stream of Austrians and Germans who had migrated to Mexico City in the 1920s and 1930s and had settled in a section that came to be known as the Fourth Reich. He had fallen in love with his adopted country and had turned to anthropology as a way of becoming more intimately acquainted with the land and its people. He was the kind of anthropologist, like John R Harrington, who had a passion for fieldwork. He seemed to have traveled over every inch of Mexico, often by foot, and had Mexican and Indian friends everywhere. In another venture I collaborated with Robert Barlow in preparing a set of lessons, using the Army Method, for teaching Nahuatl to Mexican students. Although born in the United States, Barlow, like Weitlaner, had lived in Mexico most of his life and had adopted the country as his own. He had become a nahuatlista, a specialist in Nahuatl history and culture, who spoke the language fluently. Those of us who were in the Institute of Social Anthropology kept abreast of changes in the political climate of the United States, for the Institute was established in response to the 'good neighbor' policy initiated by Roosevelt. With his death and with new concerns arising in the postwar period, the New Deal shifted its emphasis in Latin America from collaborative activities in the sciences to joint ventures in the arts, and even these began to fade. The future of the Institute appeared precarious. George Foster, who headed the Institute team in Mexico City, was called to Washington to take charge of Institute affairs there. One of the other members of our group left to take another position. When I received an offer of a post in anthropology at the University of New Mexico, I decided to accept. I began teaching there in 1949, after more than three years in Mexico.

26

Stanley S. Newman

At New Mexico I had a more varied schedule of teaching, with courses in introductory anthropology, Latin American ethnology, social structure, and primitive literature as well as a series of courses in linguistics. During the 1950s the Department of Anthropology was small, consisting of six or seven faculty members. The university had only about 5,000 students and grew slowly during the decade. But in the 1960s a flood of students transformed the university in New Mexico as it did elsewhere. The department and the university expanded to about four times its earlier size. During our years in Albuquerque, my wife taught in the public schools. At first she was assigned to a rural school some fifteen miles outside the city, where most of the children came from Spanish-speaking homes. Some of the youngsters spoke little or no English, and my wife had the challenge of improvising techniques that would lure them to participate in English-language activities and even in some reading. Her ability to speak Spanish, though it was Mexican Spanish rather than the local variety, helped the children over their initial difficulties. After several years in the rural school, she was transferred to a city highschool. In her Latin classes she introduced a version of the Army Method with practice in spoken Latin and the use of tapes. For my research I wanted to work on some of the pueblo languages. But the pueblos were experiencing internal and external conflicts at this time, and it was becoming difficult to conduct field studies within the native villages. A large number of Indian students, however, were attending the university, and I was able to work with several Laguna Keres and, later, Zuni students, who made excellent linguistic informants. Although there were no other linguists at the university when I began, students from the School of Education and from the Departments of Spanish and Psychology were encouraged to enroll in my classes. I was invited to join doctoral committees in these departments when the candidate's dissertation involved linguistic problems. It was gratifying to find these evidences of interdepartmental cooperation, though some of the language-and-literature departments maintained their traditional hostility to linguistics. One of my extracurricular duties was to act as adviser to the Indian students at the university. All of them spoke English fluently, but most had experience only with colloquial English and were unable to follow class lectures, where the instructor used six-syllable words and elaborate sentences. The required English courses often raised impossible hurdles. The frustration was poignantly expressed by a Navajo student who told

How I discovered linguistics

27

me, "Gee, Doc, that English! That's for the birds." For two years I played the same role for a group of about twenty-five students from the Ryukyu Islands. I expected these Oriental youngsters to encounter severe problems in adapting to Western practices. But most of them came from the urbanized towns of their country and had been selected for scholarships to the United States because they had attended an American University in Naha, the capital of the Ryukyus. Their experience in urban and university life made it easier for them than for the Indian students to adjust to their new environment. In activities outside the university, I taught in the Linguistic Institute at the University of Michigan in 1950 and at the University of Chicago in 1954. In the summer of 1951 I participated in an eight-week workshop, organized by John B. Carroll. It was held at Cornell and consisted of only three psychologists and three linguists. We met daily to talk to and at each other, presenting what we regarded as the basic concepts of our discipline, attempting to clarify the possible connections between the two fields and exploring some promising research applications. This was perhaps the most intensive learning experience I ever had. Another interdisciplinary venture, also spearheaded by Carroll, was housed at the University of New Mexico in the mid 1950s. During the first summer a group of about fifteen young psychologists and linguists met with a senior staff to survey the literature on the Indian cultures and languages of the Southwest and to outline empirical investigations for the following year. During the next summer the group split up into small work parties, and I took charge of five of the young participants to conduct research projects through fieldwork among the Zuni. The decade of the 1960s was memorable at the university not only for the swelling tide of students but also for a new aggressive generation of noisy protest and confrontation. We had our share of sit-down strikes, protest marches, attacks on the prexy's house, bombs set in the ROTC building, and general pandemonium. Many of the students and some of the younger faculty assumed the symbols of a proletarian revolution, the men wearing tattered clothes and appearing with long unkempt hair and beards, the women dressed in faded bluejeans or African, East Indian, or other exotic get-ups, and both sexes often going barefoot during the warm weather. As the classes grew larger, it became difficult to maintain a semblance of order. Students would bring their dogs to class and women their babies or small children, with the result that the class hour would be punctuated by howling kids and barking dogs. Mercifully, the mood of confrontation and conflict began to dissipate around 1970. There was

28

Stanley S. Newman

Figure 1. Members of the Southwestern Project in Comparative Psycholinguistics, University of New Mexico (Albuquerque), Summer, 1955. Front row, left to right: Mary Woodward f, Sally Sperling, Stanley Newman f , Edward Dozierf, Sol Saporta. Second row, left to right: James J. Jenkins (not a project member), Herbert Landar, Lois Lawrence (Elliott), Arnold Horowitz, Donald Sola. Third row, left to right: James Knowlton, Eric Lenneberg Susan Ervin (ErvinTripp), Howard Maclay, George Suci. Back row, left to right: Jarvis Bastian, Charles Solley, John B. Carroll (Project Director). Photograph from Maxwell Museum of Anthropology.

How I discovered linguistics

29

hope that, once again, the university could return to its function of teaching, learning, and research. In linguistics an upheaval also took place in the 1960s with the spread of the transformational-generative (TG) movement. During its early years I welcomed the trend for bringing a fresh approach and vitality to the discipline. But I soon had increasing doubts about its beneficial potentialities. It became TG policy to pour abuse and ridicule on 'taxonomic' and other linguistics that did not follow TG doctrine. The confrontations began to take on the emotional tone of a religious war, and I was reminded of Bloomfield's observation on the harmful intrusion of the odium theologicum in scholarly discussions. During the decade of rapid university expansion most of the new and growing departments of linguistics were captured by TG linguists, and many linguistic journals gave all or most of their space to papers on TG theory. Several generations of students were led to believe that linguistics had no history, that nothing worthwhile had been accomplished until the advent of TG linguistics. Their training failed to introduce them to the wide range of linguistic activities from which they could choose the fields that would fit their interest and temperament. But, fortunately, enthusiasms and fashions in linguistics, like language itself, cannot escape change. In the last few years the dominating influence of TG linguistics has been dissipating, and a healthier eclectic spirit is discernible today. In the early 1960s the death of Leslie Spier left the Southwestern Journal of Anthropology (later renamed the Journal of Anthropological Research) without an editor. Because the department and the university wanted to continue publishing the quarterly begun by Spier nearly 20 years earlier, Harry Basehart, a social anthropologist, and I coedited the journal for nearly ten years. In 1970 Basehart assumed the full duties of editor, and I stayed on the editorial staff for another year or two after retiring in 1971. Since retirement I have returned to one of my early research areas, the Northwest Coast. When I did my fieldwork with the Bella Coola in the summer of 1932, only a few of the Salish languages had received attention, primarily by Boas and some of his students. The Depression halted this trend, and no further progress took place for more than 20 years. Linguistic activity in Salish was renewed in the late 1950s, and by the 1970s a respectable body of linguistic material was available. Because the reconstruction of Salish phonology was fairly well advanced, my interest turned to comparative morphology, particularly to reconstructing the history of the complex pronominal system.

30

Stanley S. Newman

Retirement for me has its deprivations and compensations. Although I visit my office and the library about once a week, I no longer have the day-to-day contact with colleagues and departmental affairs. I also miss the challenge of meeting students, especially graduate students working on their theses. Some of them found the ordeal nearly unbearable and needed constant help and encouragement; others seemed to make their way almost effortlessly. When I meet or hear from my former students, it is deeply satisfying to realize that continuity in academic affairs endures in human contacts and is not confined to the printed page. I am grateful to be free of the multiple demands that pressed on me before retirement and the deadlines nipping at my heels. I particularly enjoy the leisurely pace of retirement, the luxury of having ample time to think about what I am doing and to 'recollect in tranquillity'.

SONGS FOR A WINDY DAY Stanley S. Newman EARTH-BOUND No wind disturbs things underground; And wind bends not the stubble. For those whose stems are tightly bound To earth, whose roots are tough and sound — These the wind does not trouble. But those that fling their slim shoots high To catch the sun's sharp flashing, That reach to the receding sky — These will the wind beat endlessly, These will the wind be lashing. SCIENTIST He walked in dark oblivion Where he had never seen the sun. Then down came the hard white light Like showering manna in the night. His senses ate of this new blessing. Each stone and tree for new possessing Was made apparent to his sight. Here was a room he could explore Where questions never shook the door, He saw all that there was to see. But he had come from mystery. And soon he left his pale room To wander down the corridor Of endless night, Where sounds he never heard before Rocked in the gloom, And sudden blossoms sprang to bloom Glowing white.

32

Stanley S. Newman

RETENTION Although my ears cannot retain The whispered reticence of rain, Or the wind's tremulous wavering As it shook the budded trees last spring, And though my eyes cannot recall White lilacs splashed against a wall, Or little pools that caught the sun On dappled waves when the wind would run, Still there is singing in my mind The fragile music of the wind That I heard long ago; and here White lilac blooms all through the year.

Bibliography of Stanley S. Newman

1926

Review of A marriage with space by Mark Turbyfill. The Forge: A Journal of Verse, Spring 1926:22-23. 1926 — 28 Reviews of The heart of Emerson's journals by Bliss Perry and Bill Nye, His own life story by Frank Wilson Nye, December 3, 1926; 7 P. M. by Mark Van Doren, April 1, 1927; Oxford Poetry, 1926, ed. by Charles Plumb and W. H. Auden, November 25, 1927; and others. Chicago Evening Post, Literary Section. 1927 a. "Song for a constant season." The Forge: A Journal of Verse, Spring 1927:3. b. Review of Streets in the moon by Archibald MacLeish. The Forge: A Journal of Verse, Summer 1927:15-17. c. "Prague." The Forge: A Midwestern Review 2,5 (Summer): 14 — 20. d. "Songs for a windy day: The John Billings Fiske Prize Poems." (University of Chicago) The University Record 13,4: 302 — 6. 1928 "Swedenborgianism and Emerson." Unpublished A. M. thesis, University of Chicago, 89 pages. 1930 "Specimens." Poetry: A Magazine of Verse 36,1:26-28. 1932 a. "A grammar of Yokuts, an American Indian language of California." Unpublished Ph. D. dissertation, Yale University, b. "The Yawelmani dialect of Yokuts." International Journal of American Linguistics 7,1 - 2 : 8 5 - 8 9 . 1933 "Further experiments in phonetic symbolism." American Journal of Psychology 45: 53 — 75. 1934 "Some orthographic recommendations: arising out of discussions by a group of six Americanist linguists." By George Herzog, Stanley S. Newman, Edward Sapir, Mary Haas Swadesh, Morris Swadesh, and Charles F. Voegelin. American Anthropologist 36,4:629 — 31. 1935 a. Review of Semitic and Hamitic origins by George Aaron Barton. American Journal of Sociology 41:391 —92. b. Review of Meaning and change of meaning, with special reference to the English language by Gustav Stern. Psychological Bulletin 32: 104-5. 1936 a. Review of The psycho-biology of language by George Kingsley Zipf. American Speech 11:170 — 71. b. Review of California Indian folk lore by F. F. Latta. Journal of American Folk-Lore 49:270. 1937 "An American phonetic dictionary: comments and suggestions." By S. Newman, M. Swadesh, and G. L. Trager. American Speech 12: 138-39.

34

Bibliography of Stanley S. Newman

1938 1939 a. b. 1940 a.

b.

1941 a. b. c. d.

e. 1943 1944 1945 1946 a. b.

1947

(with Vera G. Mather) "Analysis of spoken language of patients with affective disorders." American Journal of Psychiatry 94,4:913 — 42. Review of A concise dictionary of the Bini language of Southern Nigeria by Hans Melzian and The phonetics of the Hottentot language by D. M. Beach. American Anthropologist 41,1:147 — 49. "Personal symbolism in language patterns." Psychiatry 2,2:177 — 84. Yokuts and Western Mono myths. By A. H. Gayton and Stanley S. Newman. Anthropological Records, vol. 5, no. 1. Berkeley and Los Angeles: University of California Press. Reprinted as "Linguistic aspects of Yokuts style." In Language in culture and society, edited by Dell Hymes, pp. 372-77. New York: Harper & Row, 1964. (with other members of The Committee on the Function of English in General Education) Language in general education. New York: D. Appleton-Century Co. Review of The language of gesture by Macdonald Critchley. Journal of American Folklore 54:87. Review of Psychology of English by Margaret M. Bryant and Janet Rankin Aiken. Language 17,2:158 — 60. "Speech analysis and interpretation." In The adolescent personality: a study of individual behavior, by Peter Bios, pp. 214—16. New York: Appleton-Century Co. "Behavior patterns in linguistic structure: a case study." In Language, culture, and personality: essays in memory of Edward Sapir, edited by Leslie Spier, A. Irving Hallowell, and Stanley S. Newman, pp. 94—106. Menasha, Wis.: Sapir Memorial Publication Fund. Language, culture, and personality: essays in memory of Edward Sapir. Edited by Leslie Spier, A. Irving Hallowell, and Stanley S. Newman. Menasha, Wis.: Sapir Memorial Publication Fund. "Phonetic pattern." In Ethnography of the Yagua, by Paul Fejos, pp. 118 — 19. Viking Fund Publications in Anthropology, no. 1. New York: Viking Fund. Yokuts language of California. Viking Fund Publications in Anthropology, no. 2. New York: Viking Fund. [ = Wenner-Gren Foundation]. "Cultural and psychological features in English intonation." Transactions of the New York Academy of Sciences, 2d ser., vol. 7,2:45 — 54. "On the stress system of English." Word 2,3:171-87. "The Yawelmani dialect of Yokuts." In Linguistic Structures of Native America, by Harry Hoijer, L. Bloomfleld, M. R. Haas, A. M. Halpern, F. K. Li, S. S. Newman, M. Swadesh, G. L. Trager, C. F. Voegelin, and B. L. Whorf. Viking Fund Publications in Anthropology, no. 6. New York: Viking Fund, pp. 222-248. "Bella Coola I: phonology." International Journal of American Linguistics 13,3: 129-34.

Bibliography of Stanley S. Newman

1948 a.

b. 1950 a.

b. c.

1951 a. b. c. d. e.

f. g.

1952 a. b. 1953 1954 a. b.

c.

35

Review of Kwakiutl grammar, with a glossary of the suffixes by Franz Boas (ed. by Helene Boas Yampolsky with the collaboration of Zellig S. Harris). Studies in Linguistics 6,3:69 — 74. "English suffixation: a descriptive approach." Word 4,1:24 — 36. Review of Kwakiutl grammar, with a glossary of the suffixes by Franz Boas (ed. by Helene Boas Yampolsky, with the collaboration of Zellig S. Harris). International Journal of American Linguistics 16,2:99 — 101. (with Robert Weitlaner) "Central Otomian I: Proto-Otomi reconstructions." International Journal of American Linguistics 16,1:1 —19. (with Robert Weitlaner) "Central Otomian II: primitive Central Otomian reconstructions." International Journal of American Linguistics 16,2:73-81. Review of Language of the Sierra Miwok by L. S. Freeland. American Anthropologist 53,4:555. Review of The Language of the Papago of Arizona by J. Alden Mason. American Anthropologist 53,4: 556. Review of Leave your language alone! by Robert A. Hall, Jr. International Journal of American Linguistics 17,1:54 — 55. Review of Indians of the urban Northwest ed. by Marian W. Smith. International Journal of American Linguistics 17,1:56 — 57. Review of Selected writings of Edward Sapir in language, culture, and personality ed. by David G. Mandelbaum. International Journal of American Linguistics 17,3:180 — 86. Review of Language of the Sierra Miwok by L. S. Freeland. Language 27,4:606-10. Report and recommendations of the interdisciplinary summer seminar in psychology and linguistics. By John B. Carroll (Chairman), Frederick B. Agard, Don E. Dulany, Stanley S. Newman, Leonard D. Newmark, Charles E. Osgood, Thomas A. Sebeok, and Richard L. Solomon. Mimeographed. Ithaca, N. Y.: Cornell University. Review of Methods in structural linguistics by Zellig S. Harris. American Anthropologist 54,3:404 — 5. Review of The sky clears, poetry of the American Indians by A. Grove Day. International Journal of American Linguistics 18,2:105 — 6. Review of An objective psychology of grammar by J. R. Kantor. International Journal of American Linguistics 19,4:312 — 13. Review of The category of person in language by Paul Forchheimer. American Anthropologist 56,5:926 — 27. Review of Anthropology Today ed. by A. L. Kroeber and An appraisal of anthropology today ed. by Sol Tax and others. International Journal of American Linguistics 20,2:154 — 60. "American Indian linguistics in the Southwest." American Anthropologist 56,4:626-44.

36

Bibliography of Stanley S. Newman

d. e.

f.

1955 a.

b.

c. d. e.

1956 a.

b. ?

1957 a. b.

c.

"Suggestions on the archiving of linguistic material." International Journal of American Linguistics 20,2: 111 —15. "Semantic problems in grammatical systems and lexemes: a search for method." In Language in culture: conference on the interrelations of language and other aspects of culture, edited by Harry Hoijer, pp. 82 — 91. Memoir of the American Anthropological Association, no. 79. Also published as a book under the same title. Chicago: University of Chicago Press, 1954. "A practical Zuni orthography." In Zuni Law: A field of values, by Watson Smith and John M. Roberts, Appendix B, pp. 163 — 70. Papers of the Peabody Museum of American Archaeology and Ethnology, vol. 43, no. 1. Cambridge, Mass.: Harvard University. Review of Psycholinguistics: a survey of theory and research problems ed. by Charles E. Osgood and Thomas A. Sebeok. American Anthropologist 57,5:1096-98. Review of California Indian linguistic records: the Mission Indian vocabularies of H. W. Henshaw ed. by R. F. Heizer. American Anthropologist 57,6:1338-39. Review of Power of words by Stuart Chase. International Journal of American Linguistics 21,1:77 — 80. Review of Theory of names by Ernst Pulgram. Language 31,2:247—49. "Vocabulary Levels: Zuni sacred and slang usage." Southwestern Journal of Anthropology 11,4:345 — 54. Reprinted under same title in Language in culture and society, edited by Dell Hymes, pp. 397 — 402. New York: Harper & Row, 1964. Review of Language in relation to a unified theory of the structure of human behavior: Part I by Kenneth L. Pike. International Journal of American Linguistics 22,1:84 — 88. Review of The origins and prehistory of language by G. Revesz. New Mexico Quarterly 26,3:305. "Dos quentos en el dialecto de Milpa Alta." Archivos Nahuas. (This article was to be published in Archivos Nahuas and was in fact announced in the journal, but to the best of our knowledge, it never actually appeared.) Review of The evolution of an American prose epic by Paul Radin. American Anthropologist 59,4:1135 — 36. Review of Language in relation to a unified theory of the structure of human behavior: Part II by Kenneth L. Pike. International Journal of American Linguistics 23,1:48 — 50. Review of Language, thought, and reality: selected writings of... Benjamin Lee Whorf ed. by John B. Carroll. Romance Philology 11,1:41-44.

Bibliography of Stanley S. Newman

1958 1959 a. b. 1960 a. b. c. d.

1961 a. b. c. 1962 a. b. 1963 a. b. c. 1964 a.

37

"Zuni dictionary." International Journal of American Linguistics 24,1, pt. 2: 1 — 117. Publication 6 of the Indiana University Research Center in Anthropology, Folklore, and Linguistics. Review of Migrations in New World culture history ed. by Raymond H. Thompson. Language 35,4:715 — 17. Review of Essays in linguistics by Joseph H. Greenberg. Romance Philology 12,3:283-86. Review of The content and style of an oral literature: Clackamas Chinook myths and tales by Melville Jacobs. Journal of American Folklore 73:81-82. Review of Dictionary of the American Indian by John L. Soutenburgh, Jr. New Mexico Historical Review 35,3:263. Review of L'emprunt linguistique by Louis Deroy. Romance Philology 13,3:342-43. "Two variables affecting the message in communication." By Howard Maclay and Stanley Newman. In Decisions, values, and groups 1: reports from the first interdisciplinary conference in the behavioral science division held at the University of New Mexico, edited by Dorothy Wilner, pp. 218 — 28. New York: Pergamon Press. Review of Language in relation to a unified theory of the structure of human behavior: Part III by Kenneth L. Pike. International Journal of American Linguistics 27,1:63 — 64. Review of Yarn dictionary by Edward Sapir and Morris Swadesh (ed. by Mary R. Haas). International Journal of American Linguistics 27,3:267-68. Review of God had a dog: folklore of the dog by Maria Leach. New Mexico Quarterly 31,4:375 — 76. Review of Style in language ed. by Thomas A. Sebeok. Journal of American Folklore 75:84 — 85. "Comment on 'Oceanic linguistics today' by Arthur Capell." Current Anthropology 3,4:417. Review of A William Cameron Townsend en el vigésimoquinto aniversario del Instituto Lingüístico de Verano ed. by Manuel Gamio and Raúl Noriega. American Anthropologist 65,2:420 — 21. Review of Thought and language by L. S. Vygotsky (ed. and trans, by Eugenia Hanfmann and Gertrude Vakar). Romance Philology 17,2:442 - 44. "Language and culture." NAFSA (National Association of Foreign Student Advisers) Studies and Papers no. 1, English Language Series nos. 3 & 4, pp. 1—4. Review of Friendly people: the Zuni Indians by Bertha P. Dutton and Sun Chief: The autobiography of a Hopi Indian ed. by Leo W. Simmons. Western Folklore 23,4:283.

38

Bibliography of Stanley S. Newman

b. 1965 a. b. c. 1966 a.

b.

1967 a.

b. c. d. e.

f.

g.

1968 a. b.

"Comparison of Zuni and California Penutian." International Journal of American Linguistics 30,1:1 — 13. Review of Studies in California linguistics ed. by Wm. Bright. International Journal of American Linguistics 31,4:361 —63. Index. 1945 — 64. Vols. 1 — 20. Prepared by Harry Basehart and Stanley Newman. Southwestern Journal of Anthropology. Zuni grammar. University of New Mexico Publications in Anthropology, no. 14. Albuquerque: University of New Mexico Press. Review of Exact methods in linguistic research by O. S. Akhmanova, I. A. Mel'chuk, R. M. Frumkina, and E. V. Paducheva (trans, from the Russian by David G. Hays and Dolores V. Möhr). Romance Philology 19,4:594-96. "Note on 'Inflectional class and taxonomic structure in Zuni' by Willard Walker." International Journal of American Linguistics 32,2: 226-27. Review of Los mil elementos del Mexicano clàsico: Base analitica de la lengua Nahua by Mauricio Swadesh and Madalena Sancho. Canadian Journal of Linguistics 13,1:54—55. "Zuni grammar: alternative solutions versus weaknesses." International Journal of American Linguistics 33,3:187 — 92. "Obituary of Morris Swadesh." Language 43,4:948 — 57. "Yokuts." Lingua 17,1 - 2 : 1 8 2 - 9 9 . Relativism in language and culture. University of New Mexico Fourteenth Annual Research Lecture, May 10,1967. Albuquerque: University of New Mexico Press. Also appeared in New Mexico Quarterly 37,3:197-210. "Introduction." In Studies in Southwestern ethnolinguistics: meaning and history in the languages of the American Southwest, edited by Dell H. Hymes and William E. Bittle, pp. 1 — 12. The Hague: Mouton. "Classical Nahuatl." Handbook of Middle American Indians, Vol. 5, Linguistics, ed. by Norman A. McQuown, pp. 179 — 99. Austin: University of Texas Press. Review of Obligatory constructions of Isthmus Nahuat grammar by Howard W. Law. Language 44,1:189 — 91. Review of Excavations at Snaketown by Harold S. Gladwin et al., The rock paintings of the Chumash by Campbell Grant, The three-C site by Gordon Vivian, They sang for horses by LaVerne Harrell Clark, Run toward the nightland by Jack Frederick Kilpatrick and Anna Gritts Kilpatrick, Ayers rock by Charles P. Mountford, Words and things by Roger Brown, Zuni grammar by Stanley Newman, Studies in shamanism ed. by Carl-Martin Edsman, and The dream in human societies ed. by G. E. von Grunebaum and Roger Caillois. Western Folklore 27,2:129-30.

Bibliography of Stanley S. Newman

c.

d.

1969 a.

b. c. 1971 a. b. 1973 a.

b. 1974 1976 1977 1979 a. b. 1980 a. b.

39

"Zuni equivalents of English 'to be'." In The verb 'be'and its synonyms. Part 2. EskimoI HinduI ZuniI Modern Greek/Malayalam/Kurukh, edited by John W. M. Verhaar, pp. 60 — 70. Foundations of Language, supplementary series, vol. 6. Dordrecht: D. Reidel Publishing Co. "A comparative study of Salish lexical suffixes." Working paper for the Third International Conference on Salish Languages, University of Victoria, August 26-27, 1968. Dittoed. Review of The enduring Navajo by Laura Gilpin, The smoke hole by Charles G. Newcomb, Kinaaldà by Charlotte Johnson Frisbie, Pima Indian legends by Anna Moore Shaw, Once upon an Indian tale by Helen S. Carkin and Norman H. Lerman, The savage mind by Claude Lévi-Strauss, The teachings of Don Juan by Carlos Casteneda, and New Echota letters ed. by Jack Frederick Kilpatrick and Anna Gritts Kilpatrick. Western Folklore 28,4:288-90. "Bella Coola grammatical processes and form classes." International Journal of American Linguistics 35,2:175 — 79. "Bella Coola paradigms." International Journal of American Linguistics 35,4:299-306. "Bella Coola reduplication." International Journal of American Linguistics 37,1:34—38. Review of The prehistory of languages by Mary R. Haas. Lingua 27,1:97-100. Review of Finding the center: narrative poetry of the Zuni Indians trans, by Dennis Tedlock. International Journal of American Linguistics 39,4:261-63. Review of The Zunis: self-portrayals by the Zuni People (trans, by Alvina Quam). New Mexico Historical Review 48,2:176. "Linguistic retention and diffusion in Bella Coola." Language in Society 3,2:201-14. "Salish and Bella Coola prefixes." International Journal of American Linguistics 42,3:228-42. "The Salish independent pronoun system." International Journal of American Linguistics 43,4:302 — 14. "A history of the Salish possessive and subject forms." International Journal of American Linguistics 45,3:207 — 23. "The Salish object forms." International Journal of American Linguistics 45,4:299-308. "Functional changes in the Salish pronominal system." International Journal of American Linguistics 46,3:155 — 67. "Harry W. Basehart: an appreciation." The versatility of kinship: Essays presented to Harry W. Basehart, ed. by Linda S. Cordell and Stephen Beckerman, pp. xiii —xvii. New York: Academic Press.

40

Bibliography of Stanley S. Newman

1982 1983 1984 1985 f 1986 f

In press

"Toward a history of American linguistics." [Review article on Boyd H. Davis and Raymond K. O'Cain, eds. First Person Singular, 1980.] Historiographia Linguistica 9,1 —2:135—43. Review of Psychiatrist of America: The life of Harry Stack Sullivan, by Helen Swick Perry. Historiographia Linguistica 10:135 — 38. "Edward Sapir (1884—1939)." International Journal of American Linguistics 50,4:355-57. "Reinterpreting primary data: the Salish passive." International Journal of American Linguistics 51,4:521 —23. "The development of Sapir's psychology of human behavior." In New perspectives in language, culture, and personality: Proceedings of the Edward Sapir Centenary Conference, ed. by William Cowan et al., pp. 405 — 27. Amsterdam: Benjamins. "Zuni." Handbook of North American Indians, Vol. 17, Languages, ed. by Ives Goddard. Washington, D. C.: Government Printing Office, [ca. 1974],

Inventory of Stanley Newman's linguistic materials Mary Ritchie Key

The personal linguistic files of Stanley Newman were presented to the Maxwell Museum of Anthropology, University of New Mexico, by his wife, Lucy Newman. The following is based on the list made by the Museum, which I visited in order to review the collection and prepare the list for publication. I could not help but enjoy the human and the quaint, the nostalgic and the realistic, the concrete evidence of Professor Newman's personal experience of linguistics for six decades. His university life began at a newly established university in California: the "University of California, Southern Branch", now known as UCLA. Some years later, he was job-hunting and his files contain a carbon copy of a letter of recommendation by Edward Sapir (letter to Paul Klapper, June 3, 1938). Newman went from the time when there was a shortage of positions and money to the time when there was a shortage of linguists. Correspondence with "Morry" Swadesh reflects the economic times in 1938, when Swadesh was willing to work two months for two hundred and fifty dollars (letter from Morris Swadesh, January 13, 1938). In 1953 Newman wrote to William R. Bascom at Northwestern University, "I am sorry that I do not know of any young linguists who would be available ..." (letter to Bascom, February 2, 1953). In 1936 Newman received a letter from Boas, Bloomfield, and Sapir, inviting him to explore with them the idea of organizing the "Society for American Indian Linguistics" and asking for suggestions for the proposed society. They had in mind such things as the possibility of adopting "The International Journal of American Indian Languages" as the official organ. Newman's correspondence shows that he was in touch with linguists who were on the leading edge of linguistics. Correspondence with Charles Hockett refers to the marvelous story about the editor of the German philological journal returning Boas' manuscript on the Bella Coola language (Newman to Hockett, February 23, 1953; Hockett to Newman, February 27, 1953). Influence across the Atlantic shows in the three-way correspondence among Daniel Jones, Morris Swadesh, and Stanley Newman in 1954, about "word-junction" and internal juncture.

42

Mary Ritchie Key

That he was concerned about the condition of linguistics in these later years is evident throughout his correspondence. These concerns were expressed as early as 1969, when he wrote, "There seems to be a widespread fashion among intellectuals to make militant assertions that they've got the final answers to all kinds of complex problems, and it often slips over into a battle of the Angels against the Forces of Evil. Politics seems to have set the tone, which has invaded various academic disciplines besides linguistics. It worries me." (letter to Julian Granberry, February 26, 1969). He was a flexible person; he was tolerant of varying positions. In a letter about discoveries of language relationships, he wrote, "Each of us ought to keep hacking away at concrete problems, extending old procedures and devising new ones, without expecting too much. Some of the little pieces do eventually fall into place in the larger picture. But there seems to be no way to predict which little pieces are going to fit and which turn out to be bogus leads. The safest thing to do is to use multiple procedures, checking one against the other as much as possible, and to hope for the best." (letter to William W. Elmendorf, April 5, 1964). The matter of distant relationships of languages is very much evident in the collection, with correspondence from many others besides Swadesh, who kept finding similarities among languages and making lists between various languages. This remains to be sorted out. One notices some sociolinguistic rules evidenced in the correspondence, when titles and naming indicate the relationship. The younger Newman wrote to Leonard Bloomfield at Yale University, on June 10, 1945, addressing him as "Dr. Bloomfield", and the answer came back, "Dear Newman". But Boas had kept a more formal address when he addressed his much younger colleague in the early 1940s as, "Dear Dr. Newman" (letters from Franz Boas, September 24, 1940, October 5, 1942). Letters between Newman and Dwight Bolinger began, "Dear Newman" and "Dear Bolinger" (February 29, 1948, and February 28, 1955). And likewise, a letter from Boas to Sapir starts out, "Dear Sapir". A precious part of the collection contains the personal notebooks and slips of Franz Boas. It must have been an appreciative Boas to be able to leave his treasured Bella Coola notes with this eager young researcher. I have listed the linguistic material below. In addition the collection contains bunches of Newman's reprints, books and other publications by other linguists, and Ph.D. dissertations by others, which I have not itemized. The value of these other items rests on Newman's remarks in the margins. The 4 x 6 files are in metal file drawers and shoe boxes (without which no proper linguist could have functioned!). Here I will

Inventory of Stanley Newman's linguistic materials

43

simply refer to either container as a "box". One of my favorite moments in reviewing Newman's papers was when I came across one file labeled "Unclassified; Unclassifiable"!

Inventory Large brown envelope/folder entitled "Materials on language teaching". [Apparently notes for classes in English grammar at the University of Texas, Austin, 1928-1929.] Large box of scholarly correspondence, filed in alphabetical order. Separate file box of correspondence from the more recent years, filed in alphabetical order. Collection of letters between Edward Sapir and Stanley Newman. [Newman was in the process of transcribing these from the hand-written originals.] Collection of letters from Edward Sapir to Robert H. Lowie, with Introduction and Notes by Robert H. Lowie. Typescript, 1965. Folder of material on Bella Coola. Seventeen manuscripts and three published articles by Ross Saunders and Philip W. Davis, with handwritten notes by Stanley Newman in the margins. Folder of Morris Swadesh' "Amerindian Noncultural Vocabularies" in various languages. Hand-written lists of "Vocabulary (50 words) in 36 languages" [of Mexico], Consejo de Lenguas Indigenas, Biblioteca. Folder of Zuni Miscellany. Folder of Salishan notes. Boas' notes and a typescript, "Cœur D'Alene locative suffixes". Fourteen notebooks with linguistic transcriptions of texts: One each of Isleta, Carrier, Keresan, Zuni Bella Coola I, and Bella Coola II Eight of Yokuts File Boxes of 4 x 6 cards and slips: Bella Coola — four boxes Zuni and Zuni-Penutian — four boxes Yawelmani — three boxes Nahuatl — one box Miscellany — two boxes [Phonology of various languages, Persian notes] Bibliographical references — one box

44

Mary Ritchie Key

Boas' field notes: Bella Coola 3 x 5s [and even smaller!], labeled "Boas slips" Two small notebooks (blank paper) with hand-written notes, 1886 and 7 August 1879. [A kind of diary or log, with notes on the weather: "clear" or "light breeze", and linguistic data.] Maps: J. W. Powell, "Map of Linguistic Stocks of American Indians", Annual Report of the Bureau of Ethnology, Vol. 7. C. F. Voegelin and E. W. Voegelin, "Map of North American Indian Languages," American Ethnological Society, Publication No. 20, in collaboration with Indiana University. A. L. Kroeber, "Native Tribes of North America," University of California Map Series, No. 13, 1939. University of California Publications in American Archaeology and Ethnology, Vol. 38. "Figure 15 — Linguistic Map of Middle America". Sapir map, printed for the Ottawa conference [in color].

"Singularly like our ideal of a scientist" Michael

Silverstein

To me, who knew him on a daily basis during two southwestern summers, Stanley Newman will always stand out as a figure of cool and inviting alertness against a ground of meltingly hot Albuquerque adobe and asphalt. In July and August of 1969, and again in June and July of 1980, my work was situated at the University of New Mexico. Particularly during the first of these periods, before Newman became emeritus professor, I was privileged to be welcomed into one of the most congenial of movable intellectual feasts in the Department of Anthropology, presided over by Stanley and his close friend Harry W. Basehart. I think of it as a fairly continuous seminar of some seven or eight weeks, carried on in the offices of the department, the lunch spots around campus, or at someone's home for dinner. But mostly, as I recall, it flourished in the small, pea-green (or was it dreary institutional tan?) ground-floor canteen in a building across campus in which faculty and staff drank late afternoon tea and coffee, mostly hot, out of styrofoam cups. Every weekday, on their way out of Maxwell Hall, Harry and Stanley would stop by the open door of Bruce Rigsby's office, where we were puzzling out Gitksan (Tsimshian) syntax, and invite us to join them. And over tea would be reviewed the history of American anthropological theory, the unpublishable stories of various figures io the subfields — which included Amerindian and even general linguistics in the departmental worldview — and other matters that these two editors of one of the major anthropological journals thought about aloud in conversation. In a sense, Rigsby and I were an audience consisting of another collaborating pair of colleagues at a generation or so remove. Soaking all this up and enjoying every minute as the tea cooled and was drunk, we learned about Kluckhohn and Spier and Steward, about Sapir and White and Gayton, and about other ancestors of our collective past: we were being socialized professionally. In a number of private conversations — I daresay, they must have seemed to the long-suffering Stanley to be more like interviews! — I was able to engage him about Sapir's Yale group of the 1930s, in which he, Newman, had figured so prominently, about linguistic theory and linguis-

46

Michael

Silverstein

tic anthropology, and about various technical matters that the dean of California Penutianists would certainly have encountered in his work on Yokuts. His patience and kindness seemed boundless; he even arranged for the U N M photoduplication services to make copies for me of his 1930 — 31 field notebooks of Yokuts texts in the various dialects. (They later proved to be invaluable in my dissertation work, and I hope they will someday be properly published.) Touched in this way some years ago by Stanley's welcoming collegiality, at the invitation of William Bright, then editor of Language, I attempt here to set the documentary record of Newman's career in the context of the personal and institutional circumstances that seem to have shaped it, guided by my profound admiration for him as well as for his œuvre. *

*

*

*

*

Had he not suddenly died on 26 August 1984 in his eightieth year, Stanley Newman would have been President of the Linguistic Society of America in 1986. Deprived of the synthetic statement that would articulate his own views, in the form of a Presidential Address, we can attempt to understand the character of his achievements as one of the most gifted of Sapir's successors, whose career in a way typifies the experience of a whole group of that academic generation. 1 Stanley Stewart Newman was born on 18 July 1905, the second child and only son of Rudolph and Julia (Hübsch) Newman. Both Newmans had been immigrants to Chicago's south side from Czech-speaking Bohemia, part of then Austria-Hungary. In Chicago, they ran a jewelry business. The Newman household, like a number of others in the preWorld War I era about which we have read in these pages, 2 was multilingual (Czech, German, English) and conducive to learning in the central European manner and to that Culture (with a capital C) that flowed across borders and languages with equal ease.3 Newman himself ("How I discovered linguistics," in this volume) recalled vividly two family trips back to the Prague area (see also Newman 1927 c, and this volume, for a fictionalized impression), the second of which involved the adventure of returning to America in wartime. At the time Newman was graduated, in 1923, from Robert Lindblom High School on Chicago's southwest side, the family moved to Los Angeles, and Newman enrolled in what was then the "University of California, Southern Branch" during 1923 — 24. Newman described it as "primarily a teacher training institution," and, wanting a broader education, he was attracted back to The University of Chicago, where he majored in English literature (specifically, American) and minored in

"Singularly like our ideal of a scientist"

47

French. He was awarded his Ph. B. with honors in English (and election to Phi Beta Kappa) in June of 1927. Then he stayed on in the English department as a graduate student during 1927 — 28, offering for the A. M. degree a thesis on Swedenborgianism and Emerson (1928). Like so many of our predecessors who were attracted into the then fledgling academic discipline of linguistics in America, Newman attributed his shift of fields in large measure to his personal encounter with a great figure,4 in this case Edward Sapir, who had come to Chicago's Department of Sociology in Autumn quarter 1925 as Associate Professor. Newman had clearly been a very good, though not outstanding, student of American literature in his first graduate year. He was something of a poet, having won as a College senior a local poetry contest (the John Billings Fiske Prize) and having had his poems published. He reviewed literature for the Chicago Daily News and co-edited a poetry magazine (The Forge), enjoying membership in the students' Poetry Club at The University. His poems, like his M. A. thesis, reveal a neo-Romantic young scholar with deeply ambivalent feelings about the ultimate value of the world of learning, and in particular about the scholarship in which he was engaged. There is a longing for something more permanent, some satisfaction to be found in a kind of scientific and more absolute insight into otherwise hidden patterns. (This had been, in fact, one of the powerful claims of Emanuel Swedenborg that appealed so to intellectuals of Emerson's time, as Newman's thesis (1928) documents.) And here was Sapir, a noted scientist, the only member of the faculty at Chicago regularly to attend the Club and, like the other members, to submit his poems anonymously for critical discussion by the group. Here is how Newman describes it: From our perspective, furthermore, anthropologists were people who dug in the ground for buried treasure or studied apes, and language specialists were preoccupied with subjects like Old English and Beowulf, which seemed remote, even antagonistic, to an interest in writing poetry. We could recognize the poems that Sapir submitted, a polished classical type, more restrained than the frankly lyrical verse that most of us wrote. We were particularly impressed by his incisive comments on the poems read aloud and by his broad acquaintance with modern poetry; I doubt that any of us had heard of Gerard Manley Hopkins until Sapir mentioned his name and quoted a few striking lines of a Hopkins poem. We were impressed, too, by the elegant fluency of his speech. His spontaneous choice of words had the characteristics of carefully crafted prose. ("How I discovered linguistics," this volume.)

48

Michael Silverstein

Newman was indeed reacting to what others, too, have described as Sapir's intoxicating, brilliant, verbally-centered personality, indeed to his seductive way — if we may be quasi-psychological for a moment — with audiences and interlocutors. In the first quarter of his graduate career, Autumn 1927, along with the usual first-year fare of the English literature student — "The bibliography and methods of English literary history," "Shakespeare," and "Studies in the literature of New England in the nineteenth century" — Newman enrolled as a "visitor" (what we now call an auditor) in Sapir's famous course "The Psychology of Culture," where, he recalled, he was introduced "to a world of fresh ideas and new writers — Jespersen, Vendryes, Croce, Boas, Jung, Spengler, Veblen." After another year (1928 — 29) — "a searing experience" — as an instructor in English composition and literature at the University of Texas, Austin, and an additional stab at English courses back at Chicago in the Summer quarter of 1929, Newman decisively switched to the newlyorganized Department of Anthropology in its first year of operation, Autumn 1929. And in the 1930 — 31 academic year, he learned that he was to be among the whole group of advanced linguistics students in the anthropology department who would accompany their mentor to New Haven, where Sapir was to set up a Department of Anthropology at Yale University as Sterling Professor and Chairman. Because of this academic migration, Newman received his Ph. D. in anthropology from Yale the next year, in Spring, 1932, when he completed his dissertation, "A grammar of Yokuts, An American Indian language of California."5 Sapir seems to have embodied for the young Newman a scintillating demonstration that one could be a scientist and at the same time not lose that concern for Culture which had been their common personal background; that one did not have to give up concern for the poetry of language in order to study language scientifically; and that anthropology as the systematic understanding of socialized systems of symbolism, language included, grounds one's concern for Culture in his own society in a scientific study of cultures in the comparative perspective. And indeed, in submitting himself to apprenticeship with the older scholar, it seems clear that for the younger man it was important to see the lessons given voice and articulate form. The triteness of the metaphor notwithstanding, it seems clear that there was a kind of extended academic family at Yale6 composed of Newman and the others who removed there with Sapir, plus those who early on joined them there as students or postdoctoral scholars.7 But Newman always seems to have been a favorite of

"Singularly like our ideal of a scientist"

49

Sapir, and one senses a fatherly reverse transference onto, or identification with, Newman by the older scholar, of whom Sapir spoke in terms that are clearly those of an idealized self and its accomplishments — terms with which, indeed, Sapir himself is frequently described in various encomia over the years.8 Funded by the then ubiquitous ACLS Committee on Research in American Native Languages (part of the Boasian academic network), Newman conducted two summers (1930 and 1931) of field research on Yokuts dialects, principally the Yawelmani dialect, then the lingua franca of the Tule River Reservation east of Porterville, California. After the academic move to Yale, this resulted in his 1932 doctoral dissertation as well as a later series of grammatical sketches of varying scope (Newman 1932 b, 1946 b, 1954 d, 1967 d), language-in-culture studies (1940 a, 1967 e), and, of course, his magnificent Yokuts Language of California, completed in 1936 but published, despite many efforts of no less formidable a figure than Boas, only in 1944 as the second number of the Viking Fund Publications in Anthropology.9 This grammar is, we can now appreciate, the monument by which Newman will ever be remembered. It is also the finest achievement of the kind of approach to linguistic description of the Sapir group of the 1930s (cf. Harris 1944:198), a model that Sapir himself was to recommend to others for emulation (Norman McQuown, p. c.). Sapir pronounced it "perhaps the most beautiful ... story of an American Indian language that has ever been written, beautiful not only because the language is beautiful, but also because the treatment is highly balanced" (National Research Council, ms. 1935). Indeed, there is something of the linguist's specifically professional scientific aesthetic with respect to linguistic structure that shines through in Newman's various general characterizations of Yokuts as well as in his description of specifics.10 Newman's description succeeded in putting Yokuts, so to speak, on the theoretical map for his and for every succeeding generation of linguists. Certainly, among its virtues are the lucid discussions of grammatical categories, such as aspect (1944:82 — 83; 96 — 113), then being discovered and investigated by the group at Yale (e.g., in Navajo, in Hopi, in "Wishram" [ = Kiksht]) as a universal category of fundamental import in grammar. (See the tribute in Friedrich 1974: S31 — S32.) But Newman's grammar has remained the object of study and analysis particularly for its treatment of the ubiquitous processes of stem-vowel ablaut, vowel epenthesis and deletion, and vowel harmony, that interact in such complex yet rigid ways in the structure of all inflectable word classes. Yokuts has

50

Michael

Silverstein

become, in the current terminological idiom, the case par excellence of so-called "absolute neutralization," around which as a fulcrum theoretical views have seesawed for some time on the possibilities for neatly dichotomizing the phonetics — phonemics — morphophonemics cline of abstraction (to use Newman's 1944 terminology) in some principled way.11 Even Newman's treatment of the vowel system over the period 1930 — 1944 reveals changes reflecting theoretical movements within the views of the "First Yale School." For example, the final description, presumably ultimately revised for publication in the late 1930s, speaks in terms of a differentiation of phonemes and morpho-phonemes (see, e. g., 1944:20), while the earlier treatments speak in terms of phonemic units simply (as in the sketch published in 1946, written for the volume in honor of Boas that Sapir was organizing around 1935),12 and the earliest treatment (1932 b) speaks in terms of Yokuts "phonetics," with the then recent and radically new concept of the phoneme as the basis for Newman's transcriptional practice, much like the examples in Sapir's "Psychological reality" article of 1933.13 In the summer of 1969, I questioned Newman about the achievement of the solution to this extraordinarily difficult problem in phonological alternation. He explained that Sapir took an ongoing interest in the working out of the system, and had to intervene with some real arm-twisting of a very phonetically-committed Franklin Edgerton — one of the examiners of the dissertation when it was presented at Yale in 1932 — in order to have it accepted, at least by quiescence!14 In the current political and economic climate in the United States, it is perhaps possible empathetically to appreciate how the 1930s brought to a halt the painfully slow uphill struggle of academic linguistics and its scholars, both institutionally and personally. For the group of linguists around Sapir at Yale, those Depression years must have been particularly crushing, since the only prospects for almost a decade were fellowships and grants that left them without normal academic positions far into their post-doctoral careers. Increasingly frustrated and ultimately overcome by conditions in the academic polity of Yale itself, Sapir and his increasingly advanced graduate and post-doctoral group scraped together funds for their work from a number of different sources. To be sure, the group sustained and even revitalized the professional level of Boasian linguistic research on American languages. At the same time, they pushed forward an intellectual program of making professional-quality linguistics a central part of numerous interdisciplinary and applied ventures in the social sciences, for which Sapir's position at the apex of the national social

"Singularly like our ideal of a scientist"

51

science "establishment" could translate into modest amounts of support, despite the Depression. The record of Newman's professional titles after his Ph. D. attests to the way in which even this most favored student of Sapir's had to manage to continue his linguistic work mostly on what we now term "soft money." During the academic years 1932 through 1937, Newman held an appointment as Research Assistant in Yale's Institute of Human Relations (which later became the Human Relations Area Files), then serving as the research arm of a number of social science departments. From 1932 — 34, he was a Research Fellow on a grant from the Social Science Research Council; from 1934 — 36, on a grant from the American Council of Learned Societies; and in the years 1935 — 36 and 1936 — 37, Newman was appointed as Instructor in Field Linguistics at Yale, where he taught courses part-time. In 1937, Newman moved to New York City, where for two years he was a Research Fellow working on the Study of Adolescents of the Commission on Secondary School Curriculum, Progressive Education Association, an interdisciplinary project with teachers, psychiatrists, and social workers. ("I hope that you succeed in injecting some measure of true science into the thinking of the ... group," wrote Sapir encouragingly on 30 September 1937!) One senses something of the anxious and insecure mood of such times in Sapir's letters to Newman during periods, like summers, when they were not both present in New Haven.15 Each letter is variously a commentary on projects then underway, a solicitation of reports on work done, a suggestion for "leads" on whom to see, what to propose to secure funding for new projects, an expression of hopes for new projects, and a complaint about dealing with the administration at Yale. For all of the generation that came of scholarly age in the early 1930s, the hand-to-mouth economic pressures of the Depression years, followed by the mobilization during World War II, and the only gradual reemergence of the universities as sites for their scholarship in the late 1940s, were overwhelming factors in shaping their professional activities and thereby the existence of linguistics as a discipline over a period of almost twenty years. So, for all of the ongoing disciplinary and interdisciplinary team activities in which Newman played a role during the 1930s, by 1941 Newman's work, too, was focused in the war effort of the whole linguistic profession, providing language materials and teaching foreign languages, first under the ACLS Intensive Language Program (1941 —43), and then (1943—45) in the Language Section, Information and Education Division, of the War Department, better known as "165 Broadway." First, housed

52

Michael

Silverstein

at Columbia University, collaborating with a native Persian speaker, Kazem Kaihani, Newman prepared materials and taught officers on their way to Teheran. At 165 Broadway, Newman produced a number of different types of materials in Persian and other languages, and perhaps most notably the "reverse English" courses for speakers of other languages, El inglés hablado for groups of Latin American mechanics, L'inglese parlato for Italian prisoners of war in the United States, and Englisch wie man's spricht for German-speaking ones ("How I discovered linguistics," this volume). And after the war, Newman, still in government service, was posted to Mexico City by the Smithsonian's Institute of Social Anthropology, where he had the title of Anthropologist from 1945 to 1949. In this post-war version of Rooseveltian "Good Neighborly scientific and educational aid to Latin America, Newman taught (in Spanish) and conducted research as Professor of Linguistics at the Escuela Nacional de Antropología e Historia, increasingly becoming aware of how shifting was the support for this kind of venture in the United States. It was thus only in 1949 that Newman secured a regular, permanent university appointment in the Department of Anthropology at the University of New Mexico in Albuquerque, becoming full professor in 1953 and emeritus professor in 1971. To be sure, Newman had had a varied number of teaching appointments up to that time, including, in addition to those already mentioned, a stint as an adjunct anthropology instructor in the Sociology Department of Brooklyn College in 1942 — 43 (concurrent with his linguistic research). But it is clear that in Albuquerque, Newman was finally able to settle in and to sense the long-term rewards of being a senior professor with undergraduate and graduate teaching and advising, long-term research development, and editorial work on a fine general anthropological journal, the Southwestern Journal of Anthropology (now called the Journal of Anthropological Research) housed in his department. Newman was the first linguist to be hired to a regular appointment in the University of New Mexico, and happily he lived to see a separate department established (which sponsored a Linguistic Institute in 1980), and many interdisciplinary and applied linguistics projects and programs flourishing in one or another unit of the university. Newman's research followed a parallel course. During the 1930s, he brought to empirical realization several of Sapir's theoretical and essayistic formulations on the "psychiatric"16 functions of language in human affairs and in the expression of personality. Already at Chicago, Newman had used the considerable statistical sophistication he had gained studying under L. L. Thurstone to complete an innovative analysis of "Further

"Singularly like our ideal of a scientist"

53

experiments in phonetic symbolism" (1933), following up on Sapir's (1929) tentative study; some aspects of this large-scale experimental work he continued in New Haven. But in a more sustained program, several of Newman's projects during this period focused on the (clinical) psychiatric notion of personality, which, with its quasi-ethnographic case-study orientation, was the particular line of culture-and-personality investigation consonant with Sapir's formulations about "Language as a form of human behavior" (1927 a) and "Speech as a personality trait" (1927 b), etc. With the psychiatrist Vera G. Mather at Yale, Newman did a pioneering study (1938) of forty patients with various disorders of affect, such as clinical depression. Based on the innovative technique of sample recordings, they attempted a differential linguistic characterization of the overall speech style of these patients in terms of indicators cutting across all components of linguistic structure (see also Newman 1945). In the New York City adolescence study, Newman carried out longitudinal casestudies of such stylistic indicators characteristic of the Sturm-und-Drang of adolescence (Newman 1939 b, 1941c, 1941 d). His interest in the whole area of what became, for another generation, "psycholinguistics" reemerged after the war, when Newman was one of the original participants in this SSRC-backed movement in the early 1950s,17 and one of the senior advisory staff of the famous Southwest Project in Comparative Psycholinguistics (1955 — 56), field-based at the University of New Mexico (see Newman 1967 f: 6 — 7 and the bibliography [by John Carroll] in Hymes-Bittle 1967:452-4; cf. also Carroll 1980:48 — 9). [See photograph in this volume.] In this project, many of the now senior people in linguistics and psychology got early crosscultural research experience in cognitive and other studies. It must have been gratifying to Newman, the student of Sapir and colleague of Whorf,18 to see the cross-cultural dimensions of language and the interdisciplinary dimensions of linguistics being once more taken seriously, however transformed the construal of the problems were in the predominantly behaviorist psycholinguistic atmosphere of the 1950s. Much earlier, when Sapir's and Whorf s formulations of such concepts as 'formfeeling' and 'fashions of speaking' were being first explored, Newman (1940 a) had written about the relativity of aesthetic consciousness instanced in Yokuts (as opposed to English) narrative "style." Though framed in terms of a stylistic analysis of literary sensibility — examining the relationship between Yokuts phonological, grammatical, and semantic structure and the range of their utilization in narrative text — this was fully in the Boasian tradition that Sapir and Whorf also

54

Michael

Silverstein

had drawn upon. The problem is one of the historically-specific system of cultural (including linguistic) categorizations in each social group that forms an implicit reference space — a kind of collective unconscious or intuition — and the unawares backdrop for what surfaces through the operation of each individual group member's consciousness, what surfaces as products and creations of human agency. When Whorf was rediscovered under the guise of "metalinguistics" or "ethnolinguistics" in the early 1950s, Newman's was one of the almost lone voices of modulatedly gray Boasian sanity on the otherwise black-and-white behaviorist issue of "linguistic relativity." He participated in the notable 1953 University of Chicago interdisciplinary conference "on the interrelations of language and other aspects of culture" (Hoijer 1954), contributing a paper (1954 e) and to the very lively discussions of Whorf. And he returned many times to the theme of culturally-conditioned relativity of conscious, value-laden individual experiences, such as aesthetic production and affective display, most notably in his Annual Research Lecture at UNM — one of the highest local faculty honors — in 1967, "Relativism in language and culture" (Newman 1967 e). But Newman's psychological and cultural interests and his wartime duties notwithstanding, for his entire scholarly career he maintained a continued production of both descriptive and historical scholarly monographs on a number of languages and language families of North America and Mexico. Each is marked by meticulous attention to detail, and great methodological consistency. Immediately upon getting his Ph. D., Newman worked for the summer of 1932 on Bella Coola, the Salishan outlier in British Columbia. For a number of reasons — one of them, I gather, his negative aesthetic reaction to the language, which lacks the crystalline elegance and regularity of word structure to which he had so responded in Yokuts and would find again in Zuni 19 — he published a phonological sketch in UAL only in 1947,20 and morphological articles only much later (1969 a, 1969 b, 1971), by which time others (Ross Saunders and Philip Davis) had already taken up the modern, syntactically based study of the language. Newman took advantage of his presence in Mexico while at the Escuela Nacional to work on Otomi, publishing (with Robert Weitlaner) the fundamental comparative study, with reconstructions of Proto-Otomi forms (Newman—Weitlaner 1950 b, 1950 c). He also did fieldwork on a dialect of Nahuatl, and prepared an "Army Method" teaching text in Spanish to teach the language (collaborating with Robert Barlow). This work was useful in Newman's descriptive sketch of Classical Nahuatl

"Singularly like our ideal of a scientist"

55

(1967 a) for the Handbook of Middle American Indians. And when he took up residence in Albuquerque, Newman added southwestern American Indian languages to his repertoire, in particular, starting in 1951, Zuni, when he published a dictionary (1958) and grammar (1965 c), as well as early on developing a practical orthography (1954 f). As an accomplished practitioner of the genre, Newman has left us a compact and informative grammatical sketch (In press, [written ca. 1974]) for the linguistics volume of the new Handbook of North American Indians. Finally, in his emeritus years, Newman returned to Salishan, producing the foundations for a comparative and historical inflectional morphology in a series of elegant papers (1976, 1977, 1979a, 1979b, 1980a, 1985) that demonstrate his mastery of this family and the intricate chronology of grammatical (morphosyntactic) innovations in its subgroups. The autobiographical essay (this volume) commissioned by Konrad Koerner that Newman has left evidences the equanimity and grace with which he carried on his various assigned and self-assumed scholarly tasks amid all the upheavals and, simply put, less-than-ideal conditions to which he and his coevals were frequently subject (see also Newman 1982, esp. pp. 139 — 41). Part of this accomplishment lies in Newman's being, as Harry Stack Sullivan himself put it, "singularly like our ideal of a scientist"; Newman was modest, never self-promoting or aggressive, penetratingly thoughtful and fair-minded. A great part of it no doubt lies in the atmosphere of life-long mutual devotion of Newman and his wife, Lucy (Price-Benedict), whom he married in 1930, and who survives him.21 And part of it can perhaps be understood from the sustaining qualities of Stanley Newman's even-tempered personal relationships with colleagues, with students and advisees, and with linguistic informants; offering his characteristic, empathetic graciousness, it returned to enrich its source. It is clear that Sapir responded to this quality; as did Swadesh, his friend from Chicago times on (Newman 1967 c); as did Harry Basehart, his colleague and coeditor, his closest friend at New Mexico (see Newman 1980 b); and, as did, one way or another, all of us who knew him.

Notes 1. I would like to acknowledge the invaluable kindness of Garland D. Bills, Regna Darnell, Konrad Koerner, Norman A. McQuown, and James M. Nyce, who have supplied me with otherwise unavailable biographical material of various sorts, and of Paul Friedrich, for several suggestions that have improved the presentation.

56

Michael Silverstein

2. Compare Joos (1967:3) writing about Bernard Bloch, for example. 3. One might surmise that this feature of the milieu around Sapir at the Yale anthropology department of the 1930s must have convinced Whorf even more of the correctness of his intuition about a "Standard Average European" outlook and world view embodied in speakers of whatever particular language of the area. 4. The most striking example is Bloomfield's report of his encounter with Eduard Prokosch (Bloomfield 1938:311 — 12), reprinted in Bloch's obituary of Bloomfield (Bloch 1949). But see also Swadesh's (1939:132) recounting of a similar experience of Sapir's in Boas' office; Halle's (1983:72—73) of his own in Jakobson's hospital room; Harold Allen's (1980:111) of his own in Charles Carpenter Fries' class; Victoria Fromkin's (1977:172) of hers and others' in Harry Hoijer's classes; etc. This form of personal memory has something of the character of a (sub)cultural regularity, as, perhaps, do whatever experiences that underlie them. 5. For the publication history, see below and cf. Newman 1944:5 — 6 and autobiographical essay, this volume. 6. The only one among this group who eventually was in a position to repeat this is Mary R. Haas at Berkeley, who similarly has been the inspirational parental figure to a generation of Americanists of the 1950s and 1960s. Compare also Boas himself, of whom Sapir was a student among a group of devoted yet warring siblings. And observe the phenomenon of Roman Jakobson and his students moving from Columbia to Harvard en masse in 1949 (see Halle 1983:74). 7. This is what Hymes (1971) has called 'the first Yale school'. It is, perhaps, because the "second Yale school" succeeded in its turn in becoming the center of linguistic theory, explicitly founded on reverence for, and exegesis of, Leonard Bloomfield's Language (1933) and related works only after Bloomfield himself had removed to Yale in mid1940, that linguists' professional historical consciousness registers the impression of a succession, Sapir-then-Bloomfield. Of course, it was only the order of disciplinary hegemony of the latter's self-styled disciples that suggests succession; it is a succession of influence, or of mana. 8. Some of these characterizations are in fact ironic, since later writers — especially Bloomfieldians — have used them in negative evaluation of Sapir. Sapir says, for example, that Dr. Newman is a very exceptional man; I have never met anyone like him. He is a kind of mystic. ... He is strangely objective, and yet you can feel he has a lot of genuine feeling. ... I can't find the slightest trace of neurosis in him. ... I can see direct connection between his writing of poetry and [his] development of [Yokuts grammar in his description]. (National Research Council ms. [1935]) It should be pointed out that even Newman (e.g., 1951 e: 185 —6) felt obligated to apologize for many of these perceived characteristics of his master, in the changed Bloomfieldian climate of later years. 9. Compare also the fate of Sapir's Southern Paiute, completed in 1917 (except for some work on the dictionary part), but published — through Boas' efforts — only in 1930; or that of Lucy Freeland's Sierra Miwok, with introduction dated 1936 but publication date of 1951, through Carl Voegelin's efforts (rumor having it that the manuscript was languishing in Kroeber's office safe). In the days before Xerography, publication lag was much more deleterious to scholarly progress than now, of course. But fashions change as to what linguists consider genres of writing important enough to print on priority schedule.

"Singularly like our ideal of a scientist"

57

10. Some of his characterizations are wonderfully evocative: Yawelmani is characterized by a singular economy and consistency of form. ... In the formal [morphological — MS] hierarchy of the language, every element seems to be rigidly clamped into place: each suffix is allocated to a specific type of stem; each ablaut change conditioning the stem vowels is part of a symmetrical system of vowel changes; with its vowels altered according to an ablaut formula, each stem has its place among the set of stems formed from a given type of root; each root, in turn, has membership in the configuration of roots comprising a word class. (Newman 1946 b: 222) Some conception of the consonantal rigidity and the vowel plasticity of Yawelmani can be gained from the following ... (223) In its syntactic style, as a matter of fact, Yawelmani is even more loose-jointed than the formal description of its unambitious syntax might suggest, for the few devices offering a means of syntactic elaboration tend to be used sparingly. (246) ... there is a scrupulous and unremitting attention given to form for its own sake in Yokuts. Throughout its grammatical system Yokuts reveals a degree of neatness and consistency in organization, a formal balance and symmetry, that is rare among languages. Although this formal emphasis contributes feebly to the creation of [denotational—MS] meanings, it is by no means insignificant as a stylistic factor. Formal movements that take place below the level of tangible meanings carry their own esthetic satisfactions. But they are most subtle and indigenous aspects of language; they can never be captured in a translation. (1940 a: 5) The language does not permit its words to cut metaphorical capers. ... The broad area of reference covered by Yokuts words gives them a wide range of application, and their literal significance holds them austerely within their proper boundaries of reference. (6) Observe that this is a linguist's technical aesthetic, based on the analysis of structure that is then considered comparatively in a kind of connoisseurship of grammar and style. We must differentiate this from the aesthetics of native speakers innocent of technical models for language description, such as those Newman discussed in 1940 and 1967 d. Sapir, too, clearly had aesthetic preferences among languages he worked on, based on his contemplation of their grammars. For example, of Yana, Sapir wrote to Kroeber on 17 August 1907, that the virtually limitless number of verbal suffixes makes the complete mastery of Yana grammar practically an endless job. ... Some ... are so alarmingly special and concrete in significance that one hesitates to group them together with the more purely modal and local suffixes, of which alone the number is disturbingly large. ... [I]t is a very queer language and quite unlike what seems to be the most common Californian type (Yuki, Yokuts, Maidun). (Golla 1984:22 [letter 23]) Or again, he writes Kroeber concerning Uintah Ute (7 September 1909) that on the whole I must confess that I am more interested in Ute phonetics than morphology; the latter is not nearly as interesting as in the case of Takelma or Yana, for instance. It seems to be a somewhat gelatinous affair, leaving no decided impress on the mind. (Golla 1984:44 [letter 49]) 11. Some notable treatments in the 'Yokuts restatement industry' — as one wag has termed it — are Zellig Harris' famous IJAL review article (1944), "Yokuts structure and Newman's grammar"; Hockett 1967; Kuroda 1967; Kisseberth 1969; Archangeli 1984.

58

Michael Silverstein

See the bibliography in Hockett 1973 and the addenda thereto in Pullum 1973. Hockett (1973:64) speaks of such works as "products of this twenty-year period of analytical puzzlement and experimentation" with the facts as presented by Newman's work. 12. See the letter of solicitation from Sapir to Rroeber of 17 June 1935, asking for a Yuki sketch for the volume (which emerged as Hoijer etal. 1946) "intended chiefly for linguists in general who are not particularly well acquainted with American Indian linguistics or ethnology." See also his letter to Newman of 12 August of that year, encouraging him to finish his Yokuts sketch for the volume (Newman 1946 b): "So far I have only Morris' Eskimo in the docket." 13. The problem concerns what are phonetically the short and long versions of back mid rounded vowels, like German voll and English law. Viewed only in terms of distribution with respect to syllabic structure within words, these appear to be two "phonemic" units, with an anomalous, exception-riddled status in the scheme of otherwise regular vowel harmony and ablaut. However, viewed in terms of grammatically (semantically) identifiable paradigms of morphologically analogous forms, the long vowel sound [o-] represents two structurally differentiable units, o- and a-, and this distinction makes the vowel harmony and ablaut aberrances fall into place. (There are other interactions as well, on which see the sources in n. 11.) Thus, only as abstract conditioning environments for various phonological processes do the two otherwise phonetically identical vowels manifest themselves, not in any straightforward "surface" paradigmatic contrast, even in an adjacent phonetic segment. They "intersect" and "overlap" totally, in Bloch's (1941) terms. Observe that in the 1932 1JAL sketch, Newman speaks of the "phonetic patterning of Yawelmani" and writes o- and o- for the long units, noting that "Although o- and o• are the same phonetically, they are distinct phonemes; they are written differently to suggest this phonemic distinction" (1932 b: 86). The terminological usage and transcriptional practice are the same as Sapir's in his 1933 article. In the sketch from ca. 1936 (though we cannot be certain that there was not some editorial change by the time of publication as 1946 b), Newman speaks of ten vowel phonemes: i, i-, e, e-, a, a-, o, 3-, u, u-. These ten phonemes, however, do not have the same phonological status: i-, u-, and e are ablauted forms of root vowels and appear only in stems; the remaining seven vowel phonemes occur in roots as well as in stems. These seven phonologically fundamental vowels of Yawelmani are patterned as follows: Vowel Series 3 a i u Light a a i u Heavy o- a- e- dThe o- phoneme, it will be noted, plays two distinct roles in this configuration. It is a striking feature of Yawelmani that the only two phonological entities actualized in a single vowel phoneme (o- of the o series and o- of the u series) are maximally differentiated in their treatment (1946 b: 226). Note that "phonemes" in this usage are to be differentiated from "phonologically fundamental vowels." The first set are distributional units by criteria of word-boundaries and syllable structure; the second by criteria of distribution at various degrees of morphological abstraction — stem-structure and ultimately root-structure — based on paradigmatic tests. Note also that heavy u and heavy 3 are "two phonological entities"

"Singularly like our ideal of a scientist"

14.

15.

16.

17.

59

that are "actualized in a single vowel phoneme." (Compare the latter wording to Sapir's in 1925 [Mandelbaum 1949:40, n.2].) By the 1944 publication of the full Yakuts grammar, these entities had received both the distinctness of name, as "morphophonemes," and of transcription, where they appear as o- and o• respectively: In order to keep this distinction clear, this grammar will depart from a strictly phonemic orthography: the strong member of the d series will be written o-, but the strong member of the u series will be indicated by o-. (1944:20) It should be observed that there are many more interesting problems, during this crucial period of American linguistics, in the history of the theory of levels of abstractness in phonological structure in relation to methods of representing linguistic form. There are, for example, what have come to be termed "diacritic" (Kiparsky ms. [1968]) aspects of phonological representations in Sapir's methods, seen, for example, in the special juncture notations in his Southern Paiute (1930:62—63), in Sapir and Swadesh's Nootka (1939:236-239), and most interestingly, in Newman's Yokuts (1944:15-17), where three kinds of glottal stops, the segmental the "protective" glottal segment inserted intermorphemically (?), and the "floating" glottal stop of various suffixes' ... (that glottalizes resonants in second consonant position in stems), all are differentiated by special notations, some segmental, some diacritic. Were these all written as segments in some very abstract representational scheme, these also would be cases of absolute neutralization. See Edgerton's remarks (1933 [Hockett 1970:258-60, esp. 259 f.]) about the phonemic transcription in Bloomfield's 1933 Language, which remind one, in less negative form, of Boas' attitude (see, e.g., Boas 1947:208, n.5) towards the problem of the proper representation of phonemic vs. phonetic matters. Newman himself, in reviewing Boas' posthumous Kwakiutl Grammar of 1947, comments on the anti-phonemic transcriptional attitude (1948 a: 72). Bloomfield's well-known statements are, of course, prophonemic. I am greatly indebted to Regna Darnell for supplying me with photocopies of copies of these letters that Stanley Newman gave her in connection with her biographical project on Sapir. Sapir's very special use of the term psychiatric, which emerged with his growing concern for the relationship between the individual and sociocultural phenomena, must be understood by contrast with the inherited Boasian terms psychic and later psychological that have reference to the Völkerpsychologie tradition and the analysis of collective representations like language structure and the role they play in human consciousness. Sapir's usage in such titles as "The psychological reality of phonemes" (1933) is — and, alas, by current commentators has in fact been — rather seriously misconstrued by simply equating the adjective with current (or even coeval) professional usage within the psychological discipline. Cf. also Newman 1983:136. Mobilized by John B. Carroll, the Social Science Research Council sponsored a workshop at Cornell in 1951, consisting of three linguists and three psychologists (of whom Newman was one), that fashioned the influential and perhaps "paradigmatic" view of the emerging area in pre-"cognitive science" days; it led to the formation of the SSRC Committee on Linguistics and Psychology in 1952 that incrementally changed the intellectual landscape. See Carroll 1980:47—48 for his perceptions. Also cf. Osgood—Sebeok 1954 (reviewed by Newman 1955 a); and Newman 1957 c, 1963 b; Diebold 1964.

60

Michael Silverstein

18. As part of Newman's prodigious and diligent output of major book reviews, he wrote about both Sapir's and Whorfs Selected Writings, for UAL (1951 e) and Romance Philology (1957c) respectively. In the 1951 piece, Newman spends nine out of 13 columns talking about Sapir's interdisciplinary view of language, having in essence to defend Sapir's "venture beyond the strict confines of linguistics" in order "to seek new perspectives for the phenomena of language that would relate it to other forms of human behavior" (1951 e: 182), at the same time assuring his decidedly post-Bloomfieldian readers that Sapir, too, was as thoroughly committed as Bloomfleld to the view that a valid linguistic science must be a coherent and self-consistent body of concepts. ... [Sapir] did not try to select or construct a linguist's psychology, which, like a psychologists's version of linguistics, would be neither fish nor fowl, but a spurious body of doctrine irrelevant to both disciplines. This is not merely a matter of good manners, of the linguist's being gentleman enough not to muscle in on the psychologist's territory. It is the much more serious matter of recognizing that the consistency and integrity of any science must be maintained by the trained workers in that field, not by outsiders who may sometimes feel they can view the field with godlike omniscience (1951 e: 185). Things had been changing in the interim between the two reviews, as Newman's remarks in various places (e.g., 1954b, esp. 160; 1956a, esp. 87 — 88; 1957b, esp. 50) indicate. By the time Newman reviews the Whorf book — in the context of the decidedly more catholic Romance Philology, let us observe — he speaks of sectarian disciplinary narrowing of the 1930s and 1940s in confidently retrospective voice, celebrating the heroic and "stubborn persistence [with which Whorf] wrote and talked about his ideas" (1957 c: 43) and evaluating the collection reviewed as "a timely book, which should be read by those who wish to understand a growing trend in presentday linguistics" (44). 19. Compare n. 10. 20. Compare the parallel fate of two excellent articles — one on English stress (1946 a) and one on Bnglish suffixation (1948 b) — that also emerged only postwar from a collaborative project (with Morris Swadesh) on a "descriptive" grammar of English in 1934—36, sponsored by ACLS. Both appeared, with rather marked indexical value at that time, in the journal Word. To be sure, Swadesh, Newman's coworker and friend, was on the editorial board at that time. But perhaps more importantly, Language at this period of high post-Bloomfieldian ascendancy was hardly a place that welcomed approaches to structural statement such as these articles represent (see Newman's apologetic and historicizing note at the head of the article on suffixation, 1948 b: 24). And most of all, Language hardly welcomed phonemic transcriptions and other treatments of English that were not in accord with Bloch and Trager 1942 and later, SmithTrager derivatives, which by then had hardened somewhat into scientific dogma. To get some sense of the relations of the group around Bloch et al. to Swadesh, see Swadesh 1948. Unfortunately, Bloomfield himself may have given the next generation at least scholarly license (leaving aside the politics) for their treatment of Swadesh in his review of Swadesh's La nueva filologia in Language of 1943. That it stung very badly is seen in Swadesh 1948:255 — 56. 21. As does their son, David S. Newman (b. 1936), of Seattle, Washington, together with his wife and their children.

"Singularly like our ideal of a scientist"

61

References Allen, Harold B. 1980 A report for the archives of linguistics. In Davis-O'Cain 1980, 111-20. Archangeli, Diana B. 1984 Underspecification in Yawelmani phonology and morphology. Unpublished Ph. D. dissertation, MIT. Bloch, Bernard 1941 Phonemic overlapping. American speech 16.278 — 84. 1949 Leonard Bloomfield [obituary article], Lg. 25.87-94. [Reprinted in Hockett 1970, 524-32.] Bloch, Bernard—George L. Trager 1942 Outline of linguistic analysis. Special publication of the Linguistic Society of America. Baltimore. Bloomfield, Leonard 1933 Language. New York: Henry Holt. 1938 Eduard Prokosch [obituary article]. Lg. 14.310—13. [Reprinted in Hockett 1970, 347-50.] 1943 Review of La nueva filologia, by Mauricio [=Morris] Swadesh. Lg. 19.168-70. [Reprinted in Hockett 1970, 406 - 7.] Boas, Franz 1947 Kwakiutl grammar, with a glossary of the suffixes, ed. by Helene Boas Yampolsky, with the collaboration of Zellig S. Harris. Transactions of the American Philosophical Society, vol. 37, no. 3, pp. 200—377. Philadelphia. Carroll, John B. (ed.) 1956 Language, thought, and reality: Selected writings of Benjamin Lee Whorf. Cambridge, Mass.: MIT Press. 1980 The tale of a theoretical and applied linguistic psychologist. In Davis—O'Cain 1980, 3 1 - 5 2 . Davis, Boyd H. — Raymond K. O'Cain (eds.) 1980 First person singular: Papers from the conference on an oral archive for the history of American linguistics (Charlotte, N.C., 9 — 10 March 1979). Amsterdam studies in the theory and history of linguistic science, ser. 3, vol. 21. Amsterdam: John Benjamins. Diebold, A. Richard 1964 Review of Psycholinguistics: A book of readings, ed. by Sol Saporta. Lg. 40.197 - 260. Edgerton, Franklin 1933 Review of Language, by Leonard Bloomfield. Journal of the American Oriental Society 53.295-7. [Reprinted in Hockett 1970, 258-60.] Freeland, L. S. 1951 The language of the Sierra Miwok. Indiana University publications in anthropology and linguistics, Memoir 6. Baltimore. Friedrich, Paul 1974 On aspect theory and Homeric aspect. IJAL Memoir 28. [=IJAL 40(4), part II] Chicago: University of Chicago Press.

62

Michael Silverstein

Fromkin, Victoria 1977 Harry Hoijer [obituary article]. Lg. 53.169-73. Golia, Victor (ed.) 1984 The Sapir—Kroeber correspondence: Letters between Edward Sapir and A. L. Kroeber 1905—1925. Reports from the Survey of California and other Indian languages, no. 6. Berkeley: Survey ..., University of California. Gray, Paul E. et al. 1983 A tribute to Roman Jakobson, 1896-1982. Berlin: Mouton de Gruyter. Halle, Morris 1983 [Homage and reminiscence by] Morris Halle. In Gray et al. 1983, 72 — 75. Harris, Zellig S. 1944 Yokuts structure and Newman's grammar. I JAL 10.196—211. Hockett, Charles F. 1967 The Yawelmani basic verb. Lg. 43.208-22. Hockett, Charles F. (ed.) 1970 A Leonard Bloomfield anthology. Indiana University studies in the history and theory of linguistics. Bloomington: Indiana University Press. 1973 Yokuts as testing ground for linguistic methods. I JAL 39.63 — 79. Hoijer, Harry (ed.) 1954 Language in culture: Conference on the interrelations of language and other aspects of culture. Chicago: University of Chicago Press. Hoijer, Harry et al. 1946 Linguistic structures of Native America. Viking Fund publications in anthropology, no. 6. New York: Viking Fund. Hymes, Dell H. 1971 Morris Swadesh: From the first Yale School to world prehistory. The origin and diversification of language, by Morris Swadesh, ed. by Joel Sherzer, 228 — 70. Chicago: Aldine-Atherton. Hymes, Dell H.-William E. Bittle, Jr. (eds.) 1967 Studies in southwestern ethnolinguistics: Meaning and history in the languages of the American southwest. Studies in general anthropology, no. 3. The Hague: Mouton. Joos, Martin 1967 Bernard Bloch [obituary], Lg. 43.3-19. Kiparsky, Paul ms. [1968] How abstract is phonology? Mimeo, MIT Department of Linguistics and Modern Languages. Kisseberth, Charles W. 1969 Theoretical implications of Yawelmani phonology. Unpublished Ph.D. dissertation, University of Illinois, Champaign-Urbana. Kuroda, S.-Y. 1967 Yawelmani phonology. Reserach monograph no. 43. Cambridge, Mass.: MIT Press. Mandelbaum, David G. (ed.) 1949 Selected writings of Edward Sapir in language, culture, and personality. Berkeley: University of California Press. National Research Council, Division of Anthropology and Psychology, Committee on Personality in Relation to Culture, Subcommittee on Training Fellowships ms. [1935] [minutes]. In Archives, National Academy of Sciences, Washington, D. C.

"Singularly like our ideal of a scientist"

63

Osgood, Charles E. —Thomas A. Sebeok (eds.) 1954 Psycholinguistics: A survey of theory and research problems. Indiana University publications in anthropology and linguistics, Memoir 10. Baltimore. Pullum, G. K. 1973 Yokuts bibliography: An addendum. I JAL 39.269-71. Rigsby, Barbara — Bruce J. Rigsby 1975 Bibliography (1926-1974) of Stanley S. Newman. IJAL 41.399-405. Sapir, Edward 1925 Sound patterns in language. Lg. 1.37 — 51. [Reprinted in Mandelbaum 1949, 33-45.] 1927 a Language as a form of human behavior. English journal 16.421 —33. 1927b Speech as a personality trait. American journal of sociology 32.892—905. [Reprinted in Mandelbaum 1949, 533 - 43.] 1929 A study in phonetic symbolism. Journal of experimental psychology 12.225-39. [Reprinted in Mandelbaum 1949, 61-72.] 1930 Southern Paiute, A Shoshonean language. Proceedings of the American Academy of Arts and Sciences, vol. 65, no. 1. 1933 La réalité psychologique des phonèmes. Journal de psychologie normale et pathologique 30.247—65. [Reprinted, in English, in Mandelbaum 1949, 46-60.] Sapir, Edward—Morris Swadesh 1939 Nootka texts; Tales and ethnological narratives with grammatical notes and lexical materials. William Dwight Whitney Linguistic series. Special publications of the Linguistic Society of America. Philadelphia. Swadesh, Morris 1939 Edward Sapir [obituary article], Lg. 15.132-5. 1948 On linguistic mechanism. A centenary of Marxism, ed. by Samuel Bernstein et al., 254 — 9. New York: Science and Society.

Obituary Stanley S. Newman (1905-1984) Philip Bock and Harry

Basehart

Stanley Newman's sudden and unexpected death at the age of 79 has deprived anthropology of a rare spirit — one who was able to mediate between linguistics and ethnology as few people since Edward Sapir have done. His contributions to theoretical linguistics will be described and evaluated elsewhere. Here we shall sketch Newman's academic life and his influence on several generations of anthropologists, drawing in part on a brief autobiography that he prepared for publication (Newman 1984 a, in this volume). Born in Chicago on July 18, 1905, Stanley Newman grew up speaking Czechoslovakian and English: he learned German during an extended visit to Prague when he was eight, and he later studied Latin and French in school. He entered the University of Chicago in 1924 to study English and American literature, and he took classes there in Italian and Russian. Newman also wrote poetry and, for two years, he and Sterling North edited The Forge. It was at a meeting of the University Poetry Club that he met Edward Sapir (the only faculty member to attend), and he later took Sapir's course on the "Psychology of Culture." Newman finished his B. A. in 1927 and started graduate work in English at Chicago. His first academic job was as an instructor at the University of Texas in Austin, where for one "humbling" year he taught literature and composition. He returned to Chicago for additional classes with Sapir and soon was converted to anthropology: he did course work with FayCooper Cole, William Krogman, and Robert Redfield. At that time there was no department of linguistics at Chicago, but Newman's fellow students included future luminaries such as Mary Haas and Morris Swadesh (see Newman 1967 a). The summer of 1930 saw Newman's first field trip to California to study the Yokuts language. In Berkeley he met Alfred Kroeber and Paul Radin, after which he spent several difficult months collecting data on Yawelmani Yokuts that formed the basis of his dissertation. Back in Chicago for 1930 — 31, he married Lucy Price-Benedict. When Sapir moved to Yale the following year, Newman and his wife followed, along

66

Philip Bock and Harry Basehart

with several other advanced students. By the next year, Newman had completed his doctorate. He also published a paper (1933) on phonetic symbolism, using experimental techniques and statistical analysis; but at this point in the depression no teaching jobs were to be had, and he remained at Yale for the next five years, "subsisting on research fellowships" (1984 a: 11, in this volume). During this time, Newman worked on English grammar, phonetic symbolism, and the speech of psychotic patients. He also gathered data on different Yokuts languages and on Bella Coola. In 1937 — 38 he worked in New York City on a study of adolescent behavior and, with the encouragement of Franz Boas, analyzed part of his Bella Coola data — a task that still engaged him at the time of his death. Starting in 1939, Newman worked in the Intensive Language Program of the ACLS, preparing a series of lessons on spoken Persian which he later taught at Columbia. At the same time he became a (substitute) instructor of ethnology and general anthropology at Brooklyn College. When the war began, he worked under Henry Lee Smith in the Language Section of the Army Service Forces, where he prepared pocket-size Language Guides for use by American troops overseas. He was put in charge of courses in "Reverse English" (today, TESOL), and helped to prepare a course in spoken English for speakers of Spanish as well as manuals for German and Italian prisoners of war who wished to learn English. At the end of the war, Newman was transferred to the Institute of Social Anthropology of the Smithsonian Institution. He worked briefly with Julian Steward on the Handbook of South American Indians; then, with his wife and young son, David, he was sent to Mexico City to teach linguistics at the Escuela Nacional de Antropologia e Historia and to conduct research on Nahuatl and Otomi. This work was part of the "good neighbor policy" of the Roosevelt administration; but the changing political climate in the United States made such positions precarious, and after three productive years in Mexico, Newman accepted a position at the University of New Mexico in 1949. During his early years in Albuquerque, he taught a variety of courses in general anthropology as well as a series of courses in linguistics. Working with Indian students, he also undertook studies of two Pueblo Indian languages of New Mexico (Laguna Keresan and Zuni). In addition to anthropology students, he worked with graduate students in Spanish, Psychology, and in the School of Education. Newman spent many hours advising undergraduates and always took a serious interest in their

Obituary

67

problems. He worked closely with foreign students and with American Indians, acting for several years as adviser to the Kiva Club. During the 1950s, Newman was invited to teach at the summer institutes of the Linguistic Society of America (e. g., in Michigan and in Chicago). He also participated in an early conference of linguists and psychologists organized by John B. Carroll at Cornell, which he recalled as one of the most stimulating and intensive events of his career; later he supervised some pioneering psycholinguistic work at Zuni Pueblo. Newman's course on Language and Culture became a favorite of students at New Mexico, for he gently helped them to understand the relevance of linguistics to ethnological and archeological problems. His patience was legendary, and former students still chuckle at the memory of him holding up a class to explain over and over (to one who probably would never get it) some point about Penutian syntax or glottochronology. His classroom excellence was recognized when he was elected U.N.M. Teacher of the Year in 1966. Newman was not a man to seek disciples, but he had many devoted students (including Ed Ladd, Howard Maclay, Carol Stout, Morton Sloan, and Francesca Merlan) who acknowledged their deep intellectual debt to his teaching. With the rapid growth of the university and the department in the 1960s, his teaching responsibilities narrowed to more specialized linguistics courses. He continued his course in Language and Culture, though he was skeptical of some of the superficial methods then being employed. He also offered a seminar on Comparative Linguistic Structures (using his own materials on Penutian, Nahuatl, Salish, and Zuni); following Sapir's example, his students were required to write brief sketches of phonology and syntax to encourage them to grasp the basic patterning of languages. During the 1960s, Newman became a coeditor of the Southwestern Journal of Anthropology (now the Journal of Anthropological Research). Together with Harry Basehart he continued the policies established by the late Leslie Spier, evaluating and editing manuscripts in all fields of anthropology. Because of the journal's low budget, the editors also had to do all copy editing and to oversee production and distribution, but despite competition from a growing number of more specialized serials, they maintained very high standards. Newman enjoyed grappling with novel editorial problems raised by authors, including those posed by idiosyncratic writing styles that jarred with his ingrained training in language and literature. Administrative decisions and planning were of less interest, though his depression experi-

68

Philip Bock and Harry Basehart

ence made him acutely aware of the financial pressures endemic in hard times. Perhaps the most valued perquisite of his Journal activities consisted of the cancelled stamps obtained from business and editorial correspondence; Newman was an ardent collector, specializing particularly in U. S. issues. Stanley Newman was thought of by many as a conservative in his approach to linguistic analysis, yet one of us (Bock) vividly remembers the 1964 summer institute of the LSA at Indiana University when, in his series of exciting lectures on transformational theory, Noam Chomsky mentioned only one other living linguist: in an aside, he commented that "of course, Stanley Newman knew all of this years ago." Zellig Harris also believed that Newman could have been an important theorist if he had chosen to pursue that direction. When Newman's grammar of Yokuts was finally published in 1944, many readers considered it "too abstract." Yet it is today recognized as a classic and viewed as abstract "only in the simple sense of postulating an underlying vowel not realized as such in sound, and 'generating' or 'explaining' occurrent forms in terms of an underlying system, but keeping close indeed to the occurrent language when compared with the levels and leaps of linguistic models since Chomsky" (Dell Hymes, personal communication). Newman was initially sympathetic to the transformational-generative movement, but his doubts grew as some proponents increasingly ridiculed all other approaches and as a new generation of students with no sense of the history of the discipline and with an extremely narrow view of the nature of language came along. He was relieved to see, late in the 1970s, the return of a "healthier eclectic spirit" (1984 a: 23, in this volume). Despite a heavy teaching load plus committee and editorial duties, Newman maintained a steady productivity of scholarly reviews and writings. His Zuni Dictionary (1958) was complemented by a Zuni Grammar (1965); numerous articles dealt with more specific topics, but always in their full social context (e.g., Newman 1955). His research was honored by the university in 1967, when he was chosen to deliver the Annual Research Lecture at the university, later published as "Relativism in Language and Culture" (1967 b). This article (which is less available than most of Newman's publications) makes an elegant case for the differential use of linguistic resources in specific styles (using mainly Yokuts examples), and ends with a plea for relativism in approaching other languages and peoples. Stanley Newman retired from teaching in 1971, but he continued to frequent the Department of Anthropology and the library at UNM,

Obituary

69

working away at his Salish materials and taking on some editorial duties. He published a series of papers (e.g., 1976, 1979) in which he examined historical and areal features of Salish, focusing on the Bella Coola case in greatest detail. Throughout this time he conducted an extensive correspondence with U. S. and foreign scholars, frequently responding in meticulous detail to queries about aspects of his published work or field data. His contribution to linguistics was recognized when a special issue of the International Journal of American Linguistics was dedicated to him in 1975, and in 1984 he became President-elect of the Linguistic Society of America. Such honors were long overdue, but Newman's shyness and modesty were such as to prevent academic social climbing, and many at UNM never realized his true stature or his international prominence. Shortly before his death, Newman was invited to Toronto for the Sapir Centenary. He had prepared a paper on Sapir's contributions to psychological anthropology (1986) to honor his former teacher. But Newman's own academic life — his imaginative research and teaching, always grounded in a mastery of linguistic details and cultural patterns — was the best kind of tribute that a teacher can have. His students and colleagues will long remember this man of rare modesty, kindness, and imagination.

References Newman, Stanley S. 1933 1944 1955 1958 1965 1967a 1967 b 1976 1979

Further Experiments in Phonetic Symbolism. American Journal of Psychology 45: 5 3 - 7 5 . Yokuts Language of California. Viking Fund Publications in Anthropology, Vol. 2. New York: Wenner-Gren Foundation for Anthropological Research. Vocabulary Levels: Zuni Sacred and Slang Usage. Southwestern Journal of Anthropology 11:345 — 354. Zuni Dictionary. Indiana University Research Center in Anthropology, Folklore, and Linguistics, Vol. 6. Bloomington, IN. Zuni Grammar. University of New Mexico Publications in Anthropology, No. 14. Albuquerque. Obituary of Morris Swadesh. Language 43:948—957. Relativism in Language and Culture. New Mexico Quarterly 37:197 — 210. Salish and Bella Coola Prefixes. International Journal of American Linguistics 42:228 - 242. A History of the Salish Possessive and Subject Forms. International Journal of American Linguistics 45:207—223.

70

Philip Bock and Harry Basehart 1984 a 1986

How I Discovered Linguistics. Mimeo manuscript. In this volume. The Development of Sapir's Psychology of Human Behavior. Pp. 405—427 in New Perspectives in Language, Culture, and Personality. William Cowan, Michael K. Foster, and Konrad Koerner, eds. Amsterdam Studies in the Theory and History of Linguistic Science. Series 3, Studies in the History of the Language Sciences. Philadelphia: John Benjamins.

Stanley Newman and the Sapir school of linguistics Regna Darnell

Abstract Stanley Newman was one of the graduate students who followed Edward Sapir from Chicago to Yale in 1931. He was a key member of the research group which formed around Sapir in New Haven (along with Morris Swadesh, Mary Haas, Benjamin Whorf, Carl Voegelin, George Herzog, Walter Dyk, Murray Emeneau, Zellig Harris, and George Trager). Newman was the one of Sapir's students and associates who most nearly bridged the gap between his interests in linguistics and in culture and personality. Newman's initial specialization was in psychology of language and much of his work was done in collaboration with Sapir, who never wrote his own long promised textbook on psychology of culture. A close examination of Newman's work in the 1930s, then, sheds light not only on his own biography, but on the Sapir school of linguistics and on the interdisciplinary studies of language which flourished in this period and languished for many years thereafter.

Stanley Newman holds a virtually unique position among the students of Edward Sapir in that he worked seriously in both of Sapir's major areas of interest — American Indian linguistics and the nascent field of personality in its relation to culture. The two men collaborated closely in what was then called "linguistic psychology", with Newman quite independently following up Sapir's exploratory studies from the Chicago years (Sapir 1927, 1929) and both looking toward an interdisciplinary social science whose links would be primarily to psychiatry as represented in the interactional perspective of Harry Stack Sullivan. The research tradition which all three men envisioned was abortive, for reasons which included Depression cut-backs in research funding for the social sciences, Sapir's death in 1939, and the Second World War (which led culture-andpersonality studies into the characterization of whole cultures in an increasingly less isolationist North American culture). Nonetheless, both Newman and Sapir clearly took this work very seriously at the time. This paper will retrieve the context of that interdisciplinary effort, thereby

72

Regna Darnell

illuminating a little-known side of Stanley Newman and, simultaneously, clarifying the integration of the later thought of Sapir himself.1 The collaboration between Newman and Sapir took place in the context of what Dell Hymes, in an obituary of Morris Swadesh (in Hymes 1983) has referred to as the "first Yale school," centered around Sapir and preceding the better-known dominance of Yale linguistics from 1940 on by Bloomfieldian structuralism. When Sapir moved from Chicago to Yale in 1931, five of his students moved with him — Walter Dyk, Mary Haas, George Herzog, Stanley Newman, and Morris Swadesh. Research funding for work in American Indian linguistics and the impact of culture on personality was part of the package deal which lured Sapir to Yale's Institute of Human Relations to organize Rockefeller Foundation sponsored seminars on the latter topic. Fang Kuei Li, although his degree was from Chicago, was at Yale from 1937 — 39. Murray Emeneau, Carl Voegelin, George Trager, Charles Hockett, and Benjamin Whorf joined the Sapir group at Yale during Sapir's tenure there. Zellig Harris was unofficially part of the inner circle around Sapir after they met at the 1937 Ann Arbor Linguistic Institute. After Sapir's death in 1939, most of these individuals came to terms, in one way or another, with the changing emphasis of linguistic science, particularly with the structural and anti-"mentalist" tenets of Bloomfieldianism. Research in American Indian linguistics continued, but Sapir's former students did not challenge the dominant paradigm of the period. The result is a strange hiatus in the history of linguistics, eclipsing Sapir's processual approach to language structure (perhaps most elegantly exemplified in Newman's 1944 The Yokuts Language) and insistence on the necessity for links between linguistics and other social sciences, particularly psychology and psychiatry. Sapir's students have been remembered as "linguists" in a much more narrow sense than that of Sapir himself. The broader context for the study of language, to which Sapir devoted most of his later career, has been effaced in the process. In the context of current directions in linguistic theory, however, there is much to be gained from reconstructing what it was Sapir and Newman wanted to study and how they proposed to go about it. Sapir and Newman operated in a mutual admiration society, based on both personal affection and intellectual style. Sapir's comments about Newman to a subcommittee of the National Research Council on training fellowships for the interdisciplinary study of culture and personality (21 December 1935: NRC) were designed to obtain funding for Newman's

Stanley Newman and the Sapir school of linguistics

73

work but clearly reflect the enthusiasm Sapir felt for the work of his young colleague: Dr. Newman's career started off with the humanities. When I first came across him at the University of Chicago a number of years ago, he was concerned with literature. He wrote very acceptable poetry and was working in the Department of English. ... For reasons which he could probably explain better than I, he ultimately found himself dissatisfied with the traditional academic training interests that he was in contact with in the Department of English, and turned to anthropology. He took a greal deal of work from various men at the University of Chicago and did very well. At the same time he took up work in linguistics, chiefly Russian, and developed a great interest and rather remarkable technique in that regard. His first field trip was a study of the Yokuts language of South Central California, and although that language had been studied in a preliminary way by Dr. Kroeber he was able to make a new study of it which completely supercedes the older work. The grammar which was eventually his thesis at Yale University is practically finished now, in revised form, and is, I think, perhaps the most beautiful — perhaps that is not the word some of you might use — story of an American Indian language that has ever been written, beautiful not only because the language is beautiful, but also because the treatment is highly balanced. I mention this to show that Dr. Newman is always aesthetic even when he seems most technical and formal. I can see direct connections between his writing of poetry and this development of the American Indian language. After his first field trip, he took up further work among the Yokuts, covering, I think, six dialects in all and then worked in B. C. He hasn't had time to work up this material as yet. Sapir explained that Newman had found himself "gradually getting more and more interested in the psychological problems affecting language" and "the unconscious symbolism of vowels and consonants, a theme I had developed somewhat." Newman learned statistics (producing an "interesting" paper), although Sapir was highly skeptical of statistical analysis because of its inability to handle the role of the individual in culture. At Yale, Newman had pursued "social psychology, whether linguistic or not, speech being his chief theme." Sapir had been able to obtain a two-year Social Science Research Council fellowship for that work.

74

Regna

Darnell

Newman was currently working with Morris Swadesh "on a new English grammar" for the American Council of Learned Societies (1934—36), "the idea being to profess complete ignorance of English grammar and take it from the ground up as though it were an unknown language." Newman had unearthed "many original things" in the stress patterns of English. Swadesh (personal communication, Mary Haas) considered the English grammar project primarily as a means for funding his American Indian linguistic work. Newman, in contrast, was intrigued by Sapir's efforts to apply the discovery procedures of American Indian fieldwork to a written, well-studied language. The project had psychological as well as linguistic implications.) Sapir stressed that Newman was "a great man for patterns." He felt "the relations of things, not merely the facts in a sensory sense." (Sapir's own recent work on the phoneme (1925, 1933) had crystallized his longstanding concern with form and pattern in language. This was high praise indeed.) The way Dr. Newman came into the particular field we are discussing is as follows: I will begin with myself. For years I have been interested in the possibility of extending linguistic studies to include the psychological factors involved, and it seemed to me that a very careful study of all the unconscious personalities in the speech situation would be very revealing not only to the linguists but perhaps even more to the psychologists. ... it occurred to me that (Newman) might be a man to have in mind for personality studies with a speech theme, and I asked him whether he would be interested in tying up with a psychiatrist who was interested in the larger aspects of personality formation and had in mind the cultural point of view. Newman had agreed and Sapir sent him to Harry Stack Sullivan (also present at the meeting in 1935). Newman was presently "enthusiastic" about the possibility of funding for a training analysis, then just beginning to be considered a sine qua non in psychiatric training. Sapir was concerned only that the analysis would not provide sufficient material, since Newman was "a thoroughly normal person" without "the slightest trace of neurosis." Further, Newman was willing to postpone his longer-range ambitions to acquire the training for serious work in the interdisciplinary field between linguistics and psychiatry. Sapir went on to speculate about Newman's personality: "I have never met anyone like him. He is a kind of mystic." The strongest criticism was a certain lack of practicality. Newman was married and had a family to

Stanley Newman and the Sapir school of linguistics

75

support. Although he had finished his Yale degree, "he doesn't seem to realize that it is up to him to try to get connected with something immediate." Newman followed instructions to apply for things and continued his work on the resulting grants. His interests were not political but intellectual: He is quietly taking things in and has no animosities that one can see. He is strangely objective, and yet you feel he has a lot of genuine feeling, you feel he would like nothing better than to be let alone and given something intensely interesting, preferably of an aesthetic type, to do. He wants to understand people. Sapir went to a great deal of effort in this period to get Newman an inside track on what he envisioned as the research program of the future. Unfortunately, the National Research Council declined to finance training analyses for anthropologists to learn psychoanalysis and then study personality variation cross-culturally in the field. The scheme, though supported on the National Research Council committee by Johns Hopkins psychiatrist Adolph Meyer, was largely a product of the joint theorizing of Sapir and Sullivan. Meyer was, in fact, impressed by the discussion of Newman, writing to Sapir (23 December 1935: NRC): I am particularly fascinated by the necessity of drawing history and linguistics as obligatory considerations into the field of the training of our own men, although perhaps a little more from the point of view of principles and opportunities for organizing material than in any hope that the amateur can do any good to the fields himself. Psychiatry was just emerging as a discipline with standardized professional credentials, among them the training analysis, and anthropologists could not, in Meyer's view, be permitted to challenge the elitism of psychiatry. Sapir also sent Newman to see Sullivan and Henry Murray of Harvard, a more scientifically-oriented student of personality. He clearly expected that Newman would prefer the intuitions of Sullivan to the more formal methods of Murray (to Newman, 12 August 1935): Harry Stack is not a systematic thinker and he is a hornet's nest of prejudices but I know few people who have as keen perceptions of a personalistic character as he. He's really a sort of involved, lumbering, sensitive, inarticulate artist on the subject of personal relations. I feel you ought to get a lot out of him — more, probably, than out of Murray, who has systematic ideas and a comprehensive viewpoint but

76

Regna Darnell

suffers, I guess, from the system-maker's disease. The two together should be worth your while. In spite of Sapir's prestige in interdisciplinary social science circles and among granting agencies, however, funds were drying up and academics were defending their existing empires rather than expanding into new terrain. Newman's retrospective account of Sapir's "psychology of human behavior" (in Cowan etal. (eds.) 1986:405) began with his interests in music and poetry, reflecting Sapir's very early "dissatisfaction with conventional treatments and his restless search for more meaningful treatments." When Mary Haas was planning a biography of Sapir, Newman wrote to her (16 November 1971) about the power of the aesthetic in his view of Sapir: It you're planning to deal with the multiple interests that Sapir had, would you consider including some of his poetry? I may be prejudiced on this matter because my first meeting with Sapir was at the Poetry Club at the University of Chicago, where his remarks on the poetry being read seemed to me more incisive and apt than the things I'd been hearing from my English profs. (At the Poetry Club the meetings were conducted by the president's reading the unsigned poems written and submitted by the members, and after each poem those present would go at it hammers and tongs. We had some good arguments. Sapir, as you can imagine, entered into the spirit of things with his usual zest.) But he did write a lot of poetry and was serious about it. Certainly his experience with word polishing in poetry influenced his prose style. One of his poems is called "Zuni." Maybe there are others with anthropological relevance. It was not, however, entirely because of anthropological content that Newman proposed poetry as part of the essential Sapir. They shared this concern for the aesthetic. Newman's respect for the intellectual breadth of Sapir is reflected in various of his writings. In a review of Selected Writings of Edward Sapir in 1951, Newman emphasized that the "absorbing interest" of Sapir's life went beyond his descriptive American Indian studies to seek "new perspectives for the phenomena of language that would relate it to other forms of human behavior." The concept of the phoneme, for example, had intrigued Sapir "not so much as a methodological tool for the linguist, but rather as a powerful and clear demonstration of the unconscious patterning of human behavior" (1951 in Koerner (ed.) 1984:61).

Stanley Newman and the Sapir school of linguistics

77

This review devotes a great deal of attention to Sapir's prose style, honed by poetry but increasingly controlling the "evocative overtones" of his topic. Unlike most academic writers, Sapir wanted to catch the imagination and attitudes of his audience, therefore alternating humor and sober argument. Sapir's use of language was "holistic" — "an attempt to write for the reader as a person rather than as a disembodied intellect" (1951:63). Newman believed that language had become, for Sapir, less subject matter than method to approach something else: "In short, language provided the clearest and most easily described evidence of the fundamental human tendency to mold behavior into unconscious patterns of form" (1951:63). Newman lamented the selection of Sapir's work being divided into language, culture and personality segments, because Sapir himself did not acknowledge such boundaries and they detracted from the ability of readers to see his thinking as a whole. In 1951, it was necessary to relate Sapir's work to that of Bloomfleld. Newman stressed that Sapir had wanted linguistics to become more "cosmopolitan" among the sciences, but that he also shared Bloomfield's commitment to the development of an autonomous level of linguistic analysis. Indeed, Sapir acknowledged that interdisciplinary work legitimately rested in the specialized competences of its various contributors. Thus, language, with its "complete detachment from other types of cultural patterning" had enormous potential for both Gestalt and experimental psychology, the evidence being the phonetic symbolism studies (though Newman does not mention his own contributions to these) (1951:64). Newman was not alone among Sapir's students in arguing implicitly that Sapir did what Bloomfleld did, but also went beyond him in the scope of linguistics. This argument did need to be made and the tone is often apologetic, with Sapir's students trying to see the validity of both approaches. (See, for example, Zellig Harris's review of Selected Writings of Edward Sapir, reprinted in Koerner (ed.) 1984.) Newman's final reassessment of Sapir's contributions to linguistics and psychology (1986) avoided reference to his personal contacts with Sapir. He argues on textual evidence that Sapir had been interested in the individual, culture, personality and language at least since 1917 when he argued against the common anthropological concept of culture as superorganic. The "scientific psychology" of his day "left Sapir cold" (1986:409) because it couldn't account for the submerged formal systems (largely linguistic) or elaborated symbolic constructs which characterized both individuals and cultures.

78

Regna Darnell

Newman elegantly clarifies the terminology Sapir was using in this period. Increasingly, he preferred the term "psychiatry," in a non-medical sense of total personality in relation to behavior and culture; "psychology" he reserved for studies which segmented behavior and treated it statistically (Newman 1986:422). Sapir did not trust statistics. Newman wrote to Sapir's son Philip (29 March 1984) about his puzzlement over this apparent indifference to the bandwagon of the Chicago sociology department in 1930: I didn't know Edward Sapir very well. I'd shifted from English ... to Anthropology only the year before. In talking to me in his office one day, he gave me a copy of his "A Study in Phonetic Symbolism." He asked me to read it and see if I'd like to carry on with the study. At the time, I was taking a course in statistics from L. L. Thurstone, a young Prof, in the Psychology Department, who later became one of the major figures in American statistics. I had read one of Thurstone's recent papers, in which he used an experimental technique very much like the one in Edward Sapir's article. ... I was very enthusiastic about this clever statistical device. When I told Edward Sapir about Thurstone's statistical procedure and said that I'd like to apply it to further work with phonetic symbolism, I was disappointed that he didn't share my enthusiasm. In fact, he seemed rather cool toward the idea. But he didn't raise any objections, so I continued the phonetic symbolism experiments using the Thurstone statistics. Not until Newman, while preparing his paper for the Sapir Centenary Conference in Ottawa in 1984 (Cowan et al. (eds.) 1986), read transcripts of the Second American Psychiatric Association Colloquium on Personality Investigation (1930) did he fully understand Sapir's principled objection to Thurstone's emphasis on scientific laws as the priority for the social sciences. Until then, it still puzzled Newman why Sapir used such an elaborate statistical design, "his first and also his last venture into statistics." Interestingly enough, Fred Eggan (1974) also attempted to interest Sapir in Thurstone's statistics when he was a graduate student at Chicago — and was met with the same lack of enthusiasm Newman describes. The explanation is to be found in Sapir's changing audience for his theoretical writing after his move to Chicago in 1925. He was talking to psychologists, psychiatrists and sociologists on the interdisciplinary conference circuit (Darnell 1986), who equated the objective, the formal

Stanley Newman and the Sapir school of linguistics

79

and the statistical with the significant. At the first American Psychiatric Association colloquium in 1928, Sapir was very quiet throughout the discussions, though he did try to summarize the range of disciplinary views in a way that would make them potentially compatible. The following year, however, he chose to discuss his own forays into the experimental (1927, 1929). Clearly he intended to establish his credentials as a spokesman for the synthesis all participants hoped for. He stressed the "scientific" character of his work in linguistics and refused to be placed in the role of the anthropologist expert-on-the-exotic with a bizarre example for the theoretical points of others. This battle for the status of research in the social sciences had to be fought on alien ground; Sapir was definitely in a minority. In "Speech as a Personality Trait" (1927), Sapir noted that people always produce intuitive judgments of the personalities of others on the basis of their speech. He defined five levels of such potential evaluation, ranging from voice characteristics to stylistic arrangement and proposed that the individual and the culture had to be differentiated at each level. The result would be "a valuable lever in psychiatric work" (Newman 1986:418). Sapir did not propose how this differentiation might be made and would have been hard-put to demonstrate even potential application of this cumbersome scheme to clinical diagnosis. In "A Study of Phonetic Symbolism" (1929), Sapir argued that linguistics was useful for the analysis of speech because of its established methodology. His experiment presented five hundred Chicago high school students with a hundred "non-sense words" and measured their meaning changes when one sound was modified. He was dissatisfied with this methodology (acknowledging the help of a colleague in psychology, not Thurstone) because it obscured significant individual variation in response which he found more intriguing. When Sapir reported this research to the second American Psychiatric Association conference, he stressed a further experiment in which a single subject responded spontaneously to changes in sound with changes in meaning. Sapir was trying to study unconscious symbolism rather than surface meaning, and he saw this as integrated for each individual subject. Sapir quickly dropped this experimental stance, leaving it to Newman to pursue in more detail. Nonetheless, Newman (1984:421) attributes to the American Psychiatric Association colloquia of the late 1920s Sapir's compulsion, in collaboration with Harry Stack Sullivan, for "planning interdisciplinary projects that would permit making precise observations of interpersonal behavior." Unfortunately, the Depression was well

80

Regna Darnell

underway and money was not available to Sapir and Sullivan "even for their research projects" (1986:422). Newman, of course, was to suffer repeatedly from the vagaries of funding and felt that the research he did in the 1930s was determined far more by the vicissitudes of what was available than by the evolution of his own interests; he did not continue with the linguistic psychology work because he could not support himself by doing so (personal communication). The next follow-up of these interests was McQuown (1957) who cited Sapir but not Newman in arguing that psychiatry might eventually expect diagnostic help from the linguist. There is no evidence of a continuous research tradition from Newman to current emphasis on the psychiatric interview as a method in social science. In fact, during this whole period, Newman moved back and forth between American Indian linguistic fieldwork and analysis and the speech psychology research. Newman's Yokuts work was sponsored by the Committee on American Indian Linguistics dominated by Franz Boas, although Sapir and Leonard Bloomfield were nominally his equals in decision-making. Boas had felt strongly his responsibility as self-appointed mentor of salvage linguistics at least since he sent Edward Sapir to study Takelma in 1906. Boas's patience for Sapir's forays into interdisciplinary social science was distinctly limited, and his irritation extended to Newman on occasion. In 1933 when Newman was working on Bella Coola, Boas wrote to Sapir (11 December) protesting Newman's obligations to the English grammar project and his grant applications in language psychology that prevented him from writing up the Amerindian material. He wanted Sapir to "arrange" for the Bella Coola to be completed. Sapir's handwritten note on this letter declines to accept his former teacher's priorities for Newman, though pacification was in order: "Could you fix up a brief statement that I could pass on to the old man? Perhaps an engagement to put in half an hour, before going to bed, on Bella Coola?" Both Sapir and Newman wanted the work done, but Newman could not afford to complete it without salary from some other source. The Committee on American Indian Linguistics could only support fieldwork, so the work had to be sandwiched between other commitments (Leeds-Hurwitz 1985). In his text contribution to Anna Gayton's collection of Yokuts and Mono Myths (1940:4), Newman expressed the aesthetic dimension of his Yokuts grammatical work, while presenting only a sample of the material he had collected:

Stanley Newman and the Sapir school of linguistics

81

In translating we come to the unhappy realization that each language, instead of shaping itself to our will, governs and directs the trend of our expression. We are sharply reminded that languages have an inner resistance. Their materials are already shaped into a system of formal and conceptual patterns. Within the patterns of a language other than our own, we are forced to make uncongenial distinctions that seem imperative to us. Newman went on to discuss the grammatical "potentialities of style" in Yokuts and to emphasize that the traditional (Boasian) notion of an American Indian grammar did not come to grips with the expressive power of a language or its psychological reality for speakers. A grammar (1940:4): ... tells what a language can do but not what it considers worth doing. To the native, a grammar is always unconvincing, for it ignores the most vital and ultimate part of his language — the intricate network of values, of attitudes and expectancies that guides his selection of expressive tools. The Yokuts "feeling for simplicity," "scrupulous and unremitting attention ... to form for its own sake," "formal balance and symmetry ... rare among languages" were as important as the grammar in the more traditional sense. Like Sapir, and Whorf, Newman valued in a language the "collective quest" for stylistic integrity which was virtually untranslatable (1940:8). Admittedly, such statements were excluded from grammars (cf. Newman and Whorf in Hoijer (ed.) 1946, a collection of grammars by Sapir's students but appearing some years after his death). Whatever the topic at hand, it was framed in the Sapir view of language as indivisible from its context and use. Like Sapir, Newman was not above purple prose. In the memorial volume for Sapir (Spier etal. (eds.) 1941:94), for example, Newman addressed the "incredulity" of the average man's conception of linguistics (in a paper reporting on experiments in linguistic psychology): Those who have not been initiated into the somewhat esoteric concepts and methods of linguistics frequently express their incredulity at the picture of language presented by linguists. They find it very strange, for example, that a grammatical description can get along without any reference to the persons using the language. ... As naive observers, they see language as it takes place in the context of human behavior, where it is merely one of the expressive and communicative activities

82

Regna Darnell

of human beings in their everyday business of living. But the linguist seems to regard language as a huge autonomic mechanism and the language-using person as a passive thing coerced by the relentless operations of phonemic patterns, morphological processes, syntactic configurations, conceptual categories, and all the other paraphernalia of linguistic systems. Linguists were too often begrudging of the fuller picture of language. Moreover, the tendency of linguistics to define itself in terms of "isolated methodology" rather than the subject matter of language itself was to be regretted. The English grammar work had made it clear that the linguist could not ignore usage (1941:95). Tone and other language phenomena: ... undeniably intrude themselves upon the structure of English, and they are therefore within the province and responsibility of the linguist, even in his most restricted capacity as a student of formal linguistic structures. Newman proposed to make a start by studying the deviations in linguistic aspects of adolescent speech behavior. This work was sponsored by the General Education Board in New York City, but Newman approached it with the holistic view imbibed from Sapir. In fact, he considered this study of the personality problems of individual adolescents to be "the closest I've done to the kind of interest that Sapir had in the Yale years" (Newman to Stephen Murray, 31 May 1984). Newman's language psychology work was loosely based in the Yale Institute of Human Relations from 1932 — 37, although Sapir obtained various outside funding for much of that period. Mark May, Director of the Institute, was less than enchanted with the research program, failing to understand how Sapir and Newman wanted to approach "the more realistic and practical problems of speech" (Newman to May, 6 September 1933: YU). But since Newman had a fellowship renewal from the Social Science Research Council, May's objection was essentially irrelevant. Newman suggested, in his proposal for the study of speech behavior that the psychologists had avoided speech almost entirely, except in the most superficial impressionistic way, and had failed to utilize the techniques of linguistics to their potential benefit. Linguists, in turn, had failed to study "the realistic factors involved in language as a type of human behavior." The clinical observations might later be useful for psychiatric diagnosis. (Certainly the results tended to confirm diagnosis, though clinicians were usually minimally concerned with experimental

Stanley Newman and the Sapir school oj linguistics

83

controls.) Newman proposed that the applications for the study of individual speech might be pursued into several empirical areas: 1. Neurotic individuals tended to have "more extreme and conspicuous" speech characteristics than normal individuals, providing a rich research site with independent evidence from clinical treatment. 2. Newman proposed to teach English to foreigners by "adapting the teaching method to the individual situation," including attitudes toward the two languages. Culture, as Sapir insisted, had its locus in the individual. 3. English composition had been taught on an intuitive and haphazard basis. Newman wanted to describe the characteristics of acceptable prose style in "an objective, non-literary way." Writing classes were compared to a clinic where experiments could be conducted into corrective methods. The use of the psychiatric model for modification of normal behavior through learning is extremely interesting in this proposal (YU). May was unable to understand how the speech analyses were to be related to personality. He also had difficulty with the notion that "unconscious mechanisms reveal themselves in speech phenomena" (May to Newman, 27 April 1937: YU). The unconscious, for May, had nothing to do with linguistics. Part of the difficulty was that Sapir and Newman took these matters for granted, while May, then heavily influenced by the behaviorism of Clark Hull and attempting to organize Institute research around its integration with Freudian psychology, saw no way to provide "objective" proof of such fuzzy concepts. The positions were simply irreconcilable. In an undated version of his proposals (YU), Newman argued that a clinical study could not be bound by "an over-scrupulous insistence on a stereotyped approach." I. e., the features of speech which were significant in one case might be completely irrelevant in the next one. His method consisted of describing whatever proved to be relevant in a particular case. Ultimately, the study would lead not to larger samples for verification and replicability but to larger questions: The isolation of... fragments of speech behavior is frankly an artificiality, imposed for the purpose of developing an analytic technique. Even beyond this, the investigators fully appreciate that speech behavior is, in its turn, merely an arbitrarily delimited sphere within the larger context of interpersonal behavior. In response to May (31 March 1937: YU), Newman stressed that manifestation in speech of "unconscious mechanisms" was "my methodological assumption." Confirmation would come not from particular iso-

84

Regna

Darnell

lated details, but from "a study of general tendencies actualized in a group of details." That is, Newman was applying a notion of patterning derived from the phonemic model in Sapirian linguistics. His justification for the use of linguistic rather than other behavioral data was the "greater refinement" of methodological tools already available than for "any other type of overt behavior." Behavior was taken to be inherently symbolic, and language to be the key to the symbolism. Even the distinction between normal and abnormal had to be subordinated to that of "how unconscious mechanisms can, on the one hand, be expressed or disguised within the limits of culturally acceptable forms and, on the other, be actualized through cultural deviations." Beginning in 1937, Newman collaborated with Vera Mather, a psychiatrist trained by Adolph Meyer, in studying patients at the New Haven Hospital associated with Yale. The aim was to aid in clinical diagnosis by offering more precise terminology for speech aberrations which could be associated with particular affective states. The sample of forty cases was treated in a case history rather than statistical manner. In spite of the collaboration of the researchers from two disciplines, however, the study did not lead psychiatrists to linguistic studies. In 1937, Newman accepted a position with the General Education Board in New York City, finally removing himself from his affiliation to the Institute of Human Relations. He wrote to May (8 May 1937: YU) that he was eager to attack language problems in secondary schools and "get some realistic notion of the range and types of language problems encountered in such an applied field as education." Newman described "Personal Symbolism in Language Patterns" (1939:177) using his case material from the adolescent study to show "... how an individual's use of language, like his expression in other forms of interpersonal behavior, is symptomatic of his functioning and his adjustments as a person in a cultural setting." The paper dealt with written compositions of individual students and clearly addressed a practical need of the school system. Nonetheless, the rhetoric in which Newman describes the "depersonalized approach to human behavior" in the academic disciplines is straight out of Sapir and Sullivan. Moreover, the paper appeared in the second volume of Sullivan's journal Psychiatry. The adolescent study ended in 1939 and Newman described his situation as "desperate" (to Murray, 31 May 1984). Because he had gotten to know Boas during his two years in New York, he was supported briefly to write up his Bella Coola material. Although this was Newman's perception of the matter, it seems clear that Boas felt an obligation to

Stanley Newman and the Sapir school of linguistics

85

the Indian linguistic work and to a student of Sapir's; it is typical of him that personal acquaintance with Newman would have been somewhat secondary to these larger professional obligations. In any case, Newman eventually accepted an American Council of Learned Societies position intended to contribute to the war effort, and worked on Iranian in New York City, describing in retrospect (to Murray, 31 May 1984) his choices at the end of the Depression as "ruthlessly limited." Newman himself reflected on the lasting importance of the language psychology work (to Murray, 31 May 1984), distinguishing his own position from that of his mentor. Newman had: ... as much faith as Sapir did that the individual personality develops in contact with other people and has continuity through time. My main difference with Sapir (based, of course, on our different temperaments and histories) is that I would worry more about objective method and reliability. ... I sometimes felt uncomfortable in the adolescent study, feeling that I had reached conclusions in a kind of imaginative, freewheeling flight, more appropriate and valid for writing poetry or for doing applied work, like treating psychiatric patients, where the practitioner is forced to make decisions as best he can. I would have preferred to be part of a more elaborate interdisciplinary research team. An analysis of an individual's language would give only part of the picture; it would be more useful if combined with a life history. ... It is significant, I think, that when I took over Sapir's phonetic symbolism study, I concentrated on the statistical approach in his paper and improved the methodology. But as I got to know him better and read his paper again later in the light of his other writings, I realized that he had a profound dislike of statistics ... and would have preferred for me to continue with the more open-ended approach. Much of Newman's reassessment of the language psychology work seems to have revolved around the importance of Sapir as his teacher and mentor, both professionally and personally. In retrospect (Newman to Haas, 16 November 1971), Newman stressed that Sapir had consistently used his professional contacts on behalf of his students. "On looking back, I realize that I didn't appreciate the loyalty and generosity that Sapir showed to his students." In his own case, Sapir had handed over his experimental data on phonetic symbolism; Newman had thanked him for the problem and experimental approach only in "a perfunctory footnote." When Swadesh and Newman had failed to produce a book on English grammar, Sapir "should perhaps have taken a whip to us,"

86

Regna Darnell

but instead he "ran interference for us" and got another grant to continue their own work. Sapir apparently did not expect any more elaborate thanks. In any case, over the years of Newman's career, he became increasingly aware of the broad and less easily characterized influence of the first Yale school of linguistics, and particularly of the personality and intellect of Edward Sapir, on his thinking. Newman never returned to language psychology in the Sapirian sense and this early work has been largely ignored by scholars who know him only as an Americanist. The grand dreams of an interdisciplinary social science proved to be built on flimsy ground indeed, and disintegrated entirely with Sapir's death, having depended on his synthetic and charismatic abilities (as well as his reputation for these, which obtained funding for work in language psychology). Unsurprisingly, given its systematic lack of interest in language, Newman was never particularly intrigued by the culture and personality work which coalesced around Ruth Benedict, Margaret Mead, Ralph Linton, and Abram Kardiner during the war years. In terms of his own career, language psychology was a dead-end for Newman. Nonetheless, his broad approach to the nature of linguistics and social science finds its roots here. Moreover, increasingly, in the developing histories of both anthropology and linguistics, the scope of the Sapirian tradition is again intellectually congenial.

Note 1. This discussion is drawn from material on Sapir for a biography now in press. I never met Stanley Newman but I had extensive correspondence with him during the year before his death. I came to appreciate both his personal kindness and the scope of his intellect. More than any other of Sapir's students, he was able to understand the essential irrelevance of disciplinary boundaries. It is a great pleasure to pay tribute to him through this paper. I have profited from discussions of Newman's language psychology work with Judith Irvine, Stephen Murray, James M. Nyce, Richard Preston, Philip Sapir, and Michael Silverstein. Murray (1986) is the best treatment available of this work, which he discusses in relation to Chicago sociology. None of these colleagues, of course, are responsible for my interpretations. I have cited documents from the National Research Council (NRC), Washington, D. C. and Yale University Archives (YU).

References Cowan, Willian — Michael Foster—Konrad Koerner (eds.) 1986 New perspectives in language, culture and personality: Proceedings of the Edward Sapir centenary conference. Amsterdam and Philadelphia: John Benjamins.

Stanley Newman and the Sapir school of linguistics

87

Darnell, Regna 1986 Personality and culture: The fate of the Sapirian alternative. In G. W. Stocking, ed., Malinowski, Benedict and others. History of Anthropology 4:156-183. In press Edward Sapir: Linguist, Anthropologist, Humanist. Berkeley: University of California. Press. Eggan, Fred 1974 Among the anthropologists. Annual Review of Anthropology 3:1 —19. Harris, Zellig 1951 Review of Selected writings of Edward Sapir. Language 27:288-333. Reprinted in Koerner, ed. 1984:69-114. Hoijer, Harry (ed.) 1946 Linguistic structures of native America. Viking Fund Publications in Anthropology 6. Hymes, Dell 1983 Morris Swadesh: From the First Yale School to World Prehistory. In Essays in the History of Linguistics, ed. by Dell Hymes. Amsterdam and Philadelphia: John Benjamins: 273 — 330. Koerner, Konrad (ed.) 1984 Edward Sapir: Appraisals of his life and work. Amsterdam and Philadelphia: John Benjamins. Leeds-Hurwitz, Wendy 1985 The Committee on Research in Native American Languages. Proceedings of the American Philosophical Society 129. McQuown, Norman A. 1957 Linguistic transcription and specification of psychiatric interview materials. Psychiatry 20:79-86. Murray, Stephen 1986 Edward Sapir and the Chicago school of sociology. In Cowan etal. 1986:241-292. Newman, Stanley (with Vera Mather) 1938 Analysis of spoken languages of patients with affective disorders. American Journal of Psychiatry 94:912 - 9 4 7 . 1939 Personal symbolism in language patterns. Psychiatry 2:177—182. (with Anna Gayton) 1940 Yokuts and Western Mono myths. Anthropological Records (University of California) 5 (1). 1941 Behavior patterns in linguistic structure: A case study. In Spier et al.: 94—111. 1944 The Yokuts language of California. Viking Fund Publications in Anthropology 2. 1951 Review of David Mandelbaum, ed., Selected writings of Edward Sapir. International Journal of American Linguistics 17:180—186. Reprinted in Koerner, ed., 1984:59-65. 1986 The development of Sapir's psychology of human behavior. In Cowan et al. 1986:405-31. Proceedings 1928 First colloquium on personality investigation. American Psychiatric Association, Committee on Relations with the Social Sciences, New York.

88

Régna Darnell 1930

Second colloquium on personality investigation. American Psychiatric Association, Committee on Relations of Psychiatry and the Social Sciences. Baltimore, Johns Hopkins. Sapir, Edward 1925 Sound patterns in language. Language 1:37 —51. 1927 Speech as a personality trait. American Journal of Sociology 32:892—905. 1929 A study in phonetic symbolism. Journal of Experimental Psychology 12: 225-239. 1933 The psychological reality of the phoneme. Journal of Normal and Abnormal Psychology 30: 247 - 265. 1949 Selected writings of Edward Sapir. Edited by David Mandelbaum. Berkeley: California. Spier, Leslie—Stanley Newman—A. Irving Hallowell (eds.) 1941 Language, culture and personality: Essays in memory of Edward Sapir. Menasha: Sapir Memorial Publication Fund.

Sapir's panoramic view (1926) of recent advances in linguistics Yakov Malkiel

Abstract One of the least known papers from Edward Sapir's prolific pen is a piece he contributed to the so-called 13th edition of the Encyclopaedia Britannica — actually, a supplement to the 11th. The item, somewhat bizarrely titled "Philology", in fact involved a critical digest of research in linguistics and adjacent disciplines, covering a period of fifteen years, i.e., 1910 — 25. The paper, though left undated, made its appearance in 1926 and must have been written the year before, thus coinciding with the author's transfer from Ottawa to Chicago. The overall sweep of the survey is most impressive; as noteworthy as certain positive comments and commitments are a few manifestations of Sapir's studied indifference.

The 13th edition of the Encyclopaedia Britannica — which combined, according to the slightly cumbersome statement of the title page, Vols. 1 - 2 8 of the "latest standard edition" (i.e., the 11th) with "three new volumes covering recent years", plus the index volume — is undated, but Vol. 31 (P —Z), which concludes the Supplement, on indirect evidence, made its appearance by 1926. Several contributions to that volume refer explicitly to events that took place, or to books and monographs that were published, in 1925. Edward Sapir's piece (112 b —115 b), which is of direct concern to us here, is no exception in this respect. Titled in somewhat archaizing (and, at the same time, Briticizing) fashion "Philology", chiefly because this piece was, in the first place, elicited and organized as an elaboration upon, or a supplement to, an older one written by a team of European scholars, for a preceding edition of the same reference work,1 it actually involved a critical survey or digest of research in straight, undiluted linguistics and adjacent areas of knowledge conducted over a period of fifteen years, with few if any concessions to philology taken in the technical sense of that word ("textual criticism").

90

Yakov

Malkiel

From the outset, Sapir — undoubtedly prodded by the editors to do so — selects the year 1910 as the starting point for the panorama that is to unfold. Most of the publications that he cites belonged to the early twenties,2 one of the "roaring" decades of modern times, especially on this side of the Atlantic, the latest item being his own celebrated paper, "Sound patterns in language", the only one for which he fails to supply any accurate bibliographic details, presumably because the piece was still in press at the cut-off point. As every tyro recalls at present, it appeared in No. 2 of the opening volume of the newly-founded quarterly, Language (1925 a: 37-51). We have, then, a loose string of bibliographic notices from Sapir's pen, written at a memorable turning point of his career. He may well have jotted down this set of stray addenda to his predecessors' bold synthesis — scarcely an enviable task — while he was still a staff member at Ottawa's Victoria Museum, and could easily have corrected the printer's proofs after his subsequent transfer, in 1925, to the Department of Anthropology at Chicago — in all likelihood, the most dramatic twist of his entire academic life. Temperamentally, Sapir — a pioneer and experimentalist par excellence — was hardly suited for an assignment so narrowly circumscribed;3 the fact that he nevertheless managed to cope with it adds to the present-day reader's curiosity about this highly atypical venture. The material commands interest for yet another reason: As early as 1924 his publisher announced that Sapir was busy preparing what sounded like a revision and expansion of his 1921 book, which had made him famous. If this conjecture holds water, the plan called for a change in the title of the book, in the direction of increased explicitness.4 This project, ill-timed, fell through, conceivably in connection with the new round of administrative and teaching duties that Sapir, a newly-appointed department head and associate professor, had meanwhile assumed in transferring to a graduate school bubbling with activities and located in a crowded metropolis. However that may be, the six-column item before us, which could run to about twenty standard-sized pages, may well afford a hint as to the slant in which the revision toyed with in the mid19208 might have been carried out. One final excuse for briefly reverting to this almost forgotten, not to say obscure, shorter writing by Sapir is the fact that it was excluded, doubtless for valid reasons, from David G. Mandelbaum's deservedly influential selection (1949) of his teacher's scripta minora — although it figured duly in the appended bibliography. Admittedly, the piece, as a consequence of its coming close to being,

Sapir 's panoramic view (1926) of recent advances

91

through no one's fault, a patchwork of additions, lacks the elegance, in structure and wording, of Sapir's more characteristic contributions to encyclopaedias of a broader or narrower scope, a feverish activity which started with a slender essay on Vancouver Island Indians (1922:591 — 5) and reached its peak in such classics — squeezed into a lustrum of feverish productivity — as "Central and North American languages" (1929: 138 — 41), "Communication" (1931: 4.78-81), "Custom" (1931: 4.658-62), "Dialect" (1931: 5.123-6), "Fashion" (1931: 6.139-44), "Group" (1932: 7.178-82), "Language" (1933: 5.155-69), "Personality" (1934: 12.857), and "Symbolism" (1934: 14:492-5). The digest under scrutiny, then, belongs to Sapir's earlier commitments along this line and for this very reason lacks the dash and polish of most of its successors; it is also somewhat weighted down through use of supererogatory footnotes and of a concluding bibliographic guide, a feature by definition more useful than exciting. But if one overlooks such tiny cosmetic flaws as well as certain minor inconsistencies in the stringing of topics selected for discussion,5 a specialist is apt to be richly rewarded for his effort in examining this neglected opuscule under a microscope lens. Unlike Sapir's usual practice from before World War I — recall "The history and variety of human speech" (1911:45 — 57), Language (1921), and the aforementioned article "Language" (1933) — here he refrains from dividing the aggregate of his ideas about ongoing linguistic research into a synchronic (descriptive) and a diachronic (historico-comparative) perspective. Instead, one discovers, after a few introductory remarks, an implied fairly sharp division of the whole into three major chunks. The first of these addresses general linguistics viewed in its overlap with a selection of adjoining disciplines, specifically psychology, philosophy, and an ensemble of sociology and anthropology. The second concerns itself with a lower order of generalities, especially with linguistic change, either internally or externally induced, and with the moot issue of its consistency or regularity. Still within the confines of the second major category, the author examines, as has been his wont, morphology ahead of phonology (which he still elects to call phonetics); and morphology turns out to attract him almost entirely as a clue to typological classifications of languages. The third chunk, to which some unifying comments on genetic relations have paved the way, is devoted to individual language families, or to groupings of languages by their speakers' grossly-sketched habitat. The parade opens with a rapid survey of North and South American tongues; the reader is next invited to switch his attention to Austric and Australian, then to African languages. Only at

92

Yakov Malkiel

this point are Indo-European (not 'Indo-Germanic', as in Sapir's earliest writings), Semitic, and Hamitic, with special attention to newly-discovered Hittite and Tokharian, woven into the report. The survey ends with a glance at problematic or tentatively posited groupings, such as Dardic or Pisacha in India and the Karen languages in Burma. The concluding lines, impressionistically dashed off, make passing mention of, typically, isolated languages, or of language families of undemonstrated affiliation (Burushaski, Dravidian, Sumerian, Basque, etc.). The overall sweep of this survey could not have been more impressive; yet symptoms of a certain one-sidedness of approach cannot be hidden. When one detects the virtually complete absence of syntax (except for allusions to word order) and of lexicology (including its tantalizing etymological prong) from the procession of linguistic disciplines, or as soon as one recognizes Sapir's bizarre indifference to dialectology (culminating in dialect geography at that time), as the polar opposite of inquiries into often highly stylized literary languages, or upon realizing how pitifully scant attention the progress report bothers to pay to undeniable accomplishments, in that decade and a half, of several cis- and transAtlantic Indo-Europeanists, not to mention Romanists (even H. Schuchardt's late academy memoirs, undeniably germane to Sapir's cosmorama, go unreported), one cannot repress the suspicion of a certain partiality of information, or subjectivity of taste, or else sheer haste, on the part of Sapir, 6 despite, I repeat, the astonishingly encyclopedic range of the whole. It would, obviously, be unprofitable to expatiate on such lacunae, not all of which the author, even at discernibly later dates, managed to fill. The positive value of the piece under study consists in this: First, it gave Sapir a unique opportunity, or excuse, to review with as much candor as he wanted in every instance, the achievements (or, in some cases, the failures) of a long line of contemporary experts, both linguists and tillers of contiguous terrains; and second, it stimulated him to make a number of pronouncements of his own, which often show him occupying, in 1925 — 26, a patch of middle ground between his earlier positions (e.g., in Language, as of 1921) and his later attitudes (as displayed in, say, the concluding string of his encyclopaedia articles, especially those of the years 1929 — 1933). Let us examine separately these two parallel claims of the 1926 chronicle even on the present-day reader's attention. Sapir's report contains provocative statements, made doubly challenging through their enforced succinctness, on the following thinkers and social scientists (fellow linguists included): F. Boas, E. Cassirer, M. De-

Sapir's panoramic view (1926) of recent advances

93

lafosse, R. B. Dixon, Sir G. Grierson, O. Jespersen, Sir H. H. Johnston, J. R. Kantor, A. L. Kroeber, B. Malinowski, N. Marr, A. Meillet (in the company of M. Cohen), H. Meller, C. K. Ogden, R Radin, I. A. Richards, R Rivet, Father W. Schmidt, J. R. Swanton, A. Trombetti, J. Vendryes, J. B. Watson, D. Westermann, plus Sapir himself (the author showed not the slightest inhibition about referring readers to his own inquiries). The roster of names is spectacular, but does not exhaust the ground covered in a critical vein. At least in one instance the chronicler refers, globally, to an influential school of thought, the pacesetters of Gestaltpsychologie or configurational psychology, without condescending to mention a single name. Ordinarily all these investigators, who reached the peak of their respective activities in the 1910 — 25 segment of time, were judged as individuals, except of course where they had performed as a team on the special occasion that Sapir was focusing upon (as holds for Meillet and Cohen as co-authors, or rather co-editors, of a tone-setting reference book). However, the North American ethnologists responsible for the exploration of the New World's indigenous languages and underlying cultures, i.e., Dixon, Kroeber, Radin, Sapir himself, Swanton, plus for good measure the Frenchman Rivet, were depicted as a (loosely-organized) team or group and placed among those who have sought to bring order out of the linguistic chaos that still largely prevailed in America (114 b). 7 Sapir's verdicts are, as one might expect, particularly noteworthy where controversial personalities and doctrines applauded by some qualified observers while violently rejected by others are involved. A few examples will do. To Sapir's mind, A. Trombetti deserves recognition, because he "among serious linguists ... is perhaps the only one who has ventured to commit himself to the theory of monogenesis and has actually attempted ... to show in what manner the various groups of languages that are generally recognised are related to each other" (114 b). One gathers that the relevancy of the Italian theorist's and generalist's revolutionary approach — impugned in many quarters — was rooted in his audacity, which, in turn, was in tune with the Zeitgeist (beside showing an affinity to Sapir's own personality?): "There is an increasing tendency now to make larger syntheses and to suggest as at least probable, if not entirely demonstrable, relations that at first blush seem far-fetched." Another highly characteristic remark, similarly tilted (115 a), will be found apropos H. Moller's "very systematic and detailed attempt to connect Indo-European and Semitic (the relationship of which to Hamitic is now generally recognized) seems not to have been cordially received by either the Indo-

94

Yakov Malkiel

Europeanists or the Semiticists, but to the general linguist who studies [Moller's] works his demonstration seems highly suggestive, not to say convincing."8 Sixty years later the problem is still in abeyance.9 Of course, Sapir also knew how to insinuate or express skepticism ("less solid seems to be the attempt of N. Marr ...," (115a))10 or, at least, indifference, not to say a suspended judgment; to achieve such goals, he would simply have recourse to such formulaic phrasing as: "X. attempts to demonstrate ...," or "X. asks himself whether ...," withholding any explicit endorsement or rejection. Given the forcibly epigrammatic style of these verbal silhouettes (plus the pressure of time under which the author was presumably toiling), a certain unevenness of critical performance was bound to ensue. The brevity of the lone statement (113 a) on Boas, however respectfully phrased, may at first blush cause surprise: "Much of the most competent and authoritative ethnological information that we possess has been obtained as material ancillary to linguistic studies; one may instance, for America, the Kwakiutl publications of F. B.;" but then, the opinions of teacher and student on several pending issues are known to have started drifting apart rather sharply with the passage of time, and sixty years ago public attempts at open magistricide were not yet socially sanctioned. The characterization of Jespersen (112 b) remains a shade enigmatic ("J. ... takes into consideration a good many matters of practical interest ordinarily neglected by the scientific linguist"). Does this laconic remark touch upon the invention of two artificial languages by the Danish linguist, and should not other, more relevant, aspects of his many-faceted œuvre, e.g., his concern with phonosymbolism and his proposal of a new-style phonetic transcription, have been mentioned instead? To appreciate such fleeting hints, one is well-advised to hold them against certain more starkly nuanced book reviews from Sapir's pen, e.g., the one that he published precisely that same year (498 — 9), of a book by Jespersen not singled out for mention in the encyclopaedia article under scrutiny. Similarly, Sapir's brash description of Meillet and Cohen's perspective as "conservative" comes to life when held against his practically contemporaneous full-scale assessment of the book they had launched (1925: 373 — 5). Even the far more circumstantial and downright enthusiastic statement on Ogden and Richards's The meaning of meaning (112b —113a) gains in depth when collated with Sapir's slightly earlier response to it, which cast him in the role of formal reviewer (1923 b: 572 — 3). There may be a certain irony in the fact that, on balance, philosophers and sociologists, among the latter especially B. Malinowski

Sapir's panoramic view (1926) of recent advances

95

(113 a), emerge from Sapir's survey as having been, with rare exceptions, far more sparkling personalities than the spokesmen for what he calls "Indo-European philology"; i. e., hard-core linguists. To take up, as promised, the thread of the alternative analysis: Sapir engages, not to say indulges, in a good deal of parenthetical positiontaking, especially in championing, not infrequently in side-remarks or subordinate clauses, several of his favorite ideas about what linguistics, ideally, should be all about, while commenting on the triumphs or, if need be, shortcomings of fellow scholars extra-active over the last decade and a half. These ideas it would definitely be worthwhile to distill, so that they can be put to use for a variety of defensible purposes, above all, to illustrate the author's relentless growth from the ebullience of his 1921 book to the imposing maturity of his 1933 article, either one parsimoniously titled "Language". Within the limits of space available to me, I cannot go beyond offering a few samples of loosely-strung pronouncements, sometimes detached from their immediate context. Let us start out from the author's pithy self-characterization (112 b): "Sapir, a specialist in American Indian languages, is chiefly concerned with formal and psychological fundamentals, and uses freely examples taken from the languages of primitive peoples." 11 Next, one comes across miscellaneous basic ideas, such as these: a) "Other social and humanistic sciences [i.e., 'Geisteswissenschaften'] than linguistics itself have taken a serious interest in the data of language" (H2b); b) "Language always implies a selective organization ..., the individual sound being defined not merely as an articulation, but as a point in the pattern" (112 b); c) "The relativity of all knowledge and all experience to the habitual symbolism by means of which they are expressed has become more and more clearly understood" (112 b): d) The study of exotic forms of speech "discloses the possibility of markedly distinct analyses of experience where one might naively suppose that our customary analysis via speech is resident, as it were, in the nature of things" (113 a); e) A subtle dividing line is drawn between "idea" (i. e., a concept which may be present in one yet absent from another language), and "thought," "a total psychic sequence ... or unit of communication" (113a), which almost invariably can be expressed in any language by devices peculiar to it.

96

Yakov Malkiel

f) "The most important single factor making for an increased or retarded rate of linguistic change would seem to be the formal set of the language itself. ... Less important than the inner 'drift' of a language is the tendency to change as a result of external contact" (113 b).12 A few additional problems of comparable breadth and equally exciting formulation could doubtless be distilled by dint of effort. For b), the author supplies a reference to his own paper, of very recent vintage: "Sound patterns in language." As need not be labored, phonemics in an embryonic and rudimentary shape was briefly anticipated in his 1921 book and later explicitly postulated in the 1933 article; here we are facing the mid point, devoid of any allusion yet to the nearly parallel structuralist thinking that was then going on simultaneously in the minds of some forward-looking Europeans. Point c) preempts the essence of the SapirWhorf hypothesis.13 The concluding round of queries revolves about these two pivotal issues: To what extent were Sapir's (a) selections of the agenda and (b) the verdicts that he returned on that memorable occasion, certain aforementioned details apart, subjectively colored? And, to what degree, if at all, can his judgments be viewed as prophetic, insofar as he may have extrapolated, from close examination of goings-on of the 1910 — 25 era, certain trends that actually triumphed in the subsequent period; in other words, did the events enacted on the linguistic arena prove his implied prognosis right or wrong? At first glance, the margin of authorial subjectivity indeed seems to be wide (even excessively wide) for a mere attempt at updating. But the actual source of the trouble appears to have been not any deliberately programmatic bias or onesidedness, but rather, for the reasons already stated, a severe pressure of time and the need for husbanding resources, circumstances which were jointly apt to force Sapir to squeeze into the scope of his assignment preferably the topics he had known best all along and could trust himself to cover most competently. These topics, not surprisingly, almost invariably turned out to be the same that were closest to his heart in his personal research. Thus, anyone familiar with the strides etymology was taking on the European continent in 1910 — 25, with the advent to influence of Swiss-style dialect geography, may be taken aback by Sapir's failure to pay the slightest attention to these faraway gropings, until he remembers that the author's inquiries, including his two consecutive attempts (1913:617 — 46, 1923a:36 —74) to bracket genetically Wiyot and Yurok (California) with distant Algonkin, even

Sapir's panoramic view (1926) of recent advances

97

though crowned with success on the whole, remained quaintly vulnerable on the side of specific "lexical correspondences," i. e., of undergirding etymological equations (I. Goddard, forthcoming, in the wake of — distinctly earlier — withering criticism from the pen of T. Michelson 1915). In the end, Sapir awoke to the charms of etymology pursued for etymology's sake, painstakingly and in an almost artistic vein, at the height of his years spent in Chicago, in the late 1920s, and especially at Yale, throughout the 1930s (Malkiel, forthcoming), after having veered his attention, passionately and dramatically, to the Eastern Mediterranean in Antiquity. Within this context of the prevalent personal scale of values one gains the impression that Sapir, toward the end of his Ottawa period, was gripped by a certain surfeit of old-style Indo-European linguistics, including the technique used for the reconstruction of the underlying protolanguage. To be sure, one detects in this quindecennial report stray references to Greece and Italy at the dawn of history (113 a b), but made chiefly for the sake of prehistorically relevant substratum civilizations (Minoan, i. e., Mycenaean, and Etruscan), to use the currently fashionable term. One example, fleetingly tossed off, of parallel independent development, supplied from the rich store of Indo-European (the widespread erosion of the case system) is immediately matched by a similarly contoured tendency toward isolation in Sinitic (113 b), etc. Apropos diffusion and bilingualism (113 b, 114 a) — two complex issues which the author passed over lightly — a wealth of discoveries were being made, in the time segment under discussion, in regard to the overlaying of Celtic by Latin and Germanic; Sapir's prize illustration is the superimposition of Indo-Aryan (without mention of Sanskrit) over indigenous Dravidian, etc., on the subcontinent (114 b). Almost predictably, the Baltic and Slavic domains do not once come up for mention. The only islets in this vast ocean of unknowns whose exploration genuinely prodded Sapir's alertness in 1926 were Hittite and "Tokharian", i.e., Tocharian (A and B).14 As regards the latter, Sapir — conservative or conventional for a change — confined himself to reporting that in certain "phonetic" (we would, once more, be tempted to say, phonological) respects it agreed 'more closely with Greek and Latin than with the Indo-Iranian languages that were geographically nearer to it', a paradoxical truth as valid then as it still is sixty years later. What had not yet matured into a defensible conjecture by 1926 was Sapir's lingering set of suspicions that there may have existed a barely recognizable link between Tibetan and Tocharian (1936 a:

98

Yakov Malkiel

259 — 71; left unfinished) as well as one between Tibetan and certain autochthonous New World languages. Only in the thirties, after having thoroughly familiarized himself — not least, through a young visiting scholar, namely J. Kurylowicz's, intervention — with the "laryngeal theory", did Sapir allow this long-drawnout rule of apathy toward current research in Indo-European to be broken, presumably because he then recognized in its ranks a newborn avant-garde. As the summit of this eventual revival of a long-dormant latent interest one may well cite that relatively late paper in which Sapir allowed for evidence from Amerindian (specifically, from Navaho, Nootka, and Kwakiutl) and insights gained from paleo-Indo-European analyses in a new key to lend each other mutual support, at least on the typological plane (1938 b: 248 — 76). We are yet at a considerable distance from this state of bliss and reconciliation in 1926! Also to be credited, or charged, to Sapir's private predilections is the reappearance here of the ever-haunting issue of the classification of languages. Strictly speaking, two separate problems are involved: first, the establishment of criteria devised for the general ordering of languages on formal grounds (114 a) — here the author simply compresses into a single, tightly-phrased paragraph the stimulating, if slightly wordy, sixth chapter, typologically slanted, of his 1921 book; and second, the specific configurations of certain favorite language families, including those of aboriginal America and Central beside South Africa (114ab). One will do well to remind oneself that Sapir, until shortly after his death, was chiefly noted for precisely this skill: the imaginative classificatory schemata that he had proposed.15 Sapir's 1926 statement, so far as it bears on North and Central America, deserves to be assessed by an expert as part of the background for his celebrated piece written, under a different heading, for a later edition of the same encyclopaedia (1929:5.138—41), In the past Sapir has occasionally been praised for his clairvoyance. The paper under examination bears out this belief only in part. Sapir was indeed inspired in foreseeing North America's more spontaneous involvement with broad-gauged issues in linguistics than with the painstaking elucidation of spatio-temporal details as well as the gradual shift of linguistics from its nineteenth-century alliance with humanities in the direction of close affinity with social or, better still, with cognitive sciences, a term still alien to his vocabulary. He also made the right bet in solidarizing himself with the configurationalists and the moderate behavioralists rather than with a J. B. Watson or an A. P. Weiss, whom L. Bloomfield, in contrast, learned to admire almost uncritically. On

Sapir's panoramic view (1926) of recent advances

99

other scores, Sapir did not quite succeed in avoiding miscalculations. Thus, he overlooked the true potential of Slavic, in free alliance with Caucasic, research; his tentative classification of South-East Asian languages was not, one suspects, entirely borne out by subsequent, more specialized probings;16 he may have gone too far (this time, in the company of Bloomfield) in categorically denying any influence of script on the tempo of the development of speech. With due allowance for the possibility that linguists of divergent persuasions may bristle at certain statements and certain silences in Sapir's 1926 interim report, there is no denying that it is a document of the highest importance for the study of the author's own intellectual growth.

Notes 1. With the limited bibliographic resources at my ready disposal, I can only report that the 1902 edition (Twentieth Century Edition) of the Britannica offered, under the unifying lemma "Philology", two mutually complementary, separately subtitled, articles, the first by William D. Whitney, the second by Eduard Sievers (18.765—90); while the 11th edition (1910-11) - of which the 12th (1922) and the 13th (1926), except for the new set of supplement volumes, are in essence mere reprintings — contained, again under the same key word, a composite article credited to the Cambridge scholar Peters Giles (Cambridge University Press being, perhaps not insignificantly, at that time responsible for the entire venture) and, once more, Eduard Sievers (21.414 b—437 b), followed by an unsigned Summary of philological articles (437 b—438 b). It was this Giles — Sievers piece, heavily biased in the direction of Indo-European ('Aryan'), that Sapir received an invitation to supplement and bring up to date. 2. Note, however, that the manuscript of J. Vendryes' book was completed in 1914; its publication was grievously delayed by World War I (1921:5). Under these circumstances, it deserves to be chronologically bracketed with L. Bloomfield's 1913 venture, which Sapir, unaccountably, refrains from mentioning at all. Equally astonishing, even though for a different reason, is the absence of any reference to F. de Saussure's posthumously published Cours (1916; 21921). For a deftly-painted pen portrait of such a trail-blazing foreign linguist see Newman (1963:443—4). Even though the model that sat for him on that occasion was L. S. Vygotsky, almost every statement applies, with a vengeance, to Sapir. 3. Sapir's favorite scope of an encyclopaedia article can be gleaned from his actual broad-gauged contributions over the lustrum 1929 — 34: "Central and North American languages"; "Communication", "Dialect", "Fashion", "Custom"; "Group"; "Language"; "Personality"; "Symbolism". 4. From the paper jacket of J. C. Gregory's book (1924), to cite one source randomly, one gathers that the International Library of Psychology, Philosophy, and Scientific Method (general editor: C. K. Ogden) was to contain, among books solidly commissioned ('in preparation'), E. Sapir's Language as symbol and as expression, in the

100

5.

6.

7.

8.

Yakov Malkiel company of writings by I. A. Richards, B. Malinowski, and P. Radin. It is independently known, on epistolary evidence, that, as late as the mid 1930s, Sapir toyed with the idea of expanding Language into a book twice its original length. This time, surely, it was poor health that, tragically, prevented him from carrying out this project. Thus, a long paragraph subtitled "Forms of speech" was wedged in between "Psychology", "Philosophy", on the one hand, and, on the other, "Sociology and Anthropology". A medium-sized statement on "Place names", incongruously, found a niche immediately after the analysis of contacts with philosophy; some remarks subtitled "General tendencies" were left dangling in the middle of the piece (113 b) rather than being placed, more conspicuously, either toward the beginning or at the end. One wishes one could blame such gaffes and inconsistencies on some editorial amanuensis. The truth of the matter, however, is that Sapir, for all his agility with the pen, was less than careful about cogent paragraphing. One detects isolated slips of this sort, embarrassingly enough, even in the celebrated seventh chapter ("Language as a historical product: Drift") of Language. It would be pedantic to cite supporting examples; suffice it to state that such infelicities may have stemmed from the author's oft-attested habit, especially in the early and mid-1920s, to dictate part of his papers. The clustering of such non sequiturs in the piece under discussion may reflect the pressure under which he was toiling in 1925. I shall not expatiate here on a fairly long procession of cross references to other encyclopaedia articles of the same vintage (Archaeology, Central America, Folklore, Phonetics, Psychology, Western Asia), which are almost certainly due to routine interferences by the editor in charge. Boas's œuvre is alluded to in a different passage; see below. Sapir had a perfectly valid excuse here for leaving Leonard Bloomfield's name unmentioned, since his counterpart, by 1922—24, had just begun to delve into the study of Central Algonkin (Fox, Menomini) and, moreover, showed no leanings toward specializing in language classification. Stranger is the fact that Sapir found no slot for bringing up Bloomfield's trailblazing Tagalog Texts (1917). On the coolness and mutual aloofness in the relations between these two pioneers see C. F. Hockett's crisp comment which precedes the reprinting of Bloomfield's studiedly reserved 1922 review of Sapir's Language (1970:41-4). As if by way of recoil from the pedantry of bibliographic explicitness (be it only to save editorial space), Sapir here seldom bothered to supply publication dates, or so much as the exact titles of the studies alluded to, or vital biographic data on the key figures mentioned. With respect to the Danish Germanist and Anglicist H. M0ller (1850 — 1923) it is useful to remember that his pivotal publications go back to the years 1906, 1911, and 1917. Sapir's remark reveals that he was an avid reader of critical assessments, too. This habit may have stood him in good stead in regard to Trombetti's (1866—1929) two-volume 'magnum opus': Even though Sapir had an elementary knowledge of Italian, he might have hesitated to give too much of his leisure to the perusal of Trombetti's not easily assimilable book without some preliminary guidance from book reviewers. To revert to Mailer's pet idea: Sapir narrowly missed A. Cuny's opening gambit (1924) in the game that French scholar played for, at least, twenty years in trying to make a case for the 'langues nostratiques', a label he had invented as a common denominator for Indo-European and what is at present known as a branch of the Afro-Asiatic family.

Sapir's panoramic view (1926) of recent advances

101

9. One scholar living in this country and currently active in the domain of contacts between Semitic and Indo-European is Saul Levin. Positing 'contacts', clearly, is not tantamount to reconstructing common descent. This must have become crystal-clear to Sapir himself throughout the 1930s, when his attention turned to isolated Hittite and Philistine (i.e., Indo-European) words absorbed into Biblical Hebrew, and to similar situations involving Greek (1936:272-81; 1937:73-77; 1938a:86-8). 10. This adverse judgment bears on the earlier phase, relatively less radical, of Marr's approach. For details see L. L. Thomas (1957). 11. Sapir at this stage freely alternates the labels 'exotic languages' and 'languages of primitive peoples'. In previous writings he did not hesitate to have recourse to the more audacious tag 'primitive languages,' which had meanwhile become a bugaboo among social scientists. 12. This instance of Sapir's crucial appeal to the metaphor 'drift' eluded my attention on two earlier occasions: (1981:535-70 and 1982:479-84). 13. For — possibly — the latest summation of this heatedly debated issue see P. Kay and W. Kempton (1983), with a useful if woefully incomplete bibliography appended to it (30-2). 14. Sapir's comments on Anatolian languages (to use a currently more apposite term) candidly reveal the initial imbroglio caused, in the separate headquarters of paleoOrientalists and comparativists, by the sensational archaeological discoveries made in Asia Minor toward the beginning of this century. Further research before long gave rise to new theories; yet, throughout the 1930s, despite his personal friendship with E. H. Sturtevant, his next-door neighbor at Yale, Sapir refused to subscribe to the latter's Indo-Hittite hypothesis, which at that time enjoyed its first vogue. 15. One eloquent example of the solid reputation which that particular facet of Sapir's scholarship had gradually acquired is the structural design of that fine commemorative volume which Stanley Newman, in collaboration with two friends, launched in 1941. That venture fell into four major sections, of which the first — hence the one, beyond dispute, most prominently displayed — was entitled: "Problems of linguistic classification", with individual contributions by H. Hoijer, C. F. Voegelin, and M. R. Haas, all three bearing on the Amerindian scene. While Newman's own piece was skewed in an entirely different direction (1941:94—106), he unquestionably was coresponsible for the adopted architecture of the entire volume. It almost goes without saying that, in other extensions of his active curiosity, Newman, as a true "Sapirian", became deeply steeped in matters of possible genetic kinship between two or more indigenous American tongues (1964:1 — 13) and in the kindred matter of reconstructed New World ancestral languages (1950:1 — 19, 73 — 81, with R. Weitlaner). (Let me here thank Victor Golla for some useful bits of information.) 16. Without pretending to be an expert in this area, one cannot help noticing that a book reflecting present-day knowledge in this technique of indexing, such as C. F. and F. M. Voegelin (1977), deviates considerably from Sapir's insights and hunches, as of 1926. If Sapir, following in the footsteps of Grierson, set off a group of Pisacha or Dardic languages (115 a), leaving them somewhere halfway between archaic Indo-European and non Indo-European Burushaski, the Voegelins, guided by Morgenstierne, Emeneau, and Rand, credit Dardic to Indie = Indo-Aryan rather than to Iranian, and altogether omit mention of the rival label Pisacha (163). Even sharper are the discrepancies between the two views of Man (1926, 1977), which apparently has ceased to rank as the name of a language family.

102

Yakov Malkiel

References Bloomfield, Leonard 1913 Introduction to the study of language. New York: Henry Holt and Company. 1917 Tagalog texts with grammatical analysis. University of Illinois Studies in language and literature 3.2—4. Cuny, Albert 1924 Etude prégrammaticale sur le domaine des langues indo-européennes et chamitosémitiques. Paris: Société de linguistique de Paris. Giles, Peter—Eduard Sievers 1911 Philology. Encyclopaedia Britannica (11th edn.) 21.414b-437b. Goddard, Ives 1986 Sapir's comparative method. Transactions of the E. S. Centenary Conference, Ottawa, 1 - 3 October 1984, pp. 191-214. Gregory, Joshua C. 1924 The nature of laughter. New York: Harcourt, Brace, & Company, Inc.; London: Kegan Paul, French, Trubner & Co., Ltd. Hockett, Charles F. (ed.) 1970 A Leonard Bloomfield anthology. Bloomington and London: Indiana University Press. Kay, Paul—W. Kempton 1983 What is the Sapir— Whorf hypothesis? Berkeley cognitive science report 8. Malkiel, Yakov 1981 Drift, slope, and slant. Language 57:535-70. 1982 Mutual attraction, typological convergence, multiple borrowing, parallel independent development, spiral-shaped curve, periodicity, advanced state, slope, drift, drag, slant. Romance Philology 35:3.479—84. 1986 Sapir as a student of linguistic diachrony. Transactions of the E. S. Centenary Conference, Ottawa, 1 - 3 October 1984, pp. 315-40. Mandelbaum, David G. 1949 (1958) Bibliography [of Edward Sapir's writings]. In: Selected writings of Edward Sapir in language, culture, and personality. Berkeley and Los Angeles: University of California Press. Meillet, Antoine—Marcel Cohen 1924 Les langues du monde, par un groupe de linguistes. Paris: Société de linguistique de Paris. (New edn., rev. by M. Cohen: 1952). Michelson, Truman 1915 Rejoinder to Sapir's "Algonkin languages of California," PAES 17.194-198. Meiler, Hermann 1906 Semitisch und indogermanisch, 1: Konsonanten. Kopenhagen: H. Hagerup. 1911 Vergleichendes indogermanisch-semitisches Wörterbuch. Göttingen: Vandenhoeck & Ruprecht. 1917 Die semitisch-vorindogermanischen laryngalen Konsonanten. Avec un résumé en français. Kebenhavn: A. F. Hest & Son. Newman, Stanley S. 1941 Behavior patterns in linguistic structure: A case study. In: Language, culture, and personality; Essays in memory of Edward Sapir, 94—106. Menasha, WI: Sapir Memorial Publication Fund.

Sapir's panoramic view (1926) of recent advances 1963

103

Review, L. S. Vygotsky, Thought and language. Romance Philology 17:2.442 - 4 . 1964 Comparison of Zuni and California Penutian. IJAL 30.1 —13. Newman, Stanley S. —R. Weitlaner 1950 Central Otomian (1. Proto-Otomi reconstruction; 2. Primitive Central Otomian reconstruction). IJAL 16.1 — 19, 73—81. Sapir, Edward 1911 The history and varieties of human speech. Popular Science Monthly 79.45 — 67. Twice reprinted. 1913 Wiyot and Yurok, Algonkin languages of California. American Anthropologist (N.S.) 15.617-46. 1921 Language: an introduction to the study of speech. New York: Harcourt, Brace. 1922 Vancouver Island Indians. Encyclopaedia of Religion and Ethics (ed. J. Hastings) 12.591-5. New York. 1923 a The Algonkin affinity of Yurok and Wiyot kinship terms. Journal de la Société des Américanistes de Paris (N. S.) 15.36 — 74. 1923 b Review of C . K . Ogden and I.A. Richards, The meaning of meaning, The Freeman 7.572—3. 1925 a Sound patterns in language. Language 1:37 — 51. 1925b Review of Meillet-Cohen (1924), in Modern Language Notes 40.373-5. 1926 Review of O. Jespersen, Mankind, nation, and individual, from a linguistic point of view, in: American Journal of Sociology 32.498—9. 1929 Central and North American languages. Encyclopaedia Britannica (14th edn.) 5.138-41. 1931—4 Contributions on Communication, Dialect, Fashion, Custom, Group, Personality, and Symbolism to Vols. 4 to 14, Encyclopaedia of the Social Sciences, New York. 1933 Language. Encyclopaedia of the Social Sciences 12.85 — 7. New York (twice reprinted). 1936a Tibetan influences on Tocharian (1). Language 12.259 — 71. 1936 b Hebrew 'argäz, a Philistine word. Journal of the American Oriental Society 56.272-81. 1937 Hebrew 'helmet', a loanword, and its bearing on Indo-European phonology. Journal of the American Oriental Society 57.73—7. 1938 a Hittite siyanta and Gen.l4:3. American Journal of Semitic languages and literatures 55.86-8. 1938 b Glottalized continuants in Navaho, Nootka, and Kwakiutl (with a note on Indo-European). Language 14.248 — 74. Saussure, Ferdinand de 1923 Cours de linguistique générale, edited by C. Bally—R. Riedlinger—A. Sechehaye, 2d edn. Paris: Payot. (1916) Sievers, Eduard 1902 Philology 2: Comparative philology of the Aryan languages. Encyclopaedia Britannica (Twentieth Century Edition) 18.781 - 9 0 . Thomas, Lawrence L. 1957 The linguistic theories of N. Ja. Marr. University of California Publications in Linguistics 14.

104

Yakov Malkiel

Trombetti, Alfredo 1912 Elementi di glottologia. 2 vols. Bologna: N. Zanichelli. Vendryes, Joseph 1921 Le langage: introduction linguistique à l'histoire. Paris: La Renaissance du Livre. Voegelin, Cari F. — Florence M. Voegelin 1977 Classification and index of the world's languages. New York, etc.: Elsevier. Whitney, William D. 1902 Philology, 1: The science of language in general. Encyclopaedia Britannica (Twentieth Century Edition) 18.765 — 80. Cambridge: University Press.

B. American Indian Studies

Institutional language maintenance resources of American Indians in the early 1980s Joshua A. Fishman

Abstract Nationwide data is presented on the numbers of non-English periodical publications, radio programs, community schools, and local religious institutions sponsored by various Amerindian groups in the early 1980s. The lion's share of all of these language maintenance resources consisted of local religious institutions. The implications of this finding and of the distributions observed are discussed and follow-up institutional censuses are advocated.

Rise and Fall of the Ethnic Revival (Fishman et al. 1985) provides far more data on Amerindians than did the earlier Language Loyalty in the United States (Fishman et al. 1966). Nevertheless, both works focus on the colonial and immigrant languages of our country and, as a result, Amerindians receive scant attention. In the present paper, in honor of Stanley Newman, I plan to concentrate entirely on Amerindians and, thereby, at least partially, to begin to right the balance. I can think of no more appropriate honoree than Stanley Newman to help foster greater attention to the institutional/communal bases of Amerindian language maintenance resources in the United States, and no more interested a group of readers for an article on this topic than those whose appreciation of Stanley Newman has brought them together around this volume in order to honor him so deservedly.

Number and geographic distribution of units in four institutional fields There have been a few relatively recent efforts to gauge the numbers of individuals speaking various Amerindian languages (see, e. g., Chafe 1962, Chafe 1965, Fishman 1981, Gundlach and Busch 1981, Martin 1975, U. S. Bureau of the Census 1973, Veltman 1983, Waggoner 1978) but, to

108

Joshua A. Fishman

my knowledge, no one else has tried, as yet, to gauge the institutional bases (in terms of local religious units, community sponsored schools, radio/television programs and periodical publications) of Amerindian language maintenance. However rough the data may still be in connection with such a first attempt, I hope that it will be of interest, blemishes and all, as a step along a necessary and ultimately promising path. Counts of individual speakers provide, at best, estimates of individual language maintenance; counts of community institutions provide estimates of the sociocultural underpinnings of language maintenance and, accordingly, provide better estimates of language vitality than those derivable from censuses of individuals alone. In the early 1980s there were roughly three hundred and fifty thousand Amerindians in the USA (out of a grand total of nearly a million) who claimed non-English mother tongues (Fishman et al. 1985:148). These claimants represented roughly 1% of all 34 million claimants of nonEnglish mother tongues at that time. Thus, if Amerindians were to be represented in connection with language maintenance related institutions to the same degree that they were represented in the total population of non-English mother tongue claimants, then they would sponsor roughly 1% of the non-English language maintenance institutions as well. Our first query, therefore, is whether this is indeed the case. All in all, my associates and I were able to identify 23,812 institutions (of the four different kinds enumerated above) as being operative somewhere in the USA during the early 1980s (Fishman et al. 1985:192-200). Of these, 1,503 were sponsored by one or another of over ninety different Amerindian groups. Clearly, Amerindians, although only 1 % of all of the nonEnglish mother tongue claimants in the USA, were apparently supporting more that 6% of the non-English language maintenance institutions (of the above mentioned four types) in the country, that is they were represented in the latter universe nearly six times as frequently as might have been expected on the basis of their representation in the former universe. This is an anomaly, one of many, that merits the attention of sociolinguists and Amerindian specialists alike. Table 1 reveals some of the basic data concerning Amerindian institutional resources of non-English language maintenance in the early 1980s. In terms of geographic distribution, Amerindian language maintenance institutions are primarily Western and Southern, 80% of all such institutions being located in these two regions which (according to Fishman et al. 1985:179) account for 84% of all claimants of Amerindian mother tongues. The South is the most strongly over-represented region in this

Institutional language maintenance resources

« a w < HH Z S Z > S ci u t— © o v>

109

a,

^ z^ 2 < z

B c

3-2

Cu olo pa-pas píelas lato sawalow simaría'' siya sumlela7 tuma-thi pa-ka sapha-to.

I recorded only one true verb, limi 'file' or 'to file', and I recorded various paradigmatic forms of this verb over several different years. The occurrence of this borrowed verb form is unusual because it is the only known verb borrowed from Spanish into any Yokuts language. Like many other languages, Yokuts does not borrow verb forms. This restriction on borrowing in Yokuts was carefully noted by Newman: as opposed to the comparatively free borrowing of foreign nouns, no example of a borrowed verb has been found. In short, nouns are easily and freely made in Yokuts; but once formed, they move primarily within the narrow morphological limits of case formation. (1944:168-69) There were few phonological surprises in the forms recorded; however, the occurrence of the whole series of the retroflex stop including glottalized [t'], aspirated [th], and plain [t], as in hit'wa-yo9 'devil', wulathifa 'drunk, drunkard', and ko-ti 'pig' is very interesting from the areal/ historical perspective (see Langdon and Silver 1984). The appearance of glottalized stops in loan words, particularly in medial or initial position as in hit'wa-yo9 'devil' or k'u&a-la 'spoon' is unusual and recorded rarely or not at all for words borrowed into other central California languages. Although I have chosen to phonemicize the fricatives in Wikchamni into /s s h x/, I did see some variation among the sibilants with one individual. The variation involved alternations of [§/§]. These alternations also appeared in initial position in some Spanish loan words: sumlela9 'hat', sawalow 'Saturday', sa^a-flo 'shoes', sulta-thi 'soldier', and santiya9 'watermelon'. The other forms in Table 1 follow the phonological behavior one might expect for a Yokuts language borrowing Spanish words.

126

Geoffrey Gamble

What differences that do occur can be diagnostic for determining the manner in which the word entered Wikchamni. A few morphological quirks appear in the data. The word for 'Fresno' phelesnuw was borrowed into Wikchamni with the appropriate locative case marker {-w}. I had suspected that this form was also used as the nominative case form, but I had not recorded other examples, so I could not be certain. In the summer of 1985 I heard the ablative form for 'Fresno' phelesnuwnuth which confirmed that the locative form was also used as the nominative. Thus, the nominative, accusative, and locative case forms for 'Fresno' are identical. The word for 'pocket' wulus is a perfect match with the Spanish form bolsa. The word was undoubtedly borrowed in its accusative case form wulsa as a Class IIA noun theme, giving the appropriate nominative case wulus. One of the more morphologically complex forms is the word for 'trousers'. This form includes the suffix {-iwis} marking a reflexive or reciprocal verbal noun on an assumed verb base *khalsu, obviously from the Spanish word calza.5 After vowel harmony and an a dissimilation the resulting form becomes khalsuwas. The semantic derivation is unclear, but must have been something like 'to trouser yourself or each other'. The word wulathitha and a variant form wulathic'a are agentive forms of an assumed verb base *wulath 'to be drunk'. Unfortunately I never recorded the verbal form, but the morphological history of the noun is transparent. The second form, wula&i-^ic, is the neutral agentive of a verb theme and clearly related to the previous forms, but with a different derivational history. In summary, several points can be made about Spanish loans in Wikchamni. First, there are at least two levels of loan words. There are those that came into Wikchamni through an intermediate native language. These include 'button, devil, drunk, horse, pocket, Saturday, tobacco' and many others where the borrowed forms display phonological/morphological consistencies with the intermediate languages. For example, the consistent correlation of Spanish b with the native w (see Shipley 1962 for a detailed discussion). There is also a set of words that seem to have come into Wikchamni directly from Spanish. This set includes words like 'cracker, grapes, soap' and others that display phonological/ morphological differences with possible intermediate languages and also closely mirror the Spanish forms. As a corollary to this point, one can assume that the forms borrowed directly from Spanish are much more recent acquisitions. Both historical and linguistic arguments can be brought to bear on this issue. Direct Spanish influence upon the Tule-

Spanish loans in Wikchamni

127

Kaweah group was probably fairly late in coming. The earliest possible contact would have been in 1772 when three Spanish deserters took up residence among the Palewyami and Bokninuwad. The existence of the deserters was reported by Father Garces in April 1776 when he travelled as far north as the White River at the southern boundary of the TuleKaweah territory. Except for occasional deserters from the Spanish military, there was little or no contact with outsiders for these Sierra Foothill people. The late 1820s saw an influx of American traders, including Jedediah Smith, Ewing Young and Kit Carson and by the 1850s many of the tribal groups had been moved by the American military to reservations like that at Fort Tejon. Linguistically, the more recently borrowed forms display phonological correspondences that differ from those Spanish words borrowed through an intermediary language. An early correspondence of Spanish / to Wikchamni [p] becomes the correspondence/to [ph]. The later correspondence was still active when I started working with Wikchamni speakers as evidenced by their pronunciation of my name ce^ci. The data support both the northern and southern diffusion chains. An examination of data from nearby languages, such as Salinan, Chumash, Kitanemuk, and Serrano also demonstrates the strength of evidence for the southern diffusion chain particularly. With the broad diffusion patterns that seem likely for Central California, it is not surprising that connections with Southern California appear in the data. Although I will not present the evidence here, it is clear that Kroskrity and Reinhardt are correct in postulating ties between the central and southern California groups. These ties are evidenced by both linguistic and cultural data and also point to the broad patterns discussed above. Like its neighbors, including Monachi, Wikchamni borrowed many Spanish words. For the most part, these designated items of material culture and almost half of the forms (47%) designate food or household items. It is clear that Wikchamni, like the other Yokuts dialects and the neighboring languages, was an active member in the California diffusion network. Notes 1. My first contact with Stanley Newman was by mail shortly after I had started working with Wikchamni speakers who did not live on the Tule Reservation. Dr. Newman encouraged me to continue with Wikchamni because he had so little time to work with that particular dialect. In fact he had only two weeks at the end of his summer's work in 1931 to devote to Wikchamni. Given the fact he had worked such a short time with

128

2.

3.

4.

5.

Geoffrey Gamble Mrs. Alto and Mrs. Tripp, his Wikchamni speakers, it was delightful to hear from those ladies' relatives about the "kind and thoughtful gentleman" who had visited them at the Tule Reservation in the 1930s. Although they seemed to have forgotten his name, their memories of his being a perfect gentleman were still vivid in the 1970s. The most striking phonological features of the Tule-Kaweah Group include the partially rounded central vowels [i] and [e], the velar nasal [ij], and a retroflex series of stops. Morphological traits that differ from other members of the Yokuts family include root forms, derivational and inflectional suffixes, and slightly different ablaut processes. Details specific to Wikchamni can be found in Gamble 1978. Although the bulk of these data were collected prior to 1976, I also spent parts of 1984 and 198S gathering material from a Wikchamni speaker living near Dunlap, California. The meanings of the Wikchamni words mostly agree with standard Spanish: needle, alfalfa; American, ring; rice, sugar; pack of cards, yearling calf; pocket, drunk; sheep/ lamb, button; horse, coffee; money/money box (used by only one speaker — the normal word is k'e-xa), big box; sock/stocking, trousers; undershirt, coat; carp, onion; hundred, nail; pig, sack; cross, spoon; peach (the second form from English influence?), Fresno (possibly from English); beans, wagon; cracker, cat; turkey, flour; devil, thread/ string; soap, cup; milk, file; corn, table; honey, mosquito; mule, gold; bread, potatoes; paper, pears; comb, plate; cheese, Saturday; watermelon, week; seed, saddle; soldier, hat; tobacco, tomato; grapes, cow; ditch, shoes. Compare this to the Miwok form kalso- (Broadbent 1964).

References Broadbent, Sylvia M. 1964 The Southern Sierra Miwok language. UCPL 38. Berkeley and Los Angeles: University of California Press. Crowe, Earle 1957 Men of El Tejon. Los Angeles: The Ward Ritchie Press. Gamble, Geoffrey 1978 Wikchamni grammar. UCPL 89. Berkeley and Los Angeles: University of California Press. Kroskrity, Paul—Gregory Reinhardt 1984 Spanish and English loanwords in Western Mono. Journal of California and Great Basin Anthropology Papers in Linguistics. 4.107 — 138. 1985 On Spanish loans in Western Mono. IJAL 51.231-237. Langdon, Margaret—Shirley Silver 1984 California t/t. Journal of California and Great Basin Anthropology Papers in Linguistics. 4.139—165. Newman, Stanley S. 1944 Yokuts language of California. VEPA Number 2. New York: Viking Fund Publications in Anthropology. Shipley, William 1962 Spanish elements in the indigenous languages of California. Romance Philology

16.1-21.

Some principles of Alaskan Athabaskan toponymie knowledge1 James Kari

Abstract With the recent documentation on Athabaskan toponyms and geographical terms from various parts of Alaska, we can begin to recognize some of the fundamental principles of Athabaskan place naming and environmental perception. The oral place names seem to be quite stable and conservative over time. Within Athabaskan languages, a single set of place names extends across language boundaries with regular phonological adjustment; speakers repeatedly confirm sequences of names; names are strictly reported from memory and are not coined, (i. e., places with forgotten names are left unnamed); and the earliest written sources on the names can usually be reelecited. Athabaskans abhor borrowing place names from other languages, even when they are recently incursive into an area (such as the Dena'ina of Cook Inlet, Alaska, and also Hupa in California and Apache in the Southwest). Athabaskan place names are usually quite transparent in structure, that is, they are composed of recognizable Athabaskan roots and affixes. Athabaskans avoid using personal names in place names for religious reasons. In spoken discourse toponyms and a set of directional terms and locative adverbials interact in a manner analogous to a surveyor's use of a local plane coordinate system. The profound impact of the Athabaskans on prehistoric northern and western North America is demonstrated by the distribution of Athabaskan place names in vast, continuous networks of ecosystems.

0. Introduction The Athabaskan languages are a family of about forty closely related languages that are distributed in vast territories in northern and western North America. In his book Language, Edward Sapir made a few important summary statements about Athabaskan. The Athabaskan languages form as clearly unified, as structurally specialized, a group as any that I know of. The speakers of these languages belong to four distinct culture areas — the simple hunting culture of western Canada and the interior of Alaska (Loucheux,

130

James Kari

Chipewyan), the buffalo culture of the plains (Sarcee), the highly ritualized culture of the southwest (Navaho), and the peculiarly specialized culture of northwestern California (Hupa). The cultural adaptability of the Athabaskan-speaking peoples is in the strangest contrast to the inaccessibility to foreign influences of the languages themselves. (1921:228) The attention to dialect geography and comparative reconstruction that Sapir inspired has continued to be a major theme in Athabaskan linguistics. Krauss (1973:943-949, 1979:848) has emphasized that throughout its history Athabaskan has been spoken in large dialect chains. In one or more directions there is a boundary shared with another Athabaskan dialect or language, and within the chain intermarriage, polylingualism, and linguistic and cultural diffusion are routine. The distinctions between the languages, from an academic perspective, are often ambiguous. As data in the dispersed languages become more de-

'

Nenono

ALASKÀ

TAN AN A AREA

Contweil y ^ M t Mi

McKinley

KHskon

/7

River „-etn"

'Talkeetna r- Vilage

Pedro Bay Seldovio

Map 1.

Lake

Mentasto Chistochina 1 R. GulkancL/aV

Block "Little Oshetna

>

R.

T V

Copper Center^* ontana I» Kroto CI ^i5M Sutton i t n a • Chitlni Alexander* JJg»Knik—, /r^ Eklutna , SANCHORAGE, Kustatani

Nondalton

lliemna

^

Alaskan Athabaskan toponymie knowledge

131

tailed, regional patterns, which are often wave-like, can be mapped for morphological and phonological processes (e.g. Krauss and Golla 1981), as well as for cultural traits, lexical innovations, or semantic archaisms. For the past fifteen years I have had a special fascination with place names in the various Athabaskan languages in which I have done field work. Ethnogeographic data are highly important, both to the local peoples and for a range of academic purposes, whether one's goal is merely to maintain Boasian — Sapirian standards in the description of culture-history, or to analyze topics such as foraging strategies or longterm land use patterning. In this paper I will sketch some of the ways in which the "symbolic environment" (Basso 1983:49) is expressed in two contiguous Athabaskan languages of Southcentral Alaska, Dena'ina (or Tanaina) of Cook Inlet and Ahtna of the Copper River basin. In the following sections I will 1) describe the context of my field work on ethnogeography; 2) comment on the distribution and the use of Ahtna and Dena'ina place names; 3) summarize some of the structural and semantic patterns in the Native toponomy; and 4) make a few remarks about comparative issues in Athabaskan ethnogeography and territoriality.

1. Fieldwork on ethnogeography in Alaska Detailed, linguistically adequate ethnogeographic information cannot be obtained quickly as this requires a good familiarity with both the native language and the regional geography. Except for the unpublished work of Jette (1910) on Koyukon geography and Bergsland's (1959) Aleut materials, most data on Native place names in Alaska have been recorded in the past 15 years. (Some of these sources are surveyed in Kari, forthcoming a.) This work has gone through gradual stages of refinement, from the basic dialect geography, to, in some areas, annotated ethnogeography. Since 1979 the Alaska Department of Fish and Game, Subsistence Division, has sponsored the listing and mapping of Native place names in some rural areas. Place names in six of the Alaskan languages are being recorded and analyzed in the context of dictionary projects now in progress at ANLC (Alaska Native Language Center). However, keep in mind that most of the recorded Alaska Native place names have been documented after the 1971 Alaska Native land claims act. There has never been the coordinated research on Native territory in Alaska such as has taken place in Canada and Australia in the past decade prior to land claims settlements.

132

James Kari

Most of the local subsistence territories within the Ahtna and Dena'ina language areas that were in use prior to 1918, with only one exception (Lime Village on the Stony River), are now either depopulated of Native people (e. g., the Susitna basin), or are being used in highly modified and restricted ways (e.g., Lake Clark, Copper River). In the early 1970s I realized that there were very few people who knew Native names in areas such as the Kenai Peninsula, Iliamna Lake, and Knik Arm. Consequently my goal has been to record all place names remembered by the remaining speakers. The place names are sequenced in files by water flow, downstream to upstream. The files are actively updated, and the place names are analyzed and filed with the other lexical data for the language. The field techniques I have used include having informants list names in contiguous areas, recording annotations about specific places and narratives about travel; reeliciting names from maps and other sources; numbering and then labelling and posting Native place names on USGS (United States Geological Survey) maps; having speakers draw local sketch maps; and eliciting about lacunae and contradictory information. There is now a corpus of about 1700 Dena'ina place names recorded from more than 60 speakers from all four of the Dena'ina dialect areas and about 1500 Ahtna place names recorded from about 50 Ahtna speakers from all four of the Ahtna dialect areas.2 Publications include Kalifornsky 1982 on Kenai Peninsula Dena'ina names; Kari and Kari 1982, an introduction to the territory of the Dena'ina; P. Kari 1983, which includes a list of names with a wall map in the Lime Village area; Reckord 1983, a description of some of the major Ahtna historic sites; Kari 1983, lists of 1383 Ahtna place names with wall maps; Kari 1986a, which includes a series of travel narratives of the Upper Ahtna; and Kari and Fall 1987, an annotated ethnogeography of the Upper Cook Inlet Dena'ina.

2. Notes on the distribution and use of Ahtna and Dena'ina place names The Ahtna and the Dena'ina inhabit the drainages to the south of the Central Alaska Range, the tallest mountains in North America. The Ahtna language area is compact, being centered on the valley of a single major stream, the Copper River. The swift and basically unnavigable Copper River arcs around the massive Wrangell Mountains, known as

Alaskan Athabaskan toponymie knowledge

133

Q'elt'aenO These two features dominate every aspect of Ahtna life. The four Ahtna dialects are quite homogeneous, especially in contrast to those of Dena'ina (Buck and Kari 1975; Kari 1977, 1986a: 19). One key feature of Ahtna ethnogeography is that almost all of the speakers know several hundred place names that are distributed all throughout the language area. These are high frequency names such as physiographic provinces, the major streams and mountains, points on the Copper River, and the famous village and camp sites. The remainder of the recorded names are in the local subsistence territories and are known by persons who have travelled and lived in these micro-environments. It appears that most Ahtna men knew two to four local band territories in fine detail (Kari 1986a: 153-217). In striking contrast to the Ahtna area, the territory of the Dena'ina is marked by "changes in environment which are extraordinarily great for so limited an area" (Osgood 1937:13). (Note that Dena'ina territory as documented in the 1970s is about 25% larger than as reported in Osgood 1937.) The Dena'ina are the only Northern Athabaskans with coastal territory. But note that Dena'ina bands are located on both sides of the Southern Alaska Range as well as on both sides of Cook Inlet. The Iliamna-Lake Clark and lower Cook Inlet areas are considered to be some of the most productive environments in Alaska. In fact, a case can be made that the country of the Dena'ina includes a greater range of climates and ecological zones than any other indigenous language area of North America. The Dena'ina language is marked by its extreme dialect diversity (Kari 1975). This reflects a sequence of migrations into Cook Inlet basin followed by periods in which bands were relatively isolated due to the extremely hazardous nature of the travel corridors (Kari 1988). Perhaps there was no single cultural center for the Dena'ina. 4 The speakers report the names in several localized clusters. A set of over 700 names has been recorded in the Inland area west of the Alaska Range, and all the speakers here know a majority of these names. In the Upper Inlet over 650 names are recorded, many of them having been reported by a single brilliant man, Shem Pete. The record for Lower Cook Inlet, with about 400 names in Iliamna and around Cook Inlet, is fragmentary and has been recorded mainly from four speakers. Today none of the speakers knows many names on both sides of the Alaska Range. The place names reported by Dena'ina and Ahtna speakers should be viewed as an autonomous realm of verbal knowledge. The names are known exclusively by the speakers, and they are reported strictly from

134

James Kari

memory and are not taken from maps or written records. The last speakers of Ahtna and Dena'ina are quite adamant that, for the most part, the names they know are from the aboriginal period. When speaking the Native language, most speakers make the effort to use the correct form of the Native name, keeping the post-contact English or Russianorigin names distinct. Amongst many speakers there is a sense that the emerging English-based names (Kari 1984) are insidious and part of some larger plot to eradicate their history and claims to the land. As one Lime Village woman said, "That map, he lie!" Note that the name for an Athabaskan person in the northern languages, e.g., Dena'ina quht'ana, Kutchin gwich'in, literally means 'the one who possesses a territory', and place names knowledge is a key measure of one's credentials. The Ahtna and Dena'ina display place names and territorial information with openness, affection, and consistency (cf. Basso 1983 for the Western Apache). There is no tabooing of place names, not even for the many villages that have met some tragic fate. This openness in the use of place names is in direct contrast to the conspicuous avoidance of personal names. I present here a profile of the noted authority on the Dena'ina country of Upper Cook Inlet, Shem Pete, who originally was from Susitna Station and now lives in Tyonek (Kari —Fall 1987). Shem is now 91 years of age. He has travelled on foot or by boat in an approximately 13,000 square mile area in Upper Cook Inlet. The territory which Shem knows secondarily through stories embraces much of the Dena'ina and Ahtna language areas. He knows over 600 Dena'ina names in the Upper Inlet dialect area and many more in areas to the east and south. Shem is fond of listing the sequences of names, and he is entirely consistent about the pronunciation, the order, and the location of the names. In some areas he has reported these names on a dozen or more separate occasions. Shem also consistently states that he does not know certain areas, which is another measure of the integrity of the data he does report. In the Rainy Pass area of the Alaska Range, for example, Shem lists the names of streams and camps up to the point to which he has travelled (before 1918), but the sequence stops abruptly at a stream called Q'ulch'ishtnu at Ptarmigan Pass, and he will not confirm the first major name in the Kuskokwim drainage, Idzitnu, the South Fork of the Kuskowkim River, because he simply does not know this name.5 There is a remarkable degree of concurrence amongst the speakers who know mutual territory in the pronunciation and location of these 3200 names. Clearly the place names are mental maps that have been learned

Alaskan Athabaskan toponymie knowledge

135

by careful memorization, and are a stable, conservative portion of the lexicon. There is no temptation for speakers to coin place names. The names are reported as they were learned from the ancestors, and if places have been forgotten in a depopulated area, such as the southern Kenai Peninsula, they are left unnamed (Kalifornsky 1982). Furthermore, when we have documentary information on Athabaskan place names, these names can usually be reelicited with today's speakers. Of the sixty or so Dena'ina names that were recorded prior to 1850 by Russians in Alaska, all but four or five have been reelicited. The most important source on Native place names in Southcentral Alaska from the Russian period is the Wrangell map of 1839 (Wrangell 1980). Sixteen of nineteen of the Native-origin Ahtna or Dena'ina place names on this map are used by today's speakers (Kari 1986a: 104; Kari — Fall 1987). Generally speaking the Ahtna-Dena'ina place names system emphasizes linear features, especially streams and ridge lines, while it deemphasizes minor landforms, cultural features, and high country. There is a small but comprehensive system of regional physiographic province names. All major streams are named as well as all lakes on the stream systems that have some useable resources. A stream name often is derived from the name for a nearby hill or mountain or vice versa. Names for stream mouths, headwaters, and glaciers are derived from the basic stream name. A stream name never changes in mid-course. The prominent accessible ridges, hills, and points below 5000 — 6000 feet in elevation are named. These include the hills in the upland hunting areas on the margins of the Alaska Range and the navigational points around the shores of Cook Inlet. However, in the high mountain regions only a few of the major mountains, such as Denali/McKinley, are named, and the other high mountains are subsumed into one of the regional names. Typically, cultural features, such as villages and established camp sites, are subsumed into the stream/land form names and are not a distinct structural or semantic category of names.6 There are clear principles of economy and memorizability to the AhtnaDena'ina place names system which make it well suited for long-distance overland travel. As we show in section 3, a single set of names extends across language and dialect boundaries. The names are dispersed quite evenly, and there is no haphazard clutter. There are no cases of two or more names used for the same place (cf. the multiplicity of names for the same place reported by Kuipers 1984 for the Weyewa of Indonesia). There is a very low incidence of repetition in the names. (Of the 3200 names,

136

James Kari

one Ahtna name, Taltsogh Na\ 'Yellow Water Stream', occurs just five times.)7 Speakers note that features that have low visibility from an Athabaskan point of view are not named. These include the high mountains, land-locked ponds with poor fisheries, and short isolated streams that do not readily fit into the regional trail network. There are numerous issues that are relevant to the prehistoric record here and which should be compared in other Athabaskan areas. For example, there seems to be no partitioning of the village sites into named "neighborhoods". The directional system is invoked for this purpose, not the place name system. When a village site has shifted because of erosion or scarcity of timber, there is no new place name.8 In the Dena'ina area we can detect from the place names clusters of known village sites which seem to be strategically placed for defense against Eskimo groups approaching from downriver. It is significant that most of the band names are derived from the physiographic province names or from a major stream name, e. g. the Western Ahtna are called Hwtsaay Hwt'aene, 'The Small Timber People' after the name for the upper Susitna-upper Gulkana lowlands, Hwtsaay Nene' 'The Small Timber Country'. In general, these band names must have great antiquity. In the Kenai Peninsula area, the regional name Yaghenen, and the band name Yaghetna, both derive from the place name Yaghetnu, 'Good Stream' for the Swanson river, and this cluster of terms suggests that this area was the first core occupation area for the Kenai Peninsula Dena'ina. It is possible that the few village-derived band names, such as the so-called Dashq'eht'ana, the Kroto Creek band on the middle Susitna, are not bands from the aboriginal period. (Cf. Ritter 1976:114.) The ethnogeographic data are the best sources of information on Alaska Athabaskan trails and passes, especially in areas that have been depopulated. Many of the individual place names refer to trails, e.g. a stream into the upper end of Lake Clark, Chuqutenghehtnu 'Stream of Cache by the Ascending Trail'. A stream into the Copper River near Lower Tonsina, Tatl'ahwt'aene Qaltende 'Where the Headwaters People's Trail Ascends'. Most trails have no distinct name because they are either coterminous with the streams or are identified by naming the points traversed. All accessible mountain passes, however, are named for the associated stream. Control of mountain passes is a key theme in this region's prehistory. In recent years information on over twenty-five passes around the perimeter of Dena'ina territory has been salvaged (Kari and Kari 1982:55 — 57). Upper Ahtna territory is situated in a web of mountain passes (28 have been documented) which no doubt contributed to

Alaskan Athabaskan toponymie knowledge

137

the reputation of the Tatl'ahwt'aene, 'The Headwaters People', as the warriors and gatekeepers of the Copper River. One feature of Ahtna territory that is highly symbolic of Ahtna culturehistory is a system of inherited chiefs titles that appears to be unique in Alaskan Athabaskan (Kari 1986:15). The chiefs of at least twelve of the most important villages had an inherited title that was based upon the place name of the village plus a term denen or ghaxen, e. g. the chief of the Gulkana area village is Hwkae'e Denen, 'Person of the River Mouth'. The distribution of the titled chiefs reflects important nodes in the territory: seven are in the lower Ahtna dialect area, the cultural and economic center of the Ahtna; one (Gulkana) is at the hub of the Copper River in the Central dialect; two are in the Western dialect (Tazlina and Tyone lakes), the main villages on the trade corridors leading to Cook Inlet; and two (Slana and Mentasta) are in the mountainous headwaters, the stronghold of the specialists in warfare, the Upper Ahtna. For the speakers I have worked with the place names are a kind of index system to which they attach all sorts of factual, historical, and mythical information. The connection between places and stories is very much like what Basso (1983) describes for the Western Apache. Here are two annotations by Shem Pete from Kari —Fall (1987). Nutnasdlenitu 'River that Flows Opposite' uppermost fork of Talachulitna River from Beluga Mountain (translation from Dena'ina) On the other side of Hnidi (Beluga Mountain) is a stream that flows that way, toward the rising sun. It flows down around Hnidi and then flows back in an upriver direction. That is Nutnasdlenitnu. Hnidi is its headwaters. Htal 'Wide Place' Rainy Pass (English quote) From Bak'nin'iy to Ptarmigan Pass there was a village. There were many people there, lotsa people. Nobody knows how many thousand Natives that catch caribou there. Not much timber there. The people there they call Shandala Nununk'dnelghuyna ['people who cause the migrating birds to fall by hollering']. Those geese when they fly over, fall time, they all holler. Maybe geese go up 20 to 30 miles. Nichil [houses] is just up close, all along that river, I guess. They make scare

138

James Kari

the geese. They holler for miles, and the geese get all scared. Pretty soon they just get tired and they fall down. They pick em up. They eat em up too. Of course. Yeah. Good geese. I wouldn't let em go. Shandala Nununk'dnelghuyna were year-round citizens up there. An especially valued genre in these languages is the travel narrative. In the Upper Ahtna book (Kari 1986a) a series offivetravel accounts in overlapping areas is presented. In the book 370 different places are mentioned and considerable effort was devoted to locating the places and translating the names consistently. Fred John, chief of Mentasta village, in one 25-minute travel narrative in which he summarized eight travel routes in his area, mentioned 120 Ahtna place names and 76 different places. Fred also mentioned one English place name (Kari 1986a: 194-206). I have noted that the genre of mythical stories of man-animal interaction, called sukdu in Dena'ina, and yenida'a in Ahtna, are conspicuously ageographic and lack place names (e.g. sections I —III in Tenenbaum 1984). In fact one major criterion for defining historical or non-fiction stories in Alaskan Athabaskan, in contrast to myths, is the grounding of events at named places (Kari 1986:9).9

3. Structural and semantic patterns in Ahtna-Dena'ina place names There is a fair amount of evidence in Northern Athabaskan areas that indicates that a single set of place names extends across language and dialect boundaries with regular phonological adjustment. In the past bilingualism between Ahtna and Dena'ina was common (Kari 1977), and on their common border about 70 shared place names have been recorded, but not a single pair of independent names for the same feature has been recorded. In Table 1 is a set of bilingual Dena'ina-Ahtna names from the border area and a set of Ahtna-Upper Tanana names from that shared border. In each of these names the differences in pronunciation can be explained systematically, e. g., by a shift in the Dena'ina vowel system, or the use of a different root for 'stream' in Upper Tanana. This kind of systematic variation has gone unrecognized in the names Tenada::Denali for the tallest mountain in North America. Tenada, the earliest recorded name (Wrangell 1980 [1839], is from Ingalik [deijads], whereas Denali is from Koyukon [di: nae: li], both meaning 'The Tall

Alaskan Athabaskan toponymie knowledge Table 1.

139

Bilingual Athabaskan place names

place Talkeetna R Matanuska R Chickaloon R

Dena'ina name K'dalkitnu Ch'atanhtnu Nuk'din'itnu

Ahtna name I'delkuut Na' Ts'itonh Na' Nay'dini'aa Na'

place Nabesna R Jacksina R Cooper Cr

Ahtna name Nabaes Na' Tsik'elggodi Na' Tsighaan' Na'

Upper Tanana name Nabia Ndiign Tthich'ehgodi Ndiign T t h i x ^ ' Ndiign

One'. How sad it is that persons bent on debunking Native traditions in Alaska, such as Alfred H. Brooks (1911:22),10 have used linguistically naive variant spellings of Native place names to assert that the Native toponomy is amorphous and ever-changing. However, the linguistic evidence is very much to the contrary. The variant pronunciations of Athabaskan place names are almost always systematic, and are strong proof of the antiquity of the names. In fact, in Alaska the Native place names from the oral tradition seem to be vastly more stable than the official map names. As Sapir (1921:209) noted, Athabaskan languages are "highly resistant" to borrowing from other languages. This is especially evident in the lack of borrowed names in the place names systems for groups that have been incursive into their modern territories in the proto-historic period, such as the Dena'ina of Cook Inlet (Kari 1988) or the Navajo and Apache of the Southwest. We find that the Ahtna and Dena'ina place names are virtually all Athabaskan in structure, i.e., they are composed of transparent Athabaskan roots and affixes. Although there are some unanalyzable or partially analyzable names, usually these look Athabaskan in structure. There are virtually no exceptions to this pattern in the Ahtna corpus. Dena'ina historical traditions indicate that there has been southern incursion into Cook Inlet basin on two fronts (Kari 1988). Yet there is no noticeable substratum of non-Athabaskan place names in Cook Inlet. One interesting exception is the name for Kasilof River on the Kenai Peninsula, Ggasilatnu, which is a blend of a non-Athabaskan word iggasila) with an atypical intervocalic voiceless fricative s, and the Dena'ina stem for 'stream'. Dena'ina and Ahtna place names appear in several canonical patterns. These are summarized in Table 2.

140

James Kari

Table 2.

Ahtna-Dena'ina place name patterns Ahtna

Dena'ina

-na' -ina' (rare)

-tnu

1. specific + generic stream

-

river mouth straight portion of river headwaters glacier pass riverbank lake head of lake lake outlet land, country mountain

-tu', -tuu' (rare) -kaeq'e -tayene' -tl'aa luu, -luti' -tates bes, -bese' ben,-bene' -q'etl'aa -q'estsiiq'e nen,-nene' dghelaay, -dghelaaye'

ridge hill

dzei, -dzele' ses, -yese' tes, -tese'

flat, plain, base trail

ken, -kene' -tene

-

-niq' (rare) -

-kaq' -tayena, -tuyana -tl'u, -tl'ughu Ii, -/i'a -tustes, -sustes bes, -besa ben, -bena -q'atl'a -q'estsiq' einen, -nena dghili'dghelay, -dghU'ti-dghil'u -dghelaya dzel, -dzela ses, -y its'a lex íes h, -texei-tesha ch'vaq'da, -ch'vaq'da (rare) ken, -kena -ten

Other nominalizations, without generic noun compound postpositional adjectival relative, 'the one which, that which' at specific place

noun + noun noun + pp noun + adj -i

noun + noun noun + pp noun + adj -i

-den

-t

The first type is a name composed of a specific plus a generic, where the generic term is usually possessed. About twenty geographical nouns are used generically, e. g., Ahtna Tsiis Tl'edze' Na' 'Blue Ochre Stream', Tsiis Tl'edze' Kaegge 'Mouth of Blue Ochre', Tsiis Tl'edze' Tl'aa 'Headwaters of Blue Ochre'. Also note that there are a number of other geographical nouns commonly used in the languages which rarely occur generically in place names, e. g., nuu/ni 'island', tabaaghe/tubugh 'beach'. There are very few names which consist only of an unmodified generic, e. g. 'hill', 'outlet',

Alaskan Athabaskan toponymie knowledge

141

etc. The second group consists of five types of nominalizations without one of the generic terms: a) compound, Ahtna Tsa Qaen' 'Beaver Lodge', b) noun followed by postposition, Ahtna T'aghes Tah, 'Among the Cottonwoods', c) noun plus adjective, Ahtna Ben Ke'e, 'Big Lake', d) nominalized verb with -i relative suffix, Ahtna Tsiniggets'i, 'Head that is Twisted', and e) nominalized verb with -den/t locative suffix, Ahtna Tsedi Qulaendert, 'Where There is Copper'. A prime area for comparative work is to assemble tables such as this for all Athabaskan languages. The congruity in the Ahtna-Dena'ina system is striking. Even on the shores of Cook Inlet the Dena'ina use riverine toponyms, e.g., the deep bays in lower Cook Inlet are named with the term -kaq\ 'river mouth'. The generic terms used in toponyms in Interior Alaska Athabaskan are distinct from those in Ahtna-Dena'ina to some degree. For example, north of the Range the Ahtna-Dena'ina term for 'glacier', is used to mean 'slab of ice'. Some languages north of the Range use the noun gat to refer to glacier. The cognate noun in Ahtna and Dena'ina means 'frozen overlow' or 'aufeis'. Through comparison of toponymic structure, we can sometimes detect signs of prehistoric changes in language boundaries. In Kari 1988 I show how the unique Dena'ina root for 'stream', -tnu, with a -tn- cluster, is used in restricted areas both in the Western Ahtna dialect and in the Kuskokwim dialect of Inaglik which suggests former areas of occupation by Dena'ina speakers outside of the modern language area. Furthermore, it may be possible that Dena'ina occupation of the Susitna River preceded that of the Ahtna because the Ahtna name for this major river, Sasutna', with the -tn- cluster, has probably been borrowed from the Dena'ina name, Suyitnu. It is very interesting to trace how place names are used in spoken discourse. Place names interact with both directional words and locative adverbials. In Ahtna there is a set of nine directional roots, which can be derived into 216 directional words (Kari 1985; also Kari and Kari 1982:43—46). Two-word directionals are also possible. One compound sentence by Fred John (Kari 1986:203) is presented here with the directional words in regular type and the place names capitalized. Yihwts'en nahtedel ts'en' xona DESUUN' NA' tsiidze qanqedel ts'en' 'uniit TADINILTS'AEGGE yet nahnedel. From there they would start back and they came up from the stream, GOOD AREA STREAM, and they camped at a place upriver, NARROW WATER.

142

James Kari

It seems that the Athabaskan use of place names in combination with directional words is somewhat analogous to a cartographer's use of a local plane coordinate mapping sytem. There are hundreds of options for precise description of locations in relation to the named features. Graphic representation of Athabaskan travel narratives is needed to convey this. There have been several semantic typologies of Northern Athabaskan place names data in recent years (Ritter 1976; Caulfield — Peter — Alexander 1983; Cruikshank 1984). These are important contributions to the record of long-term land use in these areas, especially since all signs indicate that Athabaskan toponymy is quite static. Most of the Ahtna and Dena'ina place names have concrete, straightforward meanings that refer to some aspect of the regional natural history. Hydrology is a very prominent category. Many names refer to the critical food resources such as salmon, ground squirrel, and caribou. It appears that the animals whose actual names are tabooed to promote the luck of the hunter such as marten, wolf, and wolverine (De Laguna 1969 — 70), are rarely named in the place names.11 There are very few place names that refer to the personal names of Ahtna or Dena'ina people, at least not by their proper teknonymic name. It would be an overt insult to the ancestors for Athabaskans to intermix personal names and place names. It is also interesting that the names of Ahtna-Dena'ina mythical figures rarely occur in the place names. A rather small number of names (7% of the Ahtna names) refer to man-made structures such as caches, houses, and caribou fences. One interesting question is the nature of historicity in the place names. Less than 4% of the Ahtna names denote prehistoric or historic events. Some of these are customary events, such as a fording place, Ts'aqae Nanalyaesde 'Where the Women Are Carried Across'. Others are single events. For example, the Ahtna name Uti'sneldziits'i, 'The One We Danced out To', is a point on the lower Copper River that figures in the upriver migration of one of the clans from the coast. In Cook Inlet sites of wars or former Alutiiq (or Sugpiaq) presence are noted in six Dena'ina names with the word Ulchena, the name for the Alutiiq. It seems that most of the event names are aboriginal. The few post-contact event names we have detected that refer to activity by Russians or Americans are interesting. Katie John states that the place near Batzulnetas, K'ekenn'gha 'By the Stumps', refers to the site where a group of Russians was cremated after they were killed by the Upper Ahtna. This incident may involve the Samoilov party in the winter of 1794-95 (Kari 1986a: 75-87). The first

Alaskan Athabaskan toponymie knowledge

143

subterranean burial in the Upper Ahtna area is noted in the place name for a hill, Dzii Qoley, 'The Deaf One'. We speculate that this was in about 1830 (Kari 1986a: 188). The first American on the Susitna River (probably in the early 1870s) is commemorated in the Dena'ina name Sdigel Qak'ghalnikt 'Where Window Pane (nickname) Pitched a Tent'. Five Dena'ina names and three Ahtna names contain Russian loan words and are therefore, post-contact, e.g. Dena'ina Valatga Q'elchini, 'The One that Looks Like a Tent'. None of the Ahtna-Dena'ina names is a Native translation of a Russian or English place name. Religious meanings in the Athabaskan landscape are sometimes overt as in the mourning place the Dena'ina in Upper Inlet call Ch'chihi Ken 'Ridge Where We Cry' (Kari and Kari 1982:1). Usually these associations are covert and require some explanation before we, as outsiders, can recognize their existence. One such name is Ahtna Neqets'alyaexden, 'Where We Turn Around', a hill on the upper Matanuska River where the last view of the sacred mountain Q'ell'aeni, Mts. Wrangell/Sanford, is available. It was to this point that the Ahtna would carry a handful of soil from their homes as they started down the trail to Cook Inlet. At Neqets'alyaexden they turned around, scattered the soil, and made a prayer for a safe journey (Kari 1983: vii —viii). No Dena'ina has told me that Nduk'eyux Dghil'u, 'Mountain Which Animals Enter', Telaquana Mountain west of the Alaska Range at the heads of the Stony and Mulchatna rivers, is a sacred mountain. We can infer that it is sacred because 1) this mountain figures in a major story in which two non-game animals, the pika and the northern shrike, release the game animals from here (Tenenbaum 1984:178 — 92) and 2) because Nduk'eyux Dghil'u is in the very center of an ancient homeland of the Dena'ina west of the Alaska Range. (Kari 1988:328). Table 3.

Athabaskan place names as folk poetry

language

place name

translation

location

Ahtna

Nay 'dliisdini 'aaden

Ahtna Dena'ina

Saghani Ggaay Detl'uule' Nanelna'den Qichi Qinghilneqt

site near Copper Center point on lower Copper R ck into Chulitna R

Dena'ina Dena'ina

Q'in Tetl'i Yuyan Ach'edelt

'Where Songs Extend Across' 'Where Raven Forgot His Rope' 'The Old Lady Made It that Far' 'Exploded Fish Egg' 'Where We Walk to the Sky'

hill W of Lime Village pass N of Lake Clark

144

James Kari

Athabaskan speakers frequently refer to the poetic dimensions of their place names, which evoke for them images of the landscape and the ancient past. It is possible with proper translations and formats to present these names as folk art. I present in Table 3 a list of some of my favorite place names from these two languages.

4. Some comparative issues in Athabaskan ethnogeography Ethnogeographic data, in the Sapirian mode, should be analyzed both in terms of the local native language in a specific context and in a comparative and diachronic perspective. The topics I have surveyed here for two Alaskan languages, if applied elsewhere, could clarify issues such as the strategic character of the Athabaskan continental advance, and the nature of the original homeland prior to the exodus of bands from the north. In the southern Athabaskan area, for example, are there shared or independent Apachean toponyms for intervisible features? Is the structure of the toponyms the same throughout Apachean? What can be salvaged about the southern periphery of Apachean territory? There are several good sources on Navajo place names that are available for comparison (Van Valkenburgh 1941; Jett 1970; Young —Morgan 1980). Almost all Navajo place names are transparently Navajo in structure. There is very little borrowing despite the complex language contact situation between Navajo and neighboring groups. As in the north, in the Navajo names there is little reference to post-contact history, and personal names are very rare. The specific + generic pattern for Navajo is not as elaborate as in Table 2 for Ahtna-Dena'ina. The most common type of Navajo place name has the -i relative suffix. Some interesting semantic shifts can be detected in some geographical terms in Navajo; e. g. Navajo -kooh 'arroyo' is cognate with Carrier -koh 'stream' and the Dena'ina postposition -qugh 'along, while'. The Navajo term for 'hillside, slope', naniid, is cognate with the Ahtna directional naniit 'at a distance upstream'. There has been little work to date on the nature of border territory along the vast interface between Athabaskan and Eskimo peoples. There are a number of features about Eskimo place naming that seem quite different from that of the Athabaskans. For example, there seems to be a much higher density of names in Eskimo areas and more attention to cultural features. It seems that Eskimo names can consist of an unmodi-

Alaskan Athabaskan toponymie knowledge

145

fied generic, e.g., 'hill', 'creek', etc. Pete (1984) states that 22% of 223 Central Yupik names in the Stebbins-St. Michaels area are unmodified generics, and this raises several questions of contrast between Athabaskan and Eskimo toponomy. It also appears that Eskimos may borrow Athabaskan place names when they are incursive into Athabaskan territory, or at least they do so more often than do Athabaskans in the reverse situation. The Alutiiq place names on Kachemak Bay, most of which are of Dena'ina origin, are a good illustration of this pattern. Just surrounding the Dena'ina language area several different patterns in border relations are found. On the middle Kuskokwim River there is a large zone of bilingual place names and evidence of shared territory and much cultural exchange between the Central Yupik and speakers of three Athabaskan languages. Not far away, on the Mulchatna River there is an abrupt boundary and no shared territory or dual place naming. The coastal territory of the Iliamna Dena'ina was probably the last portion of Cook Inlet that the Dena'ina occupied (Kari 1988). In 1981 I recorded names from the last Dena'ina who knew this coast, the late Nick Kolyaha, then age 89. One of the southern-most names that he reported is where the border with the Alaska Peninsula Alutiiq should be placed. This is a mountain on Kamishak Bay called Naq'ezhch'en, 'Our Side, Our Half.

Notes 1. It is a special honor for me to dedicate this paper to the memory of my former professor, Stanley S. Newman. I would like to thank Keith Basso, Hal Conklin, and Bob Young for comments made on an earlier draft. I acknowledge support for this research from a fellowship from the National Endowment for the Humanities in 1985 — 86 while I was based at the Newberry Library in Chicago. 2. The actual inventory of toponyms could be increased by 10 — 20% if I were to generate all the names for stream mouths and stream headwaters. 3. The practical writing systems for Ahtna and Dena'ina, which follow conventions typical of Athabaskan alphabets, such as that of Navajo, are used in this paper with these modifications to facilitate comparison. The Ahtna and Dena'ina front and back velars are represented here as g, k, k ' : : gg, q, q'. The voiceless back velar fricative [x] is written here as x. The Dena'ina voiceless front velar fricative is written here as x. Note that e in Dena'ina has the quality of [a], schwa, but in Ahtna is higher, [e], or epsilon. Ahtna long vowels are written with doubled letters, e. g. ii high front, ae low front, uu high back, aa low back. See Kari 1975 on Dena'ina phonetics and orthography, and Buck and Kari 1975 on Ahtna. 4. The most central location in the oddly shaped Dena'ina language area is the Kustatan Peninsula on the western shore of Cook Inlet. A series of passes through the Alaska

146

5.

6.

7.

8.

9.

10.

James Kari Range meet here; this is a key strategic location offering good visibility up and down the inlet; this is the best place for crossing the inlet by boat. Called Qezdeghnen, 'Point Land', this was most likely one of the first areas on the inlet occupied by the Dena'ina. A series of lifetime maps and battery of statistics based upon the oral place names data will be especially instructive. The distance between the end points of the areas known by experts on Ahtna and Dena'ina territory is often 400 to 500 miles. Speakers from the Yukon-Tanana basin, such as Chief Andrew Isaac from Tanacross, know Athabaskan place names that are well over 600 air miles and probably over 1000 overland miles apart. Almost none of the village names employ the common word for village, Ahtna qayax, Dena'ina qayex, as a generic term, e. g. ' Village'. The only exceptions to this are three ancient village sites in Dena'ina, all called Suk Qayex 'Ancient Village' and one Ahtna place called Dala Qayax 'Dala's Village', which is said to have been a postcontact settlement of people with leprosy. Keith Basso (personal communication) reports that Western Apache place names also have a very low rate of repetition. In a corpus of 600 place names in a 40 miles square area around Cibecue, Arizona, only two names are repeated. Archaeologists from the University of Alaska Museum, amidst 248 sites discovered in five field seasons, failed in three or four attempts to find the site all the older Ahtna call Nilben Kaegge, somewhere near the confluence of the Tyone and Susitna rivers (Dixon et al. 1985: chapter 3, p. 41). This site may now exist exclusively in the ethnohistoric record. This absence of place names in mythical stories is in sharp contrast to what McLendon (1977) reported for the Eastern Pomo where virtually all myths are grounded at named places. Note, for example, this misleading but often quoted statement by Alfred H. Brooks, the prominent geologist in early 20th century Alaska, on Native geographical knowledge:

The Alaska Indian has no fixed geographic nomenclature for the larger geographic features. A river will have half a dozen names, depending on the direction from which it is approached. The cartographers who cover Alaska maps with unpronounceable names, imagining that these are based on local usage, are often mislead [sic] (Brooks 1911:22). 11. An exotic exception to this is the Ahtna name Naitsiis Ltaex Na' 'Stream of the Wolverine Lying Dead'.

References Basso, Keith 1983

"Stalking with stories": names, places, and moral narratives among the Western Apache. In Text, Play, and Story: the Construction and Reconstruction of Self and Society. Edited by Edward M. Bruner. Washington: Proceedings of the American Ethnological Society. Bergsland, Knut 1959 Aleut dialects of Atka and Attu. Transactions of the American Philosophical Society, 49, 3.

Alaskan Athabaskan toponymie knowledge

147

Brooks, Alfred H. 1911 The Mount McKinley region. USGS Professional Paper, 70. Washington: U.S. Government Printing Office. Buck, Mildred—James Kari 1975 Ahtna noun dictionary. Fairbanks: Alaska Native Language Center. Caulfield, Richard A.—Walter J. Peter—Clarence Alexander 1983 Gwich'in place names of the upper Yukon-Porcupine region, Alaska. Alaska Department of Fish and Game, Subsistence division. Technical Paper, 83. Cruikshank, Julie 1984 Tagish and Tlingit place names in the Southern Lakes region, Yukon Territory. Canoma 10.1:30-35. De Laguna, Frederica 1969—70 The Atna of Copper River, Alaska: the world of men and animals. Folk 11-12:17-26. Dixon, E. James—George S. Smith—William Andrefsky—Becky M. Saleeby—Charles J. Utermohle 1985 Susitna hydroelectric project, cultural investigations, 1979—1985. Anchorage: Alaska Power Authority. Jett, Stephan C. 1970 An analysis of Navajo place names. Names 18:175 — 184. Jette, Jules 1910 On the geographical names of the Ten'a. Spokane: Gonzaga University, Ms. file 14, drawer 13. Kalifornsky, Peter 1982 Kahtnuht'ana Qenaga, The Kenai people's language. Second edition. Fairbanks: Alaska Native Language Center. Kari, James 1975 A classification of Tanaina dialects. Anthropological Papers of the University of Alaska 17(2): 4 9 - 5 3 . 1977 Linguistic diffusion between Tanaina and Ahtna. International Journal of American Linguistics 43:274 — 288. 1983 Ahtna place names lists. Fairbanks: Copper River Native Association and Alaska Native Language Center. 1984 Place names at Lime Village: Native vs. non-native names. Alaska Native News 2(7): 8 - 9 . 1985 A note on Athabaskan directionals. International Journal of American Linguistics 51(4): 471-473. 1986b The Tenada-Denali-McKinley controversy. Names 34.2: 241 —244. 1988 Some linguistic insights into Dena'ina prehistory. Late prehistoric development of Alaska's native people. R. D. Shaw, R. K. Harritt, and D. E. Dumond, eds. Aurora, Alaska: Anthropological Association Monograph Series 4: 319 — 339. forthcoming Native place names in Alaska: Trends in Policy and Research. Names. Kari, James—James A. Fall 1987 Shem Pete's Alaska, the territory of the Upper Cook Inlet Dena'ina. Fairbanks: Alaska Native Language Center. Kari, James—Priscilla Russell Kari 1982 Dena'ina elnena, Tanaina country. Fairbanks: Alaska Native Language Center.

148

James Kari

Kan, James (ed.) 1986a Tatl'ahwt'aenn Nenn', the Headwaters people's country. Fairbanks: Alaska Native Language Center. Kari, Priscilla Russell 1983 Land use and economy of Lime Village. Alaska Department of Fish and Game, Subsistence Division, Technical Paper, 80. Krauss, Michael E. 1973 Na-Dene. Current Trends in Linguistics 10:903-978. 1979 Na-Dene and Eskimo-Aleut. In The Languages of Native North America. Lyle Campbell and Marianne Mithun, eds. Austin: University of Texas Press, 803-901. Krauss, Michael E.—Victor K. Golla 1981 Northern Athapaskan languages. In Handbook of North American Indians, Vol. 6, Subarctic, June Helm, ed. Washington: Smithsonian Institution, 67-86. Kuipers, Joel C. 1984 Place, names, and authority in Weyéwa ritual speech. Language in Society 13:455-66. McLendon, Sally 1977 Cultural presuppositions and discourse analysis: Patterns of presupposition and assertion of information in Eastern Pomo and Russian narrative. In Georgetown University Roundtable on Languages and Linguistics. Muriel Saville-Troike, ed. Washington D.C.: Georgetown University Press, 153 — 190. Osgood, Cornelius 1937 The ethnography of the Tanaina. Yale University Publications in Anthropology 16. New Haven. Pete, Mary 1984 Yup'ik place names: Tapraq, a case study. Unpublished M. A. Thesis. University of Alaska, Fairbanks, Dept. of Anthropology. Reckord, Holly 1983 Where Raven stood, cultural resources of the Ahtna region. Fairbanks: Cooperative Park Studies Unit, University of Alaska. Ritter, John 1976 Kutchin place names, evidence of aboriginal land use. In Dene Rights, Vol. 3, Indian Brotherhood of the Northwest Territories, Yellowknife, 111 — 135. Sapir, Edward 1921 Language, an introduction to the study of speech. New York: Harcourt, Brace and Company. Tenenbaum, Joan 1984 Dena'ina sukdu'a, Tanaina stories. Fairbanks: Alaska Native Language Center. Tom, Gertie 1987 Ekeyi: Gyo Cho Chú, my country: Big Salmon River. Whitehorse: Yukon Native Language Centre. Van Valkenburgh, Richard F. 1941 Diñé bikéyah. Window Rock, Arizona: United States Department of the Interior, Office of Indian Affairs, Navajo Service.

Alaskan Athabaskan toponymie knowledge

149

Wrangell, Ferdinand Petrovich 1980 [1839] Russian America, statistical and ethnographic information. Translated from the 1839 edition by Mary Sadouski. Richard Pierce, ed. Kingston: The Limestone Press. Young, Robert W.—William Morgan 1980 The Navajo language, a grammar and colloquial dictionary. Albuquerque: University of New Mexico Press.

Deified mind among the Keresan Pueblos Jay Miller

Abstract The Keresan belief in a supreme deity, who has been called Thought Woman in the literature, is explored here. The belief revolves around many dimensions with symbolic oppositions involving Man/Woman, Sacred/Secular, and Nature/ Culture. As an androgynous being, with the potential to assume any form but particularly that of a spider, this creator truly presides over a rational universe.

Stanley Newman did not introduce me to the literature on the Keresan Pueblos, but he certainly made me more aware of its complexities. As my undergraduate advisor, he had the difficult task of guiding me through shifting loyalties to archaeology, linguistics, and symbolic anthropology. My rudder through it all was a fascination with the Keres that was continually reinforced by my exposure to their prehistory as a member of the Anasazi Origins Project, their ethnography through reading, their public rituals as an observer, and their language via examples used by Newman from his fieldwork with Laguna. Eventually, the Keres took me to graduate school and provided a dissertation. They continue to fascinate me, but now in comparison with other Native American tribes where I have done more sustained field work (Miller 1980, 1985). Yet one aspect of Keres culture stands out above all others. This is their belief in a high god or supreme deity who has been often called Thought Woman in the literature. Recently, several authors, themselves belonging to Laguna Pueblo (Parsons 1923), have also provided commentary on this being. Among the things that impressed me in Newman's classes was his interest in the more subjective side of linguistics, particularly his interest in the sound symbolism and psychiatric dimensions of speech. Assessing both the older literature on this deity and modern commentary, our understanding of this "female" deity should improve. It is not that such a being is unique to the Keres, after all the Shawnee pray to Grandmother and the southern Numic to Ocean Woman.

152

Jay Miller

The earliest account in this century occurs in the curious volume by Gunn (1917:89): Their theory is that reason (personified) is the supreme power, a master mind that has always existed, which they call Sitch-tche-na-ko. This is the feminine form for thought or reason. She had one sister, Shro-tume-na-ko, memory or instinct. Their belief is that Sitch-tche-na-ko is the Creator of all, and to her they offer their most devout prayers, but never to Shro-tu-me-na-ko. They say it is bad to do so. This shows that they know of the two divisions of the mind, reason and instinct, and also that they are aware of the apparent uselessness, and possible evil consequences, of cultivating the subjective mind. While there is much in his book which is questionable, as in some of this paragraph, his basic statement does hold. The creator at Laguna is something like personified thought or reason. This was confirmed a few years later in the work of Parsons (1920:114), where she names the four sisters who figure in the Origin Saga: iyetiku, tsichinnako, naustiti, ushstiti. In a footnote (1920:114, n. 3) she adds, for the second, "The etymology given is tsichu, think, chinnaku, femaleness." Again from Laguna, Boas (1928:7, 276), building on the fieldwork of Parsons, makes reference to ts'its'i.'na.k'o and to "Thought Woman (the Spider?)." He includes the only available text in Keresan (Keresic) making reference to her role as creator of the universe. At the end of their discussion of the ceremonial calendar at Laguna, Dutton and Marmon (1936:20) call attention to "Reason, a great power — the Great Spirit, we might call it — had created earth, the sun, the stars, and all living creatures." This is again the deity in the guise of an English equivalent. The most careful statement, however, has come from a member of the pueblo, attempting to compare Keres and Christian notions of the deity (Purley 1974). For him, the hallmarks of Thought Woman are the concepts of a supreme being, a female, a fused godhead, and a denial of human dominion over the world. In the course of this, he reports Tse che nako is all-comprehensive and in no need to be worshipped, therefore she does not demand worship for herself to satisfy "Her Own." "Her Own" includes all life possibilities within herself, (p. 30) The Keres people believe that Tse che nako has more female than male attributes: therefore she is referred to and approached as if she is female, (p. 30)

Deified mind among the Keresan Pueblos

153

Tse che nako is not limited to a female role in the total theology ... she is both Mother and Father to all people and to all creatures. She can function in whichever role she chooses and very often does ... (p. 30) She did not restrict the process of creation only to herself. Tse che nako included the power to create individual thought in all human beings and all creatures. In other words, all living things can create, although it is a matter of degree, (p. 31) Keres holy men hesitate to mention Tse che nako's name, especially for purely secular discussions. Thought Woman's name is reserved for use only in sacred ceremonies. In secular discussions and teachings, Tse che nako is often symbolically referred to as Old Spider Woman or Spider Woman, (p. 31) ... evil, while it is a separate, recognized force, is not so strong that it must be blamed for mankind's wrongdoing. Mankind is responsible for its own behavior, (p. 32) Curiously, Purley does not mention that aspect of the Christian tradition which is closest to Keresan notions of divinity, namely Sophia as Wisdom and feminized Mind. Lastly, we have the poem that begins the famous novel by Silko (1977), herself of Laguna ancestry, paying homage to "Ts'its'tsi'nako, ThoughtWoman" (line 1), Thought-Woman, the spider (line 10), who is "sitting in a room" (line 14) thinking the story we are about to be told. Laguna is one of seven contemporary Keres pueblos in central New Mexico, and, to some extent, the most divergent. About 1870, after the railroad came through their lands, the pueblo split into different groups. While all of the other Keres are nominally Roman Catholic, Laguna had a sizeable Protestant segment, the outcome of intermarriage with American men of authority. The more traditionally conservative and Catholic party left Old Laguna and settled at Isleta, a Tiwa Pueblo just south of Albuquerque. After a few years, many of them came back to Laguna land and built the town of Mesita, although their religious paraphernalia, or much of it, stayed at Isleta (Harvey 1963). As I have reconstructed it, Keres social organization is binary at many levels, all of them pervaded by an axiomatic concern with gender. Thus, of the seven modern towns, three (Cochiti, Santo Domingo, and San Felipe) are manly, characterized by leaders drawn from the priesthoods and by the initiation of only boys into the cult of the masked Katsina.

154

Jay

Miller

The other four (Santa Ana, Sia, Acoma, and Laguna) are womanly, with leaders selected for both matri-clan and priesthood and the Katsina initiation of both boys and girls. Further, there is also something like an intensity scale in which Cochiti and Santa Ana are the most strongly consistent with this ideology and Santo Domingo and Laguna, both of which have many external links with other cultures, are the most diffuse. Hence, we need to consider the belief in this deity among other Keres before we can properly appreciate all of these references from Laguna. To date, the best published statement on this deity appears in the last volume of Leslie White's (1962:113) comparative study of the Keresan towns. The most important deity in Sia cosmology is Tsityosti.nako, "Prophesying Woman" ... This deity is found at Santa Ana ... and at Laguna ... also. But everywhere the conception appears to be unclear and even inconsistent. Stevenson treats this deity as a male, but in her emergence myth Sussistinnako is addressed as "our mother" (in Keresan pueblos the cacique, a man, is ceremonially addressed as "mother"). The ending -nako means 'woman.' But at Laguna she "looked like a man" (Boas, 1925, pp.221, 228). Stevenson says that Sussistinnako was a spider; my informants, that Tsityostinako "had the shape of a certain kind of spider." Tsityostinako is called Prophesying Woman because "she knows [rather than deciding or determining] what is going to happen;" one informant added: "when a person is thinking about something that is Tsityostinako expressing herself in him." Tsityostinako lives at Shipop in the Yellow world, "but she is everywhere, like God," one informant said. She is the creator in Sia cosmology as she is at Laguna (Gunn, 1917, p. 89). She bore two daughters, Utctsiti, the mother of the Indians, and Naotsiti, the mother of other races and peoples. As this quote confirms, this deity is associated with thought in several modes, with spider, and with creation as an ongoing process. The 'thought' etymology is supported by recent grammars. For Santa Ana, Davis (1964:170 #455) lists cidnistA as 'to think, to worry.' From Acoma, Wick Miller (1965:84, 109) has ?icitistaani 'mind, willpower' and -'ucitistaaN 'to think'. In a text, Spider Woman gives aid to the War Twins (Miller, 1965:253). For the manner in which the original creation took place we must refer to two volumes dealing with the Acoma account. By a curious twist, the

Deified mind among the Keresan Pueblos

155

account of the saga was published by Stirling (1942), but the most important aspect of the process, the songs which enabled creation to occur, was published by Densmore (1957). Together, these versions make clear that thought has a pulsating vitality most like song. It is this parallel that has been overlooked in previous accounts of the saga, such as the summary by White (1960). The spider attribute relates both to the form of the arachnid body and its web: a center with extensions. The web is a particularly apt metaphor for thought because the rhythmic pulse of thought waves mirrored in the songs of creation and curing flows out as ripples and rays from the source and summary which is this deity. The presence of two terms for this deity is a reflection of the important distinction between the ordinary Keresan language used by everyone and the ceremonial vocabulary used by men in the kivas, first reported for Laguna by Hrdlicka (1903) and since confirmed for all Keres (White 1944, Fox in Lange 1959: 558). All of the evidence assembled indicates that Thought Woman, as Mind, is not so much an entity as a nexus of many important dimensions, the crux of a series of symbolic oppositions, involving Man/Woman, Sacred/ Secular, and Nature/Culture. As a manlike being with female attributes, with the potential to assume any form but particularly that of a spider, this creator truly presides over a rational universe.

References Boas, Franz 1928

Keresan texts. Memoir 8. Part I: English. New York: The American Ethnological Society.

Davis, Irvine 1964 The language of Santa Ana. Bureau of American Ethnology, Bulletin 191, Anthropological Paper 6 9 : 5 3 - 1 9 0 . Densmore, Frances Music of Acoma, Isleta, Cochiti and Zuni. Bureau of American Ethnology 1957 Bulletin 165. Dutton, Bertha—Miriam Marmon 1936 The Laguna calendar. University of New Mexico Bulletin 283. Anthropological Series 1(2): 1—21. Gunn, John 1917 Schat-Chen: history, traditions and narratives of the Queres Indians of Laguna and Acoma. Albuquerque: Albright and Anderson. Harvey, Byron 1963 Masks at a maskless pueblo: the Laguna Colony Katsina organization at Isleta. Ethnology 2.4:278-289.

156

Jay Miller

Hrdlicka, Ales 1903 A Laguna ceremonial language. American Anthropologist 5:730—2. Lange, Charles 1959 Cochiti. A New Mexico Pueblo, past and present. Austin: University of Texas Press. Miller, Jay 1980 High-minded high gods in North America. Anthropos 75(4): 916-919. 1985 Shamanism in western Native America: Numic, Salish, and Keres Pueblo. In Woman, poet, scientist: Essays in New World anthropology, honoring Dr. Emma Louise Davis. Compiled and edited by the Great Basin Foundation. Los Altos, California: Ballena Press. Miller, Wick 1965 Acoma grammar and texts. Berkeley: University of California Publications in Linguistics 40. Parsons, Elsie Clews 1920 Notes on ceremonialism at Laguna. Anthropological Papers of the American Museum of Natural History 19 (Part 4): 85-131. 1923 Laguna genealogies. American Museum of Natural History, Anthropological Papers 19.5:131-282. Purley, Anthony 1974 Keres Pueblo concepts of deity. American Indian Culture and Research Journal 1(1): 2 9 - 3 2 . Silko, Leslie Marmon 1977 Ceremony. New York: New American Library. Spencer, Robert 1940 A preliminary sketch of Keresan grammar. University of New Mexico, M. A. Thesis. Stirling, Mathew 1942 Origin myth of Acoma and other records. Bureau of American Ethnology Bulletin 135. White, Leslie 1944 A ceremonial vocabulary among the Pueblos. International Journal of American Linguistics 10:161—67. 1960 The world of the Keresan Pueblo Indians. In Culture in history essays in honor of Paul Radin, pp. 53—64. New York: Columbia University Press for Brandeis University. 1962 The Pueblo of Sia, New Mexico. Bureau of American Ethnology Bulletin 184.

Creek curing in academe1 Willard Walker

Abstract This essay describes my relationship with a college instructor, who was also a Creek Indian, and goes on to summarize his experiences with Creek curers and conjurors. It makes inferences as to the current vitality and prevalence of Creek and other Indian American curing and conjuring practices and identifies certain domains into which these practices appear to be expanding. It also attempts to provide a new perspective on the roles of traditional medical personnel and their patients in the context of urban institutions and a reinterpretation of the complex interdependencies that may develop between the fieldworker and the informant in non-traditional settings.

1. The fieldwork Good social science research is usually the result of careful planning and assiduous fieldwork. There are times, however, when a rich vein of data opens up quite unexpectedly, of its own volition, and in unlikely circumstances. Such was the case when Ha jo told me his story. It happened in a very institutional setting, on a college campus in Oklahoma, where I was assigned in 1978 to develop a curriculum for an Indian Studies program. In retrospect, it is not so surprising that an academic colleague, who happened also to be a Creek Indian, should gratuitously blurt out a stream of recollections that would turn out to involve the practice of Indian medicine on college campuses and have a direct bearing on my notions of the nature of culture contact and change and the interdependence of Euroamerican and Indian American institutions. Nor is it surprising that he should do so in college classrooms and at odd moments between classes and other academic appointments. Indeed, the fact that much can be learned about such matters from 'acculturated' people like Ha jo and that institutional settings can be fertile grounds on which to collect certain sorts of data may be the chief conclusion to be drawn from his testimony and the most crucial point of this essay. Data on cultural continuity and change must be sought where cultures collide,

158

Willard Walker

e. g., in schools, jails, bureaucratic agencies, housing projects and military services. I could scarcely have selected a more natural setting in which to elicit information from Ha jo than the campus classroom and office building to which he regularly commuted to work. I did not seek Ha-jo out. He happened to have a desk just down the hall; and I soon became aware that he had a considerable following among the Indian students, to many of whom he was a friend as well as instructor, athletic coach, and all purpose advisor. He was curious about the Indian Studies program I was designing and offered a number of suggestions. In the course of our conversations I gradually began to learn random bits of information relating to his experiences with curing and conjuring, practices with which I also had some familiarity and in which I had an intense interest. I was eager to learn more and asked him for a comprehensive account of his experiences. He tried to oblige me, but our conversations were invariably curtailed or interrupted by a variety of academic obligations and unanticipated emergencies. And there were other obstacles to communication. He spoke with great excitement and emotion, leaping about in time and space and making profligate use of third person pronouns whose referents were seldom identifiable. I soon began to take notes at these interviews, which continued over a period of three months. After each encounter I would try to construct a chronologically ordered narrative of what had happened and a set of questions to be asked at our next meeting that would clarify temporal and spatial relationships, specify the participants in each episode, and supply other details of what was said and done at various times and places. As my version of his experiences took shape I continually reviewed it with him for confirmation and further clarification; and he was given a final draft, which he examined and approved. This version of his account differed in several respects from his own, oral account, however. It was typed on paper and could be read as a single unit. It was chronologically ordered. It included some information that he had not volunteered, but that I had elicited, like the nature and amount of fees paid to medical practitioners. The referents of pronouns were specified, and the dramatis personae for each scene were enumerated. Moreover, my own version was intentionally concise, whereas Ha-jo's had been expansive and repetitive. Finally, I had inserted fictitious names for people and places, in order that the cast of characters might not be identified, or not be identified easily, or in any case, not be identified with absolute certainty. One of the fictitious names I inserted was, of course, that of Ha-jo himself, who was mildly amused and rather pleased, I think, to find that

Creek curing in academe

159

I had assigned him that particular alias. He was not particularly interested in disguising his identity, or that of his parents or the medical personnel, or even that of a certain acquaintance who had conjured him. But Hajo is a goodnatured fellow and made no effort to dissuade me from assigning him his alias. He knew that /ha-jo/ was a component of many traditional Creek war names, and that Crazy Snake, the leader of the Creek resistance movement which opposed the dissolution of the Creek Nation at the turn of the century, was known in his own language as 'cito ha-jo.' He would not have known that /ha-jo/ was glossed as 'crazy, mad, furious in battle' by Swanton (1928:101); but, as one who had at least a good passive knowledge of Creek, he could not have denied that Ha-jo was an apt description of his own behavior in crisis situations, such as that reported in (my version of) his story that is soon to follow. The version that follows, however, differs even more radically from Hajo's oral account than the typed version I gave him in December, 1978. Of the six episodes he reported to me, only the last is described here in detail. The others have been reduced to mere abstracts, partly to conserve space, partly because, in this form, they serve my present purposes. It is, I think, all right to edit his text, so long as we all remember that what follows is not what Ha-jo said. It may bear some resemblance to what he might have written, had he been of a mind to write it in this way.

2. The American dream fulfilled Ha-jo's daddy learned carpentry in the 1930s at Chilocco Indian High School, which is near Ponca City and the Oklahoma-Kansas line. After graduation he worked as a carpenter for some years and later became a construction supervisor and building inspector. He was also a devout Christian and life-long member of a certain Creek Indian Baptist Church in which his own father (Ha-jo's grandfather) had been the pastor. At various times in his life he had found it useful to consult Indian doctors. In the 1960s, for example, he took his oldest son, who was about to go to Viet Nam, to a hilis haya 'medicine maker', who talked with him and gave him a charm which he wore while he was in military service. He credited his charm with protecting him during the Tet offensive, when his friends were killed all around him but he survived unharmed. He lost his charm after returning to the United States and, soon after, was almost killed in an automobile crash that put him into a coma for two and a half days.

160

Willard Walker

In the fall of 1975 the two youngest sons in the family were leaving home to become teachers at an Indian school in the southwest. On this occasion Ha jo's daddy retained a Creek kila 'knower' to prepare them for their journey and protect them from any dangers to which they might be exposed living among strange Indians so far away from home. Ha-jo, a fourth son, was 27 years old at that time and was also teaching, but at another institution, close to home, in eastern Oklahoma. It was not thought necessary to consult a medical practitioner in his case until a series of events that began in January 1976 caused Ha-jo and his daddy to consult the same ki la for counsel in what they considered to be a crisis situation. Ha-jo had graduated from high school in 1965 and had gone to an eastern Oklahoma college in the fall of that year, but dropped out of college in 1968 to get married and go to work. He was employed at a naval ammunition plant, then in road and building construction, and later yet as a roughneck in the Texas oil fields. He returned to the college in the spring of 1972 and graduated in December of 1974. In this second stage of his college career he was so successful that, even before he graduated, he was offered a job as instructor at his old high school, where he began teaching in January 1975. The following summer he was offered a job as an instructor at an eastern Oklahoma college, where he began teaching in the fall of 1975, having arranged for a friend from his undergraduate days to replace him at the high school. In the fall of 1975, then, when his two younger brothers had gone to the southwest, Ha-jo was teaching at the college, developing an academic program, and acting as faculty sponsor for several student athletic activities and clubs involved in Indian music and dance. While teaching at the college he and his wife continued to live in a Creek Nation housing development, not far from his parents' home and in the town where he had been teaching at the high school. All during the fall semester he drove his 1975 Pontiac Grand Prix to and from the college, a distance of some 80 miles each day.

3. The ki la All went well for Ha-jo during the fall semester at the college; but in January, 1976, some unusual events occurred which will be summarized below. The first incident occurred one evening on his way home from work. He was unable to steer his Pontiac around a curve in the road and proceeded on a straight course across the path of an oncoming Mack

Creek curing in academe

161

truck, which passed on his right with its horn blaring. "I just couldn't think anything or say anything. I was sort of in a trance. I couldn't move my wheel." A few days later Ha jo was driving home when his Pontiac unexpectedly veered to the left and crossed the path of an oncoming police car. He was unable to turn the wheel. He came to a stop on the lefthand shoulder and remembers seeing the revolving red light of the police car, as it came up behind him, in his rearview mirror. The third incident occurred at about 7:15 one morning as he drove to work, not in his gray Pontiac, but in his wife's blue Gremlin. I must have driven onto the railroad track there and stopped. I guess I just sat there for a while. I remember hearing the train whistle; and I guess that brought me back to consciousness, because I looked up and saw the train coming down the track; and that's when I jerked the stick into low gear and gunned that old blue Gremlin off the track. During this period Ha jo saw his parents frequently and discussed these incidents with them. They suggested that he was probably being conjured with Indian medicine. Ha-jo responded that he was not in contact with any Indians who had either the incentive or the esoteric knowledge necessary to conjure him. His parents were unconvinced, but nothing was done. A few days later Ha-jo worked late at the college and left for home at about 9:30. Evidently he drove off the highway on an old, discontinued road leading into a lake which the Army Corps of Engineers had made some years earlier when they built a dam farther down the valley. The old road was used only as a boat ramp. Ha-jo evidently drove down into the lake, because he found himself at the wheel of his Pontiac which was surrounded by water. He came to his senses when he heard truck horns up on the highway. He put the car in reverse and backed out of the lake. After this incident, Ha-jo began to give more credence to his parents' counsel. His wife was suffering from a prolonged illness at this time and the family thought that her illness and his difficulties might both be caused by conjuring. It was decided that they should see the ki-la who had been consulted a few months before when Ha-jo's brothers had gone to the Southwest. Before Ha-jo saw the ki-la, however, there was a fifth incident. Ha-jo was driving home in the Pontiac and came over the crest of a hill and skidded into a pickup parked in the road ahead. He is unable to remember anything that happened after he went into the skid until a few moments later when he found himself running down the road in pursuit

162

Willard Walker

of the pickup, which had been pushed forward by the impact. He had a scratch over one eye and cuts on his arms from the shattered windshield. He had evidently been thrown against the steering wheel, had gotten out of the car through the righthand door, and gone running down the road after the pickup. After this accident the Pontiac was taken to a garage for a 'complete overhaul', and Ha-jo drove a car borrowed from the garage for some six weeks without incident. On the Sunday after the pickup incident, a cold day in February, HaJo went with his wife and his daddy to see the ki-la. Like Ha-jo and his family, the ki-la was a Baptist. They had to wait for him to return from church. The ki-la also shared their faith in 'education', having insisted that each of his children finish high school. The ki-la asked whether Ha-jo understood Creek, and was evidently pleased to learn that he did. He asked Ha-jo a number of questions, most of them in Creek. Ha-jo responded in English. His father interpreted, although it is likely that both Ha-jo and the ki-la understood most of what the other said. Ha-jo was asked where he worked, the nature of his job, and what sorts of people he had contact with at work. He was asked whether he believed in Indian medicine. The ki-la said that Indian people sometimes get jealous of people they think are too successful and will try to bring their adversaries down even at the risk of their own welfare. He also said that he had had a premonition that something bad would happen to one of the boys in Ha-jo's family and had had a 'real bad feeling' on the day that Ha-jo skidded into the pickup. Ha-jo's daddy told the ki-la that Ha-jo's wife had been told by White doctors that her illness was 'terminal' and asked him if he thought her medical problems were related to Ha-jo's problems. The ki-la asked if she had had any experience with Indian medicine. She had not. Then he said he thought her disease was not terminal and said, Tell her she can be helped by Indian doctors. Whiteman's medicine can't help her much. Her troubles may be exaggerated by the problems that Ha-jo's having; so maybe an Indian doctor can help her. The ki-la began deliberating on Ha-jo's case, chanting and nodding his head rhythmically with his eyes closed. Meanwhile, Ha-jo's daddy prayed softly, but audibly, for a successful diagnosis and resolution of the case. Ha-jo listened to his daddy pray and listened also to the ki-la, chanting and nodding in his chair. The ki-la interrupted his chanting from time to time to ask questions or ask for confirmation of some point. He began to outline the course of events, asking for further elaboration on various points. Had Ha-jo's wife been with him on any of these occasions? (She

Creek curing in academe

163

had not.) He said he could see a big "metallic gray car" and made specific reference to all or most of the four incidents involving the Pontiac; but he did not mention the Gremlin or the incident involving the Gremlin and the railroad track. They had to tell him about that. The ki-la said H a j o was very lucky to have survived. "My mind sees a good friend as the cause of this. These things were caused by jealousy." The ki la sometimes interrupted his chanting to engage Ha-jo's daddy in small talk. He inquired about the health of Ha-jo's grandmother and the activities of various mutual acquaintances. But then he would return to his chanting or ask Ha jo another question. "Who are you close to that might be jealous? Is there anyone that you have trouble with or have trouble communicating with?" Finally the ki la asked whether Ha jo had a friend who lived near him and was in the same "line of work." He's not a Creek. His medicine is weaker than Creek medicine. You've fought off his medicine with just your will power alone. He's not a Creek. He might be a Cherokee. After considerable prompting of this sort, Ha jo's daddy finally, said, softly to Ha jo, "Isn't a Cherokee?" The ki-la said, "Don't tell me the name. I know the name already. Don't tell me the name." In that way they succeeded in identifying the conjuror. The ki la then asked a great many questions about this man, his family, employment record, and relationship with Ha jo. In particular, he wanted to know if there had been a change in this relationship during the past month or two. Ha jo then told the ki la everything about the Cherokee and the history of their relationship in minute detail. By the end of the visit no one present had any doubt as to the identity of the conjuror, his motives, or the fact that he had both the ability and the opportunity to conjure Ha-jo. Before they left, the ki la recommended a hilis haya 'medicine maker', not the one who had given the charm to Ha-jo's older brother, but one that the ki la thought might be able to help Ha jo and perhaps his wife also. The ki-la told Ha-jo, in parting, to "get over there as soon as you can."

4. The hilis haya Ha-jo and his daddy visited the hilis haya that same evening, right after supper. He lived in a 'camp house' at a Creek Baptist church ground about 10 miles away. He was not home, but at about 10:00 p.m. his

164

Willard Walker

son and apprentice, a man in his late forties, came out and invited them in. A few minutes later the hilis haya arrived, rubbing his hands for warmth as he entered. He had been out digging roots, which he had in a sack. He said the ground was not completely frozen. They had coffee and Ha-jo's daddy talked in Creek with the hilis haya. Ha-jo couldn't follow the conversation very well, because the hilis haya, a man in his seventies and an ordained minister, well known for his sermons, used verb forms and syntactic constructions with which Ha-jo was not familiar. Finally, the hilis haya got around to asking Ha jo's daddy the purpose of their visit. He was told that they had been referred to him by the kila, who thought he would verify that Ha-jo was being conjured. The hilis haya listened as Ha-jo's daddy recounted their tale, smoking Ha jo's daddy's cigarettes or sometimes his pipe, that he filled from a can of Prince Albert in his hip pocket. Occasionally he asked questions in Creek, most of which Ha-jo answered in English, his daddy interpreting. Was Ha jo's wife involved in any of the mishaps? Did she believe in Indian medicine? Ha jo's daddy told him about Ha jo's wife, her illness, and what the White doctors had told her. The hilis haya sat listening and smoking and looking at the floor. Sometimes he talked, telling about Creeks who had achieved prominence long ago and had been attacked by jealous conjurors. He also told how he had protected some of his own patients and about the herbs that he had used. All this was interspersed with the account of Ha jo's experiences and his relationships with his wife, his friends and neighbors, and the people he knew at work. Finally, the hilis haya gave them his analysis. He said that one of Hajo's friends was conjuring him with medicine that he had gotten from a Cherokee hilis haya. He said Ha jo's whole insides, from the eyes on down, were filled with this medicine and that Ha-jo was blinded. There was probably a lot of medicine in the car too, as well as in Hajo. It was not a Creek that was using this medicine. Creeks don't use medicine against an individual. They put it in the house; and it works against the whole family. But this medicine was outside, in the car probably, and in Ha-jo himself. Ha-jo's wife's illness and the marital problems that they had been having of late were not caused by Indian medicine, although they might have been exaggerated by Ha-jo's condition, which was the direct result of medicine. The hilis haya asked about Ha-jo's Cherokee friend. In particular, he asked whether the Cherokee smoked dope and whether Ha-jo had ever smoked tobacco or marijuana with him before the incidents occurred.

Creek curing in academe

165

Ha jo answered that he had smoked with him, but not since the first incident in early January. The hilis haya said this was how Ha-jo had got filled up with medicine, smoking cigarettes and grass with the Cherokee, and that there must be medicine in the car. It controls the car. He's probably using medicine on things outside the house. When you drove the Gremlin onto the railroad you had control of the Gremlin; but you didn't have control of yourself. I know it's a Cherokee by the way he's using the medicine. Cherokee medicine is good against Cherokees, but not very effective against Creeks. Ha-jo has a strong will; and he's been able to fight off the medicine. The hilis haya promised to provide medicine for Ha-jo and his wife. He had all the ingredients except one root that he could not dig until 4:00 a. m. He had to get it in the woods and pray and ask for a blessing when he got it. He said he would take his son along to dig for him, that it was hard work for an old man and that, in any case, he wanted his son to know how to get this kind of medicine. Before they left, the hilis haya told Ha-jo how to use the medicine he would fix. He asked what kind of cigarettes Ha-jo smoked and was pleased to learn that Ha-jo had a preference for 'Salem Greens,' a kind of mentholated, filter cigarette. He said the mentholated tobacco would mask the odor of the medicine that he would put in the filter. He said the doctored cigarettes would be in the back of the package of cigarettes with unmodified cigarettes in the front, where the package was opened, in case anyone asked Ha-jo for a smoke. The doctored cigarettes would have a slight stain in the filter, although they would not have been lit. Ha-jo was to walk all around the building where he worked, blowing medicine smoke on it from each direction. He was also to blow smoke in each corner of the classrooms where he taught and all around the casing of the front door. He was also to give his office a thorough smoking. Finally, he was to keep the cigarettes at hand and, before mingling in any crowd of people, smoke his clothing and his hand, so he could rub the smoke over his hair. There were also to be two liquid medicines in which he and his wife and little daughter were to bathe every morning before sunrise for four weeks and every night before bed time. He was to mix half a cup of each with a tub of water to bathe in. He was to use this solution with a bar of lye soap, which was also doctored. After bathing he was to dry himself in the air without using a towel. As this medicine had a disagreeable

166

Willard Walker

odor, resembling that of old leaves that have lain in stagnant water, he might take a shower after drying in the air. He was told to use the liquid medicine also on his house and car. He should dip his fingers in it and sprinkle the corners of the house outside, the corners of all the rooms inside, and the floor, seats, engine, and trunk of the Pontiac, as well as its window frames and, as best he could, the chassis, wheels, brake drums, etc. Ha jo got the medicine the next day and followed his instructions for some three weeks, until he left on an overnight trip and forgot to pack the medicine. He felt so healthy by this time that, on his return, he failed to return to his strict regimen. (His wife stopped bathing herself and her daughter after only one week.) Some time after this, perhaps in April, the Cherokee conjuror was convicted and placed on probation for possession of narcotics, resisting arrest, and contributing to the delinquency of minors. He was told that he would lose his job as a high school teacher at the end of the spring semester.

5. A Creek antagonist Hajo's most alarming experience occurred in May of 1978 and is believed to be the result, not of the efforts of the Cherokee conjuror, but of those of a new antagonist who arrived at Ha jo's college as a transfer from Haskell in the spring of 1978. He was, according to Ha-jo, a "twofaced, boisterous, braggart"; he and Ha-jo developed a warm and mutual antagonism. They were of about the same size; and both were Creeks, but from localities that were some distance apart. The student was known to Ha-jo as being from "a long line" of Creeks who used medicine. He claimed he was my friend; but he told other people he was going to whip me. He hated my guts. He was always trying to beat me. He used to box over in the gym and he asked me to come over there and spar with him to help him train. He told people he was going to whip me when he got me in the ring; but, when the time came, he couldn't touch me. I kept jabbing; and he never could get past my left to get at me. Then a few days later Ha-jo came back to his office late, after umpiring a softball game. It was already dark when he got there. He was going to pick up some things before going home; and, as he was getting his things together, he heard people arguing outside the window.

Creek curing in academe

167

I went out to see what was the matter and it was this Creek student arguing with a girl. It was not his girlfriend, either, that he was engaged to marry, but another girl. He had been drinking and he was getting ready to slap this girl around, so I told him to leave her alone. He said it was none of my business and told me to back off. He said he'd let me off easy in the ring but that this time he'd really let me have it if I didn't back off. I told the girl to take off. He grabbed at her, but she slipped away. I told her to go on and that I'd hold onto him. She left and then I let go of him. He said, "Don't put your hand on me again or I'll kill you." Then he came at me with his fists and said he was going to "whip my ass." I said I wasn't going to fight with him, but if he kept on coming I'd hit him just once and drop him. He swung and missed and came on again and I knocked him out with one punch. I got some water and put it on his face to revive him and got him up and asked if he was Okay. I said I didn't mean to hurt him, but that he was out of his head and I had to hit him. He said he wasn't out of his head and that he was still going to beat up that girl. Then he said that, when he was at Haskell, a boy had died up there. "I can kill people without having to touch them," he said. "If I can't, I've got some strong medicine that can. You're going to be finding out what I can do." I said he was talking nonsense. Then he said he didn't want anything more to do with me, but that something was going to happen to me, and that I had forced him to do this. It was about two weeks later that Ha jo had the accident. It was a Wednesday night. He was living in an apartment near campus and was at home typing up a test. His wife was away at the time; and a Creek friend who was staying with him had gone to bed at about 12:30. Hajo was still typing. After a while one of his students, a Navajo named Claude, came by with his girlfriend and knocked on the door. They said they were hungry and would I take them somewhere for something to eat. I was hungry too. It was about 3:00 a. m. so I took them across town to where there was a 'Seven-Eleven' place that stayed open all night. On the way back to the college, Claude asked me to take him down to the river. His roommate was down there with a bunch of students who were having a party. The roommate had all Claude's money and Claude wanted to get it before his roommate lost it or spent it.

168

Willard Walker

Ha-jo, the Navajo, and the girlfriend were all in the front seat as they approached the river. The road goes right down within 12 or 15 feet of the river bank and then makes a 90° lefthand turn. They were talking as they approached the turn and Ha-jo kept thinking the turn was farther away; but Claude said, "You're going to miss the turn." Then Ha-jo slammed on the brakes; but his foot slipped off the brake pedal and onto the accelerator. The car slowed down and then leaped ahead when he hit the accelerator; and they went off the ten foot bank and into the river. There were two creosote poles there. The grill hit them and pushed them apart, and the car went right between them into the river. The lights were still on in the car but the buttons that open the windows wouldn't work and they couldn't get the doors open. The lights were on and we could see the water rising inside the car. I tried to bust the window with my fist and elbow; but it wouldn't give at all. So I got my feet up and started to kick the window out. Finally the glass sort of 'laid over' so I could push it out; and when I pushed it out I remember being surprised because the water came in against my face. I swallowed a lot of water. Then I grabbed the window frame with both hands and pulled myself out through the window hitting the side of my face on the window frame. I came up to the surface thinking Claude and his girlfriend were still down in the car there, that I had to get them out. It was dark and I was coughing because I'd swallowed the water; but I heard voices saying "This way, this way." It was people on the bank. There's a strong current there, so I landed quite a way down river from where the car had landed. Ha-jo came ashore, still thinking that his friends were trapped in the car and that he had to get them out. The people were trying to tell him that the other two were safely ashore already; but Ha-jo couldn't hear them. He kept breaking away from them and diving back in the river, desperately trying to find the car, drifting ashore again, and diving back in, despite all attempts to restrain him. Finally Claude, himself, got a hold of Ha-jo and yelled at him that they were all okay. But even then Ha-jo thought only of his car in the river and that he must get it out somehow. He dove back in the river in a vain attempt to find it. It was only after that that they overpowered Ha-jo and took him home. The next morning he and his Creek friend, who had been asleep at his apartment during this episode, returned to the river bank. Ha-jo dove in again but still couldn't find his car. He went back to his apartment, showered, and went to his 10:30 class. After class he called the Dean and

Creek curing in academe

169

told him what had happened. He thought he might lose his job and get Claude and his girlfriend into trouble; but the Dean told him not to worry. Commencement exercises were held that same night. Ha-jo was reluctant to go, thinking everyone would have heard about his accident. It appeared, however, that very few people had heard the news, only the students who had been at the river, the Dean, and a very few others. No one had been hurt, except Ha-jo himself when he hit his face coming out of the car. The Creek student who had threatened Ha-jo two weeks earlier left school on Thursday morning, the day of the accident. Just before leaving, he went to the campus security and to the Dean and reported Ha-jo for being drunk and partying with the students down at the river. It turned out that he had not been going to classes for some time and that he had his bags all packed when he went to the party on the river bank on Wednesday. On the same Thursday a Cherokee friend of Ha-jo's went home and talked to a Cherokee doctor who advised that Ha-jo should see a medicine maker and also said that he thought this accident was supposed to have killed Ha-jo and that he was lucky to be alive. Not long after, Claude also sought advice. He went to see a Navajo Peyote man who told him that he hoped Ha-jo would see a doctor, that the medicine was too strong for Ha-jo or Claude to deal with alone. He told Claude not to talk about this incident with anyone. The Creek friend who had been staying over night with Ha-jo also sought advice. His doctor, a Creek, said that whatever was in Ha-jo's car had been intended to kill him, but that its power had since been neutralized by the water. After the May accident Ha-jo's parents urged him to see the hilis haya. He planned to do so; but it was a busy time of year. He had to teach again in summer school; and, for one reason or another, he never visited him again prior to the time I last spoke with him in December, 1978. He felt that he needed to see him and expected to do so, but had not gotten around to it. His wife made no attempt to dissuade him from going, although she said it would do no good. "What's done is done." Ha-jo still had his medicine. He was doubtful about the efficacy of the liquid medicine, since it had been intended for protection against Cherokee medicine, not Creek. But he continued to use the cigarettes occasionally as a prophylactic medicine, as when he was unexpectedly called in to the college president's office. False rumors that Ha-jo had been remiss in his duties had reached the president. Ha-jo smoked his clothing and his hair, went into the president's office, gave a spirited

170

Willard Walker

defense of his curricular and extra-curricular activities, demanded an opportunity to face his accusers, and spoke to such good effect that the interview lasted some three hours and ended cordially.

6. Afterword Those who doubt the vitality of native curing and conjuring in the 'Five Civilized Tribes' of eastern Oklahoma should note that, when Ha jo totalled his Pontiac in 1978, his three closest friends all consulted medical practitioners in their own ethnic groups. One was a Creek, one a Cherokee, and one a Navajo; and all were associated with the college, one as a student and the other two as administrative staff. All sought medical advice and reported it to Ha jo. It seems evident that, while native curing and conjuring may have declined in such contexts as hunting, horticulture, and weather control, the complex remains viable in the context of social relationships and has expanded to embrace such relatively new activities as living with strangers in housing projects, driving with them on the highway, and competing with them for jobs. Another conclusion that may be drawn from Ha-Jo's testimony is that such White institutions as schools and colleges need not always be viewed as aggressive forces assaulting and seeking to dominate passive Indian enclaves. Sometimes they serve as arenas where Indians vie with one another in accordance with their own goals and understandings and adapt old techniques to new conditions, just as their ancestors must always have done. Significantly, these modern antagonists continue to rely upon elderly medical practitioners with little formal education in White schools to analyze situations and prescribe remedies on the basis of a system of beliefs that would seem quite alien and bizarre to any ordinary White dean, faculty advisor, or guidance counselor. Thus, schools, jails, courtrooms, hospitals, and military barracks can be viewed as new theaters of activity in which Indian American cultures are given expression and in which elderly traditionalists exercise a degree of real, if remote, control. The ki la and then the hilis haya listened attentively to Ha-jo's automotive problems, elicited additional information on certain matters that seemed critical to them, and developed a plausible explanation to account for the strange occurrences. They helped Ha jo identify the conjuror and provided him with a set of medicinal substances and explicit directions for their use. By so doing they succeeded in dispelling Ha-jo's anxieties, focusing his attention on a well defined, external enemy, and imbuing him with a sense of confidence and control based largely on his faith in

Creek curing in academe

171

the efficacy of the medicine and his certain knowledge that more medicine, expertise, and counseling would be available in time of need. The hilis haya's magic permitted Ha jo to stride manfully into the president's office, enveloped in his invisible armor and proof against all perils. There is one final point to be made. It often seemed that, when Ha jo was recounting his story and I was trying to rework it in a form that I could understand, we both felt, and communicated to one another, a sense of urgency and excitement about our mutual enterprise. I was not simply performing the role of an anthropologist gathering data from an informant; there was far too much affect involved. And I feel certain that Ha-jo too, had a very different image of our relationship. He did not go back to the hilis haya, as his parents had urged him to do. He came to me instead, not to play the role of informant in the service of the science of anthropology, but for some intangible reward that he could probably not have specified. I think he saw in me a surrogate hilis haya who could, by manipulating a different set of symbols, give him a massive dose of reassurance, confirm his identity as a Creek, and validate the Creek values, categories, and institutions which he and I, together, constantly endorsed in each of our discussions. I was, after all, a kind of academic hilis haya. I had come there to design an Indian Studies program. I had read Swanton and Mooney. I could swap tales about Indian doctors and their miraculous cures and soul-searing conjurations. I knew how to listen attentively and how to ask questions that were oblique enough not to be intrusive. Most importantly, I talked to Ha-jo, not as I would to a college instructor with a college degree, a credit card and a less than perfect ability to speak Creek, but as a fellow communicant with whom I shared a faith in the power and authority of Indian medicine and in its ability to restructure and restore the perceived world. As we talked, the power flowed within, between, and around us. How could I have ever doubted the power of the hilis haya when the charmed circle he had cast around Ha-jo came to embrace me as well? And how could Ha-jo have ever doubted the power of those mentholated cigarettes resting in his pocket, ready to hand, when I, a stranger from the east, so obviously believed?

Note 1. The interpretive section at the end of this article, for which I bear sole responsibility, owes much to the insightful commentary of Robert Conkling, who read a preliminary draft.

172

Willard Walker

Reference Swanton, John R. 1928

Social Organization and Social Usages of the Indians of the Creek Confederacy. Forty-Second Annual Report, Bureau of American Ethnology, Smithsonian Institution 1924-1925, pp. 23-472. Washington, D.C.: U.S. Government Printing Office. Reprinted, Johnson Reprint Corp., 1970.

C. Grammar and Discourse

On the unit of paragraph analysis in formal monologue discourse1 Betty Lou Dubois

Abstract To determine how biomedical scientists make their first public communication of research results to professional peers, transcriptions of fifty-two slide talks from the 1979 annual meeting of the Federation of American Societies for Experimental Biology were first divided into major discourse divisions. Then, to discover the structure of the major divisions, the Longacre apparatus for identification of paragraph types and constituents was employed, in preference to the traditional semantic paragraph, the graphic paragraph, and the conceptual paragraph. The question of the basic unit of paragraph analysis was resolved against the tone unit, in favor of the syntactic unit.

To gather data for an ongoing study of how biomedical scientists make their first public communication of research results, I am studying poster sessions and slide sessions at annual meetings of the Federation of American Societies for Experimental Biology (FASEB), slide session being the name for the ones at which papers are read. I have called these papers speeches, since few are literally written out in full and read. Already included, or to be included eventually, are studies of vocabulary and syntax of all fifty-two papers which I taped at the 1979 FASEB meeting and for one session of twelve papers from the 1979 meeting, detailed intonation and discourse structure. Considering a discourse to be a piece of verbalization complete for its own purpose (Moffett 1983) which has structure (Halliday and Hasan 1976), I have analyzed the major divisions of the biomedical discourses (Dubois 1980), comparing them to biomedical journal articles. The next step is to discover the structure of the major divisions, for which purpose I have chosen the Longacre apparatus for identification of paragraph types (Longacre 1980). The Longacre paragraph is different from the traditional semantic paragraph of freshman composition, which consists of topic sentence with supporting sentences; from the graphic paragraph, which is a visual entity; and also from the conceptual paragraph (Lackstrom, Selinker, and Trimble

176

Betty Lou Dubois

1973) of English scientific and technical writing, which is a core generalization with support and which can be (1) a semantic paragraph, (2) a semantic paragraph so long it is broken into two graphic paragraphs, or a short essay, complete with thesis statement and two or more semantic paragraphs, the whole embedded into a longer discourse. The canonical Longacre paragraph consists minimally of a pair of sentences judged by the analyst to relate to each other in specified ways. Figure 1 shows examples of such relations, or paragraph types,

Paragraph marker

Paragraph type

Constituent marker

Constituent

(AA

Antithetical

(CE (EA

Coordinate list Amplification

(EP

Paraphrase

(TCP

Process/ Chronological Process/ Simultaneous

(AAT (AAA,...,, (CEMln (EAT (EAA, ...„ (EPT (EPP,...„ (TCPSj „

Thesis Antithesis Item Text Amplification Text Paraphrase Step

(TSPS! ..„

Step

(TSP

Figure 1.

Code for selected paragraph types and constituents

paragraph constituents, and the coding scheme I have devised. There is an occasional orphan sentence, which forms a simple paragraph. Longacre (1980) notes that the "constituents of a paragraph are not necessarily sentences," but rather slot classes into which paragraphs may be embedded. He is silent, however, on whether paragraph constituents can have less than sentence length. The question of paragraph constituents shorter than sentences has arisen as a byproduct of my decision to use the word tag and syntax analysis scheme of Svartvik et al. (1982) in order to obtain a measure of comparability with the large scale Study of Spoken English. Svartvik explains the choice of tone unit as the basic analytic unit (Aarts 1984): "At stage one, the analysis is carried out on four levels within the boundaries of prosodic chunks called 'tone units', one at a time. This approach is based on the assumption that tone units are valid communicative units in spoken English." Because I too am dealing with spoken materials, I found it reasonable to take the tone unit as the analytic entity.

Paragraph analysis in formal monologue discourse

177

Some relevant characteristics of tone units follow. The upper limit of length of the tone unit (TU) is surely physiological, i. e., we simply cannot speak 487 words in one TU, and TU length in narrative bears a relation to the clause (Chafe 1980). Whether from psychological or sociological causes, there is individual variation in length of TU in the twelve speeches of Session 414 of 1979 FASEB, which I consider here. The average for the session is 5.47 words/TU, which is also the median; the range of average word/TU for individual speakers is 4.38 to 7.5 (Dubois 1982). The word count is slightly problematic and may even appear smaller than it actually is, since I count such expressions as delta PC02, in vitro, Morelis melotus varieagatum, formalmethianilphenolalanine, and three point nine seven as one word each, but the count has a reasonable degree of internal consistency. For present purposes, a more interesting figure than average words per tone unit is the range of words/TU among speakers, which is 1 — 18. The lower limit, one word, is obviously not determined physiologically. I hope to show that one of the determinants of TU length is discourse structure, as revealed by a Longacre paragraph analysis. Sometimes, of course, a one-word unit simply corrects a slip, as below, for speaker C1S4A1. (Note that what appears to be punctuation at the end of the TU's is the intonation contour code, separated from the TU by a space, to indicate its function.) C1S4A1 routinely delivers much In their pattern of breeding, Breathing, longer tone units, as indicated by some representative ones below (not in sequence). And the low ventilation exhibited during the early morning hours. That the equation was not accurately describing gas exchange. Associated with the analysis of air convection requirements in garter snakes, Function of length of tone unit is sometimes neatly revealed by Longacre paragraph typing, as in the following example from the same speaker, i. e., C1S4A1. He makes some three and four word TU's — much shorter than he is physically capable of producing and less than a full clause — whose function is demonstrated by showing them to form a coordinate list, a paragraph type indicated by (CE, whose constituents, called items, are indicated as (CEM1, (CEM2, and (CEM3. (Note that the paragraph type and constituent markers precede.)

178

Betty Lou Dubois

Experiments were performed, (CE (CEM1 In an environmental chamber, (CEM2 In total darkness, (CEM3 At ten twenty and thirty degrees centigrade. The coordinate list paragraph is embedded into another paragraph, the details of which I do not present here. Similar instances from this and other speakers are abundant. Another virtue of the Longacre analysis is shown by the following example from a different speech. In order to make it comprehensible, I will put the text into its setting. In the course of testing a hypothesis, the scientist, a marine biologist, made an interesting and quite unexpected discovery, for which he designed a subsequent experiment to verify. Since as a group, the fifty-two biomedical speeches which I have studied have a much more pronounced narrative character than journal articles do (Dubois 1980), it is appropriate to refer to a peak, the most exciting point, the culmination, which is typically marked in several ways. Notionally, the paragraph below is the peak of the narrative in that it offers evidence that the accidental discovery was not an artifact of the original experimental setting. Speaker involvement in the passage is shown in the shift from past to present. The peak marking device which concerns us most here is multiple embedding of paragraphs. The text of the example to be discussed reads as follows: This is consistent with the PH data. But for the animals that were cold, PC02 appears to be relatively flat. It's relatively constant, At about fifteen And then only slowly comes back to normal, As this animal warms up. Bicarbonate, On the other hand = Falls, Appears to fall as the animal warms \ The animal is dumping C02; He is becoming more alkaline and then it comes back Starts to come back to normal as the animal warms up.

Paragraph analysis in formal monologue discourse

179

(AA (AAT This is consistent with the PH data. (AAA1 (AA (AAI But for the animals that were cold, (AAT (TCP (TCPS1 PC02 appears to be,

(AAA Bicarbonate,

(EA (EAT Relatively flat. (EAA1 It's relatively constant, At about fifteen (TCPS2 (TSP (TSPS1 And then only slowly comes back to normal, (TSPS2 As this animal warms up. (EP (EPT On the other hand = (EA EAT Falls, (EAA1 Appears to fall as the animal warms \ (EPP1 The animal is dumping C02; (EPP2 (TCP (TCPS1 He is becoming more alkaline, (TCPS2 (TSP (TSPS1 (EA (EAT And then it comes back (EAA1 Starts to come back to normal, (TSPS2 As the animal warms up.

Figure 2.

Peak marking, with six levels of embedding, using tone units

Figure 2 shows the paragraph analysis of the passage, with indentation to indicate embeddings. The indentations show that there are six paragraph levels here, the deepest embedding of the entire speech. (For a more complete discussion, see Dubois 1984, from which figure 2 is taken.) Moreover, two paragraph types at deep levels (embellishment and paraphrase) show, in the former case, an amplification of detail and, in the latter, a restatement of ideas, two peak marking devices which together convey the speaker's judgment of the importance of the information.

180

Betty Lou Dubois

Because the biomedical speeches are monologue discourses, planned beforehand although almost never fully written out, delivered for the most part in formal scholarese by experienced lecturers, they are more susceptible to parsing into syntactic sentences than, for example, casual conversation. Figure 3 shows our passage analyzed with the sentence as paragraph unit. Two levels of embedding are lost, and although the paraphrase is retained — the paraphrase appears to be the key peak marker here — some of the embellishments, the curlicues which mark the important passage, are concealed.

(AA (AAT This is consistent with the PH data. (AAA1 (AA (AAT (EA (EAT But for the animals that were cold, PCO appears to be, Relatively flat. (EAA1 It's relatively constant, At about fifteen And then only slowly comes back to normal, As the animals warms up. (AAA1 (CE (CEM1 (EP (EPT Bicarbonate, On the other hand = Falls, Appears to fall as the animal warms \ (EPP1 The animal is dumping CO; (EEP2 He is becoming more alkaline, (CEM2 And then it comes back Starts to come back to normal, As the animal warms up. Figure 3.

Peak marking, with four levels of embedding, using syntactic sentences

In the preceding discussion, two instances have been shown in which imposing the Longacre paragraph apparatus onto TU's appears to offer a discoursal explanation for the division into TU's. There are other instances, nevertheless, when the apparatus is too coarse a sieve to catch details, when, in effect, it gives inconsistent results. Take, for example, the case of conjunctive adverbials, specifically, however. When however

Paragraph analysis in formal monologue discourse

181

constitutes the only member of the TU, it can be treated as the setting or introduction to a paragraph. Below, however is the (LR (LRS However, LRT The ventilation, The total ventilation, During the sleeping state, Is about one-third of the resting state, setting [(LRS] for the text [(LRT] of a logical relations [(LR] paragraph, which is itself embedded into an antithetical paragraph signalled by however. In the speech under consideration, in most instances of conjunctive adverbial however, the word is not alone in the TU and hence cannot be treated as setting. Three instances will illustrate. However recent modifications in our technique, However resting frequency, However tidal volume does. The paragraph analysis forces two different treatments of functionally identical uses of the word. A further problem is that sometimes a paragraph constituent stretches across several TU's, an unsatisfying situation but less troublesome than the treatment of adverbials. Such noncongruence of analytic units is a common linguistic phenomenon, as when a syntactic sentence is out of phase with a phonological sentence, or, as in the speeches, where sometimes the syntactic unit fails to coincide with the TU. Of the following illustrations of such noncoincidence, the first two are from C1S4A5, the third from C1S4A1. Compared with, Perhaps a second, For inspiration. As you probably know the, The unidirectional flow model of, Of the avian lung is well accepted And it is this topic, That I would like to, Focus the remainder of, Of this talk.

182

Betty Lou Dubois

It is perhaps a superficial concern, but there is something intuitively unsatisfactory about taking semantic sentence relationships from traditional rhetoric and applying them to TU's which consist of only words or phrases. More important is the problem of how — and indeed whether — multiple embeddings are perceived by listeners. I myself had a great deal of difficulty coding embeddings at the depth of, say, four or five, so much that I finally resorted to a complicated system of color coding. Although I am certainly willing to concede that biomedical scientists, possessing schemata by which to decode the incoming information encoded in the slide talks, might be able to process deeper embedding than I, still, my difficulty in keeping the levels straight in the calm and unlimited time I had makes me question whether such embedding is perceptible in the same ways to listener as to linguist. For my part, I prefer to deal with perception motivated concepts (Palkova and Palek 1978) rather than analytic units which have no psychological reality for language users; pending some demonstration that very deep paragraph embeddings are processed by the latter, I prefer to err on the side of caution. I have thus reluctantly abandoned the TU and adopted the syntactic main clause as the analytic unit for paragraphs, with the fragment as default unit. On the positive side, the decision will enable me ultimately to compare paragraph types and sequences of spoken biomedical English with those of biomedical journal articles. In conclusion, there is here the classic linguistic problem: living language in living use always escapes the analytic tools, forcing linguists to choose the alternative which most nearly suits their purposes, recognizing all the while its limitations. Note 1. The paper was presented to the 1985 Annual Meeting of the Linguistic Association of the Southwest and was financed in part by the National Institute of General Medical Sciences and minigrants from the Arts and Sciences Research Center and a sabbatical leave, New Mexico State University. I am deeply honored to have the opportunity to pay public tribute to Stanley Newman and to acknowledge a continuing personal debt to him. Although I had had many courses in Romance philology and linguistics, it was Stanley who introduced me to Linguistics, in a 1961 course in phonetics and phonemics. During that semester, I felt increasingly that I had come home intellectually, a sentiment I was finally able to act on in 1970, when I enrolled as a doctoral student at U. N. M. The following spring, along with others of my friends, I was privileged to be a student in Stanley's lastbefore-retirement course in historical and comparative, among the most difficult but

Paragraph analysis in formal monologue discourse

183

best courses I ever took. In addition to the expected content, Stanley gave us the complete Uto-Aztecan bibliography and, at our urging, personal accounts of that first luminous generation of American linguists: Sapir, Bloomfield, Whorf. An ever great privilege was having Stanley on my dissertation committee, where, with his quizzical and exquisitely gentle way of expressing disapproval, he taught me a lesson which grows in value each year. In an era in which linguists were defiantly manufacturing their own examples for analysis, I felt decidedly empirical in taking mine from a published account of a trial. After I explained the procedure, Stanley cocked his head, raised his eyebrows, stepped back slightly, and wondered if I couldn't find a better data sample. I took the hint. Although it necessitated weeks more work, using authentic linguistic data not only improved my dissertation, but showed me how to do linguistic research. While controversies swirled at his feet and while theories ballooned up and were deflated, Stanley kept at his work: analyzing the real linguistic output of real people. Having never succumbed to mere novelty, he never had to apologize for right-angle shifts in theory. I count it my great good fortune to have learned methodology from Stanley Newman. And while I could never contemplate approaching his genius, I am grateful, like many others, to have had his example.

References Aarts, Jurg (ed.) 1984 Conference on the use of computers in English language research. ICAME News. No. 8, Pt. 1:9-29. Chafe, Wallace 1980 The deployment of consciousness in the production of a narrative. In Wallace Chafe, ed. The Pear Stories. Cognitive, Cultural and Linguistic Aspects of Narrative Production. Norwood: Ablex: 9—51. Dubois, Betty Lou 1982 Genre and structure of biomedical speeches. Forum Linguisticum. 5(2): 140— 169. 1982 The function of intonation contours in biomedical speeches. In Shiro Hattori and Kazuko Inoue, eds. Proceedings of the XHIth International Congress of Linguists. Tokyo: Proceedings Publishing Committee, under the Auspices of CIPL. 1004 - 8. 1984 Paragraph structure of biomedical speeches: Preliminary report. LACUS Forum X. Columbia: Hornbeam. Halliday, Michael A. K.—Ruqaiya Hasan 1975 Cohesion in English. London: Longman. Lackstrom, John—Larry Selinker—Louis Trimble 1973 Technical rhetorical principles and grammatical choice. TESOL Quarterly. 7(2): 127-37. Longacre, Robert E. 1980 An apparatus for identification of paragraph types. Notes on Linguistics. 15:5-22.

184

Betty Lou Dubois

Moffett, James 1983 Teaching the Universe of Discourse. Boston: Houghton-Mifflin. Palkova, Zdena—Bohumil Palek 1978 Functional sentence perspective and textlinguistics. In Wolfgang U. Dressier, ed. Current Trends in Textlinguistics. Berlin: Walter de Gruyter. 212—28. Svartvik, Jan et al. 1982 Survey of Spoken English. Report on Research. Lund Studies in English 63. Lund: Gleerup.

Object agreement in the Halkomelem Salish Passive: a morphological explanation Donna B. Gerdts

Abstract Davis (1980) and Gerdts (1981) point out that in two Coast Salish languages the Passive construction involves a nominal which tests syntactically to be the final subject but, in the case of pronominals, surfaces as an objective suffix. Here I provide an explanation for this phenomenon which correctly predicts two types of Coast Salish languages — those with objective passive subjects (e. g. Sliammon, Halkomelem, Sechelt) and those with subjective passive subjects (e.g. Straits, Squamish, and Lushootseed). Crucially, the Coast Salish passive is formed from a transitive stem; transitivity is overtly marked in Coast Salish. In the former group of languages — but not the latter — the transitive suffixes have fused with the objective suffixes making it impossible to mark transitivity without also marking the person and number of the object. The morphological requirements of stem formation take priority over the syntax thus interfering with the syntactic conditioning of case.

1. The Halkomelem "funny" passive1 In Halkomelem, as in other Central Coast Salish languages, passive morphology involves an overtly marked transitive verb form followed by an intransitive suffix, as in (1 b), the Passive corresponding to (1 a):2-3

(la) b)

ni

q'wal-dt-3s

0a sienP fid sce.itsn

aux bake -tr-3erg det woman det salmon 'The woman baked the salmon.'

ni

q'wal-dt-dm

0a sienP fid sce.itan

aux bake -tr-intr obi det woman det salmon 'The salmon was baked by the woman.' Other than this rarity, the Halkomelem Passive seems straightforward: the passive agent is presented in the oblique case and the clause is finally intransitive, as seen by the lack of 3rd person ergative agreement in (1 b).

186

Donna B. Gerdts

However, in Halkomelem (and two other Coast Salish languages [cf. section 2.2]), a 1st or 2nd person final subject in the Passive is represented by an object suffix, as in (2) and (3).4 (2)

ni

hm -aOehm

">a

ia

siertP

aux look-tr+lobj+intr obi det woman 'I was looked at by the woman.' (3)

ni

hm -dOa.m

ia

sieni^

aux look-tr+2obj + intr obi det woman 'You were looked at by the woman.' This last property is strange since a final subject is expected to be represented by a subject — not object — marking; subjects are represented as clitics and objects as suffixes, as seen in (4). (4)

ni

can

hm

-adama

aux lsub look-tr+2obj 'I looked at you.' However, the passive subject cannot be represented by subject marking, as exemplified in (5) and (6). (5)

*ni

can

hm

-at-am

aux lsub look-tr-intr ('I was looked at.') (6)

*ni

c

hm

-at-am

aux 2sub look-tr-intr ('You were looked at.') The purpose of this paper is to provide an explanation for object marking in the Passive. First, some alternative analyses involving a revised syntactic analysis or agreement rule are rejected. Then a proposal is made: object marking is present because it is fused to the transitive marker which is required in the Halkomelem Passive. 1.1 Syntactic approaches Two attempts (Hukari 1980, Gerdts 1981) have been made to reconcile a syntactic analysis of Halkomelem Passives to the morphology. Hukari (1980) posits Spontaneous Demotion; that is, Halkomelem Passives involve demoting the subject without promoting the object. Gerdts (1981) suggests Impersonal Passive following Perlmutter (1978); an inserted (invisible) dummy, which places the passive nominal en chomage, is the

Object agreement in the Halkomelem Salish Passive

187

final subject. Under these analyses the object suffix in (2) and (3) would be representing a final object or object-chomeur. However, there are two ways in which the passive nominal behaves like a subject. First, although subject marking in a main clause Passive is not acceptable (see *(5) — (6)), if the Passive is embedded as a nominalization, the nominal may also be represented as a possessor, the case used for subjects in nominalizations. (7)

sk^ey k^a na -s -c'ew -a9e.lt impossible det lpos -nom -help - t r + l o b j + st 'It's impossible for me to get helped.'

Second, as Gerdts (1981) discusses, the passive nominal can raise to object, as exemplified in (8 b); Raising to Object is a property of subjects and not objects in Halkomelem.5 (8 a)

b)

n can xaxci-t [?u n-fas leVam? aux lsub wonder-tr Ink aux-3ssub look -aOa.m">] -tr+2obj + intr 'I'm checking out to see if you are being watched.' 9 can xexci -Oama i -9as le^lam9 aux lsub wonder -tr+2obj Ink aux -3ssub look -aOa.m?] -tr+2obj+intr 'I'm checking you out to see if you are being watched.

Since possessive marking in nominalizations and raising-to-object are subject properties, analyses involving Spontaneous Demotion or Impersonal Passive are less than satisfactory.6'7 Therefore, I take Halkomelem Passives to be Personal — that is, the passive nominal advances from object to subject — and turn to the problem of explaining why a syntactic subject should be represented by object morphology. 1.2 Case/agreement approaches Other cases of passive subjects appearing as objects have been cited, albeit rarely. A quick look at two other examples — Icelandic and Kashmiri — shows that the solutions posited for non-nominative subjects in these languages are inadequate as an explanation of the Halkomelem "funny" Passive.

188

Donna B. Gerdts

Icelandic Passives, as Zaenen et al. (1985) discuss, involve "quirky" case: objects whose case is lexically determined by the verb (see (9 a)) retain this case in the Passive (see (9 b)) even though they are final subjects. (9 a) b)

Eg hjalpadi 'I helped Honum 'Him (DAT)

honum. him (DAT).' var hjalpad. was helped.'

Their explanation is that "quirky" case is assigned lexically and nominals assigned lexical case do not undergo further case marking via the normal rule assigning NOM to final subjects, ACC to final object, etc. It would be difficult, however, to posit a "quirky" case analysis for Halkomelem, since all 1st and 2nd person subjects of Passives, not just those of a lexically determined class of verbs, appear as object suffixes. In Kashmiri (Altaha 1985), the subject of a Passive, although it determines subject agreement and in other ways tests to be syntactic subject, appears in the Accusative case [in the present tense], as seen in (10).8 (10)

mardas cu yiwa.n hicnawne mastarni sindi zeryi man-ACC aux coming teach-pass teacher-obl of by 3. m. sg. 'The man is taught by the teacher.'

Altaha posits a metastratal case rule (that is, a rule which does not refer to a specific syntactic level) where ACC case takes precedence over NOM case. Although the passive nominal, being both a subject and object (at different levels) would qualify for either case, it is first marked ACC making further case marking unnecessary. The case rule applies generally; final subjects of Unaccusative and Inversion constructions also take ACC case, for example, the Unaccusative in (ll). 9 (11)

hrkss cu log boy-ACC aux hurt 3. m. sg. 'The boy is hurt.'

However, formulating a metastratal agreement rule along these lines would be difficult for Halkomelem. The subjects of Unaccusatives (Gerdts 1981) are represented by subject clitics — not object suffixes — as in (12):

Object agreement in the Halkomelem Salish Passive

(12) ni

can

189

r hihkw < wac'a-X 1 c'iwal9

( happy aux lsub < fall v. annoyed

{

was happy.' fell.'

was annoyed.' Although the above solutions are inappropriate for Halkomelem, the general approach is useful. What is required is a reason for object marking to occur before subject marking in Passives but not Unaccusatives. 1.3 A morphological explanation My explanation hinges on the morphological fact that Coast Salish languages form Passives — but not Unaccusatives — from a transitive base. I claim that in Halkomelem the transitive marker is "fused" to the object marker, what I refer to as T-obj fusion, making it impossible to meet the morphological condition of having a transitive base without also including the object suffix. Sections 2 and 3 argue for this viewpoint. Evidence comes from comparative data and from various syntactic constructions from within Halkomelem. As section 4 discusses, the object agreement in Passives takes priority over subject agreement. Once the subject of the Passive is represented by the object suffix, subject agreement is unnecessary.

2. T-obj Fusion This section gives phonological evidence for T-obj Fusion and discusses the implications of Fusion for the distribution of "funny" Passives in Coast Salish. 2.1 The phonology of the T-obj Evidence for Fusion comes from the unexpected phonology of the T-obj in Halkomelem. The transitive suffix is clearly -t (often preceded by an epenthetic schwa), as is obvious in forms with 3rd person objects.

190

Donna B. Gerdts

Comparative evidence (cf. Newman 1979) allows us to posit the base forms for the Halkomelem object suffixes as in column (a) which combine with the transitive suffix to give the forms in (b). (13)

(a) objects

(b) T-objs

1 sg.

-samPs

-Qam?s

2 sg.

-sama

-Oama

3 sg./pl.

0

-t w

1 pi.

-aPx

-taPxw

2 pi.

-ah

-tab

In 1st and 2nd sg. forms, the -0 which arises from the sequence of -t and -s is unexpected, since this sequence is allowed elsewhere in Halkomelem, for example, in (14). (14)

kwd3

s

-amss-t

-s

det nom -give-tr -3pos 'what he gave him' I take the fact that the transitive and object markers "share" a consonant as evidence that the forms are closely associated in the morphology. 10 Once this merger takes place, the form functions as a unit in the morphology; the elements are not available independently but must appear in tandem. 2.2 Fusion and "funny" Passives in Coast Salish Some support for this view comes from a comparison of Halkomelem to other Central Coast Salish languages. The prediction is clear. Since these languages all form a Passive from a transitive base, if a language has a T-obj, as evidenced by phonological fusion, then, it should have a "funny" Passive.11 Although the available data is sparse, it appears that three Coast Salish languages have "funny Passives" — Halkomelem, Sliammon (Davis 1978, 1980) and Sechelt (Beaumont 1985). Two of these — Halkomelem and Sliammon — clearly have phonological fusion. Davis uses S to represent an s which fuses with the transitive suffix -t in his paradigm for object and Passive suffixes. (15) 1st person 2nd person

object suffixes -S -Si

passives suffixes -Say-Si-

Object agreement in the Halkomelem Salish Passive

191

The Sliammon data in (16) [from Davis (1978)] illustrates fusion and "funny" Passive. (16)

active 'to give me' 'to know you'

passive 'someone gives it to me' 'someone sees you'

xanaO t'sgaOi

xanaQayam k^dOim

In contrast, Lushootseed and Squamish do not have "funny" Passives; rather the passive subject is represented as a subject clitic, as illustrated in the Lushootseed data in (17) (from Hess 1973); the nominal in the passives is marked like the subject in (18) and not like the object in (19): ?m c'axwat-b

(17)

caxw

ti

cacas

'You were helped by the boy.' kwaxwat-s

(18)

caxw

'You helped me.' 9

(19)

u

c'axwa-t-sid

ti

c'ac'as

'The boy clubbed you.' In these languages a fused form does not occur. The regular phonological reflex of a sequence of /t/ and /s/ in these languages is /c/. The Straits languages (e.g. Saanich and Lummi) also have subject clitics in the Passive. In these languages, the sequence of -t and -s does not occur, rather -t is deleted. Information concerning Passives in Pentlatch, Nooksack, and Twana was not available to me. Table 1 summarizes the survey of Coast Salish languages below; the table shows subgroupings as posited by M. Dale Kinkade and Laurence Table 1.

Fusion and Passives in Coast Salish Fusion

Funny Passive

Comox/Sliammon Sechelt Pentlatch

yes no

yes yes

Squamish Halkomelem Straits (Saanich) (Lummi) Nooksack

no yes

no yes

no no

no no

Lushootseed

no

no

Twana

192

Donna B. Gerdts

C. Thompson (p.c.) and basically gives the languages from north to south.12 Thus, the comparative data support the T-obj analysis. Sliammon — the only other Coast Salish language with phonological fusion — also has "funny" Passives.

3. The syntax of the T-obj That the transitive and object suffixes form an inseparable unit is observed in several syntactic situations besides the Passive. Both or neither of the T-obj elements are present in all constructions even if the presence of the object suffix is semantically redundant or the transitive marker would be otherwise expected. Two cases — Extraction and Object Cancellation constructions — are given here. 3.1 Extraction Halkomelem allows the direct extraction of subjects and objects, as seen in the cleft constructions in (20) —(23):13 (20) (21) (22) (23)

sienP da ni qwdl-3t t°a see.ton woman det aux bake-tr det salmon 'The woman is the one who baked the salmon.' nawa ni q'waqw -at (*-axw) 2emph aux club -tr (*2ssub) 'It's you who clubbed it.' sce.itan fa ni q'wal-at-as ia sienP salmon det aux bake-tr-3erg det woman 'The salmon is what the woman baked.' nawa ni lam -Oama -?e.n? 2emph aux look-tr+2obj -lssub 'It's you that I looked at (you).'

When subjects are extracted, as in (20) and (21), subject agreement is not present in the embedded clause. The subject agreement remains when the object is extracted, as in (22) and (23).14'15 In contrast, when objects are extracted, a copy is left in the embedded clause, as in (23). What is the motivation for the object copy? It is semantically redundant since potential ambiguity is resolved by the presence (or absence) of subject agreement. T-obj fusion provides an explanation. The embedded

Object agreement in the Halkomelem Salish Passive

193

clause is finally transitive and so transitive marking is required. It is impossible to mark transitivity without also giving the person marker due to T-obj fusion, hence *(24).16 (24)

*K3ws

ni

q'waqw

-tr-lssub 2emph aux club ('It's you that I clubbed.') An alternative account which would stipulate that extracted objects leave copies while extracted subjects do not runs afoul of the Passive data. When the subject of a Passive is extracted, as in (25), it leaves a copy.17 (25)

nawa

ni

lam

-aOamat

2emph aux look-tr+2obj + st 'It's you who was looked at (you).' The parallelism between the extraction of objects and passive subjects is captured under this account since both involve T-obj fusion. 3.2 Object cancellation T-obj fusion also provides a solution to another mystery in Halkomelem. In periphrastic causatives with the predicate cset 'tell someone to do something', if the object of the embedded clause refers to the subject of cset as in (26 a), then it may be deleted, as seen in (26 b); the general intransitive marker -am replaces the object suffix. (26 a)

cse -t

can

ce9

is

sieni9

9

u

q'wal

-die

-6am9s

tell-tr lsub fut det woman Ink bake -adv - t r + l o b j -as

?a

kw6a

sce.itan

-3ssub obi det salmon 'I'm telling the woman to bake the salmon for me.' b)

cse -t

can

ce9

ia

sienP

9

u

q'wal

-aic

-am

-as

tell-tr lsub fut det woman Ink bake -adv -intr -3ssub kw9a

sce.itan

obi det salmon 'I'm telling the woman to bake the salmon for me.' The transitive marker is also omitted in the complement clause, even though the clauses are clearly based on transitive forms. This follows from T-obj fusion: if the object suffix does not appear neither will be transitive suffix.18

194

Donna B. Gerdts

A curious feature of object cancellation is that it appears not to effect 3rd persons. While 1st and 2nd persons make good deletion targets, as in the above data, 3rd person does not, as is seen by comparing (27 a) and (27 b). cse -Qam^s -as q'wal -aic -t ">a kw0a tell -tr + lobj -3erg Ink -bake -adv -tr -lssub obi det sce.itan salmon 'He's telling me to bake the salmon for him.' 9 b)?? cse -Oam9s -as u q'wal -aic-am -9e.n9 9a kwQa tell -tr + lobj-3erg Ink bake -adv-intr -lssub obi det sce.itan salmon 'He's telling me to bake the salmon for him.'

(27 a)

This would be mysterious under a view of object cancellation as a syntactic phenomenon. However, if what is involved is an ellipsis of the redundant person marking, it makes sense that 3rd person objects would be exempted, as they are 0 anyway.

4. The Halkomelem agreement rule The previous sections presented arguments for the unitary structure of the T-obj: the transitive and object markers are phonologically fused and they function in tandem in various syntactic situations, one of which is the Passive. T-obj fusion provides an explanation for object agreement in the Halkomelem Passive: since Passives must be marked transitive, and since the transitive suffix is fused with an object suffix, object marking is obligatory. What remains to be accounted for is the presence or absence of subject marking in the Passive. As discussed in section 1, simple clause Passives have only object marking; subject marking is not possible, as (28) shows: (28)

ni (*can) lam -aOelam aux lsub look - t r + l o b j + intr 'I was looked at.'

However, when the Passive is embedded as a nominalization, the subject may be doubly represented by possessive and object marking.

Object agreement

(29)

in the Halkomelem

Salisti Passive

195

sk^ey k^a [na -s f-c'ew -ade.lt]] impossible det lpos -nom -help - t r + l o b j + st 'It's impossible for me to get helped.'

The first case, where doubling is not possible, is consistent with a principle for agreement put forth by Davies (1986:168). (30)

Given a set of agreement rules making a predicate agree with the same set of properties a, b, ... n of nominals, the rules apply disjunctively to any given nominal.

The Halkomelem subject clitics, the object suffixes, and the possessive prefixes reference the same semantic features (person and number), so (30) requires them to be mutually exclusive in a local environment (in the domain of the same predicate), hence the ungrammaticality of (28). However, in a non-local environment, (30) is not relevant and doubling is possible. For example, in (29) above, the agreement locality of the object suffix is the passive predicate "be helped" and that of the possessive prefix is the entire nominalized clause. The Personal Passive analysis together with the agreement principle in (30) make the right range of predictions. The passive subject will agree like a subject only outside of the domain of object agreement.

5. Conclusion The approach of this paper has been to seek a morphological rather than a syntactic explanation for the presence of object marking in the Halkomelem Passive. The benefit of such an approach is that it allows a maximally simple grammar. Passives in Halkomelem may be analyzed as run-of-the-mill Personal Passives. The final subject, after all, behaves like a subject in all respects except agreement. The approach taken here also allows a straightforward view of agreement. Basically, objects cue agreement whenever there is transitive marking, as T-obj fusion necessitates. Final subjects cue subject marking as constrained by the putatively universal agreement principle (30), that is, except in the domain of object agreement. Finally, an explanation should be given for why transitive marking and hence object marking takes priority over subject marking. If morphological ordering reflects the levels of syntactic analysis (as stipulated by

196

Donna B. Gerdts

the Satellite Principle [Gerdts 1981]), then transitive marking, which references an early level of the Passive, has precedence over subject marking, which would reference the final level.

Notes 1. The Halkomelem data herein are from the late Arnold Guerin of the Musqueam Reserve, Vancouver, British Columbia. My research has been supported by the Melville and Elizabeth Jacobs Research Fund, the Phillips Fund, and the National Museum of Man, Ottawa. 2. The following abbreviations are used in the glosses of the Halkomelem data: adv advancement suffix aux auxiliary det determiner emph emphatic person forms erg ergative fut future intr intransitive Ink linker obj objective obi oblique marker pi plural sg singular sub subjective ssub subordinate clause subjective st stative tr transitive 1 first person 2 second person 3 third person 3. See Gerdts 1981 for a general description of Halkomelem. Halkomelem is a verb initial language (predominantly VSO). Subjects, objects, and common noun possessors are unmarked for case. Other nominals — obliques, chomeurs, possessors — are flagged by the all purpose preposition V Pronominal subject agreement is presented by clitics which appear in 2nd position in main clauses. Pronominal object agreement is suffixed to the verb. Halkomelem is a "pro-drop" language: independent pronouns are used only for emphasis. Halkomelem is a split ergative language; ergative agreement is marked only for 3rd person ergatives in main clauses (in all tense/aspects). 4. Throughout this paper, I have limited the data to examples involving the transitive suffix -t, the marker for control [Newman's neutral paradigm]. Discussion of data with limited control forms and causatives forms has been excluded due to space limitations. There is not a perfect match between the object suffixes (which are given in (13) and the passive paradigm, which I give in (i) below. Furthermore, main clause passives, which have the intransitive suffix -am are different from subordinate passives, which have the stative suffix -at.

Object agreement in the Halkomelem Salish Passive

(i) lsg 2sg 3 lpl, 2pl

5.

6. 7.

8. 9.

10. 11. 12.

13.

14.

15.

16.

Passive Suffixes -Oehm -Oa.m -tsm -tahm

197

Subordinate Passive Suffixes -6e.lt -Oanwt -tewat -taht

The major difference is with the lsg, where the object suffix in the active is -9am9s. As Newman (1979) points out, there are several Salish languages with two forms for the neutral object so apparently two forms can be reconstructed for Proto-Salish; the Halkomelem Passive form is clearly a reflex of one of these. The Halkomelem active object is a reflex of a Causative object form in Proto-Salish. The passive paradigm shows a neutralization of 1st and 2nd plural object forms. A common phonological change of a sequence of vowel + resonant + vowel + consonant changes to long vowel + resonant + consonant can be seen in the forms for 2sg, and lsg subordinate. This argument was first given for Sliammon by Davis (1980). Hukari (1980) discovered the phenomenon in Halkomelem with the verb cset. Gerdts (1981) points out that passive agents also raise. Hukari gives an argument based upon extraction in support of his analysis. An alternative account of this phenomenon is presented in section 3.1 below. See note 17. Gerdts (1981) suggests a treatment of the doubling phenomenon within an Impersonal Passive analysis in terms of brother-in-law agreement (see Perlmutter 1978). However, such a device would be inappropiate for Raising phenomena. The case rule used by Altaha parallels Davies' treatment of Choctaw, a language which does not appear to have a Passive. According to the Unaccusative Hypothesis (Perlmutter 1978), the nominal in Unaccusative Advancement constructions is an initial object which advances to subject. Since the ACC case rule takes priority over the NOM case rule in Kashmiri, and since agreement is blind to syntactic level, the nominal in the Unaccusative construction, since it is an object, qualifies for ACC case. An analogous situation would be the merger of subject and object prefixes into a synchronically unanalyzable unit as in Iroquoian or Karok. Of course, it would be possible for a language to have T-obj fusion without showing the phonological effects. Sechelt may be such a language. Halkomelem is neither geographically adjacent nor genetically subgrouped with the other Coast languages with "funny" Passive making an areal or historical source for this phenomenon unlikely. Also, irrelevantly, certain possessors extract directly. Other nominals, if they can extract, must extract via nominalization. See Gerdts (1981) and references therein for discussion. Since Halkomelem is a VSO language which does not differentiate subject and object via nominal case, a potentially ambiguous situation arises in extraction. This is resolved by the deletion of subject agreement in the embedded clause in the case of subject extraction, as seen by the absence of -as in (20). This dichotomy between subjects and objects complicates the agreement rules. Subject agreement is with a "surface" nominal while object agreement is with a "final" object. See Gerdts 1981 for discussion. If the explanation I have given here is correct, it suggests that Coast Salish languages without T-obj fusion, e. g. Squamish or Lushootseed, will not require a copy when a

198

Donna B. Gerdts

pronominal object is extracted. The information necessary to check this prediction was not available to me. 17. This fact that Passive subjects pattern with objects and not with other subjects with respect to copying in extraction led Hukari to argue that they actually are objects. The T-obj fusion account given here provides a reply within a Personal Passives analysis to that argument. 18. In fact, a sentence like (26) but with -t rather than -6am?s on the embedded verb is possible, but it would have the meaning "I'm telling the woman to bake the salmon for him/her."

References Altaha, Fayez 1985 Non-Nominative subjects in Kashmiri. Dissertation, State University of New York at Buffalo. Beaumont, Ronald C. 1985 She Shashishalhem: the Sechelt language. Penticton, British Columbia: Theytus Books. Cinchor, Nancy 1975 A treatment of passives and pronouns in Lummi matrix sentences. In Papers from the 10th International Conference on Salishan Languages, Ellensburg, Wa. Davies, William 1986 Choctaw verb agreement and universal grammar. Dordrecht: Reidel. Davis, John 1978 Pronominal paradigms in Sliammon. In Papers from the 13th International Conference on Salishan Languages, Victoria, B. C. 1980 The passive in Sliammon. Proceedings of the Sixth Annual Meeting of the Berkeley Linguistics Society, ed. by Bruce R. Caron et al. 278—286. Gerdts, Donna B. 1981 Object and absolutive in Halkomelem Salish. Dissertation, University of California, San Diego. Hess, Thom 1973 Agent in a Coast Salish language. IJAL 39.2, 89-94. Hukari, Thomas 1976a Person in a Coast Salish language. IJAL 42.4, 305-318. 1976 b Transitive in Halkomelem. In Papers from 11th International Conference on Salishan Languages, Seattle, Wa. 1980 Subjects and objects in Cowhichan. Paper presented at the 15th International Conference on Salishan Languages, Vancouver, B. C. Kuipers, Aert 1967 The Squamish language. The Hague: Mouton. Montler, Timothy 1986 An outline of the morphology and phonology of Saanich, North Straits Salish. Occassional Papers in Linguistics No. 4. University of Montana.

Object agreement in the Halkomelem Salish Passive

199

Newman, Stanley 1979 The Salish object forms. IJAL 45:4, 299 - 308. Perlmutter, David M. 1978 Impersonal passives and the unaccusative hypothesis. Proceedings of the Fourth Annual Meeting of the Berkeley Linguistics Society, ed. by Ten T. Tager et al. 157-89. Zaenen, A.—J. Maling—H. Thrainsson 1985 Case and grammatical functions: The Icelandic passive. Natural Language and Linguistic Theory 3:4, 441 —484.

Argument obviation and switch-reference in Hopi* LaVerne Masayesva Jeanne and Ken Hale Abstract The Hopi system of switch reference forms a part of a cross-categorial theory of obviation which governs relations of coreference between the subject of a sentence and the principal argument of a dependent category. Following work by Jeanne and Finer, the Hopi system is integrated into the general binding theory of the Government and Binding framework.

In this paper we examine the suggestion made in Jeanne (1978) to the effect that the phenomenon of subject obviation, more popularly termed "switch-reference" (Jacobsen 1967), belongs properly to the general system of grammatical principles which govern coreference relations among the arguments of a sentence, i. e., to what is now known as the Binding Theory within the framework elaborated by Chomsky in his Lectures on Government and Binding (1981). We reexamine Jeanne's original treatment of Hopi obviation in the light of recent formulations of the Binding Theory and in the light of recent work by Finer (1984, 1985) which also seeks to account for switch-reference within that theory. The generally accepted version of the Binding Theory is that presented in (1) below: (1)

The Binding Theory (Chomsky 1981:188): (A) An anaphor is bound in its governing category. (B) A pronominal is free in its governing category. (C) An R-expression is free.

The following Hopi sentences, and their English counterparts, illustrate the functioning of each of these principles: (2a)

Mi' tiyo'ya [ec] naa-tuhota that boy [ec] REFL/RECIP-hurt 'The boy hurt himself.'

202

LaVeme Masayesva Jeanne and Ken Hale

b) c)

Mi' tiyo'ya pu-t tuuhota that boy him-ACC hurt 'The boy hurt him.' Pam mi-t tiyo'ya-t tuuhota he that-ACC boy-ACC hurt 'He hurt the boy.'

In (2 a), the object of the verb tuuhota (hurt), represented in syntax as an "empty category" (symbolized here as [ec]), is identified as an anaphor (by virtue of the reflexive/reciprocal marking on the verb). Being an anaphor, this object must be bound in its governing category, in accordance with condition (A) of the Binding Theory. The notion "governing category", in the sense relevant here, can be identified with the concept "Complete Functional Complex" of Chomsky (1986), defined very roughly, and minimally, as follows: (3)

Complete Functional Complex (CFC): A is a CFC for B if it is the minimal category containing (i) B itself, (ii) a governor of B, and (iii) a subject.

Assuming, as is usually done, that the verb governs its object and that the subject is a constituent of S, not VP, then it is the category S which constitutes the governing category of the object in (2 a), in the sense relevant to the Binding Theory. The latter is satisfied in (2 a), if we can assume that the object there is bound to the subject, as we customarily do assume to be the case for the reflexive relation. We will adopt the usual definition and notation for the relation bound: (4)

Bound: A binds B if A c-commands B and B is "referentially linked" to (co-indexed with) A.

If an argument is not bound, in the sense of (4), then it is free, as required by conditions (B) and (C) of the Binding Theory. Sentence (2 b) conforms to condition (B), since the object NP, the pronominal pu-t (him), is free in its governing category (S). And sentence (2 c) conforms to condition (C), since the object, the R-expression object — i.e., the overt NP expression mi-t tiyo'ya-t (the boy, accusative) — is free. The structural configuration shared by the sentences of (2) is essentially that deplicted in (5) below:

Argument obviation and switch-reference

in Hopi

203

£A

& / 7

/

i \ VP A \\

/ ' \\

NP / NP

\

\\ INFL

V

Following Jeanne (1978), we assume that INFL (formerly AUX) is the head of the category S in Hopi (as is now assumed, within the GB framework, to be the case for languages generally) and that the verb heads an autonomous projection VP. That is to say, S and VP are projections of the categories INFL and V, respectively. The subject argument and the VP which is predicated of it are introduced as complements of INFL. And since the object argument is introduced as a complement of V, it follows that the subject asymmetrically c-commands the object, as can be seen in (5), where NP is the subject, and NP the object. Given these structural relations, according to the Binding Theory, if NP is an anaphor, it must be bound by NP in (5). By contrast, if NP is not an anaphor, i. e., if it is a pronominal or an Rexpression, it must be free in (5) and, therefore, not bound to NP. Thus, given the grammaticality judgments and coreference interpretations attached to the Hopi sentences of (2), it follows that the object of the verb must be an anaphor in (2 a) and a non-anaphor in (2 b — c). The parallel which exists between binding relations of the type represented in (2) and the coreference relations involved in subject obviation, or switch-reference, is illustrated by the sentences of (6) below: (6 a)

'I-pava paki-t puu' pam qatuptu my-OBro enter-PROX then he sat 'When my brother entered, he (my brother) sat down.' 'My brother came in and sat down.' b) 'I-pava paki-q puu'pam qatuptu my-OBro enter-OBV then he sat 'When my brother came in, he/she (not my brother) sat down.' c) Pam paki-q puu' 'i-pava qatuptu he enter-OBV then my-OBro sat 'When he/she (not my brother) came in, my brother sat down.' We assume here, as is usually done, that the first clause in these Hopi sentences is subordinate to the second. The relevant coreference relations are those holding between the subordinate-clause subject and the matrix-

204

LaVerne Masayesva Jeanne and Ken Hale

clause subject. In (6 a), the first is necessarily interpreted as coreferential with the second, while in (6 b—c), the two subjects are necessarily disjoint. This situation is, of course, closely similar in nature to the binding relations which hold in the sentences of (2), and it is reasonable to expect that the Binding Theory is at work here as well. It is evident, however, that the Binding Theory cannot carry over directly and straightforwardly to the coreference relations involved in socalled switch reference. First, while the NP expression 'i-pava (my brother) is necessarily coreferential with the subject of the matrix clause in (6 a), we cannot say that it is bound, in the sense of the Binding Theory. It is inherently an R-expression and not an anaphor. Therefore, if it were bound, we would have a straightforward condition C violation — condition C is otherwise inviolate in Hopi, as can be seen by the ungrammaticall y which results if, for example, an R-expression is substituted for the empty category object in (2 a). Second, it is most unlikely that a ccommand relation (in either direction) holds between the arguments involved in subject obviation of the type represented in (6) — hence, the binding relation is technically impossible there. And, as just noted, ccommand between the subjects would lead immediately to condition C violations in cases of subject coreference. It seems most reasonable to assume that the dependent clauses in the sentences of (6) are adjuncts, rather than daughters, to S (cf. Jeanne 1978, for a detailed discussion of Hopi x-theory). Simplifying somewhat, without loss of essential detail, the structure shared by the sentences of (6) is as depicted in (7) below: S'

(7)

•

VP

CP • /

•/

\ S

/

C(OMP)

S /N / i VP NP

NP

INFL

INFL

Argument obviation and switch-reference in Hopi

205

Under "strong" c-command (according to which A c-commands B if the first branching node dominating A also dominates B; cf. Reinhart 1976), NP does not c-command NP. The latter cannot, therefore, be bound to the former. This is fortunate, since otherwise the perfectly grammatical Hopi sentence (6 a) would be in violation of condition C, in defiance of the facts regarding that condition elsewhere in Hopi and in defiance of what has been found to hold quite generally in languages of the world. We are left, however, with an as yet unexplained similarity between subject obviation (i. e., switch-reference) and the canonical situation, illustrated by (2), in which the Binding Theory applies straightforwardly. To put the matter in question form: How do we account for the coreference facts of (6) and the like? Is the Binding Theory involved in Hopi subject obviation at all? And, if so, precisely how is it involved? In attempting to answer these questions, we will develop a conception of subject obviation which combines aspects of the analysis suggested for Hopi in Jeanne (1978) and aspects of the theory outlined for switchreference generally in Finer (1984, 1985). Let us first examine other Hopi constructions in which the Binding Theory appears to apply in the expected way. The sentences of (8) below each contain a postpositional phrase complement of the main verb yu'a'ata (speak). In (8 a), the object of the postposition is bound by the subject of the main verb. In (8 b), on the other hand, the object of the postposition is free: (8 a) b)

Taqa Taqa 'Taqa Taqa Taqa 'Taqa

fee J naa-mi yu'a'ata [ec] REFL/RECIP-to speak is speaking to himself.' pu-t 'a-w yu'a'ata him-ACC 3SG-to speak is speaking to him/her.'

The structure shared by these two sentences is essentially as follows:

i NP

/ /

/

/ \ / \ \ \

PP /

/

/

/

NP

/

A

\

N INFL

V \

\

\

\

P

206

LaVerne Masayesva Jeanne and Ken Hale

The CFC for NP is clearly S, since that is the minimal category containing the governor of NP (i. e., P) and a subject (i.e., NP). Therefore, according to the binding theory, anaphoric NP must be bound within S, as is the case in (8 a), while pronominal NP must be free within S, as it is in (8 b). The sentences of (8) conform perfectly to the Binding Theory. The same can be said of the sentences of (10) below, in which the object of the main verb kun'tuva (kick) is a nominal possessive construction: (10a)

b)

Taqa Taqa 'Taqa Taqa Taqa 'Taqa

fee] moosa-y kun'tuva [ec] cat-ACC :PROX kicked kicked his (Taqa's) cat.' pu-t moosa-y-at kun'tuva him-ACC cat-ACC -OBV kicked kicked his/her (not Taqa's) cat.'

The structure involved here is essentially that set out in (11) below:

(»)

§

'

/

• • 1iii vN% '' s I

\

VP t \

NP

INFL

/ /

\

/

\

NP /

/

/

NP

/

\

\

\

\

N

The structural relations in (11) are perfectly analogous to those in (9). Accordingly, the CFC for NP is S. Anaphoric NP is bound in S, as in (10 a), and pronominal NP is free in S, as in (10 b), in conformity with the Binding Theory. In relation to the Binding Theory, the Hopi structures (5), (9), and (11) all share a common characteristic: The morphological properties of the governor of NP signal the anaphoric status of the latter. Thus, for example, if the verb in (5) is marked with the reflexive/reciprocal prefix naa-, then its object, NP, is an anaphor; if the verb is not so marked, then its object is not an anaphor. Similarly, if the postposition in (9) is marked with the prefix naa-, its object is an anaphor; otherwise, the object is a non-anaphor. Finally, in the possessive construction of

Argument obviation and switch-reference in Hopi

207

(11), if the head noun is marked with the obviative element -at (plural -am), the possessor expression governed by the head noun is identified as a non-anaphor; if the obviative element is absent, the possessor is an anaphor. This observation led Jeanne (1978) to propose that a single general principle is at work. Briefly, the principle is as stated in (12) below (slightly modified from the original to conform to the conception of Hopi anaphora being developed here): (12)

The Hopi Rule of Coreference: If an XP structure is identified (by the morphological properties of its head, X) as a-proximate, then its principal argument (governed by the head X) is a-anaphoric and, accordingly, bound (if [+anaphoric]) or free (if [ — anaphoric]) in relation to the subject of the minimal CFC containing that argument.

The term proximate here is borrowed from the original Algonquianist terminology of obviation (see Jeanne 1978, for discussion). In the UtoAztecan usage, adopted by some linguists, the plus, or proximate, value of the opposition (glossed PROX) designates the relation in which one argument is identified as coreferential to another; by contrast, the minus, or obviative, value (glossed OBV) designates the relation in which a particular argument is identified as disjoint in reference to another. In using this terminology here, we posit a feature [proximate] which we claim is generally associated with the heads of syntactic categories (being determinable from the morphological properties thereof) and which serves to identify the anaphoric status (i.e., status as bound or free) of their arguments. The Hopi Rule of Coreference, in conjunction with the general Binding Theory (1), correctly accounts for the coreference facts of the sentences of (2), (8), and (10). It does not, however, immediately account for the coreference facts of subject obviation sentences of the type represented by (6 a —c). In particular, we are left with the apparent paradox that the subject of (6 a), for example, is necessarily coreferential with the matrix subject while not, technically speaking, bound to the latter. Our task now is to determine the manner in which the Hopi Rule of Coreference extends to subject obviation. In each of the sentences of (6), the dependent clause is marked to indicate the coreference relation which its subject must bear in relation to the subject of the matrix clause. This, we contend, is simply an extension of the general Hopi principle of marking the head of a construction for obviation (i. e., for a value of the feature [proximate]), thereby

208

LaVerne Masayesva Jeanne and Ken Hale

defining the anaphoric status of its complement (i.e., its "principal argument", in the terminology of (12) above). The elements which effect this marking in (6) are the [+proximate] complementizer -t (appearing in (6 a); this is just one of a larger system of proximate complementizers; cf. Jeanne 1978, for more detail), and the obviative, or [ — proximate] complementizer -q (appearing in (6 b—c)). We follow Finer (1985) in identifying these elements with the category "complementizer" (COMP, or simply C), and we assume further, as Finer and others have done, that the complementizer fits into the x-theory in the usual way. Thus, the category C(OMP) heads a phrasal projection CP, and its complement (i. e., the principal argument of CP) is the sentence, S. These structural relations are included in (7) above; we repeat the essential substructure in (13) below: (13)

CP

S

C

1

/ VP INFL NP If C is [+proximate] (e.g., -t, as in (ta)), then its complement is anaphoric, and therefore bound in its CFC. Notice, however, that the complement of C is not the subject of the dependent clause (NP) but rather S, the clause itself. What does it mean to say that the S in a proximate structure is boundl This is the key question in arriving at a conception of subject obviation as a natural extension of the Binding Theory, which normally concerns itself with coreference relations among arguments in the standard sense (i. e., NP arguments of verbs). The answer lies in the nature of the category S. If, as originally suggested in Jeanne (1978), the head of S is INFL, then, to say that S is bound in a given structure is to say that INFL is bound. But what does it mean to say that INFL is bound? In Hopi, as in most languages in which it is an autonomous category, INFL contains elements identifying certain properties of its two complements, the VP and the NP subject. Thus, it contains an element, sometimes abbreviated Tns, defining the tense category of the verb; and it contains an element, often labeled AGR for "agreement", which identifies one or more (pro)nominal features of the subject. In the Hopi sentences used so far in illustration, the category

Argument obviation and switch-reference in Hopi

209

INFL is, in fact, phonologically non-overt, due simply to the fact that the tense category of the illustrative sentences is consistently non-future (always marked by zero) and the subjects are consistently singular in number (also marked by zero). There are, however, overt occurrences of INFL in Hopi — e.g., in the category of tense, -ni 'future', -ngwu 'usitative'. The INFL element which is relevant to our discussion is AGR. In Hopi, the number category of the subject is marked in INFL. Although it has a number of distinct morphophonological realizations (see, for example, Hale, Jeanne, and Pranka 1984), the regular pattern marks the opposition by means of the suffix -ya, for plural, and the zero element, for non-plural, as illustrated in (14) below: (14 a) b)

Puma hohonaq-ya they play-PL 'They are playing.' Pam hohonaqa he/she play:NONPL He/she is playing.'

We wish to suggest that it is the AGR element in INFL which is identified as bound or free by the obviation system inherent to the governing C in Hopi. That is to say, if C is [+proximate] in (13), then the AGR component of INFL is [+anaphoric]; and conversely, if C is [—proximate], then AGR is [—anaphoric]. The question now is this: if AGR is bound in a given instance, what is it bound to? And if AGR is free, what is it free in relation to? The answer to this is relatively straightforward, given a slight, though rather natural, extension of conventional notions. First, the binder of AGR must c-command it, by definition. If we look at the structure given at (7) above, we see that the matrix S stands in precisely the right structural position to bind the subordinate S. The relevant portion of (7) is repeated here for convenience: (15)

S' / CP / \ ' \ N S C

X S

Clearly, S c-commands S. The former could, therefore, bind the latter. Of course, we are claiming that the relevant binding relation here

210

LaVerne Masayesva Jeanne and Ken Hale

involves AGR. But this is a feature of S, by virtue of the fact that S is a projection of INFL, of which AGR is a component. And this is the sense in which we maintain that S is bound (or free, as the case may be) in relation to S. If AGR of S is identified as anaphoric, it is bound to AGR of S; if AGR of S is non-anaphoric, then it is free in relation to AGR of S. If this is the correct idea for subject obviation in Hopi, then it will follow automatically from the Hopi Rule of Coreference (12), provided we extend the notion subject to include not only [NP, S] but also AGR (cf. Chomsky 1981, in which the notion "SUBJECT" (upper case), introduced into the Binding Theory to handle certain problems of English, subsumes AGR as well as the traditional "subject"). This extension is not unreasonable in view of the fact that AGR does, in fact, represent certain features of the subject. We assume that the agreement relation which holds between AGR and the subject NP is a form of identification, of AGR with the subject. It is by virtue of this identification that the coreference relations found in sentences of the type represented by (6 a—c) are defined. Thus, for example, the subject of the dependent clause of (6 a) is necessarily preferential with the matrix subject because it is identified with the dependent AGR which, in turn, is bound to the matrix AGR with which the matrix subject is identified. Note * We take great pleasure in dedicating this paper to the memory of Professor Stanley Newman who has, both directly and indirectly, influenced our work in the linguistic study of languages of Native America. His integrity as a scholar has been inspirational for us, and the products of his investigations over the years have been an endless source of extraordinarily rich material for research by countless subsequent scholars engaged in the general program of furthering our understanding of the human capacity for language.

References Chomsky, Noam 1981 Lectures on government and binding. Dordrecht: Foris. 1986 Knowledge of language: Its nature, origins and use. New York: Praeger. Finer, Daniel 1984 The formal grammar of switch-reference, University of Massachusetts, Doctoral Dissertation, Amherst, Massachusetts.

Argument obviation and switch-reference in Hopi 1985

211

The syntax of switch-reference, Linguistic inquiry, Cambridge, Massachusetts. Hale, Ken—LaVerne Masayesva Jeanne—P. Pranka 1984 Where's Suppletion? Paper presented to GLOW Conference, Copenhagen. Jacobsen, William H. 1967 Switch-reference in Hokan-Coahuiltecan, in Dell H. Hymes and William E. Bittle, eds., Studies in Southwestern ethnolinguistics. The Hague: Mouton, 238-263. Jeanne, LaVerne Masayesva 1978 Aspects of Hopi grammar, MIT Doctoral Dissertation, Cambridge, Massachusetts. Reinhart, Tanya 1976 The syntactic domain of anaphora, MIT Doctoral Dissertation, Cambridge, Massachusetts.

Some agent hierarchies in Upper Chehalis M. Dale Kinkade

Abstract Agent hierarchies have been previously described for four Salishan languages (Lushootseed, Lummi, Halkomelem, Squamish), all related in some way to ergativity. Upper Chehalis, in another branch of the Salishan family, also has person hierarchies, but no ergative patterning in the language. One hierarchy prohibits the use of a second person object suffix with a first plural subject. The other requires that a human object with a non-human subject must be expressed with a special (otherwise obviative) suffix.

1. The intent of this paper is to present some information on agent hierarchies in Upper Chehalis.1 Because of the nature of the data on this language, this information is almost certainly only fragmentary, but perhaps worth looking at in relation to other studies of hierarchies in nearby Salishan languages (Jelinek and Demers 1983; Gerdts 1983). Since the hierarchies in Upper Chehalis are apparently quite different from those in the languages to the north (Lushootseed, Lummi, Halkomelem, Squamish), further studies will be in order to try to understand how these various patterns came about, and why they differ from language to language, even when the languages are quite closely related. Among other things, the hierarchies in the languages north of Upper Chehalis seem to have a relationship to ergativity in those languages; that is not the case in Upper Chehalis. Studies devoted specifically to hierarchies in language seem to be relatively recent (besides the references above, see particularly Silverstein 1976, Zwicky 1977, and, for a recent study of Algonquian hierarchies, Jolley 1983). I know of none from the early 1960s when I was doing field work on Upper Chehalis, and when it would have been possible to elicit specific pertinent information. Since this was not possible at the time, I must now rely on my field notes (which concentrate on other topics) and on texts, and some information on hierarchies may thus have to be only inferential. Much remains to be done on Upper Chehalis texts; Boas left

214

M. Dale Kinkade

some 580 pages of texts, but only published a third of one story. My own collection of texts is much smaller. There is undoubtedly a vast amount of syntactic information recoverable from these texts, but that remains to be done. I have noted two agent hierarchies in Upper Chehalis, one having to do with second person objects, and the other with third person. I see no connection between the two. 2. The first of these hierarchies is a prohibition on the co-occurrence of a second person object suffix (singular or plural) with a first plural subject. This restriction applies in both aspectual sets of person markers, where continuative aspect subjects are suffixes and non-continuative aspect subjects are enclitics; object suffixes precede both types of subject markers. Table 1 shows the Upper Chehalis object and subject markers, and is included here to facilitate recognition of these components in example sentences given later. Table 1.

Pronominal Markers objects

1st sg 2nd sg 3rd 1st pi 2nd pi 3rd pi

subjects

possessive

continuative

noncontinuative

continuative

noncontinuative

-cal-, -mal-ci-, -mi-y-tul-, -mul-tul-, -mul-y-

-c, -mi -ci, -mi -n, -xw -tuli, -muli -tuli, -muli yams (or = 3rd)

-ans -s -n -stawt -alp -Ut

in c 0 ci calp yams

n?a-s -ci -nalp -s

Third person object is also rarely -l-/-i, and in causatives is -stw- or -tw-j-xw.

When the second person-first plural sequence is called for semantically, we find instead a third person object suffix, and second-personness is indicated by a syntactically coreferential object complement. 1. 2. 3. 4. 5.

sle'a-l'stustawt fit Xa-lixw 9

it mayxw

ci tit

nt mayxw

ci tit

smaytustawt

'we are looking for you' 'we looked for you' nawi 'we took you in' ?ilapa 'we took you (pi.) in' 9 ilapa 'we are taking you (pi.) in'

tit nawi ci tit nswi

tit

Some agent hierarchies in Upper Chehalis

215

6. 9 it 9axim ci tit 9ilapa 'we saw you (pi.)' 7. 9amu i c'dcn ci tit nawi 'if we defeat you' (from Boas; my informant would have said c'acm) This restriction appears to be total — I have come across no examples whatever of a first plural subject preceded by second person object suffixes. But other constructions involving second person are variable. I find a number of instances of avoidance of second person objects with first singular (8) or third person (9,10) subjects, and one avoidance each of a first plural object with a second singular (11) or plural (12) subject. 8. 9. 10. 11. 12.

9

it d-lixw cn tit niapa 'I looked for you (pi.)' s£a-l'stwn tit 9ilapa 'he is looking for you (pi.)' ssawlayn tit 9ilapa 'he is asking you (pi.)' ssawlays tit 9inim 'you are asking us' 9 it Xa-lixw calp tit 9inim 'you (pi.) looked for us'

But these constructions are inconsistent, and second person affixes are usually used. Thus the completive aspect form of sentence 9 was given as sentence 13, and sentence 14 is exactly parallel to sentence 8. 13. 9it Xal'stumuli 'he looked for you (pi.)' 14. 9it sawlamuli cn 'I asked you (pi.)' And parallel to sentences 11 and 12 we find sentences 15 and 16. 15. smaytomols 'you are taking us in' 16. 9it Xa-l'stumuli calp 'you (pi.) looked for us' I have observed no similar avoidances of other persons, so second person seems to be in the weakest position in hierarchies. The same avoidance occurs in dependent constructions, where the subject is indicated with a possessive affix. Since the first person plural completive subject clitic and possessive suffix are phonologically identical, these constructions differ from those in 2, 3, 4, 6, and 7 in that I leave a space before a subject clitic, and their dependent status is marked by a prefixed s-.2 17. miita t s9axanci tit nawi 'we didn't see you' Contrast this with 18 and 19. 18. miita t 9as9dxtuli 'you didn't see us' 19. miita t s99xtuls 'he didn't see us' 2.1 Besides involving a different set of avoidances, there is another aspect of this Upper Chehalis hierarchy which differs from languages to the

216

M. Dale Kinkade

north. Where those languages resort to passive constructions (Jelinek and Demers 1983, Hess 1973), Upper Chehalis simply uses a third person object with an object complement to indicate second person. I have no evidence that passives are used there for this purpose, but it may be possible. Passives are common in Upper Chehalis, and occur frequently in texts, but I have made no specific study of them to see what restrictions in their usage there may be. Any person can be subject/patient of a passive construction: 20. 21. 22. 23.

Hi fit ?it 9 it

txwiawaistm txwiawaistm txwiaw'aistm txwiaw'aistm

'he was left' cn 'I was left' c 'you were left' calp 'you (pi.) were

left'

I have far fewer examples of passives in the continuative aspect in my Upper Chehalis data, but have no evidence that there are any other restrictions on them. Their construction is different, however; the subject/ patient is indicated by an object suffix, not a subject clitic, as in noncontinuative aspect constructions: 24. sqakwslyanisicalsts 25. sxwo-?xwu?icalsts

'my teeth are chattering' 'I am catching cold'

(These are very peculiar constructions, even for Salish. The passive marker here is -sts, for which I know no cognates. Nearly all my examples of continuative passives have third person subject/patients, where the subject is often zero, but states of health are typically continuative passives as in these examples.) 3. The other Upper Chehalis agent hierarchy has to do with human vs. non-human third person subjects and objects. Here the restriction seems to be that a predication with a non-human subject may not have a human object expressed by one of the usual object suffixes. But again, passive is not the technique used to avoid these constructions. Instead, Upper Chehalis has an obviative suffix which is used in these instances to express a human object. (As far as I know, among Salishan languages only Upper Chehalis, Cowlitz, and Columbian have obviative suffixes.) The obviative suffix is -(t)wal-l-(t)wali, and simply replaces the usual third person object suffix. As an obviative, it is used as a device to index multiple third persons in a sentence, as in Algonquian. Boas (1933:109) gives clear examples of this usage:

Some agent hierarchies in Upper Chehalis

26. 27.

tit q'iwts, tit q'iwts,

217

'he called her and she saw it' dxtwali 'he called her and she saw the one who

huy n ta 9axtn huy n ta

9

called' Such constructions are relatively straightforward, and occur frequently in texts. But the obviative is also used when a human is the object of a sentence with a non-human subject.3 The constructions always puzzled me until I recently realized that they reflect a hierarchical distinction. Most examples I have are with an animal as subject: 28. 29. 30.

9

it xay'twali s-k'al'stwaln 9 acqwantswali

tat qaxa9 tat qaxa9 tat qaxa9

'the dog growled at him' 'the dog is looking for him' 'the dog was afraid of him'

A human object with an indefinite subject also seems to call for the obviative: 31. 32.

9

acwe acwe

9

x cutcn wa• t ya-^'twain 'somebody took him home' x t wa- t sq'iwtswaln 'somebody is calling him'

How all this fits together I do not know, and further study of the Upper Chehalis obviative is needed. 4. I draw no conclusions from these data. It seems clear that agent hierarchies in Upper Chehalis are different from those in at least some other Salishan languages. To understand how such differences might have arisen will require much more knowledge about hierarchies in Salish and other languages of the Northwest. More also needs to be known about hierarchies in other languages of the world. The Upper Chehalis system shows some parallels to the Kiksht system described by Silverstein 1976, but the relevance of these parallels is unclear to me, since the Kiksht hierarchy is coordinated with the split-ergative case system of that language, whereas case is scarcely relevant to Upper Chehalis (unless one wishes to view the pronominal markers as expressing case, in which instance the language is strictly nominative-accusative).

Notes 1. My Upper Chehalis data were collected from 1960 onward under the auspices of the American Philosophical Society Library in Philadelphia, Indiana University, and the National Science Foundation. Forms cited here were obtained from the late Silas Heck.

218

M. Dale Kinkade

2. Unfortunately my field notes have no examples of non-dependent possessives with subjects of the type discussed by Hess 1974, 'we are your fathers', 'you are our father(s)'. Perhaps examples will turn up in texts. 3. Note that with an obviative object suffix, a complement, when present, refers to the subject of the predication. With a regular third person object suffix, such a complement would refer to the object, and with an intransitive predicate the complement refers to the subject; this is the usual Salishan pattern. a. m ?ax3n tit qaxa> 'he saw the dog' b. 9it wiq'i tit qaxa"> 'the dog ran' This, as far as I can tell, is as close as Upper Chehalis comes to ergative-type constructions.

References Boas, Franz 1933 A Chehalis text. UAL 8, 103 -110. Gerdts, Donna B. 1983 A surreptitious person/animacy hierarchy in Halkomelem Salish. Paper read at the 22nd Conference on American Indian Languages, Chicago, Illinois. Hess, Thom 1973 Agent in a Coast Salish language. IJAL 39, 89-94. 1974 How do you say, "You are our father." in Salish? Paper presented to the 9th International Conference on Salishan Languages, Victoria, B. C. Jelinek, Eloise—Richard A. Demers 1983 The agent hierarchy and voice in some Coast Salish languages. UAL 49, 167-185. Jolley, Catherine 1983 Algonquian person hierarchy: morphosyntactic or semantic? Actes du quatorzième Congrès des Algonquinistes, ed. by William Cowan, 281—291. Ottawa: Carleton University. Silverstein, Michael 1976 Hierarchy of features and ergativity. Grammatical categories in Australian languages, ed. by R. M. W. Dixon, 112—171. Canberra: Australian Institute for Aboriginal Studies, Linguistic Series 22. Zwicky, Arnold M. 1977 Hierarchies of person. Papers from the thirteenth regional meeting, Chicago Linguistic Society, ed. by Woodford A. Beach, Samuel E. Fox, and Shulamith Philosoph. Chicago Linguistic Society 13, 714—733.

Vowel ablaut and its functions in Yuman1 Margaret Langdon

Abstract This paper describes the various functions of vowel ablaut in Yuman languages. These include various types of verbal plurality, and a number of derivations. The latter produce verbs with attenuated meanings, verbs derived from nouns with meaning 'have the characteristics of N', various derivatives from numeral verbs, causatives, nouns describing typical objects associated with the source verb, subject and oblique relativizations. These are argued to reduce to two main functions: a semantic one of "extension" and a grammatical one of "nominalization".

Newman's (1944, 1946) masterful description of Yawelmani and other Yokuts dialects has been in recent years the subject of much discussion as linguists considered the theoretical implications of the phenomenon of vowel harmony which he so lucidly presented. Many other aspects of the structure of Yokuts are worthy of attention and again Newman's work is rich in information. The phenomenon of vowel ablaut has attracted less attention in recent scholarship, not because it is inherently less interesting, but because current theory has had little to say about this rich area of language structure, even though it is attested in a great variety of language types (including Indo-European, among others). In lucidly concise fashion, Newman (1946:227) introduces the topic as follows: Stems are formed from roots by processes of vowel change. These processes are to be regarded as operating on two planes: on the one hand, dynamic vowel processes effect ablaut changes that are to be defined in terms of morphological conditions; on the other hand, a few phonetic processes introduce additional vowel changes of a mechanical nature. In the formation of stems these two planes interact; a stem which has undergone dynamic vowel changes may, in turn, be subjected to secondary phonetic changes.

220

Margaret

Langdon

Processes of a similar nature are found in Yuman languages; their nature and functions are investigated in this paper. There are ten Yuman languages still spoken, normally considered to divide into four subgroups.2 In these languages, words are built by a variety of processes around a "root" consisting of a stressed syllable, typically of the shape CVC, sometimes only CV, YC, or even V. The root vowel may be either long or short, so root types are as follows: CV(:)C,

C V(:),

V(:)C,

V(:)

In addition to the common processes of prefixation and suffixation, an important characteristic of these languages is that they abound in rootmodification processes. These consist of two main types: consonant alternations and vowel alternations. Consonant alternations typically involve sound symbolism, a productive phenomenon, which I have described elsewhere (Langdon 1971). Vowel alternations (the topic of the present paper) also abound and serve a great variety of functions. From the phonological point of view they consist of vowel quality and vowel quantity alternations, sometimes both simultaneously. Historically, however, the basic alternation can be shown to be one of length rather than quality, the vast majority of quality alternations being the result of language-specific developments involving vowel lowering and raising, as well as diphthong coalescence (Langdon 1976). Length alternations, however, which cannot be accounted for by rules, are attested in all the languages of the family, and pervade the word-formation mechanisms in intricate ways. By far the most common phenomenon is one in which a basic short vowel is lengthened in a derived form. However, each language also shows alternations in the opposite direction, i. e. basic long vowels are shortened in derived forms. The latter are far less numerous than the former and individual languages vary considerably in the number of forms undergoing the long to short alternation. Nevertheless, the process is best characterized as one of length ablaut which may affect any root vowel and whose effect is to lengthen a short vowel or to shorten a long one. While length ablaut combined with the various diachronic processes affecting individual languages accounts for the phonological shapes in a large number of cases, it should be noted that analogical developments have created cases whose history is far from clear, while in some languages new phonological developments have negated the effect of the ablaut process. For example, in Diegueno, there is a general rule by which short root-final vowels are lengthened, so that, in some cases, the ablauted and

Vowel ablaut and its functions in Yuman

221

non-ablauted form may have the same vowel. In some other languages, certain stems based on long vowels have short forms conditioned by discourse phenomena rather than derivational ones. The situation is made even more opaque by the fact that, while vowel ablaut is a common phenomenon, it is extremely difficult to characterize its exact function. Its best known use is in the formation of verb forms conveying plurality, but it is also widely attested in nominalizations, and in derivation of verb stems from other verb stems. While ablaut can be the only difference between forms, this is actually the least common case, the most common ones involving in addition various prefixes and/or suffixes. As if this were not complicated enough, there is the maddening fact that there is no way to predict on phonological, syntactic, or semantic grounds whether or not the ablaut process will affect a particular root, even if other items which appear to be members of the same class are so affected. In the following discussion, therefore, it should not be assumed that the observations made can be translated into a system of airtight rules. By far the most numerous examples involve some aspect of plurality in verbs. Some nouns are affected also, but most nouns in Yuman do not take special plural forms. Those that do are typically transparently derived from verbs, so that plurality is a basically verbal notion in Yuman. On the other hand, plurality is a misleading term, useful only to characterize a common feature of the translations of these forms into English, but differentiating, depending on the language, such notions as paucal subject, multiple subject, plural object, collective or distributive subject, distributive object, repeated action, not to mention ambiguity between one or more of these categories. From a formal point of view, the inflectional morphology of pronominal marking distinguishes only three persons, with no distinction of number, except that some languages allow a peripheral prefix denoting plural object. A simple verb paradigm illustrates these facts: (1)

DI 3

'-mui nt-mui

'I gather acorns' 'you gather acorns'

'-mu:i m-mu:i

w-mui

'he/she gathers acorns'

w-mu:l

'we gather acorns' 'you all gather acorns' 'they gather acorns'

Many stems are capable (with semantic and grammatical restrictions) of forming more than one plural stem, up to four in some languages. The above discussion should make it evident that the situation described here is much more akin to derivation than to inflection and, in fact, I have been in the habit of thinking of these facts as part of the derivational

222

Margaret Langdon

component of the grammars of these languages (although this may not be the best synchronic characterization in all cases). An argument for this position is that, as opposed to languages where "plural" is a grammatical category, there are no obligatory number agreement rules in most Yuman languages, in that semantically plural subjects do not require an appropriately marked plural verb and are just as likely to take what one would assume to be a "singular" stem, which I prefer to call the "non-plural" stem, i. e. one that is not overtly marked by one of the plurality markers, but is semantically neutral with respect to plurality. All Yuman languages have examples of non-plural stems forming plural stems simply by ablaut of the root vowel even in those languages where vowel length is otherwise somewhat problematic (a separate issue I will not deal with here). HU MO CO KI

'to 'to 'to 'to 'to 'to

fall' dig' break' walk' do all night' crack open'

non-plural man hwal uPe:s p'aw anya:k car

plural ma:n hwa:l uPes p'a:w anyak ca:r

More frequently, however, the plural stem is formed by a combination of prefixation, ablaut, and suffixation, sometimes resulting in a plural stem differing from its non-plural stem by as many as four distinct morphological processes. (3) HA MO DI KI

'to pull by the hair' 'to catch a disease' 'to sleep' 'to rise from prone position'

non-plural sico ahno: xwrta: tip

plural tisco: cu:hnu:c p»xemac titi:p

Needless to say, in these cases, it is far from clear what increment of meaning each individual modification contributes to the final form. Even in forms marked only by ablaut, an attempt to characterize more precisely the function of ablaut fails. For while one language may lean toward distributive subject or object, the next language may in fact point to collective as a basic function. We can only conclude that each language has reshaped an already complex system in idiosyncratic ways, leading to a system even more complex because so erratic. Another function of ablaut is to derive verb stems from other verb stems. A subset of these are stative verbs deriving new stative verbs with "attenuated meanings", as in

Vowel ablaut and its functions in Yuman (4) YU

MO DI CO HA

basic

attenuated

arfor 4a\'or xacor 'akuc 'axwet 'akor haco:r xi:cur xma:P nym6a:v

'be 'be 'be 'be 'be 'be 'be 'be 'be 'be

striped' sleepy' winter' old' red' distant' winter' winter' white' at dawn'

anyu:r 4av 'u:r xacu:r 'aku:c 'axwa:t 'aku:r hacu:r X0cu:r xmwaP n?mOav

223

'be varicolored' 'be drowsy' 'be cold' 'grow old' 'be bloody' 'be near(time)' 'be cold' 'be cold' 'be off-white' 'be faded'

Note the MO alternation here is one of quality, not quantity. This is probably due to the fact that vowel length is somewhat erratic in this language. Historically, however, it is probably an *u/u: alternation. Verbs can be derived from nouns by adding a suffix -y which is often accompanied by ablaut. This suffix derives from nouns verbs with the approximate meaning 'have the characteristics of N'. This process is typically accompanied by ablaut in Yuma, but not in other languages. Thus (5)

YU

axwe 'akwe yavume ikwe

enemy 'cloud' 'beard' 'horn'

axwa:y 'akwi:y ya:vumi:yv kwi:kwa:yv

'to 'to 'to 'to

'be big' 'fog' 'cloud'

ny-eku:y xayu:ma:y kwi:y

'be an old woman' 'be foggy' 'be cloudy'

be hostile' be cloudy' be bearded' have horns'

but, without ablaut Dl HA

'i:ku: xa:ma: kwi:

Note that the DI forms may in fact be cases of ablaut masked by the final vowel lengthening rule and HA kwi: may be, at least historically, from *kwiy. A benefactive suffix -y with ablaut of the root is attested deriving verbs from verbs, mostly in Yuma, some in Mojave, but no examples have been found outside of the River subgroup. (6)

YU

MO

awe ama asi a'we iya:

'to 'to 'to 'to 'to

do' eat' drink' do' get'

awe:y u:ma:y u:se:y u: 'wi: u:yay

Note that Mojave u: 'wi: is probably from u:'wi:y.

'to 'to 'to 'to 'to

do for' feed' give to drink' do for' get for'

224

Margaret Langdon

Another process involving ablaut derives nouns from verbs. Semantically, these nouns denote the object most prototypically associated with the meaning of the verb. (7)

YU

MO DI

CO

takyet axer a:sva:r ama isva.r ak«na: -nak -si: c'ur cusi cukat

'to chop' 'to tie up' 'to sing' 'to eat' 'to sing' 'to tell a story' 'to sit' 'to drink' 'to sit on' 'to comb' 'chop'

ta:kya:t a:xi:r sa:var 'icama:vc sa.var 'a:knac 'a:na:k 'a:sic cwa'u.r cwasi.s caka.t

'axe' 'rope' 'song' 'food' 'song' 'story' 'chair' 'cup' 'chair' 'comb' 'axe'

Numerals, which are verbal in Yuman, are bases for deriving by ablaut other verbs, sometimes nouns, with meanings such as 'one by one', 'in pairs', 'twins' , 'double', etc. (8)

HA YU

sit 'asent xavik

'be one' 'be one' 'be two'

MO DI CO

havik xmvak xwak

'be two' 'be two' 'be two'

si:t 'asi:nt xava:k xu:va:k xa.vaka havak xa'wa:k xa.wak

'be one by one' 'be one by one' 'be two by two' 'double' 'twins' 'twins' 'twins' 'twins'

Note the lack of root vowel ablaut in the Cocopa example. With other morphological apparatus, ablaut also is involved in the derivation of causative and other transitive stems, particularly in Yuma. (9)

YU

anex qol 'aPaqwaq

'be quiet' 'be long' 'be bitter'

ta:ni:x ta'aqo.I ta'aPa:qwa:q

'do quietly' 'lengthen' 'make bitter'

'be warm' 'be dry'

ta:pin ta:say

'to warm up' 'to dry'

but, with no ablaut, DI

-pin -say

Finally, and once again with other morphological marking, ablaut is involved in the formation of nominalized verb forms appearing in relative clauses, with different forms systematically required for subject relativization as opposed to oblique. Subject relativization always involves the presence of the prefix *kw-, while oblique relativization is marked in a variety of ways.

Vowel ablaut and its functions in Yuman 225 Subject relativization: (10)

CO i:ma n"xat DI '#c-yu:w -ma:w wa: nur

'to dance' 'to have a pet' 'to sing' 'not to be' 'to sit, be' 'to know'

kwyama:s kn?axa:t '«ckwecya:w kumuw kuwac k"«nu:r

'he who dances' 'he who has a pet' 'singer' 'he who is not' 'he who is there' 'know-it-all'

'shoot' 'be with' 'to drink'

akyay u:da: v u.de.t

'to eat'

ama

'that which he shoots' 'the one he is with' 'that which he drinks' 'that which he eats'

Oblique relativization: (11)

MO ok^a: idaw iOe ama:

Can anything more general be said? All the evidence points to the basic function of ablaut being one of derivation, even in the case of so-called "plurals". As early as 1973, Munro, in the only attempt I know of to deal with some aspects of the problem, proposed for Mojave a special relationship between pluralization and the nominalizations in relative clauses based on arguments revolving around the notion that a plural sentence has an underlying structure where it is embedded as the subject of an abstract predicate PLURAL, thus reducing plurality to a special case of nominalization. I would like to take here another approach which is both comparative-historical and semantic. I start with the observation that, except in very surface ways, the distinction between nouns and verbs is far from straightforward in Yuman languages. As noted above, "plurality" involves various kinds of derivations and, reviewing the material presented in this paper, it is evident that all other instances of ablaut involve derivation. Furthermore, two kinds of nominalization, i. e. those in (7) "prototypical objects" and (11) "oblique relativization", share the meaning conveyed by oblique relativization, with the nouns of (7) in addition requiring interpretation as having an unspecified agent. So while standard oblique relativizations involve a specified subject, e.g. 'that which he shoots', the items in (7) convey such notions as 'that which one eats', 'that which is eaten'. The unspecified subject notion is conveyed, not by the ablaut which both classes of derivatives share, but by an infix a:- and other additional morphology. Another observation is that the 'benefactive' suffix -y is limited to the River languages, and may represent an innovation and not be part of the general topic. It should also be noted, in connection with the other suffix -y 'characterized by', that Yuma is the only language where it has ablaut

226

Margaret

Langdon

associated with it and that Yuma has a tendency to lengthen vowels in a number of situations where other languages do not. So perhaps this is not a true part of the basic problem either. I now wish to propose that "plural" instances of ablaut and the items of (4) "attenuated meanings" and (8) "numeral derivatives" share an important semantic component. If so-called plurality is redefined in Yuman as expressing the extension of the state or process denoted by the verb stem in a distributed or collective fashion to its various arguments or to the state or process itself, the so-called "attenuated meaning" verbs convey the extension of the base meaning from the simple description of a state (they are all stative) to a process involving the acquisition of that state (e.g. 'be old' vs 'become old'), which certainly includes in its inception an attenuated form of the state in question. As to the forms derived from numeral verbs, they also extend the state in distributive fashion, e. g. 'one by one', or collective, e. g. 'twins'. It should also be noted that, with the exception of plural stems which match their nonplural sources in transitivity and stativeness, most derivational processes I have discussed involve stative or intransitive notions. The only exceptions are the special causatives of Yuma in item (9), which could be accounted for in one of two ways: either they are again one of the cases where Yuma alone lengthens vowels (an innovation in that language) or they are doubly derived forms adding by ablaut to the stative notions of the basic verb the process of becoming as in the items in (6), but in addition becoming causative by means of the t- prefix (a typical causative marker in Yuman) with the usual attendant effects of transitivizing and the addition of an argument. We are thus left with two basic functions: a semantic one of "extension", and a grammatical one of nominalization. I will refrain for the moment from trying to reduce these further though I think it might be possible. If this hopes to characterize the functions of vowel ablaut in Yuman, its origins must surely be sought at a deeper Hokan level. While a number of Hokan languages have instances of vowel ablaut, it is certainly premature to try to relate these to the Yuman facts. Encouraging, however, is the suggestion which I owe to Halpern (personal communication), that Yuman ablaut phenomena may be relatable to Pomoan stem alternations which also include among their functions notions of plurality. While this suggestion needs to be supported by a much more finegrained study than the casual observations presented here it seems to me that only by making hypotheses such as these can we hope to gain a point of entry into the complex problems posed by such recalcitrant families as Hokan.

Vowel ablaut and its functions in Yuman

227

A final note will allow me to acknowledge information supplied to me by colleagues who kindly commented on an earlier version of this paper. It concerns the typological presence and function of ablaut in other American Indian languages. Specifically, Brent Galloway tells me that Salish languages abound in ablaut phenomena which are uncannily like the Yuman facts in their many functions, and Catherine Callaghan points out that Miwok also has alternations very similar to the Yuman ones in function. This suggests that a systematic survey of ablaut types and their function may reveal unexpected generalizations and contribute another dimension to typological concerns.

Notes 1. A preliminary version of this paper was read at the 1984 meeting of the American Anthropological Association. 2. Subgroups with language names and abbreviations used in this paper are as follows: Pai: Upland Yuman (Havasupai HA, Hualapai HU, Yavapai YA), Paipai River: Mojave MO, Yuma YU, Maricopa MA Delta-California: Cocopa CO, Diegueno DI Isolate: Kiliwa KI 3. Sources of the data in this paper are: DI, Couro and Hutcheson 1973; MO, Munro and Brown 1976; KI, Mixco 1985; HA, Hinton 1984; YU, Halpern 1946, 1947; HU, Watahomigie, Bender, and Yamamoto 1982.

References Couro, Ted—Christina Hutcheson 1973 Dictionary of Mesa Grande Diegueno. Banning, CA: Malki Museum Press. Halpern, A. M. 1946 Yuma I, II, III. International Journal of American Linguistics 12.25 — 33, 147-151, 2 0 4 - 212. 1947 Yuma IV, V, VI. International Journal of American Linguistics 13.18 — 30, 9 2 - 1 0 7 , 147-166. Hinton, Leanne 1984 Havasupai songs: a linguistic perspective. Tübingen: Narr. Langdon, Margaret 1971 Sound symbolism in Yuman languages. Studies in American Indian languages, ed. by Jesse Sawyer. (University of California Publications in Linguistics 65) 149-173. 1976 The Proto-Yuman vowel system. Hokan Studies: Papers from the first conference on Hokan languages, ed. by Margaret Langdon and Shirley Silver. (Janua Linguarum, Series Practica 181), 129 — 148. The Hague: Mouton.

228

Margaret Langdon

Mixco, Mauricio J. 1985 Kiliwa Dictionary. University of Utah Anthropological Papers 109. Munro, Pamela 1973 Nominalization and plurality in Mojave. You Take the High Node and I'll Take the Low Node, Proceedings, CLS comparative syntax festival, ed. by C. Coram et al., Chicago. Munro, Pamela—Nellie Brown 1976 A Mojave dictionary (private edition). Newman, Stanley S. 1944 Yokuts Language of California. Viking Fund Publications in Anthropology 2. New York: Viking Fund. 1946 The Yawelmani dialect of Yokuts. In Linguistic Structures of Native America, Cornelius Osgood, General editor. Viking Fund Publications in Anthropology 6. New York: Viking Fund, 222 - 248. Watahomigie, Lucille—Jorigine Bender—Akira Yamamoto 1982 Hualapai reference grammar. American Indian Studies Center, UCLA.

Aspect in Isthmus Zapotee1 Velma B. Pickett

Abstract The aspect prefixes of Isthmus Zapotec have been listed in various articles and books but have not been described thoroughly in terms of their semantics and usage. This article attempts to fill that lacuna, and thus to provide better data as a basis for comparative study of the tense-aspect systems of various languages of the Zapotec family.

The Zapotec languages of the state of Oaxaca, Mexico all have a system of verbal prefixes which represent varying kinds of effect on the action. Most of these are clearly aspectual, rather than tense, assuming the following definitions: "Aspect (the kind of action)". Footnote: "Aspect may also identify a type of process, a quality, or a state" (Nida 1949:167). "Aspects have to do, not with the location of an event in time, but with its temporal distribution or contour" (Hockett 1958:237). "Aspects are different ways of viewing the internal temporal constitution of a situation" (Comrie 1976:3). "Tense relates the time of the situation referred to to some other time, usually to the moment of speaking" (Comrie 1976:1). In the Zapotec languages, aspect is an obligatory category. With very few exceptions, every verb must include one of a set of aspect prefixes. Among the twenty-nine Zapotec languages for which we have data thus far, the set of aspect prefixes number from three (habitual, potential, and completive) to nine. Isthmus Zapotec (IZ) has eight (including stative, which does not occur on all verbs). Several of the languages do not have a contrast between habitual and progressive. Bybee (1985:142) lists 'Zapotec' as one of only seven in a list of fifty nonrelated languages which include the distinction habitual/ progressive. Her sample for 'Zapotec', however, was the IZ language, only one of a large family of languages by the same name. I know of seven others which make this distinction, but twelve which do not.

230

Velma B. Pickett

The following are the eight IZ prefix morphemes, identified by the underlying form in each case (for the current analysis of the allomorphy of these prefixes, see Marlett and Pickett, 1987). The glosses which are used below reflect the conventions currently in use among Zapotecanists. Form Aspect (abbreviation) r(nonfuture) habitual (H) gbcompletive (C) kaprogressive (PR) repetitive past (RP) waz(future) incompletive (F) niunreal (U) kpotential (P) nastative (S) This paper presents each of the eight IZ aspect prefixes in relation to its various meanings and usages. 1. Habitual r-. In its unmarked usage, the r- prefix has a somewhat present tense usage similar to the English use of the present in 'I go to campus three times a week' or 'I go to work early.' Its meaning in IZ, however, is clearly habitual as opposed to an action going on in the present. It implies repeated or on-going action in past or present time, the context (either within the same sentence or a more extended context) giving the clue as to the time of action.2 (1) ci'ke r-ye -be temprano, yanna ma r-ye -be old = time H-go 3h early now already H-go 3h wa-jt3 RP-day 'S/he used to go early, but now s/he goes late.' ke r-anda r-ye -be guuze la, (2) b-iiya si ka -0 ma C-see only PL 3 already not H-able H-go 3h hunt,4 b-i-s-aana ka -0 laa -be C-ThV-CA-remain PL 3 X 3h 'As soon as they saw that s/he could no longer hunt, they left him/her.' (3) gu-diibi nana s-payu -be geela ke C-wash grandma POS-handkerchief 3h night DEM ti ri-nda9 benda because H-smell fish 'Grandma washed his handkerchief that night because it smelled of fish.'

Aspect in Isthmus Zapotee

231

The habitual morpheme is not used for future time. Note the use of the future prefix where the habitual might be expected: (4)

iza ri r-ye -be skwela sya'dos-ti iza year DEM H-go 3h school morning; other-one year ke la, z-ye -be wa-ji DEM , F-go 3h RP-day 'This year s/he goes to school in the mornings; next year s/he will go in the afternoons.'

2. Completive gb-. In its unmarked usage, the gb- morpheme has a past tense interpretation: (5)

pa ji b-eeda -be what day C-come 3h 'When did s/he come?'

In a future context, however, it carries the meaning of future perfect: (6)

ora g-eeda -be la, ma bi-luze -lu? when P-come 3h , already C-finish 2s 'When s/he comes, you will already have finished.'

In addition, this form cannot be used with a negative morpheme which actually negates the action. For this meaning, the unreal is required: (7) (8) (9)

b-eeda -be C-come 3h 'S/he came.' ke n-eeda -be not U-come 3h 'S/he didn't come.' *ke b-eeda -be not C-come 3h 'S/he didn't come.'

But compare these with the following examples in which the negation has a narrower scope. The gb- morpheme may occur in such sentences, but with a different negative form: (10)

kadi not -be 3h 'S/he

para g-uni b-eeda -be g-uuya -be lii sinuke C-come 3h P-see 3h you but = rather for P-do jiiha? work didn't come to see you, but to work.'

232

Velma B. Pickett

(11)

kadi b-eeda -be yannaji sinuke neege9 not C-come 3h today but = rather yesterday 'S/he didn't come today, but rather yesterday.'

In all the Zapotec languages, the completive prefix is also used for singular imperative. In this usage, no person subject clitic is used: (12) (13)

gu-ni9 -be C-speak 3h 'S/he spoke.' gu-ni? C-speak 'Speak!'

3. Progressive ka-. In its unmarked usage, the ka- form implies present progressive action: (14) (15)

zi kay-uni -tu what PR-do 2p 'What are you (pi) doing?' kay-ui -du diija9 PR-tell lp word 'We are chatting.'

In a past context it has a past progressive interpretation: (16) (17)

(18) (19)

ma za'ji kay-uuna -me, already long PR-cry 3a 'It (animal) has/had been crying for a long time.' laga kay-ui nana ne ta Hwan diija ye while PR-tell grandma and Mr. John word C/go yubi ka ba9du ni g-o ka -be look = for PL child REL P-eat PL 3h 'While Grandma and John were talking, the children went to look for something for them to eat.' b-i-zu'lu -me kay-o -me C-ThV-begin 3a PR-eat 3a 'It (animal) began eating.' gu-ri tobi de ka guuze ke ra'ke ka-zii la9ji9 C-sit one of PL hunter DEM there PR-* liver 'One of those hunters sat down there and was resting.'

Aspect in Isthmus Zapotee

(20)

233

b-i-zu'lu -me b-i-kaa -me riji kay-uuna -me C-ThV-begin 3h C-ThV-hit 3a voice PR-cry 3a 'It (animal) began to shout and was crying.'

In a future context, ka- has a future progressive interpretation: (21)

na'ga ora ci nda -ya ra lij -e9 Tamara later when P/go arrive Is LOC home Is Tamara ma kay-uni ta'rea sti? -0 already PR-do assignment of 3 'When I get home, Tamara will already be doing her assignment.'

4. Repetitive Past wa-. This prefix has previously been labelled perfect. However, neither the term perfect nor perfective as explained by Comrie 1976 are precisely applicable. In its unmarked usage, wa- is interpreted as present perfect and in past context as past perfect. However, it is not frequently used as present perfect. The completive form is used in many contexts where present perfect would be used in English or Spanish: (22)

b-iiya -lu Pedru la C-see 2s Peter Q 'Have you seen (or did you see) Peter?'

The wa- prefix always implies repeated action over a span of time: (23) (24) (25)

(26)

ma way-eeda -be conna tiru already RP-come 3h three time 'S/he has already come three times.' But not: *ma way-eeda -be already RP-come 3h 'S/he has already come.' wa-ba?na -be nisi liji -be b-yaana ne i'ra ni only house 3h C-remain with all REL RP-steal 3h 'Only his/her house remained, with all that s/he had stolen.' (The context makes it clear that the action had been repeated on various occasions.) ma conna iza way-eeda -be ra'rP already three year RP-come 3h here 'She has been coming here for three years.'

One common use of the repetitive past is in the negative, with a temporal expression indicating a span of time in which the action has not taken place:

234

Velma B. Pickett

(27) (28)

ma conna gu'bija ke way-aw -a9 already three sun not RP-eat Is 'I haven't eaten for three days.' ma r-aka sono gu'bija de b-eeda -be ne ke already H-pass eight sun since C-come 3h and not g-o -be wa-dii ka -0 laa -be gasti RP-give PL 3 X 3h nothing P-eat 3h 'It's been eight days since s/he came, and they haven't given him/her a thing to eat.'

5. (Future) incompletive z-. The z- form presents a problem in a system of aspects. It is used only in a future time context, with implication of action not yet begun. In order to fit this morpheme into a pure aspect system, I have sometimes labelled it only incompletive aspect. The term is not the best, however, since that would seem to imply an action begun but not completed. At other times I have labelled it future tense and the system as a whole a tense-aspect system. Here I combine the two concepts in the term (future) incompletive. It implies action not yet begun as well as not yet completed: (29) (30) (31)

in'tiika na -lu z-uni -du -ni whatever S/say 2s F-do lp 3i 'Whatever you say we will do.' pa kaa z-uni de -du lii if not F-do ashes lp you 'If not, we will tear you to pieces.' pa z-u-ndaa -tu laa -me o pa za-gibi -du if F-ThV-CA/escape 2p X 3a or if F-shake lp yaga ri za dyaga -tu stick DEM below ear 2p 'Either you (pi) let it (animal) go or we'll beat you (pi) in the head with this stick.'

6. Unreal ni-. The ni- prefix has some features of a past subjuntive mode, and for this reason I have labelled it 'subjuntivo' in the dictionary published for readers with Spanish grammar background (Pickett et al. 1965). However, to cover its varied usages it is better considered an unreal (or contrary to fact) aspect: (31)

bi-la n-yaba -be C-escape U-fall 3h 'S/he almost fell.'

Aspect in Isthmus Zapotee

(33)

(34)

(35)

(36)

235

pa n-uuya -be -ni h-yuu la?ji -be -ni if U-see 3h 3i U-enter liver 3h 3i 'If s/he had seen it, s/he would have liked it.' or: 'If s/he were to see it, s/he would like it.' ke ni-na -0 n-u-yaa -0 lu doo ne nin not U-want 3 U-ThV-dance 3 face rope and neither ke ni-na -0 ni-naaze myati laa -0 not U-want 3 U-grab anyone X 3 'It (the monkey) didn't want to dance on the rope, and neither did it want anyone to take hold of it.' b-iiya -be n-anda -ni ra'ka la, bi-jibi -be n-yuu tu C-see 3h S-hang 3i there , C-fear 3h U-enter who ni-baPna -ni U-steal 3i 'When s/he saw it hanging there, s/he was afraid that someone who would steal it would come in.' z-eda g-uuya si -a9 pa n-u-sig-a?de -lu naa PR-come P-see only Is if U-ThV-CA-given 2s me kaaji nisa some water 'I've just come to see if you might give me some water.'

Note also the use of unreal with a negative morpheme to express a past (non)event, illustrated in (8) and (9). 7. Potential k-. The prefix k- is a type of future, related to morphemes labelled in other Zapotec languages as 'indefinite future'. Semantically, it usually implies a future action, in relation either to the moment of speaking or to the time of the main verb in the sentence. Syntactically, it frequently marks a subordinate clause. The following are some of the distributions which require the k- prefix. a) following an auxiliary motion verb to express coming or going in order to perform an action. Only one subject occurs in these constructions: (37) (38)

b-eda g-uuya -be -ni C-come P-see 3h 3i 'S/he came to see it.' ye g-uuya -be -ni C/go P-see 3h 3i 'S/he went to see it.'

236

Velma B. Pickett

b) following the verb 'be able' in future form: (39)

z-anda g-uuya -be -ni i'zi9 F-able P-see 3h 3i tomorrow 'S/he'll be able to see it tomorrow.'

c) following the verbs for 'want', 'like', or 'need',5 and in purpose clauses: (40) (41) (42) (43)

gu-ka latji -be g-uuya -be -ni C-happen liver 3h P-see 3h 3i 'S/he wanted to see it.' r-yuu la?]i -be g-uuya -be -ni H-enter liver 3h P-see 3h 3i 'S/he likes to see it.' na-kiine g-uuya -be -ni S-need P-see 3h 3i 'S/he needs to see it.' b-eda-ne -be -ni ti g-uuya -lu -ni C-come-with 3h 3i so = that P-see 2s 3i 'S/he brought it for you to see.'

d) following the adverbs ka9ru 'not yet' and kasi 'as soon as', 'when', whether in past or present setting: (44) (45)

(46)

ka9ru g-uuya -be -ni not = yet P-see 3h 3i 'S/he hasn't seen it yet.' ma ka-winni stusu gu'bija kaPru g-yasa Linu already PR-appear rays sun not = yet P-get = up Lino 'The sun's rays were already shining, and Linu hadn't gotten up yet.' kasi g-uuya -me z-eeda tuuza ra liji -me when P-see 3a PR-come someone LOC home 3a maka zi u-kaaci -me gendaro sti -me already PR/go ThV-C/hidden 3a food of 3a 'When it (animal) would see someone coming to its house, it would quickly go hide its food.'

e) to express uncertain future action, following zandaka 'maybe': (47)

zandaka gi-naaze ka -be laa -0 maybe P-grab PL 3h X 3 'Maybe they would/will/might grab him.'

Aspect in Isthmus Zapotee

237

f) a mild form of command, negative commands, and hortatory: (48) (49) (50)

g-eeda -lu i'zP P-come 2s tomorrow '(Please) come tomorrow.' kadi g-uuya -lu -ni not P-see 2s 3i 'Don't look at it.' gu-da? gi-do? -no C-come P-eat lp 'Come, let's eat.'

g) in plural imperatives, following the plural imperative proclitic la: (51) (52)

la g-uuya9 PLimp P-see 'Look!' la g-yuu PLimp P-enter 'Come in!'

h) A very few verbs take the potential with the negative to indicate a present (negative) state or action, in contrast to the habitual for a longer period situation (the latter is much less frequently used): (53) (54)

ke g-apa not P-have 'S/he doesn't ke g-anna not P-know 'S/he doesn't

-be -ni 3h 3i have it.' -be -ni 3h 3i know it.'

8. Stative na-. The stative prefix occurs in the same position as the active prefixes, but is more limited in its distribution. It occurs with about half of the verb stems, with adjectives which have no other verbal forms, and also occurs with a few nouns to give a stative predicate: (55 a) b)

na-guundu ka gye9 ka S-wilt PL flower DEM 'Those flowers are wilted.' nageenda bi-guundu ka gye? ka rapidly C-wilt PL flower DEM 'Those flowers wilted very quickly.'

238

Velma B. Pickett

(56) (57)

na-bata -ni S-wide 3i 'It is wide.' kadi na-bene ra'ri? not S-mud here 'It's not muddy here.'

In the proper context, the na- prefix may also indicate a state in past or future time: (58)

(59) (60)

(61) (62) (63) (64)

orake b-i-ndaa ja-zawi bya?ki ma then C-Th-CA/escape that-bluejay crow already na-zi'na rini n-uu lu S-red blood S-be face 'Then that bluejay let go of the crow, whose face was already very red.' ma za'ji na-beza -me stubi -me already long S-live 3a alone 3a 'It (animal) had lived alone a long time.' ne wa-sinni, pa na-kiine -be g-uni -be jiina ri-ji if S-need 3h P-do 3h work at-day and RP-night z-uni -be -ni F-do 3h 3i 'If s/he needs to work day and night, s/he'll do it.' neege na-wiini ru -be wiji ma na-soo -be yesterday S-little still 3h future already S-tall 3h 'Only yesterday he was still little; before long he'll be tall.' ti sya'do9 ro9 b-yasa -me na-yece n-uu -me one morning big C-arose 3a S-happy S-be 3a 'One morning very early it (animal) got up feeling very happy.' b-iiya si -be n-anda -ni ra'ka la, bi-jibi -be C-see just 3h S-hang 3i there , C-fear 3h 'As soon as s/he saw it hanging there, s/he was afraid.' na-yece b-yaana ka -me S-happy C-stay PL 3a '(So then) They were happy.'

9. Special aspects. The verbs for 'come' and 'go', and their respective auxiliaries, (see 10) have a special progressive form for singular reference, the prefix z- accompanied by tone changes in the stem. For 'come' it implies that the person is coming within sight or hearing. Note the

Aspect in Isthmus Zapotee

239

comparison of the tone pair with future (low tone is unmarked, high tone marked with acute accent, rising tone with a wedge). (65) (66)

z-eeda -be i'zP F-come 3h tomorrow 'S/he will come tomorrow.' ma z-eeda -be already PR-come 3h 'S/he's already coming.'

The time of the action is determined by the context. In the following example it is in past time: (67)

b-iiya C-see lade beside 'When bush.'

-me z-eeda cupa guuze ra'ke b-yuu -me 3a PR-come two hunter there C-enter 3a ti gPsP a bush it (animal) saw two hunters coming, it hid beside a

For 'go', this use of the progressive refers to having left the place of locutionary action and not yet returned. Note the contrast with completive, which refers to a round trip accomplished; (68) (69)

ma z-e -be already PR-go 3h 'S/he's already gone (and isn't back yet).' ma gu-ye -be already C-go 3h 'S/he already went/has already gone (and has returned).'

The normal progressive form ka- is used with the verbs 'come' and 'go' only with plural subjects, with a distributive interpretation: (70)

kay-eeda ka -0 de i'ra ladu PR-come PL 3 from all side 'They are coming from everywhere.'

10. Auxiliary verbs. There are three auxiliary verbs which precede the main verb. The most commonly used indicates movement in order to perform an action. These auxiliaries are followed by the main verb in potential aspect.6 10.1 Movement away from speaker7 has forms related to the full verb 'go'. It occurs in six aspects, as seen in the following examples:

240

Velma B.

Pickett

(71)

Habitual:

%UUya £ ? H/go P-see 3h 3i 'S/he goes (regularly) to see it.'

(72)

(72)

g UUya Completive- y e ' ~be completive. c / g Q p _ s e e 3 h 3 i 'S/he went to see it (and has returned).' _ . zi g-uuya -be -ni Progressive: , „ -, & PR/go P-see 3h 3i 'S/he has gone to see it (and hasn't returned).'

(74)

(Future) incompletive:

(75)

'S/he will go to see it.' ni g-uuya -be -ni TTnrp „,. Unreal. u / g Q p . s e e 3 h 3 i

Zl

^ uuya be ni F/go P-see 3h 3i

'S/he was going to go see it (implied didn't go).' (76)

PotentialPotential.

p"/ g Q

8 UUya ' see p_

be 3' h 3i

'S/he is going to see it.' 10.2 Movement toward speaker is expressed by the full verb 'come' except for loss of the laryngeal features of phonological phrase head. This auxiliary also occurs in six aspects, with the following forms: Habitual:

reda, as in reda g-uuya -be -ni 'S/he comes to see it.' Completive: beda Progressive: zeda (Future) incompletive: zeda Unreal: neda Potential: geda 10.3. A third auxiliary verb indicates going around performing the action, without indication of direction related to the speaker. This verb occurs only in the progressive aspect and is therefore always ka-na; its temporal interpretation may be past, present or future, depending on the context. It is followed by the stem of the main verb without an aspect prefix. The main verbs which occur with this auxiliary are limited semantically. (77)

ka-na u-too -be byaji PR-go = around ThV-CA/sold 3h plum 'S/he goes around selling plums.'

Aspect in Isthmus Zapotee

(78)

241

ka-na cesa -be PR-go = around jump 3h 'S/he goes around jumping.' b-yasa -me zi g-uuya -me pa j-a'ndi C-arose 3a PR/go P-see 3a if true ka-na za tuuza ra'ke PR-go = around walk someone there 'It (animal) got up and went to see if someone was walking around there.'

(79)

The other two auxiliaries may also be used with this special form of the verb, limited to a small lexical list, to indicate movement while accomplishing the action rather than with purpose of action: (80)

ri yuuna -be H/go cry 3h 'S/he goes around crying.' zeda yuuna -be PR/come cry 3h 'S/he's coming along crying.'

(81)

11. Relation of aspect to time. As indicated throughout the paper, IZ verbal prefixes are primarily aspectual rather than tense marking. Table 1 presents the relationship of these prefixes to the time of the event. Only two are limited to one time orientation. Table 1.

Aspect-time relations

Prefix rgbkawaznikna-

Past H C PR RP F U P S

X

Present X

X X

Future

X X

X

X X X X

X

X

X

X

X

X

12. Conclusions. This article has been written with three purposes in mind: to give more information on the IZ aspects than has been available in previous presentations; to clarify some misunderstanding as to Zapotec

242

Velma B. Pickett

as a family of languages rather than a single language; to present the more fully developed IZ system as a basis for comparison with other Zapotec systems. With sufficient similar information from the remaining languages, historical developments will be interesting to reconstruct.

Notes 1. I am especially grateful to Stephen Marlett for his comments on this paper, in light of his work of an overall nature on other Zapotec languages. 2. The segmental phonemes of IZ are given in Pickett 1967 and in Marlett and Pickett, 1987. Tone is contrastive, but is indicated here only in cases where the contrasts are important to the material being presented. Stress most commonly falls on the first syllable of the root. It is marked here only in the exceptional cases. The following glosses are used for the interlinear morpheme identification in addition to those for the aspect prefixes given in the introduction: CA Causative DEM Demonstrative (enclitics) Location marker (proclitic) LOC PL Plural (clitic) PLimp Plural imperative POS Possession marker prefix Question marker (enclitic) Q ThV Theme vowel (stem formative) lp, 2p First, second person plural Is, 2s First, second person singular 3h Third person human Third person animal 3a 3i Third person inanimate 3 Third person, identification understood from the context * Not glossable X Stem of complex pronouns 3. The personal pronoun enclitics are written as separate words preceded by a hyphen to distinguish them from independent pronouns. Other clitics are written separately without a hyphen. The following are nonpersonal clitics included here: Proclitics ka ta ra

Enclitics

plural (PL) si 'just, only' 'Mr.' la question marker (Q) la subordinator(,) (note 3) locative marker (LOC) ru 'still' ri la plural imperative close demonstrative (DEM) (PLimp) ka farther demonstrative (DEM) za 'below/at foot of auxiliary verbs ke out-of-sight demonstrative (DEM) 4. The morpheme la, glossed with a comma, optionally follows preposed dependent clauses and topicalized phrases.

Aspect in Isthmus Zapotee

243

5. Unless the verb is in the unreal aspect, as in ke n-aka la?Ji -be n-uuya -be -ni not-happen liver 3h U-see 3h 3i 'S/he didn't want to see it.' 6. This type of construction in another Zapotec language is treated in Speck and Pickett 1976, followed by a comparison with IZ in Pickett 1976. 7. Possible terms for these constructions and those of 10.2 (used for Miwok by Freeland 1951 for a different type of construction, but similar semantically) might be andative and venitive respectively. The constructions of 10.3 might then be labelled ambulative.

References Bybee, Joan L. 1985 Morphology. Amsterdam: John Benjamins. Comrie, Bernard 1976 Aspect (Cambridge textbooks in linguistics 2). Cambridge: University Press. Freeland, L. S. 1951 Language of the Sierra Miwok. IJAL Memoir 6 (Indiana University Publications in Anthropology and Linguistics). Hockett, Charles F. 1958 A course in modern linguistics. New York: Macmillan. Marlett, Stephen A.-Velma B. Pickett 1987 The syllable structure and aspect morphology of Isthmus Zapotec. International Journal of American Linguistics 42: 398 —422. Nida, Eugene A. 1949 Morphology. Ann Arbor: University of Michigan Press. Pickett, Velma B. 1967 Isthmus Zapotec. Handbook of Middle American Indians 5: Linguistics. Austin: University of Texas, pp. 291—310. 1976 Further comments on Zapotec motion verbs. International Journal of American Linguistics 42.162 — 64. Pickett, Velma B. et al. 1965 Vocabulario zapoteco del Istmo. Mexico City: Summer Institute of Linguistics. Speck, Charles H.-Velma B. Pickett 1976 Some properties of the Texmelucan Zapotec verbs go, come, and arrive. International Journal of American Linguistics 42.58—64.

A later view of Gitksan syntax1 Bruce Rigsby

Abstract The paper emends and revises "Nass-Gitksan: an analytic ergative syntax" (Rigsby 1975). It argues that Gitksan is a language separate and distinct from Nisgha (Nass), and it provides general descriptions and discussion of some basic syntactic constructions of Gitksan, including independent order, right-extraposed and focused sentence constructions, incorporated verb themes, antipassives and passives.

1. In my 1975 paper, 'Nass-Gitksan: an Analytic Ergative Syntax', I described the basic surface syntax of independent and dependent clauses in the language and argued that it was ergative at the levels of both surface and deep syntactic structure and that its underlying basic constituent order was Verb — Ergative — Absolutive (called Nominative there). Rood (1977) made some thoughtful criticisms of the paper, particularly its placement of the ergative noun phrase within the verb phrase in deep structure. The present paper is intended to emend and revise the earlier one, but it retains the objective of providing a brief general description of some basic syntactic constructions of the language. 2. Some sociolinguistic matters merit our early attention so that we may identify the language that is the main subject of our discussion because I cite Gitksan examples exclusively in the two papers. In Rigsby (1975:346), I said the Nass-Gitksan was 'the same language whose Nass dialect was described by Franz Boas' in his Tsimshian sketch in Part 1 of the Handbook of American Indian Languages (Boas 1911 b), and in so doing, I followed the identification made in an earlier paper (Rigsby 1970:212, which cited both Gitksan and Nisgha2 examples): ... Nass-Gitksan, as a language name, subsumes the dialects which are spoken today in a number of villages located in the Nass and Skeena River valleys. These dialects appear to fall into two major groupings; the Nass dialects of the Nass valley and the Gitksan dialects of the Skeena valley.

246

Bruce Rigsby

At that time I followed conventional Americanist practice by conceptualizing the relationship between dialect and language in terms of mutual intelligibility, viz., dialects are varieties of a language that either are mutually intelligible or are connected by mutually intelligible varieties (Hockett 1958:321 - 330; Gleason 1961:441 -442). It is true that competent speakers of Gitksan and Nisgha find each others' speech to be mutually intelligible for the most part upon first contact or hearing of it, although some words may be unfamiliar. Today, however, I would no longer speak of the Gitksan dialect(s) of the Nass-Gitksan language, but instead I accept and use the Gitksan people's own phrasing in English that speaks of the Gitksan language, not the Gitksan dialect. Theoretical, methodological, and practical considerations have led me to this position. The theoretical objection to the mutual intelligibility approach is that it assumes the intelligibility of two language varieties is primarily a function of their structural similarity. The closer two language varieties are in phonology, grammar, and vocabulary, the more they are intelligible to each other. Yet the literature on attempts to operationalize and measure intelligibility shows the difficulties of distinguishing intelligibility due to structural similarity from that due to normal language learning. It is important to keep in mind that language varieties are intelligible to people, to speakers who display a range of language skills and knowledge; language varieties or codes in themselves are not intelligible to each other. And as Wolff (1959, 1967) pointed out, matters of social and cultural evaluation may lead speakers to deny or affirm that one or another language variety is intelligible to them. I believe that the interesting and important matter of intelligibility should be kept separate from the question of whether language varieties are to be considered dialects of the same language or not. (See Rigsby and Sutton 1980 — 82:17 —18 for a critical overview of the literature on intelligibility). There is a second usage, common in Europe and among Englishspeaking laypeople, that distinguishes languages from dialects in terms of their status and functions. In this view, essentially, only standard language varieties qualify as full languages, because they have official recognized status and a full set of speech functions, including written ones. By this criterion, neither Gitksan nor Nisgha can be considered to be full languages, although both are in the early stages of developing a distinctive standard language variety. Such functional definitions of language and dialect often include an implicit negative, denigrating connotation in that a full language is considered to be better than a mere dialect. It is not surprising that Gitksan people insist on calling their

A later view of Gitksan syntax

247

language variety a language and not a dialect, as anthropologists and linguists sometimes tell them they should. A third view is that of modern sociolinguistics and linguistic anthropology (e.g. Hymes 1968; Rigsby and Sutton 1980-82; Romaine 1982), which accords primacy to community perceptions and standards. Thus, Danish, Norwegian, and Swedish count as separate languages because of the autonomous status of the Danish, Norwegian, and Swedish standard languages, even though their speakers find one another's speech (and writing) mutually intelligible to considerable degrees (Haugen 1966, 1976:61 —2). Similarly, the mutually intelligible Germanic varieties that are spoken on either side of the Netherlands / West German border (Moulton 1985:404) are assigned differentially to the Dutch and German languages because of their heteronomous status with respect to the standard Dutch and German language varieties. That is, their speakers regard them respectively as kinds of Dutch or German. If we adopt this perspective, then it is quite in order to concur with native speakers that Gitksan and Nisgha are separate languages. Certainly the Gitksan and the Nisgha villages belong to separate social communities with respect to ethnic (or tribal) identity and cultural norms, including norms of grammar, vocabulary and accent. Objectively, the Gitksan community can be defined by such sociological criteria as marriage patterns (Kasakoff 1970, 1974,1976) and participation in a common system of ceremonial exchange (Adams 1969, 1973, 1974). Subjectively, it is defined by the people's own belief and public statements that they are Gitksan and speak their own language — they are not Nisgha nor are they Tsimshian, who each have their own distinctive native language. (This section is adapted from a longer one in Rigsby 1986:6 — 12). 3. In section 1 of Rigsby (1975:346 — 348), I discussed some features of grammatical type, then surveyed the basic syntax of independent order transitive and intransitive sentences. Both matters require attention. Grammatical type In the morphological complexity of its words, Gitksan is analytic to synthetic. It displays more synthesis than described in the title of Rigsby (1975) and where I characterized it (p 346) more precisely as 'analytic to mildly synthetic, much like modern French or English'. Tarpent (1983:123) describes Nisgha as like the other members of the Tsimshianic language family in having 'a mixed analytic-synthetic morphological structure', and she suggests (p. c.) that German with its many com-

248

Bruce Rigsby

pounded nomináis and verb phrases with separable and inseparable particles provides the more apt comparison. I agree with her. Gitksan also has a rich derivational morphology, and while words of one and two morphemes abound in discourse and narrative, words of five or six morphemes are not infrequent. An example is seen in: (1)

haguxwsgalt'amdinsxw /h9-kw3xws-tqal-t'am-t9n-sxw/ INST-self-up. against-mark-TRN-ANTI 'camera'

In its technique of synthesis, Gitksan is for the most part agglutinative, rather than inflective or symbolic. However, it does make use of a range of grammatical processes, which include affixation (including also reduplication and clisis), ablaut, compounding, suppletion, selection, and order. Of these, order is no doubt exploited to the greatest extent. The order or sequencing of constituents within words, phrases, and clauses is fairly rigid. Among its grammatical categories (in the classical sense of Boas 1911 a), the distinction between common and proper nouns and the marking of number in nouns and verbs merit brief mention. Proper nouns include personal names, (most) senior kinterms, personal pronouns, and demonstrative pronouns; all these share reference to a definite or specific individual or set of individuals.3 The singular proper noun proclitic and prefix is /t = /, and it occurs in a restricted set of syntactic environments, while its plural counterpart /tip = I displays no such restrictions. Examples of singular proper noun number-marking are seen in (2), (3), and (5). As well, proper noun phrases display their class membership by selecting an / = s/ connective enclitic, while common nouns select /=i/ — see examples (2) —(5). Number — and specifically, plural number — is the Gitksan grammatical category par excellence. Common noun and verb theme constructions fall into six major plural-marking classes, which Rigsby (1986:89 — 116) labels as invariable, regular, irregular, pleonastic, suppletive, and passivized plurals, and there are many more subclasses that can be identified within these.4 Independent order constructions The exemplary sentences (1) —(13) in section 1 of Rigsby (1975:347 — 348) are correct, but they provide an incomplete range of constructions because they do not include any non-pronominal proper noun phrase constituents. And so, the proper noun connective enclitic and the proper

A later view of Gitksan syntax

249

noun number proclitics and prefixes went unnoticed and unmentioned. In order to extend the analysis to them, I now move to a somewhat different consideration of the major syntactic constructions, transitive and intransitive, which touches also upon matters of basic syntactic relations (or functions), case-notions, focusing, and voice. Taken together, these all provide the formal means for the creative text-building power of the language. Gitksan texts, whether they be of conversations, traditional narratives, or political oratory, display formal structure and coherence that is achieved in large part through these grammatical means. 5 Transitive active (or ergative) construction (2)

(3)

(4)

Xhlii haksis Bruce t Barbara /xiihaks-a = s Bruce t = Barbara all.over scold-TRN = CNN Bruce PSG = Barbara gaxxw. kaxxw/ last.night 'Bruce bawled out Barbara last night.' Xhlii haksihl gat t Barbara gaxxw. /xiihaks-a = i kat t = Barbara kaxww/ all.over scold-TRN = CNN man PSG = Barbara last.night 'The man bawled out Barbara last night.' Xhlii haksis Bruce-hl hanak' /xiihaks-a = s Bruce = i hanaq all.over scold-TRN = CNN Bruce = CNN woman gaxxw. kaxxw/ last.night 'Bruce bawled out the woman last night.'

Intransitive construction (5)

(6)

Limx t Bruce gaxxw. /limx t = Bruce kaxww/ sing PSG = Bruce last.night 'Bruce sang last night.' Limxhl gat gaxxw. /limx=4 kat kaxxw/ sing = CNN man last.night 'The man sang last night.'

250

Bruce Rigsby

Sentences (2) —(6) all have a fixed constituent order that is verb-initial. The transitive verbs are marked by a stem-forming suffix /-(y)a-/> (which occurs before all agents except third person plural pronoun ones). Nonpronominal proper noun transitive subject (agent or A) phrases are marked by an / = s/ connective that is enclitic to the preceding word, while common noun agent phrases are marked by a / = !/ connective enclitic. (2) —(6) also display a surface ergative / absolutive syntax, in that transitive object (or O) phrases are treated the same as intransitive subject (or S) ones, viz., neither are marked by a connective enclitic, and in the case of singular proper noun absolutive phrases, the singular /t = / proclitic appears. Right-extraposed constructions Rigsby (1975:348) also commented on independent order transitive sentences with right-extraposed non-pronominal agents and personal pronominal absolutives, as seen in (7), but failed to point out that these are found only with first or second person absolutives (which, of course, are always pronominal). I analyzed them then as deriving from underlying Verb — Agent — Object configurations that underwent obligatory rightextraposition of the agent, leaving behind a trace in the third person pronoun suffixed to the verb. I hold to that analysis — because I believe it correct to represent all transitive configurations underlyingly as Verb — Agent — Object in their order — but note how that the operation of the rule is more complex. Younger speakers commonly produce sentences that display the same surface constituent order as the non-extraposed underlying one, as seen in (8). (7)

(8)

Hlimooyit 'nuu'm t Mary. /iamo-ys-t nu-m t = Mary/ help-TRN-3SG ABS-1 PL PSG = Mary 'Mary helped us.' Hlimooyis Mary 'nuu'm. /iamo-y3 = s Mary nu-m/ help-TRN = CNN Mary ABS-1 PL 'Mary helped us.'

I have discussed the two constructions — the right-extraposed one and the simpler one — with some thoughtful middle-aged and older speakers. They say that the former is 'good' or 'proper' Gitksan, but they recognize that the latter occurs, and sometimes they use it themselves. There

A later view of Gitksan syntax

251

appears to be no difference of propositional meaning between the two constructions, and I would expect the (derivationally) simpler, non-extraposed construction to remain frequent or become more so in the speech of younger bilingual speakers. My friends also tell me that the nonextraposed construction seems better when there is an oblique phrase or complement present, as in (9). (9)

Xhlii haksis Barbara 'nii'y /xiihaks-9 = s Barbara ni-y all.over scold-TRN = CNN Barbara ABS-1SG he'y. hi-y/ saying-1SG 'Barbara bawled me out when I mentioned it.'

wil wil CMP

The situation evidently differs in Nisgha (see Tarpent 1981), where rightextraposition is obligatory, at least if there is no oblique phrase or complement present. Jelinek (1986) correctly interprets such constructions in Nisgha as evidence for an argument hierarchy where first and second person rank higher than non-pronominal arguments, but she proposes their underlying constituent order is Verb — Noun Phrase — 1 or 2 Pronoun and that the third person pronoun suffix on the verb is an agreement marker, not the main agentive argument. As she remarks (p 4), even if the extraposed agentive phrase is regarded as an adjunct, as in my analysis, the hierarchy still holds. Focused constructions Any major phrasal or clausal constituent 6 of an underlying independent sentence may be focused or topicalized by moving it to the front of the sentence. The propositional content of a focused sentence is the same as its non-focused independent order counterpart; i. e., they are propositionally synonymous and thus they share a common underlying structure representation, but they are not pragmatically or indexically equivalent. Focusing is one means by which the textual structuring of discourse and narrative is accomplished. However, focused constructions are not simply stylistic alternatives that speakers are free to make use of or not. In the case of underlying sentences containing a question word constituent, it must be focused and fronted in a way that is reminiscent of English WHquestion movement. Similarly, demonstrative S-phrases and O-phrases

252

Bruce Rigsby

seem to require focusing. And focused sentences (or their elliptical fragments) are the usual appropriate answers to focused question word interrogatives. Beyond these cases, though, focusing provides the means for speakers to maintain and shift topics over sentences. Focused agent question (10)

(11)

(12)

Naa ant xhlii haks Barbara gaxxw? /na- 9 an = t x4ihaks = s Barbara kaxxw/ who AREL = 3 all.over scold = CNN Barbara last.night 'Who (is the one who) bawled out Barbara last night?' Bruce ant xhlii haks Barbara gaxxw. /Bruce 9 an = t xiihaks = s Barbara kaxxw/ Bruce AREL-3 all.over scold = CNN Barbara last.night 'Bruce is the one who bawled out Barbara last night.' Bruce ant xhlii haks-t. /Bruce ?an = t xiihaks-t/ Bruce AREL = 3 all.over scold-3SG 'Bruce is the one who bawled her out.'

4. In section 2 of Rigsby (1975:348-350), I argued that the ergative / absolutive syntax existed also at the level of underlying structure in Gitksan, and I noted some evidence that transitive objects and intransitive subjects, i. e. absolutives (called 'nominatives' there), could not be differentiated at any level of syntactic representation, whether surface, intermediate, or underlying. I observed that verbal number agreement operated on an ergative / absolutive basis, but that this was not good evidence because number agreement operated after passivization. Then I remarked that imperative constructions undergo ergative-deletion, rather than subject-deletion as in English. That is true, but it too does not constitute good evidence because the rule could be a late one, similar to number agreement. And finally, I pointed out the relevance of incorporated verb themes and one deverbative nominal construction. Focusing provides the strongest evidence that the absolutive syntactic relation is not a unitary undifferentiated one in underlying structure. Rather, the absolutive represents the surface neutralization of different underlying syntactic relations, as can be clearly seen in the separate treatments of focused transitive object phrases and intransitive subject phrases.

A later view of Gitksan syntax

253

Focused object question (13) Naahl xhlii haksis Bruce gaxxw? /na- = i xiihaks-a = s Bruce kaxx w / who = CNN all.over scold-TRN = CNN Bruce last.night 'Who(m) did Bruce bawl out last night?' Focused object construction (14) Barbara-hl xhlii haksis Bruce /Barbara = i xiihaks-3 = s Bruce Barbara = CNN all.over scold-TRN = CNN Bruce gaxxw. kaxww / last.night 'Barbara is the one Bruce bawled out last night.' (15) Barbara-hl xhlii haksit. /Barbara = 4 xiihaks-a-t/ Barbara = CNN all.over scold-TRN-3SG 'Barbara is the one he bawled out.' Focused subject question (16) Naahl limit gaxxw? /na- = i limx-a-t kaxxw/ who = CNN sing-SREL last.night 'Who (was it who) sang last night?' Focused subject construction (17) Bruce-hl limit. /Bruce = i limx-at/ Bruce = CNN sing-SREL 'Bruce (is the one who) sang.' Incorporated verb themes Earlier (Rigsby 1975:349 — 350), I derived incorporated noun — verb constructions, such as guxwsmax 'shoot bears' and ishun 'stink like fish', by an obligatory transformation triggered by the absence of complex determiner marking on underlying absolutive nouns. That analysis is surely wrong, and such constructions should be derived lexically in view of such facts as that only a minority of verbs (only a few of them simple intransitives) occur in the complex derived incorporated forms. The same analysis should hold for Nisgha, but it

254

Bruce Rigsby

makes more use of incorporated forms, which occur in more construction types (Tarpent 1982: 81-83). Antipassive constructions I also earlier (Rigsby 1975:350) derived certain deverbative nomináis, such as jagwasxw 'animal' and bahasxw 'wind', transformationally from underlying verbal predications. In fact, these nomináis have the same form as regular antipassive verb themes, a point that Boas (1911 b: 344— 345) commented on for Nisgha. I now regard antipassive themes and their deverbative nomináis as lexically derived. Antipassive verb themes are built upon roots that generally occur also in transitive themes, but not all transitive active themes have antipassive counterparts. Sometimes the antipassive form has a more specialized meaning, and some antipassive themes contain roots that are no longer used in transitive active constructions. Examples of the antipassive construction are seen in (20)-(21). Ergative construction (18) Yukwt gibas Bruce /yukw = t kapa = s Bruce PROG = 3 wait = CNN Bruce 'Bruce is waiting for Barbara.' (19) Ts'o'odis Bruce-hl /cu^-t3 = s Bruce=i skin-TRN = CNN Bruce = CNN 'Bruce skinned the bird.'

t Barbara. t —Barbara/ PSG = Barbara ts'uuts'. cu-c/ bird

Antipassive construction (20) Yukwt gibee'esxws Bruce as /yukw = t kap-e-9-sxw = s Bruce *>a = s PROG = 3 wait-SFX-ANTI = CNN Bruce OBL = CNN Barbara. Barbara/ Barbara 'Bruce is waiting for Barbara.' (21) Ts'o'osxw t Bruce. /cu">-sxw t = Bruce/ skin-ANTI PSG = Bruce 'Bruce did some skinning.'

A later view of Gitksan syntax

255

In the case of antipassive themes that have regular transitive active counterparts, the semantic distinction appears to be one of undirected action as contrasted with directed goal-oriented action (see Comrie 1981:225 on the same distinction between effect-orientated from subjectorientated action in Georgian). Passive constructions Gitksan also has a passive construction. With an exception or two, roots occurring in passive themes also are found in antipassive ones, but there are many more antipassive themes that lack passive partners. Some passive themes too have specialized meanings. There are no deverbative passive nominals, as there are deverbative antipassive ones. And passive themes govern only a subject / absolutive phrase, i. e. oblique agents are not permitted. (22) is an example of a passive construction, and (23) — (25) exemplify a family of lexically related transitive, antipassive, and passive verb themes.7 Passive construction (22) Tso'osxwhl ts'uuts'. /c u 9 -tx w =4 cu-c/ skin-PASS-CNN bird 'The bird got skinned.' (23) gubi/kwap-a-/ eat-TRN 'eat (something)' (vt sg) (24) gubasxw /kwap-asxw/ eat-ANTI 'eat berries while picking' (vi sg) (25) gupxw /kwap-xw/ eat-PASS 'be edible' (vi sg) 5. In section 3 of Rigsby (1975), I treated complementation by discussing antipassive-type transformations and deriving surface possessive sentences from underlying structures with a dummy verb 'possess'. I pointed out the surface parallelism between the form of sentential complements

256

Bruce Rigsby

and that of possessed nouns, which I thought was not accidental, but I offered no explanation for that. In retrospect, my discussion of antipassive-type constructions was misdirected, and I would now analyze the possessive constructions differently.8 Nonetheless, my exemplary sentences were correct, as was also Figure 7, which showed the relationships of the independent order and dependent order (subordinate) pronoun schemata to the three series of personal pronouns. However, Table 1, which presented the three pronoun series, contained some transcriptional errors and overlooked small differences between certain eastern and western Gitksan personal pronouns. Rigsby (1986:412—414) treats these pronoun series with their several alternants and variants. 6. The final section (4) of Rigsby (1975:353-354) argued that the surface verb-initial independent order schema of phrasal constituents is the same as the underlying one, using evidence from particular dependent order configurations that display an Ergative — Verb — Ergative — Absolutive ordering. I still hold to this analysis, which differs from Belvin (1984), working in the government and binding framework, who argues that Nisgha is syntactically accusative and underlyingly Subject — Verb — Object, having a rule that moves the verb into initial position. Jelinek (1986) too disagrees with Belvin, noting that he does not treat dependent order constructions, and she analyzes Nisgha as being a predicate-initial non-configurational language whose nominal and pronominal arguments are ordered positionally by an argument hierarchy. I close the paper now on matters that remain uncertain to me: whether Gitksan is a configurational language or whether it has "flat syntax." I assume that the underlying order of basic constituents in Gitksan (and Nisgha) is Verb — Agent — Object or Verb — Subject, and in the past, I have also assumed that there is a Verb Phrase constituent that includes an internal agent argument. These assumptions were related in that they permitted argument roles to be uniquely specified by constituent structure. I am no longer sure that one can justify the existence of such a verb phrase constituent, and I do not know what kind of structural data might provide evidence for or against it, Jelinek's (1986:8 — 10) discussion notwithstanding. As well, I am more inclined now to regard argument roles (basic semantic-syntactic relations or functions) as primes themselves, not defined by constituent structure. Jelinek's analysis, in which the positional arrangement of arguments follows from the argument hierarchy, is appealing, but in its details it needs to attend to the nonextraposed sentences of Gitksan.

A later view of Gitksan syntax

257

Notes 1. It is an honor to have this paper included in a memorial volume for Stanley Newman. We were friends and colleagues at the University of New Mexico, 1966—1975, and with Harry Basehart, we worked together in editing the Southwestern Journal of Anthropology, later The Journal of Anthropological Research, 1970—1975. Funding for my research on Gitksan and Nisgha since 1966 has come at various times from the National Museum of Canada, the University of New Mexico, the Phillips Fund of the American Philosophical Society, Harvard University, the 'Ksan Association, the Center for Applied Linguistics, the British Columbia Provincial Museum, the University of Queensland, and the Gitksan-Wet'suwet'en Tribal Council. I appreciate their support, but I am alone responsible for the deficiencies of this paper. Most examples are in four lines — the first is in the Gitksan practical alphabet, as described in fn. 2 of Rigsby (1975:346), except that the italicized uvular consonants there are here underlined; the second line gives the underlying phonological representation, including word, clitic, and morpheme boundaries; the third line provides interlinear morphemic glosses with grammatical morphemes in capitals; and the fourth is an English translation. I owe special thanks to Marie-Lucie Tarpent, who has shared her insights into Nisgha with me freely and rescued me from a number of dead-end positions. Among her original ideas that I have appropriated are her discovery of the /-(y)a-/ transitive suffix, her identification of /t = / as the singular proper noun marker (rather than as a connective enclitic), and her elucidation of the focusing system. 2. The Nisgha people, who live in the Nass Valley of northern British Columbia, call their language Nisgha in English. Nisgha is the spelling preferred and used by the Nisgha Tribal Council and School District 92 (Nisgha). It derives from the Nisghas' own ethnonym /nasqa?/ or [nisGx*7]. Niska and Nishga are spellings that are also found in the anthropological and linguistic literature. 3. Boas 1911b: 297 speaks of 'special human individuals and common nouns'; Tarpent 1981, 1982 prefers the labels 'determinate' and 'indeterminate' or 'non-determinate'; while I follow Schachter 1985:8—9 in using 'proper' and 'common' for the two classes. 4. Tarpent 1983, who does much internal reconstruction, recognizes a different set of plural classes and subclasses in Nisgha. 5. Many of the exemplary sentences make reference to my own family situation. It is not the case that I customarily abuse my wife verbally, but rather that one should be circumspect in selecting personal names when making up exemplary Gitksan sentences because sentences may have pragmatic force that goes far beyond their propositional content. All the examples have been checked with native speakers in workshop and other settings. 6. Only examples of focused agents, objects, and subjects are given here, but datives, time adverbs, place adverbs, etc. may also be focused. See Rigsby (1986:281—306) for a fuller account. 7. There are also other complex verb thematic constructions having different transitivity and voice values, e. g. causatives, detransitives, jussives, medials, etc. Transitivity and voice are semantically interrelated generic categories concerned with the thematic roles of speech event and narrated event participants to actions, states, processes, etc. Rigsby (1986:317 — 361) surveys verb theme derivation more fully.

258

Bruce Rigsby

8. Tarpent (1986) surveys a range of Nisgha possessive constructions; they differ somewhat from those of Gitksan. As well, Rood (1977:222—3) was correct in noting that possessed nouns both are and have the form of subordinated constructions in Gitksan.

References Adams, John W. 1969 The politics of feasting among the Gitksan. Ph.D. Thesis. Department of Social Relations, Harvard University. 1973 The Gitksan potlatch: population flux, resource ownership and reciprocity. Toronto and Montreal: Holt, Rinehart and Winston of Canada, Ltd. 1974 Dialectics and contingency in "The story of Asdiwal": an ethnographic note. In The unconscious in culture. The structuralism of Claude Lévi-Strauss in perspective. Ed. by Ino Rossi. New York: E. R Dutton and Co., Inc., 170—8. Belvin, Robert S. 1984 Nisgha syntax and the ergativity hypothesis. M.A. Thesis. Department of Linguistics, University of British Columbia. Boas, Franz 1911a Introduction. Handbook of North American Indian Languages, Part 1. Bureau of American Ethnology. Bulletin 40, 5—83. 1911b Tsimshian. Handbook of North American Indian Languages, P a r t i . Bureau of American Ethnology. Bulletin 40, 283 - 422. Comrie, Bernard 1981 The languages of the Soviet Union. Cambridge: Cambridge University Press. Gleason, H. A., Jr. 1961 An introduction to descriptive linguistics. New York: Holt, Rinehart and Winston. Haugen, Einar 1966 Semicommunication: the language gap in Scandinavia. Sociological Inquiry 36:280-97. 1976 The Scandinavian languages. London: Faber and Faber. Hockett, Charles F. 1958 A course in modern linguistics. New York: The Macmillan Company. Hymes, Dell 1968 Linguistic problems in defining the concept of 'tribe'. Essays on the problem of tribe, ed. by June Helm, 23—48. Proceedings of the 1967 Annual Spring Meeting. American Ethnological Society. Jelinek, Eloise 1986 The ergativity hypothesis and the argument hierarchy in Nisgha. Working Papers for the 21st International Conference on Salish and Neighboring Languages. 14—16 August 1986. University of Washington, Seattle, 7 — 17. Kasaicoff, Alice B. 1970 Explicit and implicit marriage rules among the Gitksan. Ph.D. Thesis. Department of Social Relations, Harvard University. 1974 Lévi-Strauss' idea of the social unconscious: the problem of elementary and complex structures in Gitksan marriage choice. In The unconscious in culture.

A later view of Gitksan syntax

259

The structuralism of Claude Lévi-Strauss in perspective, ed. by Ino Rossi, 143-69. New York: E. P. Dutton and Co., Inc. 1976 Open and closed marriage systems on the northwest coast. Unpublished paper presented at the Northwest Coast Studies Conference, May 12—16, 1976, Simon Fraser University. Moulton, William G. 1985 Review of Handbuch zur Niederdeutschen Sprach- und Literaturwissenschaft. ed. by Gerhard Cortes und Dieter Móhn. Language in Society 14:3:403 — 7. Rigsby, Bruce 1970 A note on Nass-Gitksan speech-play. International Journal of American Linguistics 36:3:212-15. 1975 Nass-Gitksan: an analytic ergative syntax. International Journal of American Linguistics 41:346 —54. 1986 Gitksan grammar. Unpublished manuscript prepared for the British Columbia Provincial Museum. 442 pp. Rigsby, Bruce—Peter Sutton 1980 — 1982 Speech communities in Aboriginal Australia. Anthropological Forum 5:1:8-23. Rood, David S. 1977 Against artificial tree branches: another look at Nass-Gitksan syntax. International Journal of American Linguistics 43:3:218—24. Romaine, Suzanne 1982 What is a speech community? Sociolinguistic variation in speech communities, ed. by Suzanne Romaine, 13 — 24. London: Edward Arnold. Schachter, Paul 1985 Parts-of-speech systems. Language typology and syntactic description. Volume 1. Clause structure, ed. by Timothy Shopen, 3—61. Cambridge: Cambridge University Press. Tarpent, Marie-Lucie 1981 Major features of Nisgha syntax. Unpublished paper. University of Victoria. 1982 Ergative and accusative: a single representation of grammatical categories based on evidence from Nisgha. Working Papers of the Linguistic Circle of Victoria, pp. 50-106. 1983 Morphophonemics of Nisgha plural formation: a step towards Proto-Tsimshian reconstruction. Studies in Native American Languages II. Kansas Working Papers in Linguistics 8:2:123—214. 1986 Nisgha possessives. Working Papers for the 21st International Conference on Salish and Neighboring Languages. 14—16 August 1986. University of Washington, Seattle. Pp. 1 8 - 3 4 . Wolff, Hans 1959 Intelligibility and inter-ethnic attitudes. Anthropological Linguistics 1:3:34-41. 1967 Language, ethnic identity and social change in southern Nigeria. Anthropological Linguistics 9:1:18—25.

The Kuna verb: a study in the interplay of grammar, discourse, and style Joel Sherzer

Abstract The Kuna verb is characterized by the potential use of many different suffixes. Like Yokuts, as described by Stanley Newman, speakers do not use as many suffixes on each verb as the theoretically potential system would seem to indicate. Rather, the verb suffix system is used to differentiate cultural and individual styles and genres of speaking and chanting.

The purpose of this paper is to present a preliminary analysis and overview of the structure of the Kuna verb. Kuna is an American Indian language spoken by about 30,000 people who live, mainly, on island and mainland communities in San Bias, Panama. Like many, though not all, American Indian languages, Kuna is a polysynthetic language. It is a language in which boundaries between morphology and syntax, word, clause, and sentence are problematical. In this paper I will discuss the relationship between the morphological structure of the Kuna verb and Kuna discourse, the various styles, forms, and genres of speaking. My own starting point for this paper is dual. First there is my interest in the relationship between grammar and discourse and, more generally, in the crucial and creative role of discourse in the language-culture relationship. Kuna discourse speaks quite nicely to this interest. Second there is a very provocative statement by Stanley Newman concerning the Yokuts language of California: There are no more than a hundred suffixes in Yokuts. Although English possesses about the same number, it does not rely primarily upon suffixation; such techniques as word order, compounding, prefixation carry as much, if not more, of the functional load. Yet, in spite of the fact that Yokuts depends almost exclusively upon suffixes, it is remarkably restrained in exploiting the possibilities of suffix combination. Every word in Yokuts, except the uninflected particle, must have at least one suffix, and the great majority of words occurring in the

262

Joel Sherzer

several volumes of text dictated by my informants do not go beyond this minimum requirement. Words containing two suffixes are fairly common; but words with three suffixes are relatively rare, and those with more than three are practically nonexistent in the texts. This is not a matter of mere statistics; it is a manifestation of selective forces in Yokuts that limit the free application of its grammatical resources... An instructive exercise that I indulged in during my Yokuts field work was to construct words having four or five suffixes and to ask the informant for a translation. Although such words complied with the grammatical rules and could be translated by my informant without any difficulty, they seldom failed to provoke his amusement. It was obvious that these words were impossibly heavy and elaborate. To the Yokuts feeling for simplicity they were grammatical monstrosities. (Newman 1940, reprinted in Hymes (ed.) 1964: 373-374.) 1 While not a discourse perspective per se, in the contemporary sense of the term, these remarks by Newman do raise fascinating questions about the relationship between grammar as a statement of resources in a language and grammar as a description of patterns of actual language use. They also express a belief that it is in discourse rather than grammar where a language's spirit or genius, in the Sapirian and Whorfian sense, is expressed. Some of what Newman has to say about Yokuts is quite different from what I would say about Kuna. The concepts expressed by the suffixes are 'abstract in nature.' 'Particularized and concrete meanings are not expressed through the suffixes in Yokuts' (Newman 1964: 373). At the same time what Newman has to say about the relationship between grammar as a resource and actual discourse or texts, in particular the fact that while the grammar seems to predict the possibility of the stringing along of a large number of suffixes while in actual usage this is not found, is quite similar to my experience with Kuna, as is Newman's description of informant responses to his created verbs. The most characteristic feature of Kuna morphology is suffixation. Most words, usually \ orbs, which are the sentence core of Kuna sentence, are formed by adding several suffixes to the stem. The Kuna verbal structure consists of a few prefixes and many suffixes. The grammatical description of the verb thus involves a statement of each of these prefixes and suffixes, their meanings, their grouping into classes, their possibilities of order, and their possible co-occurrences. Kuna phonology and morphology intersect and interact in that most Kuna forms, including suffixes,

The Kuna verb

263

can occur in a long form, characterized by a final vowel, or a short form, in which the vowel is deleted, causing the operation of various morphophonemic rules affecting consonant clusters. With regard to transcription, I use two different levels, (1) the actual spoken or chanted form and (2) the underlying morphological form, with the potentially deletable vowel of the long form of a stem or suffix placed within parentheses and morpheme boundaries indicated by dashes. Thus: (1) parittocurmala 'they didn't feel anything' or (2) paritto(e)-s(a)-sul{i)-mal(a) feel-PAST-NEG-PLURAL2 For the purposes of this discussion, it is useful to group the verbal suffixes into various grammatical/semantic classes.3 Suffixes indicating tense are: s(a) 'past' and o(e) 'future', which is also a modal, with mandative force. Aspect is highly developed and integrated with tense, it also has subcategories within it. Suffixes indicating a temporal perspective or timing, with regard to the event described are: sokkal(i) 'about to', al{i) 'just beginning', t{e) 'at that moment, then', pinn(e) 'just occurred', ti(i) 'ongoing', and ta(e) 'always'. Suffixes indicating movement and direction (sometimes in conjunction with tense) are: natapp{i) 'walking along', na(e) 'go to a location', tak{e) 'come', tanikki 'come', nonikki 'came', tapp(i) 'did on arrival', and app(i) 'went there and came back'. Suffixes indicating position are: ma(i) 'horizontal', kwic(i) 'vertical', si(i) 'sitting', and na(i) 'hanging or perched'. Suffixes indicating number are: mal(a) 'plural' and pukkw(a) 'many'. Suffixes which are distributive in meaning are: mok(a) 'also' and pal(i) 'again'. Suffixes indicating modality are: su(i) 'negative', pi(e) 'want to', tipa 'possible, indirect question', and ye 'optative, emphatic'. le(k[e]) indicates passive voice, kal(a) 'in order to', ku(a) 'when (usually) in the past', le 'when in the future, if, and sokku(a) 'since' mark clause linkage and subordination, sunto (perhaps a combination of two suffixes: sunna 'in truth' and to 'occurred') is a narrative marker. Two suffixes are derivational: t(J) is a nominalizer and al(e) is a gerundive. And there are more. As a guide for the reader, these suffixes can be organized as in Table 1. These verbal suffixes are optional, rather than obligatory, in the BoasSapir sense of this opposition.4 Thus they can be viewed as a rich set of choices that speakers can draw on. Notice that many of these suffixes are quite concrete and specific in their meaning, in contrast to what Newman has to say about Yokuts.

264

Joel Sherzer

Table 1.

Kuna suffixes: grammatical/semantic classes

Tense s(a) o(e)

past future

Aspect Temporal perspective sokkal(i) al(f) tie) pinn(e) ti(i) taie)

about to just beginning (inceptive) at that moment, then just occurred ongoing always (usitative)

Movement and direction natapp(i) naie) tak(e) tanikki nonikki tapp(i) app(i)

walking along go to a location come come came did on arrival went there and came back

Position ma(i) kwic(i) si(i) na(î)

horizontal vertical sitting hanging, perched

Number malia) pukkw(a)

plural many

Distributive mok(a) pal(i)

also again

Modality (in addition to o(e)) sul(f) pi(e) tipa ye

negative want to (desiderative) possible, indirect question optative, emphatic

Voice le(k[e])

passive

The Kuna verb Table 1.

265

Continued

Clause connectors kal(a) ku(a) le sokku(a)

in order to (purposive) when (usually) in the past when in the future, if since

Narrative sunto Derivational

m

al{e)

nominalizer gerundive

There are certain basic rules of order, selection, and co-occurrence that govern how these suffixes are strung along. In the first position after the verb stem occur suffixes that can also themselves stand alone as verbs. These include the four positionals, the temporal perspective suffixes tii 'ongoing' and tae 'always', and many of the movement and direction suffixes, such as natappi, nae, tanikki, and nonikki, and pukkwa 'many'. Leke 'passive' can also occur in the first position, as can kala 'in order to'. Tense and other temporal perspective suffixes occur in second position, as do some of the movement and direction suffixes. The future suffix oe can occur after te and after tappi. Pinne, suli, mala, pali, moka, and pie occur in third position. When the future suffix oe occurs along with pali, mala, or moka, it follows them. Ku, le, and sokkua occur in fourth position. tipa occurs in fifth position. The narrative suffix sunto occurs in sixth position. And the optative, emphatic ye occurs in seventh position. These position classes are set out in Table 2. Given this general statement of the structure of the Kuna verb, one might imagine that there are many verbs which contain seven or even more suffixes. But in actual practice, and this is what reminds me of what Newman had to say about Yokuts, fewer suffixes are used with each verb than would seem theoretically possible. In addition, and this is different from Yokuts, or at least what Newman says about Yokuts,5 the different styles and genres of Kuna differ in their utilization of the suffix potential. Each of the many styles and genres of Kuna (See Sherzer 1983) makes use of the resources provided by this set of verbal suffixes in a particular, somewhat unique way. There are interesting relations between and among

266

Joel Sherzer

Table 2.

Kuna suffixes: position classes

Verb stem Position I: a. suffixes that can also stand alone as verbs 1. four positional: mai, kwici, sii, nai 2. temporal perspective suffixes: tii, tae 3. movement and direction suffixes: natappi, nae, tanikki, nonikki, take b. number suffix: pukkwa 'many' c. leke 'passive' d. kala 'in order to' Position 2: a. tense suffixes: sa, oe b. temporal perspective suffixes: sokkali, ali, te (oe can occur after te) c. movement and direction suffixes: tappi, appi {oe can occur after tappi) Position 3: a. temporal perspective suffix: pinne b. number suffix: mala c. distributive suffixes: moka, pali (oe occurs after pali and moka) d. modality suffixes: suli, pie Position 4: a. clause connectors: kua, le, sokkua Position 5: a. modality suffix: tipa Position 6: a. narrative suffix: sunto Position 7: a. modality suffix: ye

styles and genres in this regard. There are suffixes which have a greater frequency and greater range of meaning in particular styles or genres. Ye, which in colloquial Kuna serves an optative and emphatic function, occurs with great frequency in magical and curing chants, often marking the ends of lines, sometimes artistically embellishing these texts by being repeated two or three times. This is an illustration of a process which might appropriately be called the poeticization of grammar. Magical and curing chants make use of a relatively small subset of the full set of Kuna verbal suffixes. At the same time, there are suffixes unique or almost unique to this genre, such as kua and tar, which seem to have little or no referential meaning and serve mainly as markers of the magical-curing style.

The Kuna verb

267

Another way in which the use of verbal suffixes differs across styles and genres is in the number used per verb. In the most everyday, colloquial speech, verbs often occur with one, two, or no suffixes. Verbs in magicalcuring chanting have at least one and usually two or three suffixes. The same is true in the chanting of chiefs, which makes use of a different but overlapping subset of suffixes from the magical-curing subset. Narration within everyday speech and especially in formal contexts uses more suffixes per verb than these other styles and genres and exploits the full potential set more fully and more widely. Still another way to look at the relationship between styles and genres and verb suffixes is to note certain constellations of suffixes which tend to go together in each genre. There are many such constellations in each style and genre and a complete statement will ultimately require both qualitative and quantitative rules. At this point the best way to provide a sense of the situation I have been discussing is through the presentation of illustrative examples. All of these examples were recorded in the context of naturally occurring speech events. The first numbered line of transcription is the form actually spoken or chanted, along with an English translation. The second numbered line of transcription represents the full morphological form. From everyday conversation (1) namo: it will also go (2) na(e)-mo(k[a]) go-also (1) soysit: he is saying (2) sok(e)-si(i)-t(i) say-sitting-NOMINALIZER (1) amis: did you get (2) ami{e)-s{a) get-PAST From magical-curing chanting a set of parallel lines: (1) akkuekwiciye: are reaching (2) akku(e)-kwic{i)-ye reach-vertical-OPTATIVE (1) upoekwiciye: are being placed (2) upo(e)-kwic(i)-ye be placed-vertical-OPTATIVE

268

Joel Sherzer

(1) pioklekekwiciye: are being nailed (2) piokle(ke)-kwic(i)-ye be nailed-vertical-OPTATIVE In this example the three verbs, akkue, upoe, and piokleke, are in the final position of three parallel lines. The morphological parallelism is expressed by the repetition of the structure: verb stem-kwici-ye. From chief s chanting (1) an parmialimarye: he sent us (2) an(i) parmi(e)-al{i)-mal(a)-ye us send-coming-PLURAL-OPTATIVE from two parallel verses: A: (1) urpisaye: he left (2) urp{e)-s{a)-ye leave-PAST-OPTATIVE (1) nonimar: we came (2) noni(kki)-mal(a) come-PLURAL B: (1) urpisa: he left (2) urp(e)-s(a) leave-PAST (1) nonimarye: we came (2) noni{kki)-mal{a)-ye come-PLURAL-OPTATIVE These two parallel verses, A and B, consist of two lines each, the first line ending in urpe and the second in nonikki. Notice that in the first verse, urpe has ye suffixed to it while nonikki does not, while in the second verse nonikki has ye suffixed to it while urpe does not. This sort of artistic play with ye is typical of both chiefs chanting and magical-curing chanting and is another example of the poeticization of grammar. From narration within conversation (1) sokkarsunto: he said to me (2) sok(e)-al(t)-sunto say-INCEPTIVE-narrative (1) noar: it (the wind) came up (2) no(e)-al{i) rise-INCEPTIVE

The Kuna verb

269

From formal narration of a personal experience (1) an pankusmarmoye: I left you (2) an(i) panku(e)-s(a)-mal(a)-mo(k[a])-ye I leave-PAST-PLURAL-also-OPTATIVE (1) sunmaysaku: when they had finished talking (2) sunmakk(e)-s(a)-ku(a) talk-PAST-when (1) soymosunto: I said also (2) sok(e)-mo(k[d\)-sunto say-also-NARRATIVE From formal narration of a traditional folktale (1) natparsunto: then she went again (2) na(e)-t(e)-pal(i)-sunto go-then-again-NARRATIVE (1) amiarkua: when he began to look (2) ami{e)-al(i)-ku(a) look-INCEPTIVE-when (1) ekwacasunto: she threw (2) ekwan(e)-s(a)-sunto throw-PAST-NARRATIVE (1) ekiciarsunto: she began to ask (2) ekic(e)-al(i)-sunto ask-INCEPTIVE-NARRATIVE This constellation of verb stem-tense or aspect marker-sunto is quite typical of this genre. In particular, the sequence al(i)-sunto is so common that it sometimes can be considered a sociolinguistic marker of the genre. I offer this preliminary overview of the Kuna verb in memory of Stanley Newman and as a contribution to a continuation of his work and the particular themes and spirit within anthropology and linguistics that he so well represented and which I found in that very provocative statement about Yokuts with which I began this paper. The Kuna data speak to some of these themes, in the following ways. A central aspect of culture is the logic, the world view, and the symbols by means of which a group of people perceive and conceive of the world. Language is central to this view of culture. Language is the verbal means by which individuals conceive of, express, and transmit their culture. But by language I do not mean only grammar, in the abstract structural sense, but rather discourse, the actual narrations, myths, speeches, and conversations with which people communicate and express themselves.

270

Joel Sherzer

Grammar provides the resources, potentials for the creation of meanings in discourse. In order to study this process in which grammatical resources and potentials are actualized in discourse, we must examine the forms of discourse themselves. It is not through Kuna grammar alone, but through its exploitation in various forms of discourse, often artistic discourse, that we can appreciate the distinctive ways Kuna culture, through individual voices, organizes, creates, recreates, and perceives the world.6 Translation plays an interesting role here. Since as we make sense of Kuna suffixes, from grammatical, social, cultural, and discourse points of view, we strive to translate the subleties of Kuna into our own language, in this case English, and thus come to appreciate the particular efficiency and economy of the Kuna system. Finally, Newman's statement and the Kuna data raise interesting questions with regard to the relationships and boundaries between grammar and discourse, between linguistic analysis and stylistic analysis. For what is Kuna grammar: the theoretical system of potentials that I can construct and which generates forms that Kuna individuals find amusing, though interpretable, or the set of systems and rules that underlie the actually occurring forms in different genres and styles? This is a false dichotomy, as most are. Grammar and discourse must be seen as ill-defined realities, in fact realities whose boundaries and structures emerge in social contexts, including both contexts of language use by speakers in communities around the world and contexts of analysis of language, by natives as well as professional linguists. And competence in language use no doubt includes aspects of all of what I have been describing here — suffix meaning, suffix order, potentials actualized, potentials not actualized, differential suffix usage in various styles and genres, and humorous uses and interpretations of suffix combinations. The situation described here has a relevance that goes beyond Kuna and beyond suffixes. With regard to affixation, the issue of the position and number of affixes theoretically possible vs. those that actually occur in a word will be familiar to individuals who have struggled with the analysis of native American languages, though the sociolinguistic and linguistic anthropological implications of such situations have rarely been drawn in the literature.7 But syntactic structures and possibilities, in languages such as English, can also function to distinguish styles and genres.8 Most generally, the point is that function and style rather than being marginal to linguistic theory and analysis should actually be quite central to them.9

The Kuna verb

271

Notes 1. Compare this statement with Newman's more technical description in his 1944 grammar of Yokuts (1944:81—82), which includes a statistical statement. Newman is talking about thematizing suffixes, which are derivational and not inflectional. Most of the Kuna suffixes I discuss here are inflectional. 2. See Sherzer (1983:36 — 39) for a discussion of Kuna phonology and a description of the orthography used here. Glosses of certain grammatical categories are written in capitals in my examples. 3. Other classificatory schemes than the one presented here are of course possible. 4. This is a quite problematical distinction. No doubt further analysis will reveal it to involve much more complexity than a simple dichotomy would seem to indicate. Furthermore notice that I have not dealt with the quite challenging question of the meaning of the absence of particular forms and whether and how many zero forms need to be included in the analysis. 5. Newman provides data from one style, myth dictation. 6. For further elaboration of this position, see Sherzer (1987). 7. For an exception and a description of a sociolinguistic situation nicely comparable to the Kuna one, see McDowell (1983). 8. See Labov's (1972) discussion of the relationship between narration and conversation. 9. On this matter the work of Dell Hymes has been quite insistent. See, among others, Hymes (1974). The subject matter of this paper seems to have been provocative enough to generate a remarkable set of trenchant critiques and commentaries. I am most grateful to John Baugh, James Howe, Dell Hymes, Michael Silverstein, Brian Stress, Randy Valentine, and Anthony Woodbury in this regard. Their comments will remain with me as I continue to wander through the complexities of the Kuna verb.

References Gayton, Ann H. — Stanley S. Newman 1940 Yokuts and Western Mono myths. Anthropological Records, volume 5, no. 1. Berkeley and Los Angeles: University of California Press. Pp. 4 — 11 reprinted as Stanley S. Newman "Linguistic aspects of Yokuts style." In Language in Culture and Society, edited by Dell Hymes, pp. 372—377. New York: Hapar & Row, 1964. Hymes, Dell 1974 Linguistic theory and functions in speech. In Foundations in Sociolinguistics: an ethnographic approach, 145 — 178. Philadelphia: University of Pennsylvania Press. Labov, William 1972 The transformation of experience in narrative syntax. In Language in the Inner City: Studies in the Black English Vernacular, 354—396. Philadelphia: University of Pennsylvania Press. McDowell, John H 1983 The semiotic constitution of Kamsä ritual language. Language in Society 12:23-46.

272

Joel Sherzer

Newman, Stanley 1944 Yokuts language of California. New York: Viking Fund Publications in Anthropology, No. 2. 1964 (Partial reprint of Gayton and Newman — see above.) Sherzer, Joel 1983 Kuna Ways of Speaking: an ethnographic perspective. Austin: University of Texas Press. 1987 A discourse-centered approach to language and culture. American Anthropologist 89:295-309.

D. Word Formation

Navajo stem variation Herbert Laudar

Abstract The Navajo verb stem varies systematically across modes, allowing partial similarities to be identified. In this paper, Stanley Newman's contributions to stem classification and description of root variation are adapted to a prototype approach with analysis and classification of the Navajo stem as composed of root and suffix. Five major categories of verb stems are described, with 19 classes defined by mode or mode combinations, and subtypes defined by consonantal and vocalic features.

1. Stanley Newman's studies of stems in Bella Coola and stem classes in Bella Coola and Zuni (1965, 1969, 1971) have led me to a better understanding of Navajo stem variation. Of particular interest is how Newman (1971) compares Bella Coola stems with their modifications. In Bella Coola one moves from a stem to a diminutive or continuative with consonant and vowel symbolism and some insertions of consonants or vowels; suffixation confirms formation of the diminutive. Here Newman sets focus on feature analysis of phonemes. He sets focus on morphology in classifying stems, classifying Zuni stems, for example, by shapes for present or past and selection of suffix alternants. To follow Newman is to part company with Pike and Becker (1964), Sapir and Hoijer (1967), and others who take the Navajo stem to be a single morpheme or part of a discontinuous morpheme. Pike and Becker describe classes of verb stem sets with different inflectional patterns across two, three, and four alternative stem forms within a given set. They propose eight classes of the four alternative type, twelve of the three alternative type, five of the two alternative type, and four of the single stem type. Despite some faults in handling their data, Pike and Becker were the first to show, as I prefer to express it, that Navajo active verb stem sets fall into classes partitioned by vowel and suffix features. My own slant on such partitioning is given below. I maintain that morphological analysis is independent of homophony and Pike and Becker's internal reconstruction.

276

Herbert Landar

If Pike and Becker favor internal reconstruction, Kari puts faith in external reconstruction. Kari (1976:300) postpones full analysis of Navajo stems until better comparative Na-Dene data are available. He shows, nevertheless, how stem variation relates partly to prefixes and partly to shifting, competing patterns in popular usage. Systematic shifts are apparent when one compares sets collected by Sapir and Hoijer from Albert Sandoval (Hoijer 1974) and more modern sets of Young and Morgan (1943, 1951, 1980). Regardless of shifting patterns, the Navajo system involves morphological conventions distant from anything Salishan. Each Navajo stem consists of a root conforming to a basic shape of CV (in an orthography suited to Newman's for Bella Coola) and a stem suffix usually of shape -C, though -CV or -V may occur. The basic vowel may be lengthened, elevated in pitch, or nasalized. Taking quality from the imperfective stem, there is, moreover, some symbolism in vowel changes for the perfective, future, iterative and optative stems. (In citing stems, I follow the SapirHoijer sequence, ipf., pf., fr., iter., opt.). The stem suffix for a given mode has numerous variant forms. In a sample of the first 152 stem sets in Hoijer (1974), there are eleven variants of the imperfective stem suffix, twelve of the perfective, four of the future, four of the iterative, and eight of the optative. These are supplemented by -d, a ninth optative variant, by Pike and Becker (1964). These variants rarely establish meaning, modal or aspectual, by themselves. Three suffix shapes, indeed, -h, -t, and -s, complete the stem of any of the five categories. The richest treatment of Navajo stems from a statistical point of view is that of Hoijer in Sapir and Hoijer (1967:56 — 67). Hoijer, using 647 active bases, some with fewer than five modes, identifies favorite patterns, including (1) 11111 (70 cases with one stem form); (2) 11221 (54 cases with ipf.-pf.-opt. identical in one way and prog.-iter. identical in another); (3) 12111 (38 cases with all stems identical except the pf.); (4) 12331 (213 cases, the prime favorite); (5) 11231 (51 cases); (6) 12311 (46 cases); (7) 12341 (67 cases); (8) 12345 (24 cases). Hoijer does not use such modal patterns in classifying verb stems per se, preferring rather to identify four categories based on matched pairs of stems. He matches the "arbitrarily basic" imperfective with identical or changed forms of the other modes. The first category (ipf.-pf.) has 39 subtypes, counting inverse patterns as a single type; the second (ipf.-opt.), 17; the third (ipf.-prog.), 36; the fourth (ipf.-iter.), 30. Within these categories, subtypes are identified by how stem endings coincide or vary across modes and how root vocalisms do the same.

Navajo stem variation

277

Hoijer's classification is valuable, in my view, partly in showing which stem endings are replicated or changed in going from the imperfective to counterparts in other modes; partly in supporting feature analysis of roots; and partly, as I would put it, in identifying classes of roots which are environments for grammatical alternants of stem suffixes. Hoijer shows a decomposed, shifting consonantal and vocalic symbolism, complex and opaque. Bella Coola symbolism is more complex and opaque. The diminutive of a CVC- stem form qinx 'shoe', for example, is made (on Newman's pattern 1213 ... where the basic image to be modified is qin...) into qiqnq-i. The diminutive suffix -i follows the reduplicative prefix qi- and the reduced underlying image -qnx- which has been altered symbolically to -qnq-. The shift from x to q is of Newman's type C-l, velar fricative to corresponding velar stop, seen also in tasa lxi 'nape of the neck' > ?asa-lqi-0 dimin., and spi xw- 'hit the head with a stick' > spipqw- contin., where a symbolic vowel change of Newman's type V-2 occurs, long vowel to short in a derived stem having one vowel (Newman 1971:35). Shifts of sound features do not correlate clearly with particular meanings. They do not support easy learning, any more than do shifts in Germanic strong verbs or Navajo stems. 2. It helps, in learning Navajo stem sets, to assume that every monosyllabic stem reflects or develops from an underlying shape, CV-C. Consonants which fill the -C part of this formula are shown below. Ipf.

Pf.

-0

-0

-d -h -t -s -z

-d -h -i -s -z

-S

-z -n -g

-z -n -I -g

F T .

Iter.

Opt.

-h -t -s

-h -t -s

-d -h -t -s -z

-S

-s

-S

-z -n

The vowel of the imperfective generally is kept for all members of a stem set, but sometimes the quality changes. In the sample from Hoijer (1974) mentioned above, these changes were encountered:

278

Herbert Landar

Ipf.

Pf.

i

e

i

a

e

i

Fr.

Iter.

Opt. a

i

i

e

a

a

e

o

o

Types of changes which occur in stems, in the provisional perspective of this paper, are Type C-l C-2 C-3 V-l V-2 V-3 V-4 V-5 V-6 V-7 V-8

Change Stem suffix consonant goes to zero in word final position (-C > -0). Stem suffix consonant goes to vowel (-C > -V). Stem suffix consonant, unvoiced, goes to voiced counterpart (e. g., -t > -I). Stem initial consonant, unvoiced, goes to voiced counterpart (in correlations of noun and verb stems). Stem vowel goes from short to long (V > V). Stem vowel goes from low to high pitch (V > V). Stem vowel goes from oral to nasal (V > Y). Stem vowel goes from high front to mid front (i > e). Stem vowel goes from high front to low central (i > a). Stem vowel goes from mid front to high front (e > /). Stem vowel goes from mid front to low central (e > a). Stem vowel goes from mid front to high back (e > o).

Sapir and Hoijer (1967:56—67), characterizing a sample of 647 active bases, show the need for additional features: V-9 (o > i), V-10 (a > i), V-ll (a > o) and V-l2 (a > e) as in -Xo (ipf.), -Xi (pf.) 'become cold,' -t'q (ipf.), -t'i i (fr.) 'ripen', -ya-s (ipf.), -yos (fr.) 'go to sleep', -na (ipf.), -ne-t- (fr.) 'capture'. 3. The meaning of a verbal expression depends in large measure upon prefixes which anticipate the stem of the verb. The run of prefixes, in a given pragmatic context, constrains possibilities of stem selection. Thus stem variation is determined in part by morphology. Stem initials and vowels are affected by various assimilations, indicated in Hoijer (1945:31-59) and Young and Morgan (1980:391-93); for differences see Kari (1976:285 — 301). Certain enclitics affect stem shape (Hoijer 1945:31 — 32, Landar 1963:24). The nominalizing enclitic /• 'the one (who,

Navajo stem variation

279

which, what ...)', for example, depresses high pitch of a long stem vowel before the stem suffix. In illustration, consider -lé-h, -la-, -li t, -9[ ~ -9ih, -le9 'create, make'. Assimilated to sg. 1 s-, stem initial I becomes t, as in be9eslé-h 'I am copying it' (ipf.), bPi sta• 'I have copied it' (pf.), bi9de-sH-l 'I shall copy it' (fr.), bipòste9 'I wish I could copy it' (opt.). The stems with change V-6 are usitative or customary -9i and iterative -9i-h with sg. 1 forms be9es9i 'I copy it customarily (now and then)' and bìnà9i-s9{-h 'I copy it repeatedly (again and again)'. The nominalizing enclitic i- creates another stem alternant pair, -le-h and -iè-h, as in beveste-h i- 'what I am copying'. Nominalized forms in the future are bPde-sli-t i• 'what I shall copy', bPdi-li-l i- 'what you (sg.) will copy', yi9do-li-i i- 'what he will copy', bizdo-li-t i- 'what one will copy', bi9di-lni-l i- 'what both of us will copy', bi9do-hli-l i- 'what both of them will copy'. The stem classifier /- in the next to last form dissimilates the stem initial to rt. Note also 9àlnéh 'it is being made' (-né-h) beside 9àsié-h 'I am making it' (-lé-h) and 9ànilé-h 'you are making it' {-lé-h). The perfective listed as -la-, similarly, may occur also as -la- (?àsta- or 9i-sta- 'I made it') or -ya- (9i-lya• 'both of us made it', 9àlya- 'it was made'). Another area of variation involves stress. In speech, not every stem is marked by strong stress; some have medium or, rarely, weak stress. Thus one might list variants with different stress markings, if such markings are systematic. Variability in stress is illustrated with - 9 l 'see', whose shape is conditioned by the enclitic da of do- ... da 'not'. The stem has strong stress in Wo- yo- '9{- da 'he can't see', sdo- yis'9i da 'I can't see it (sg. 3)', "do- 9is'9ì da 'I can't see it' (indef.). In my notes, the nominalized forms of the future given above have strong stress on the stem and weak stress elsewhere, except for bi9di'lil i- 'what you will copy'. Another kind of variation preserves -n intervocalically (Hoijer 1945:40) beside a stem-final nasalized vowel elsewhere. Note with - 9 i 'see' the forms do- yis9p da 'I can't see her', do- yis9i-n i- 'the one that I can't see'. An assimilated form of this last phrase, du- (i)Pi-ni-, means 'my motherin-law' (man speaking), in allusion to the taboo against a man looking at his mother-in-law. 4. Hoijer (1974) and Young and Morgan (1980) have noted some correspondences of shape for nouns and verbs. Before we consider classification of active verb stems, it is useful to see how the noun base ties in with one or more of the members of an active set as well as with any neuter stems. The independent stem noun ba 9 'raid, war', for example, correlates with the perfective stem only of two sets: -ba-h, -ba-9, -bah, -bah, -ba h

280

Herbert Landar

'raid', -ba h, -ba-9, -bah, -bah, -ba h 'be raiding around'. As with interjections, most of our unique correlations are with perfective stems (Landar 1985). There is, however, one other unique correlation and some cases of correlation with several stems. The noun base or one of its alternants shares identity with a perfective stem only, given an active stem set, in (1) ba-9 'raid, war', -ba-9 'raid, make war'; (2) ba-l 'tent, tarpaulin', -ba l 'hang like a curtain; drape on or over'; (3) ca 'weeping, crying', -ca 'cry, weep, make loud noises'; (4) cq-9 'excrement', -cq-9 'defecate'; (5) ci-l 'snowstorm', -ci-l 'snow'; (6) coz 'boil (ulcer)', -coz 'become covered or afflicted with boils'; (7) -ci in nilci 'air, breeze, spirit', -ci 'blow (as a light wind)'; (8) da n 'springtime', -da n 'become or be springtime, pass (a springtime)'; (9) diz 'floating trash or driftwood (in a stream)', -diz 'twist (a slender flexible or slender stiff object); roll, wind, tangle, spin (a flat flexible object)'; (10) xai 'winter', -xai 'pass (a winter)'; (11) koi 'vomit', -koi 'vomit'; (12) sq 'old age', -sq 'grow'; (13) se-9 ~ -ze-9 'saliva, spittle', -ze-9 'spit'; (14) liz ~ -liz 'urine', -liz 'gush, leak; urinate'; (15) s( 'summer', -si 'become summer'. Not includable here, but instructive, is (16) cil 'plant, weed', identical in shape not with pf. -ci-l of the active set -ci-l, -ci-l, -cil, -cil, -ci-l 'curl, move in a curve or curl' but with the neuter pf. -cil 'be curly, creased, wrinkled'. A noun matches the shape of the imperfective and (in these cases) optative stem in (17) bq-s 'hoop', -bq-s 'roll'; (18) ci-h 'red ochre', -ci-h 'redden'; (19) Xe s 'white clay', -Xe-s 'paint, smear'; (20) j;o-l 'ball', -jot 'drift or tumble (e.g. like tumbleweed)'; (21) jeh 'resin, pitch', -je-h 'stick, adhere, smear'; (22) with V-l, xai 'club', -xa i 'beat'; (23) with V-2, -ka-h in 9i-ka-h 'sand painting', -ka-h 'make a sand painting' (24) te-l 'cattail, rush', ? -te l 'spread in masses (around a lake)'; cf. neuter ipf. -te-l 'be broad, wide'; (25) t'e-s 'charcoal', -t'e-s 'blacken with charcoal'; (26) with V-2, -Xi-s 'snake', -Xi-s 'wriggle, writhe; be zigzag'. A noun matches the shape of the future in (27), with V-l, nit 'stone axe', -ni l 'split (as by an axe)'. A noun matches the future and iterative in (28) Xoh 'mirth, laughter', -Xoh 'laugh, smile'; (29) jil ~ -ji-l 'mountain', -jil 'be or become firm or strong'; cf. -jil 'stiffen, form as a hard mass' (e. g., dough, concrete), one shape for all modes; (30) xit 'night', -xit 'become dark'; (31) toh ~ -loh 'noose, loop', -loh 'act using a loop of rope'; (32) with affrication to [tz], zas 'snow', -jas 'snow, have snowfall'. A noun matches the perfective and iterative in (33) Uai9 'willow', -Kai9 'be forked, have limbs spread'. (Here I follow Young and Morgan (1980). Hoijer (1974) has -kai9 uniquely perfective.)

Navajo stem variation

281

A noun matches the perfective and optative in (34) tid ~ -lid 'smoke', -lid 'burn, give off smoke'. A noun matches the imperfective and future in (35) with C-3, sis ~ -zi-z 'belt', -zis 'put on one's belt'. (Here I follow Young and Morgan (1943); Young and Morgan (1980) show -zis, however, as ipf., fr., iter., opt.) A noun matches the imperfective and perfective in (36) with C-3, sa-z —za-z 'knot, joint, mended place', -za-s 'knot, mend'. A noun matches the imperfective, perfective and optative in (37) tin 'ice', -tin 'freeze'; (38) kid"fart', -kid (ipf., opt., with V-l), -kid (pf., V-l, V-2) 'throw off objects, spurt (liquid), fart'; (39) ¿in ~ -¿in 'bone', -¿in 'become hard, ripe or rigid; strike with fist or (of a horse) front leg' (pf. also -¿id). A noun matches a stem invariant for all modes in (40) cos 'bug, worm, maggot', -cos 'get wormy, maggoty; cluster into a bunch'; (41) tea-9 ~ -Ka9l-1ca 9 'arrow', -kah 'shoot it once with an arrow' (with glottalized and plain initials); (42) sei zei 'sand', -sei 'crumble'; (43) kah 'salve, ointment', -kah 'anoint, put on salve, grease or cold cream'. I have found two cases, in addition to (41), of glottalized vs. plain initial, (44) cat 'frog', -ce-h, -ca-cah, -cah, -ce h 'hop, jump' and (45) ke9 'night', -ki-s, -ki-z, -¿is, - , - 'become dark'. On (45) see Hoijer (1974:244); Young and Morgan (1980:436); Haile (1950:335). A noun matches the future and optative in (46) with C-3, xet ~ -ye-t 'pack, load, burden', -ye-t 'handle or move a pack, burden or load'. We saw with (16) that a neuter verb may match a noun base perfectly, while no stem of the active set matches as closely. Another example is (47) $ 'day', -j£h, -jid, -ji t, -j,i-h, -¡fch 'carry a ceremony to its final day', neuter ipf. -ji 'be in the final day of a ceremony'. In general, however, the match of noun and neuter stem is close but not perfect. A noun matches the neuter imperfective in (48) bes ~ -be z 'flint, metal, knife', -be s 'be flinty'; (49) with C-3, xos ~ -yos 'thorn, cactus', -yoz 'be thorny'; (50) 9it — ? / 7 'evergreen needle, bough', -9il(i) 'be hairy, bristly'; (51) -Ki-z 'body half, side of carcass', -Hi-z 'have a side, half; (52) Sg? ~ -Zg? 'star', -zgs —sgs 'glitter'; (53) cah 'awl, needle', -cah 'be like an awl, poke'; (54) -¿id in ^¿id 'sinew', -¿id 'be tough, sinewy.' Thirteen classes of nouns which relate to verb stems are shown in Table 1. 5. Five major categories of verb stems are shown in Tables 2 — 6, based on full active stem sets found in the first 152 sets in Hoijer (1974).

282

Herbert Landar

Table 1.

Classes of Navajo nouns related to verb stems.

Noun class

Matching active stems

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

ca 'crying, weeping' j e /i 'pitch, resin' nit 'stone axe' loh 'mirth, laughter' xe-l 'burden, load' Rai? 'willow tree' lid 'smoke' sis 'belt' sa-z 'knot' tin 'ice' ¿ah 'salve, ointment'

Matching neuter stems

M 'day' cil 'plant'

These classes have been identified by matches shown: 1. pf.; 2. ipf.-opt.; 3. fr.; 4. fr.-iter.; 5. fr.-opt.; 6. pf.-iter.; 7. pf.-opt.; 8. ipf.-fr.; 9. ipf.-pf.; 10. ipf.-pf.-opt.; 11. ipf.-pf.-fr.-iter.opt.; 12. ipf.; 13. pf.

Category A contains unique shapes; B, shape duplication for two modes, e. g., ipf.-opt.; C, the same for three modes; and so on. Within the major categories are 19 classes, five for A, four for B, five for C, four for D, and one for E. These 19 classes are defined by mode or mode combinations. A stem belongs to class Al, for example, if it is an imperfective whose shape in its stem set is unique, or to class C2 if it represents three modes in its set, imperfective, perfective, and iterative. Types within a class are defined by the shape of the stem suffix. A stem belongs to type Al(2), for example, if it ends in -d, as with -ta d 'sprinkle (mom.)', or to type C2( 1) if it ends in -h, as with -dg h 'shoot, make a bang (repet.y. Stem subtypes are defined by attributes C-l through C-3 and V-l through V-8, as detailed after Table 6. Abbreviations with glosses, such as mom. or repet., indicate aspects. The abbreviations of Tables 2 — 6 are mom., momentaneous, indicating start, finish, or other transitory stage of a sequence; cont., continuative, indicating moving along, being moved along, or some similar continuation; repet., repetitive, indicating multiple acts in a series, as in twisting, pounding, scratching or kicking; div., diversitive, apportioning action or setting here and there; distrib., distributive, apportioning individual objects here and there; and semelf., semelfactive, limiting an act to a single instance.

Navajo stem variation Table 2.

Category A. Unique shapes.

Class 1.

Imperfective

1. b. 2. 3 a. b. c. 4 a. b. c. 5. 6.

-ta 'look for, search' -?/ 'look at' -ta d 'sprinkle' (mom.) -ba-h 'raid' (cont.) -be-h 'fade, shrivel, wear out' -9i h 'hide' (mom.) -9a-i 'chew' -di i 'take place (noise)' (mom.) - 9 i i 'copulate' - 9 a s 'move (as a herd or flock)' -9in 'hide' (cont.)

Class 2. 1 a. b. c. d. e. 2 a. b. c. d. e. 3 a. b. 4. 5. 6a. b. c. d. 7 a. b. c. d. 8 a. b. 9 a. b. c.

283

Perfective

-da 'sit (one living being)' -dq- 'become springtime' - 9 q 'move (as a round solid object)' -bq 'win (a game)' -ti 'handle a corpse or carcass' -ba 9 'be kind, take pity' -ba-9 'raid' (mom.) - 9 i 9 'steal' -bq9 'fade, shrivel, wear out' -bi9 'swim' (mom.) -9a-d 'direct, command' (cont.) - 9 i d 'become light (to the eyes)' -tah 'snap, jab, move quickly or suddenly' - 9 H 'look, watch' -9a-z 'move (as a herd or flock)' - 9 a-i 'move (as two living beings)' -9e-z 'tuck, crimp, ruffle' -taz 'tap, peck at' -diz 'spin, twist' (mom.) -ta-z 'bend, spin, twist' (mom.) -ma z 'roll like a ball' (mom.) -bq z 'roll' -9in 'hide' (cont.) -dq-n 'become springtime' -ba-l 'hang, swing (as fabric)' -9e-l 'float' -de-l 'move freely through space, as ropelike objects or two or more objects of a set' (repet.) 10. -doi 'become hot'

284

Herbert Landar

Table 2. Continued Class 3. 1. 2 a. b. c. d. e. f.

-9ih 'look, watch' -tai 'snap, jab, move quickly or suddenly' -be t 'pluck, pick (e. g., fruit)' -bii 'several sit down' (mom.) -9H 'bark like a dog' -bit 'fill' -Whang'

Class 4. 1 a. b. c. d.

Future

Iterative

-be-h 'swim' (cont.) -te-h 'handle a corpse or carcass' -bi-h 'several sit down' (cont.) -di-h 'get used to, fond of

Class 5. Optative 1 a. b. c. 2 a. b. c.

-9a9 'move as a round solid object' (cont.) -fi-f 'do to, treat so, become' -ba-f 'win (a game)' -ba-h 'raid' (cont.) -ta-h 'attempt, try to' (cont.) -9a-i 'move (as a round solid object)' (div.)

Table 3. Category B. Duplicates Class 1. Imperfective-Optative 1. 2. 3. 4 a. b. c. 5 a. b.

-te? 'animate being moves' (distrib.) -ba-h 'raid' (mom.) -9e-i 'float' -9a-s 'two living beings move' -?e-s 'lead or guide several' (mom.) -tq-s 'tap, peck at' -di s 'spin, twist' (mom.) -bq-s 'roll'

Class 2. Imperfective-Iterative 1 a. -9a h 'move as a round solid object' (mom.) b. -9a h 'move as a round solid object' (div.) c. -9i~h 'become light'

Navajo stem variation Table 3.

Continued

Class 3.

Imperfective-Perfective

1. -bp 'build house or hogan' 2. -dog 'fill to a bulge' Class 4.

Future-Iterative

1 a. -bah 'raid' (mom.) b. -dah 'several persons move freely through space' (mom.) c. -tih 'run' 2 a.-bai 'hang or swing as fabric' b. -?oi 'float' 3 a. -9as 'two living beings move' b. -bis 'braid' c. -"is 'lead or guide several' (mom.) d. -tq£ 'tap, peck at' 4 a. -mas 'roll like a ball' (cont.) b. - % 'move feet, move with feet'

Table 4.

Category C. Triplicates

Class 1.

Imperfective-Perfective-Optative

1 a. b. 2 a. b. 3. 4.

-if? 'talk, speak' (cont.) 'look, watch' -biz 'braid' -be-z 'boil, bake, cook' -diz 'spin, twist' (repet.) -bin 'several sit down' (mom.)

Class 2.

Imperfective-Perfective-Iterative

1. -dg-h 'shoot, make a bang' (repet.) Class 3.

Imperfective-Future-Iterative

1. -tas 'bend, spin, twist' (mom.) 2. -9is 'tuck, crimp, ruffle' Class 4.

Perfective-Future-Iterative

1. -dih 'limp' 2. 'copulate'

285

286

Herbert Landar

Table 4.

Continued

Class 5.

Imperfective-Iterative-Optative

1 a. -be-h 'swim' (mom.) b. -dih 'experience, live through'

Table 5.

Category D. Quadruplicates

Class 1.

Imperfective-Future-Iterative-Optative

1. -tah 'sprinkle' (distrib.) 2. -dU 'move freely through space as ropelike objects or objects in a set' (repet.) 3. -dis 'spin, twist' (cont.) Class 2.

Imperfective-Perfective-Future-Iterative

1. -tah 'sprinkle' (cont.) Class 3.

Imperfective-Perfective-Iterative-Optative

1. -dg-h 'straighten' (semelf.) Class 4.

Perfective-Future-Iterative-Optative

1. -tah 'sprinkle' (mom.)

Table 6.

Category E. Quintuplicates

Class 1.

Imperfective-Perfective-Future-Iterative-Optative

1. -tah 'sprinkle' (repet.) 2. -tai 'move legs or feet; kick' (semelf.) 3. -dis 'spin, twist' (semelf.)

Categories A —E were set up after analysis of features, illustrated here for two classes of category A. Category A, class 1, type 1 a, -0 ipf., C-l, V-2; l b , -0 ipf., C-l, V-2, V-3; 2, -d ipf., V-l, V-2; 3 a, -h ipf., V-l; 3 b, -h ipf., V-l, V-2; 4 a, -t ipf., V-l\ 4 b, -i ipf., V-l, V-2; 4 c, -i ipf., V-l, V-2, V-3; 5, -s ipf., V-l; 6, -n ipf. Class 2, type 8 a, -n pf.; 8 b, -n pf., V-l, V-3; 9a, -/pf., C-3, V-l; V-2; 9c, -/pf., C-3, V-l,V-2, V-8; 10, -ipf., C-2.

Navajo stem variation

287

References Haile, Berard 1950 A stem vocabulary of the Navaho language. Vol. 1, Navaho-English. St. Michaels: St. Michaels Press. Hoijer, Harry 1945 Navaho phonology. University of New Mexico Publications in Anthropology 1. Albuquerque: University of New Mexico Press. 1968 Navajo reference verbs and verb expressions made up of two verb forms. IJAL 34.176-82. 1974 A Navajo lexicon. UCPL 78. Berkeley and Los Angeles: University of California Press. Kari, James M. 1976 Navajo verb prefix phonology. New York and London: Garland Publishing, Inc. Landar, Herbert 1963 Navaho syntax. Lg. Diss. 57, Supp. to Lg. 39 (no. 3, pt. 2). 1985 Navajo inteijections. I JAL 51.489 - 91. Newman, Stanley 1965 Zuni grammar. University of New Mexico Publications in Anthropology 14. Albuquerque: University of New Mexico Press. 1969 Bella Coola paradigms. I JAL 35.299 - 306. 1971 Bella Coola reduplication. IJAL 37.34-38. Pike, Kenneth L.—Alton L. Becker 1964 Progressive neutralization in dimensions of Navaho stem matrices. IJAL 30.144 - 54. Sapir, Edward—Harry Hoijer 1967 The phonology and morphology of the Navaho language. UCPL 50. Berkeley and Los Angeles: University of California Press. Young, Robert W. — William Morgan 1943 The Navaho language. Phoenix: U.S. Indian Service. 1951 A vocabulary of colloquial Navaho. Phoenix: U.S. Indian Service. 1980 The Navajo language: A grammar and colloquial dictionary. Albuquerque: University of New Mexico Press.

Lexical morphemes in Bella Coola Ross Saunders and Philip W. Davis

Abstract It may be argued that Bella Coola opposes only two form classes, one of lexical content and a second of grammatical content. That is, Bella Coola is one of those languages which syntactically does not distinguish Noun from Verb. In this context, we find that between these extremes of Root and Affix, there lies a continuum of forms which partake in varying degrees of properties of both classes. We identify three mediating steps in the transition: Roots — Bound Roots — Predicative Prefixes — Referential Suffixes — Affixes. The behavior of several otherwise aberrant forms, e. g. Hii 'to lack', become sensible in that they can now be viewed as caught in midstride between stages. Finally, the existence of such a continuum adds support to the claim that Roots and Affixes constitute the major grammatical form classes in the language.

Introduction Bella Coola is a Salishan language1 in which the primary opposition in the system of form classes (or parts of speech) is between a class of lexical elements and a class of grammatical ones (cf. Kinkade 1983). Lexical morphemes designate entities, concepts, qualities, and events and in Bella Coola they may function either predicatively or referentially. Morphologically they are either roots or affixes. Like other Salishan languages, Bella Coola is structurally polysynthetic: rich in affixes which reflect not the semantic modulation of grammatical categories, but rather the semantic content of terms in construction with a predicate. For the most part, these affixes are suffixes and have been given the generic name of 'lexical suffixes'2. In order to include other affixal types we refer to the same set of lexical morphemes as 'referential affixes' and include therein: anatomical and non-anatomical lexical affixes, their metonymic extensions, classifiers, and quantifiers.

290

Ross Saunders and Philip W. Davis

Referential affixes Referential affixes refer to the subset of lexical morphemes which reflect the semantic content of PARTICIPANTS of an EVENT.3 They are mostly suffixes as in (1 —4). (1) (2) (3) (4)

cp -ai -ic tiialas- tx* J/wipe -vehicle -it I Prox -boat- Prox5 'I wiped the boat.' lis -ak -m -ic [/push -hand -ContraPass -it/I 'I pushed it with my hand.' lis -ixwlayx-ak -m-ic J/push -knee-hand -ContraPass-it/I 'I pushed his knee with my hand.' kma -lie -ak -c ]/hurt -skin -hand -I 'The skin on my hand hurts.'

In (1), the referential affix -at- is a 'classifier' reflecting the semantic content of the class of objects which serve as means of conveyance. The specific reference here is to the EXPERIENCER ialas 'boat'. The root cp specifies the EVENT-relationship between the PARTICIPANTS 'I' and 'boat'. In (2), the speaker is again the EXECUTOR of the ACTION (lis- 'to push') but the referential affix -afc-'hand' refers to the IMPLEMENT used by the EXECUTOR to push an unspecified EXPERIENCER. (3) contains two referential affixes -i:xwlayx 'knee' representing the EXPERIENCER and -ak- 'hand' the IMPLEMENT. In (4), a complex referential affix -lic-ak- 'skin of hand' composed of two referential affixes -lie- 'skin' and -ak- 'hand' represents the EXPERIENCER of the predicate kma- 'to hurt'. A smaller number of referential affixes are prefixes. With the exception of those which refer to body parts, they differ from other referential affixes in occurring only with ACTION roots. Cf. (8) and (9) in that regard. (5) (6)

qui -kma -0 penis -j/hurt -he 'His penis is sore.' 9 it -kic' -cinu clothes -J/wash -you/I 'I'll wash your clothes.'

Lexical morphemes in Bella Coola

(7) (8) (9) (10) (11)

291

9

it -9ip -ic ti -nup-s-tx clothes -|/take -it/I Prox -jacket-his-Prox 'I took his jacket.' * nt -t'aws -c clothes -|/damp -I 'My clothes are damp.' t'aws -ant -c ]/damp -clothes -I 'My clothes are damp.' ?£s -tux -is pelt -J/pull apart -it/he 'He is skinning it.' stam -xwism co-participant -|/play 'playmate'

(5) provides an example of a referential prefix reflecting the semantic content of a body part. (6 — 8) contain the referential prefix ?it'clothes'. (8) illustrates the non-occurrence of the non-body part referential prefixes with STATIVE roots while (9) illustrates how the corresponding referential suffix can occur with STATIVES. (10) contains the referential prefix which copies the semantic content of 'pelt' or 'animal covering'. Example (11) shows how a referential prefix is used in the derivation of a PARTICIPANT.

Predicative morphemes Lexical morphemes with predicative function occur in the manifestation of the EVENT. They are, for the most part, roots. Most roots can occur with or without referential affixes as in the following pairs: (12 a) b) (13 a)

cp -a -0 |/wipe -AntiPass -he 'He is wiping.' cp -at -a-0 x-a qxwmtimut-c [/wipe -vehicle -AntiPass-he Prep-Prox-car-Prox 'He is wiping cars.' cp -is J/wipe -it/he 'He is wiping it.'

292

Ross Saunders and Philip W. Davis

b) (14 a) b)

cp -ai -is ti-qxwmtimut-tx |/wipe -vehicle it/he Prox-car-Prox 'He is wiping the car.' ">al -cp -0 FCR -J/wipe -it 'It has been wiped.' 9 ai -cp -ai-0 ti- qxwmtimut-tx FCR -J/wipe -vehicle-it Prox-car-Prox 'The car has been wiped.'

On the other hand, the language contains a number of roots which, while not bound to a single referential affix, nevertheless do not occur without one. ci-']/ to extend' as in: (15 a) b)

(16 a)

b) (17 a) b) (18 a) b) (19 a)

ci -ak -m -ic ]/extend -hand -ContraPass -it/I 'I reached for it.' *ci-ic ti-suxa-c-tx J/extend-it/I Prox-j/hand-I-Prox HI- '|/to pass by' m -us -cxw J/pass by -face -me/you 'You passed me by.' *ni -cxw J/pass by -me/you nu -il -u:c -cinu central -|/pass by -door -you/I 'I passed by your door.' -ic ti -numucta -nu-tx |/pass by -it/I Prox -door -your-Prox pi- 'J/to lack' pi -i:xw -0 J/lack -head -he 'He is headless.' *pl -0 ti -t'nxw -s -tx |/lack -he Prox -head -his -Prox pi -ak -c J/lack -hand -I 'I lost my hand.'

Lexical morphemes in Bella Coola

b) (20 a) b)

*pl ylack qui penis 'Have *pl |/lack

293

-c ti -suxa -c -tx -I Prox -|/hand -my -Prox -pi -nu -a -|/-lack -you -Interrog you no penis?' -nu -a ti -xli -nu -tx -you -Interrog Prox -|/penis -you -Prox

One of the characteristics of these 'bound roots' that differentiates them from predicative prefixes is that constructions containing them never contain other morphemes identifiable as roots. Thus one never obtains forms such as (21) (22)

*pl ylack 'He is *pl |/lack 'I lost

-t'nxw -0 -]/head -he headless.' -suxa -c -1/hand -I my hand.'

or (23)

*pl -xli -nu -a ]/lack -|/-penis -you -Interrog 'Have you no penis?'

Between free and bound roots is the set of roots displaying intermediate degrees of freedom or boundedness. The root kal- 'to do correctly/be successful' occurs with referential affixes as in: (24) (25)

kal J/do correctly 'I did it right.' kal |/be successful 'He will find a

-ak -m -ic -hand -ContraPassive -it/I -ai -m -is -foot -ContraPass -it/he trail.'

and occurs with grammatical affixes in the absence of referential affixes as in: (26)

kal -liwa -nu ]/do correctly -Sim -you 'You are graceful.'

294

Ross Saunders and Philip W. Davis

(27)

kal -tmaxw -aw J/be successful -Recip -they 'They met/found each other.'

but the root is not totally free as it cannot occur with simple pronominal desinences. (28 a) b) c)

*k:al-c ]/do right-I *kal-nu j/do right-you *kal-0 |/do right-he

Predicative prefixes Another set of lexical morphemes with predicative function is the set of predicate prefixes. They are precisely identifiable by their co-occurrence in constructions with the other class of lexical morphemes having predicative function — roots. The roots may be independent or part of lexical stems as in tam- 'to make/manufacture'. This prefix occurs before a STATIVE predicate and its specific gloss varies according to the nature of the predicate. (29) (30)

(31)

(32)

tam -swata -0 make -|/sweater -she 'She is knitting a sweater.' tam -lulusta -0 make -J/mask -he 'He is carving a mask.' tam -sui -it make -J/house -we 'We are building a house.' tam -cla -nu -a make -J/basket -you -Interrog 'Are you weaving a basket?'

xi- 'to have/possess' also occurs only before (33)

(34)

xi -wac -c have -J/dog -I 'I have a dog.' xi -q lsxw -0 have -|/-rope -he 'He has a rope.'

STATIVES.

Lexical morphemes in Bella Coola

295

ns 'to eat' (35)

ns -">at' -c eat -j/herring roe -I 'I eat herring roe.' 9 is -kusi -c eat -j/potatoes -I 'I eat potatoes.' 9 is -yaik -0 x-a -sqat eat -|/be too much -he Prep -Prox-meat 'He ate too much meat.' n t r 'to wear' 9 it -nup -nu -a wear -j/-shirt -you -Interrog 'Are you wearing a shirt.' Vtr 'to speak a language' 9 it -nuxalkmx -aylayx -nu -a speak -J/Bella Coola -NC -you -Interrog 'Can you speak Bella Coola?'

(36)

(37)

(38)

(39)

Table 1 summarizes the co-occurrence relations between the various kinds of lexical affixes and root types. Table 1. STATIVES

Summary of co-occurrence only

predicate Prefixes

only

STATTVES a n d ACTIONS

ACTIONS

Anatomical Referential Prefixes Referential Suffixes

Non-anatomical Referential Prefixes

Predicative prefixes occur only with STATIVE root stems because the prefix represents the EVENT and the STATIVE root with which it occurs represents a PARTICIPANT in that EVENT. The relationship can be seen in similar sentences where the PARTICIPANT occurs independently. Compare for example, examples (35), (40), and (41). (35)

9

is -Vat' -c eat -J/herring roe -I 'I eat herring roe.'

296

Ross Saunders and Philip W. Davis

and (40)

knix -ic ti -9at' -tx |/eat -it/I Prox -j/herring roe -Prox 'I am eating herring roe.'

(41)

9

or atps -c x -ti -fat' -tx J/eat -I Prep -Prox -J/herring roe -Prox 'I had a meal of herring roe.'

While (40) and (41) speak to specific instances of eating herring roe, (35) is a timeless expression of the EVENT. The referential prefixes which occur only with ACTIONS appear anomalous with regard to the remainder of the referential affixes which co-occur with both ACTIONS and STATIVES.

Problem cases To this point, the decision whether a form is a root or predicative prefix has been fairly easy. We now consider three morphemes where the distinction is not as clear cut. kii- 'to lack'. This is the predicative prefix correlate of the root pi- 'to lack'. (42 a) b) (43 a) b) (44 a) b)

-t'nxw -0 kit lack -J/head -he 'He is headless.' *kii -i:xw -0 lack -head -he kit -suxa -c [kis-suxa-c] lack -|/hand -I 'I am missing a hand.' *kii -ak -c lack -hand -I kit -xli -nu -a lack -J/penis -you -Interrog 'Have you no penis?' *qui -kii -nu -a penis -lack -you -Interrog

However, unlike the other predicative prefixes listed above, all of which occur in construction with STATIVE roots, kii- prefixes ACTION roots as in:

Lexical morphemes in Bella Coola

(45) and (46)

297

Hit -pi:x -ic lack -J/steam -it/I 'I have nothing to steam.' Hit -X'ap -0 lack -J/go -it 'There is no-one to go.'

In this regard, Hit- behaves somewhat like the referential prefixes listed above which represent the semantic content of PARTICIPANTS in ACTIONS. Consider as well txw- 'to go (in)to' in (47)-(51). (47) (48) (49) (50) (51)

txw -stuwa -nu -a go (in)to -1/store -you -Interrog 'Are you going to the store?' txw -X'uK-als-m go (in)to -J/be high-inside-ContraPassive 'He is going upstairs.' fas -txw -c Located -go (in)to -I 'I am inside.' 9us -lxw -c Direction -go (in)to -I 'I am going inside.' nx -?us -txw -c Distributed -Direction -go (in)to -I 'I entered many houses.'

-0 -he

In (47) and (48), txw- appears to be a predicative prefix of the type tam-, xi- etc. It prefixes a STATIVE root and represents the EVENT with the root identifying a PARTICIPANT in the EVENT. In (49) through (51), however, txw- occurs in constructions where it appears to be prefixed by another predicative prefix, 9as- 'located at' or 9us- 'directed to'. In other words, it looks very much like a root in these constructions. If we follow this path, we then reflect that (47) and (48) have predicates containing two roots. Another option is to consider that txw- is a predicative prefix in (47) and (48) but a referential suffix in (49) —(51). In this alternative, 9as- and 9us- then become roots. Aside from the anomalous behavior of txw-, this is an attractive solution as it solves what would have been a problem in a large number of otherwise rootless forms.

298

Ross Saunders and Philip W. Davis

9

as- 'to be located.'

(52) (53)

9

as -ank -0 ¡/located at -front -he 'He's in front.' 9 as -nalus -a:x-il ^/located at -between -trees-we 'We are between the trees.'

Compare (52) and (53) with: (54) (55)

9

us -ank -am -0 [/directed at -front -Devi -he 'He's going in front.' 9 us -nalus -a:x -am -it [/directed at -between -trees -Devi -we 'We are going to a spot between the trees.'

9

as- as a root in (52) and (53) is bound to referential affixes. In this regard it parallels ci- 'extend' and pi- 'lack'. In addition to these sentences there are a large number of body parts whose names consist of 9as- plus an array of anatomical referential affixes. These include: 9as-als-i:xw 'buccal cavity', 9as-a:ls-ik-an 'meatus of ear' and 9as-ik-lqs-ak 'middle finger'. Likewise many points of spatial and geometric reference are constructed with 9as-. These include forms such as: 9as-an 'corner', 9asals-a:q 'inside bottom of a container', and 9as-i:xw 'head of valley/river'. If 9as- is not analyzed as a root, then Bella Coola has hundreds of lexemes that are rootless. On the other hand, if we assume that 9as- is a root with the semantic content 'to be located at', we must now consider sentences such as (56 a - c ) . (56 a) b) c)

9

as -su:m -c [/located at -J/pants -I 'I am wearing pants.' 9 as -qayt -0 ylocated -]/hat -he 'He is wearing a hat.' 9 as -X'ayal -apsm -nu -a [/located -J/necklace -throat -you -Interrog 'Are you wearing a necklace?'

The paraphrase that gets one from the semantic content 'to be located at/on' to that of 'wearing' is, in the case of (56 a), 'Pants are located on

Lexical morphemes in Bella Coola

299

me.' For (56 b), it is 'A hat is located on him.' and for (56 c) 'Is there a necklace located around your neck?' That is to say, there is no semantic reason to assume immediately that there are in fact two 9as- morphemes here; one meaning 'to be located' and the other 'to wear.' The latter can be seen as a reasonable semantic extension of the former. Our problem now is that we have predicates containing two roots.

Possible solution A possible solution is to see 9as- as being in the process of semantic change wherein its general locative meaning is narrowed to 'to have on/to be wearing' and accompanying this narrowing of meaning is a concomitant shift from root to predicative prefix. This could be seen as a part of a pattern which includes the other problem morphemes as involved in similar shifts, i. e., 9as- Bound root to Predicative prefix with narrowing of semantic content from general location to location on body, i.e., 'wearing'. (57) (58)

as -ank -c j/be located -front -I 'I'm in front.' 9as -su:m -c have -|/pants -I 'I'm wearing pants.' 9

Then kii- represents the form class shift from Predicative prefix to Referential prefix with the semantic shift from the STATE of lacking something to the STATE of being the thing lacking. (43 a)

(46)

Kit -suxa -c lack -]/-I hand 'I am missing a hand.' M -X'ap -0 no-one -J/go -he 'There is no-one to go.'

and finally, txw- exemplifies a shift from Predicative prefix to Referential suffix with semantic shift from the ACTION of moving into or toward a goal to the inside/interior itself.

300

Ross Saunders and Philip W. Davis

(59)

(60)

(61)

txw -stuwa -c go to -J/store -I 'I'm going to the store.' mn -txw -uc -cinu ]/dip -in -liquid -you/I 'I'll dip water for you.' ?as -txw -c |/be located -in -I 'I'm inside.'

We see in Bella Coola that the apparent opposition of root — affix is better viewed as a gradient from one pole to the other. Examples (17) — (20) show a first step in the evolution from root to affix in those bound roots which must occur with a lexical suffix. The so-called 'rootless' forms (e.g., 52 — 55) can be viewed as the complementary first step from affix to root when the 'affix' appears without the requisite co-occurring root. Thus the 'rootless forms' are but one point on the scale from 'root' to 'affix'.

Notes 1. Bella Coola is an isolate of the Salishan language family spoken in the Bella Coola Valley of central, coastal British Columbia. We would like to acknowledge those who have helped us to an understanding of their language: David Moody, Charlie Nappi, Andy Schooner, Annie Schooner, Margaret Siwallace, Charles Snow, and Felicity Walkus — all deceased. The work reported on here was supported in part by National Science Foundation Grants SOC 73-05713 A01 and BNS 73-05713 A02, Social Science and Humanities Research Council of Canada Grant 410 — 79-0297, and a grant from from the British Columbia Provincial Museum. 2. The literature on Salishan 'lexical suffixes' is extensive. Of general interest are Stanley Newman (1968) and M. Terry Thompson (1974). Articles specific to Bella Coola lexical suffixes are Davis and Saunders (1973) and Saunders and Davis (1975 a, 1975 b, and 1975 c). 3. The terms printed in capitals are those relevant within a semantic/cognitive system. Space constraints preclude specific elaboration of the categories; but interpretation of them in their common senses will not lead the reader far astray. Generally, DOMAINS of knowledge/memory are "selected" to compose a NARRATED EVENT, with the PARTICULARS of this selection being formed into a PROPOSITION that reflects the contingencies of the conversation — what the speaker and listener know to that point and what they each may be assumed to be aware of. Among the terms of the PROPOSITIONAL formation are the EVENT and the ROLES (AGENT, EXECUTOR, EXPERIENCER, and IMPLEMENT) that the PARTICIPANTS (i. e., PARTICULARS cast in ROLES) will fulfill. This frame is explained in some detail in Saunders and Davis (forthcoming).

Lexical morphemes in Bella Coola

301

4. The forms ti-...-tx are elements of the deictic system of Bella Coola. The system is moderately complex and for masculine and feminine gender in the singular and a plural number (that merges gender distinctions) ranges across a proximal, middle, and distal space-time, plus a nondemonstrative-demonstrative contrast within each. The system has no direct correlate to any English usage, and the forms will be glossed here as 'the' with no additional comment (see Davis and Saunders 1975 a, 1975 b, 1976 for details). 5. Some of the abbreviations used in the analytical representation of sentences are: AntiPass = Antipassive, FCR = Full Control Resultative, NC = No Control, Interrog = Interrogative, ContraPass = Contrapassive, Sim = Simulative, Recip = Reciprocal, and Prox = Proximal, Prep = preposition, and Devi = Developmental.

References Davis, Philip 1973 1875 a 1975 b 1976 Kinkade, M. 1983

W.—Ross Saunders Lexical sufiix copying in Bella Coola. Glossa 7:23 — 52. Bella Coola nominal deixis. Language 51:845 — 58. Bella Coola deictic usage. Rice University Studies 61:13—35. Bella Coola deictic roots. UAL 42:319 - 30. Dale Salish evidence against the universality of 'noun' and 'verb'. Lingua 60:25-40. Newman, Stanley 1968 A comparative study of Salish lexical suffixes. (Paper presented to the III International Conferences on Salish Languages). 1969 Bella Coola grammatical processes and form classes. IJAL 35:175 — 79. 1969 Bella Coola paradigms. IJAL 35:299 - 306. 1976 Salish and Bella Coola prefixes. IJAL 42:228 - 4 2 . Saunders, Ross—P. W. Davis 1975a The internal syntax of Bella Coola lexical suffixes. IJAL 41:106—13. 1975 b Bella Coola referential suffixes. IJAL 41:355 - 68. 1975 c Bella Coola lexical suffixes. Anthropological Linguistics 17:154—89. Forthcoming A Grammar of Bella Coola. Manuscript. Victoria, British Columbia: British Columbia Provincial Museum. Thompson, M. Terry (ed.) 1974 Herman Haeberlin's distribution of the Salish Substantival [Lexical] Suffixes. Anthropological Linguistics 16:319 — 350.

Lexical elaboration in Navajo Robert W. Young

Abstract Navajo culture has borrowed freely from surrounding peoples over the course of some four centuries, but loanwords have not accompanied alien cultural elements and institutions. Fundamental differences between Navajo and the neighboring languages impose barriers that effectively prevent the integration of extraneous verbs into Navajo, and discourage the integration of extraneous nouns and other word classes. Consequently, Navajo has long relied on internal resources for the generation of new vocabulary. Navajo verb and noun morphology are analyzed and the formulae applied for lexical derivation are described.

For the Navajo the past four or five centuries have been marked by widespread cultural borrowing, first from the Puebloan peoples with whom the ancestral Navajo first came in contact following their arrival in the American Southwest, and subsequently from Spanish, Mexican, and Anglo-American sources. It has included the institution of a sophisticated religious-philosophical-ceremonial system, the adoption of a new economic base, first in the form of agriculture and stockraising and, in recent times, wagework and professional pursuits; revolutionary changes in mobility beginning with the acquisition of the horse and followed by an assortment of vehicles; the introduction and spread of institutionalized training for the young; the rise of tribal nationalism accompanied by the adoption of a tribal governmental organization patterned on AngloAmerican models; the advent of modern medicine, with a complex of new concepts relating to the cause and treatment of disease and, of course, a wealth of tools and artifacts. A cursory analysis of the noun inventory, as represented by about 6600 nouns (exclusive of placenames and ceremonial-religious terms), reveals the fact that well over one third are labels attached to elements, institutions or concepts borrowed from Spanish, Mexican, or Anglo-American sources. Borrowed elements and institutions have enriched Navajo culture greatly across the years, and the language has kept pace with cultural growth, historically. However, enrichment of the language owes little to

304

Robert W. Young

linguistic borrowing, for Navajo, on a par with its sister languages, has responded to the demands for linguistic change by drawing primarily on its own internal resources; Spanish and English loanwords integrated into the language historically aggregate little more than fifty terms, all of them nouns. Navajo is an Athapaskan language, and more than a half century ago Sapir (1921:196) observed that "the Athabascan languages of America are spoken by peoples that have had astonishingly varied cultural contacts, yet nowhere do we find that an Athabascan dialect has borrowed at all freely from a neighboring language. These languages have always found it easier to create new words by compounding afresh elements ready to hand." Haile (1941:1) echoed the same observation twenty years later when he wrote to the effect that "Pueblo contact has not influenced Navajo to a noticeable degree, while Spanish elements in the language are comparatively few, and English elements practically none." Young and Morgan (1943:16) listed a small number of borrowed nouns in The Navajo Language, with the notation that "despite the long association of the Navajo with various European groups, word borrowing has been very light, probably due to the dissimilarities that exist between Navajo and the Indo-European languages. A few words have come in, largely from Spanish." World War II gave impetus to a previously lethargic trend toward Navajo-English bilingualism. Spurred by service in the armed forces, by off-reservation work experience, and by expansion of the reservation school system, bilingualism, in turn, gave rise to a growing trend toward the incorporation, but not necessarily the integration, of a substantial number of English loanwords, largely in the form of nouns, into the speech of Navajo adults who had learned English, and this development was reflected in the speech of their monolingual children. Holm, Holm, and Spolsky (1971:3) observed that "the children now beginning school are the first generation most of whose parents were educated in Englishmedium schools." A 1971 study of the speech of six-year-old Navajo children, conducted by the University of New Mexico and based on over 200 taped interviews listed 508 English loanwords, including 453 nouns and 26 phrase-embedded verbs in the speech of monolingual children involved in the study. Nearly all of the borrowed nouns were in the form of labels attaching to the outside culture from which they were borrowed, although one child chose to refer to his younger sister as 'shilittle sister' in lieu of shideezhi,

Lexical elaboration in Navajo

305

and some of the children used the English, rather than the Navajo, number system. The few English verbs were uninfected, phrase-embedded elements of the type 'jump' in jump 'iyiilaa 'he made a jump', a statement that could as readily have been phrased as dah neeshjiid (Holm, Holm, and Spolsky 1971:6-7). Today, in 1987, Navajo bilingualism is widespread and language mixing among bilinguals is of common occurrence, especially in urban or other locations where English is dominant, and in the context of informal conversation. Mixing, however, remains distinct from integration so far as most of the extraneous terms are concerned, and the loanwords are more often than not complemented by terms generated, as in the past, by the application of purely Navajo derivational formulae, utilizing internal resources. An exception probably attaches when the subject matter includes highly specialized elements that are not shared generally by the Navajo speaking public — areas of concern where most speakers would use English anyway (e. g. psychology, chemistry, higher mathematics). Sapir made the observation that the speakers of the Athapaskan languages "have always found it easier to create new words by compounding elements ready to hand", and Young-Morgan attributed the paucity of extraneous loanwords in Navajo to "dissimilarities" that obtain between Navajo and neighboring languages. Sapir did not explain why the creation of new words was 'easier' than borrowing, and Young-Morgan did not identify the 'dissimilarities' involved. Phonological difference aside, an examination of the structure of the Navajo language, albeit superficial, suffices to identify some of the major areas of dissimilarity betweeen Navajo, on the one hand, and the surrounding languages on the other, that impose constraints on lexical borrowing; and at the same time to explain why it is easier 'to create new words by compounding elements ready at hand' than to borrow terminology along with cultural loans. Upon examination it becomes quite apparent that the verb occupies a position of central importance, and that it differs radically from the same word class in Spanish and English. The Navajo verb is a composite construction, closely paralleling a simple sentence in Spanish or English. It is based on a monosyllabic stem which functions as the surface representation of an underlying root. There are about 550 verbal roots, expressed lexically in the form of about 2100 derived stems which, in combination with a small corpus of non-verbal roots attaching to other word classes (stem nouns, postpositions, pronouns and various particles), constitute the foundation upon which the language is built. The verbal roots convey

306

Robert W. Young

meaning in general terms, and the derived stems, as inflected forms of the root, retain the root meaning but add grammatical information relating to mode and aspect. The verbal root inventory, here represented for purposes of illustration by a Perfective (mode) stem form, includes two broad semantic categories: (1) those whose meaning is focused on a narrow conceptual range as yqq' 'ingest food, eat', ghqq' 'kill plural objects', and cha 'cry, weep'; and (2) those whose meaning is broad, often to the point of abstraction, as wod 'flex, bend', kaad 'expansiveness, spreading, scattering', and 'q, basically concerned with the movement, usually by continuing manual contact, of an object class identified conventionally as 'solid roundish', although the derived stems function lexically with reference to objects and concepts that are not distinguished by these physical characteristics. The verb stem functions as a component of a meaningful lexical unit when it is preceded by essential prefixes that mark subject, direct object (if the verb is transitive), number and potentially mode and aspect. In addition, most verbs include adverbial prefixes that serve to modify the broad meaning conveyed by the stem and derive specialized lexical meaning. For example, nl'q 'I arrived carrying it, I brought it (a solid roundish object)' is a simple lexical form with the literal meaning 'I caused it to move terminatively'. With the addition of adverbial elements this basic meaning is redirected, as ch'i- 'out horizontally' + -ni'q 'I carried it out', 'alts'a- 'away from each other, apart' + -ni'4 'I divided it in two parts', ni- 'cessation' + -ni'4 'I stopped carrying it, I set it down', and ni'cessation' + ni-, generic classifier for roundness + -nisht'q 'I laid my head down', literally 'I laid down my own solid roundish object, moved my own solid roundish object terminatively'. Type (1) roots offer little latitude for lexical derivation, but those of type (2) are generally highly productive. The root wod 'flex, bend', for example, with its stem derivatives wod, wo', wol, wood, yeed ( < ProtoAthapaskan *ghut' 'flex, bend)1, serves as the foundation for more than 130 primary lexical derivatives, including verbs with such disparate meanings as nishwod 'I arrived running' (verbs that describe the act of running, on the part of a single subject, are based on the root wod and refer literally to the flexing action of the legs of the runner), 'ahaawod 'it became disjointed', k"rnilhod 'I broke it in two by flexing it,' naa'iiwod 'it toppled over (as a tree in the wind)', bighalwod 'it went through him (a bullet or arrow — a secondary derivation based on 'run' in the extended sense 'move swiftly')', yiwod 'he is rocking or swaying', yilhod 'he is pumping it,' literally 'he is rocking it (as the handle of an old-fashioned

Lexical elaboration in Navajo

307

pump)', na'nilhod 'he is limping about', and in modern usage the English caique yiyoolwol 'he is running it (a business, school, train)'. Similarly, the root kaad 'expansiveness, spreading, scattering,' with its derived stems kaad, kad, ka\ kal ( < Proto-Athapaskan *kat' 'be spread out, fall scattered, slap, sew')1 appears in such disparate lexical derivatives as nanishkaad 'I'm herding them, driving them about (as a herd of sheep, the components of which move as a scattered mass)', hdiikaad 'I gave him a slap (action involving the movement of a flat object),' naa'aniikaad 'it collapsed, fell over (scattering, as a stack of dishes or a masonry wall),' naakaad 'it spreads (as a plant)', 'adishka' 'I play cards, gamble (repeated movement involving flatness)', 'aheshkad 'I'm applauding, clapping my hands (literally I am moving flatness together against each other)', 'ahldiilkad'l sewed them together (literally I joined flatness against each other)', and sikaad 'it stands or lies spreading (a bushy tree, a rug, a thin expanse of water)'. Well over 100 primary lexical derivatives employ the verbal root kaad. And the highly productive root 'q 'move a solid roundish object', with its stem derivatives 'q, 'a, 'a', 'aah, 'aah, 'aal, 'aal ( < Proto-Athapaskan *'a-n y , classificatory stem for a compact object) 1 yields well over 200 primary lexical derivatives, ranging from verbs that describe the movement of the sun to those concerned with the handling of a broad class of objects (one bead, seed, berry, coin, book, knife, hat), verbs relating to song and news, and neuter verbs that describe an object of the 'solid roundish' class at rest. Thus: ha'ii'q2 'the sun came up' (literally 'something roundish climbed up'), 'i'ii'q2 'the sun set' (literally 'something roundish moved away out of sight'), 'alni'ni'q2 'it became noon' (literally 'something roundish moved half way'), ch'ini'q 'I carried it out,' hadii'4 'I started a song' (literally 'I started to move oral noise'), hane' bitaase'q 'I carried the news among them,' si'q 'it sits, lies, is (in position, as a book on the table)'. Verbal actions, such as those represented by the separate, non-interrelated verbs carry, lift, bring, insert, extract, put, separate, be (in position), keep, in English, are represented in Navajo as directional or stative facets of the meaning of a shared root such as 'q 'move a solid roundish object'. Carry = 'move along', lift = 'move upward from a surface', bring = 'move terminatively to a destination', insert = 'move into', extract = 'move out', put = 'move into position', separate = 'move apart', be (in position) = 'lie at rest (as the sequel to cessation of movement)', and keep = 'cause to lie at rest'.

308

Robert W. Young

The approximately 550 verbal roots, and especially those whose meaning is broad and general in scope, are capable of generating terminology that is eminently adequate to communicate the full range of actions, events, relations, attributes, conditions, and states of being in the world of human experience. The derivation of terminology, old and new, involves primarily the identification of an appropriate root and stem complex, and the choice of appropriate adverbial-derivational and paradigmatic prefixes with which to construct a required lexical form; or, secondarily, extension of the meaning of an applicable existing lexical unit. There is no evidence that Navajo has ever attempted to incorporate a verb from any neighboring language, in integrated form, to function after the fashion of a root or stem, and to be inflected after the manner of a Navajo verb. The parameters or basic principles that govern the Navajo verb constitute effective barriers against borrowing from non-Athapaskan language sources, so far as this word class is concerned, although loan verbs from related (Athapaskan) languages could be — and perhaps have been — accommodated by Navajo, historically. The verb lexicon generated around the transportational complex that has developed since acquisition of the horse some 300 years ago, and around the more recent institution of an elective tribal government, illustrate the formulae and procedures involved in the elaboration of new vocabulary with which to communicate actions and events relating to these innovations and, by extension, the formulae and procedures that govern verb derivation generally. Acquisition of the horse was followed by a succession of vehicles including wagons, buggies, automobiles, trucks, tractors, bicycles, motorcycles, boats, trains, and airplanes and, in an extended sense, elevators, escalators, tramcars, Ferris wheels, and carousels. The choice of an appropriate verbal root or roots to describe the use of each of these conveyances first involves determination of the manner or manners in which the conveyance is capable of moving, and a lexical derivative may need to reflect the relationship that is to exist between the conveyance and the person or persons conveyed inasmuch as the user's function may be either active or passive — i.e. he may be either a passenger or an agent. A wide range of verbal roots are relatable to the movement of the several types of conveyance, including ya 'single subject goes' (including an animal at a walk, or a car or truck at a slow or unspecified rate of speed), wod 'single subject runs or moves rapidly' (including an animal at a run, or a fast-moving vehicle as a car, bicycle, motorcycle, train,

Lexical elaboration in Navajo

309

bus), dloozh 'move on all four' (any quadruped, as a horse, burro, donkey), te' 'dash, race' (a horse), tqq' 'lope, gallop' (a horse), 'eel 'float, move on or within water' (a boat, raft, submarine), t'a' 'fly' (an airplane), 'na 'crawl' (a caterpillar tractor, army tank), bqqz 'hooplike object rolls' (a wheeled vehicle, as a wagon, car), maaz 'spherical object rolls' (a barrel, a gourd in which a sorcerer is said to travel), t'i 'move swiftly' (a car, bicycle, motorcycle), taal 'dash, accelerate' (animal, motorized vehicle), zhood 'move in a heavy dragging manner' (a sleigh), ch'qql 'move suspended' (a passenger elevator, cable car, parachute, hot air balloon), do 'move by wafting, floating in air' (an escalator, hot air balloon), deel 'slender flexible object moves abruptly and independently through the air' (a rainbow, sunbeam, lightning, in mythology). The formulae applied for the derivation of verbs relating to transportation fall into three categories, most with a full range of directional derivatives for each root (i. e. go out, go in, go up, go down, go around, go away, go back, go again). The categories include: Passive transportation. The derivational formula requires a comitative construction in which the subject of the verb is the conveyance, which moves 'in company with' the passenger as indicated by the postposition -il 'with, in company with.' The vehicle may be represented by a noun, in which event the subject of the verb is a definite 3rd person; or the vehicle may be unnamed but represented as an unspecified animal or thing by 'a- 'someone or something unspecified' as the pronoun subject of the verb. Thus ![[' shil yildloozh 'I arrived horseback' (literally 'horse arrived moving on all four with me'), shil 'ildloozh 'I arrived horseback' (literally 'something arrived moving on all four with me'), telii bil niya 'he arrived riding on a burro' (literally 'a burro came with him'), bil 'aniya 'he arrived (by slow conveyance)', chidi shil yilwod 'I came by car' (literally 'a car arrived running or moving swiftly with me'), shil 'ilwod 'I arrived by fast moving conveyance (as a car, bicycle, train)' (literally 'something came running or moving swiftly with me'), chidi naat'a'i shil yit'a 'I arrived in the airplane' (literally 'an airplane arrived flying with me'), shil 'it'a' 'I came by air' (literally 'something came flying with me'), tsinaa'eel shil ni'eel 'I arrived in the boat' (literally 'the boat arrived floating with me'), shil 'ani'eel 'I came by boat' (literally 'something arrived floating with me'), tsinaalzhoodi shil yilzhood 'I arrived in the sleigh' (literally 'a sleigh arrived sliding with me'), shil 'ilzhood 'I came by sleigh' (literally 'something came sliding with me'), tsinaabqqs shil nibqqz 'I arrived in the wagon' (literally 'the wagon arrived rolling with

310

Robert W. Young

me'), lii'/dzi'izi/chidi shil yaaltaal 'I dashed off on a horse/on a bicycle/ in a car' (literally 'a horse/bicycle/car dashed off with me'), shil ya'iiltaal 'I dashed off or accelerated (by any of the foregoing conveyances)' (literally 'something dashed off or accelerated with me'). A department store elevator is variously designated as hoi haadaalch 'qligH 'that which repeatedly ascends suspended with people', hoi haadaalwo'igii 'that which repeatedly ascends rapidly with people', or hoi haadaalzho'igii 'that which is repeatedly dragged up with people', and a person who takes an elevator to an upper floor may refer to the action as hoi haadaalch

'qligiijhol

haadaalwo

'igii/hol

haadaalzho

'igii shil haashch

'qql/

'I went up in the elevator', or if the conveyance is not named he may say simply shil ha'ashch 'qql¡shil ha'aswodjor shil ha'ashzhbod 'I took an elevator up' (literally 'something climbed up suspended with me/something climbed up swiftly with me/or something was dragged up with me'). A passenger on an escalator seems to float in air as he is conveyed up or down, a feature that led to use of the root do 'waft or float through the air', to describe the action, as hoi haddaaldohigii 'that which repeatedly wafts up with people' shil haasdo 'I went up on the escalator' or if the conveyance is unnamed shil haasdo 'I went up by escalator'. The distinctive feature attaching to a Ferris wheel and a carousel is the fact that it revolves at a relatively slow rate of speed. Movement of this type is expressed by the root yiz (and the derived stems yiz, yees, yis). A Ferris wheel is identified as hoi ndeiyesigii 'that which turns around with people', and a rider describes his ride as hoi ndeiyesigii shil naayis 'I am riding on the Ferris wheel (or carousel)', or if the vehicle is not named shil na'ooyis, 'I am taking a Ferris wheel ride' (literally 'something is turning around at a slow rate of speed with me'). The root deel includes the meaning 'slender flexible object moves independently (flies) through the air', and in mythology rainbows, lightning, and sunbeams transport cultural heroes, usually to mountaintops. Their swift movement is described by the root deel, as in naats'iilid Tsoodzil bilatahjf hoi dah yizdeel, 'a rainbow whisked him to the top of Mount Taylor' (literally 'a rainbow sped with him to the top of Mount Taylor'). The root na' describes the independent movement through the air (flying) of a flat flexible object. If, in a Navajo translation of the Arabian Nights there were an account of travel by flying carpet, the translator might choose na' to describe the action involved: 'he flew away on a flying carpet' beeldlei naat'a'i bil yob' 'una'. haaswod/haashzhood

Lexical elaboration in Navajo

311

Travel by any conceivable means can readily be described by drawing upon an appropriate root that describes the action in terms of the physical characteristics of the vehicle involved, the rate of speed, manner of movement or other distinctive features. Agentive transportation. With reference to inanimate vehicles of transportation some (as a wagon, boat, car, or airplane) may be represented as caused to move by an agent in the form of the person transported. This extended meaning is produced by application of an established formula for causative-transitivization. The formula is not applicable if the 'vehicle' is animate (as a horse or burro), possessing a will of its own, nor is it universally applicable to the motion stems. However, transportation verbs based on the roots bqqz 'hooplike object moves by rolling (a wheeled vehicle)', 'eel 'move by floating', and t'a' 'move by flying' are commonly transitivized to represent the person transported as the driver or pilot who causes the movement. Thus tsinaabqqs nilbqqz 'I drove the wagon, I came by driving the wagon', tsinaa'eel nil'eel 'I came by rowing or sailing the boat', chidi naat'a'i nilt'a' 'I came by flying the plane', and 'anilbqqz 'I drove, I arrived by driving an unspecified wheeled vehicle', 'anil'eel 'I came by boat, I arrived by causing an unspecified floating conveyance to arrive at a destination', and 'anilt'a' 'I flew, I caused an unspecified flying conveyance to arrive at a destination'. An indirect approach is taken by verbs of transportation that are based on the root lo', whose area of meaning includes 'move suspended by a cord or handle (as a bundle carried by means of a string or a purse by its handle)', and simply 'act by means of a cord or rope (as in snare, lasso, hang up)'. Verbs of transportation derived from the root lo' were first used with reference to the driving of a horse-drawn vehicle, where the agent controlled the movement by means of the reins ( = a cord). A verb such as haselo' acquired the dual meaning 'I climbed up carrying it (by its handle)' as in bikoohd§§' to bee naakahi haselo', 'I climbed up out of the canyon carrying the waterbucket', and 'I drove it up out', as in bikoohdgg' tsinaabqqs haselo', 'I drove the wagon up out of the canyon'. With the advent of motor-driven vehicles the meaning of lo' and its transportational derivatives was easily extended in the general sense 'drive a guided vehicle (car, boat, snowmobile, wagon)' despite the fact that reins (representing the cord or rope) were no longer involved. In addition, the braking system in use on early wagons was controlled by means of a rope on which the driver pulled to bring the brake shoe

312

Robert W. Young

into contact with the rim of the wheel. Since the act of applying a wagon brake also involved a cord, the root lo' was again drawn upon for necessary vocabulary, as in tsinaabqqs bidiilo' 'I applied the wagon brake' (literally 'I moved it into position against the wagon by means of a rope'), and 'idiilo' 'I put on the brake' (literally 'I moved it into position against something unspecified by means of a rope'). These terms continued in use to describe the application of a wagon brake long after the primitive rope controlled braking mechanism was replaced by a lever and, with the advent of motorized vehicles, bidiilo'/'idiilo' acquired the general meaning 'apply a brake', although again no cord or rope is involved. Agentive/passive transportation. The agent who is represented as causing the movement of a wagon, car, boat or airplane can be described as causing it in company with a passenger, as in chidi bil nilbqqz 'I took (conveyed) him in the car' (literally 'I caused the car to arrive rolling with him'), bil 'anilbqqz 'I drove him, I took him by unspecified wheeled vehicle', tsinaa'eel bil nil'eel 'I rowed him, took (conveyed) him in the boat', bil 'anil'eel 'I took him by boat, I took him by unspecified floating conveyance'. To describe the parking of a wagon or automobile Navajo drew upon a formula involving an adverbial particle dah 'up' in combination with a Progressive mode form which, together, represent the action of the verb as static, as yish 'aal 'I'm carrying it along (a solid roundish object)' / dah yish'aal 'I'm holding it up', 'eeshkggl 'I am swimming along' / dah 'eeshkggl 'I am treading water'. Applied to Progressives derived from the root bqqz the formula expresses the meaning 'be parked', as in tsinaabqqs dah yisbqs 'I have the wagon parked', dah 'eesbqs 'I am parked (in an unspecified car or wagon)'. And to describe the landing of an airplane Navajo drew upon the root da 'single subject sits down,' as in chidi naat'a'i neezda 'the airplane landed' (literally 'the airplane sat down'), chidi naat'a'i shil neezda 'I landed in the airplane' (literally 'the airplane sat down in company with me'), and in unspecified form shil 'aneezda 'I landed' (literally 'something unspecified sat down with me'). With reference to motorized vehicles audibility, represented by the root ts'44' (with stem alternants ts'a', ts'fth, ts'uf) 'make sound, hear, listen', was selected as the distinguishing feature to describe the running and starting of the motor, as in chidi diits'a 'the car (motor) is running,' chidi diselts'qq 'I started the car motor (literally 'the car is making sound', 'I caused the car to make sound').

Lexical elaboration in Navajo

313

In 1921 an elective tribal governmental organization, known as the tribal council, was introduced to the Navajo. The tribe had not been organized politically before, but as this cultural innovation took root a body of specialized vocabulary was soon generated spontaneously. The new items of vocabulary were, for the most part, descriptive terms attaching, not only as noun labels to the organization and its components, but as verb labels to the actions and events associated with governmental activities as well. Verb terminology took two forms: terms based upon root meanings that reflected a salient feature of the action or event, and caiques representing direct translations of English terms — an expedient introduced by English speaking officers and delegates. Among the many concepts that required expression were terms of the type nominate, run (for office), vote, preside, move or make a motion, second a motion, table, approve, disapprove, inaugurate (swear into office), assemble, recess, adjourn. Metal badges were issued to the early council delegates. These were worn pinned to the breast and, since they were made of metal they were designated simply as beesh 'metal'. The wearing of the badges was chosen as the distinguishing feature with which to name the organization as well as the delegates, both of which came to be known as beesh bqqh dah naaznili, literally 'those upon whom metal objects sit up at an elevation' or, with closer focus on the individuals beesh bqqh dah naaz'ani, literally 'those (individuals) upon whom metal sits up at an elevation'. At council meetings the chairman, as presiding officer, sat on an elevated dais, with the result that he came to be designated as 'a dah nanidaahii, literally 'the one who repeatedly sits down up at an elevation for the benefit of someone unspecified' = 'the one who takes an elevated seat and presides'. Legislative bills considered and acted upon by the council, known in English as resolutions, were identified functionally as naaltsoos bee 'add nihodiit'aahigii, literally 'the paper with which planning for self is done'. Candidates for council office were nominated by oral acclaim at conventions held for that purpose so a root dzii' 'breathe', in the form of a lexical derivative meaning 'speak out,' was chosen to represent the concept 'nominate', as in shee ha'oodzii' 'I was nominated' (literally 'someone unspecified spoke out with me') or beesh bqqh dah naaz'ani ya dah nanidaah dooleelgo biniiye shee ha'oodzii' 'I was nominated for council chairman' (literally 'someone unspecified spoke out with me for the purpose of becoming the council chairman').

314

Robert W. Young

The concept 'run for office' came to be expressed by a caique from English, as in beesh bqqh dah naaz'ani biniiye 'atah ndiishwod 'I ran for council delegate' (literally 'among others I leapt up running or started racing for the purpose of council delegate'). Likewise the concept 'move, make a motion,' as applied in a parliamentary sense, was expressed by a direct translation from English, based on the root naa 'move, make a movement (a stationary object, as a tree in the wind, a sleeping person, an earthquake'). Thus dil naaltsoos bee Iq 'adooleel dishniigo bee hidishnaah, 'I move for approval of this resolution' (literally 'by making a movement I'm saying that this paper will be approved'). 'Second a motion' was expressed by ya 'single subject goes,' in a lexical derivative meaning 'stop going, take a position,' as 'akee' niniya 'I followed, took a position behind' in dil naaltsoos bee Iq 'adooleel bee hideesnaigii bee 'akee' niniya 'I seconded the motion that was made for approval of this resolution' (literally I took position behind on the motion that was made for approval of this resolution'). The root tsooz 'move a flat flexible object' was applied for the derivation of t'oo ni' niiltsooz 'it was tabled' (literally 'it was merely set down'). Council delegates take a standing vote on resolutions so the concept 'vote' was rendered by the root jee' 'plural subjects run or lie,' in a lexical derivative meaning 'rise, stand up,' as in bee h'diijee' 'it was voted upon' (literally 'unspecified plural subjects stood up with it'). Council elections are by paper ballot and voting involves an action that is describable by the root 'ah 'toss or abruptly move a single flat flexible object', in speaking of the casting of a single ballot, or by the root nil 'move or toss plural objects', when more than one ballot is involved. The voter says John naaltsoos ba 'ii'ah 'I voted for John' (literally 'I tossed a paper away out of sight for John') or, referring indirectly to the ballot John ba 'i'ii'ah 'I voted for John' (literally 'I tossed an unspecified flat flexible object away out of sight for John') and, with reference to the casting of multiple ballots in an election 'atah naaltsoos 'ahee'nil 'we voted in the election' (literally 'among others we tossed papers away out of sight one after another'). And the inauguration of new council delegates and officers involves taking the oath of office which, in turn, is an action distinguished by the raising of the candidate's hand. The root nii' describes 'action performed by the hand or arm', so yadidiilnii' 'he raised his arm' is equivalent to 'he was inaugurated'.

Lexical elaboration in Navajo

315

The entire verb lexicon, including modern day coinages, is drawn from the approximately 550 unit root reservoir by application of appropriate derivational formulae. And, to a very large extent, the noun lexicon is derived, directly or indirectly, wholly or in part, from the same root reservoir as reflected by the fact that, of about 6600 nouns, roughly two thirds derive directly from verbs, or they are composite constructions that include a verb. And nouns classed as non-verbal include many stem elements that perform a dual function as noun and verb (stem). Verbal nouns, like the verbs from or in combination with which they are derived, are descriptive terms that function, in fact, as extensions of the meaning of the verbal element. Production involves the application of a variety of derivational formulae, the choice of which is contingent on the nature of the noun — i. e. the choice depends upon whether the description is based on function, purpose, appearance, quality, or other criteria. Verbal nouns can be classed in two broad groups, semantically and in terms of derivational formula as: those that serve as labels for tangible or material entities (persons, things, substances), and those that serve to identify ideas, actions, and qualities. The several formulae that govern the production of verbal nouns are identified and illustrated in detail in The Navajo Language (Young — Morgan 1980). Verbal nouns that relate to persons and things range from the ancient to the modem, and from simple third person verbs, usually nominalized, to complex descriptive phrases involving words of two or more classes, but in which the verbal component plays a key role. Compare 'adilgqshii 'wizard, sorcerer' ( < 'adilgqsh 'he shoots evil with his eyes' + -ii, the one), bee'ak'qqsi 'arrow straightener' ( < bee 'by means of it' + 'ak'qqs 'something is straightened' + -i 'the one'), 'atseeltsoii 'yellow tail hawk' (< 'atsee- 'tail' + -Itso- < litso 'it is yellow' + -ii 'the one'), ba ndaa'ii 'the patient in a War Dance Ceremony' ( < ba 'for his benefit' + ndaa 'War Dance' + -ii, the one), bee'atl'ihi 'sling (for throwing missiles)' (< bee 'by means of it' + 'atl'[h 'something is slung' + -i 'the one'), Binaa Yee 'Agha.nl 'the Evil-Eye Monster' ( < Binaa' 'his eyes' + Yee 'he with them' + 'Aghq 'he kills things' + -(n)i 'the one'), 'ani'Uhi 'thief (< 'ani'Hh 'he steals things' + -/ 'the one'), na'albqqsi 'driver (of a wheeled vehicle)' ( < na'albqs 'he drives something unspecified about' + -i 'the one'), chidiltsooi 'schoolbus' ( < chidi 'car' + -Itso- < litso 'it is yellow' + -i 'the one'), 'olta'i 'pupil, school boy or girl' ( < 'olta' 'he/she counts or reads'), ba olta'i 'schoolteacher' ( < ba 'for his benefit' 'olta' 'something

316

Robert W. Young

unspecified is counted or read' + -i 'the one'), 'aheejolyeedi 'baseball game' ( < 'aheejolyeed 'a person runs around in a circle' + -i 'the one'), bee na'alzoli 'spray gun' ( < bee 'by means of it' + na'alzol 'something is blown about' + -i 'the one'), tsinaa'eel 'ship, boat' ( < tsi(n)- 'wood' + naa'eel 'it moves about floating'), leejin 'coal' ( < lee- leezh 'soil' + -jin 'black'), galbahi 'cottontail rabbit' ( < ga(h)- 'rabbit' + -Iba- < liba 'it is gray' + -hi 'the one'), kg' na'aibqqsii 'locomotive' ( < kg' 'fire' + na'albqqs 'it causes something to move about rolling (on wheels)' + -ii 'the one'), naaltsoos bii' yistl'ini 'bookcase' ( < naaltsoos 'books' + bii' 'inside of it' + yistl'in 'they are piled' + -i 'the one'), 'awoo' bee yich'iishi 'toothbrush' ( < 'awoo' 'teeth' + bee 'with it' + yich'iish 'they are scrubbed (abrasively)' + -i 'the one'), nilch'i likoni bee ch'iyaart 'al'ini 'gas stove' ( < nilch'i 'air' + likon 'it is flammable' + -i 'the one' = natural gas + ch'iyaan 'food' + al'[ 'it is prepared' + -(n)i 'the one'), ch'osh doo yit'iinii 'microbe, bacterium' ( < ch'osh 'bug' + doo 'not' + yit'l 'it can be seen + -(n)ii 'the one' = 'invisible bug'), ch'osh doo yit'iinii naalkaah 'bacteriology' ( < naalkaah 'they are tracked, investigated'), ch'osh doo yit'iinii ndeilkaahii 'bacteriologists' ( < ndeilkaah 'they track or investigate them' + -ii 'the ones'), ch'osh doo yit'iinii bee nil'ini 'microscope' ( < bee 'with it' + nil'l 'they are looked at' + -(n)i 'the one'), ch'osh doo yit'iinii bee naatseedi 'antiseptic, bactericide' ( < bee 'with it' + naatseed 'they are killed' + -i 'the one'). Verbal nouns that represent ideas or qualities generally take the form of impersonal verbs in which the subject of the object is represented as an unspecified person or thing ('a- 'someone or something'), or as impersonal 'it', 'things' (ha- ho- 'space, area, impersonal it, things'), as nahagha 'religion, ceremonial system' (literally 'impersonal it goes about, things go about'. Cf. naagha 'he goes about'), 'oodlq 'belief, creed' (literally 'something unspecified is believed'), 'ihoo'aah 'learning, education' (literally 'something unspecified is learned'), na'a'ne 'play' (literally 'someone unspecified plays'), 'alah 'aleeh 'meeting' (literally 'alah 'together' + 'aleeh 'unspecified subjects become'), beeso 'a'i'niil 'money-lending' (literally 'plural unspecified moneys are loaned to unspecified recipients'), ntsekees 'thought' (literally 'thinking takes place' or 'he thinks'), hak'az 'coldness' (literally 'space or impersonal it is cold'), chahalheel 'darkness' (literally 'impersonal it is dark, space is dark'). Non-verbal nouns constitute about one third of the total within a corpus of about 6600 nouns, including simple stems many of which function both nominally and verbally. For the most part the stem nouns represent natural objects, body parts, kinship relations, plants and ani-

Lexical elaboration in Navajo

317

mals, and they are quite widely shared throughout Athapaskandom. Included in this category are terms of the type tó 'water', sg' 'star', lóó' 'fish', tsé 'stone', tsin 'wood', béésh 'flint, metal', lid 'smoke', 'éé' 'clothing', -gaan 'arm', -jáád 'leg', -kee' 'foot', -zid 'liver', -zee 'mouth', -náa 'eye', -woo' 'tooth', -má 'mother', -taa' 'father', -ye' 'son', -tsi' 'daughter', biih 'deer', dzeeh 'elk', shash 'bear', tl'iish 'snake', tl'oh 'grass', c/j 77'plant', ch'ó 'spruce', dláád 'mold, lichen'. Some simple nouns include a prefix as 'azee' 'medicine', 'ásaa' 'pot', diñé 'man, person, Navajo', dichin 'hunger.' And some stem nouns combine with elements of the same or a different word class to form compound nouns, as tókg 7 'oil, kerosene' ( < tó 'liquid' + kg' 'fire' + -i 'the one'), tsékg 'í 'coal' ( < tsé 'stone' + kg' 'fire' 4- -i 'the one'), tsiizis 'head basket' (< tsii- 'head' + -zis 'sack, bag, container'), táyi' 'interior of the water' (< ta- 'water' + -yi' 'inside'), leeyi' 'underground' ( < lee- < leezh 'soil' + -yi' 'inside'). Non-verbal nouns sometimes appear in the form of phrases consisting of two or more components, as béésh bii' kg'i 'stove' ( < béésh 'metal' + bii "inside of it' + kg" fire' + -í'theone'), béésh bii'kg 7 bizool 'stovepipe' (< bizool 'its windpipe'), béeso bá hooghan 'bank' ( < béeso 'money' + bá 'for it' + hooghan 'home'). And some compounds classed as verbal nouns include a non-verbal component plus a verb stem, as leejin 'coal' (< lee- < leezh 'soil' + -jin 'black'), tsénil 'axe' ( < tsé 'stone' + -nil 'pound', originally a stone axe but now transferred to designate a steel axe), tl'ohchin 'onion' ( < tl'oh 'grass' + -chin 'smell'), or a non-verbal component plus a verb, usually in nominalized form, as ch 'osh doo yit 'iinii 'microbe' ( < ch'osh 'bug' + doo 'not' + yit'i 'it can be seen' + -(n)ii 'the one'), 'awoo' bii yich'iishi 'toothpaste' ( < 'awoo' 'teeth' + bii 'in company with it' + yich'iish 'they are scrubbed (abrasively)' + -i 'the one'). The limited inventory of nouns that have been borrowed from Spanish and English and integrated historically into the Navajo language comprise a scattering of terms, most of them culture related, associated with card playing, food, money, numbers, animals, and ethnic designations. Nouns relating to playing cards include borrowed terms of the type 'aldóós 'the deuce' ( < Span, el dos), 'áás 'ace' ( < Span, as), zhéí 'king' ( < Span. rey), sóoda 'jack' ( < Span, sota), Ift' 'queen' (l[i 'horse' is a caique translating Spanish caballo 'horse, queen in cards'), báasdos 'clubs' ( < Span, bastos), góobas or báabas 'hearts' ( < Span, copas), 'óola 'diamonds' ( < Span, oros), 'asbáala or 'esbáada 'spades' ( < Span, espadas), 'alóos 'rice' ( < Span, arroz), basdéél or masdéél 'cake' ( < Span, pastel),

318

Robert W. Young

bilasáana 'apple' ( < Span, manzana), géeso 'cheese' ( < Span, queso), gohwééh or 'ahwééh 'coffee' ( < Span, café), mííl or miil 'thousand' ( < Span, mil), 'óola 'hour' ( < Span, hora), 'óola 'gold' ( < Span, oro), béeso 'money, dollar' ( < Span, pesó), yáál 'bit' ( < Span, real — as in naakiyáál 'two bits'), dóola 'bull' ( < Span, toro), béégashii 'cow' ( < Span. vacas), jélii 'jelly', nóomba 'number', Sooléí 'Salt Lake City', Báshdo 'Barstow'. Inflection, for the Navajo noun, is limited to possession, a distinction that, for most nouns, involves nothing more than the addition of possessive pronoun prefixes to the noun base, as ch'ah 'hat' / shich'ah 'my hat', béésh bii' kg'i 'stove' / shibéésh bii' kg'i 'my stove', béégashii 'cow' / shibéégashii 'my cow', and jélii 'jelly' / shijélii 'my jelly'. Noun inflection places no obstacle in the way of syntactic accommodation of borrowed terms belonging to this word class. Historically phonological differences between Navajo and neighboring languages posed something of a problem, as reflected by the Navajo rendition of Spanish and English nouns such as 'alóos 'rice' for Spanish arroz, zhéí 'king (in cards)' for Spanish rey, yáál 'bit (money)' for Spanish real, gááboleida for English 'carburetor' and Késhmish for English 'Christmas'. However, as we have pointed out, a major constraint on linguistic borrowing stems from the immense chasm that separates Navajo from English or Spanish, as language systems. For Navajo the verb, as a lexical unit, is essentially a facet of the meaning of one of the approximately 550 verbal roots that comprise the foundation upon which the verb lexicon and, by extension, a large part of the noun lexicon, rests. As facets of the abstract meaning of a shared root many verbs, conceptually separate and distinct in Spanish or English, are closely interrelated in Navajo. The verb occupies a central position in Navajo; the Navajo verb system cannot accommodate extraneous elements in integrated form; and most Navajo nouns are generated from verbs or include verbal components. As a result, it is not only easier to generate new terminology by utilizing internal resources and purely Navajo derivational formulae, but mandatory for the verb. Patently, the internal resources are quite adequate to meet future needs as they have met needs in the past. Alice Neundorf, a native speaker of Navajo, long interested in the promotion of vocabulary development holds to the view that "new terminology can be developed by anyone who is a good speaker of the language" (Neundorf 1982:271), a point that she demonstrates quite

Lexical elaboration in Navajo

319

profusely in her illustrated children's dictionary (Neundorf 1983), and in her lectures on word coinage. Extraneous nouns can and do function syntactically in Navajo, especially in the informal speech of bilinguals who find an English term to be a shortcut when the Navajo equivalent is unwieldy or when the subject under discussion is of a technical nature, involving concepts that are not generally shared by the Navajo speaking public. However, so far as unwieldy terminology is concerned, Neundorf points out that with general use, or in contexts where a detailed description is not required, overlong noun phrases tend to shorten. A label such as chidi bidaahgi naabalgo to yineesk'azigii describes the radiator fan in an automobile (literally 'that which cools water by whirling rapidly at the front of the car'), but in contexts or situations where the referent is clear the term is reducible to chidi bidaahgi ndabaligii ('that which whirls at the front of the car') or simply ndabaligii ('the whirler'), and similarly tsinaa'eel biyaadi naabalgo yee naaghahigii ba ni'aii 'ship propeller shaft' (literally 'that which extends horizontally for that with which whirling under a ship it goes about') can reduce, in given situations or contexts to stinaa'eel biyaadi ndabaligii ba ni'aii ('that which extends for that which whirls under a ship') or simply biyaadi ndabaligii ba ni'aii ('that which extends for that which whirls under it'). And, as new institutions and concepts become integral features of Navajo culture, the tendency remains one of generating Navajo terminology in preference to borrowed terms that are more symbolic than intrinsically or descriptively meaningful. Bikaaadani 'dining room table' (literally 'that upon which eating takes place'), and bik'idah'asdahi 'chair' (literally 'that upon which sitting takes place') are in closer harmony with Navajo preferences in the naming of utilitarian objects than are the symbolic nouns 'table' and 'chair', neither of which describes the object in terms of its function. Generally speaking, and despite bilingualism, Sapir's observation with reference to the Athapaskan languages will probably continue to hold true for Navajo in the future as it has in the past. Notes 1. Proto-Athapaskan reconstructions by JefFry Leer and Michael Krauss, University of Alaska. Personal communication. 2. Active verbs derived with the root '4 are transitive constructions. Hence a more accurate literal translation of terms that describe the apparent diurnal movement of the sun would reflect this fact. Constructions of the type ha'ii'q / 'alni'ni'q / 'i'ii'q might

320

Robert W. Young be treated as transitives in which the direct object (rather than the subject) is an indefinite or unspecified solid roundish object represented by the indefinite pronoun prefix 'a-~ 'i-~'-, 'something.' The subject of verbs of this type might be construed as the mythical Sun-Bearer whose function it is to carry the solar orb across the sky each day. Ha'ii'q 'the sun came up', would thus be, in literal translation, 'he carried an unspecified solid roundish object up out;' 'alni'ni'4 'it became noon', would be 'he carried an unspecified solid roundish object half way,' and 'i'i'i'q 'the sun set', would be 'he carried an unspecified solid roundish object away out of sight.'

References Haile, Berard OFM 1941 Learning Navaho, vol. I. St. Michaels, Arizona. Holm, Agnes—Wayne Holm —Bernard Spolsky 1971 English loanwords in the speech of six-year-old Navajo children. Navajo Reading Study progress report No. 16. The University of New Mexico. Neundorf, Alice 1982 Terminology development in Navajo. IJAL 48.3:271-276. 1983 'Aichini Binaaltsoostsoh A Navajo/English bilingual dictionary. Native American Materials Development Center. Albuquerque, New Mexico. Sapir, Edward 1958 [1921] Language. New York: Harcourt Brace. Spolsky, Bernard 1971 Navajo language maintenance II: six-year-olds in 1970. Navajo Reading Study progress report no. 13. The University of New Mexico. Young, Robert W.—William Morgan 1943 The Navajo Language. Phoenix: Bureau of Indian Affairs. 1980 The Navajo Language. University of New Mexico Press. 1987 The Navajo language, revised edition. University of New Mexico Press.

E. Phonology

Loss of contrast between voiced and voiceless alveolar flapped stops in American English Donald Bowen f

Abstract This article deals with the pronunciation of "t" and "r" in American English. The matter of contrasts between dialects is relevant to language change, as well as teaching English as a Second Language.

American English is distinguished from other standard variations, and notably British English, by a relatively small number of contrasts. One of the most conspicuous is the behavior of the sound /r/, which from an American perspective sometimes appears when it is not expected and at other times is missing when it is anticipated. Thus in the expression "law and order" there is an /r/ attached to law, giving /lor/, but in order, where we expect two /r/ sounds, we get none, and the form /5da/ occurs. But the distribution is even more complicated. Some American dialects, notably New England and Southern, follow the British rather than the American pattern. Another pattern almost as conspicuous is the glottal release of /t/ in American English when this sound occurs between a /t/ and a syllabic /n/, as in button, cotton, fountain, curtain, etc. This pronunciation clearly marks a speaker as American. A third contrast between British and American English is the subject of this paper. It is the flapped version of American English /t/ when the /t/ is between two syllables with the preceding syllable strong-stressed, and the following syllable weak-stressed. An example of this pattern is atom and atomic, the same sequence of phonemes appears in a-t-o-m and a-t-o-m-i-c. The form with the suffix -ic is stressed after the -t-, and this has an effect on the pronunciation of the /t/. Note the following comparison: atomic:

The /t/ is aspirated, the normal, basic pronunciation of the /t/: /atamik, atamik, atamik/.

324

J. Donald Bowen

aiom:

The /t/ is flapped, a result of the location of the strong stress before the /t/: /¿tarn, ¿tarn, aetom/.

Now listen to them together: /atamik, aetam, atamik, aetam/. The minimal-pair concept can be applied to words, phrases, sentences, and theoretically to larger contexts. Examples of minimal-pair words are: map pan pal in/ernal historical

nap pen pad infernal hysterical

The contrasts between minimal pairs are shown in the spellings in these examples and are further underscored to pinpoint specific details. (Minimal-pair words are widely used in the teaching of phonology and pronunciation, though in my opinion there is a limited gain with their use.) The spellings and the language mores indicate minimal-pair status for end and and, even though and is most often pronounced /an/ or even /n/. Here are two minimal-pair sentences: , , trader. We don t need another ^ . traitor. „ ^ rider. TT , He s an excellent writer.

(merchant) ; . (spy) (horseman) lL > (author)

The main difference between atom/atomic and traitor/trader comes from the fact that atom/atomic are related words and traitor/trader are not. That does not mean we cannot do something useful with a pair like atom/atomic. We can, but the result is an illustrative pair of sentences, not a minimal pair. Minimal-pair sentences have one different word, as in trader/traitor or riifer/wrifer shown above. These words are not related to each other in any meaningful way. More numerous are pairs that set the form pattern but are unrelated to each other. It is very difficult to get minimal-pair sentences, though it is fairly simple to get illustrative sentences. Note the following pair with baddie and batty, which have almost no lexical items in common: The guy with the black hat is a baddie. It's enough to drive a person batty.

(hoodlum) (insane)

Voiceless alveolar flapped stops in American English

325

Now we are ready to discuss the problem for which this paper is written. The flapped /t/ appears to be voiced, and if so, it is possibly not distinguished from the flapped /d/, which occurs in patterns similar to the /t/. The primary data consists of possible minimal-pair sentences that may or may not be different: bitter/bidder latter/ladder atom/Adam Maybe instead of having minimal pairs, we have by coalescence the same sentence twice. This is not easy to determine, partly because there is a regular distinction in the spelling, and partly because the flaps occur only in allegro (informal) forms, and because any native speaker can substitute lento (formal) forms, which clearly distinguish /t/ and /d/. To maintain the existence of the flaps one has to restrict the corpus to informal style. Many native speakers claim they can tell the difference between forms with /d/ and forms with /t/, and so they can if the form of the corpus is controlled. Of course the corpus could be controlled in the other direction, imposing sameness where differences exist, for the sake of producing a corpus where phonemic differences are suppressed. How do we produce, or identify, a fair corpus? The traditional procedure for studying the /t-d/ pair has often been to record minimal-pair words in random order and then have test subjects identify which one they heard. One trouble with that procedure is that the members of a pair don't have the same probability of occurrence (e. g. butter and budder, where butter is much more likely to occur than budder and consequently the subject's interpretation of budder might be deflected in the direction of butter). There are reasons to avoid word pairs that might be affected by balance, and an effort to preserve balance has been made in the present study. I have proposed to build an evaluation procedure that will measure a possible distinction between voiced and voiceless flaps, assuming these exist in informal American English. One way to minimize the probability weakness is to present the test words in context. This is helpful, but the frequency problem is only partially taken care of. First of all, the favorite contextual sentence is a minimal-pair sentence frame such as those cited previously in this paper. It is possible, though, that the voice reading the data tapes may be

326

J. Donald Bowen

influenced by the recording. The procedures worked out for this exercise are designed to minimize errors of this kind. Two charts accompany this paper (pp. 328 — 329). The first is Chart One, the Recording Script, since this test is administered from a tape recording. The Recording Script consists of two parts: twenty-five each of test sentences and test words. Column One heads twenty-five brief sentences, each containing a word with an alveolar flap, distributed between /t/ and /d/ sources. To prepare the script the recording voice reads Sentence 1, allows for a pause of 3 seconds, and then reads Word 1. (Word 1 is the alveolar-flap word from Sentence 1, whether derived from /t/ or /d/.) This pattern of recording is continued until all 25 sentences and test words are recorded. To finish, it is necessary to remove the test sentences when preparing Chart Two. The test sentences and words are designed to preserve the bias that is inherently present in the meaning of words. Let me illustrate: Suppose I have chosen the pair of alveolar flaps represented by tidal/ title. In preparing the test tape the recording voice will see and will be instructed to repeat the sentence "A tidal wave threatened the island." Then, by itself he/she repeats the word "tidal." The words wave, threatened, and island all strongly support the interpretation tidal over title. To be more sure the recording voice produces tidal instead of title, the word tidal is read onto the tape as a separate part of the item, so a complete item appears as follows: A tidal wave threatened the island.

Tidal.

The test subject taking the test later on hears only the word tidal, the sentence has been physically removed from the tape. Both alveolar-flap words are present in written form on the answer sheet, one of which is heard. The task for the test subject is to check the form he thinks he hears. If the test subject is consistently correct, the data which distinguishes tidal and title are present in the recorded alveolar flap. If the test subject makes numerous and random errors, the alveolar flaps can be assumed to have fallen together, no longer carrying the feature voiced/voiceless. The test sentences have done their part by biasing the meaning of the test words. Thus the test words will be read with meanings clearly in the mind of the recording voice, and the test sentences are now part of the history of the test. The question to be answered is: Does a conscious awareness of the meaning of the test sentences transfer to the test words?

Voiceless alveolar flapped stops in American English

327

Chart Two also has two columns. Some of the test words from Chart One appear correctly in Column One of Chart Two; others fit correctly in Column Two. If /t/ appears in Column One, for example, /d/ appears in Column Two, and vice versa. The test tape has Column One of Chart One removed, so it is not heard by the test subjects. All that remains are the test words, and these are what the subject hears and judges. Twenty-five of the fifty blanks will be marked with a check to indicate what the test subject believes are the correct answers. Each test item gives a choice of a /t/-word or a /d/-word, with judgment made from the two. The number of correctly-selected test items is the test score. A score around 50% indicates the /d/ and /t/ forms have fallen together, with no surviving features to mark them as different. A high score, say 80 — 85%, indicates there are surviving clues that allow a native speaker to distinguish between the /d/-word and the /t/-word. I plan in later studies to analyze test results, both of my own test administrations and of those colleagues who may be interested in joining me for comparative studies. For this purpose I have included details of test construction and analysis. This procedure should validate the presence of a distinction, or once and for all prove that alveolar flaps do not coalesce. This modest attempt at research has been explained in detail, with full forms, for a purpose. It is to invite you to join me in finding out whether or not the flapped variants fall together — and if so, under what set of circumstances this happens. Yet to be decided is what percentages of correct usage and of error constitute a reliable answer. Note that I am not providing a set of recordings for test subjects to use. I leave this to you. We might make comparisons later, if it seems it would be productive. It might be helpful to tell you how I proceeded to remove the full sentences when I wanted only the single words. I start by finding a good voice to make the recording from the script. This voice must be sensitive to naturalness of informal English. (There would be no point doing a project on formal English.) This recording then serves as a master, with a clean tape in another machine, a slave. Ride the pause control switch so that only the numbers and the items which are found on the answer sheet appear. This tape becomes your test tape. I hope this invitation attracts some takers. Good luck.

328

J. Donald Bowen

Appendix i Voiced and Voiceless Alveolar Flapped Stops in American English Chart One:

RECORDING SCRIPT

Instructions: You are preparing a tape recording for a language test in English. On your recording script (this chart) you will see 25 items, each consisting of a short sentence and, after a pause, one of the words from the sentence. Be sure to leave about 3 seconds between the sentence and the single word. For recording voice only Examples: Have you seen Etty? Have you seen Eddie? 1.

Don't meddle.

(Heard by test subjects) Etty. Eddie. 1.

Meddle.

2. Pedal faster.

2. Pedal.

3. He gave his caddie a generous tip.

3. Caddie.

4. What's the title of his new book?

4. Title.

5. Theirs was a feudal society.

5. Feudal.

6. He's a writer of novels.

6. Writer.

7. He's odder than a raccoon.

7. Odder.

8. That tree's a cedar.

8. Cedar.

9. Adam was the first man.

9. Adam.

10. He's deader 'n a doornail.

10. Deader.

11. He was the auction's best bidder.

11. Bidder.

12. That's utter nonsense.

12. Utter.

13. Up 'n at 'em.

13. At 'em.

14. It tastes bitter.

14. Bitter.

15. A kiddy does not say meow.

15. Kiddy.

16. I'll bring a liter of milk.

16. Liter.

17. He's certainly on his mettle.

17. Mettle.

18. You're an utter nuisance.

18. Utter.

19. John gets madder and madder.

19. Madder.

20. Take us to your leader.

20. Leader.

21. A tidal wave threatened the island.

21. Tidal.

22. He's a high bidder.

22. Bidder.

23. Adam was Eve's husband.

23. Adam.

24. Oh, a kiddy car.

24. Kiddy.

25. He's even madder today.

25. Madder.

Voiceless alveolar flapped stops in American English

Appendix

329

ii

Voiced a n d Voiceless Alveolar Flapped Stops in American English Chart Two:

ANSWER SHEET

Instructions: You will hear a series of two-syllable words or expressions each with an alveolar flap between its syllables. Decide whether the flap is voiced or voiceless and check the corresponding word in the numbered blank provided. Examples:

Etty

Eddie

Eddie

Etty

1. metal

meddle_

2. pedal

petal

3. catty

caddie_

4. tidal

title

5. feudal

futile

6.

rider

writer

7. odder

otter

8. seater

cedar

9. atom

Adam_

10. debtor

deader_

11. bidder

bitter

12. udder

utter

13. Adam

at'em

14. bitter

bidder_

15. kitty

kiddy

16. leader

liter

17. mettle

medal

18. udder

utter

19. matter

madder.

20. leader

liter

21. tidal

title

22. bidder

bitter

23. atom

Adam

24. kitty

kiddy

25. matter

madder.

Some environments which may condition vowel length1 Eunice V. Pike

Abstract A vowel may have various lengths in accordance with the environment in which it occurs. Although an environment may affect vowel length in the same way in various languages, the effect is probably not universal. In this paper I have listed fourteen environments which sometimes affect the length of a vowel, and have given an example of a language in which this lengthening occurs.

The length of a vowel is often conditioned by the environment in which it occurs. This paper gives fourteen environments which sometimes affect the length of a vowel. For each environment, languages are cited which have lengthened vowels in that environment. For example, a vowel in an open syllable may be longer than a vowel in a closed syllable. Robbins (1961:245) says that in Quiotepec Chinantec of Mexico a checked syllable is normally slightly shorter than a vowel in a free syllable. A vowel preceding a voiced consonant may be longer than a vowel preceding a voiceless consonant. Kenyon (1943:67) says that the same vowel is longer when before a voiced consonant than it is when preceding a voiceless consonant. 1. Occurrence in a stressed syllable lengthens the vowel in many languages. For example: English (Jones 1940:217, sixth edition), Chuave of Papua New Guinea (Swick 1966:35). Of French, Haden says (1962:38) "... prolongation rather than loudness is the main feature of so-called stress in French." On the other hand, Lehiste points out (1970:41) that in Estonian, "the duration of a vowel in an unstressed second syllable is, as a rule, longer than that of a vowel in a stressed short first syllable."

332

Eunice V. Pike

2. A vowel preceding a voiced consonant may be longer than a vowel preceding a voiceless consonant. Numerous people have pointed out this feature in English. For example: Kenyon says (1943:67), "The same vowel, if stressed, is longer when final or before a voiced consonant than it is before a voiceless consonant." O'Kane says (1978:318), "Durational vowel differences must still be seen as the major perceptual cue used to distinguish voiced from voiceless following consonants." Mitleb (1984:23) says that "a vowel before a voiced consonant is significantly longer than that preceding a voiceless one." But he warns that that is not a universal. He points out that it is not true of Arabic. Abercrombie also says (1967:81) that vowels preceding voiced consonants are longer than vowels preceding voiceless consonants. 3. A vowel preceding a glottal stop may be shortened. Concerning the language Chatino of Mexico, Upson says (1968:3), "When other environmental factors are the same, the nucleus of a syllable with a syllable-final glottal stop is shorter than the same nucleus without a syllable-final glottal stop." 4. A vowel in an open syllable may be longer than a vowel in a closed syllable. Concerning Tamil of India, Balasubramanian says (1981:153) that "vowels ... are longer in syllables of the structure V and CV than they are in syllables of the structures VC and CYC." Concerning Quiotepec Chinantec of Mexico, Robbins says (1961:245), "A long checked syllable is normally slightly shorter than a long free syllable." Abercrombie says (1967:81) that a vowel in an open syllable is longer than a vowel in a CVC syllable, and that a vowel in a CVC syllable is longer than a vowel in a CVCC syllable. 5. A word-final vowel may be longer than a word-initial vowel. Concerning Awa of Papua New Guinea, Loving (1966:24) says that word-final vowels are longer than word-initial or word-medial vowels. Also Rich (1963:198) says that words in Arabella of Peru end in a lengthened vowel. 6. A prepause environment may lengthen the vowel. Upson (1968:4) says of Chatino of Mexico that a syllable in a prepause environment is longer than a syllable in a non-prepause environment.

Some environments

which may condition vowel length

333

Homma (1981:272) says of English that "vowels are longest before pauses." Berkovits (1984:256) says, and quotes others as saying, that final syllables are longer in English, German, Spanish, and French. 7. Nasalized vowels may be longer than oral vowels. O'Shaughnessy (1981:389) says that a nasalized vowel in French is longer than an oral vowel. 8. Vowels in a controlled syllable may be longer than vowels in a ballistic syllable. Merrifield (1963:3) says that in Palantla Chinantec of Mexico, "Ballistic syllables are shorter in duration than controlled syllables." Mugele (1984:2) adds that the ballistic syllable is also shorter in Lalana Chinantec. 9. A vowel preceding a lenis consonant may be longer than a vowel preceding a fortis consonant. Nellis and Hollenbach say (1980:98) that in Cajonos Zapotec, "Simple vowels with primary or secondary stress are lengthened in open syllables and preceding lenis consonants" (but not when preceding fortis consonants). 10. A vowel which is the only one in a syllable may be longer than a vowel in a syllable with two or three vowels. Pike and Pike (1947:78) say of the Huautla dialect of Mazatec of Mexico that, "The nucleus of the syllable takes about the same length of time ... regardless of the number of vowels it contains, whether one, or two, or three." 11. Vowels in a word with only one or two syllables may be longer than vowels in a word with more syllables. That is, a syllable in a word with several syllables is pronounced more rapidly than a syllable in a word with only one or two syllables. This characteristic whereby a word with several syllables takes no more time than a word with only two syllables might be called "word timing". Concerning Chuave of Papua New Guinea, Swick says (1966:37), "A one syllable word may be as long or nearly as long as a polysyllabic word 99

Concerning the Molinos dialect of Mixtec, Hunter and Pike say (1969:24), "... a phonological word of several syllables tends to be said

334

Eunice V. Pike

with the same length of time as one with two syllables, and a word with numerous syllables is said very fast ...." Lehiste says (1970:40), "It appears that in some languages the word as a whole has a certain duration that tends to remain relatively constant, and if the word contains a greater number of segmental sounds, the duration of the segmental sounds decreases as their number in the word increases." She gives Hungarian as an example. Balasubramanian says of Tamil (1981:160), "Vowels in monosyllabic words are longer than those in words of more than one syllable." Ladefoged says of English (1975:80), "... the vowel in the stressed syllable gets progressively shorter as a result of adding extra syllables in the word." He demonstrates that with the words, "speed, speedy, speedily." 12. A low vowel (for example [a]) may be longer than high vowels (for example [i, u]). Jones (1950:124) points out that in English the low vowel /a/ is longer than the high vowel /i/. Lehiste (1970:18) uses English as an example, but she also refers to other linguists who have reported that low vowels are longer than high vowels in Danish, German, Swedish, Thai, and Spanish. O'Shaughnessy (1981:398) says of French that in closed syllables the low vowel is definitely longer than the high vowels. 13. Close (or tense) vowels may be longer than open (or lax) vowels. Abercrombie (1967:81) says that the vowel /i/ as in 'beat' is longer than the vowel /i/ as in 'bit', "when they occur in the same circumstances." 14. A vowel with a low tone may be longer than a vowel with a high tone. In Villa Alta Zapotec of Mexico, high tone is shorter than mid or low tone (Pike 1948:167). Also in Western Popoloca of Mexico, vowels with high tone are short, but with low tone are long (Williams and Pike 1968:379). Abramson says of Thai (1962:108) that the mid and low tones are longer than the high tones. Note 1. This paper was presented orally at the Thirteenth LACUS Forum held at the University of Texas in Arlington, Texas, August 12—16, 1986.

Some environments which may condition vowel length

335

References Abercrombie, David 1967 Elements of general phonetics. Chicago: Aldine Publishing Co. Abramson, Arthur S. 1962 The vowels and tones of Standard Thai: Acoustical measurements and experiments. Indiana University Research Center in Anthropology, Folklore and Linguistics Publication no. 20. Bloomington: Indiana University. Balasubramanian, T. 1981 Duration of vowels in Tamil. Journal of Phonetics 9:151 —61. Berkovits, Rochele 1984 Duration and fundamental frequency in sentence-final intonation. Journal of Phonetics 12:255-65. Haden, Ernest F. 1962 Accent expiratoire. Studies in Linguistics 16.1:23 — 39. Homma, Yayoi 1981 Durational relationship between Japanese stops and vowels. Journal of Phonetics 9:273-81. Hunter, Georgia G.—Eunice V. Pike 1969 The phonology and tone sandhi of Molinos Mixtec. Linguistics 47:24 —40. Jones, Daniel 1940 (Sixth edition) An outline of English phonetics. New York: E. P. Dutton and Co. 1950 The phoneme: Its nature and use. Cambridge: W. Heffer and Sons Ltd. Kenyon, John Samuel 1943 (Ninth edition) American pronunciation. Ann Arbor: George Wahr. Ladefoged, Peter 1975 A course in phonetics. New York: Harcourt Brace Jovanovich. Lehiste, Ilse 1970 Suprasegmentals. Cambridge, Mass.: M.I.T. Press. Loving, R. E. 1966 Awa phonemes, tonemes, and tonally differentiated allomorphs. Linguistic Circle of Canberra Publications, Series A, 7. Canberra: Australian National University. Merrifield, William R. 1963 Palantla Chinantec syllable types. Anthropological Linguistics 5.5:1 — 16. Mitleb, Fares M. 1984 Voicing effect on vowel duration is not an absolute universal. Journal of Phonetics 12:23-27. Mugele, Robert L. 1984 The phonetics of ballistic and controlled syllables. Texas Linguistic Forum 23:1-43. Nellis, Donald G. — Barbara E. Hollenbach 1980 Fortis versus lenis in Cajonos Zapotec phonology. IJAL 46:92 — 105. O'Kane, Donald 1978 Manner of vowel termination as a perceptual cue to the voicing status of postvocalic stop consonants. Journal of Phonetics 6:311 — 18.

336

Eunice V. Pike

O'Shaughnessy, Douglas 1981 A study of French vowel and consonant durations. Journal of Phonetics 9:385-406. Pike, Eunice V. 1948 Problems in Zapotec tone analysis. IJAL 14:161—70. Pike, Kenneth L. — Eunice V. Pike 1947 Immediate constituents of Mazateco syllables. IJAL 13:78—91. Rich, Furne 1963 Arabela phonemes and high-level phonology. Studies in Peruvian Indian Languages. Mildred L. Larson and others, Norman, Okla.: SIL, 193 — 206. Robbins, Frank 1961 Quiotepec Chinantec syllable patterning. IJAL 27:237 - 50. Swick, Joyce 1966 Chuave phonological hierarchy. Papers in New Guinea Linguistics, 5. Linguistic Circle of Canberra Publications, Series A, 7. Canberra: Australian National University, 33—48. Upson, Jessamine 1968 Chatino length and tone. Anthropological Linguistics 10:2.1 —7. Williams, Ann R. — Eunice V. Pike 1968 The phonology of Western Popoloca. Lingua 20:368 - 80.

What is a 'register' language? Richard S. Pittman

Abstract Phonetics and Phonemics were developed primarily by speakers of European languages, using Greek and Roman alphabets, and their derivatives, for symbols. Contrastive articulations for these languages are primarily oral. A large number of languages, however, particularly in Asia and Africa, exploit contrastive pharyngeal articulations for lexical and grammatical purposes. These articulations, based principally on constricted versus expanded pharyngeal cavity, give resultant phonetic effects variously labelled as squeaky vs. breathy, emphatic vs. non-emphatic, retroflexed vs. non-retroflexed, high vs. low pitch, open vs. close vowel articulation, and glottalized or imploded vs. non-glottalized and nonimploded consonant types. These, and vowel harmony languages, are typically 'register' languages.

Phonological awareness, like that of other disciplines, has gone through several stages. One of the earliest was identification of consonantal segments, followed by distillation of vowels from consonant-marked syllables.1 Much later came factoring out of tones and other 'short' prosodic components, followed by writing of intonation and 'long' prosodic features of utterances. Only well into this century did there emerge a well-defined perception of what could be called 'a word-base phonology of contrastive pharyngeal effects.' J. R. Firth's 'prosodies' concept pioneered the way in this for the English-speaking world. Semitic phonologists, Indian grammarians, and Chinese rhyming specialists had already done it for their respective languages but had not generalized their results into linguistic universals. Now there is sufficient data in hand from languages all over the world to make possible a universal definition of 'register' language. That which distinguishes a 'register' from a 'non-register' language is the exploiting of pharyngeal effects, principally a more open vs. a more closed pharyngeal cavity, for lexical and/or grammatical contrast. Obviously any language may do this to a limited extent, as when English story tellers exploit

338

Richard S. Pittman

a narrowed pharyngeal cavity for dramatizing the word 'wee', or an expanded pharyngeal cavity for sepulchral effects in ghost stories. But a 'fair dinkum' register language is one in which the contrast is pervasive, affecting very large areas of the lexicon and/or grammar. Following are some illustrations. 1. All of the Semitic languages, so far as I know, are register languages. The so-called 'emphatic' consonants of Arabic and Hebrew, for example, are ones which constrict the pharyngeal cavity by retracting the tongue root. Syllables initiated by non-emphatic consonants have a more open pharyngeal cavity. 2. All tone languages are register languages. This is because the laryngopharyngeal articulations which are exploited for contrastive tones give contrastive registers also, and register contrast is more fundamental than tone contrasts. However the converse is not true. Not all register languages are tone languages. 3. Languages with major vowel harmony systems are register languages. This is because pharyngeal articulations, as Perkell has pointed out, 'drag' across syllable boundaries. Turkic languages are ideal illustrations of this phenomenon. 4. Languages such as Tamil, with a major retroflexed/non-retroflexed syllable contrast are register languages. The retroflexion accomplishes for the pharyngeal cavity what tongue-root retraction does in Semitic languages. 5. The so-called 'voiced aspirates' of some languages of India are diagnostic of register languages. They are an artefact of enlarged pharynx articulation. 6. If a language has a set of prenasalized stops (mb, nd, ngg, etc.) contrasting with a set of voiceless stops, expect the prenasalized ones to be symptomatic of the 'open mode' articulation of the pharynx. 7. Glottalized and preglottalized consonants, as well as 'glottal-creak' syllables are features of the constricted pharynx mode of register languages. 8. A major contrast between /a/ and schwa in accented syllables is a register contrast. Schwa is an open-pharynx articulation, while /a/ is a narrow-pharynx articulation. Many languages of India have this contrast. 9. In languages which have a correlation between a voiced-voiceless stop contrast and tone or voice quality contrasts, the voiced-stop initials represent open pharynx register and the voiceless stop initials represent narrow pharynx register.

What is a 'register' language?

339

10. Imploded consonants reflect narrowed pharynx articulation of register languages. Three major corollaries must be added: A. No one of the identifying clues for either register is an infallible artefact of that register. Every one of them may be overridden (dominated) by a higher-ranking articulation which will put a syllable, wordbase, or word into a register which seems to be the opposite of that expected. The ultimate determination of the register of a given syllable, word-base or word depends not on the phonetic details of its composition but on the register-harmony rules which govern the phonology of the language. B. Either of the two most common registers may be 'marked' and the other 'unmarked'. While one may be typically marked and the other unmarked in a given language, it is a mistake to seek to specify either one as the most typically marked in all register languages. C. In addition to the two principal registers which characterize every register language, there is always a residue of 'neutral' syllables, wordbases and/or words which are non-contrastive with regard to register. It is tempting to regard this third group as a third register. In some cases there may indeed be a third register. In most cases, however, those syllables, word-bases or words of a register language which do not fit contrastively into either the first or the second register set are simply neutral with regard to register. Todrah is an example of a three-register language (Gregerson and Smith 1973). Assuming this description of register languages, however incomplete, to be essentially correct, a serious question immediately arises: How does it happen that within a single language family some of the languages are register languages and others are not? There are two parts to the answer. 1. As Henry Sweet pointed out long ago, one of two cognate languages may focus most of its articulations near the front of the mouth, while the other may exploit the back of the mouth, including the pharyngeal cavity, to a much greater extent. Some dialects of Hebrew, for example, are mostly 'front of the mouth' dialects, while cognate Arabic makes much use of the far back of the mouth. Those which involve the pharyngeal cavity for lexical and/or grammatical contrasts are register languages. 2. Contrastive register, as a major, pervasive feature of a given language may splinter, shatter, fragment into scores of apparently unrelated non-

340

Richard S. Pittman

register contrasts, few if any of which retain much resemblance to the original, consistent register contrast. Bahnar, for example, while unquestionably related to the large Mon-Khmer family of register languages, now reveals few traces of the major register contrast which characterizes most other languages in the family. Note that non-register languages may have various bits and pieces of register contrast in their phonologies. A quite different, but also highly relevant question is: What is the relation between phonemics, long (Harris) components, Firthian prosodies, and register? Classical phonemics focuses primarily on consonants and vowels, with suprasegmental features of stress, length and tone recognized as relevant to the syllable. More recent phonemics accepts contrastive features of any length up to the utterance and the sentence, but does not specifically focus on register contrasts. One reason, perhaps, is that Pike used the term 'register' as referring to contrastive pitch levels, which he distinguished from 'contour' glides from one level to another. A 'register' language, by that definition, was a tone language whose tones were primarily ones of different levels rather than ones of different glides. Harris, because of his familiarity with Semitic languages, and Firth, because of his exposure to languages of India, were much closer, with their 'long components' and 'prosodies' to the definition used in this paper. There is not a one-to-one correspondence, however, between their definitions and those given here. A principal problem for all theories of phonology, whether addressed or not, is the representation of registers in orthographies. This has traditionally been done in three different ways: 1. By distinctive single symbols, such as /h/ for open pharynx register and /'/ for narrow pharynx register, written before or after a vowel or consonant. 2. By distinctive sets of symbols, such as the voiced aspirates of Devanagari for open pharynx and the retroflexed consonants of Devanagari for narrow pharynx register. 3. By distinctive combinations of symbols, such as is done in the writing of Thai. Each of these conventions has both advantages and drawbacks. The commonest tradeoff is that between redundancy and economy. Economy is sometimes achieved at some cost in ambiguity. Crystal clarity often requires some redundancy.

What is a 'register' language?

341

Without presuming to disparage any of the earlier work which did or did not take register (as here defined) into account, we would summarize by affirming that a better understanding of voice register phonology should be of great value to all studies of the sound systems of natural languages.

Note 1. I am indebted to David Thomas and Kenneth Gregerson for reading earlier drafts of this paper, and for suggestions, some of which I have been able to incorporate into the final version.

References Abercombie, David 1967 Elements of general phonetics. Edinburgh: Edinburgh University Press. Catford, John 1964 Phonation types. In honour of Daniel Jones, ed. by David Abercombie et al. London: Longmans. Firth, J. R. 1957 Papers in linguistics, 1934—51. London: Oxford University Press. Gregerson, K. J. 1976 Tongue-root and register in Mon-Khmer. Austroasiatic Studies, ed. by Philip N. Jenner, Laurence C. Thompson and Stanley Starosta. Oceanic Linguistics Special Publication No. 13, 1:323—60. Honolulu: The University Press of Hawaii. Gregerson, K. J. — Kenneth Smith 1973 The development of Todrah register. Mon-Khmer Studies IV. ed. by David Thomas and Nguyen Dinh Hoa, 143 — 84. Harris, Zellig S. 1944 Simultaneous components in phonology. Language 20:181—205. Manley, T. M. 1976 Pharyngeal expansion: Its use in Sre vowels and its place in phonological theory. Austroasiatic Studies //:833—41. Honolulu: University of Hawaii Press. Perkell, J. S. 1969 Physiology of speech production. MIT Research Monograph No. 53. Cambridge, Mass.: MIT Press. Pike, Kenneth L. 1948 Tone languages. Ann Arbor, Michigan: University of Michigan Press. Pittman, Richard S. 1978 The explanatory potential of voice register theory. Mon-Khmer Studies VII: 201 - 2 2 6 . Sweet, Henry 1877 A handbook of phonetics. Oxford: Clarendon Press.

F. Comparative Studies

The Proto Otopamean vowel system and the development of Matlatzinca Doris A. Bartholomew

Abstract The reconstruction of Proto Otopamean (*OP) vowel clusters helps explain the unusually large number of vowel correspondences in the daughter languages. This solution was anticipated by Newman and Weitlaner (1950) to account for Otomi and Mazahua vowel correspondences. Present day vowel clusters in North Pame (NP) permit the recovery of *OP clusters in most instances although some are secondary clusters and some *OP clusters have been modified under certain conditions. Reflexes of *OP clusters in Matlatzinca generally show loss of the second member but in certain cases show a coalescence. In cases where the NP reflexes are not clear the *OP cluster can often be reconstructed from the correspondences of Matlatzinca with Otomi and Mazahua.

Introduction Matlatzinca is an Otopamean (OP) language of the Otomian subfamily. It and closely related Ocuilteco form Southern Otomian, whereas Otomi and Mazahua constitute Northern Otomian. The Pamean subfamily includes several dialects of Pame and Chichimeco Jonaz. The Otopamean languages are spoken in Central Mexico in the states of Mexico, Hidalgo, Queretaro, Guanajuato and San Luis Potosi, with some spillover into Michoacan on the West and Puebla and Veracruz on the East. Matlatzinca and Ocuilteco are spoken in the southwestern part of the State of Mexico. Reconstruction by the comparative method was initiated by Stanley Newman and Robert J. Weitlaner in 1950 with the reconstruction of Proto Otomi and Primitive Central Otomian (Otomi-Mazahua). These papers resulted from a seminar on the comparative method conducted by Newman in Mexico at the National School of Anthropology and History. The Otomian family was a logical laboratory for the comparative method because of its proximity to Mexico City and because of the extensive field work which Weitlaner had done in these languages. In

346

Doris A.

Bartholomew

1953 Weitlaner published a reconstruction of Proto Otomi-Matlatzinca, which also included Mazahua, and which constituted the third article in the series begun by Newman. In 1960, Juan Hasler published a reconstruction of Matlatzinca and Ocuilteco. In 1960 I published some revisions of Proto Otomi consonants based on field work done by me and other members of the Summer Institute of Linguistics. In 1959 I had compared Otomi and Pame for my master's thesis at the University of Pennsylvania; in 1965 I compared all the languages of the family for a reconstruction of Proto Otopamean as a doctoral dissertation at the University of Chicago. In 1965 as I was reconstructing Otopamean, I had less material on Matlatzinca and Ocuilteco than on the other languages, although I had a long word list by Henrietta Andrews of the Summer Institute of Linguistics, done in 1945. I also had a grammar and vocabulary of Matlatzinca by Miguel de Guevara, a 1638 manuscript from the convent in Charo, Michoacán, published in 1862. While I was at Chicago I consulted a photo copy of Basalenque's 1640 grammar and dictionary at the Newberry Library. I also had occasion to track down an earlier manuscript dictionary of Matlatzinca which was hand-written by Fray Andrés de Castro in the margins of the 1555 edition of Molina's SpanishAztec dictionary, with a note that he finished the Matlatzinca entries in 1557. I was interested to see that Castro had a much better treatment of glottal elements in Matlatzinca words than Guevara had in 1638 (Bartholomew 1967). On a Phillips Foundation grant I later copied out all of Castro's Matlatzinca entries (which I did over a two year period). Apart from the extensive lexical data in the Castro manuscript, the years since 1965 have seen quite a bit of work done and further publications in Otopamean languages. In 1967 Daniel Cazés published his professional thesis on the Matlatzinca still spoken in San Francisco Oxtotilpan; in 1982 Martha Muntzel published her thesis on Ocuilteco. Other Mexican linguists have published theses on Chichimeco Jonaz, Pame, Mazahua, and Otomi. I myself did some field work in Matlatzinca, putting special emphasis on tone analysis and contemporary grammar (Bartholomew 1968). In 1975 Basalenque's 1640 grammar and dictionary of Matlatzinca was published. Today it should be possible to do a much more detailed work in comparative Otopamean. As I started this paper I entertained thoughts of doing this to some degree, but it soon became apparent that such a detailed study will have to be the work of the next generation of scholars. I decided then to take a portion of my 1965 thesis and highlight the

The Proto Otopamean vowel system

347

insights gained from the vowel clusters in present day Pame for the reconstruction of the vowel system of Proto Otopamean and for the explanation of the sound changes that have taken place in Matlatzinca and in Otomi and Mazahua. It is instructive to note that Newman and Weitlaner (1950 b) reconstruct diphthongs for Proto Central Otomian (Otomi and Mazahua). They reconstructed *Vi diphthongs on the evidence from Otomi, attributing the VmV correspondences in Mazahua to an inserted m in the environment of vowel nasalization. Further evidence from Matlatzinca and North Pame indicate that Otomi lost the stem-formative m and the resulting vowel cluster is a secondary diphthong. The other diphthongs that Newman and Weitlaner reconstruct are formulas for the irregular OtomiMazahua correspondences. The present day diphthongs in North Pame permit recovery of Proto Otopamean (*OP) diphthongs in many cases and allow for better recovery of the *OP diphthong system. There are still some residual cognate sets which remain to be explained by future research. Some may never be completely resolved because of intrinsic limits to the comparative method, such as duplicate merger.

1. The vowel system of Proto Otopamean The Otomian languages have many vowel phonemes but no vowel clusters (except those in Otomi which have resulted from the loss of *m from stem-formative suffixes). The Pamean languages on the other hand, and North Pame in particular, have few vowel phonemes but many vowel clusters. If one assumes that Otopamean had four oral vowel phonemes, paralleled by four nasalized vowels, and vowel clusters similar to those found in North Pame today, the development of the extra vowel contrasts in the Otomian languages can be explained as fusions of vowel clusters. In addition, the irregular vowel correspondences in the Otomian languages can be explained as differing developments from the original vowel clusters. The reconstruction of vowel clusters for *OP has further support from the phonological typology of some other Otomanguean languages: Mazatec and Popoloca, as well as Amuzgo, have small vowel inventories and those vowels combine in clusters of two or even three vowels (Pike and Pike 1947, Eunice V. Pike 1954, Kalstrom and E. Pike 1968, Bauernschmidt 1965). In fact, Rensch (1976) reconstructs such a system for all of Otomanguean.

348

Doris A. Bartholomew

The vowels of *OP are distinguished for front and back and for high and low and for oral or nasalized: Oral vowels i o e a

Nasalized Vowels T o e a

The two-vowel clusters are displayed in Tables 1 and 2. There are no geminate clusters and there are no clusters of front vowels or of low vowels. Nasalized *e does not enter into clusters of nasalized vowels. Table 1.

i e a

0

Clusters of oral vowels. i

e

a

0

—

—

ia

-

-

ai Ol

-

-

io eo ao

oe

oa

—

Table 2.

T a o

Clusters of nasalized vowels. l

a

O

—

ia

To ao

ai Ol

-

oa

Clusters of three vowels were bi-morphemic, resulting from a metathesized *i prefix in what I have called the C form of the paradigm. (The C form involves a mutation of the stem initial consonant, presumably caused by a pre-*OP prefix vowel **i.) The syllable nucleus of a *OP root contained a vowel or vowel cluster and an optional glottal element or h). If the vocalic element was a single vowel, the vowel was rearticulated after the glottal element: V V or VhV; if it was a vowel cluster, the glottal element was articulated between the two vowels. A third glottal element, or laryngeal, is posited for *OP which had a centralizing effect on the vowels of the Otomian languages but which does not have any laryngeal reflexes in the daughter languages; this is symbolized in the reconstructions as HC. The symbol HC indicates that the element had a centralizing effect (C) and that it was probably some sort of laryngeal (H).

2. Reflexes of *OP single vowels Matlatzinca shares with the other Otomian languages the development of the high vowels under the influence of the centralizing laryngeal HC. It shares with Ocuilteco the merger of *OP *e and *a to a, the loss of

The Proto Otopamean vowel system

349

phonemic nasalization and the development of *a to central vowel a. Table 3 displays the reflexes of ""OP single vowels. Table 3.

Reflexes of *OP single vowels.

Otomi Mazahua Matlatzinca Ocuilteco North Pame Chichimeco

*j

*iHC *o

*oHC *e

i i i i (e)i e

i i i i e »

3 3 3 3 O 0

o 0 0 0 0 0

Examples for *OP *i: 'chile' Ot n *?i Mtz mi NP 9ei

e e a a

»

*a a a a a a a

Ì I i i I è

*o

*e

*ä

Q ü u u 5 ü

X è e e &

ä à 3 3 ä à

Maz m Oc mi Ch

'blood' *khi

Ot khi Mtz ci-hyabi NP kkhi

Maz khm Oc nci-hya Ch khe

'squirrel' *mi-na

Ot minä Mtz mini NP meigq

Maz Oc mi Oc

'tell' *si-p

Ot siphi Mtz si NP sep

Maz siphi Oc Ch se

Examples for *OP *iHC: 'pull' Ot khPmi Mtz kUza *ki9HC-m9 NP kke^edn 'chin' *khiHC-n'> 'beans' *khiHC-9

—

Maz khibn Oc nkibi Ch ke?en9

Ot khtni Mtz nctri NP khweP

Maz kWi Oc ncili Ch

Ot khi Mtz cht NP nkhwe?

Maz khm Oc cht Ch khe9

—

T

350

Doris A. Bartholomew

Examples for *OP *o: 'louse' Ot t^oni *t9o Mtz Po NP Po 'house' *t9o-s

'stone' * — to

Maz Po9o Oc Po Ch r9o

Posi 'bed' Mtz Posi 'supper' NP Pos

Maz Posi 'bed' Oc -

Ot do Mtz nto NP to

Maz ndo Oc nto Ch ro

Ot

Examples for *OP *oHC: 'flower' Ot (foni Mtz tani * = toHC-n NP togy

Ch

r9os

Maz rufohna Oc neh Ch ro

'hen' *9oHC-n

Ot Mtz NP

3ni 9 ani ? ogV

Maz 9ana Oc ? a Ch -

'hear' *9oHC-t

Ot ?zde Mtz ?dti NP

Maz 9ara Oc 9ati Ch 9o

9

Examples for *OP *e: 'water' Ot dehe * = teh Mtz tawi NP tiE

Maz ndehe Oc nta Ch ri

'to weave' *pe

Ot pe Mtz pa NP ppas

Maz pe9e Oc mpa Ch pi

'mouth' *ne

Ot ne Mtz na

Maz ne9e Oc si-na 'lips' Ch ni

NP nae

The Proto Otopamean vowel system

'people' *te

Ot te Mtz ta NP Is

Maz te9e Oc ta Ch ri

'spider' *me-s

Ot mese Mtz masi NP mass

Maz mese Oc masi Ch -

351

Note: There is a set of correspondences where Mtz-Oc e corresponds to Ot-Maz e, and thus presumably from *e. It is found in two cognate sets, each having to do with 'cough' or 'sneeze' . The Mtz form contains a y, which suggests that a metathesized *i may have been involved. 'to cough' *hyeh

'a cold' *t-heh

Ot hehe Mtz hyewi 'a cold' NP hhae

Maz hehe Oc he

Ot thehe Mtz hyewi NP ri-thyae9se

Maz the Oc Ch -

Examples of *OP *a: Ot 9na 'one' 9 Mtz dawi *n a NP nda

Ch

-

Maz daha Oc la Ch nf>a

'plain' *m9a-tha

Ot 9matha Mtz bati NP nibia

Maz b?athi Oc bati Ch -

'horse' *pah-n9

Ot phani Mtz pari NP wahal9

Maz phad9t Oc bali Ch -

Examples for *OP *1 'straw mat' Ot si-mphi *pih-9 Mtz mphiwi NP ppëhi9 'drunken' *ti, rii

Ot ntï Mtz nimi

Maz phi-ngwa Oc phi Ch Maz tPi Oc ni

352

Doris A. Bartholomew

Examples for *OP *o: 'to look' Ot nu *nd Mtz nu

Maz

nu?u

Oc

nu

'wake up' 'to tie up' *to-t9/c?

NP

nnd?o

Ch

Ot Mtz

tut

Maz

tunt?i

tunc^i

Oc

tunc?i

Ch

tucf 'knot' tu?u

nW>u

'knot' NP

ttoc?

'knot' 'to die' *tu

'owner' *mho

Ot Mtz NP

tu

Maz

tuwi

Oc Ch

Ot Mtz NP

hmu

Maz

hmu^u

mhu

Oc Ch

mhu

Maz Oc Ch

pe

Examples for *OP *e: 'to steal' Ot Mtz *pe NP 'tortilla' *mhe

'tamale' *the-t

Ot Mtz NP Ot Mtz NP

tto

mho9

pit pewi ppx

ppT

thSbdi

Maz

theti

Oc

Ihab^ae

Ch

Ot t?aei Mtz t9emi *t e-m NP V&J xamples for *OP *t?: 'to roast' Ot hasi *ha-s Mtz hasi NP has Ot Mtz NP

pe

mhab

9

*ma-m

mhu

Maz — Oc mhe Ch -

hmab mhewi

'atole'

'to say'

tu ru

Maz

Oc Ch

thenti r?Thi t9ehme t?e r9i

Maz

hasa

Oc

hasi

Oc

hus

ma

Maz

mami

Oc Ch

mmarj

theze

mama

ma ma

The Proto Otopamean vowel system

'to know' *pa-t

Ot padi Mtz paya NP ppa 'visit'

353

Maz para Oc mpaya Ch pa 'see'

3. Reflexes of vowel clusters There are three main processes which simplify a vowel cluster: 1) coalescence of the two into a single vowel, 2) loss of one member of the cluster and 3) desyllabification of one member of the cluster. Coalescence is favored by Otomi and Mazahua whereas loss of the second member of the cluster is favored by Matlatzinca and Ocuilteco. The main lines of development of the *OP vowel clusters are quite clear and will be outlined in this paper. There are, however, some problem sets of correspondences where the data are not sufficiently numerous to permit satisfactory explanations at this time. Though the vowel clusters in NP cognates point to *OP clusters in most cases, there are some NP clusters which are caused by the copying of a high vowel in the prefix to become the first member of the new cluster. In such cases the second member of a *OP cluster may be lost in NP or it may coalesce with the original first member. Some *OP clusters, then, are reconstructed on the basis of the "irregular" correspondences in the Otomian languages. It is to be hoped that the next generation of scholars will be able to sharpen the reconstructions on the one hand and also to identify the phonological environments in which the "irregular" reflexes took place; this paper will suggest some possible explanations. The reflexes of *OP vowel clusters are presented in three groups: 1) Clusters involving *a and *o, 2) Clusters involving *a and *i and 3) Clusters involving *o and front vowel *e or *i. Table 4 (p. 354) presents the reflexes of *OP clusters involving *a and *o.

Coalescence of *OP *ao and *ao to low rounded back vowels o and 5 is common to Otomi (Ot) and Mazahua (Maz). Matlatzinca (Mtz) cognates generally show the loss of the second member of the cluster, yielding the normal reflexes for the first member: a and a. In some words, however, the Mtz reflex for *OP *ao is o. Two of these cognate sets have an initial *c, and one set has an initial but the data are too limited to identify with certainty the phonetic environment in which this reflex developed. The Mtz o reflex could have resulted from simple loss of the first member

354

Doris A. Bartholomew

Table 4.

Reflexes of *OP clusters involving *a and *o.

Otomi Mazahua Matlatzinca Ocuilteco North Pame Chichimeco

*ao

*oa

0 3 a/ o a/ o ao / a a

wa wa wa wa oa u

/ / / /

a o o o

*ào

*5a

fi

wà / à wà / 6 u u da ù

6 3 3 ào / à à

of the cluster (contrary to the usual loss of the second member) or it could have resulted from ;a coalescence of *ao to a (as in Ot and Maz) and a subsequent raising and merger with i0. Examples for *OP *ao: 'to buy' Ot *tao-m Mtz NP 'mucous' *m?ao-s

Ot Mtz

toi tomi ttaogy 9

mosi

masi

Maz Oc Ch

tomi ta

Maz Oc

b9osi

'nose' 'to ask' 9

9

* ao-n

'to try' *cao

'avocado' 9

*c ao-n

'to stand up' 9

*m?ao-m

ta masi

'nose'

NP

mbaos

Ch

Ot Mtz NP

toni fari fahodn9

Maz Oc Ch

9

Ot Mtz NP

CO copi c W

Maz Oc Ch

S090 co

—

oni ali

9 9

an

—

Ot Mtz NP

c oni c9oni c?aogrj

Maz Oc Ch

s9ohni c9o

Ot Mtz

9

Maz Oc

b9ob9i nbohobi

9

moi

—

—

The *OP cluster *oa shows desyllabification of *o to w in Otomi and Mazahua after *k, *h, and * 9 . In Matlatzinca the desyllabification takes place only after *k; elsewhere the second member of the cluster has been lost. Mazahua reflexes show the same coalescence as for *ao, o. Ot reflexes show loss of the first member after nonback consonants.

The Proto Otopamean vowel system

'white' *t?oa-s

Ot t9asi Maz t9osi NP ddoa

Maz fasi Oc Ch nu9u

'plow' *t?oa-p

Ot t9abi Mtz t?opi NP

Maz t9obi Oc t9opi Ch r9u

Ot cac9i Mtz coc^i NP coac9

Maz Oc coc9i Ch

'foot' *koa

Ot gwa Mtz kwahtu NP koa

Maz ngwa9a Oc Ch gu

'wing' *hoa

Ot hwa Mtz nhowi NP hoa

Maz hwa9a Oc nho Ch

Ot ?wada Mtz NP ddoa

?

—

'bat' *coa-c9

'maguey' *n9oa

—

355

—

—

—

—

—

Maz wadi Oc nlooti Ch n9u

The reflexes of *OP *ao in Otomi and Mazahua show a coalescence which parallels that of the oral cluster. The phonemic system for oral vowels is simpler in Otomi than in Mazahua and this typological fact influences the high vowel m reflex in Otomi as opposed to the mid vowel reflex o in Mazahua. Otomi nasalized vowels are: T = eu and a; Mazahua nasalized vowels are let S u and o. If we assume that the phonetic result of the coalescence was 5 then in Otomi it was higher than its a and was merged with m; in Mazahua a was a low central vowel which paralleled i and the 5 fit in as the low back vowel (phonetically they are low vowels but function as mid vowels with regard to vowel harmony, Bartholomew 1975). The Matlatzinca reflex shows loss of the second vowel and the centralization of *OP *a as nasalization was lost: a. Examples of *OP *ao: 'bitter' Ot *khao Mtz NP 'to sing' Ot *tàoh Mtz NP

khu cha khao tùhu tawi ndào

Maz Oc Ch Maz Oc Ch

kho9o cha kha tdho ta -

356

Doris A. Bartholomew

moon

Ot Mtz mba NP m>ào^

Maz Oc Ch

bd

There is a pair of cognates where Matlatzinca or Ocuilteco have the reflex u instead of the expected a; no explanation is apparent at this point. 'night' ~sao-m

Ot Mtz NP

sui Sdirti sào

Maz Oc Ch

somt Isù sa

'shadow'

Ot

sudi

Maz

sort

*sào-t/m'?

Mtz

subì

Oc

chubi

NP

sàot

Ch

-

*OP *oa has developments parallel to those for the oral cluster: Otomi and Mazahua desyllabify the o to w after back consonants *k, *h and * 9 . Maz coalesces the cluster to o whereas Otomi shows loss of the first vowel after nonback consonants. Matlatzinca and Ocuilteco reflexes show loss of the second member of the cluster and raising of the vowel as the nasalization is lost, u. Examples of *OP *da: 'corn' Ot Mtz *thda NP 'cornfield' *nhda-hm

'to confess' *kdah-

thà thuwi Ihda

Ot Mtz NP

hwàhi

Ot Mtz NP

khwàni

nomi nhòa

kunya kkòaho

Maz Oc Ch

thòfo thu —

Maz Oc Ch

hwàhma

Maz Oc Ch

kkwàm?a

nu nhu

— —

Table 5 presents the reflexes of *OP clusters involving *a and */. The Otomi and Mazahua reflexes for *OP *ai show a coalescence to the low front vowel ¿e. Matlatzinca and Ocuilteco reflexes show the coalesced mid front vowel e in those cases where the C-form i prefix has been desyllabified to y; elsewhere the second member of the cluster has been lost, leaving the reflex a. Pame has the reflex a in the case of the C-form i prefix, showing a simplification of a three-vowel cluster *OP *i-ai to Pame i-a. The Chichimeco Jonaz reflex is e, except in the case of the C-form i prefix where it is preserved at the expense of the *OP root vowels.

The Proto Otopamean vowel system Table 5.

357

Reflexes of *OP clusters involving *a and *i.

Otomi Mazahua Matlatzinca Ocuilteco North Pame Chichimeco

*ai

*ia

*âi

*Ta

X X (y)e / a (y)e / a ai / a e/i

ya ya ya ya ia

—ë ë 9 3 ài ë

yâ / â yà / T / ï i / 3/ u i/s/u Ta / ë ï/ë

Examples of *OP *ai\ 'to beat' Ot Mtz *paih-?m? NP 'to abandon' *hai-k

/ m?

'abandoned' *hi-ai-k

'hand'

Maz Oc Ch

pk&bn

papi ppahP

Ot Mtz NP

haegi habi haigy

Maz Oc Ch

haezi habi he

Ot Mtz NP

hyxgi

Maz Oc Ch

hyxzi heci hi

Maz Oc Ch

dy?as7ae ye

Ot Mtz NP

*n?i-ai

-

phx9mi

heci nhiagij fyx ye n9ia

—

pe

nn

The cognate sets for *OP *ia are not very clear, but it looks like the first vowel is desyllabifled in all the Otomian languages. iiver'

Ot Mtz

*nia

ya ya

'heart' NP

nia

Maz — Oc ya 'heart' Ch -

Otomi and Mazahua have the coalesced reflexes = e and e, respectively; Matlatzinca and Ocuilteco have the reflex a, normal for the first vowel of the *OP cluster *ai. Ch has the coalesced reflex *e. Examples of *OP *ai: 'to dance' Ot *naih-m

?

nei

Mtz

mbi

NP

nnahP

Maz mm?e Oc Ch

nahsbi nëhe?

358

Doris A. Bartholomew

'to spin' *hai-t? / m

Ot het9i Mtz hati NP haiy

Maz hec9e Oc hanti Ch rhe

The reflexes of the presumed *OP *ia present an especially complex picture. Each of the daughter languages has at least two reflexes for this cluster. It shows up in Otomi and Mazahua correspondences a : t and a : T. Mtz u corresponds to Maz i and Mtz i seems to correspond to Maz f but there is one case where Mtz a corresponds to Maz t. The data are so spotty that I hesitate to propose any conditioning factors for the varied reflexes. In fact, the NP cognates never show an Ta cluster; my reconstruction *ia is based on the Otomi-Mazahua correspondences. Perhaps Otomi routinely dropped the initial *?whereas Mazahua dropped the second vowel, *a. The difference between Maz t and f might be explainable by the presence or absence of the centralizing HC. In this case, Matlatzinca could have followed the Mazahua process and later have shifted the i to u. I suspect, however, that there are better answers to be uncovered. Examples of *OP *ia: 'rope' Ot nthàhi *thi-ai-mh Mtz thimi NP Ihwè

Via/ ththmi Oc thi Ch nthè

'to marry' *tiah-

Ot thàhti Mtz Onye NP ttehe9t

Maz chthti Oc Ch tèher

'head' *ni-ao

Ot yà Mtz nu NP nàoij

Maz ni Oc nu SP nyào

'expensive' *mla-t

Ot mòdi Mtz mu NP

Maz mizi Oc NP mèhe

Ot nsà Mtz suti NP

Maz sPi Oc sunti Ch -

—

'to smell' *sTa

—

Table 6 presents the reflexes of *OP clusters involving *o and front vowel */ or *e.

The Proto Otopamean vowel system

359

Table 6. Reflexes of *OP clusters involving *o and front vowel *i or *e.

Ot Maz Mtz Oc NP Ch

*io

*oi

*eo

*oe

*Io

*5i

yo/i ye / i o/i o/i io e

wi / i wi / i u wi / u oi / ei u/ ù

9 3 e e o e

wae / ae / a wae / ae / e we / e we / e oae / ae u/u/i

yù / a I yu/u yu / u èo / è T

we / ù wae /1 / fl u ei / u oi / èi a /s

The Otomian reflexes of *OP *io show desyllabification of *i to y after *n and *n9; Mazahua fronted *o to e and Matlatzinca and Ocuilteco lost the y. In other environments *io coalesced to i in the Otomian languages. Examples of *OP *io: 'two' Ot yoho Mtz nowi *nioh NP noi 'cane' *n9io

Ot 9yo Mtz ro NP

Maz dy?e9e Oc lo Ch r9uhu

Ot hisi Mtz htsi NP nheos

Maz hist Oc htsi Ch nhes

—

'to whistle' *hio-s

Maz yehe Oc no Ch nehen?

The *OP cluster *oi has desyllabified *o after *h in Otomi, Mazahua, and Ocuilteco. Elsewhere Otomi has the reflex * and Mazahua has the reflex i. Matlatzinca has the reflex u, which shows a raising of *o under the influence of *i before it dropped. Pame also shows a simplification of the *OP cluster; the reconstruction is based largely on the OtomiMazahua correspondences. Examples of *OP *oi: 'to blow' Ot hwiphi Mtz hupi *hoiNP heiky 'to bathe' *soi-

Ot sihki Mtz suti NP sigi]

Maz hwiphi Oc hwipti Ch hur Maz si(-ndy?ae) Oc cubi Ch -

360

Doris A. Bartholomew

'to nurse'

Ot

*coi

Mtz

NP

ci cuPu ci?

Maz sPi Oc ci Ch cu?

*OP *eo has coalesced to 3 in Otomi and Mazahua; in Matlatzinca and Ocuilteco it has lost the second member, giving e. Examples of *OP *eo: 'to ride horseback' Ot *teog

Mtz NP

tage — ll&ogy

'to burn'

Ot

nz3

*ceo

Mtz

ceti

NP

-

Maz caga Oc nde Ch M a z

Z3T3

Oc Ch

-

""OP *oe has differing reflexes according to the presence or absence of the centralizing element HC. When HC is present the Otomi reflex is 3, otherwise it is x. The Maz reflex is consistently e as is the Matlatzinca reflex. The Matlatzinca reflex in this case shows the loss of the first member, but the second member has been prevented from merging with the reflexes of *a as it does elsewhere. Examples of *OP *oe: 'meat' Ot Mtz *nkoeHC NP 'star' *coeHC

'cold' *coe

'to split wood' *coe-n?

ngs -

ygwx?

Maz Oc Ch

nge —

gu sehe ce

—

Maz Oc Ch

Ot Mtz NP

cas ce cse

Maz Oc Ch

sae?se

Ot Mtz NP

caeni seri

Maz Oc Ch

Ot Mtz NP

C3 nce?e

ccae?sedn

—

ce ci —

celi citi

The reflexes of *OP *oi show desyllabification of *5 in Otomi and Mazahua after 9 and lowering of *f to ar, elsewhere Ot u corresponds to the palatal T reflex in Mazahua. The Matlatzinca reflex is u, showing loss of the second vowel and raising of the first vowel. After *? NP and Ch

The Proto Otopamean vowel system

361

show loss of the first vowel and raising of the second vowel; elsewhere the original cluster shows in NP. Examples of *OP *oi: 'fly' Ot *?5i Mtz NP

9

wsb u 9 ei

9

Maz 9was Oc ngwi Ch fè

'hunger' *thdi

Ot thühu Mtz NP Ihdi

Maz thihmi Oc thu Ch -

'squash' *mdih-9

Ot mü Mtz muhu NP mohi9

Maz mü9u Oc muhli Ch mühu9

—

The reconstruction of *OP *io is not very clear. The Otomi-Mazahua correspondences show desyllabification of *f to n after h and n and identical vowel correspondences with those for *di. Matlatzinca and Ocuilteco have desyllabification after h and loss of the first vowel after a nasal. NP reflexes show loss of the first member after h and n. Examples of *OP *To: 'sand' Ot -mü *mlo-m Mtz mumi NP mmèoìj 'full' *nio-t 'three' *nhfo

Maz Oc mu Ch —

—

Ot yüdi Mtz nuli NP nab

Maz ñizi Oc nuti Ch

Ot hyü Mtz hyu NP nhò9

Maz hñPi Oc hyu Ch nhü

—

References Andrews, Henrietta G. 1945 a Observaciones en fenómenas tonales de matlatzinca. Ms. 14 pp. Summer Institute of Linguistics. 1945 b Matlatzinca kinship system. Ms. 2 pp, chart. Summer Institute of Linguistics. 1945 c Long word list of Matlatzinca. Ms. file. Summer Institute of Linguistics.

362

Doris A. Bartholomew

Andrews, Henrietta—Olive Shell 1945 a A preliminary report on Matlatzinca grammar. Ms. 7 pp. Summer Institute of Linguistics. 1945 b A tentative statement and description of Matlatzinca phonemes with their distribution. Ms. 21 pp, chart. Summer Institute of Linguistics. Basalenque, Diego 1975 a Arte y vocabulario de la lengua matlatzinga vuelto a la castellana. Publicación del manuscrito del año 1640, versión paleogràfica de Maria Elena Bribiesca S., con un estudio preliminar de Leonardo Manrique C. México: Biblioteca Enciclopédica del Estado de México. 1975 b Vocabulario de la lengua castellana vuelto a la matlatzinga. Publicación del manuscrito del año 1640, revisión paleogràfica, nota introductoria y apéndice por Leonardo Manrique Castañeda. México: Biblioteca Encliclopédica del Estado de México. Bartholomew, Doris A. 1959 Proto-Otomi-Pame. University of Pennsylvania Master's Thesis. Ms. 1960 Some revisions of Proto-Otomi consonants. International Journal of American Linguistics 26.317—29. 1965 The reconstruction of Otopamean (Mexico). University of Chicago dissertation. 1967 The loss of Spanish H and its effect on the transcription of the Matlatzinca language. Ms. 26 pp. 1968 Notes on present day Matlatzinca grammar, including verb aspects, noun phrase structure. Some texts. Ms. 1975 Some morphophonemic rules in Mazahua. International Journal of American Linguistics 17.37 — 41. Bauernschmidt, Amy 1965 Amuzgo syllable dynamics. Language 41.471 —83. Castro, Andrés de 1557 Vocabulario de la lengua matlatzinga. Manuscript in the margins of Molina's Vocabulario de la lengua castellano y mexicana (México 1555). Library of the Museum of the American Indian, New York. Cazés, Daniel 1967 El pueblo matlatzinca de San Francisco Oxtotilpan y su lengua. Acta Antropológica 2 a. época, 3(2). México: ENAH, Sociedad de Alumnos. García Payón, José 1936 La zona arqueológica Tecaxic-Calixtlahuaca y los matlatzincas: Primera parte. Edición facsimilar publicado en 1974, preparada por Mario Colín. México: Biblioteca Enciclopédica del Estado de México. 1979 La zona arqueológica de Tecaxic-Calixtlahuaca y los matlatzincas: Etnología y arqueología. Textos de la segunda parte revisados y anotados por Wanda Tommasi de Magrelli y Leonardo Manrique Castañeda. México: Biblioteca Enciclopédica del Estado de México. Guevara, Miguel de 1862 Arte doctrinal y modo para aprender la lengua matlatzinga. Publication of the 1638 manuscript. Boletín de la Sociedad Mexicana de Geografia y Estadística 9.197 - 260.

The Proto Otopamean vowel system

363

Hasler, Juan A. 1960 Reconstrucciones matlatzincas-ocuiltecas. Anales de Antropología 13.269—78. Kalstrom, Maijorie—Eunice V. Pike 1968 Stress in the phonological system of Eastern Popoloca. Phonetica 18.16 — 30. Muntzel, Martha 1982 La fonología de Ocuilteco (Tlahuía). México: Escuela Nacional de Antropología e Historia. Newman, Stanley—Robert J. Weitlaner 1950 a Central Otomian I: Proto-Otomi Reconstructions. International Journal of American Linguistics 16.1 — 19. 1950 b Central Otomian II: Primitive Central Otomian Reconstructions. International Journal of American Linguistics 16.73 — 81. Pike, Eunice V. 1954 Phonetic rank and subordination in consonant patterning and historical change. Miscellanea Phonetica 2.15 — 41. Pike, Kenneth L. — Eunice V. Pike 1947 Immediate constituents of Mazateco syllables. International Journal of American Linguistics 13.78—91. Quezada Ramirez, Maria Noemi 1972 Los matlatzincas: Epoca prehispánica y época colonial hasta 1650. (INAH, Departamento de Investigaciones Históricas. Serie Investigaciones 22.) México: INAH. Rensch, Calvin 1976 Comparative Otomanguean phonology. (Language Science Monographs 14) Bloomington: Indiana University Publications. Weitlaner, Roberto J. 1953 Proto-Otomi-Matlatzinca: Reconstrucciones del Proto-Otomi III. Memorias del Congreso Ciéntifico Mexicano, IV Centenario de la Universidad Autónoma de México. 12.199 - 206.

A new look at Aztec-Tanoan Irvine

Davis

Abstract Evidence for Whorf and Trager's claim of a genetic relationship between UtoAztecan and the Tanoan languages is re-examined. A revised list of cognates is given and modifications in Whorf and Trager's reconstructions are suggested in light of fuller data now available. It is concluded that the evidence for a genetic relationship between Uto-Aztecan and Kiowa-Tanoan is valid, although the relationship may not be so close nor so exclusive as to justify the use of the term Aztec-Tanoan.

Introduction It has been half a century since Whorf and Trager (hereafter WT) published their well-known study purporting to show a relationship between Tanoan and Uto-Aztecan (Whorf and Trager 1937). Their evidence has, in general, been accepted uncritically, and most classifications of North American native languages of the past fifty years have included Aztec-Tanoan (or Azteco-Tanoan) at one level or another as a discrete group of genetically related languages. Only the more recent recognition of the close relationship between Kiowa and the Tanoan languages (Miller 1959, G. Trager and E. Trager 1959, Hale 1962) has resulted in any generally accepted major modification of WT's orginial proposal. Linguists familiar with Kiowa-Tanoan and Uto-Aztecan have, nevertheless, been cautious about accepting the purported relationship between the two language families as being proven. Newman (1954) wrote, "Sound but not entirely conclusive evidence has been presented to demonstrate the relationship of Tanoan to Uto-Aztecan." Elsewhere (Davis 1979) I have observed that the evidence for a relationship between Uto-Aztecan and Kiowa-Tanoan seems convincing in spite of the fact that many of WT's Aztec-Tanoan reconstructions can be questioned. The relationship, however, is not as close as one might suppose from a cursory reading of WT's article. I want to expand on these observations in the present paper and suggest modifications in WT's

366

Irvine Davis

reconstructions based on data unavailable to them at the time of their writing. I have assembled below evidence for those phonetic correspondences involving Kiowa-Tanoan (KT) and Uto-Aztecan (UA) which seem plausible on the basis of available data. 1 No claim is made, however, that every suggested cognate pair is actually valid.

Kiowa-Tanoan reconstructions KT reconstructions follow Hale (1967). Some of these are his own reconstructions which I have filled out by substituting specific vowel symbols in place of his Y's. Other KT reconstructions are mine based on Hale's conclusions regarding initial consonantal developments in this language family. The reconstruction of noninitial consonants remains problematical. Although few of my KT forms reconstruct with medial consonants, there are frequent morpheme-final (but often word-medial) consonants in the data. These involve, most commonly, Taos /, Tewa n, or Kiowa n, m, or /. The occurrence of these consonants seems quite sporadic, and it is difficult to determine how they should be represented, if at all, in the KT reconstructions. While the rarely occurring initial / in Taos appears to be a reflex of KT *d as suggested by Hale, I have reconstructed medial */ as well as *d for KT. This */ is reflected by / in Taos, Hopi-Tewa, Jemez, and Kiowa, and by r in Isleta and Rio Grande Tewa. It appears to be distinct from *d, which is reflected as medial d in most of the languages, but as nd in some Tewa dialects. Final n in Tewa appears to be a reflex of any one of several KT consonants, including *n, *m, *d, and */. Likewise, morpheme-final / in Kiowa seems to reflect more than one KT consonant, possibly *l and *d. Forms that show a final n in Tewa and/or I in Kiowa but which remain ambiguous with respect to their reconstruction are identified by a final (D).

My vowel reconstructions are based on the hypothesis that Proto Kiowa-Tanoan had basically a system of six oral and three nasal vowels. These are listed in Table 1 together with their reflexes in the daughter languages and the corresponding V symbols used by Hale. The evidence suggests that, of the daughter languages examined, Taos is the most

A new look at Aztec-Tanoan

367

conservative in terms of retaining the postulated KT vowel system. The only major development in this language seems to have been the merging of oral mid front with high front vowels. On the other hand, Jemez and Kiowa exhibit major innovations resulting in multiple reflexes for some of the vowels. Because of this, I do not reconstruct separate proto vowels solely on the basis of differing vowel reflexes in Jemez or Kiowa. This means that I have combined Hale's Vj and V2, his Yi and Y4, and his Y2 and Y6> even though the differing Jemez reflexes remain unexplained. Table 1.

*i *e *a *u *e *o

*