Linguistic Theory and Complex Words: Nuuchahnulth Word Formation

Linguistic Theory and Complex Words Nuuchahnulth Word Formation

John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Linguistic Theory and Complex Words

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

This page intentionally left blank

Linguistic Theory and Complex Words John Stonham University of Newcastle upon Tyne

palgrave macmiilan

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Nuuchahnulth Word Formation

© John Stonham 2004

No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1T 4LP. Any person who does any unauthorised act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The author has asserted his right to be identified as the author of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2004 by PALGRAVE MACMILLAN Houndmills, Basingstoke, Hampshire RG21 6XS and 175 Fifth Avenue, New York, N.Y. 10010 Companies and representatives throughout the world PALGRAVE MACMILLAN is the global academic imprint of the Palgrave Macmillan division of St. Martin's Press, LLC and of Palgrave Macmillan Ltd. Macmillan® is a registered trademark in the United States, United Kingdom and other countries. Palgrave is a registered trademark in the European Union and other countries. ISBN 1-4039-0348-4 This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Stonham, John T. Linguistic theory and complex words : Nuuchahnulth word formation / John Stonham. p. cm. Includes bibliographical references and index. ISBN 1-4039-0348-4 (cloth) 1. Nootka language—Word formation. 2. Nootka language-Morphology. 3. Wakashan languages—Morphology. 4. Grammar, Comparative and general—Word formation. 5. Grammar, Comparative and general—Morphology. I. Title. PM2031.S76 2004 497'.9555—dc22

2004053934 10 13

12

9 11

8 7 6 5 4 3 2 10 09 08 07 06 05

1 04

Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham and Eastbourne

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

This page intentionally left blank

Tables

ix

Figures? Abbreviations

XI

A cknowledgements

xii

1 Introduction 1.1 Purpose of the Book 1.2 Sources 1.3 Organisation 1.4 Nuuchahnulth and Wakashan 1.5 Nuuchahnulth Dialect Research 1.6 Previous Literature 1.7 Transcription System 1.8 On the Term 'Nuuchahnulth' 1.9 Coverage

1 1 1 2 3 4 6 8 10 11

2 Some Phonological Preliminaries 2.1 Consonants 2.2 Vowels 2.3 The Structure of the Syllable 2.4 Stress 2.5 Phonological Processes

13 13 14 14 15 16

3

Word Structure and Categories 3.1 On the Concept of 'Root' 3.2 The Structure of Morphemes 3.3 Stems 3.4 Bound Verbs 3.5 Numbers 3.6 Names 3.7 Word Categories 3.8 Summary

32 32 35 45 46 52 53 54 63

4 Suffixation 4.1 Inflection versus Derivation 4.2 Paradigmatic Inflection 4.3 Non-Paradigmatic Inflection 4.4 Aspect

64 64 66 82 106

Vll

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Contents

Contents

4.5 Derivation 4.6 Summary

109 119

5 Reduplication and Infixation 5.1 Reduplication 5.2 Infixation 5.3 Summary

121 121 139 143

6 Stratal Segregation in the Nuuchahnulth Lexicon 6.1 Derivation/Aspect vs. Inflection 6.2 Arguments for Stratal Separation 6.3 Double Reduplication in Southern Wakashan 6.4 Mono- versus Multi-Stratal OT 6.5 Conclusions

144 145 147 162 172 176

7

178 178 185 192 195 205 206 208

Templatic Morphology 7.1 Reduplication and Templatic Morphology 7.2 Fixed Segmentism 7.3 Infixation and Templatic Morphology 7.4 Hypocoristic Formation 7.5 Vowel Length Patterns 7.6 Abnormal Speech 7.7 A Final Word about Foot-based Templates

8 Morphosyntax 8.1 Clitics 8.2 Incorporation 8.3 Compounding 8.4 Classifiers 8.5 Inflection as Clisis

210 210 213 234 246 258

9 Conclusion 9.1 Review 9.2 Theoretical Issues

272 272 273

Appendix

277

Notes

279

Bibliography

285

Index

293

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

viii

Table 1.1 Table 1.2 Table 2.1 Table 2.2 Table 2.3 Table 2.4 Table 2.5 Table 2.6 Table 3.1 Table 3.2 Table 4.1 Table 4.2 Table 4.3 Table 4.4 Table 4.5 Table 4.6 Table 4.7 Table 4.8 Table 4.9 Table 4.10 Table 4.11 Table 4.12 Table 4.13 Table 4.14 Table 4.15 Table 4.16 Table 4.17 Table 4.18 Table 4.19 Table 8.1 Table 8.2

IPA equivalents Idiosyncratic Symbols Consonant Inventory Vowels Glottalisation Lenition Quality Coalescence Quantity Coalescence Distribution of Roots Word Class Properties Indicative Paradigm Absolutive Paradigm Quotative Paradigm Interrogative Paradigm Purposive Paradigm Dubitative Paradigm Inferential Paradigm Conditional Paradigm Relative Mood Paradigms Subordinate Paradigms Indefinite Relative Paradigms Relative Dubitative Paradigm Present Imperative Paradigm Future Imperative Paradigm 'Come' Imperative Paradigm 'Go' Imperative Paradigm Use of - 'at Order of Inflectional Suffixes Aspectual Categories Lexical vs. Syntactic Incorporation Noun-Noun Collocations in Nuuchahnulth

IX

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

8 9 13 14 18 20 22 22 37 57 67 68 69 70 71 72 73 74 75 76 76 77 78 79 80 81 90 105 107 233 245-6

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Tables

Figure 1.1 Map of the Southern Wakashan Groups Figure 1.2 Example Format

x

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

6 9

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Figures

# lp. Is. 2sg. 3s. ABS Alv.. MW CAUS CND CVcdup CVXdup CVtdup DEF DIM DIST DUB DUP DUR O FUT gl. GRD.CAUS HYP.FUT IMPER INAL INC DSfD INDF.REL

INF

word boundary first person plural first person singular second person singular third person singular absolutive alveolar contemporaneous/ 'meanwhile' causative conditional reduplication with -c-infix reduplication with -X-infix reduplication with -t-infix definite diminutive distributive dubitative reduplication durative foot future glottal(ised) graduative causative hypothetical future imperative inalienable possession inceptive indicative indefinite relative inferential

INTENT INTER IRR ITER L lab. LOC M/MOM m.n. MC NOM NOW PASS PAST PL pin. POSS PURP QT RC REF REL REL.DUB REP RL rnd a son SUB SUF TRANS V VOC w.n.

intentive future interrogative irrealis iterative long vowel labial(ised) locative momentaneous aspect man's name momentaneous causative nominaliser temporal passive/switch reference past tense plural place name possessive purposive quotative relative clause referential stem relative relative dubitative repetitive aspect reduplication w/length round syllable sonorant subordinate suffix-triggered reduplication transitive variable length vowel vocative woman's name

XI

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Abbreviations

Many people have contributed to this work, either directly or indirectly, and it would be impossible to acknowledge everyone here. I hope I'll be forgiven if I do not name them all, but there are several acknowledgements that I must make because they stand out as special contributions to this work. First of all, I would like to acknowledge the extensive support given to me by the Arts and Humanities Research Board through a research grant, no. B/RG/AN7953/APN12323 to investigate the Nuuchahnulth language and, before this, the Hong Kong Research Grants Council research grant No. HKU 7183/97H to develop a 'Textual Database of Nootka'. Without the support of these funding bodies, this work would have taken substantially longer to complete. Secondly, I would like to acknowledge the support that I have received from the University of Newcastle and, in particular, the School of English Literature, Language and Linguistics. This made it possible for me to have the research leave that has proven crucial to the completion of this monograph. Winnie S.M. Yiu has been working on both the above-named research projects since the beginning and is responsible for much of the inputting, analysis, and translation of the Nuuchahnulth texts associated with the project. She has also been a colleague and co-author of a number of research papers cited here. Ben Braithwaite and Ben Thorp, members of the Nuuchahnulth project, have read previous versions of this manuscript and provided useful insights and comments on it. Many Nuuchahnulth people, past and present, have generously shared their knowledge of their language and this work, and many others, would have been impossible without their contribution. Jill Lake, Commissioning Editor for Palgrave, read through the entire manuscript, catching a number of potential gaffs and confusing wordings and made this a far more polished document. If many people have contributed to this work, one person in particular has had a significant influence on both the final form and content of this monograph: Eun-Sook Kim. She read over the entire manuscript and

xn 10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Acknowledgements

Acknowledgements xiii

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

commented on everything,fromthe accuracy of the data, to the style, to the theoretical substance. This work has benefited much from her involvement. Any errors herein are my own responsibility and due, no doubt, to my own stubborn resistance to valuable comments.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

This page intentionally left blank

1

1.1

Purpose of the Book

There has been an increasing amount of interest in the consequences of Nuuchahnulth morphology for theoretical linguistics. Numerous phenomena have been discussed in theses, monographs, conference papers, and journal articles, but there has been no overall presentation of the scope of Nuuchahnulth morphology and of how the various properties discussed in individual analyses might interact. The aims of this book are to draw together the various phenomena, to highlight their significance for linguistic theory, and to provide tentative analyses for many of the issues. Along the way, the general properties of Nuuchahnulth morphology will be elucidated in order to provide a fuller picture of the grammar of the language. As such, this book should serve as a sourcebook of Nuuchahnulth morphology, as an exploration of the range of challenging morphological phenomena that are encountered in the investigation of Nuuchahnulth, and as a first attempt at providing a consistent analysis of the numerous issues that arise. Hopefully it will encourage others to delve further into the nature of word-building processes in this highly polysynthetic language. 1.2

Sources

Every effort has been made to provide sufficient data to allow the reader to appreciate the structures under investigation and to permit alternative possible interpretations of the facts, so that one can evaluate the system for oneself. Naturally, there are space limitations on the number of examples that can be provided, but the hope is that these will not inhibit comprehensibility or opportunities for reanalysis.

1

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Introduction

Linguistic Theory and Complex Words

The bulk of the research presented here is based on the fieldnotes of Edward Sapir, gathered over a period of several years (1910-1914) from the Nuuchahnulth people of Port Alberni, supplemented by the subsequent materials sent to him by two of his native speaker consultants, Alex Thomas and Hamilton George, over the period from 1914-1924. This material represents a rather detailed account of the culture and language of the Nuuchahnulth people around the turn of the century, when the culture and language were still little influenced by the Europeans. There are over 140 texts of different lengths, comprising over 50,000 lines and 150,000 words in total. The majority of the speakers are Tsishaath, but there are also speakers of Hupachasath, Ucluelet, and Huiath dialects. All speakers were male, an obvious deficiency, although there is no clear evidence of strong gender-based differences in the language. This is, however, somewhat circular and obviously more research in this area would be of great importance in clarifying this issue. In addition to this material, data and, occasionally, analyses have been drawn from the work of other scholars working in the field of Wakashan linguistics. This data has been noted where it arises and a full list of references is provided in the bibliography. 1.3

Organisation

This book is organised in the following fashion: after this introductory chapter, some preliminary discussion of Nuuchahnulth phonology and morphophonology, important for understanding the complexities of the morphology, will be presented in Chapter 2, followed by a discussion of Word Structure and Categories in Chapter 3. This will include a detailed discussion of the place of the root in Nuuchahnulth grammar, the issue of bound forms of roots and combining forms of suffixes, the distinction between bound verbs and verbal suffixes, and the status of syntactic categories and category neutrality. Chapter 4 on Suffixation will discuss the various kinds of suffixes, distinguishing between inflection, derivation and aspect. Topics introduced include action-at-a-distance, affix-triggered reduplication, lengthening, shortening and cases of fixed segments. Chapter 5 on Reduplication and Infixation provides details of the many forms of reduplication found in Nuuchahnulth, including those occurring in inflection, derivation, and aspect marking, as well as a discussion of the more limited forms of infixation. This chapter also includes a

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

2

3

discussion of double reduplication in Nuuchahnulth and other varieties of Southern Wakashan. Chapter 6 presents arguments for dividing the grammar of Nuuchahnulth into two levels, stem and word, and argues for the advantages of a multi-stratal model of Optimality Theory in dealing with the issues raised, in particular the phenomenon of double reduplication. Stratal effects, including a number of phonological and morphological phenomena relevant for arguments for a multi-stratal approach to Optimality Theory, as advocated by Bermudez-Otero (1999, etc.) and Kiparsky (2000, etc.) are the topic of Chapter 6. These clearly demonstrate the need for a multi-stratal account of Nuuchahnulth morphology. Furthermore, the issue of rule-based versus constraintbased accounts of the grammar will be addressed in this chapter. Templatic Morphology is the topic of Chapter 7 and includes detailed discussion of a number of different areas of the grammar, including reduplication, infixation, hypocoristic formation, and abnormal speech, which provide evidence of the importance of the template for the morphology of Nuuchahnulth. Chapter 8, Morphosyntax, is rather broad and discusses a number of contexts where there is strong interaction between the morphology and the syntax, including cliticisation, incorporation, and compounding. Nuuchahnulth distinguishes between lexical and syntactic types of incorporation and exhibits a complex form of this process involving not only object head nouns, but also other members of the object phrase. A discussion of the treatment of inflection as cliticisation also features in this chapter. The final Chapter, 9, summarises the findings of the book, and ends with a brief presentation of a number of further topics of investigation. 1.4 Nuuchahnulth and Wakashan Nuuchahnulth is a member, along with Ditidaht (also known as Nitinaht, from the Nuuchahnulth term) and Makah, of the Southern Wakashan branch of the Wakashan family. The Northern branch of Wakashan is represented by Kwak'wala (previously referred to as Kwakiutl), a language made famous by Franz Boas, as well as Heiltsuk, Haisla, and Oowekeno, spoken on the corresponding areas of the mainland.1 The relationships within the Southern branch are quite clear, but there has yet to be any conclusive comparative work done between the branches. For

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Introduction

4

Linguistic Theory and Complex Words

further details of the relationships, Jacobsen (1979a) is a good starting point.

The actual number of dialects or varieties of Nuuchahnulth have been estimated at between 14 and 19, but these may be roughly divided into three geographical and dialectal variants, the Northern, Central, and Southern groups, following Drucker (1951). 1.5.1 Northern Nuuchahnulth The Northern varieties of Nuuchahnulth consist of those languages spoken in the region from Ucluelet Arm on the central coastline of Vancouver Island continuing north to Brookes Peninsula at the northern tip of the island. These include the following: Tribe Name Chiclesit Kyuquot Ehetisaht/Nuchatlath Mowachath Muchalath

Location Brookes Peninsula Kyuquot Sound Queen's Cove Nootka Sound Muchalat Arm

For more discussion of the grammar and phonology of one member of Northern Nuuchahnulth, Kyuquot, consult Rose (1981), a grammar based on extensive fieldwork conducted in the 1970s. Klokeid (ms) also discusses certain aspects of Kyuquot grammar but does not go into great detail. 1.5.2 Central Nuuchahnulth Central Nuuchahnulth consists of the varieties spoken from Hesquiaht Harbour in the north to Clayoquot Sound in the south. The varieties spoken in this region include: Tribe Name Hesquiaht Manhousaht Ahousaht Clayoquot

Location Hesquiat Harbour Maanu'is Flores Island Clayoquot Sound

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

1.5 Nuuchahnulth Dialect Research

5

There was early missionary work on Hesquiaht by Father A. Brabant, including a dictionary and catechism. Turner & Efrat (1982) discuss Hesquiaht ethnobotany and provide many botanical terms in the language. Similarly, a book by Ellis & Swan (1981) focuses on Manhousaht ethnobiology, and includes many terms from this variety that has now merged with Ahousaht. There has been a recent upsurge of work on Nuuchahnulth focussing on the Ahousaht dialect, which probably has the most speakers of any variety currently. Hess (1990), Nakayama (1997a, 2001), Kim (1999a,b, 2000a,b, 2001, 2002, 2003a,b,c), Wojdak (2000, 2001) are some of the works on this variety. Boas (1890) makes brief reference to Clayoquot, citing work by N. J. Lemmens. Paik (1968) collected fieldnotes on Clayoquot and wrote an MA thesis on its phonology, but there has been little significant subsequent work. 1.5.3 Southern Nuuchahnulth These varieties are spoken between the Ditidaht lands in the south and the Ucluelet group in the north. Tsishaath as spoken in Port Alberni, British Columbia, will be the principal focus of this volume and is probably the most widely studied variety of Nuuchahnulth in general. There is more data on this member of the family than on all of the other varieties combined. Nevertheless, most of the research stems from a single source, the field notes of Edward Sapir, gathered between 1910 and 1914, as discussed above. A number of individuals have written on aspects of Tsishaath grammar, including Boas (1890), Sapir (ms, 1911, 1915, etc.), Swadesh (1931, 1933, 1939, etc.), Stonham (1994, 1999, etc.), Davidson (2002), and others. The varieties spoken in this region include: Tribe Name Ucluelet Toquaht Tsishaath Huiath Uchucklesit Hupachasath

Location Ucluelet Harbour Barkley Sound

Sproat Lake

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Introduction

6

Linguistic Theory and Complex

Words

N A Chiclesit Kyuquot

Ehedsahtv Vancouver Nuchatlath Mowachath Muchalath Hesquiaht > -

TdflnH

Ahousaht * v

lsiana

Clayoquot Ucluelet ^ \

Hupachasath

Toquaht v*~ (J Tsishaath ' ( \ P i t i d a h t Uchucklesit ^^acheedath Huiath Makah

Figure 1.1 Map of the Southern Wakashan Groups 1.6 Previous Literature There has been little work published previously on Southern Wakashan morphology, but there are some notable works that should be mentioned. The earliest treatment of Nuuchahnulth morphology may be that of Knipe (1868), in part echoed in Sproat (1868).2 Knipe shows a keen understanding of the language although many of his attempts at decomposing the morphology extend to the fanciful, as demonstrated by the example below: sinnamooxyets. The name of a berry. It contains the four roots, si, nah, mooxyeh, yets, which mean severally distance, sight, rock, walk. The name, of course, implies that the first person who discovered the berries walked over the rocks a long distance to look for them — a description which quite agrees with the locality of the berry. It may be noticed that

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

For a description of aspects of the grammar of these varieties, the reader is referred to the various works of Sapir, Swadesh, Jacobsen, and Stonham cited in the references.

Introduction

1

Clearly, this is a rather fanciful interpretation of the form sinmuxsyac 'evergreen huckleberry' but other aspects of Knipe's book show an insight into the language, recognizing a number of roots and suffixes, as well as discussing the process of reduplication and a brief examination of Ditidaht grammar. Subsequent to this, Boas (1890) presents a sketch of the morphology of Nuuchahnulth, presenting a number of grammatical paradigms and discussion of various morphological processes, including the observation that certain suffixes require reduplication (ibid: 124), the affix-triggered reduplication that will be discussed in greater detail in Chapter 5. Sapir (1911) debates the issue of noun incorporation in a series of exchanges with Kroeber (1910, 1911), while Sapir (1915) presents a description of the speech mutations involved in the production of abnormal speech. A manuscript containing detailed discussion of Nuuchahnulth morphology was written by Morris Swadesh in 1937 and both his Master's thesis (1931) and doctoral thesis (1933) provide further discussion. A published version, Swadesh (1939), provides the earliest accessible modern treatment of the morphology. Sapir & Swadesh (1939) briefly discuss the morphology in their first book of Nuuchahnulth texts. After a fallow period for morphological research in Nuuchahnulth of approximately 30 years, Haas (1969) discusses the topic of pronominal inflection in Nuuchahnulth and Haas (1972) discusses the structure of roots and stems in Nuuchahnulth. Rose (1976) discusses the issue of glottalisation and Jacobsen (1979b) presents an important paper discussing the nature of the noun and verb in Southern Wakashan. Rose (1981) constitutes an important contribution to our understanding of Nuuchahnulth grammar, and particularly the northern variety of Kyuquot. Rose & Carlson (1984) begin the debate on the nature of the suffix -'at, joined subsequently by Whistler (1985), and Emanatian (1988). Renker (1988) returns to the debate concerning noun and verb in Southern Wakashan. Stonham (1990a) discusses variable-length vowels while Stonham (1991) discusses the templatic nature of hypocoristic formation and Jacobsen (1994) discusses vocative vocalism. Stonham

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

the root yets, here meaning to walk, enters into yetseyetsokleh [yaacyaadaqXi 'kicking at the rear'?], mentioned above, with the meaning of kicking. The full word meaning to kick is yetshitl and to walk yetsook, or, as it is more frequently pronounced, yatsook. (Knipe 1868:14)

(1990b) presents analyses of several issues in Southern Wakashan, expanded upon in Stonham (1994b). Nakayama (1997a) investigates a number of topics in the grammar of the central dialect, Ahousaht. Nakayama (1997b) returns to the issue of -'at, as does Kim (2000b). Stonham (1998) discusses numeral incorporation and Stonham (1999a) the nature of noun compounding. Davidson (2002) discusses a number of issues in Southern Wakashan, while Stonham & Yiu (2002a) discuss the characteristics of the word in Nuuchahnulth and Yiu & Stonham (2002b) discuss the use of classifiers. Stonham (2003a) discusses double reduplication in Southern Wakashan, while Stonham (2004) discusses level-ordering in the morphology of Nuuchahnulth. Kim (2003c) provides a detailed account of various issues in Nuuchahnulth phonology and morphology in terms of Optimality Theory. Recently there has been an upsurge in interest in Nuuchahnulth morphology expressed by the linguistic community, with a number of theses, conference papers, and unpublished materials, as well as some published work discussing various aspects of the language, including Yiu (2000), Kim (2003a, b, c), Davidson (2002), Wojdak (2000, 2001). 1.7 Transcription System A brief discussion of the transcription system employed here, which is a somewhat modified version of the IP A, is in order at this point. The basic modifications employed here concern the treatment of the palatal glide, which will be represented here by [y], the laryngealised sonorants, and the affricates, which will be represented by single symbols in keeping with their distinct nature, as indicated in Table 1. is book c £ s \ X X

s

IPA ts tj

I

is book d <5

IPA ts'

tr j

\

y y

a

rii

tr

A h

m' n' h

s

j"

Table 1.1 IPA equivalents

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8 Linguistic Theory and Complex Words

Introduction

9

V-

variable-length vowel glottalising suffix leniting suffix clitic boundary inflectional boundary

-' = ||

Table 1.2 Idiosyncratic Symbols Where examples assume information not yet encountered, as may necessarily occur in the description of a complex system such as this one, cross-references to chapter, section, will be provided for the reader. In glossing morphemes, inflectional and aspectual morphemes appear in small capital letters, typically abbreviated as MOM 'momentaneous aspect' or PAST 'past tense'. Derivational morphemes appear as glosses, surrounded by single quotes, thus hut 'throw liquid'. Roots have no hyphens associated with them while suffixes are preceded by a hyphen, and reduplicative copies, always prefixal in position, are followed by a hyphen. Infixes have hyphens on either side, i.e., -t- PL. A potential source of confusion in looking at the examples is the possibility of root and affix allomorphy which occurs quite commonly in Nuuchahnulth, and Southern Wakashan more generally. These languages are highly polysynthetic and there are many opportunities for allomorphy to arise. In this book the most basic allomorph of a morpheme has been chosen for the morphemic breakdown, except in those cases where it would cause some confusion. In such cases, the allomorph closest to the surface form is used. With respect to the format of examples, each example is presented in the following fashion: Example Number

Phonemic Transcription qwisit?itq q^is -(m)it -?irtq

e. ?iiqhulTaX ?iiqh-uk-'aX

\

Translation

'he told what he went through' tell-DUR-NOW do so-PAST-3s.REL (Ahous)

t

Morpheme Breakdown

Morpheme-by-Morpheme Gloss

f Data Source

Figure 1.2 Example Format

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

In addition to these changes, we will employ certain non-standard symbols that have come to be used as standard in the field of Wakashan studies as well as several others which are useful but non-standard. Such symbols will be explained in greater detail as they are encountered, but the table below outlines the most common ones.

Examples are numbered consecutively throughout a chapter, using a combination of numerals for examples and letters for sub-examples. Each example consists of a phonemic transcription, a translation, a morpheme breakdown that uses the basic allomorph of each morpheme as the representative, and a morpheme-by-morpheme gloss corresponding to the breakdown. All data is drawn from Sapir's Tsishaath Nuuchahnulth materials unless otherwise noted. Shorthand representation for other data includes 'Ahous' for the Ahousaht dialect materials in Nakayama (1997a) and 'Kyu' for the Kyuquot dialect materials in Rose (1981). Other sources are noted as they arise. Glosses are typically rather terse owing to the need for space in the interlinear glossing and so may not exactly match the free translation provided. They may also vary slightly depending on the context, in order to make the relationships among morphemes clearer. Similarly, the allomorph representing the morpheme may not be the one appearing in the example, which arises from the application of various phonological and morpho-phonological rules. 1.8

On the Term 'Nuuchahnulth'

A brief word is perhaps in order here concerning the term 'Nuuchahnulth'. This is an anglicised version of a native word, nuudaanul, meaning 'along the mountains.' It is neither a traditional word for the people nor for the language they speak. In the approximately 50,000 lines of text assembled from the texts gathered by Sapir, the term appears once, designating land that belongs to the people. The traditional term for the language is taataaqsapa 'speaking true or straight' and this has been used by native speakers in both Sapir's texts and in other publications.3 Traditionally, each band would designate themselves by the placename most closely associated with the specific group, for instance the Tsishaath are the people of Tsishaa, i.e. ci§aa?ath from cisaa -'ath 'people of ...'. Sproat (1868) proposes the use of 'Aht', based on this suffix, as a descriptor for the whole group of people and their language. This has not achieved wide currency in the literature. While there have been numerous alliances and confederations formed over the years, there is no evidence of there ever having been a single unified grouping of all Southern Wakashan people on Vancouver Island, although there has always been interaction among the groups (see, for example, Drucker 1951).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

10 Linguistic Theory and Complex Words

11

The term 'Nuuchahnulth' was agreed upon by a Tribal Council decision in the late 1970s4 as a term to designate all of the Southern Wakashan people of Vancouver Island. As such it is a socio-political designator not a linguistic one and, in fact, poses difficulties in this area, since it is employed to designate not only the people formerly referred to as the 'Nootka'5 but also their more southerly neighbour, the Ditidaht. The term 'Nootka' is attributed to James Cook, who made early contact with the Nuuchahnulth people of Friendly Cove (yuk^aath), the Mowachath. There are various claims as to its origin and it is difficult to say which is right, but the name is clearly not a Nuuchahnulth one. This form of mis-appellation is not uncommon in the languages of the world, 'English' being a designation derived from the misnomer 'land of the Angles', French 'the land of the Franks' (a Germanic group), and Japan, apparently derived from a romanisation, Jih-pun, of the Chinese pronunciation, [r-p9n], of the original Japanese self-designation, Ni-pon.6 Nevertheless, it is important to be aware of the potential confusion that may arise from utilising a term as broadly construed as 'Nuuchahnulth' in the description of linguistic affiliation.7 In this book 'Nuuchahnulth' will be utilised as follows: as a term for the language of the most northerly of the Southern Wakashan groups, north of the Ditidaht-speaking people, as discussed above. As such, it covers those varieties of Southern Wakashan formerly described as 'Nootka'. Where it is deemed necessary to be more specific, the particular group will be combined with this general descriptor, e.g., 'Tsishaath Nuuchahnulth'. It will also be used in its adjectival form to produce various derivatives such as the Nuuchahnulth people, Nuuchahnulth grammar, etc. In order to describe all of the groups on the west coast of Vancouver Island, the term West Coast People will be employed.8 The term 'Nootkan' will still be used as a descriptor for the various Nuuchahnulth groups. 1.9

Coverage

In addition to those interested in the language itself, it is intended that this book will be useful to several groups of individuals, specifically: (i) typologists, interested in the fuller description of the morphological system of a Wakashan language; (ii) morphologists, seeking data and phenomena for testing and evaluating their theories of morphology; and

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Introduction

(iii) students of morphology interested in learning about the progression from raw data to the analysis of a language's morphological system. Hopefully, the data will be sufficient for the reader to evaluate the phenomena but, should this prove not to be the case, there is a wealth of additional material available in Sapir's fieldnotes (Sapir ms.), ready for the enterprising linguist to investigate in order to clarify whatever issues might suggest themselves.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

12 Linguistic Theory and Complex Words

Nuuchahnulth has a large inventory of phonemes and a complex set of phonological and morphophonological processes, including some quite unique to this language group. In this chapter we will briefly examine those aspects of the phonology that will be most useful for proceeding with an investigation of the morphology of the language. For a more detailed exposition of the phonetics and phonology of Nuuchahnulth, see Stonham (1999b). 2.1

Consonants

There are 39 consonant phonemes in Nuuchahnulth. These include the regular phonemes of the language, excluding special sounds introduced in non-nativised foreign borrowings, rhetorical modifications, song texts and the like. The following table exemplifies this system:1 Labial Alv.

Stop Gl. Stop Lab. Stop Gl. Lab. Stop Fricative Lab. Fricative Plain Affricate Gl. Affricate Sonorants Gl. Sonorants

P

i

c

a m

Velar Uvular Phar. Glot tal k k k*

t i

s

rii

Alv. Pal. Lat. Obstruents

s

i

q q

?

'
rx 0)

r

?

X


f

h \f

X 5 X t Sonorants

n > n

y y

w vv

Table 2.1 Consonant Inventory

13 10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

h

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Some Phonological Preliminaries

14 Linguistic Theory and Complex Words

As can be seen from the table, the system is quite symmetrical, with pairs distinguished on the basis of glottalisation or labialisation. Vowels

The phonemic vowel system is quite simple, involving just the three vowels, /i, u, a/, and a length distinction. There are, in addition, two mid vowels which are only encountered in the long variety, at least phonologically, and even then only under very special circumstances, such as in foreign borrowings and special processes. n

u

uu ee

00

aa Table 2.2 Vowels Phonetically, long and short mid vowels occur frequently in some varieties, for instance Tsishaath, but not in others, e.g., Ahousaht. These mid vowels are the result of coalescence and spreading. 2.3 The Structure of the Syllable All syllables begin with one and only one consonant that may be any of the consonants found in the language, /h/ can occur only word-initially and in the reduplicative copy of an /h/-initial root. The onset of the syllable is followed by a vowel that may be long or short. Following this, there may be from zero to three, or in very rare cases four, consonants in the coda, as indicated in (1). Note that all consonants involving the glottis or glottalic action are prohibited in the coda, indicated in (1) by the use of the informal feature, [± Glottal]. 1.

Syllable Structure

o

O(nset)

R(hyme) N(ucleus) V

(V)

C(oda) (C)(C)(C) r- Glottal!

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

2.2

15

One issue that is not resolved immediately by this syllable structure is the absence of long vowels before a coda nasal, as first discussed in Wilson (1985). A possible solution to this problem is to allow the nasal a special status in Nuuchahnulth, more intimately connected with the nucleus than other consonants which obey the standard restrictions on codas, for discussion of which see Stonham (1994a, 1999b). 2.4

Stress

In Nuuchahnulth the extent of the domain of primary stress is the first two syllables of the word, as demonstrated in (2). 2.

a. cuusuk^iX

'begin to suspect'

b. kuuhsinqinr?ap

'always cause hole to be in the side'

c. qahnaakaX

'someone now died'

d. hayaa?aksi?aX

'did not know now'

e. hiiyiih

'be after blood'

As can be seen from these data, stress always appears on the first or second vowel of the word, regardless of the presence of long vowels later in the form. It is also the leftmost of two long vowels that is assigned the stress in the case where there are long vowels in both of the first two syllables of the word, e.g., (2e). Now the question arises as to what happens if the word contains no long vowels. Examples of this are provided in (3). 3.

a. t'ananak

'have a child'

b. nayaqak

'baby'

c. wa?icupaXisak

'come and cause me now to sleep!'

In this case, stress is assigned to the leftmost vowel. Note that this result obtains even in the case where there is a long vowel in subsequent syllables. 4.

a. tanakmiici?aX

'turned now into mosquitoes'

b. hu?aca£iX?aaqXma

'he will come back'

c. qwayaciiktaqimi

'Wolf band'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Some Phonological Preliminaries

16 Linguistic Theory and Complex Words

5.

a. ?inksyiquu

'make a fire'

b. hisimyuup

'gather together'

c. cimsrhiit

'Bear-son'

As can be seen from this data, stress appears on the first short vowel if it is immediately followed by a sonorant in the same syllable, otherwise on the second short vowel if it is followed by a tautosyllabic sonorant. If a long vowel occurs in the first syllable, it bears the stress. If a syllable with a long vowel is preceded by a sequence of vowel-sonorant, it is the vowel-sonorant sequence in the first syllable that bears the stress. All of this seems to suggest that for the purposes of stress assignment, either a long vowel or a vowel-sonorant sequence consist of two moras and, accordingly, attract the stress from a mono-moraic syllable. The domain of the stress rule for Nuuchahnulth is the first foot. The rule can thus be described as: Nuuchahnulth Stress Rule3 Stress the first syllable, unless the second syllable is heavier. 2.5 Phonological Processes As has already been suggested with respect to primary stress placement, in Nuuchahnulth metrical structure above the level of the syllable can be important for the application of phonological rules. In particular, the first foot appears to constitute a special domain for rule application and there are a number of differences between the way rules work in this domain and outside of it. In the following sections we will examine some of the aspects of the phonology of Nuuchahnulth relevant to the investigation of the morphology. 2.5.1

Glottalisation

The process of glottalisation and that of lenition to be discussed in the following section are quite unique to the Wakashan languages and were first described for Kwak'wala, a Northern Wakashan language, by Franz

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

However, when we examine forms containing a sonorant consonant in the coda of one of the first two syllables,2 we can see that its presence triggers the attraction of stress, even when a long vowel is available to attract the stress, as in (5c).

Some Phonological Preliminaries

17

Suffixes influence the terminal sounds of stems, which they often harden or soften. When softened, surd and fortis are transformed into the sonans of the same series; when hardened, sonans and surd are transformed into the fortis. s softened becomes dz or y; hardened, it becomes ts!. x- softened and hardened becomes n; Xw softened and hardened becomes w. L softened and hardened becomes 1. n, m, 1, y, w, when softened become sonant by being preceded by the laryngeal catch. The following examples will illustrate these processes: Stem Softened Hardened L!aqw-, red L!aa'gw-atoo, red-eared L!aa'q!w-ooboo, red-breasted hanL-, to shoot ha'nl-as, place of shooting ha'nL!-aala, noise of shooting qas-, to walk qaa'y-as place of walking qaa'ts!-eenox, walker mix*- to strike mEn-a'ts!ee, drum striking mE'n€xst, to strike hind end seeXw-, to paddle see'w-ayu, paddle see'€w-eenox, paddler ts!ooX- black ts!oo'l-is, black beach ts!oo'€l-a, black rock The processes of glottalisation and lenition in Nuuchahnulth are quite similar to those found in Kwak'wala and their realisation will be described in this and the following section. For a more detailed account of these processes, see Rose (1981) for Kyuquot, Stonham (1999b) for Tsishaath, and Kim (2003 c) for Ahousaht. Glottalisation involves the interaction of a specific subset of suffixes with a preceding segment. The outcome will vary depending on the segment involved, but typically results in the change of stops to globalised stops, fricatives to laryngealised glides, and sonorants to their laryngealised counterparts. For further details of the phonetic aspects of this process, see Kim (2003c). The following chart illustrates the various changes that may occur:

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Boas (e.g., Boas 1900, 1911, 1947). Boas (1900:710) states the facts for Kwak'wala in the following fashion:

18 Linguistic Theory and Complex Words

tcxek r

Glottalisation Pfcxi k r T

qq"

—> —>

S TS

->

y —

-> —>

w rh n W y

xxhh

r x* x10 r mnwy

Table 2.3 Glottalisation Basically, glottalisation is 'triggered' by some characteristic of the class of suffixes which Sapir & Swadesh (1939) represent by a globalising mark /'/ at the beginning of the appropriate suffix, e.g., -'at PASS. Such suffixes are reasonably common in the language and instances of the process are frequent. These suffixes all begin with a vowel, although not all vowel-initial suffixes trigger the glottalisation. How one should appropriately distinguish those suffixes that do from those which do not cause this process is, perhaps, a theory-internal issue. However one does it, it will obviously be necessary to make some distinction between the two categories of suffixes.4 Examples of suffixes that trigger glottalisation (a) and those which do not (b) are provided in (6). 6.

a. -'aX

NOW

-'ap

CAUS

-'iX -'urn

'invite, go for, take' 'on the rocks'

b. -aX -ap -iX -im

'receptacle' 'on a point' MC 'thing'

Examples of how these suffixes combine with preceding elements are provided below. 7.

a. qahnaakaX qah -na-k10 -'aX5 wikapwe?in wik -'ap -weY?in

'he was dead now' dead -have -NOW 'it hadn't been ...' not -CAUS -3s.QT

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Segments

Some Phonological Preliminaries 19

wiskum wisk -'urn b. k^anux^aX k^anux01 -aX

'they started out of the woods' LOC-exit woods-go for[L]6-MOM-NOW 'angry on the rocks' angry -on the rocks 'chamber-pot' urinate (female) -receptacle

friqatapas Tiqat -apas

'dripping from a standing (bush)' drip -stand on ground

?aayimkiXqas ?aya -miik -iX -qa*s

'may I be a getter of many...' many -getter of -MC -IS.SUB

titinkum DUP- ti -nuk0 -im

'handwiper' SUF- wipe -at the hand [R] -thing

As can be seen from these examples, the result of the combination of a glottalising suffix with a preceding consonant-final base is the creation of a globalised version of the consonant at the end of the base. Where there is no available appropriate version, the nearest possible consonant is provided, i.e., for labialised fricatives, a laryngealised labiovelar glide, /w/, and for coronal fricatives, a laryngealised coronal glide, /y/. The data follow the pattern below. hasiii?aaqX'apqa hasri -?aaqX -'ap -qa*

'it will be stuck in his throat' bone in throat -INTENT-CAUS -SUB

wikin wik -'in

'do not do to us' not -lp.lMP

yaaqs5aS"in?itqak yaq* -qs -£aq -'in -?i*tqak

'one who is with you in the canoe' REL -in vessel-..-ed-treated as-2s.REL

Xiiwin?ap Xiix10 -'in -'ap

'he laughed at him' laugh -sound of -CAUS

<

'they gave us deer to eat' deer-eat-CAUS -PASS-PAST-lP.iND

Tatuyis?arjaanitin Taatus -'rs-'ap-'at-mit-min

maaXiicu?ayin 'he had been seen wearing it around his head' maX -(w)iic-(y)u?ai-'in tied -around head[L]-perceive-treated as

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

hiitahtiXsi?aX hita -ht -'iX -§iX -'aX

20 Linguistic Theory and Complex Words

When the base ends in a consonant which is not a possible candidate for glottalisation, such as the [+back] consonants /x x h h/, then no change is found to occur. ?ah?in?ak S'ah -'in -'ak

'his diver' dive -treated as -POSS

Where the base ends in a vowel, the result is different. In such a case, a glottal stop appears between the two vowels, preventing coalescence. 10.

qaahk^aTap qah -k^a -'ap

'he was really dead' dead -completely -CAUS

q^ee?^ q^aa -'in

'let us do so!' thus -lp.iMP

It should also be noted that the application of the process to fricatives and nasals occurs only within the derivational morphology. At the point of inflectional morphology, there will be no change to laryngealised glides. Sapir & Swadesh (1939:236) state: ... while globalizing formative suffixes [derivational] ... change voiceless fricatives to globalized semivowels, glottalizing incremental suffixes [inflectional] affect only nonglottalized stops The theoretical importance of glottalisation and its relevance to the arguments for stratal segregation will be discussed in greater detail in Chapter 6 on level ordering. 2.5.2 Lenition A parallel process to glottalisation is that of lenition, which affects fricatives, converting them to glides in the following fashion: Segments sis

Lenition -»

y

-»

w

xxhh

r x" x \f

Table 2.4 Lenition

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

9.

Some Phonological Preliminaries

21

11.

-'aciX -'is -'ii -'(i)saX

MOM 'on the beach' 'in the house' 'on the beach (M)'

The following examples illustrate the differences between the suffixes that cause lenition (12a) and those that trigger glottalisation (12b), specifically with respect to fricatives. 12. a. ca?uwis da?ui(0 -'is ca?uyis da?ui -'is b. qawa^is qawas -'f s caqiicqimyis caqix -qimi -'rs

'waves rolling on the beach' waves rolling -on the beach 'waves rolling on the beach' waves rolling -on the beach7 'eating salmonberries' salmonberry -eat 'it cost twenty dollars' twenty -CLS-consume

This process will prove to be important for the argument for stratal separation and will be discussed with this in mind in Chapter 6 on level ordering. 2.5.3

Coalescence

Coalescence is the merger of two or more elements into a single one in some specific environment. There are a number of cases of coalescence or similar phenomena in Nuuchahnulth that complicate the morphophonology of the language to a great extent. Vowel coalescence arises when two morphemes are juxtaposed, the first ending in a vowel, the second beginning in one. There are two aspects to the coalescence of vowels: (i) vowel quality and (ii) vowel quantity. With regard to (i), the rule appears to be that any vowel conjoined with IvJ yields /u/, any non/u/ vowel conjoined with IM yields IM and only /a/ + /a/ yields /a/, thus:

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

This process is much more restricted in terms of its occurrence, appearing with only a limited number of suffixes of which the following represent nearly an exhaustive list.

22 Linguistic Theory and Complex Words -,nd

u i i

u i a

Table 2.5 Quality Coalescence Examples of these vowel mergers are plentiful and pose particular challenges to the analysis of Nuuchahnulth morphology, creating numerous potential ambiguities in the data. Examples of coalescence appear below. 13. a. ?uuktaqaX ?u -taqa -'aX

'he did it for that reason' REF -do to because of [L]-NOW

b. ?uucaX ?u -iic -'aX

'it belonged to her' REF -belong to -NOW

c. hitinqish hita -inqis -(q)h

'they were at the beach' LOC -at beach -MW

d. ?a?aaqiyukhak DUP- ?aqi -ayuk

'why are you crying?' SUF- what? -cry for [R+L]-2s.lNTER

e. ?uunuu?aXhwe?in ?u -a-nuuX-'aX-(q)h-we'?in

'that is why ...' REF -because of -NOW -MW -3S.QT

In most cases, (ii), i.e., quantity coalescence, is best represented as a simple merger of nuclei, the result mirroring the length of the heaviest nucleus, either long or short. Variable length vowels are long only in the first two syllables of the word (represented by PI—refer to Section 2.5.5 for discussion of variable vowels), so any coalescence of such vowels occurring after that point will necessarily result only in short vowels, the reflex of both short and variable length vowels. So we have the following possible results in the case of vowels coalescing outside the first foot of the word:

V +v =v v + v =v

v- + v =v Y

V + V: = V :

Y

V + V =v VY + V: =V:

V:+V =V V: + V- = V V: + V: = V

Table 2.6 Quantity Coalescence Examples of such mergers occur frequently, as demonstrated in (14).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

u u u

Some Phonological Preliminaries 23 'he was unhappy' poor -as far as -inside[MC] -NOW

b. ?uucaX ?u -iic -'aX

'it belonged to her' REF -belong to -NOW

c. pu^ai?atuu5i?aXwe?in pu^ai?atu -ii&X-'aX-we^in

'she became sleepy' sleepy -INC -NOW -3S.QT

d. hitinqish hita -inqis -(q)h

'they were at the beach' LOC -at the beach -MW

Note that the result of coalescence may be either a variable vowel, as in (14a,b) or a long vowel, as in (14c). (14d) demonstrates the constraint on long vowels preceding a coda nasal: the result is a short vowel even though it is predicted to be long according to Table 2.6. In the first foot there is a special, additional consideration: the result of coalescence cannot be shorter than variable length. What this means is that two short vowels coalescing in the first foot result in a variable length vowel, which will appear long unless later reduplications push the vowel into the third or later syllable. Thus, we have examples such as the following, where (15a) is an example of two short, identical vowels coalescing between the first and second syllables, (15b) is an example of the same situation with unlike vowels, (15c) is a case of coalescence between the second and third syllables, and (15d) is a case where the coalescence is in the third and fourth syllables, resulting in a short vowel. Notice that in (15e) we have a case of a variable length vowel coalescing with a short one in the third syllable: the result is a short vowel. 15. a. maaidaas ma -aida -'as

'house against wall on the ground' dwell-at upright surface-on ground

b. wiis wi-as

'failing to reach' not-reaching to

c. fiapaatah DUP- pa -atah

'ready to potlatch' potlatch -ready to [R]

d. ?u?u?utah DUP- DUP- ?u -atah •

'whalers here and there' DISTRIB- REF- try to get [R]

e. q^aq^arhiyaqh DUP- q^arha* -iyaqh

'singing thus many songs' SUF- thus many -sing a song [R]

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

14. a. wirhaaqstu?aX wi -rhav -'aqstuX -'aX

24 Linguistic Theory and Complex Words This leads us to the observation that coalescence in the first and second syllables is different from that in later syllables, and that it may be tied to the difference in syllable structure between the first foot and later syllables. ?-Epenthesis

Epenthesis of glottal stop is found specifically when coalescence might otherwise occur between root and suffix, potentially obscuring the underlying nature of the root vowel. Note the following examples. 16. a.

fi?uuf3aXicas fi -u*p -'aX -'r£as

b. nuu?i?aXwe?in nuu -iX -'aX -we^Pin

'throw it to me now!' throw -MC -NOW -2>1.IMP 'they sang' sing -MC -NOW -3S.QT

As can be seen from such examples, the /?/ serves to protect the integrity of both root and suffix vowels which might otherwise be lost in coalescence. However, it is not always the case that ?-epenthesis occurs where one might expect and the exact mechanism for determining when ?-epenthesis occurs and when it does not remains unclear. See Kim (2003c) for further discussion of this issue. 2.5.5 Variable-length Vowels The phenomenon of variable-length vowels, described in Sapir & Swadesh (1939), continues to elude a viable theoretical explanation in the linguistic literature.8 The argument for variable-length vowels takes the following form. As mentioned above, Nuuchahnulth has a phonemic distinction between long and short vowels. In addition to this, the vowels of certain morphemes vacillate, appearing sometimes long, sometimes short. It is generally agreed that this chameleon-like behaviour is tied to the position of the pertinent vowel within the word. A variable-length vowel, which is long in the first two syllables of the word, is short when it is in the third or later syllable of the word as shown in the following example. 17.

?unaak Pu-na-k"

'possessing it' REF -have

capacnak capac-na^kw

'possessing a canoe' canoe -have

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

2.5.4

Some Phonological Preliminaries

25

18.

-'as -i -?ai -ui

'on ground' 'quality of 'aware of 'place'

-a*s 'daughter' -r

DUR

-?a*i 'on a surface' -u*i 'on the face'

-'aas 'at the wrist' INC.ITER -ii -?aai 'to come off -uui ITER

While this distinction has been primarily discussed in terms of suffixes, Morris Swadesh, in an unpublished manuscript (Swadesh 1937), has suggested that it can also be a property of certain roots. The difficulty with detecting such roots in Southern Wakashan is that they must undergo double reduplication (for which see Section 6.3) in order to be placed in the third syllable of the word, where the distinction may be observed.9 Swadesh provides the following examples of this phenomenon: 19. a. Piihtuup ?rh" -turp ?i?i?ihy iml DUP- DUP- ?'v\f -yimi b. yaaksi?aXat ya-k-siX -'aX -'at y ay ay aksuuh DUP- DUP- ya^* -suuh

'whale (= big thing)' big -creature ' big-shouldered' DISTRIB- big -at the shoulder [R] 'became sore' sore-MOM-NOW-PASS ' sore-eyed' DISTRIB- sore-at the eye [R]

We can see from these examples that variable-length may be a property of the root vowel as well as the suffix, although it has not proved possible to find exact minimal triplets among the roots. How do we describe such a property? The solution offered by Sapir & Swadesh (1939) is to allow a threeway distinction among vowels, representing them as short /V/, long N\l or variable /V7. While this accounts for the data, given a rule which shortens or lengthens the variable length forms depending on position, one would prefer not to allow a three-way lexically distinctive contrast in vowel length, since this is unattested in the languages of the world and is not evident in the surface form of Tsishaath vowels, where we only find a two-way distinction of long versus short.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Given this third possibility, we would expect to find minimal triplets, where there is a three-way distinction in vowel length. And we do, in fact, encounter this readily among the suffixes, as is shown in (18).

Another possible solution to the problem is in terms of stress placement, saying for instance that the vowel appears long when stressed and short when not. However, we cannot appeal to the stress system of the language as it operates independently of this phenomenon and unstressed vowels are found to occur both long and short, and in any position in the word. Furthermore, variable length vowels do not necessarily correlate with the stressed vowel of a word. It is not within the scope of this book to delve any further into this primarily phonological issue and so the reader is referred to other treatments of this topic, including Stonham (1994a, 1999b), Davidson (2002), and Kim (2003c). 2.5.6 Assimilation Rules As in many other languages, there are a number of assimilation processes that occur in Nuuchahnulth. The following sections will present the major types encountered in the data. In some varieties of Nuuchahnulth, there is a post-lexical rule of assimilation which raises and fronts a low vowel /a/ to [e] when it is followed by IM across a glottal stop /?/. Thus we have examples such as the following: 20.

rhaarhiiqsuse?i rhaarhiiqsu -sa =?i*

'only the oldest child' oldest child -just =DEF

iuucsme?i iuucsma =?iY

'the woman' woman =DEF

naacse?isim naacsa -'rsim

'look at me!' look -2s>l.lMP

There is a group of processes which share a common change, reducing a sequence of vowel + glide + vowel to a single vowel which is typically variable-length and bears the quality of the glide which separates the two vowels, thus /w/ —> /u/ and lyl —> /e/, as exemplified in (21). 21. a. naaXkuuci?aXat naXk-uwa-siX-'aX-'at b. wikiip?aaqXe?ic wik -iiyip -?aaqX -meY?ic

'they were putting their feet out together' have foot out-together[L]-MOM-NOW-PASS 'you will fail to get it' not -obtain -INTENT -2S.IND

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

26 Linguistic Theory and Complex Words

c. hinee?i?aXwe?in hina -ayii?iX -'aX -we*?in

'she entered the house' LOC -enter house -NOW -3S.QT

d. ?a?aquu?ishak DUP- ?aqi -wa- -?is -ha'k

'What did you say?' REP- what? -say -DIM -2S.INT

e. wiinapuX wiinapi -awiX

'she stopped' stop -MOM

It is also possible for vowels to coalesce across glottal stop, and this may perhaps be another instance of the assimilation discussed above. Examples of the application of this include: 22. a. wirhaaqstu?aX wi -rha~ -'aqstuX -'aX

'she was unhappy' poor -as far as -inside[MC] -NOW

b. t'anaak?i fana -'ak =?r

'his child' child -POSS =DEF

c. ?uyaaqXe?ic ?u -(y)r -?aaqX -meY?ic

'you will do it then' REF -time -INTENT -2S.IND

d. wiinaaXXaa wiina -'aX =Xaa

'they attacked again' attack -NOW =again

Curiously, there is a certain indeterminacy in the occurrence of this phenomenon: examples of it not taking place are as numerous as those where it does. For the present, the exact nature of the conditioning factors determining the applicability of the rule remains unclear. 2.5.7

Labialisation

In this section we will discuss the process of labialisation, whereby certain consonants undergo a change to their labialised counterpart. Nuuchahnulth possesses both plain and labialised variants of a number of consonants, both stops and fricatives in more or less free distribution.10 However, there is one environment where labialisation invariably appears — that is when a velar or uvular consonant appears between a preceding high back rounded vowel and a following nonround vowel, where all elements belong to derivational morphology. The following are instances of this process, and involve originally non-labial consonants becoming labial via this mechanism. 23. a. hayuqumi?aX hayu -qimi -'aX

'ten units' ten -CLS -NOW11

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Some Phonological Preliminaries 27

b. nupuqumiuk^itqa nupu -qimi -uk -(m)it -qaY

'six units' Six -CLS -POSS -PAST -3.SUB

c. haPukqathPatuFeePit ha?uk -qa*th-(y)a*tuk-'ee?ita

'he made the noise of eating' eat -pretendedly-make noise-3.PURP

The difficulty in providing convincing examples of this process is not that there are few examples available but rather that, of the large number of extant examples, the majority are difficult to establish as not involving underlying labialised consonants. For discussion of these difficulties, see Stonham (1999b) and for a theoretical examination and treatment of labialisation, see Kim (2003c). The theoretical relevance of labialisation for stratal segregation will be discussed further in Chapter 6 on level ordering. 2.5.8 ?-Deletion A deletion rule which interacts with coalescence involves sequences of V?V, where V is a short vowel. If this sequence occurs between the first and second syllable of the word, nothing happens, i.e., the sequence remains intact, as in (24a). If, however, the sequence occurs later in the word, then the /?/ disappears and the vowels coalesce according to the rules of coalescence stated above. The results of this rule are the coalescence of the vowels across a glottal stop, as in (24b, c). Xa -'as

'sticking up on the ground' stick up -on the ground

waa?aXwe?in waa -'aX -we*?in

'he said it' say -NOW -3S.QT

24. a. Xa?as

b. hawaas hawa -'as ma&iiafaX ma -ci -'ii -'at -'aX c. ?uwiihtas ?u- wiihta -'as ?ah?aayiyaXwe?in ?ah?aa -(y)iya-'aX-we-?in

'go in order to eat' eat -go in order to 'they lived there' dwell -at -in the house -PASS -NOW 'he is in the lead on the ground' REF -at the head -on the ground 'at that time' then -at...time -NOW -3S.QT

As we can see from the data above, in (24a), the sequence occurs in the first foot and nothing happens. In the second case (24b), V?V could

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

28 Linguistic Theory and Complex Words

potentially appear between the second and the third syllables, but the /?/ disappears, the vowels coalesce according to regular rules of coalescence, resulting in a variable length vowel. In (24c), the same thing happens between the third and fourth syllables and the result is a short vowel. Another ?-deletion rule occurs when a base ending with a fricative comes in contact with a suffix beginning with a glottal stop /?/. In such situations, there are two possible outcomes that depend on the category of suffix involved. If the suffix is a derivational one then the glottal stop will disappear, as in (25a). If, on the other hand, the suffix is an inflectional one, then the glottal stop will appear following the fricative. The following are some examples of this process. Those in (a) show the unaltered forms containing the glottal stop, while those in (b) show the result of the application of the rule at the level of derivational morphology. 25. a. caahaqsui caah -?aqsui hasaqsui has -?aqsui b. k^isitTaqsui k^ist -it -?aqsut ?ap?aqsui ?ap -?aqsui

'Tough-mouthed (man's name)' tough -at the mouth 'loud voice' loud -at the mouth 'other side of the mouth' different -at the side -at the mouth 'mouth' LOC -at the mouth

The issue of ?-deletion will be of relevance for the arguments in favour of level ordering and will be discussed further in Chapter 6. 2.5.9

Delabialisation

This process is a particularly interesting one given the fact that there are several possible overlapping conditioning factors for its occurrence. We will examine each of these factors in turn, beginning with the simplest. The first category of delabialisation is that of word-final position. In this situation forms will automatically be delabialised, as demonstrated in (26a). The cases in (26b) demonstrate the underlying labial status of the consonants in question when followed by a vowel. 26. a. ii?aa?ak ii?aa -ak*

'missed him' miss -NOM

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Some Phonological Preliminaries 29

hisink hi§ -ink*

'both together' both -together

t'aftanak t'ana -na-k10

'he had a child' child -have

b. qahaFas qah -ak" -'as Xahink^acistuPaX Xah -ink" -a£ist -awiX -'aX

'on the ground unconscious' dead -NOM -on the ground 'it came alongside ...' alongside-with-on sea-MOM-NOW

nuuknaakwitah?aaia 'I have a song to go with it' 0> (0 nuuk -na-k -(m)it-ma-h=?aaia song -have -PAST -1S.IND =always This is a very general rule and agrees with the phonotactic distribution of labialised consonants. The second category of delabialisation occurs when a normally labialised consonant is followed immediately by a consonant, as shown in the examples in (27). 27.

fthakhqa fih -akw -(q)h -qaY

'while crying' cry -NOM -MW -3.SUB

huq^inksap huq" -ink10 -sa*p

'she closed it up' hollowed object -together-MC

muuciiinakh?aXwe?in muu -Sri -na-k" -(q)h -'aX

'it had been four days now' four -days -have -MW -NOW

Another category of delabialisation is that wherein the vowel /u/ occurs immediately after the labialised consonant. In this situation, forms will automatically be delabialised, as demonstrated in (28a). The cases in (28b) demonstrate the underlying labial status of the consonants in question when followed by a different vowel, IM or /a/. 28. a. ricakukft k*i -cak" -uk =?r b. fja^ak^aaPaX fja -^ak" -aa?aX

'his file' file -instrument -POSS =DEF 'for potlatch gifts' give potlatch gift -tool-destined for

This rule can be regarded as a case of dissimilation, whereby the preceding element is assigned the opposite value to that of some feature of

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

30 Linguistic Theory and Complex Words

31

the following element, in this case the feature [RND]. For further discussion of this phenomenon, see Kim (2003b,c). The final category is perhaps the most interesting, since it involves a necessary distinction between derivation and inflection, in other words category or level distinctions. In this instance, labial consonants will be delabialised at the level of inflectional morphology, necessitating a distinction between levels in the grammar. This is illustrated by the following examples. 29. a. ?ayacinkaq ?aya -ink01 -aq

'a great many' many -together -very

b. qahnaakaX qah -naV -'aX

'he was dead now' dead -COMPLETE-NOW

c. hisiikatwe?in his(t) -rkw -'at -we^in

'they passed by' LOC -go along -PASS-3S.QT

In all of these cases the immediately following suffix is inflectional, indicating that this is the level at which the rule is applying and therefore the consonants are delabialised. Thus in all of the environments discussed in the preceding sections labialised consonants will appear without the expected labialisation. It should be added that any consonants previously labialised by the rule of labialisation will obey the same constraints as the inherently labialised forms above. As with a number of the previous phonological phenomena discussed, delabialisation will also be discussed in greater detail in Chapter 6.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Some Phonological Preliminaries

3

In this chapter we will examine the nature of roots and the derivative concepts of stem and base and the various strategies for distinguishing among them. We will then formulate working definitions that will allow us to proceed with our study of Nuuchahnulth morphology. In order to be able to separate affixes from other elements within the word it will be necessary to show that there is obvious motivation for such a move. One clear morpheme structure constraint is the inviolable principle of one root per word, effectively prohibiting word-internal compounding in the language. This fact should be of interest to linguists, given that the standard assumption is that all languages make use of compounding.1 It is interesting to note that it is in a highly polysynthetic language, which allows many other forms of word-building, that we encounter such a constraint. Word structure in Nuuchahnulth is very complicated, or in typological terms 'polysynthetic', employing suffixation to a great extent. There are no prefixes in the language and the only element that may appear ahead of the root is one or more reduplicative copies of some portion of the root. Words consist of roots, defined as the leftmost, non-reduplicative, non-infixal, single morpheme. Roots may not stand on their own, whereas stems, which are really extended roots, are built from roots plus any number of derivational affixes, and may stand on their own. Suffixes combine with roots to form complete 'words' containing some form of aspectual marking, either inherent or overtly marked in some fashion. 3.1

On the Concept of 'Root'

The notion of root is one often used in linguistic theory and description. Following are several quotes from classic writings on linguistic theory and description, discussing the nature of the 'root'.

32

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Word Structure and Categories

33

The broadest and most comprehensive classes of morphemes in English, and the most nearly universal in the languages of the world, are roots and affixes. ... A definition of these two classes which would be universally applicable would be immensely complex .... In general, affixes are subsidiary to roots, while roots are the centers of such constructions as words. (Gleason 1961:58-59) The roots of a language make up its most numerous class of morphological forms and accordingly bear its most varied and specific meanings. (Bloomfield 1933: 242) The root is the irreducible element common to all words of the same family. But any subjective and synchronic analysis separates material elements only by considering the share of meaning that matches each element, and the root is in this respect the element in which the meaning common to all related words reaches the highest degree of abstraction and generality, (de Saussure 1974:186) If we symbolize such a term as sing by the algebraic formula A, we shall have to symbolize such terms as sings and singer by the formula Ab. The element A may be either a complete and independent word (sing) or the fundamental substance, the so-called root or stem or "radical element" (sing-) of a word. (Sapir 192lb:25) As the quotes above attest, the definition of root is a non-trivial task. Root is a useful notion in describing the morphological structure of a language but must be considered for the most part to be a pre-theoretic one. And yet, we rely on this distinction to explain our morphological analyses. Most definitions of 'root' rely on some notion of semantic centrality, of being the 'core' of a word, and it is at this point that a problem arises for the definition with respect to Nuuchahnulth. In languages such as Nuuchahnulth in particular, the root is useful mainly in allowing the statement that there is only one root per word and for stipulating the leftmost boundary for the stem, to which suffixes may be attached. Take for example the following forms in (1).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Word Structure and Categories

1.

a. ?uca£i?aX ?u -ca -siX -'aX

'they went off REF -go to -MOM -NOW

b. histu?ai?aXwe?in his(t) -(y)u?ai -'aX -weY?in

'he saw where ...' LOC -perceive -NOW-3.QT

c. hitacswa Saapuk?i hita -swr -'aX caapuk =?r

'the canoe went through' LOC-go through-NOW canoe

=DEF

Notice that in all these examples, it is the suffix and not the root that determines the meaning of the whole word and might be considered the 'core' or central semantic unit. In fact, roots such as ?u, hist, and hita can be said to make minimal contribution to the semantics of the word, if any.2 There are a number of elements that have a similar status in Nuuchahnulth. They all share certain properties with other roots, such as undergoing reduplication and vowel lengthening and shortening, and preempting the introduction of another root, whether by compounding or incorporation. Boas (1890:126) first noticed the nature of these 'expletive' roots: In Nootka these suffixes may be made independent words by being appended to the stems 6, a certain (definite) [?u], 6csome (indefinite) [?uus], hit- and hitl- [hii], that; ap-, probably side [?ap]. In Kwakiutl the suffixes may be made independent nouns by being affixed to 6 -, ok-, 6g-, hi-, awl-, the separate meanings of which have not become clear to me. Thus roots must be defined structurally—one cannot appeal to semantic notions of 'core' vs. 'peripheral' in order to define roots, since certain elements otherwise definable as roots such as the above, have no clear semantic contribution. The root is more of a structural unit serving to 'anchor' the word and while it may contain a core semantic value, this is not a necessary requirement and the root may simply fulfil a purely structural function, defined in terms of position (leftmost non-reduplicative unit) and constrained for co-occurrence (only one unit per word). This unit acts as the target for reduplication, constrains the application of incorporation, and acts as the site for various morphophonological processes, e.g. vowel lengthening or shortening, reduplication, etc. Furthermore, there are constraints placed on the structure of roots, which do not apply to affixes, as we shall see in the following section.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

34 Linguistic Theory and Complex Words

Word Structure and Categories

35

3.2 The Structure of Morphemes

3.2.1 Morpheme Structure Constraints The use of morpheme structure constraints (MSCs) may be somewhat pretheoretic, but it provides us with some insight into the possible shapes of morphemes in a language. In the case of Nuuchahnulth, MSCs exhibit striking differences with respect to the distinction between roots and suffixes in Nuuchahnulth. In addition to the relative length of morphemes, wherein suffixes may often exceed the length of roots, the possible sequences of consonants and vowels also differ between them, with roots typically being more restricted in terms of their structure. Naturally, composite forms consisting of root + lexical suffix may extend the base to a larger number of syllables, but this is not the unit under consideration here.The basic shape of Nuuchahnulth roots appears in (2). 2.

Root Structure: CV(V)(C)(C)(C) Examples of the various possible permutations are provided in (3). Root hu nuu hus kuuq kums nuutk hastk muuqXh hustq winsh cimsk

Category v v N V N V

ADJ N V

PLN V

Gloss 'bail' 'sing' 'saltwater' 'stalk' 'herring roe' 'circle' 'loud' 'Indian Celery' 'splash' 'Gibson Cove' 'make a whistling sound'

Rose (1981) observes that a large majority of roots are monosyllabic, but that there are a certain number of bi- and polysyllabic roots (see Table 3.1). The absence of prefixes in Nuuchahnulth allows us to identify the root straightforwardly in the majority of cases.3 We can define the root as

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

In this section we will investigate the structure of morphemes, the nature of morpheme structure constraints on them, and the important issue of 'bound' versus 'free' morphemes in Nuuchahnulth, which cuts across traditional distinctions of root versus affix.

the leftmost structural unit not involving copying of adjacent phonemes, i.e., removing all reduplicative copies. Roots are rarely more than two syllables. When it comes to suffixes, we can see that they may, in fact, be more diverse in their shape than roots, as suggested in (4). Note that there is one constraint on suffixes which does not apply to roots: /h/ never appears in the onset of a suffix. 4.

Suffix Structure:

-(C)(C)(C)V(V)(C)(C)

Examples of the various instantiations of this structure include the following: 5.

-a -ee -ac -iic -ca -caqs -hwink 0 -q?i£h -cswini

INFL INFL N ADJ V LOC V ADV

ABS

LOC

'at the armpit [RL]'

voc 'receptacle' 'belong to' 4 go to' 'on...side' 'use [L]' 'for... years'

Suffixes may be monosyllabic, although there are a large number of polysyllabic suffixes, reaching up to four or more syllables in some cases. 6.

-axyin

LOC

-'aisimhi -a-nufihta

V LOC

'at the bow' 'want [L]' 'along the nose [L]'

Suffixes may be complex, lexicalised combinations of other suffixes as, for instance, in the case of -mars 'moving about in the village', derived from the merger of the two suffixes -mai and - 'as, with concomitant application of morpho-phonological rules, thus: -mai and -'as => -mayas =» -maYs. Rose (1981:29f) discusses MSCs in Kyuquot and provides estimates of the distribution of mono- and multi-syllabic roots among the various grammatical categories: Neither CVC0-3 roots nor multi-syllabic roots are distributed evenly among Kyuquot semantic classes. These semantic

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

36 Linguistic Theory and Complex Words

Word Structure and Categories

37

The table below, adapted from Rose (1981:29), shows the percentage of mono- and multi-syllabic roots found within each of the grammatical classes. Rose observes that the so-called multi-syllabic roots are virtually always disyllabic, with only 16 tri-syllabic roots attested. V

O

A

N

15% 50% 75%

85% jgggjggZ

if wmmmm

IP

Mono-syllabic Roots

Multi-syllabic Roots

Table 3.1 Distribution of Roots (based on Rose 1981:29) Note that verbal roots are predominately mono-syllabic (85%), whereas nominal roots are predominately multi-syllabic (85%), suggesting, perhaps, a historical relationship whereby nouns are derived from verbs. There are many such suggestive relationships, e.g. ha?um 'food' < ha^/a 'eat' + -im 'thing'; compare hatiackwi?i 'the food remnants' < haw a 'eat' -ckwi' 'remains of ?i r DEF. Rose goes on to state (ibid:30) that "of the mono-syllabic roots, the majority (two-thirds) are CVC, a quarter are CVCC, and the remainder are virtually all CV. Only four roots have been found to end in a triple consonant sequence." Thus, root structure constitutes a virtual subset of suffix structure. 3.2.2 Bound vs. Free Forms An important distinction encountered in Nuuchahnulth stems is the morphological property of root allomorphy. In Nuuchahnulth, there is a high degree of root allomorphy that distinguishes between derivational and inflectional levels of affixation.4 One form is chosen when the suffix is derivational or aspectual and another one when the suffix is inflectional.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

classes, which will be discussed and referred to many times in this thesis, are: verbal (V), nominal (N), adjectival (A), locative, quantifier, quantity, and temporal. The latter four are here amalgamated into one class, O ('other'), because together they constitute only three percent of all roots.

In (7) below, the first column provides the free form that occurs with inflectional suffixes and on its own. The second column indicates the basic word class of the root and the third column provides a rough gloss. The final column provides the form of the root that is found in combination with derivational and aspectual suffixes. Examples in (7a-c) involve bound roots ending in /tq/, (7d) are of borrowed words, mainly from English, (7e) involves bound forms ending in /q/, while (7f) shows forms with a shortening of the root. (7g) shows a substantially altered form, ending in Id. Free Form

7.

a.

b.

ce?inwa fia^uunis hax^inmac kiianuus makit q^aahin tiikTin

Cl. N N N N V V

v

c.

pu*iai

d.

Sikinis misin tiipin ?eepinis

N

?iitu yaasmai yayaSim

N

?athii ci£i£i ^aaq muwa£ tefif

N

e.

f.

g-

yimnqsu

ADJ

N N N

V N

N

ADJ N

ADJ

Meaning

Bound Form

'Sheep's-Feet' 'edible berry species' 'wren' 'furseal'

ce?itq Sa^utq hax^itq kilatq

'play with shells, dolls' 'fish with prong-spear' 'bake in sand, ashes'

makitq q^aahitq tikwatq

'sleepy'

pufrtq

'chicken' 'boarding school' 'table' 'apple'

5ikitq misitq tipatq ?epitq

Titu bird' 'hunt fur seal' 'Yayachim whale'

?ituq yasmaq yaacuupq

'night' 'tooth' 'long' 'deer' 'sick'

?ath CiC yaq muk ta

'male's brother-in-law'

yimac

The point here is that the choice of bound vs. free form is not a property of any particular word class or even a historical remnant. It is an active, productive process applied even to borrowed words. Furthermore, the

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

38 Linguistic Theory and Complex Words

Word Structure and Categories 39

8.

a. capitqifiakyak caapin -rnakD -yak caapin caapin

'(I) have a sawbill dance' sawbill duck -imitate in dance-tool 'sawbill duck' sawbill duck

b. Xuk^atquwis Xuuk^aana -uF - 'is Xuuk^aanaX Xuukwaana -'aX c. Xaatmaqanui?i Xatmapt -a*nui ==?r Xatmaptuk?i Xatmapt -uk =?r

'Wolf Ritual Beach' Wolf Ritual -place -on the beach 'I had given a Wolf Ritual' Wolf Ritual -NOW 'the one of yew' yew -along [L] =DEF 'his yew tree' yew -POSS =DEF

Bound roots may occur in a variety of shapes, including examples such as the following, where the first instance in each pair is the free form and the second is the bound form. Note that in some cases the only difference between the forms is vowel quantity. 9.

cuiin / <5iiitq

'potlatch money'

a. ciiitqsi?at ?ah?aa diiiin -§iX -'at ?ah?aa

'they gave potlatch goods to him' potlatch money-MOM-PASS that way

b. ciiiin (5iiiin

"Tsiithlin" is what they say' potlatch money say -NOW -PASS

wawaa?aX'at wawaa -'aX -'at

10. kuxmin / kuxmatq

'rattle'

a. kuxmatqsi?aX kuxmin -siX -'aX

'they began shaking their rattles' shake rattle -MOM -NOW

b kuxmatqTaX kuxmin -TaX

'it would make the sound of a rattle' wooden rattle -make a sound

c. hinii?aX'at kuxmin hina -ayr-'aX-'at kuxmin

'he was given a rattle' LOC -give -NOW -PASS rattle

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

exact shape of the bound form is, to a large extent, unpredictable and therefore lexically marked. The bound allomorph of the stem exists only at the level of derivational/aspectual morphology and never at the later level of inflection, suggesting that here again a distinction exists between the two domains.

40 Linguistic Theory and Complex Words

11. Ci5i5i/Cie a. £ic7ataXcuu £i£i£i -?atu -'aX -cuu

'the tooth would twist out nicely' tooth -fall off -NOW -preferably

b £i£i£irhinh £i£i£i -rhinh

'teeth' tooth -PL

12. q"iiXin / qfiiXitq

'bone chisel'

a. q^iiq^iiXitqsiPaX DUP- q^iiXin -siX -'aX

'they started chiseling' REP- bone chisel -MOM -NOW

b. qii£iX q^iiq^iiXitqa qii -siX DUP- q^iiXin -(y)a*

'for a long time, he was chiseling' long time-MOM REP- chisel-DUR

c ?uuhwai q^iiXin ?u -[L] -hwai q^iiXin

'he used a bone chisel' REF -use [L] bone chisel 'scapegoat'

13. clfcrhuu / £i£rhuq a. Si£rhuu?aaqXap SiCrfur -?aaqX -'ap

'he was forced to make sacrifices' scapegoat -INTENT -CAUS 'the scapegoat' scapegoat -DEF

6i£rhuu?i SiSifiu* -iii b. 6i£rhuqck
'what is sacrificed' scapegoat -remains

SiCrhuqiii £i£rfur -iii

'make a scapegoat' scapegoat -make

14. hamuut/hamut

'bone' 'the bones' bone =DEF

a. hamuut?i hamuut =?r Xuunim hamuutuk hamuut -uk Xuunim b. hamutck^i hamuut -ck^r

?iihtuup ?iihtuup

hamuthtin £it'uui?i hamuut -htin Sifuul =?f

'an elk's bone' bone -POSS elk 'remains of whale bones' bone -remains whale 'clubs were made of bone' bone -made of warclub =DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

'tooth'

Word Structure and Categories 41

a. maht'iimit?i mahfii -(m)it =?i* maht'iirhinh?i maht'ii -rhinh =?i*

'house' 'the former house' house -PAST =DEF 'the houses' house -PL =DEF

b. maht'iqapui-?is?i maht'ii -apui -?is =?r

'the little pretended house' house -imitate -DIM =DEF

mahfiqckwak maht'ii -ckT -'ak

'the remains of his house' house -remains -POSS

As can be seen from the examples above, where the (a) forms involve free roots combining with inflectional suffixes and the (b) forms involve bound roots combining with derivational and aspectual suffixes, there are a variety of forms that bound roots may take, diverging from the free forms to a greater or lesser extent, but the suffixes found to occur with bound forms are consistent throughout the paradigm, suggesting some close affiliation. Such suffixes include derivational and aspectual suffixes, but no members of the previously presented inflectional category. It may be useful to examine the behaviour of one or two suffixes across a number of different stems, to confirm the use of bound forms across the board. In (16) we encounter the suffix -rnakmmeaning 'to perform a kind of dance determined by the stem to which it is attached'. 16. a. capitqinak caapin -rnakf0

'a sawbill dance' sawbill duck -imitate...in dance

b. Xapisaqinak Xapisim -f Aak01

'a raccoon dance' raccoon -imitate...in dance

c. hiyiqinak hiiyi -i'nak10

'a snake dance' snake -imitate...in dance

d. Tamiqinak S'amf -r nak*0

'a horse clam dance' horse clam -imitate...in dance

e. dix^atinqiriak cixwat -rnak"

'an eagle dance' eagle -imitate...in dance

f. qu?isitqinak qu?isin -rnak10

'a raven dance' Raven -imitating...in dance

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

15. marjfii/ mafrfiq

Many additional examples of this form could be provided from the data, all employing the bound form of the stem in combination with this suffix. Similarly, the suffix -Ruk, meaning 'resemble', whatever it is attached to, is very productive and exhibits an additional interesting characteristic, to be discussed in Chapter 4 on affixation, the property of requiring reduplication of the stem to which it attaches, as shown in (17). 17. a. cadaakuk DUP- <5a?ak -kuk

'Creek-like (pin)' SUF- creek -resemble [R]

b. kikin&atkuk DUP- kinwa*t -kuk

'Mat-like (m.n.)' Suf- old mat -resemble [R]

c. XiXickuk DUP- Xic -kuk

'flour (= looks white)' SUF- white -resemble [R]

d. niftixkuk DUP- nixtin -kuk

'peas (= looks like salmon roe)' SUF- salmon roe -resemble [R]

e. papatfixkuk DUP- patSrwa -kuk

'roseberry-like' SUF- wild roseberry-resemble [R]

f. pipickuk DUP- picup -kuk

'Ninebark (=resembles cedar bark)' SUF- cedar bark -resemble [R]

g. ququ?isitqkuk DUP- qu?isin -kuk

'Oregon grapes (= resembles a raven)' SUF- raven -resemble [R]

Again there are a number of further examples of this suffix, but the point should be made by now: in all cases the base to which this suffix attaches is a bound one, exhibiting distinct properties from the free base which appears as a free form or in combination with inflectional morphology. Finally, example (18) below shows the shape of some of the same bases shown above, with an inflectional suffix, -?is DIM, in order to demonstrate the difference in stem allomorphy. 18. a. caapin?is caapin -?is b. ca?ak?is?i da?ak -?is =?i*

'little sawbill duck' sawbill duck -DIM 'a little stream' stream -DIM =DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

42 Linguistic Theory and Complex Words

Word Structure and Categories 43

Akin to the bound roots of the previous section is a distinction among suffixes between basic forms and so-called combining forms. The latter are a set of suffix allomorphs that occur in a special shape only in combination with derivational and aspectual suffixes. Since 'bound suffix' would be somewhat incongruous, the term 'combining form', first introduced by Sapir & Swadesh (1939), will be employed to refer to this class of affixes. Examples of these are provided in (19-22). Note that in the (a) examples, the suffix is in its base form, either finally or in combination with inflectional morphology, whereas in the (b) forms the suffix is followed by other derivational or aspectual morphemes. 19. -s^aqsti /-s^aqstiqa. ?uksyaqsti ?u -syaqsti yaqsyaqsti?itq?ar yaq10 -syaqsti -?rtq =?ai b. ?uksyaqstiqistasi ?u -syaqsti -ista -si* yaqsyaqstiqiiyip?itq?ai yaq" -syaqsti-iiyip-?f tq=?ai 20. -api / -apiq-

'main or leading' 'the leaders were ..' REF -leading 'the one who was the head' REL -leading -3s.REL =PL T was chief of the crew' REF -lead-persons in canoe-lsg.ABS 'the most valuable thing they had' REL -leading -obtain -3s.REL=PL 'up in the air'

a. kiicapi?i kicuk -[L] -api =?r

'a fallen tree that was slanting upward' log -in air =DEF

Xiikapirhinh Xik -[L] -api -rhinh

'they were placed stretching upward' have hands in position-in air [L] -PL

b. diisapiqsi?aX
'send a telegram' rope, line -in the air[L] -MOM -NOW

suu£aqimimaiapiqiiyip 'they got five (birds) in the air' suSa -[L]-qimi-mai-api-iiyip five -CLS -move -in air[L] -obtain 21. -'isaiap / -'isanaqa. £iisanaqyak?i £ii -'isafiap -yak40 =?r

'on the beach' 'the song for pulling up on beach' pull -on the beach [MC] -song =DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

3.2.3 Suffix Combining Forms

44 Linguistic Theory and Complex Words

b. suk^iX wihiisanap suu -siX wihi -'isanap

'they brought him to the shore' hold-MOM on dry land-on beach[MC]

22. -yu* / -yaqa. cusyaqsiX ?anah?at?itq dus -yuY-siX Panarf-'at-Prtq

'she dug the size hers was' dig -...-ed-MOM size-PASS -3S.REL

b. cusyuu?i dus -yu* =?i*

'the hole in the ground' dig -...-ed =DEF earth

cakumc cakumc

c. dusyuu?akma dus -yuY-'ak-ma*

'he has a tunnel' dig -...-ed -POSS -3.IND

The (a) examples above show what happens when the suffix completes the stem, whether or not further affixation at later levels occurs, while the (b) examples demonstrate the results within the derivational component of the grammar, where a 'combining' form is necessary for further derivational suffixation. Again, the suffixes that concatenate with these combining forms are the same ones that occur with the bound roots discussed in section 2.3. As such, it seems obvious that there is a clear distinction made in the grammar between combining forms of certain roots and suffixes and a stem-final form, which combines with aspect markers and inflectional morphemes. The example below demonstrates a case of combining forms for both the root and first suffix. Note that the word is terminated by a basic form of the final suffix, -c"ik 'go along'.5 23.

hiinaqapiq£ik hiina -[L] -api -£ik

'quartz going along in the air' quartz -in the air [L] -go along

3.2.4 A Brief Excursus into English The facts of Nuuchahnulth stem formation may seem odd at first blush, but they should not be particularly surprising to speakers of English, which exhibits its own bound/free distinctions, such as the following. 24.

Free

Bound

Example

receive permit resign leaf knife

receptpermissresignleavkniv-

receptive permissive resignation leaves knives

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

'-ed; COMPLETE'

Word Structure and Categories

-i[s]-a[s]-

specificity hyphenation

As can be seen from such cases, the existence of a bound/free or combining form should not be considered unusual or even uncommon. The reader should note that, as in Nuuchahnulth, such forms in English are a property of an early level of the grammar and bound forms are not found in later derivations, e.g. receiver, permitted, resigning, etc. 3.3

Stems

Given this clear-cut distinction between bound and free forms which exists at the level of derivational and aspectual morphology, we may wish to posit a special category to describe constructions involving such forms. This is the notion of the stem. The term 'stem' will be employed here to indicate that unit of the morphology which operates in the derivational module and which serves as input to aspect marking. Stems will be considered complete when they are assigned aspect, after which they may enter the inflectional component of the grammar, where they may be marked for the appropriate inflectional categories. Stems may consist of a simple root + aspect (25a), a root combined with a derivational suffix (25b), a reduplicated root combined with one or more derivational suffixes (25c), or various combinations of root + derivational suffixes (25d). 25. a. ciqsiX ciq -siX

'say s.t.' speak -MOM

b. ciqyak ciq -yakw

'a word' speak -tool

c. ciciqink DUP- ciq -ink™

'they speak (to each other)' SUF- speak -in conversation [R]

d. cumaanupmafiiqX cumaa -inup -mafiiqX

'he wanted to fill it up' full -GRD.CAUS-want to [M]

As one can see from this data, stems can be fairly short or quite long, the common property being that they are incomplete until assigned an overt aspect, at which point they may be considered formally complete words. Note from (25d) that a stem may contain more than one aspect marker, in which case it is the outermost aspect marker that has scope

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

-ic -ate

45

46 Linguistic Theory and Complex Words

26. a. ciqsi?aXsi ciq -siX -'aX -si

'I said it' speak -MOM -NOW -l.ABS

b. cumaanupikqas cumaa -inup -(y)ik -qars

'may I fill it' full -GRD.CAUS -FUT -ls.SUB

3.4 Bound Verbs In this section arguments will be put forward for a distinction within the Nuuchahnulth lexicon that is not the commonly accepted view: that there exists a difference between verbal suffixes and bound verbs. Note that this is a common morphological distinction in English and other languages, as demonstrated in (27). 27.

a.

Root hyphen Bowdler solid Bound Verb -ceive -mit -fer

Verbal Suffix -ate -ise -ify

Resulting Verb hyphenate Bowdlerise solidify

Resulting Verb receive, conceive, deceive remit, commit refer, confer, defer

As should be clear from (27), both contexts involve some form of bound verbal morpheme, but the status of the morpheme is different. In the (a) cases, the verbal suffix is attached to a nominal base, creating a derived verb, while in the (b) cases, the bound verb serves as the nucleus of the construction, to be modified by prefixes and suffixes. Thus, a distinction must be drawn between these two classes of bound verbal morpheme in English. This issue, as it pertains to Nuuchahnulth, will be examined in some detail, since it bears on later discussions in Section 8.2 concerning incorporation.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

over the entire word. At the point where a fully formed stem is combined with aspect marking, it is somewhat traditional to refer to the output as a base, to which inflectional morphology may be attached. These may then serve as the input to inflection, as demonstrated below.

Word Structure and Categories

47

There are approximately 1000 verbal roots and stems found in the Nuuchahnulth lexicon. This is, of course, a rough estimate, given the open nature of the lexicon and the productive compositional properties of the language. The number of bound verbal morphemes is far less, but still substantial and we may recognise two types: those which are involved in modifying an existing head by further verbal specification, and those that assume the role of head of a verbal construct, capable of incorporating one of its arguments into the overall predicate. We will refer to the former as verbal suffixes, and to the latter as bound verbs. The argument to be presented in this section is that we need to draw a distinction between the two sorts of morphologically-bound elements, for several reasons. Distinctions similar to the ones made here have been discussed for Nuuchahnulth at least since Swadesh (1933) who uses the terms governing and restrictive to distinguish between verbs which act as heads and verbs which act as modifiers. However, this distinction is insufficient to treat all cases of Nuuchahnulth bound verbals, since it fails to address issues of valence and of semantic constraints on combination, which is crucial for the process of incorporation: not all transitive verbal morphemes may incorporate. In fact, it is only a small proportion that do this in any active manner. Some of these elements will be presented here and examples of their usage will be provided, although a more detailed presentation of the nature of incorporation will be postponed until Chapter 8 on Morphosyntax. The basic differences between bound verbs and verbal suffixes may be outlined in the following. We will discuss each one of these issues in the following sections. 28. a.

Bound Verb Acts as head of construction

b.

Attaches to ?u

c.

Incorporates themes and element: of the theme argument May have inherent aspectual distinctions

d.

Verbal Suffixes Non-head element: modifies the meaning of the base Attaches only to lexical stems, never directly to ?u No incorporation is triggered by these Aspect is attached as a suffix

We shall now proceed to an examination of these various properties.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

3.4.1 Bound Verbs vs. Verbal Suffixes

48 Linguistic Theory and Complex Words 3.4.2 Bound Verbal Properties

29. a. ?unaak ?u -na^k" b. fananak fana -na*kw

'have s.t. (transitive)' REF -have 'have a child (intransitive)' child -have

As we can see, the semantics of the forms resulting from the combination of -na^ 'have' with the bases in (29) clearly indicates that it is -na'k" that is the head. Bound verbs may serve as the incorporator of arguments from the sentence, but verbal suffixes never do. An example may be useful to clarify this distinction. 30. a. -ma*s b. -?ii

'move about in the village' 'go to'

The two morphemes in (30) both involve movement and both are bound, but there is a difference between them with respect to the way that they interact with other elements of the word and beyond. Take the following examples: 31. a. wawaaimas?aX DUP- waa -maYs -'aX

'they said, moving about in village' REP- say-move about in village-NOW

b. wihii?u?aX ?iihtuup?i 'the whale reached dry land' wihi-?ii-'awiX-'aX ?iihtuup=?r dry land -go-on rocks[MC]-NOW whale=DEF As can be seen from these examples, the effect of the suffix -ma*s 'move about in the village' in (31a), is to modify the meaning of the base wawaai 'say', while the effect of the bound verb -?ii 'go' in (31b), is to act as the head of the collocation 'go to dry land'. This is the basic difference between the two types, but there is a further distinction remaining to be made. 3.4.3 The Nature and Status of ?u The difference is with respect to the possibility of attachment of the bound verbal morpheme to ?u, which is an expletive base used solely for

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

In keeping with the generally accepted view of heads as determiners of the overall category of the word and as the semantic 'core' of the word, we shall consider bound verbs to constitute heads. Thus, upon incorporating an object, they remain verbs, albeit saturated verbs.

49

the purpose of providing a site of attachment for bound elements. As such, it is semantically void of all meaning, unlike the locative bases, which may contribute various senses of position versus direction. Furthermore, it can never act as a head, either syntactically or semantically. This kind of an expletive base is quite uncommon in languages in general and, in Nuuchahnulth, it can never appear on its own, unlike most roots. It must obligatorily be made predicative, minimally by the use of the element -(q)h MW, as demonstrated below.6 32. a. cuusuk^iX ?uyi ?uhquu qaahk^aTap. dusuk -siX ?uyi ?u -(q)h -quu qah -k^a -'ap suspect-MOM whether REF-MW-CND dead-complete-CAUS 'she began to suspect what it might be that caused the dying' b. ?uh?ishak Taatusrhit? ?u -(q)h-?is-ha-k Saatus-rfift

'Is that little you, young Deer?' REF -MW-DIM-2S.INT deer-Son

There are other, similar bases in Nuuchahnulth, for instance the locative base, hita, which cannot stand on its own but must be suffixed. However, there is no requirement that it be made predicative or that it be attached by a verbal morpheme. It may stand as head itself, modified by a locative suffix, as in the following example. 33. a. hitaas?aXma yaa?ai?aX hita-'as-'aX-maY yaa?ai-'aX b. ?ustaascipXaa ?ust -a*sda -ip =Xaa

'they were now outside looking on' LOC-outside-NOW-3s.iND look on-NOW 'he put that one on the roof too' LOC -on roof -MC =again

In both of these examples, locative bases are employed without any verbal morpheme, and yet they may serve as the main predicate of their respective sentence. This is the more typical situation in Nuuchahnulth, whereas ?u, which may never act in this way, is the more unusual. Nevertheless, this type of base is not unattested in the languages of the world. West Greenlandic, for example, appears to also have such a base, pi in this case. Thus: With qar ['have'] the possibility of a corresponding nonincorporating construction exists using 'empty' stem p i plus that affix, the 'object' appearing in the instrumental case externally. (Fortescue 1984:83)

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Word Structure and Categories

50 Linguistic Theory and Complex Words Compare the following examples from West Greenlandic (Fortescue 1984:83) with those in (29) from Nuuchahnulth, repeated below.

b. ikinngutinik piqarpuq 29. a. ?unaak b. fananak

'he has a friend/friends' 'he has some friends' 'have s.t. (transitive)' 'have a child (intransitive)'

It is important to note that this language also shares other relevant properties with Nuuchahnulth, including the prevalence of incorporation and the rather large number of derivational affixes.7 However there are very few cases of this in other languages, even in those closely related to Nuuchahnulth, e.g., the Northern Wakashan language, Haisla, 8 which does not have such a base and therefore has obligatory incorporation in such situations. Which brings us to the motivation for such a base. ?u may act as a placeholder for bound verbs which for various reasons, do not or cannot incorporate in the particular context. One reason for this may be that the incorporee is a coordinate NP or a clausal element, which obviously could not be incorporated wholesale. Alternatively, there may be discourse-related reasons for not incorporating. This issue will be discussed at greater length in Chapter 8. 3.4.4

Syncretic Alternates of Bound Verbs

Many of the bound verbs discussed in this section have syncretic free forms which do not allow incorporation and which stand without the use of ?u? These forms occur in the same contexts as the ?u + bound verb forms, as in the following examples: 35. a. na?u* -pryaqh

'accompany' 'accompany'

b. Xawas -(n)as?ii

'approach' 'approach'

c. wiina -cTyuq^a

'attack' 'attack'

d. maakuk -'aap -haY

'buy' 'buy' 'buy'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

34. a. ikinngutiqarpuq

e. siqa -ca*qa

'cook' 'cook [L]'

f. qah -sawiX

'die' 'die'

g- nus -'ini

'distribute' 'distribute [L]'

h. puuwica -'itui

'dream' 'dream [R]'

i. hawa ha?uk -'rs

'eat' 'eat' 'eat'

j - na?a* -'aap

'hear' 'hear'

k. nuu -iyaqh

'sing' 'sing [R]'

51

As one can see from these examples, and there are numerous others, the bound and free forms share no physical resemblance, although semantically they perform virtually the same function. In most cases the bound verbs are transitive, although not (35f), for example. It should also be noted that, while the bound forms may always be employed as free forms via the attachment of ?u, the free forms cannot be employed as bound forms, i.e., in incorporation, for which see Chapter 8. In many cases, there are exact parallels between the usage with free syncretic form (36a), bound verb with ?u (36b), and bound verb with incorporated object (36c). 36. a. nussi?aX ?ah?aa sudiics Xusmit 'he then distributed five herrings' nus-siX-'aX ?ah?aa sud-iics Xusmit distribute-MOM-NOW then five -take along herring b. ?uu?ini?aX ... hiix^aa ?u -[L] -'ini -'aX hiixV

'he distributed... dentalia' REF -distribute [L] -NOW dentalium

c. hiix^afrn-ffaX hiix'V'ini -'aX

'they distributed dentalia' dentalia -distribute [L] -NOW

A final point about this distribution is that there are both free forms with no bound parallel, and vice versa. For instance, there is no bound

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Word Structure and Categories

52 Linguistic Theory and Complex Words

3.5 Numbers The Nuuchahnulth number system is fairly complex and worthy of note for the role it plays in incorporation (see Section 8.2) and other processes in the language. We will briefly examine the number system in the following. Numbers from 'one' to 'five' are simplex (37a), while those from 'six' to 'nine' are composed of two elements, the roots for either 'one' or 'two' together with the suffixes -pu 'more than' and -kwai 'less than' (37b), giving us:10 37. a.

cawaa ?aXa qacca muu suSa

'one' 'two' 'three' 'four' 'five'

b.

nupu ?aXpu TaXak^ai cawaakwai

'six' 'seven 'eight' 'nine'

'Ten' and 'twenty' are morphologically simplex, whereas the multiples of 'ten' fall into two categories, either single words or syntactic constructs, without overlap. They are formed by the use of the suffix -iiq 'score' in combination with the syntactic construction ?is hayu 'and ten'. 38.

hayu caqiic caqiic ?is hayu ?aXiiq ?aXiiq ?is hayu

'ten' 'twenty' 'thirty' 'forty' 'fifty'

qacciiq qaCciiq ?is hayu muyiiq muyiiq ?is hayu su&iq

'sixty' 'seventy' 'eighty' 'ninety' 'hundred'

Units between the tens are formed by the use of ?is 'and' or the related ?uh?is 'and' with the appropriate unit and decimal indicators, thus: 39.

hayu ?is dawaak hayu ?is ?aX caqiic ?is qa£ca ?aXiiq ?is dawaak

'eleven' 'twelve' 'twenty-three' 'forty-one'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

form for cax^'spear', naq 'drink' or t'iq" 'sit', nor is there a syncretic free parallel for -A a ^ ' h a v e ' , -Hi 'make', or -simd 'train for'. Thus, there is an asymmetry in the verb system that is balanced if bound verbs are regular members of the verbal system with only a phonological dependency on some base such as ?u.

Multiples of one hundred are formed in one of two ways: (a) by continuing the use of -iiq 'score' in conjunction with the conjunctive ?is, or (b) by the use of the appropriate unit combined with a suffix -]5it meaning 'times' followed by the form suiiiq 'hundred'. Given the consistency of the former set with the lower numbers we will assume that it is the original one and that the alternative system is a caique based on the English number system, i.e. 'two hundred, three hundred, ...'. n 40. a. hayuuq 'two hundred' hayuuq ?is suSiiq 'three hundred' ?aXpituuq hayuuq 'four hundred' qacdufjituuq hayuuq kiix?in?ath 'there were 600 Kiihin men' b. PaXpit suSiiq qacdu^it su&iq muupit suSiiq nupufjit suciiq PaXak^aipit suciiq

'two hundred' 'three hundred' 'four hundred' 'six hundred' 'eight hundred'

Multiples of 'one thousand' are formed with the appropriate unit designator followed by a suffix -/3if 'times' followed by the borrowed word taawisin 'thousand', again clearly calqued from English. 41.

nujMt taawisin 'one thousand' PaXpit taawisin 'two thousand' PaXfMt taawisin Pis muupit suciiq 'two thousand four hundred'

There is no overlap between root and suffix within the Nuuchahnulth number system, unlike in other areas of the grammar. Among the numbers, all are either roots or roots combined with derivational suffixes which perform other tasks elsewhere, with the exception of -iiq 'score' which only occurs with the roots from 'two' to 'ten'. 3.6

Names

Proper names are typically morphologically transparent, although there are exceptions to this, involving lexicalised forms. In terms of semantics, names are often far from transparent. Nevertheless, they appear to be subject to the same constraints as any other word in the language and, in fact, exhibit additional morphological properties relating to the formation

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Word Structure and Categories 53

54 Linguistic Theory and Complex Words

42. a. tux^siX tux^-siX

'He-jumps' jump-MOM

b. PanimiyisPaqs Pani -mai -'is -Paqs

'Woman-going-alone-on-the-beach' alone -move -on the beach -woman

c. wastqaPa wa -tqa -'a^Pa

'It-is-coiled-under-on-the-rocks' coil -underneath -on the rocks

d. t'atatwafrih DUP- faatwin -'iih

'Speckled-trout-hunter' SUF- speckled trout -hunt

As can be seen from the examples above, names are usually analysable by speakers and the meanings are typically transparent. There are a number of constraints on the operations that may be performed on names, for instance, they cannot be incorporated usually, although there are exceptions which will be discussed in Chapter 8. For discussion of the treatment of names in general, see Section 7.4 on Hypocoristic Formation. 3.7 Word Categories The subject of word categories in Native American languages has been debated at least since Boas (1911), who discusses the Northern Wakashan language, Kwak'wala in the following: The classification of suffixes here given shows that a division of words into verbs and nouns has taken place, both being fairly clearly distinguished by suffixes. We find, however, that syntactically the distinction is not carried through rigidly; nouns being treated with great ease as verbs, and verbs as nouns. It must be added here that the forms of the pronouns as attached to the noun and as attached to the verb are distinct. Since the psychological relation of sentences is included in the process of suffix formation, conjunctions are absent. For this reason, and on account of the verbal character

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

of hypocoristics, much as English has contracted forms such as Sam from Samuel, Marg from Margaret, or Fran from Frances. Names in Nuuchahnulth are morphologically complex, containing a single root as in all words in the language, and from one to theoretically any number of suffixes, including aspectual markers, as in (42a) below.

Word Structure and Categories

55

Wakashan languages have always presented somewhat of a challenge to linguists when it comes to deciphering grammatical relations and the status of categories. Claims that word class distinctions do not exist in the languages are reflected in statements such as that in Sapir & Swadesh (1939: 235-236), which emphasises the lack of distinctive word classes. ... the major word classes are normal word and particles. Normal words are made up of a stem with derivative suffixes, ranging from none to four or five, and are always defined as to aspect....This statement applies to words corresponding to English verbs, adjectives, nouns, prepositions, and adverbs. Sapir (1921:119) provides further embellishment on the topic in the following: This distinction is of such fundamental importance that the vast majority of languages have emphasized it by creating some sort of formal barrier between the two terms of the proposition. The subject of discourse is a noun. As the most common subject of discourse is either a person or a thing, the noun clusters about concrete concepts of that order. As the thing predicated of a subject is generally an activity in the widest sense of the word, a passage from one moment of existence to another, the form which has been set aside for the business of predicating, in other words, the verb, clusters about concepts of activity. No language wholly fails to distinguish noun and verb, though in particular cases the nature of the distinction may be an elusive one. It is different with the other parts of speech. Not one of them is imperatively required for the life of language. Jacobsen (1979b) takes a more circumspect approach to the question, providing concrete arguments for distinguishing between nouns and verbs, but leaves the question of the status of adjectives somewhat more open, describing them as a subclass of intransitive verbs. We will now examine Jacobsen's arguments in somewhat more detail.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

of most adverbs, there remain only few classes of words — nouns, verbs, and particles. (Boas 1911:443)

56 Linguistic Theory and Complex Words

In this section the arguments for differentiating categories both within and outside the syntax will be examined, with the aim of understanding the reasons for the opacity of word classes and the problems that this has presented for linguists. Jacobsen (1979b) provides a number of arguments for distinguishing word classes in Southern Wakashan. His principal focus is the most southerly of the languages, Makah, but the arguments are directly applicable to the other members of the group. In what follows we will briefly review his arguments, providing parallel examples from Nuuchahnulth where appropriate. Jacobsen states that 'verbs are limited to predicative function, except when nominalised with -°iq or in the nominal paradigm. Many formations are unique to them, such as the iterative and repetitive aspects and the future tense.' (ibid: 139) He furthermore states that only verbs and prepositions may be transitive and that there are a few ditransitive verbs. Intransitive verbs may not take objective nominal arguments or objective pronominal endings. They may be formed from nouns by means of a momentaneous aspect marker. Thus it will be seen that my argument basically accepts the premise that major words may all occur as predicates (while still noting gaps in the inflectional possibilities available to some classes) and bases most of its discriminations on differences of occurrence as arguments, (ibid: 107) As for adjectives, they are considered to be a special kind of intransitive verb that may occur as a modifier of a following noun, and occasionally as a nominal argument. On nouns, Jacobsen (ibid: 140) states: Nominal arguments are comparable to what would often be called noun phrases. They act as subjects and objects of verbs, and objects of prepositions. Among potentially predicative words, only nouns may directly act as nominal arguments. They may also be words bearing the article -°iq, with or without a following demonstrative ending; this article may be added to verbs (including adjectives) but not to prepositions, to nouns by virtue of their predicative potentiality, and to the first word of phrases, which may be a modifying

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

3.7.1 Jacobsen on Noun and Verb

57

adjective or adverb, as well as a verb. ... Nouns may be predicates in their own right, but are rather restricted in the inflectional categories they may take, being limited to the durative aspect and apparently not occurring in the future tense. They do not take objects, and seldom take subjective nominal arguments. They may occur as nominal arguments. Based on Jacobsen's observations and the preceding discussion, we can arrive at the following conclusions regarding the co-occurrence of properties in Southern Wakashan word classes. Verbs

Nouns

yes

no

yes only as RC's

Past Tense

yes

Future Tense Complements

Transitive Perfective Aspect Article

Demonstratives

Prepositions Adjectr yes

no

no

no

yes

no

no yes

yes

no

yes

yes

no

no

no

yes no

no yes

yes no

yes yes?

Table 3.2 Word Class Properties As can be seen from this summary of the properties of the various elements, there is significant overlap, but each category may be distinguished by the combination of properties, allowing us to separate words into their separate word classes. In concluding his investigation of category distinctions in Southern Wakashan, Jacobsen (1979b: 142) observes: The fairly traditional inventory of parts of speech that we have recognised must, after all, reflect inherent distinctions of the real world and human cognition, concerning entities and their attributes and relationships. We now move on to some of the properties discussed above. 3.7.2 Distributional Properties In determining basic word categories, it may be worth noting the distributional properties of words within the sentence. To take a classic example, Sapir & Swadesh (1939: 235) provides an argument for the lack

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Word Structure and Categories

58 Linguistic Theory and Complex Words

All normal words express a potential predication, which becomes an actual predication on the addition of paradigmatic suffixes. This statement applies to words corresponding to English verbs, adjectives, prepositions, and adverbs. The argument takes the following form: Sapir & Swadesh (1939) provides examples of words in both predicative and non-predicative positions, demonstrating apparent total overlap in distribution. He then claims that this demonstrates that any word may occur in any position, thereby concluding that all words have the same word class status, i.e. there are no distinct word classes, providing the data below. 43. Cat.

Non-predicative

N

waiaakma quu?as 'a man goes'

Adj waiaakma ?iih P

Predicative

'a large one goes'

ciqsiXma ?uuk<°ii quuPas 'he speaks to a man'

Adv waiaakma ?athiya 'he goes at night'

quu?asma

'he is a man'

?i/7jmaa

'he is large'

?uuktoiimsL quuPas 'he does it to a man' ?athiyama. waiaak 'it is at night that he goes'

This chart is adapted from Sapir & Swadesh (1939: 235-6), removing the more poetic translations that attempt to imply the predication of all parts of speech. What is clear from these examples is that they are all syntactically simple: there are no complex argument phrases nor even transitive relationships. This point is significant for the following reason: if we choose more complex phrases, we will see quite readily that there are constraints on what is allowed to become a predication in the parallel forms. For instance: 12 44. a. kaXhsiX kaXh -siX

Piih quuPas Piih10 quuPas

b. PiihweePin quuPasPi Piih" -weY?in quuPas =PiY

'a large person appeared' visible -MOM big person 'he was a big man' big -3s.QT person =DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

of syntactic categories in Nuuchahnulth that is taken up by all succeeding linguists writing introductory textbooks and the like.

Word Structure and Categories

Piih Piih10

'he was a big man' person -3s.QT big

In (44a), the adjective ?iih 'big' is acting as a modifier of quu?as 'man', while in (44b) it is acting as a predicate. However, the semantic function of this adjective is the same in both cases, as suggested by the translation. Furthermore, there are no examples in any of the data where the noun quu?as 'man' acts as the predicate and the adjective remains in situ. The prediction is that (44c) would be ungrammatical even though it is entirely possible to have either quu?asma as a predicate or ?iih as a freestanding adjective. The explanation for the grammaticality of (44a) and (44b) and the ungrammaticality of (44c) is straightforward if we assume that there are ordering restrictions imposed on constituents in Nuuchahnulth, constraining the appearance of certain elements in the initial position within the sentence. Let's take another example. 45. a. cuSuuk^afinmePitafiaX PaXa quuPas SuSuuk^afrn -ma -'ita -'ap -'aX PaXa quuPas invite -go toward [R] - down -CAUS -NOW two person 'then he had two people go about the village inviting' b. PaXaaX quuPas naacuksiX 'it was 2 people who went searching' PaXa-'aX quuPas naaCuk-siX two -NOW person look for -MOM c.*quu?asaX PaXa naacuksiX

'it was 2 people who went searching'

There are absolutely no examples in the data of the noun quu?as being a predicate while the numeral modifying it remains in situ, just as with the adjective in (44). If, in fact there is no category distinction, this is rather curious. The following example illustrates the predication of quantifiers. 46. a. hitaaqXiPaX PayePi quuPas 'the many people went into the woods' hita-'a-qXiX-'aX ?aya=?r quuPas LOC-in woods-NOW many-DEF person b. PayiiciPaX Pay a -iiCiX -'aX

quuPas quuPas

'there were many people (there)' many -INC -NOW person

c.*quu?acii£i?aX Paya quuPas -ii£iX -'aX Paya

'there were many people (there)' person -INC -NOW many

Again, it is never the case that the noun acts as a predicate while the quantifier remains in the argument position.13

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

c*quuPaswePin quuPas -we^Pin

59

60 Linguistic Theory and Complex Words

47. a. quuPasma quuPas -ma*

PiihPii Piih10 =Pi*

b. Piihmaa quuPasPi Piih" -ma* quuPas =?iY

'the large one is a man' man -3.IND big =DET 'the man is large' big -3.IND man =DET

What is interesting here is the potential ambiguity in the second example that could as easily be glossed as 'he is the big man'. This is not an option for the first case, reflecting the status of ?iih?ii 'the large one' as a nominalisation of some kind, perhaps a reduced relative. Compare Jacobsen's (1979b: 134) example from Makah: 48.

XuiuuPaXits PuPuyapui Xuiuu-PaX-its PuPuyapui

'I had a good dream' good -now-ls.iND dream

Rather than glossing this as T am good to/with the dream', Jacobsen provides the most natural translation of the meaning, which places the adjective 'good' back into adjacency with its complement, 'dream'. This then would suggest the alternative 'he is the big man' for (47b). What we have seen in this section is the fact that Nuuchahnulth, like other languages, does distinguish word classes and that there are both empirical and theoretical reasons for maintaining this distinction. 3.7.3 The Status of Word Class Categories The notion that languages may blur the distinction between different word categories is not specific to Nuuchahnulth. As a matter of fact there are numerous examples of this kind of operation in natural language. We will examine two examples of this in the following sections before returning to a discussion of Nuuchahnulth. Cantonese is a typical Chinese language in many ways, exhibiting characteristic properties shared by many varieties. One of the properties is the blurring of distinctions between verbs and adjectives in a number of contexts. For example, the following utterances are prototypical: 49. a. Neihhaamjo houdo. b.*Neihhouhaam.

'you cried a lot' 'you are very weepy'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Swadesh (1939) provides some additional examples, one of which appears to be a counter-example to the position presented here, in that the noun appears within the predicate while the adjective appears outside, and so it bears further examination (from Jacobsen 1979b: 85).

Word Structure and Categories

b. Neih hou gou.

'you have got very tall' 'you are very tall'

As can be seen from these examples, the overlap between verbs, as in (49), and adjectives, as in (50), is not perfect, but there are situations where the two coincide in their usage. Thus, one can posit a certain vagueness about the distinction and a certain degree of overlap between the two categories. Our second example comes from a language a bit closer to home: English constitutes another good example of the blurring of the distinction between word classes, in this case those of noun and verb. In the following examples one can see that nouns, even very recently coined ones, may be converted to verbs without any overt marking to distinguish them. Hence, the distinction between such nouns and verbs is not overt. 51. a. Send me the fax today, b. I'll fax you right away. 52. a. I just bought a new filing cabinet. b. I filing-cabinetted the entire wall of my office. While some examples of this operation may require a degree of imagination, there is general agreement that such constructions are valid in English. As all of the examples above should confirm, determining the exact status of word classes in virtually any language may be exceedingly difficult, and it is not clear that there is any merit in doing so. Word classes may shift back and forth without any overt marking of the differences. Thus it should not be surprising if the same is true in Nuuchahnulth. 3.7.4 Wojdak on Category Neutrality More recently, in two papers on the subject of category neutrality, Wojdak (2000, 2001) argues from a syntactic point of view that the distinction between nominals and non-nominals is not universally reducible to the argumenthood of the nominal, employing Nuuchahnulth as the basis for her arguments. Wojdak demonstrates that in Nuuchahnulth there is evidence for a [+/-N] syntactic category distinction which is not reflected in a categorial restriction on argumenthood, demonstrating first that any open-class

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

50. a. Neihgoujohoudo.

61

62 Linguistic Theory and Complex Words

lexical item may occur in the initial predicate position, as in (53) (Wojdak (2001 's (la-c)) and may bear the determiner =?r, as in (54) (Wojdak (2001 's (2a-c)). 'the man is working' work-3sg.real man =DET

b. hiix^athiPis mamuukPi hiix^athi -Pis mamuuk =?f

'the working (one) is cranky' cranky-3sg.real work =DET

c. CakupPis hiix^athiPi Cakup -Pis hiix^athi =?f

'the cranky (one) is a man' man-3sg.real cranky =DET

54. a. hiix^athiPis CakupPi hiix^athi -Pis Cakup =?f

'the man is cranky' cranky-3sg.real man =DET

b. mamuukPis hiix^athiPi mamuuk -Pis hiix^athi =?iv

'The cranky (one) is working' work-3sg.real cranky =DET

c. CakupPis mamuukPi cakup -Pis mamuuk =Pi*

'The working (one) is a man' man-3sg.real work =DET

This clearly supports the observations made previously concerning the flexibility of categories. She then goes on to demonstrate the existence of modification constructions in the language that are sensitive to a [+/-N] categorial restriction. Wojdak concludes that Nuuchahnulth is not a syntactically category-neutral language and that there is syntactic evidence for a [+/-N] distinction, allowing both nominals and non-nominals alike to serve as the base of an argument construction that is derived syntactically via a determiner. Wojdak's examination is a good example of how one must proceed in addressing the issue of category neutrality, but there remains much work to be done. There is an important point that is perhaps not clear from the discussion above. As will be discussed in Chapter 8, elements of the argument may be fronted into the predicate from object position in certain contexts.14 Where this occurs, there should be an ambiguity in the readings of the sentence that is not made clear in the examples in (53). Given the strict order of ADJ > N within the argument phrase, any sentence containing a N in predicate position will be unambiguous, as in (53c). However, any predicate containing an adjective in predicate position will necessarily be ambiguous, as demonstrated in (55).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

53. a. mamuukPis cakupPi mamuuk -Pis Cakup =Pr

Word Structure and Categories

55. a. hiix^athiPis [Cakup?i]SUBJ hiix^athiiPis [ tj CakupPi]0Bj

63

'the man is cranky' 'it is the cranky man' (i.e., NOT the happy man)

hiix athi

Cakup

This ambiguity may contribute to the confusion regarding category neutrality and should be borne in mind in any examination of the facts. It would seem a prudent approach to assume that one should not conclude the category of a form in isolation, but always within some context. This means that subcategorisation frames in Nuuchahnulth will necessarily be somewhat different from those in a language with a clear distinction among categories. It will also complicate the task of anyone building a lexicon for computational purposes, for which see Whitman (2002). 3.8

Summary

This chapter has examined the various properties, both structural and categorical, of words in Nuuchahnulth. This investigation is necessary in order to understand the nature of wordbuilding strategies in the language and will be useful in understanding the discussion to follow.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b.

4

Suffixation is the most productive means of word formation in Nuuchahnulth. There are hundreds of suffixes of various types interacting and combining to form the word. Since the bulk of affixation in Nuuchahnulth is suffixation, this will be the sole subject of this chapter, which is divided into the three categories of inflection, aspect, and derivation. 4.1 Inflection versus Derivation There has been extensive discussion in the literature of the distinction between inflection and derivation, e.g., Scalise (1988), Badecker & Caramazza (1989), who present two differing views. Bickel & Nichols (forthcoming) observe: Cross-linguistic variation in the forms and categories of inflectional morphology is so great that 'inflection' cannot be defined by simply generalizing over attested inflectional systems or paradigms. Rather, we define it as those categories of morphology that are SENSITIVE TO THE GRAMMATICAL ENVIRONMENT in which they are expressed. ... Inflection differs from derivation in that derivation is a lexical matter in which choices are independent of the grammatical environment. The relevant grammatical environment can be either syntactic or morphological. The syntactic environment is relevant, for example, when morphological choices are determined by agreement. In Nuuchahnulth there are many areas where the distinction is blurred. Early on, Sapir (1921:134) remarked upon this ambiguity:

64 10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Suffixation

What if we add the preterit tense suffix -it? Is not inikw-ihl'minih-'is-it necessarily a verb: 'several small fires were burning in the house'? It is not. It may still be nominalized; inikwihl'minih'isit-'i means 'the former small fires in the house, the little fires that were once burning in the house.' It is not an unambiguous verb until it is given a form that excludes every other possibility, as in the indicative inikwihlminih'isit-a 'several small fires were burning in the house.' We recognize at once that the elements -ihl, -'minih, -'is, and -it, quite aside from the relatively concrete or abstract nature of their content and aside, further, from the degree of their outer (phonetic) cohesion with the elements that precede them, have a psychological independence that our own affixes never have. His examples are recast in more technically accurate representations as in (1). Note that (la) provides a verbal interpretation with the past tense marker, -(m)it, while in (lb) the same marker appears on a noun, indicating a former state, i.e. 'the former ...'. This has no doubt contributed to the controversy concerning part of speech classes in the language. 1.

a. Pink^iirfiinhPisita 'several small fires were burning in the house' 10 Pink -'ii-rhinh-Pis-(m)it-ma" fire -in house -PL -DIM -PAST-3.IND b. Pinkwiiifiinh?isitPi

'the former small fires in the house, the little fires that were once burning in the house' Pinkt0-'ii-rhinh-?is-(m)it=?f fire -in house -PL -DIM -PAST =DEF

There are many similar cases of simple nouns with past tense indicating a former state, i.e., 'the late ...' for humans, 'the broken ...' for objects, 'the former ...', etc. as shown in (2). 2.

a. Sakyakuk'itPi Sakyak-uk^tm^t =?f

'what had been his knife' knife -POSS -PAST =DEF

b. qasiimitPi qasii -(m)it =?ir

'his former eyes' eye -PAST -DEF

c. yimhaaPamit yimhaaPap -(m)it

'the late Shames-Them' Shames-Them -PAST

d. tiipanit tiipat -(m)it

'the late David' David -PAST

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Suffixation 65

Linguistic Theory and Complex Words

Some of the inflectional morphology of Nuuchahnulth is marked by clitic rather than suffix, and in a sense is beyond the domain of this investigation, but it will be useful to examine its status, and this will be pursued in greater detail in Chapter 8 on morphosyntax. It should be noted that in this book the terms 'inflection' and 'derivation' will be used in a strictly descriptive sense and no theoretical significance should be assigned to them. Theoretical level distinctions will be based on stratal segregation, to be discussed in Chapter 6. 4.2 Paradigmatic Inflection In terms of inflectional suffixes, we may recognise two sub-domains, based on the regularities that hold of a set of suffixes. Suffixes that operate in sets, exhibiting characteristics of person and number, will be referred to as paradigmatic suffixes and those that occur as singletons will be termed non-paradigmatic, following terminology introduced by Sapir & Swadesh (1939). There are many inflectional paradigms that typically exhibit contrasts in number and person often, although not always, with a single member for the third person. All of these paradigms represent independent modal marking and they are invariably mutually exclusive. There is a certain amount of dialectal variation in these paradigms and for this reason paradigms from three distinct dialects, Tsishaath, representing the Southern Nuuchahnulth group, Ahousaht, a member of the Central group, and Kyuquot, a Northern Nuuchahnulth variety, will be presented where available. Below are provided sets of paradigms for the various mood/person markers appearing in Nuuchahnulth. For the most part, different varieties agree on these, but where there are differences, they are noted below. Further research may reveal additional paradigms not provided here. We will proceed through each of the categories, providing the paradigm categorised by person and number along with examples of the usage of each mood, beginning in the following section with the indicative mood. 4.2.1 Indicative Mood Perhaps the most common mood encountered is the indicative. This mood appears in conversations and is widely used in statements of fact. The form of the elements that make up this paradigm is as shown in Table 4.1.'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

66

Suffixation 67

3

Ahousaht Sing Plur -srs -nis -Pick -Picuus -Prs

Kyuquot Sing Plur -srs -nis -a*c -Picuus' -Prs

-itais

Table 4.1 Indicative Paradigm Examples of the usage of the indicative follow. Note that, as with many of the paradigms, there is no number distinction made on the affix with respect to third person. 3.

a. naPaaPaXah, waaPaX naPaa -'aX-(m)ah waa -'aX

'"I understand,' (someone) said" hear -NOW-1 S.IND say-NOW

b. PahniimePic Pahnii -(m)e?ic

'here you are wetting yourself!' that way -2s IND urinate -in house

Tuq^ii ftuf-'ii

c. PaXjMiima yaqii nuuk 'there are two of these songs' ?aX-rM*i-maY yaq^-^ii nuuk two -CLS-3.IND REL-3.INDF.REL song d. finrhiiCiXPaqXPicuus Tinrhi -iiCiX -PaaqX-Picuus

'you (pi.) will become a snail' snail -INC -FUT -2PL.IND (Ahous)

e. hiitkinPis finiiXPi hiitkin -Prs finiiX =Pr

'the dog was strange' strange -IND.3 dog =DEF (Ahous)

f. hayimhinis hiiii hayim-hr-nfs hii-(y)ii

'we don't know where he would be' not know-DUR-lND.lp there-lNDF.3(Ahous)

g. sutwaaqkwac hiisink 'you're supposed to kiss first'(Kyu) sut-wr-PaaqX-wa--ax his-ink0 you-first-FUT-QUOT-2s.lND both-with h. cawaaksis ?iqhi hii dawa -ak^-sis ?iq-hr hii

'I alone am still there' (Kyu) one-DUR-1 s.lND same-DUR there

piis 'they each ate one fish' (Kyu) i. dawaakrhinhintitais cawa-akw-rhinh-int-it-ai-s piis one-DUR-PL-PAST-lND-PL-lND fish It is worth noting a special, portmanteau, form that arises in one particular context in Tsishaath, i.e., immediately following the inflectional morpheme -(m)it PAST. Note the example in (4).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

1 2

Tsishaath Plur Sing -(m)a'h -(m)in -(m)e*Pic -(m)e-Picuu -(m)ahsuu -ma*

68

4.

Linguistic Theory and Complex Words

Pukiaamita qeePas 'her name was Keas' ?u -(k)iaY-(m)it -ma* qeePas REF -name-PAST-3s.lND Keas

4.2.2 Absolutive Mood The absolutive mood is found most commonly in the recitation of mythical tales and the like, where it performs the duties of the indicative in normal conversation. It should be noted that the third person is typically zero-marked in the absolutive, and thus often difficult to identify. The following are the suffixes associated with the absolutive mood: Tsishaath Sing 1 2 3

r uYk

Plur

Ahousaht Sing

-ni* -suu 0/-si

Plur

Kyuquot Sing -s -k 0

Plur -in -suu -?aYi

Table 4.2 Absolutive Paradigm Examples of the absolutive abound in the textual materials relating to myths and the following are typical examples. 5.

a. PahPaaPaXsi hiniisPaX PahPaa -'aX -si* hiniis -'aX

'I (shall) be carrying it' then-NOW-ls.ABS carry along-NOW

b. PahPaaPaXsuk huPacaCiX PahPaa -'aX-su'k huPaca -siX

'then you will return' then -NOW -2S.ABS return -MOM

c. PahPaaPaXsuu waaPaX ... 'then you (will) say...' PahPaa -'aX -suu waa -'aX then -NOW -2p.ABS say -NOW d. PeePinhaCPisni DUP- [L] Panaft0 -aC -Pis -ni

'we had but a small margin' SUF-small-at margin[RL]-DiM-lP.ABS

Rose (1981:226)'s comments on the distribution of the absolutive mood in Kyuquot may be worth noting at this point: In contrast to the indicative mood, all of whose persons are employed, the second person of the absolutive mood virtually never occurs. This may be in part due to the similarity of

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Nakayama (1997a:31) views the indicative mood as being used for the assertion of facts. He states that the '[indicative is by no means a default mood, and its use can add a strong assertive tone to a statement.'

Suffixation 69 second person absolutives, -k II and -su: Up to second person interrogatives -haY-k (~ -k in third or later syllables) II and -ha*-su: (~ -h-su: in third or later syllables) IIP.

The quotative mood is very common in the recitation of texts, indicating that the narrator did not actually witness or experience the events themselves, but that they are simply recounting what was said. Its most typical rendition in translation, if at all, is 'it is said' or 'it was said'. The quotative indicative mood is principally used to indicate that an event or state was reported by some party other than the speaker or listener. Hence, the assertion is an announcement of fact (namely, that something has been reported). Reported facts, whose predicates are marked by -war, are generally expressed in the indicative mood, obligatorily so if the subject is coreferential to the speaker or listener. (Rose 1981:229-230) Tsishaath Sing 1 2 3

Plur

-we*?isi -wePinni -we*?incuk -werPincuu -weY?in

Ahousaht Sing

Plur

-wa'Picas -wa?iCin -wa-Pick -wa^icuus -wa?is

Kyuquot Sing

Plur

-wa*sis -wa*nis -wax -waxuus -wa*s -wa'tais

Table 4.3 Quotative Paradigm Examples of this mood include: 6.

a. takaaPaaqXwePincuk ... takaa-PaaqX-we^Pincuk

'nevertheless you'll go and ...' nevertheless -INTENT-2S.QT

b. suk°iXPaaqXwe?incuu suu-siX-PaaqX-we^Pincuu

'you are to go after it' hold -MOM-INTENT-2p.QT

c. waaisiXwaPick wai -[L] -siX -wa*?ick

'(they say) you are going home' go home -GRAD-MOM-2S.QT (Ahous)

d. hininwaPis q^ayaacikrhit '(they say) the Son of Wolf came' hin-in-wa'?rs q^ayaacik-rhrt there-come-3.QT wolf-son of(Ahous)

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.2.3 Quotative Mood

Linguistic Theory and Complex Words

e. waisiXwass wai -siX-wa*sis

'I'm said to have gone home' go home -MOM-IS.QT (Kyu)

f. puukPaaqXwatais puuk°a -PaaqX-waYtais

'they're going to play poker' poker -FUT-3P.QT (Kyu)

Rose (1981:230) observes that there 'is no Kyuquot counterpart to the Port Alberni quotative mood -we'?in, with its own unique pronominal series.' She further observes that: There is another quotative type of construction in Kyuquot. In an argument or interchange where one wished to stress one's assertion of an utterance, the morpheme (compound) -waa(q)s is attached to an indicative predicate. Evidently, the q is found optionally in predicates with singular subjects (but never in ones with plural subjects). This morpheme typically indicates that the assertion thus marked contradicts one made previously, (ibid) 7.

'we have a nice one!' nice-have-lP.lND -QUOT.EMPH

Xuinaakniswaas Xui -na^-nrs-waas

4.2.4 Interrogative Mood The interrogative mood is used to indicate the posing of a question of either the yes/no type or the informational type. Tsishaath 1 2 3

Sing

Plur

v

-hin -ha~suu

-ha s -fork -ha

Ahousaht Sing

Plur -hin -hsuu

-hs -ha*k -h

Kyuquot Sing

Plur

-ha*s -haYk -har

-hin -ha"suu

-ham

Table 4.4 Interrogative Paradigm Examples of the use of this mood include: 8.

a. wikhaasuu Paanaqh qahsaapmihsa'do you not really want to kill me?' wik-ha*suu Paanaqh qahsaap-mihsa not-2p INT really kill -want to b. wikithas waa maatmees 'did I not say, O tribes, ...' wik-(m)it-ha*s waa maatmaas-ee not-PAST-ls.lNT say tribes-VOC

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

70

Suffixation 11 'are you not ashamed?' not -2s.INT ashamed

yimhaa yimhaa

d. waasih kapCaa waasi -h kapCaa

'where is Kapchaa?' where-3.iNT Kapchaa (Ahous)

e. Paqishithsuu siiXuk Paqish -(m)it-hsuu siiXuk

'why did you move?' (Ahous) why -PAST -2P.INT move house

f. waayaqpimith Xawaa town 'which was the closest town?'(Ahous) waayaq-api-(m)it-h Xawaa town which-most-PAST-3.iNT near town As one can tell from this data, the interrogative mood may occur with question words such as waasi 'where', ?aqish 'why', waayaq 'which' etc., or in yes/no type questions. 4.2.5 Purposive Mood The purposive mood has, as its name suggests, a function of indicating the intended use or aim of some action with respect to the interlocutors or other parties. The forms of the purposive mood are as follows: Tsishaath Sing 1 2 3

Ahousaht

Plur

-'aah(sa) -'aani -'eePic(a) -'eePicuu -'aahi /-'eePit(a)

Sing

Plur

-'aah -'aani -'aaPick -'aaPicuu -'aaPit

Kyuquot Sing -'aa(h)s -'aac -'aat

Plur -'aan -'aacuu -'aatai

Table 4.5 Purposive Paradi gm Examples of the occurrence of this morpheme include those provided below. 9.

a. hasiiCiPeePita PayePi quuPas 'so the many people may know' hasii-iiCiX-'eePita Paya=?i* quuPas know-lNC-3.PURP many=DEF person b. hinaaCiPaani hina -aCiX -'aani

'so we can set out' LOC -go out to sea[M] -lp.PURP

c. k°eePik huPacaCiPeePic k*ee?ik huPaca -siX -'eePic

'come, so you may return' come(sg.)! return -MOM-2s.PURP

d. q^iisaahaahi PucaCiX ... q^ii -sarhi-'aahi Pu-ca-siX

'so that it may go to ...' which-because-3s.PURP REF-go-MOM

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

c. wikhaak wik -hark

Linguistic Theory and Complex Words

e. wikaaPick hiixtaq wik -'aaPick hiixtaq

'so you won't have an accident' not-2.PURP have an accident (Ahous)

f. wikaaPit cicikpaia wik -'aaPit DUP- cik -paia

'so it will not be out of alignment' not-3.PURP aslant-on each side(Ahous)

g. Puk<0iidaa(h)s Puk^a -iic-'aahs

'that's so I can have my own' self-belong to-Is.PURP (Kyu)

h. waisiPaan wai -si(X)-'aan

'let's go home!' go home -MOM-lp.PURP (Kyu)

As can be seen from these examples, the purposive mood is employed to indicate the reason or purpose for taking a particular action. As such it is most commonly found to occur with perfective aspectual forms rather than imperfective ones. According to Rose (1981), the purposive most commonly occurs as a hortative in 1st plural subject predicates, as in (9h). 4.2.6 Dubitative Mood The dubitative mood expresses some degree of uncertainty, or a lack of confidence as to the factual accuracy of a statement. It is marked by the following set of suffixes, distinguished for person and number: Tsishaath Sing 1 2 3

Plur

Ahousaht Sing

Plur Y

-qa £ina -qaxsa -qa'£in/-qa'Sana -qaxa -qa"Cka -qaxuuwa -qa*Cka -qaxsuu(wa) -qa*6a -qa*6a

Kyuquot Sing

Plur

-qax -qaYSk -qar6

-qa*6in -qaxuu -qaYC?ai

Table 4.6 Dubitative Paradigm The following are examples of this paradigm. 10. a. CaaniXqaCka Caani -iX -qa'Cka b. iuChimPakqacsa PahPaa iuChim -'ak-qaxsa PahPaa

'why aren't you the one ...?' at first -MC -2s.DUB T think this brideprice consists of...' marriage payment-POSS-ls.DUB that

c. naPatahPaXqacsuuwa haweeh T think you are listening, O Chiefs' na?atah-'aX-qaxsuuwa hawiih-ee listen-NOW -2P.DUB chiefs-VOC d. PaXpupititqaSa PaXpu -pit -(m)it -qaYSa

'it might have happened seven times' seven -times -PAST-3.DUB (Ahous)

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

72

13

e. Xupaasqac Xup-(y)a*-(s)as-qa*S

'it must be hot / perhaps it's hot' hot-CONT-really-DUB (Kyu)

f. hiihPaXqacu haPuk hii -(q)h -'aX -qaxuu

'maybe you'll be where you're to eat' there -MW-NOW-2P.DUB eat (Kyu)

It should be quite clear from these examples that the notion of the dubitative is one of uncertainty, or a lack of clarity as to the factuality of a statement. 4.2.7 Inferential Mood The inferential mood shares many similarities with the dubitative, in terms of the degree of certainty. However the inferential carries the additional connotation of implying that something is true, which is not the case for the dubitative.2 The paradigm for the inferential markers is presented below. Note that there is some variation in the form of individual markers. Tsishaath Sing 1 2 3

Plur

-(c)sa'S"as -cVnaTas -CkaY
Ahousaht Sing

Kyuquot Sing

Plur

,

-(c)sa*Ss -Cka^Ts -Ca-^§

Plur -Cina^s -cuuwa^s -Ca'tais

Table 4.7 Inferential Paradigm 11. a. maPasPaaqXdanaTas 'I suppose we'll show our tribal greatness' maPas -PaaqX -ca*naTas tribe -INTENT -1P.INFER b. CaaniXsuuwaS'as Pahnii hiiiPiihsap 'you ought to just let him try' CaaniX-suuwaS'as' Pahnii hiii?iih-sap at first-MC-2p.lNF that try-MC c. hiiSaaTas XihaqsiX hii -ca'Tas Xihaq -siX

'it seems there is a lean one' there -3.INF thin, lean -MOM

d. hisxiiXCkaS'as Puustuup T guess you won't amount to anything' hisxiiX -Cka-^as Puus -tu*p amount to -2S.INF some -thing e. haCaqsiPaXsaTas ha -Caq- siX-'aX -saS"as f. payinkinaTs pink- -^- -Cina'S's

'I seem to have completed it' complete -...ed -MOM -NOW -IS.INF 'we are evidently/obviously still awake!' awake-PL-1 P.INFER (Kyu)

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Suffixation

74

Linguistic Theory and Complex Words

From the above data, we can see that there appears to be some overlap between the inferential mood and the dubitative. The distinction may be made in terms of the degree of certainty expressed.

The conditional mood is typically employed in subordinate clauses conjoined by elements expressing some sort of condition upon which the clause is predicated or a sense of habituality. This may include clauses initiated by 'if, when, as though, ...'and other indicators of counterfactuality. The paradigm takes the following form: Tsishaath Sing 1 2 3

-quus -quuk

Ahousaht

Plur -qun -quusuu -quu

Sing -quus -quuk

Plur

Kyuquot Sing

a>. 3

-quus -q in -quuk -quusuu -quu -quu Table 4.8 Conditional Paradigm

Plur -qun -quusuu -quuPai

12. a. q^aamin qaqahPaqXqun 'we are as though ours were dying' q^aa-min DUP- qah-PaqX-qun thus -lp.lND PL- die-inside-lp.CND b. wikuuXquusuu

tumiis

wik-qawiX-quusuu tumiis

'if you (pi.) do not have charcoal on your faces, then ...' not -on face[M] -2p.CND charcoal

c. Puyi SuqmaftiqstuPiikquuk 'when you need to urinate' Puyi Suq-mafiiqstuX-'iik-quuk when urinate-want[M]-HYP.FUT-2sCND d. q^aahwinkquus Pahnii q^a -hwink -quus Pahnii

'... if I used that' thus -use [L]-ls.CND

e. wikatquu TaCuuPat wik -'at -quu SaCuu -'at

'may there be no wounds' (Ahous) not -PASS-3.CND wound -PASS

huu f. PucaCiPaXq^in ?u -ca-siX-'aX-q^in huu

'we used to go way up ...' (Ahous) REF-go to-MOM-NOW-lP.CND yonder

that

It is quite clear from the examples above that the conditional is used in situations where there exists some condition for the accomplishment of a task, often indicated by conjunctions such as 'if, when, as though,...'.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.2.8 Conditional Mood

Suffixation 75 4.2.9 Relative Mood

Ahousaht

Tsishaath Sing 1 2 3

-Pi'tqas -Prtqak

Plur -qin -Prtqsuu -Pi-tq

Sing -qs -Prtk

Plur -qin -Pitqsuu -Prtq

Kyuquot Sing Y

-qa s -Prtk -Pftq

Plur -qin -Prtqsuu -Pftqai

Table 4.9 Relative Mood Paradigms Relative clauses containing these suffixes occur quite commonly in Nuuchahnulth and are exemplified by the following cases: 13. a. yaTiisPitqas yaq40 -'iis -Pi'tqas b. Paniicuq^ashuiPitqak Pana -iicuq-ashui-Pf tqak

'what I was given to eat' REL -eat -ls.REL 'which is as long as your chest is wide' that much-as long as-on chest-2s.REL

c. wikmaa ?uh yafrisqin haPuk 'that is not the one which we eat' wik-mar Puh yaq^-'iis-qin haPuk not-3s.iND is REL-eat-lp.REL eat d. yaaq^iiitPitqsuu wik haPuk 'that which you haven't been eating' t0 yaq -ii-(m)it-?itqsuu wik haPuk REL -do to [L]-PAST-2p.REL not eat e. Piiqhuk^aX q^isitPitq 'he told what he went through'(Ahous) Piiqh-uk-'aX q^is -(m)it -Prtq tell-DUR-NOW do so-PAST-3.REL As can be seen from the examples above, the relative mood marker may appear with or without the relative pronoun, yaqm, although it is more commonly found to co-occur with it. 4.2.10 Subordinate Mood This mood appears in subordinate clauses, and shares many superficial resemblances to the relative mood discussed above. They are, naturally, in complementary distribution and pose no great difficulties in terms of identification, although the actual markers overlap substantially. The relevant suffixes, marked for number and person are as illustrated in Table 4.10.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

The so-called relative mood occurs within relative clauses, typically along with the relative pronoun, yaqa. It is differentiated in much the same way as the indicative mood, with three persons and two numbers, the third person making no number distinction. This is described below:

Linguistic Theory and Complex Words

Tsishaath Sing 1 2 3

-qa*s -Prtqak

Ahousaht

Plur

Sing

Plur

-qs -k

-qin -Prtqsuu -qar

-qin -qsuu -q

Kyuquot Sing -qa*s -qa*k -qar

Plur -qin -qa*suu -qa7i

Table 4.10 Subordinate Paradigms The subordinate clauses with which these markers co-occur are typically headed by a certain set of conjunctions, including ?ani 'that', ?uyi 'if, etc. Examples of the usage include: 14. a. cuxsiXPaaqXqas qahsaap dux-§iX-?aaqX-qa*s qahsaap

'for I am going to stab and kill him' stab-MOM-INTENT-Is.SUB kill

b. ... q^isPaXPitqak ymufaa q^is -'aX-Pitqak yuuq^aa

'...just as you did' do thus -NOW -2S.REL also

c. Pani nuuPiPaX'atukqa Pani nuu-iX-'aX-'at-uk-qa*

'that its (song) had been taken up' that sing-MC-NOW-PASS-POSS-3.SUB

d. Puqhiip suuhaaq Puqhiip suuhaa -qa*

'he thought it was a salmon' think salmon-3.SUB (Kyu)

Examples such as these demonstrate the distinction between the relative and the subordinate moods, which, though sharing similar morphs, serve different functions in the clause. 4.2.11 Indefinite Relative Mood The indefinite relative mood, like the relative mood discussed above, is employed in relative clauses, often reduced and without a relative pronoun. It is distinct from the relative mood in that it expresses a degree of uncertainty derived from either lack of knowledge or the fact that the statement is projected into the future. The various markers for this mood, distinguished for person and number, appear below. Tsishaath Sing 1 2 3

-(y)iis -(y)iik

Plur -(y)in -(y)iisuu -(y)ii

Ahousaht Sing -(y)iis -(y)iik

Plur -(y)in -(y)iisuu -(y)ii

Kyuquot Sing -(y)iis -(y)iik -(y)ii

Plur -(y)in -(y)iisuu -?a-l

Table 4.11 Indefinite Relative Paradigms

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

76

Suffixation

11

Examples of the usage of these markers are provided in the following sentences. '... from that which you do?' REP- thus -2p.lNDF.REL

b. q^ishcuuyiisuu qahsaap siya 'how you may succeed in killing me' q0,ishxuu-(y)iisuu qahsaap siya how-PROB -2p.lNDF.REL kill me c. haPuksiX quurhiiPanitii 'however many people that are invited ate' ha?uk-siX qtoarha-'rX-'at-(m)it-(y)ii eat -MOM as many-invite-PASS -PAST -3.INDF.REL (Ahous) d. ?inumsaqkuka,ina niXaak 'it seems that we rarely quarreled' ?inums-aq-kuk-(y)ina niXaak seldom-very-apparently-1 P.INDF.REL fight (Ahous) e. q^iPiimitii tf\ -Pii -(m)it -(y)ii

'he went to some place' (Ahous) that.which -go to -PAST-3.INDF.REL

f. hayimh q^isi hayimh q^is -(y)ii

'they don't know what happened' not know happen-3.lNDF.REL (Kyu)

g. q^aaPaaqXi iuucma q^a-PaaqX-fy^i iuucma

'I wonder what the woman will be like' thus-FUT-3.lNDF.REL woman (Kyu)

4.2.12 Relative Dubitative Mood The relative dubitative mood, often glossed 'that X might/what might ...' appears in subordinate clauses indicating doubt or uncertainty. The paradigm, similar for all varieties, appears below. Singular 1 2 3

-(w)uusis -(w)uusik

Plural -(w)uusin -(w)uusisuu

-(w)uusi

Table 4.12 Relative Dubitative Paradigm Examples of this mood are provided in the following sentences. 16. a. hayaaPakat q^iyiihawuusi 'it wasn't known why he died' hayaa?ak-'aX q(0i-(q)h-'at-uusi not know-PASS that which-die of3.REL.DUB

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

15. a. q^aq^aayiisuu DUP- q^aa -(y)iisuu

78

Linguistic Theory and Complex Words

c. huqsiXuusin huq-siX-(w)uusin

'we might capsize' capsize-MOM-lP.REL.DUB

d. yaqwuusi haptsiX yaq^-wuusi hapt-siX

'who might hide?' which-3.REL.DUB hide-MOM (Kyu)

It should be noted that this marker typically occurs in the dependent clause of a matrix verb expressing doubt or uncertainty, e.g. 'it wasn't known why ...' or 'he asked them who ...'. 4.2.13 Imperative Mood The present imperative mood performs the hortative and command-type functions that are common to most languages. As in most languages, this paradigm is restricted to first person plural and second person as presented below, however there are special forms to indicate the various combinations of agent and theme. These are illustrated in the table with the agent listed in the first column and the theme across the top.

lpl Subj. 2sg Subj. 2pl Subj.

3 0bj -'in -'r -'rC

lsg Obj -'rs -'rCas

lplO -'in -'rCin

Table 4.13 Present Imperative Paradigm Examples of the various combinations are provided in (17). 17. a. naPaatahPis haweei naPaatah -'rs hawii -ee

'listen to me, O chief!' listen -2s.IMP chief -VOC

b. huPacaCiPin huPaca -Ci -'in

'let's turn back!' return -along with -lp.lMP

c. Cuu, Xuimaa, nuuPiPi Cuu Xui-ma* nuuPiX -'r

'very well, that's fine! sing!' very well! good-3.lND sing-2s>3.lMP

d. naaCuksiPiC yaqkwaaiuusi 'look and see who is missing!' naaCuk-siX-'iC yaq^-k^aai-uusi look for-MOM-2P.lMP REL-absent -3.REL.DUB

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. PaPaatuu yaqukuusi nuwiiqsu 'he asked them who their father was' PaPaatuu yaq^-uk-uusi nuwiiqsu ask REL-POSS-3.REL.DUB father

79

e. t'iPuupaXiCas fi -'uup -'aX -'iCas

'throw me into the fire!' throw-on the fire [MC]-N0W-2>1.IMP

g. kuhsaajMC kuh -sa*p -'rC

'you all make an opening!' open -2MC - 2p>3s.lMP (Ahous)

h. tiica^is tiiC -'ap -'i*s

'let me live!' be.alive -CAUS -2s>ls.lMP (Ahous)

i. naaPuuqstaPiCin naPuu -Vqsta -'irCin

'join us!' (Ahous) accompany -amongst -2p>lp.iMP

j . ciiqciiqaP^ai DUP- ciq-[L] -(y)a--'ap='a*i

'make them talk!' REP-speak-GRAD-DUR-CAUS=PL (Kyu)

Rose (1981:231) says of the imperative mood: The remaining mood, the imperative, is unique. Its morpheme for lip, -£, is unlike that of other moods and its future morphemes, -im FUT and -k COME, are unique as well. In addition, it is severely constrained in its privilege of occurrence with persons and modes. It can be found only in predicates with Ip or II(p) subjects and either the mode -cim MUST or no mode at all. The organization of mood and postmood affixes in imperative forms is as follows. Such complex encodings of agent and patient are commonplace in Nuuchahnulth, and we shall meet a number of others as we progress. 4.2.13.1 Future Imperative Mood The future imperative mood is, like the present imperative, a mode of command and exhortation, but the time referent for this mood is in the future rather than the more immediate present imperative. As with the present imperative, there are combinations of actor and object acted upon, as illustrated below: 3 obj lpl 2sg 2pl

-'rnirn -'im -'rCim

lsg obj -'rsim -'rCasim

1 pi obj -'rnim -'rCanim

Table 4.14 Future Imperative Paradigm

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Suffixation

Linguistic Theory and Complex Words

18. a. hawiiPaXinim hawiiX-'aX-'inim

'you can now stop ...!' stop-NOW-lpl>3.FUT.lMP

b. takCiPinim suu muustati tak-Ci-'inim suu muustati

'let us each carry a bow!' each-with-lp>3.FUTJMP hold bow

c. iaksuuqstuPisim iaksuuqstuX -'isim

'take pity on me!' take pity -2s>l.FUT.lMP

d. waweePiCim XaXuu t'aatnePis 'you are to say to other children...!' DUP-waa-'iCim XaXuu t'aatnePis REP- say -2p>3.FUT.lMP PL-other child -PL- -DIM e. XuyaCiPafim iirhaqsti Xui-'aCiX-'at-'im iirhaqsti

'let your mind become favorable!' good -MOM -PASS -2>3.FUT.IMP mind

f. Puk^iiiCipiCisim ?u -iii -Crp -'rCisim

'you'll all make it for me!' it-make., -for -2p>l.FUT.iMP (Kyu)

4.2.13.2 Directional Imperatives In addition to the two imperative moods presented above, an interesting feature of Nuuchahnulth is the existence of a set of special markers that incorporate directionality into the command, glossed as either 'come and ...' in the case of the 'Come' imperative or 'go and ...' in the 'Go' imperative. These complex imperatives exist in all the usual forms for imperative markers, including the specification of the object involved. The first case is that of the 'Come' Imperative, which incorporates the notion of direction towards into its basically imperative nature, thereby creating a new portmanteau morph combining the senses of both morphemes. The suffixes are as follows: 3 obj lpl 2sg 2pl

-'fnak -'rk -'irCak

lsg obj

1 pl obj

-'rsak -'rCasak

-'rnak -'rCink

Table 4.15 'Come' Imperative Paradigm Examples of these suffixes are fairly common although not as much as the regular imperative. 19. a. nasiPik nas -iX -'rk

k^atyeet k°atyat -ee

'come and see, Kwatyat!' look -go-come and... Kwatyat-voc

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

80

Suffixation 81 'come and take it there!' deliver -NOW -2p.come and...

c. wePiCu^aXisak Caani we?iCup-'aX-'isak Caani d. suk^iPaXiCasak suu -siX-'aX-'rCasak

'come, now, and put me to sleep first' sleep [MC]-NO W-2>1. come and., at first 'come and get me!' hold -MOM -NOW -2p>l.come and...

The 'Go' imperative is another portmanteau morph incorporating the notion of direction away from the speaker into its basically imperative nature. 3 obj lpl 2sg 2pl

lsg obj

1 pi obj

-Cr -Ca*suu -csuu

Table 4.16 'Go' Imperative Paradigm Examples of the use of these suffixes include the following: 20. a. PaakuuXcsuu PaakuuX -csuu

'you people go and borrow it' borrow -2p>3.go and...

b. waisaapCasuu PeePeePisa 'let him go home quickly' wai -sa*p -Casuu PeePeePisa go home-MC-2p>3.go and., do quickly Saatusrhit c. haiiiyasPaXCi haiiii-'as-'aX-Ci Saatusrhit

'go and ask Deer' invite -go to-NOW-go and... Deer

d. huPacaCiPin huPaca -CiPin

'let's turn back!' return -lp.go and ...

e. PunaahC ?u -na*h -C

'go and look for it!' REF -seek- 2>3.go and ... (Kyu)

There are gaps in the system for this morpheme that indicate the infrequency of certain combinations. Nevertheless, it is clear that this is a paradigmatic set of markers as are all of the other morphemes in this section. This concludes the discussion of paradigmatic inflection. The possibility that all of the paradigms discussed above may constitute part of an inflectional clitic cluster to which the verb is raised will be discussed in

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. rhawaaPaXiCak rhawaa -'aX -'i'Cak

82

Linguistic Theory and Complex Words

Chapter 8. In the following section the non-paradigmatic inflectional markers will be presented.

The non-paradigmatic suffixes include all those that behave as single, unitary markers of inflectional categories, such as tense, valencechanging, possession, diminutive, augmentative, etc. Each of these will be discussed individually, focusing on the particular context for their occurrence. 4.3.1 Tense Tense in Nuuchahnulth has been a longstanding topic of interest, going back as far as Sapir (1921), who uses the past tense to make a special point about language (Sapir 1921:134): What if we add the preterit tense suffix -it? Is not inikw-ihl'minih-'is-it necessarily a verb: "several small fires were burning in the house"? It is not. It may still be nominalized; inikwihl'minih'isit-'i means "the former small fires in the house, the little fires that were once burning in the house." Comrie (1985:13), discussing the concept of tense, comments on this example: '[I]n Nootka, tense can be shown on noun phrases, thus distinguishing "the entity that was an X" from "the entity that is an X".' In the next few sections we will examine the tense morphemes of Nuuchahnulth in their most typical use as a marker of tense on the predicate. In the final subsection we will address the issue of tense appearing on nominals. 4.3.1.1 Past Tense Past tense is not obligatorily indicated on verbs, but may be .called upon where the situation warrants it. The past tense morpheme, -(m)it, has several allomorphs, depending on the phonological environment and the dialect. Following vowels the allomorph -mit is employed (21a); after most consonants, the allomorph is -it. With /p/ and l\l in the coda of the preceding syllable there is a special allomorph which merges the oral stop with the nasal, producing a nasal that is homorganic with the oral

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.3 Non-Paradigmatic Inflection

Suffixation 83 stop (21e,f). There is also a metrically reversed allomorph, -imt, which occurs in certain contexts. Examples of all of these allomorphs follow:

b. qiik^aaiithaPaaia qii -kVi-(m)it-hav=?aaia c. q^isitwePin Pah to q is-(m)it-we*?in Pah

'were they always away for a long time?' long time-absent-PAST-3.lNTER=always 'this is what happened to him' do thus -PAST -3.QT this

d. haheePihtisPaXitah T have been around on every beach' DUP- ha-'ihta-'is-'aX-(m)it-(m)a*h DISTRIB- complete-at the end-on the beach -NOW-PAST-lS.iND e. Puuyimitqa suuhaa ?u -iiyip-(m)it-qaY suuhaa

'he had got a spring salmon' REF-obtain-PAST-3.SUB spring salmon

f. Puus^initaPaaia Puus^it -(m)it -ma- =?aaia

'it was always like that' sometimes -PAST -3s.lND=always

g. q^amihsimtquus q^aa -mihsa -(m)it -quus

'if I had wanted to' thus -want to...-PAST -lS.CND

h. wahsiPaanitqa wah -siX-'at-(m)it -qa*

'she had been abandoned' abandon -MOM -PASS -PAST -3.SUB

Further allomorphy arises in the combination of the past tense with the indicative mood paradigm, which results in a further reduction, as shown in (21 d,f) and in the combination with the passive morpheme, - 'at, resulting in the portmanteau form, -?aanit, as in (21h). Rose (1981:211) notes for Kyuquot, and this may be true for both Northern and Central varieties, that there are different allomorphs of the past tense from those in the southern varieties. The past tense allomorphs -int, -nit, and -na are in complementary distribution. The latter one, -na, accompanies subordinate and irrealis mood morphemes, i.e., the possessive, relative, subordinate, dubitative, interrogative, and conditional. With other mood morphemes, -nit is chosen when the tense morpheme follows a first-syllable vowel and -int occurs elsewhere. Contrast the following sentences.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

21. a. q^isPaX wawaamitPitq 'he did this as they had said' q^is-'aX DUP-waa-(m)it-?ftq do thus-NOW PL-say -PAST -3s.REL

84 Linguistic Theory and Complex Words (74) wa-nitrs /wa*-nit-(y)i:-s/

I said it say-PAST-lNDF-l

This variation in the nasal may be due to a process of assimilation to the final l\l historically. 4.3.1.2 Future Tense There are in fact several different markers of the future, indicating various degrees of certitude and intention. Unlike the past marker, these do often occur on the verb, although in certain contexts futurity may be expressed without the presence of a future morpheme. We will examine each of them individually in what follows. 4.1 A. 1.1 Irrealis Future. The Irrealis Future, -(y)ik, is non-committal as to probability of the action actually occurring. It occurs in predictions about what may happen, as in expressing a wish. 22. a. kimssiPatikqas kirns -siX -'at -(y)ik -qa*s

'may I be bitten' bite on bait-MOM-PASS-lRR.FUT-lS.REL

b. nuWicnahPapikin 'we will have him look for his father' nuwrqsu-na'h-'ap-(y)ik-(m)in father -seek -CAUS-IRR.FUT-1 P.IND c. t'aS'eePinyapikin 'we will leave him on the rocks' t'aS'ee?inyu-'ap-(y)ik-(m)in abandon on rocks-CAUS-iRR.FUT-lP.iND d. PuuPaisimhiCiPatikqas 'may I be desired' Pu-'aisimhi-siX-'at-(y)ik-qa*s REF-lust for-MOM-PASS-iRR.FUT-ls.REL 4.3.1.2.2 Future Intentive. The Future Intentive, -?aaqX, is more indicative of one's intentions to perform some task or other and appears in examples such as the following: 23. a. huPacaCiXPaaqXma huPaca -siX -PaaqX -ma*

'he will come back' return -MOM -INTENT -3.IND

b. PuwiiPaaqXePic caxsiX 'you will be the first to spear' Pu-wii-PaaqX-(m)ePic dax-siX REF -first-lNTENT-2s.lND spear-MOM There is a strong indication of intention to perform some act, and so this future is a clearer marker of certitude.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

(75) wa*ma^qXinti#s I wanted to say it /wa*-maS'aqX-int-(y)i:-s/ say-want to..-PAST-INDF-I

Suffixation 85 4.3.1.2.3 Hypothetical Future. The hypothetical future, -'iik", carries a much higher degree of uncertainty with it and therefore, is used in those situations were such an ambiguity is desirable.

b. kupcuqsiPiikC kup-duq^a-siX-'iik-s

'when night might come' night -MOM -HYP.FUT -CND 'they'll put their index fingers into their mouths' index finger-in mouth-HYP.FUT-lNFER

As can be seen from these data, the outcome in all cases is uncertain, and may be dependent on the tense of a higher clause, suggesting some sort of tense agreement in operation. 4.3.1.3 Co-occurrence of Tenses Rose (1981:209) first notes the possibility of future and past tenses cooccurring, indicating more complex tense interpretations. When -PaaqX occurs with -int PAST, the event can occur either in the future but prior to some subsequent event (as in (60) [=25a]) or in the past but subsequent to some other related event (as in (61) [=25b]). 25. a. hininqu Bill PusiikPaaqXintin 'when Bill comes, we'll have made it' hin-in-quu Bill Pu-siik-PaaqX-int-in there -come-CND Bill REF-finish -INTENT-PAST-lp.IND

b. PuciiCXPaaqXXint t'asiis 'he was intending to go to Tahsis' ?u -cPii-siX-PaaqX-'aX-int t'asiis REF -go to-MOM-lNTENT-NOW-PAST Tahsis Contrary to Davidson (2002:305), who states 'Rose (1981) reports that =?aaqX co-occurs with the irrealis = 'aah and the past tense clitic =(m)it in Kyuquot dialect, but examples of either are lacking in the Tseshaht [sic] corpus', there are, in fact, examples of this type of co-occurrence in Tsishaath, including the following: 26. a. hiinicsuXPaaqXitPi 'the one who would have taken her away' hina-[L]-iics-'awiX-?aaqX-(m)it=?r LOC -take along[L]-MOM-INTENT -PAST=DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

24. a. PaathsiPiikquu Pathii -[L] -siX -'iik40 -quu

86 Linguistic Theory and Complex Words

Davidson (2002:321) does note the possibility of the co-occurrence of the other future tenses with the past and provides the following example from Sapir & Swadesh (1939). 27.

XiCiPiikitqa Puyi ... 'they would have shot already if ...' Xi-siX-'iik0,-(m)it-qa- Puyi shoot -MOM-HYP.FUT-PAST-3.SUB if

Thus the combination of tenses must be seen as a means to marking more complicated tense relationships, much as in English one may say 'he will have been gone for two years next Wednesday.' 4.3.1.4

Tense and Nominals in Nuuchahnulth

Now that we have examined the various tense morphemes found in Nuuchahnulth, it is time to examine a very special property of Nuuchahnulth, first discussed in the Sapir (1921) quote at the head of this section on Tense. While this property may not be unique to Nuuchahnulth,4 it is certainly rare among the languages of the world and worthy of note, if only for the contribution it makes to the topic of category neutrality, discussed in Section 3.7. The past tense morpheme, -(m)it, is commonly applied to nouns, in which case it indicates a former status, either of an object or of a person. Examples of this situation are provided in (28): 28. a. Takyakuk^itPi Takyak -uk -(m)it =?v b. PumPiiqsakitPi ?um?iiqsu-ak -(m)it =?i*

'what had been his knife' knife -POSS-PAST =DEF 'his late mother' mother -POSS-PAST=DEF

(28a) demonstrates the use of the past tense with an inanimate noun, Takyakuk^it 'what had been his knife, his former knife', and (28b) with a human being, ?um?iiqsakit 'his late mother'. There is no theory-independent evidence that these different usages motivate separate morphemes in Nuuchahnulth: they share the same allomorphs, attach in the same order, and carry, essentially, the same meaning. Thus, the tense markers should not be viewed as only a marker of tense on the verb, but rather as indicating a state, applicable to either noun or verb depending on the situation.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. qwisPaX q^iisPaaqXitPitq 'he did what he had tried to do' w <0 q is-'aX q is-[L]-?aaqX-(m)it-?rtq do thus -NOW do thus -GRAD -INTENT -PAST-3S.REL

Suffixation

87

In the following, the (a) examples provide a plain noun without tense marking and the (b) examples involve a nominal marked with past tense.

b. wikiiX maht'iimitPi wik -'i'X maht'ii -(m)it =?i*

'the house' house =DEF 'there was no longer any house left' not -find house -PAST =DEF

30. a. kaXhsiX qaqmapt qasiiPi 'the alder appeared out of his eyes' kaXhak-siX qaqmapt qasii=Pi* appear -MOM alder eye =DEF b. qasiimitPi qasii -(m)it =?f

'his former eyes' eye -PAST =DEF

31. a. wiinapuX PiihtuupPi 'the whale stopped' wiinapi -'awiX Piihtuup =?i* stop -MOM whale =DEF b. qahsiPaX PiihtuumitPi5 'now the one who had been a whale died' qahsiX-'aX Piihtuup-(m)it=Pr die[M] -NOW whale -PAST =DEF 32. a. PucaCiX PumPiiqsakPi ?u xa-siX ?um?iiqsu-'ak=?r b. PumPiiqsakitPi PumPiiqsu -'ak-(m)it=?f

'he went to his mother' REF -go to-MOM mother-POSS=DEF 'the late mother of ...' mother -POSS -PAST =DEF

The past tense may also be applied to individuals to indicate that they have passed away, often glossed as 'the late...'. Note that this applies to both native and non-native names. 33. a. yimhaaPamit yimhaaPap -(m)it

'the late Shames-Them' Shames-Them - PAST

b. mistimt mista -(m)it

'the late Mister' Mister - PAST

c. tiipanit tiipat -(m)it

'the late David' David -PAST

It is not only the past tense that appears on nominals: the future also appears, as demonstrated by the following examples. 34. a. q^ayadiikPi q^ayaciik =?iY

'the wolves' wolf =DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

29. a. mahfiiPi maht'ii =?i*

88 Linguistic Theory and Complex Words

35. a. Pustaascip miisyiPi Pust -a*sca -ip miisyi =Pir b. miisyiPaaqXPi miisyi -PaaqX =?i* 36. a. yaa saanakPi PaXa yaa saanak =?r PaXa b. saanakPaaqXPi saanak-?aaqX=Pf

'the wolves-to-be' wolf -INTENT =DEF 'he put the spear on the roof LOC -on the roof-MC spear=DEF '(the shaft of) the future spear' spear -INTENT =DEF 'the two crawl dancers' that crawl dance=DEF two 'those who are to do the crawl dance' crawl dance -INTENT =DEF

Rose (1981:250) notes that the past affix can be attached to a nominal without the concomitant attachment of ?r DEF when the nominal is uniquely identified. 37. a. muk^inintC muk^ina-int-C b. xutaaykint xutaay-'ak-int

'Maquinna (now deceased)' Maquinna-PAST-INF (Kyu) 'her old knife (now lost)' knife-POSS-PAST (Kyu)

She further observes: When -PaaqX FUT is affixed to a nominal and when either -PaaqX or -int is affixed to a non-nominal, the stem must also be affixed by -Pi* DEF if it is to be interpreted as a non-predicative NP or NP constituent. Tense is more commonly indicated in Kyuquot NP's than in English NP's although tense is never obligatorily specified in NP's. The past morpheme occurs in NP's when the referent is lost as in (228), deceased as in (227), or destroyed as in (230). [...] Evidently, in Kyuquot, the use of-int in NP's coreferential to deceased individuals is becoming optional. The future morpheme is present in NP's when the future occurrence of an entity or of a certain state of that entity is factual and subject to someone's volition or control as in (217). (ibid) It seems quite clear from the discussion above that tense-marking of nominals as well as verbs is a feature of Nuuchahnulth grammar and one that makes the language rather special typologically.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. q^ayaciikPaaqXPi q^ayaciik -PaaqX =Pi*

Suffixation 89 4.3.2 Valence-Changing Affixes The morphemes to be discussed in the next three sub-sections all have an effect on the valency of the predicate to which they are attached.

The term passive has a special status in Nuuchahnulth grammar. What was originally described as a passive morpheme by Sapir & Swadesh (1939) has more recently been argued to be a marker of switch reference. In fact, there has been much controversy over this element in the literature, with individuals arguing for a number of different analyses: see, for example, Rose & Carlson (1984), Whistler (1985), Emanatian (1988), Nakayama (1997b), Kim (2000b). While we use the term 'passive' here for want of anything better, we will present facts about the use of this morpheme without adhering to any specific, theory-internal position. The basic allomorph of the passive is -'at, and there is another allomorph conditioned by the presence of a following past tense morpheme, -(m)it, resulting in a combined, portmanteau form -'aanit. 38. a. waaPatXaa waa -'at =Xaa

finthtin frntmis-htin

'Mucus-made was told again' say -PASS=again snot -made of

b. hita^riat Taatusrhit hita -£iq-'at Taatus -rhrt

'Deer was passed by' LOC-do in passing[M]-PASS deer-Son

c. suk^iPat haawiiaXPi suu -siX-'at haawiiaX=?r

'the young man was taken hold of hold -MOM-PASS young man =DEF

39. a. waaPaanitPitq tuuyaap 'just as she had been told to throw them' waa-'at-(m)it=?ftq tu-[L]-ya*p say-PASS-PAST-3S.REL throw-MC[L] b. yaSaanitqas yaq"-'aKm)it-qa*s

'... the one by whom I was ...' REL-PASS-PAST-1 s.REL

c. wahsiPaanitqa wah-siX-'at-(m)it-qa"

'that she had been abandoned by him' abandon-MOM-PASS-PAST-3.SUB

The sentences in (38) and (39) provide examples of the distribution of the two basic allomorphs of the passive. The (a) examples demonstrate the basic form of the morpheme, the (b) examples demonstrate its glottalising nature following obstruents, and the (c) cases present a special property of certain glottalising inflectional suffixes following aspect markers with a final /X/.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.3.2.1 Passive

Linguistic Theory and Complex Words

Examples of this morpheme abound in the language, far more than one might expect for a passive marker in general, and it seems typically to act as a marker of a change in subject across clauses. In order for this to become apparent, it may be useful to examine a piece of continuous discourse where the subject referents change frequently. The following is drawn from Sapir & Swadesh (1939), Text 4, 'Kwatyat and Wolf: 40. a. PahPaaPaXwePin qahsaa^aX k°atyaat and then killed Kwatyat 'And then Kwatyatj killed Wolf].'

Puuk°ii do to

q^ayaciik. wolf

b. Puucahtaksa dussiPaX k°atyaat hiyathPitq. and then dig a hole Kwatyat where he was living 'He; dug a hole where hej was living.' c. PucaPapaX qahakaXPitq tuuk^iiCpita^aX. put into the dead one cover up in the house 'Hej put in it hiirij who was now dead, and covered himj up.' d. pawaisiPaXatwePin. they missed him 'Now theyk (the Wolf people) missed himj (Wolf).' e. naaCuksiPaXat hisiikatwePin hiyathPitq k^atyaat. look for pass by where lives Kwatyat '(One of the Wolf people)m went out to look for himj, and hej passed by where Kwatyatj was living.' f. PaPaatuuPatwePin q^aawuusi wikyuu naCuPai. ask whether not refer to... see s.o. 'Hem asked himj whether he* had not yet seen himj.' Sentence

-'at

Kwatyatj

Wolfj

40 a b c d e e f f

— — — -'at -'at -'at -'at

Subject Subject Subject

Object

Object Subject

Object Object Object

Wolvesk Searcherm

Status No Change

Subject

Object

Subject Subject Subject

Transition Transition Transition Transition Transition

Table 4.17 Use of - t o

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

90

Suffixation

91

Table 4.17 illustrates the roles of the participants in the discourse and the transitions from one to the next. It will be noted that the points of transition equate with the use of the morpheme -'at. Where there is a transition either in terms of subject or object, this marker may appear. Causative

The causative marker - 'ap, which has the phonologically conditioned allomorph, - 'am, is used to increase the valence of a verb, converting intransitives to transitives and transitives to di-transitives. 41. a. hisimyuup Pink^aPap hisimyawup Pink-'a^Pa-'ap

'they gathered wood and made a fire' assemble[MC] fire -in hearth -CAUS

b. wahaaka^aXwePin wahaak-'ap-'aX-we^Pin

'so he had her go' go -CAUS -NOW -3S.QT

d. siyaaTajMs siyaaq-'ap-'rs

'let me do it!' I-CAUS-2s>l.lMP

These examples demonstrate the valence augmenting function of the causative morpheme. As can be seen from these examples, the causative may effect a secondary causation marked by the English expression 'let ...' as well as the primary causation marked by 'make someone do something' In certain cases the causative and the passive may co-occur, as indicated below. 42. a. qaahk°a?apat qah -k°a -'ap -'at

'they were being killed off die xompletely -CAUS -PASS

b. hamafapat hamat -'ap -'at

'they knew (were made to know)' known -CAUS -PASS

c. rftaPak^aPapaXat rha -PakV -'ap -'aX -'at

'they chewed him up' bite -in pieces -CAUS -NOW -PASS

d. Panaasuk^apaXat qinaqac 'only the guts are left(by the Wolves)' Pana-Vsuk-'ap-'aX-'at qinaqac only -left over-CAUS-NOW-PASS guts Some of these examples may be argued to be lexicalised forms, but they clearly suggest that there is more to the passive and causative than simple argument modification. 4.3.2.2.1 Lexical Causatives. There are a number of morphemes that combine the causative morpheme with the momentaneous aspect, which

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.3.2.2

92 Linguistic Theory and Complex Words

43.

-ahuup -a-nuup -'aqstup -atsap -ayiiPitap -VyaYp -aaPatup xaPap -inksfitap

'in front' 'all along [L]' 'inside' 'come' 'enter the house' 'go out to sea' 'destined for [L]' 'go to' 'back of the head'

-'ipitap -'isanap -iyuqstup 1 0>

"

»

-k a-yap -qawup -qinu'p -suup -swanup -Vp

'in the house' 'on the beach' 'put in the mouth' 'completely' 'on the face' 'on top, on the head' 'at the eye [R+L]' 'in the armpit [RL]' 'on the rocks'

Examples of the use of these aspectually-marked variants of suffixes are provided below. 44. a. rhuCquupiXsi rhuC -qawup -siX -si* b. wihiisanap wihi -'isanap

'I covered my face with mine' cover -on the face[MC]-MOM-ls.ABS 'they brought him to the shore' on dry land -onto the beach [MC]

4.3.2.3 Indirective This morpheme implies action with respect to someone else, whether this has a beneficial, detrimental or neutral impact on the party. For this reason, the term indirective, first used for Nuuchahnulth in Sapir & Swadesh (1939), will be employed for this morpheme. The form of this suffix is -dip INDIR and it is used in examples such as the following: 45. a. ciPak^aPapCip £i -Pak^a -'ap -Cip

'he cut up all its insides for them' cut -in pieces -CAUS -INDIR

b. ha?uksiXCij5aX hePiisCiXCi^aX 'they began to eat his (salmon)' ha?uk-siX-Cip-'aX ha-'rs-siX-Cip-'aX eat -MOM-INDIR-NOW complete-eat-MOM-lNDlR-NOW c. haPuksiXCip Xacaqstim haPuk-siX-Cip Xac -'aqstim

eat -MOM-INDIR fat -at the abdomen

d. suk°iXCipaXwePin suu-siX-Cip-' aX-we^?in

'she took theirs' hold -MOM -INDIR-NOW-3S.QT

e. PuuPiihsiXCipaX qasiiPat Pu-'iih-siX-Cip-'aX qasii-'at

'they took out its eyes' REF-get[L]-MOM-INDIR-NOW eye-INAL

'hej ate the fat of hisj abdomen'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

produces a combined momentaneous causative [MC] version of the morphemes. Morphemes with these combinations include those listed below:

Suffixation 93 The exact mechanism of this morpheme is still not clear but it would appear to affect the predicate argument structure of predicates to which it is attached.

Nuuchahnulth exhibits the property of marking possession on the possessee rather than on the possessor, if there is an overt argument to mark, as is typical of a head-marking language. Thus: 46. a. N ha\Viiuk?i

NP q^ayaciiktaqimi

b. hawiiukPi q^ayadiiktaqimi 'chief of the Wolf Tribe' hawii-uk=?i- q^ayadiik -taqimi chief -POSS =DEF wolf -tribe Nuuchahnulth also makes use of the typologically common distinction of alienable versus inalienable possession, the former referring to objects loosely construed with the possessor, whereas the latter involves intimate, part/whole relationships, perhaps exclusively bodyparts. We will begin with the latter, more restricted domain of inalienable possession. 4.3.3.1 Inalienable Possession The first thing to notice about the marker of inalienable possession is that it has the same shape(s) as the passive marker, i.e. - 'at. This may be coincidental, but there may, in fact, be some indirect relationship between the two concepts that are, semantically, in complementary distribution. Examine the following instances of the latter: 47. a. CimcaasPatPi Cim xa*s -'at =?r

'his right hand' right -in hand -INAL =DEF

b. tiiCmaatPi tiiC -ma -'at =?f

'his heart' alive - thing -INAL =DEF

c. cicksiPaX fisiiPatPi cick -siX-'aX fisii-'at =?r

'he struck his ankle' strike w/blunt object ankle-lNAL=DEF

d. XuiukePic sifaatquuk 'it is good for a tail for you' Xuiuk -(m)e?ic sit'a-'at-qu-k good -2s.lND tail -INAL-2S.CND

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.3.3 Possession

94 Linguistic Theory and Complex Words There are no examples of this marker with any objects other than bodyparts, and even with bodyparts, it is not always necessary, if the proper conditions are met. Take, for example, the following case:6

b. hiniicsPaX fuh&takPi hina-iics-'aX t'uhciti-'ak =?iY

'hej was carrying hisj head'7 LOC-take along-NOW head-POSS=DEF

Examples such as (48a) are the normal means to indicating possession in the case of bodyparts, whereas (48b) has a special pragmatic force: it indicates that the bodypart is no longer in a close relationship with the body to which it was previously attached, i.e., it has been removed from the body, as in this extract of a story in which the protagonist cuts off the head of the Wolf Chief and brings it home to his wife. Such cases are much less common but do occur in the appropriate context. 4.3.3.2 Alienable There are two markers of alienable possession, -'ak and -uk. The two forms are, in fact, allomorphs of a single suffix, phonologically conditioned by the immediately preceding element: if this is an obstruent, then -uk will be used (49), otherwise -'ak (50), with the further proviso that the glottal stop may drop out, triggering the subsequent coalescence of vowels, as shown in (50d). 49. a. Takyakuk^itPi Tak0 -yak10 -uk -(m)it =?r

'his knife' whittle -tool -POSS -PAST =DEF

b. quuiukPi quui -uk =Pi*

'his slave' slave -POSS =DEF

c. qawassacukPi qawas -sac -uk =PiY

'her salmonberry dish' salmonberry xontainer =DEF

d. rhusukPi rhus -uk =Pi*

'his trap' trap -POSS

50. a. haPumPakPi c haPum -'ak =PiY b. TahPinPak Piihtuup Sari -'in -'ak Piihtuup

=DEF

'their food' food -POSS =DEF 'Whale's diver' dive -treated as -POSS whale

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

48. a. PuPiiPakap t'uhcitatPi 'his (spear) went right into his head' ?u -Pii-'ak-'ap fuhditi-'at =PiY REF-go to-POSS-CAUS head-iNAL=DEF

c. taayiiPakPi taayii -'ak =Pr

'his older brother' older brother -POSS =DEF

d. iuucsmaak iuucsma -'ak

'his wife' woman -POSS

One point with respect to - 'ak is important. There is a very common morph, /ak/ that appears as the free form of a number of roots in Nuuchahnulth, but that has nothing to do with the possessive marker. This is straightforwardly demonstrated by the occurrence of a number of such forms with the anticipated allomorph of the alienable possessive attached in addition to the /-ak/ morph of the root. 51. cacawakukqa nayaqakuk haCatakuk qahakuk PuuSmaqakuk

'his one ...' 'her baby' 'all of theirs' 'his dead one' 'his terrible one'

dawa nayaq haCat qah Puusmaq

'one' 'baby' 'all' dead' 'terrible'

For further discussion of the distinction between bound and free forms, refer to the discussion in Chapter 3. 4.3.4

Pluralisation

Pluralisation is a very complex issue in Nuuchahnulth grammar; it may be realised in many shapes, dependent on a variety of factors, or it may be unmarked. Here we will examine those aspects of pluralisation that are directly related to affixation. There are a number of different morphemes that mark arguments as plural in Nuuchahnulth. The most common suffix indicating pluralisation is -minh, which might be described as the basic, regular plural, as demonstrated by the examples in (52). 52. a. Pusimmah yaqsrhinh ?u-sim-(m)a*h yaqs-rhinh b. nePiiCiPaX Paamarhinh nePiiCiX -'aX Paama -rhinh

'I need long ones' REF -need -Is.IND long-PL 'they heard loons' hear [INC] -NOW loon -PL

c. cihatsiX SiyaarhinhPi 'those who were cutting were frightened' dihat-siX ci-(y)a*-rhinh=?r in state of fright-MOM cut-DUR-PL=DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Suffixation 95

96

Linguistic Theory and Complex Words

The suffix -rhinh PL denotes plurality, specifically a collective and potentially partitive plurality. ... When attached to intransitive or passive predicates, it indicates that the subject is plural and often that this plurality is new information. 53. a. tuhukrhinh tuh -uk -rhinh

'a bunch of them are afraid / there's a bunch of scared ones' afraid -DUR -PL

b. XahaXrhinh Xah -'aX -rhinh

'there's a bunch doing it now' now-NOW -PL

c. qahsaaptrhinh qah -sa*p -'at -rhinh

'a bunch of them were killed' die -MC -PASS -PL

d. citka Pucu-rfiinh citk -(y)ar PuxV -rhinh

'the (milk) inside flowed out' flow-DUR it-inside-PL (Rose 1981:240)

Note particularly the last example (53d) where the plural is applied to what would usually be expected to be a mass noun, 'milk'. Another, more idiosyncratic indicator of plural occurring on nouns is -h, which Sapir (1921:75-76) describes in the following fashion: In Nootka, to refer to but one other language in which the process is found, the t or tl of many verbal suffixes becomes hi in forms denoting repetition, e.g., hita-'ato 'to fall out,' hita-'ahl 'to keep falling out'; mat-achisht-utl 'to fly on to the water,' mat-achisht-ohl 'to keep flying on to the water.' Further, the hi of certain elements changes to a peculiar h-sound in plural forms, e.g., yak-ohl 'sore-faced,' yak-oh 'sore-faced (people).' This process is irregular enough to be considered morphological rather than phonological at this point in time: see (54a-d). 54. a. Cuucskapuuh Cuu -ska*pui -h b. huquuh huquui -h

'comrades' well -rascal! -PL[DUR] 'head-masks' mask -PL[DUR]

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Rose claims that, at least for the Kyuquot variety of Nuuchahnulth, -mlnh, has a collective force.

c. t'apqapiih fapqimi -h

'bales' bale -PL[DUR]

d. tuk^aqapiihPi tuk^aqimi -h =Pr

'seal-skin floats' float -PL[DUR] =DEF

In addition to this marker, there are also other restricted markers of the plural, including reduplication, infixation, and combinations of the two. These will be discussed in subsequent chapters. Two other markers of a plural nature remain to be discussed, - 'aqa 'several doing...' and -?ai8 Pl(ural argument). 55. a. Puuq^iyaTaqamePicuwee 'you (pi.) are having fine weather' Puuq-(y)iya-'aqa-(m)er?icuuxe good-weather-several -2p.iND-V0C b. siyaaXaqa siX-y- -'aqa

'you're all moving (house)' move house-PL- -several..ing

c. hitacsmaSaqiiCiX hita -sma(q) -'aqa -iiCiX

'they began struggling over it' LOC xontest -severaL.ing -INC

From its ability to co-occur with inflectional plural marking, as in (55b), and its position of occurrence inside of aspectual markers such as -HdiX INC in (55c), it would appear that - 'aqa is best regarded as derivational in nature. The final morpheme we will examine under the category of plural is =?ai, described by Sapir & Swadesh (1939:241) as an enclitic.9 This morpheme always appears outside of paradigmatic inflectional morphology and is best viewed as a clitic (see Section 8.1). It affects the plurality of the subject or object of the sentence, as shown in (56) below. 56. a. miisuk^aXPai mis -[L] -'aX=?ai

'they went about smelling' smell -GRAD -NOW=PL

b. Pani PeePiiktaqsPatqaPai 'that a miracle was happening to them' Pani ?ee?iikta-s-'at-qar=?ai that miracle happen-iTER[F]-PASS-3.SUB=PL c. CapxnaakukwePinPai Capx10 -na'k^-uk-we^Pin =Pai

'theirs (their sister) was married' man -have -POSS -3.QT =PL

d. qaiaatikPisukPiPai qaiaatik -Pis -uk=Pi*=?ai

'their little brother' younger brother-DlM-POSS=DEF =PL

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Suffixation 97

98 Linguistic Theory and Complex Words Note that in (56a), the plurality refers to the subject, 'they', whereas in (56b) it refers to the object, 'them'. (56c,d) show how this morpheme interacts with the possessive to produce plural possessors.

The suffix described as past quotative should perhaps more properly appear with the paradigmatic suffixes except for the fact that it is only attested in a single form, i.e. the third person form, -(i)c*a\ This should perhaps be expected, given the nature of the morpheme, which indicates what was said to be the case. Describing it as a quotative is perhaps misleading. It obviously indicates some degree of uncertainty associated with a period of time in the past beyond the recollection of the speaker. Examples include the following: 57. a. k^iik^istupsapca k^iik^istupsap-iSa

'the supposed Transformer they tell about' Transformer -PAST.QT

b. hinasiX hinaayiica nisrha 'he reached the sky-land, as it was known' hinas-§iX hinaayii-ica nisrha reach-MOM sky-PAST.QT land c. hiPiisi£a maht'ii disaa hiPiis -ica maht'ii cisaa

'that was said to have stood at Tsishaa' on the ground-PAST.QT house Tsishaa

Common translations of this suffix include: 'reputed, supposed, said to be ..., claimed to be ...'. It is usually associated with things discussed in the distant past and not known directly by the speaker. 4.3.6 Irrealis Irrealis is used to describe what is counterfactual or hypothetical, often associated with a high degree of uncertainty. The use of the suffix - 'aah clearly demonstrates this: 58. a. hisPaahPaX hitaa^atu hi -as-'aah-'aX hitaa^atu

'she could not come down' unable to-reach-lRR-NOW move down

b. hisiikPaahwePin hitaapuPasPi 'he would have gone under the log' hisiik-'aah-we^Pin hita-'apuX-'as=?r go -irr-3s.qt LOC-underneath [L] [M]-on the ground=DEF c. hayimhiCiX q^isPaahi 'he did not know what to do' hayimhi -siX q^is -'aah -(y)ii forget-MOM do thus-lRR-3.lNDF.REL

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.3.5 Past Quotative

Suffixation

99

As can be seen from the data above, the irrealis typically occurs in situations where there is a lack of ability or knowledge on the part of the participant and so is common in expressions involving such inability. Simultaneous/Meanwhile

The suffix which is described as simultaneous, often glossed as 'meanwhile', has the shape -(q)h MW, the Iql appearing between preceding sonorants and certain other consonants and the suffix. It occurs frequently in texts and indicates a certain degree of contemporaneity or simultaneity in the actions performed within the clause, as demonstrated in the examples below: 59. a. TiihsiX hitinqish Tiih-siX hita-niq-'is-(q)h b. hiih Taatusap hii -(q)h 'Taatusap

'she burst into tears down at the beach' cry-MOM LOC-down slope-on beach-MW 'While he was there at Deer-up ...' LOC -MW Deer-up

c. qahakhPaX naPaa 'it was while he was dead that he heard...' qahak -(q)h -'aX naPaa dead -MW -NOW hear With respect to this morpheme, Rose (1981:151) observes: The temporal affix is attached to the predicate of the clause which is the 'ground' during which the event expressed in the other clause takes place. 60. a. ciiqciiq : waPiiCh 'he was talking in his sleep' DUPxiq-[L]-(y)a* wa?iC-(-q)h REP- talk -CRAD sleep-MW (Kyu) b. waiyuuqh : waPiC wai-yuY-(q)h waPiC

'while he was at home, he slept' go home-...-ed -MW sleep (Kyu)

Rose further notes (ibid: 122) that '[w]hen an event occurs while participants are at a particular place, hii is affixed by -(q)h SIM ....' Her examples follow in (61a,b). 61. a. hiihnah hii -(q)h -na -ha* b. hiinah hii -na -ha'

'Did he do it there?' there-MW-PAST-lNTER (Kyu) 'Was he there?' there -PAST-INTER (Kyu)

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.3.7

100 Linguistic Theory and Complex Words

So -(q)h is employed in general to indicate an event co-occurring with some other event in the discourse.

Nuuchahnulth makes use of both diminutive and augmentative markers, the former particularly frequently. The diminutive suffix takes the form -?is DIM, and may occur on nouns, verbs or other parts of speech as demonstrated in (62). 62. a. faatnePisukPi fafta -t- -Pis -uk =Pi*

'her little children' child -PL- -DIM -POSS =DEF

b. nuunuksiPaX nuukPisukPi 'he started to sing his little song' DUP-nuu-siX-'aX nuuk-?is-uk=?r REP-sing-MOM-NOWsong-DlM-POSS=DEF c. PapaasPisPi quuPas Papa-s -Pis =?i* quuPas

'the little person' small-DiM=DEF person

d. t'aniiCiXPicaX t'ana -iiCiX -Pis -'aX

'he became a little child' become a child-DlM -NOW

An interesting possibility with this suffix in particular is a sort of 'spreading' or 'harmony' that may occur throughout the sentence, perhaps indicating scope over an entire clause by marking each member of the clause within the domain. Examples are in (63). 63. a. PunaakPiswePin t'anePis 'she had a little child' ?u -na'k^-Pis-we^Pin t'ana-Pis REF -have-DlM-3s.QT child -DIM b. cawaakPicaXquuwePin10

± _

: Y

saa&nkPis

A

frhakPis

finiiXPisPi

_A

± (

cawa'-?is-'aX-quu-we Pin saaSink°-Pis Sihak-Pis TiniiX-?is=?f one-DlM-NOW-3.CND-3.QT always-DIM cry -DIM dog-DIM=DEF 'now there was one little dog who was always crying' Examples like this are not common, but are found to occur as part of what Sapir has described as 'abnormal speech'.11 In representing children, the morpheme -?is DIM is inserted to convey the sense of size and endearment. The site for insertion is immediately after the particular word's aspect marker and before any inflectional suffixes. Examples of the use of this marker follow: 64. a. q^isPisCi q^is -Pis -Ci

'do so, little one!' do thus -DIM -'go and...'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.3.8 Diminutive

Suffixation 101 'is that you, Deer?' is -DIM -2s.lNTER deer -Son12

The exact nature of this suffix is still unclear and further work must be done to clarify the extent of its use. Notice that the diminutive -?is may appears on virtually any word category within the sentence, including the verb, the modifiers, the head noun, and even adverbials. 4.3.9 Augmentative The augmentative suffix, -aq(aq) AUG, is also found to occur as a modifier with different classes of words. Below it co-occurs with both a verb (65a) and a noun (65b). 65. a. XacaqaqwePinfiaTasqoo Xac -aqaq -we^incaTaskfaa b. XiiPiiaTaX XiiPii -aqaq-'aX

'it turned out that it was very fat' fat-AUG-3s.QT 'to give a big feast' feast -AUG -NOW

However, its principal duty appears to be as an intensifier of adjectives, as in the cases below: 66. a. Pihaqak Piih01 -aq -ak b. XuiaqwePin haawiiaX Y Xui -aq -we ?in haawiiaX

'it was very big' big -AUG 'he was a very handsome youth' attractive -AUG-3s.QT young man

Note that there are two, apparently free, variants of this suffix, -aq and -aqaq. Furthermore, this morpheme typically triggers shortening of the root vowel, a process that is discussed in more detail in Chapter 7. This morpheme, as with the previous one, may occur in abnormal speech to represent the speech of unusually large individuals (Sapir 1915). 4.3.10 Actualis The term actualis is used here to refer to a frequently occurring suffix, -'aX, glossed as NOW, which is found to occur on the majority of predicates in Nuuchahnulth sentences. Rose (1981:152) states that '[d]uration which includes the present time is expressed by a temporal predicate affixed by - 'ak TEM "at the time".' 67. a. naacsaaX iuucsmePi naacsa-'aX iuucsma=?f

'now the woman saw' see -NOW woman=DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. PuhPishak Taatusrhit ?uh -Pis -ha*k Saatus-rhrt

b. iicwisa nayaqnaksiPaX iicwisa nayaqnak-§iX-'aX

'she became pregnant and gave birth' pregnant give birth -MOM-NOW

c. hamatsa^aX hamat -sarp -'aX

'she found out that...' know -MC -NOW that

Pani ... Pani ...

All of these examples share the common property of indicating a time concurrent within the discourse, whether past or present, rather than an account of events at other stages. 4.3.11 Specificity Here we will discuss ?i\ a frequently occurring morpheme in the language. In Chapter 8 we will discuss its status as a clitic, indicated by the use of the '=' boundary mark in this book, and in this section focus on its usage only. What has been referred to as a definiteness marker, ?r DEF, is perhaps better viewed as a specificity or particularising marker. For instance, this marker may co-occur with personal names and other definite nouns that should not require it. While Rose (1981:251) states that proper names are not affixed by ?i\ and this is the usual case, Sapir (ms) notes that certain contexts allow for the possibility. Note the examples below. 68. a. waaPaX PiiSaaPapsiiiPi waa-'aX Pii£aa?apsiii=Pr

'... said Always-lifts-up' say-NOW Always-lifts-up=DEF

b. hiiuk°iiC t'asii kapCukPi 'where the road to Kapchuk was' hii-uk-yiiC t'asii kapCuk=?r LOC-POSS-3.INDF.REL road Kapchuk=DEF maawicinPi 'Maawitsin had no one to fear' c. hiihityakaah DUP-hi-ityak-'aah maawicin=?r SUF-unable to-fear[R+S]-lRR Maawitsin=DEF Contrast the forms above with ?r with the basic forms in (69). 69. a. naCuPaiPaX PiicaaPapsiii naCuPai-'aX ?ii£aa?apsiii

'Al way s-lifts-up saw something' see -NOW Always-lifts-up

b. iuCsiXquu Puuk°ii kapCuk 'he gave a woman as ransom to Kapchuk' iuucsma-siX-quu ?uuk°ii kapCuk woman-MOM-3.CND to Kapchuk d. tuuhuk^aXatqa maawicin 'they feared Maawitsin' tuuhuk-'aX-'at-qa* maawicin afraid-N0W-PASS-3.SUB Maawitsin

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

102 Linguistic Theory and Complex Words

Suffixation 103

NP's are obligatorily specified by -?r if a tense morpheme is present (unless an otherwise particularized nominal is present)... if a nominal head has an object ... or if a multimorphemic or phrasal modifier is present .... Even bimorphemic modifiers require the presence of -?r if the affix in the stem is derivational but not aspectual. Finally, the NP is obligatorily specified for particularity if it includes a temporal, locative, or verbal stem not following the nominal head. 70. a. taaksXi huhuuniih tak-[L] -siX=?i* huuni -'iih b. t'iTasi quuPas ticf -'as=?i* quuPas

'the ones always out whaling'(Kyu) each-GRAD-MOM=DEF whale-pursue[R] 'a/the person sitting outside' sit-outside=DEF adult (Kyu)

Note that there is no specific indefinite marker in Nuuchahnulth. 4.3.12 Perspective The following morphemes provide different perspectives on the predicate to which they are attached, conveying notions of possibility, pretendedness, constancy and other notions. 4.3.12.1

Probability/Preferability

This suffix, -cuu PROB, indicates a degree of probability of success or of preferability of the action represented by the predicate, as shown in (71). 71. a. q^ishcuuyiisuu qahsaap 'how you may succeed in killing (me)' q<0is-(q)hxuu-(y)iisuu qahsaap thus-MW-PROB -2p.lNDF kill me b. jiajMicuuPatePic jJafjr xuu -'at -(m)e'Pic

'this is better fit for your ears' ear -PROB -INAL-2S.IND

c. hinsiiyapcuumin hina -as -iiyap xuu -(m)in

'we had better start bringing her to' LOC -reach-lNC.CAUS-PROB-lp.lND

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

As can be seen from these and numerous other examples, ?r may be attached to proper names, raising the question of the exact nature of this morpheme. On the topic of attachment of ?r, Rose (ibid) also notes that:

104 Linguistic Theory and Complex Words 4.3.12.2 Pretendedly

72. a. haPukqathPa^atwePin 'it was believed they were eating them' haPuk -qa'th -'ap-'at-we"Pin eat -PRETEND -CAUS -PASS -3S.QT b. PuhqaathPaXukwePin iuucsma 'he thought it was his wife' Puh-qarth-'aX-uk-we*Pin iuucsma is-PRETEND-NOW-POSS-3s.QT wife c. wiPaksiX wi?ak-siX -qa'th

'he pretended to get angry' angry -MOM-PRETEND

4.3.12.3 Supposedly The suffix -matak SUP means 'probably' or 'supposedly' and is used to indicate a particular point of view, imparting a certain degree of probability to the predicate. 73. a. qahsaapmatakqa qah -sa*p -matak -qaY b. nasaakmatakaXhsuk nasaak -matak -'aX -hsuk

'he must have killed him' dead -MC -SUP -3.SUB 'I think you are indeed glad' glad -SUP -NOW -MW -2s.ABS

c. siqaCiXmatakaXPitq Puyi 'when they thought it was cooked' siqa-siX-matak-'aX-Pi'tq Puyi cook -MOM-SUP-NOW-3s.REL when 4.3.12.4 Constantly The suffix -maqsiX CONST imposes a temporal connotation on the stems indicating a constant application of the predicate, as demonstrated below. 74. a. hawiiqXmaqsiX havViiqX -maqsiX

'remaining always hungry' hungry -CONST

b. haPuqhmaqsiX Xiicuu 'they would give feasts one after the other' ha?u-(q)h-maqsiX Xiiduu do in turn-MW-CONST give feast 4.3.12.5 Just/Only/Precisely The suffix -sasa JUST, with its allomorph -sa, modifies the base by indicating a certain degree of commitment to the action. Examples of this are provided below.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

This is another non-paradigmatic inflectional suffix that affects the modality of the stem to which it is attached, applying a sense of pretending or mistaken assumption to the whole. Its form is -qarth and examples of its usage include:

Suffixation 105

75. a. PuuhsasaXePic suu 'you will just keep them' Puh-[L]-sasa-'aX-(m)e-Pic suu is -JUST[L]-NOW-2s.lND hold

c. PeePimstipsa kicinksiX 'just as soon as he closed his eyes...' PePim-[L]-tiYp-sasa kicink-siX as soon as-do to while- JUST[L] close eyes-MOM d. Siiscyaksama Sis -[L]xyak-sasa -ma*

'he is just nasty' dirty-appear like...[L]- JUST[L]-3s.lND

4.3.13 Order of Inflectional Suffixes The following table outlines the order of inflectional suffixes which, unlike derivational ones, is rigid. From this table we can see that, with the exception of the conditional, quotative, and inferential, the paradigmatic suffixes are mutually exclusive, whereas the non-paradigmatic ones may co-occur in many combinations, in the order specified.

1

NON-PARADIGMATIC 4 7 2 6 5 3

qh PaaqX 'ap rhinh Pis ayap

'aX

uk ak 'aah

8

'at

PARADIGMATIC 9 11 10 CND

(m)it

QT

INFER { ^ a U j

INDICATIVE RELATIVE SUBORDINATE DUBITATIVE RELATIVE DUBITATIVE ABSOLUTIVE INTERROGATIVE PURPOSIVE INDEFINITE RELATIVE PRESENT IMPERATIVE FUTURE IMPERATIVE

Table 4.18 Order of Inflectional Suffixes Nevertheless, it should be noted that there are clearly co-occurrence restrictions on certain positions within the order. For instance, -?aaqX and -(m)it occupy two non-adjacent positions within the table and yet never co-occur. It should be noted that derivational and aspectual suffixes will occur inside of inflectional suffixes and clitics will occur outside them.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. waawaitipsatwePin 'just as he said this, ...' DUP- waa -ti'p-sasa-'at-we^Pin REP-say-do to while- JUST[L]-PASS-3.QT

106 Linguistic Theory and Complex Words

Aspect

Comrie (1981:3) defines aspect in the following way: 'Aspects are different ways of viewing the internal temporal constituency of a situation.' Aspect allows us to divide our actions into categories according to the nature of the action, whether it be a single event, a continuous activity or a repeated action. It is a very important part of Nuuchahnulth grammar and all predications must bear some aspect. The most common means to this is by suffixation and there are a number of aspect markers that involve this kind of realisation. In what follows, focus will be placed on those aspectual categories that are realised by suffixation, leaving the discussion of the others to Chapter 5 on reduplication. 4.4.1 Aspect Categories in Nuuchahnulth The main division in Nuuchahnulth aspect is between the two categories of perfective and imperfective. This is a common dichotomy universally and Nuuchahnulth is no exception here. Within each of these categories are a number of derived categories, dependent on the basic aspects but providing further refinements of an aspectual nature. Sapir and Swadesh (1939:241) provide the following list of aspectual distinctions in Nuuchahnulth, based on the verb mitx°- 'to turn, make a circuit': A. 1. 2. 3. 4. 5. 6. 7.

Durative mitx^aa 'turning about, circling' Inceptive mitxcoiiCi(X) 'to start turning about' Graduated Inceptive miitx<0iCi(X) 'starting to turn about' Pre-inceptive miitx(0iCiXsi(X) 'to start starting to turn about' Inceptive iterative miitx^iCiii 'to start turning about at intervals' Repetitive miitxmiitx^a 'turning round & round'(emphasized repetition) Repetitive inceptive miitxmiitxsi(X) 'to start turning round & round'

B. 8. 9. 10. 11. 12.

Momentaneous mitxsi(X) 'to make a circuit, turn' Graduative miitxsi(X) 'making a circuit, turn' Pre-graduative miitxsiXsi(X) 'to start making a circuit, turn' Iterative mitxmitxs 'to make a circuit, turn at intervals' Iterative Inceptive mitxmitxssi(X) 'to start in on a spaced series of circuits, turns'

On the opposite page, Table 19 summarises these aspectual categories, divided into structural elements found in various forms.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.4

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect 2011-03-11

Structural Analysis

Base

Base


Base Base

-iiCiX -iiCiX

[L] -siX [L]

Base

-iCiii

[L]

Base

-(y)a^

[L]

Base

-siX

M

Base

-siX

Base Base

-siX

Base

-s

Base

-s

-siX

[L] -siX [L]

-siX

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

108 Linguistic Theory and Complex Words

There is still much work to be done in working out the intricacies of the relationships between the parts of the composite aspects but this research goes beyond the scope of this book and will have to wait for further studies in Nuuchahnulth aspect.

In addition to those aspectual categories presented in Sapir & Swadesh (1939), there are several other combinations found in the texts. While the exact nature of these aspectual distinctions is not completely clear at this time, instances may arise, and will be noted, in the data as they occur. 4.4.3 Portmanteau Categories Various of the aspectual markers discussed in the preceding sections may be regarded as portmanteau elements. For example, the inceptive iterative -idiii, may be viewed as a composition of the inceptive aspect marker -HdiX with the iterative marker -siii. 4.4.3.1 Momentaneous Causative The momentaneous causative is a portmanteau morph combining the duties of the momentaneous aspect with those of the causative to render a combined form. The various forms of this morpheme include -iyap, -sap, -saYq, -'up, and -yap. The exact nature of the relationship between the various allomorphs is, as yet, unclear. Examples of these forms include: 76. a. CaCimhiyap CaCimhi -yap

caPak caPak

b. Puuxsaqyak Puux -saq -yak10 c. hisimyuup his -imi -'uup

'he fixed up a river' fixed up -MC [L] river 'tool for felling trees' topple over -L -MC -tool

hawiiPi ha\Vii =PiY

'then the chief called them together' all -in group -MC chief =DEF

As can be seen, these various morphs all serve to mark the confluence of two morphemes, the momentaneous aspect and the causative mood. 4.4.3.2 Graduative Causative The Graduative Causative is a portmanteau morph, consisting of a combination of the graduative aspect morpheme and the causative morpheme. It may take the following shapes: -'ayap, -sap, or -inup. Examples of the usage of these suffixes include (77a-c).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.4.2 Other Categories

Suffixation 109 77. a. Xuyaya^aX muciCupaX Xui-'ayap-'aX muSiC-'u'p

'they dressed her nicely' attractive-GRD.CAUS-NOW clothe-MC

c. taascipakXinupat 'he was rubbed over his hams' tas-[L]xi-pi*-'akXi-inup-'at rub on-at crotch[L]-on the back-at rear -GRD.CAUS [L+S] -INAL 4.4.3.3 Inceptive Causative Inceptive causative combines the functions of the inceptive morpheme with that of the causative morpheme, rendering something meaning 'begin to cause ...'. The form of this morpheme is -iiyap, and it is used in the following situations: 78. a. siqiiyap saacinPisPi 'she roasted the little spring salmon' siq -iiyap saacin -Pis =?i* cook-lNC.CAUS spring salmon-DlM=DEF b. hinsiiyapcuumin hina -wisa-iiyapxuu-(m)in c. PinxiiyapsiPaX Pinxaa -ii^ap-siX-'aX

'we had better start bringing her to' LOC-conscious-lNC.CAUS-PROB-lp.lND 'we started making ready' ready -INC.CAUS-MOM -NOW

As one can see from these examples, this portmanteau morpheme combines the functions of two separate morphemes, the inceptive and the causative. 4.5

Derivation

Derivational suffixes in Nuuchahnulth have, in the past, been divided into two groups based on whether or not they may change the category of the stem to which they attach. Governing suffixes behave as heads would, changing the category of whatever they attach to, while restrictive suffixes simply modify the stem to which they attach. Note the following quote from Sapir & Swadesh (1939: 236): Derivational suffixes fall into two important classes, according to their semantic relation to the underlying theme:

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. hamatsap q^iq^iiyuk^iis 'find out why I am crying' co hamat-sap DUP-q i(q)-ayuk-(y)iis know-GRD.CAUS SUF- that which xry for[R+L] -ls.IND.REL

110 Linguistic Theory and Complex Words

2. restrictive suffixes, which introduce a notion that is semantically subordinate to or coordinate with that of the underlying theme, e.g. Pih 'large' + -qo'Pa 'on a rocky point' > PrhqoPa 'large on a rocky point'. In the following sections we will forego this and examine the suffixes on the basis of the category to which they belong, determining what categories they attach to and what categories they derive. We will assume that headedness is an important feature of some of the bound morphology discussed below. 4.5.1 Concomitants of Derivational Suffixation Before examining the various derivational suffixes of Nuuchahnulth, it may be worthwhile to consider some of the morphophonological properties demonstrated by various types of suffixes of this class. Suffixes in this group, more than any other, tend to interact with the root to which they are directly or indirectly attached. The term 'root' is used, because it is always this part of the word that undergoes the various changes to be discussed, rather than subsequent elements such as affixes. It is never difficult to locate the root, since it will always be the leftmost element of the word, other than reduplicative copies of the root itself. The kinds of changes that may occur in this situation are several: vowels may be lengthened, copies may be made, or consonants may be inserted. In what follows, each one of these will be examined along with examples of the types of changes. 4.5.1.1

Vowel Length

This is by far the most common concomitant of affixation, occurring with eighty-odd derivational suffixes of various types. The attachment of one of these suffixes, either directly to the root, as in (79) or to a stem already containing several suffixes, as in (80), will automatically trigger the lengthening of the first vowel of the root, as shown in the examples below, where the (a) form is without the suffix, and the (b) form occurs with it. The symbol [L] indicates this lengthening property of the suffix.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

1. governing suffixes, which introduce a new central notion to which the underlying stem or theme becomes subsidiary, e.g. £apac 'canoe' + -o?ai 'see, perceive...' > 6apaco?ai 'see a canoe';

Suffixation 111 79. -mafiik

'one skilled in ...' [L]

a. nayaqak

'baby'

b. naayaqmaTukrhinhPi nayaq -maSuk -minh=?f

'the maternity doctors'

'move about' [L]

a. nupiriaXwePin nup -|5it -9aX -we'Pin

'it made a sound once' one -time(s) -make a sound -3s.QT b. nuupitpanaCwit'as 'they intended to make the rounds in one trip' nup -jMt -panaC -wit'as one -time(s) -move about [L] -go to As can be seen from these examples, the attachment of suffixes such as these demands that the vowel of the root be lengthened, unlike the attachment of other suffixes, such as -pit 'times' in (80a) which carries no such requirement. Naturally, should the root already contain a long vowel, there will be no additional length required, suggesting that this is a constraint on the length of the root vowel rather than a rule adding weight blindly. 4.5.1.2

Reduplication

The next most common operation associated with certain suffixes is that of reduplication, associated with around 60 different suffixes. Reduplication is encoded by means of the symbol [R]. 81. -Kuk

v

a. hawiiiicPi hawii -iic =Pr

'resemble' [R] fana fana

b. hahawiikuksiX DUP- hawii -kuk -siX 82. -Auk

LOC

'a chiefs child' chief -belong to -DEF child 'come to look like a chief SUF- chief -resemble [R] -MOM 'at or in the hand' [R]

a. caxyak cax10 -yak*0

'a spear' spear -tool

b. cacaxyak^aqXnuk DUP- caxw -yak" -'aqX -nuk

'each holding a spear in hand' SUF- spear -tool-inside-in the hand[R]

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

80. -panaC

baby -one skilled in [L] -PL -DEF

112 Linguistic Theory and Complex Words Again, the suffix may be adjacent (81) or at a distance from the root (82) with no difference in the application of the process. The topic of reduplication will be the subject of Chapter 5.

As can be seen from the examples below, reduplication may combine with lengthening, of either the original root vowel, or of the copy vowel, resulting in a number of possible shapes of stems. These are encoded as [RL], lengthening of the reduplicated copy or [R+L], lengthening of the original root vowel. 83. -mii?ak

84.

v

'fear [RL]'

a. wik

'not'

b. wiiwikmiiPakh DUP- wik -miiPak -(q)h

'having no fear that...' DUP- not -fear [RL] -MW

-ayuk

v

' cry for [R+L]'

a. PuciiihPaaqXwePincuu ?u -ci'ih-PaaqX-we'Pincuu

'you are to light fire with this' REF -use as fuel -INTENT -2p.QT

b. PuPuuyuk^aaqXwePincuu 'you are to weep for ...' DUP- ?u-ayuk-?aaqX-we-?incuu SUF- REFxry for[R+L]-lNTENT-2p.QT There are no suffixes that require, in and of themselves, that both the root and the copy be lengthened simultaneously. 4.5.1.4 Consonant Insertion Another operation which appears in conjunction with reduplication is the insertion of particular consonants associated with certain derivational suffixes, for instance, Id in the suffix -sawiX 'in or at the eye'. These kinds of operation are demonstrated in the following examples: 85.

-sawiX

LOC

'in or at the eye [M] [Re]'

PicPinksawiPaX maamaati 'the birds were blinded by the fires' DUP-Pink°-sawiX-'aX maamaati SUF- fire-in eye[MC][Rc+L] bird Modifications of this sort are encoded on the suffix by means of a diacritic such as [Re], indicating reduplication with accompanying Id, which always appears on the copy rather than the original root.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.5.1.3 Combinations offLJ and [R]

Suffixation 113 4.5.1.5

Combinations of [L] and Consonant

Insertion

86.

-supfaai

'compete [RLC]'

PuucPuksupfaaiPaX

'each tries to be the first to...'

DUP- [L] Pu -supt'aai -'aX

SUF- REF -compete [RLC] -NOW

-l?sa-

87.

v

LOC

PuucPuumahsaqhPi DUP-[L] Pu -ma'-hsa-(q)h=Pi-

'at the brink [RLC+L]' 'sit at the very edge of the bluff SUF- REF-as far-at brink[RLc]-MW=DEF

As can be seen in this example, all of these various modifications may interact in the derivation of a stem in Nuuchahnulth. 4.5.1.6

Vowel

Shortening

One final operation cited in Sapir & Swadesh (1939) is that of vowel shortening. This is a very rare operation involving only a very few suffixes and the data are not particularly clear on this issue (cf. Kim 2003c). Nevertheless, the following examples that demonstrate this process are provided. 88.

-ir^ak

v

a. PuuPutyak DUP- [L] Pu -ityak

'fear [R+s]' 'for fear that they ...' SUF- REF -fear [R+s]

b. Pah q^aaq^atyaksiXah. 'that is what I am scared o f Pah DUP-qtoaa-ityak-siX-(m)a'h this SUF-thus-fear[R+S]-MOM-lS.lND Note that [S] is the indicator of shortening in these cases, and that in both of these cases there is a combination of lengthening and shortening taking place. We are now ready to move on to a description of the various categories of suffixes that appear in the derivational component of the grammar of Nuuchahnulth. The reader should bear in mind the indicators and issues that have been discussed in the previous sections when reading what follows.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Rounding out the characterisation of stem modification, a further possibility is that of vowel lengthening combining with consonant insertion and reduplication, creating a stem with all of these properties, as in the case of -supt'aai 'compete in' or -hsariai the brink', presented below.

114 Linguistic Theory and Complex Words

We have already discussed the distinction to be made between bound verbs and verbal suffixes in Section 3.4. Given that we do not consider bound verbs to be suffixes, we will focus here on those morphemes that we do consider to be verbal suffixes. One possible area of contention in the distinction between bound verbs and verbal suffixes is the status of intransitive verbs, since they do not normally incorporate, there being no object to incorporate. However, it is possible for certain unaccusative verbs, intransitive in form, to display incorporation, as shown in the examples below: 89. a. PayasuuX Paya -sawiX

namintPath namint -'ath

b. haayiiPiX qiCinstaqimi ha -ayiiPiX qiCin -taqimi

'many Namint people died' many xlie[M] Namint -tribe 'all the Louse people entered the house' all -enter house louse -tribe

Incorporation will be discussed more fully in Chapter 8, but the examples above demonstrate that even intransitive verbal morphemes may be demonstrated to be bound verbs. The verbal suffixes constitute a fairly small class of derivational suffixes in Nuuchahnulth. These suffixes typically have the force of full verbs in other languages, but behave more as modifiers of the stem than as heads. Some examples of this behaviour appear below: 90. a. cimiiqstuPaX ci -mai -'aqstuX -'aX

'he started to cut inside of it' cut -move about -insidefM] -NOW

b. huuiaqtas hasmatqdus kiikiixPinPaqsup hui -[L] -aYqtu* -'as hasmatqdus DUP- kiixPin -'aqsup dance -GRAD-go across-on ground hill PL- Kiihin -woman 'the Kiihin women went dancing across the hill c. hiihPiitq qiiXaakapi hii -(q)h -Prtq qiXa'q -api

'where he stood fishing' LOC -MW -3S.REL fish -stand [L]

As one can see from these examples verb suffixes such as -mai 'move about' in (90a), -a-qtw 'go across' in (90b), and -api 'stand' in (90c) provide further information, modifying an existing predicate, but do not act as the head of the construction. In (90a), the focus is on the cutting up of the insides of a monster, while moving about, and not the reverse. In (90b) what is important is that the women were dancing, the verbal

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.5.2 Verbal Suffixes

suffix, -arqtu\ simply providing additional information concerning the direction of the dancing. And in (90c), it is the fishing which is central, -api simply clarifying that the protagonist was not sitting or in any other position while performing this action. Among the verbal derivational suffixes there are numerous combinations of verbal and locative senses, mostly implying location of an event. They are often phonologically opaque, but semantically transparent. Take the example of -mai 'move about' introduced in (90a) above. It has the following portmanteau forms: 91.

-mi*s -ma*s -mri

'move about on the beach' 'move about in the village' 'move about in the house'

The examples in (91) above may be reduced to a combination of the verbal suffix -mai 'move about' and various locative suffixes setting the scene of the movement, thus: 92.

-mi's < -mai + -'is -ma*s < -mai + -'as -mi'i < -mai + -'ii

'on the beach' 'in the village' 'in the house'

The merger of the two morphemes creates a variable vowel in all these cases. It should be mentioned here that the class of bound verbs is large and covers all the semantic fields of verbs in any language. This is discussed more fully in Section 3.4. 4.5.3 Nominal Suffixes Nominal suffixes include suffixes creating agentive, instrumental and locational nouns, as well as those implying temporal and attributive notions. Agentives indicate human agents of whatever action they are attached to, typically a verbal base, as -mafuk 'one skilled in, maker [L]' or -miika> 'getter of [ L ] \ The differences between some of these suffixes, e.g. -hsr and -vta, both glossed as '...-er', remain unclear and may be quite idiosyncratic. Instrumentals constitute a class of nominalising suffixes that create instrument-type nouns, such as -dak" 'instrument,' -ac 'receptacle,' -fiafu 'instrument for' or -yak* 'tool.' Locational nouns create nouns referring to localities and are often combined with locative roots to produce these results. Examples include -adus 'surface,' -dus

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Suffixation 115

'bay,'and -nit 'place of....' Temporal noun suffixes are utilised to create nouns referring to specific periods of time, such as -pai 'season of....' Nominalising suffixes create nouns with a specific semantic contribution. They are treated as nouns and may be pluralised, etc. Examples include -a?aq 'hide, skin,' -ck^r 'remains of,' -'in 'sound of [R],' -mapt 'plant, bush, tree,' or -ma 'thing.' The final group of nominalising suffixes are the attributive nominalising suffixes. There are a number of suffixes that serve attributive nominal functions in Nuuchahnulth. These describe membership in some group, status within a group, serve as an indicator of some naming function or some familial relationship. Examples of this category include -'aqsup 'woman,' -(q)a's 'daughter,' -'ath 'tribe,' or -rqsu 'relative'. 4.5.4 Adjectival Suffixes There are approximately 45 adjectival suffixes, covering a variety of contexts. There are no adjectives referring to colour, shape, or more specific categories. In general, these suffixes are quite generic in nature, but may be said to cover the general categories of size, such as -aqaq 'big,' -arqdu 'wide,' -k*in 'toy [RL],' physical qualities, such as -daq 'sore[R],' -htin 'made of,' -pwqs 'smelling of,' -st'w 'hidden,' -PiC 'clothed in,' -k"a'i 'absent, missing,' -yuk 'wrapped in, covered by,' -nit 'stocked with,' emotional qualities such as -daat 'sorry for,' -inqX 'inimical to [L],' -?ai 'aware of,' inherent qualities such as -'a-csyi 'male,' -'aqX 'gifted in,' and a few others. 4.5.5 Adverbial Suffixes There are around 40 adverbial suffixes in the data. These involve the modification of predications, typically verbal, by the attachment of a suffix to the predicate. Adverbial suffixes may be: (i) temporal, such as -(y)iya 'at...time', -piiZiX 'in...season'; (ii) manner, such as -?akmar 'in pieces,' -ckin 'slightly'; (iii) quantity, such as -api 'too, excessively,' -aqaq 'very extremely' and some others. 4.5.6

Classifiers

Classifiers are employed in Nuuchahnulth in much the same way as they are in many other languages: they co-occur with numerals and other quantifier expressions. As with most other languages with classifiers, there is a certain degree of arbitrariness in the determination of which classifier occurs with which referents.13 The most common are:

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

116 Linguistic Theory and Complex Words

-&q -Cri -6a -fctaTf -'ixh -imt -ista

93.

-i>rr -#**" -£inq -#t -,Mt

'long objects' 'days' 'size' 'containers ' 'year, season' 'group, chunk' 'persons in canoe' 'long, thin objects' 'songs' 'times around' 'handspans ' 'times

Xjimi Xjimiiya

-sa'htak10 -sa'th -taq -taqak -taqimi -fim -ya*sca -yai -yayisu

'unit, chunk' 'month' 'kind(s), varieties' 'tribes' 'times' 'units' 'bundle' 'at a time' 'rolls' 'fathoms' 'fathoms'

Classifiers are associated with specific nouns in a regular, but arbitrary, fashion, as demonstrated in the examples in (94-5).l4 94. a. nupciqPi nup-diq =?f b. muuciq muu-ciq

miTaat mffaat hiinaanuhsim hiinaanuhsim

c. PaXciqPisuk ciihati PaX xiq -Pis -uk ciihati 95. a. PaXjMiiPi ficsyin PaX -jM'i10 =?i' Ticsyin

'one sockeye salmon' one-long object=DEF sockeye salmon 'four whaling spears' four -long object whaling spear 'his two small arrows' two -long object-DlM-POSS arrow 'the two spines' two - thin long objs=DEF spine

b. muupiiiuk Xama muu-pri10 -uk Xama

'his four house posts' four - long thin objs-NOM house post

c. PaXjMiiPi saasaqi PaXa -jM'i01 =?'v saasaqi

'the two cannons' two - long thin objs=DEF cannon

The usage of classifiers will be discussed in greater detail in Chapter 8. 4.5.7

Locatives

Locatives may be viewed in general as restrictive modifiers, indicating the location of an event or action. In Nuuchahnulth, locatives may be sub-divided into a number of distinct categories, including geographical, orientative, directional, locational, and somatic suffixes, the last constituting by far the largest class.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Suffixation 117

In general, most main predicates are marked for location, serving to situate the action at some place. The three basic locative suffixes are -'is 'on the beach,' -'as 'on the ground/in the village,' and -'ii 'in the house'. Directional suffixes imply movement in some direction, as opposed to a static location, for instance -aYAawii 'up the river,' -sti'i 'into the throat,' -htaYkw'out to sea,' etc. Somatic suffixes refer to body parts, and as such are typically inalienably possessed. There are several dozen different somatic suffixes, covering most overt bodyparts. There are no specific suffixes for internal organs or the like, nor for certain bodyparts, including kneecap, nipple, umbilicus, etc., which are specified by either roots, or complex stems. There is a fairly large range however, including those referring to the abdomen (-V<*i 'in the belly'), the limbs (-?ardui 'on the palm,' -s&inX 'in the armpit [M]'), the back (-afai ^behind, on the upper back'), the front torso (-ashwi 'on the chest,' -V£/ 'in the crotch'), and the head (-inkst'as 'back of the head', -suk 'at the septum [L]'). As can be seen, there is a very developed set of such suffixes, far beyond that found in the Salish languages, for instance (cf. Hinkson 1999 for discussion of locative suffixes in Salish). 4.5.8

Completive

This is a group of morphemes that indicate completion of an action, often glossed as 'having ...-ed'. It may involve one of several different suffixes, including -ck°ir, -du\ -naYkm, and -yu\ It should be noted that these markers operate independently of the aspectual system, which also encodes the notion of completion. 96. a. faCck^ii faC x k T

siXaack°i siXaa xk°i'

'fern roots which had been chewed up' chew-having..ed fern root-remains of

b. hawiiCaqsiXwePin hawii -Cu* -siX-we^Pin

'he was finished' finish-having...ed-MOM-3.QT

c. qahnaakaX qah-na-k10-' aX

'someone died' die-COMPLETE-NOW

d. hiieePicuu qatqyuuPak hawii 'there you are with a chief whose head is cut off hii-(m)eYPicuu qatq^-yu^'ak hawii LOC-2p.lND decapitate-ed-POSS

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

118 Linguistic Theory and Complex Words

Suffixation 119

4.5.9 The Order of Derivational Suffixes The order of derivational suffixes, unlike that of either aspectual or inflectional suffixes, is relatively free, being based on the semantic exigencies of the word rather than on a formally determined fixed order. Thus two suffixes may appear in two different orders with respect to each other, based on the semantics of the stem, as in (97). 97. a. Xupyiihackwim?at?i 'her sweat' Xup -yiihaxkT-im-'at =Pi* heat-suffer from-..ed-thing-lNAL=DEF b. safihtimck^iPi saq -'ihta-imxkcor(q) =?r

'the spear prongs' dentalia spear -at end -thing-..ed=DEF

Note that -im 'thing' appears following -ckmr 'remains, ..ed' in (a), but precedes it in (b). Derivational suffixes constitute the largest group of morphological elements besides roots; there are between 600 and 700 separate morphemes. These cover virtually every aspect of the lexicon and bring with them various modifications to the roots to which they attach. Their order is relatively free, unlike that of inflectional and aspectual suffixes. A subset of these suffixes are portmanteau morphs, containing two or more suffixes in a single, unified morph. They are at least in part responsible for the large number of derivational suffixes within the language. 4.6

Summary

As we have seen in this chapter, suffixation is a very important aspect of the grammar of Nuuchahnulth and operates in all areas of the grammar. There are a large number of inflectional paradigms indicating combinations of mood, person and number. Non-paradigmatic inflection may involve a number of markers of various inflectional categories, including valency-changing morphemes, plural markers, diminutives, etc. Aspect also plays an important role in the grammar and involves a complex

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

These suffixes would appear to complement the aspectual morphemes to be discussed later on in this chapter. As one can see from examples such as (96b), they may co-occur, and when they do, the completive marker may occur inside the aspect marker, indicating that these cannot be inflectional morphemes.

120 Linguistic Theory and Complex Words

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

system of markers. Hundreds of derivational suffixes, probably more than any other language outside the family, are employed to build complex words, modifying the base to which they attach in a variety of ways involving vowel lengthening or shortening, reduplication and combinations of these.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

5

In this chapter we will examine two phenomena, reduplication and infixation, which share a rather intimate relationship in Nuuchahnulth grammar. There are numerous patterns involving the combination of these two processes. Furthermore, both processes require rather similar treatment, a topic that will be addressed in detail in Chapter 7 on templatic morphology. We will begin with the larger topic of reduplication, followed by the more restricted patterns of infixation in-Section 5.2. 5.1

Reduplication

Reduplication in its various forms has been the topic of much discussion in the linguistic literature and reduplication in Nuuchahnulth is an important and highly productive process. There are a number of different patterns of reduplication representing different grammatical requirements and one aim of this section will be to present the different patterns of reduplication attested in the language. A second aim will be to provide a systematisation of the interactions of the various reduplications. Finally, this section will discuss the nature of double reduplication in Nuuchahnulth that will lead to further discussion in the following chapter concerning the treatment of such cases within constraint-based theories. Sapir & Swadesh (1939:237-239) present the basic facts of Nuuchahnulth reduplication, introducing the terminology and diacritics still employed by Wakashanists today. They discuss briefly the various patterns of reduplication and annotate their glossary for those morphemes that require the reduplication of the base, a topic to be discussed later in this chapter. For Kyuquot, the most northerly variety of Southern Wakashan, Rose (1981) has discussed the typologically uncommon phenomenon of suffixes that appear to require concomitant reduplication of the base and the possibility of several suffix-triggers co-occurring in a single form. 121 10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Reduplication and Infixation

Stonham (1990b, 1994b) has discussed the existence of similar reduplicative processes in Ditidaht, the most southerly variety on Vancouver Island, from a theoretical perspective, arguing for a constraintbased approach to the phenomenon. Kim (2003a) discusses theoretical aspects of this process for the Ahousaht variety. We will begin with an introduction to the various patterns of reduplication found in Nuuchahnulth, followed by a discussion of the different grammatical morphemes in which these patterns occur. We will then move on to a discussion of the interaction of these various morphemes, and end with a brief discussion of the possibility of double reduplications, a topic that is relevant to the issue of stratal segregation to be discussed in Chapter 6. 5.1.1 Patterns of Reduplication There are a limited number of possible patterns which copies may follow, for the most part based on some portion of the syllable. These patterns always involve the copying of the onset followed by the vowel of the root,1 but there may be certain additional modifications, including lengthening or shortening the vowel, inserting certain pre-specified consonants, or copying the entire root. We will examine each one of these possibilities in the following sections. It is worth noting that, for the most part, Nuuchahnulth observes the constraint proposed in McCarthy & Prince (1998) regarding the possible templates available for reduplication, i.e., while reduplication may involve a prosodic unit such as a single mora, two moras, or a full syllable, it may not involve such constructs as the nucleus and one element from the coda — the onset is always part of the copy. However, several of these patterns are worthy of note. 5.1.1.1 Onset-Nucleus Reduplication This is perhaps the most common pattern of reduplication in Nuuchahnulth, involving the copying of the onset and nucleus of the root. In this case, no statement about the internal structure of the nucleus need be made, it will simply copy the structure of the original nucleus. Following are some examples of this process: 1.

a. dadaxpaisiPat DUP- dax40 -pai -siX -'at

'they speared at him' SUF-spear-several at one[R]-MOM-PASS

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

122 Linguistic Theory and Complex Words

Reduplication and Infixation 123 b. hawiiPaX cuduqhta hawiiX-'aX DUP- du-(q)hta

'he finished washing his feet' finish[M]-now SUF-wash-on foot [R]

d. waawaasCiihak DUP- waas -Cii -haYk

'where is your name?' SUF- where? -name [R] -2s.iNT

e. diidink0)iiksiX DUP- [L] dink" -iik*0 -siX

'he twisted their heads' PL- twist -at the head [L] -MOM

f. hiihimfaqXnuFaX DUP- himtak-'aqX-nukt0-'aX

'crosswise in his hand' SUFxrosswise-inside-at hand[R]-NOW

g. susudink Paanaqimi DUP- suda-riuk0 Panax-imi

'five hands long' SUF- five-at hand[R] thus much-CLS

One important thing to note about this pattern is that it involves either a long or a short vowel within the nucleus, depending on the weight of the original root vowel. Examples (la-c) involve short stem vowels, while (Id) involves a long stem vowel, (le-f) involve nasal sonorants in the coda of the stem, and demonstrate that the copy vowel is long in such cases. Note that this supports the claims concerning the moraic status of nasals discussed in Chapter 2, since [R]-type reduplication always replicates the underlying weight of the root, as shown in the previous examples. Note that (lg) provides an example of the same suffix as in (If), but inducing reduplication of a short stem vowel. 5.1.1.2 CVV Reduplication The difference between this pattern and that described above is that this one requires that the copy be long, regardless of the length of the original vowel, whereas in the previous case the copy vowel will only be long if the original vowel was long. 2.

a. kuukuhinqii DUP- [L] kuh -inqii

'with a hole in their side' SUF- hole -at the ribs [RL]

b. XiiXicswanup DUP- [L] Xi -swanup

'shoot in the armpit' SUF- shoot -in the armpit [RL]

c. PuuPumiiPak DUP- [L] Pu -miiPak

'for fear that...' SUF- REF -fear [RL]

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

c. taataaqhtinXwePin 'he squeezed her ankle' DUP-taaq^-^hta-inawiX-we^in SUF- squeeze -on foot[R] -MC-3.QT

d. suusuuhinqh DUP- [L] suu -hin -(q)h

'holding onto the end (of the spear)' SUF- hold -at end [RL] -MW

e. yaayaacsyup DUP- [L] yaac -syup

'urging them to go' SUF- go xoax [RL]

f. CiiCimcithinPasPi 'the right end of the village' DUP-[L]Cimxit-hin-'as=Pf SUF-right-end-at end[RL]-in village=DEF Examples (2a-c) show underlying short vowel stems undergoing CVV reduplication, with the concomitant lengthening of the copy vowel, whereas examples (2d-e) show underlying long vowel stems. Example (2f) shows a stem containing a nasal sonorant, which is believed to be moraic (see Section 2.4) and yet does not appear in the coda of the lengthened copy. This suggests that even though nasals may have moraic status for the purposes of stress assignment and influence the lengthening of the copy of a short root vowel, they may not participate directly in this type of reduplication. Comparing this example with (2d) shows that this suffix requires a lengthened copy. The main difference between this pattern and the previous one is thus only observable in those stems with underlying short vowels. 5.1.1.3 CVc Reduplication The next three patterns involve the insertion of a specific consonant into the coda of the syllable. Note that these are not cases of sub-syllabic copying but rather cases of pre-associated material within the syllabic structure, which in each case pre-empts the copying of material into the coda. The nature of the process involved here will be the subject matter of Section 7.2. The first case is that of the copying of onset and nucleus with the consonant Id associated with the coda of the copy, as exemplified in the following: 3.

a. haachawacsupfaai DUP- [L] ha\Va -supt'aai

'they had an eating contest' SUF- eat -compete in [RLC]

b. hiichinhsacpePi 'the ones on the beach side' DUP- [L] hina-hsa*xpaY =Pi' SUF-LOC-atbrink[RLc+L]-on..side=DEF c. XiicX'iiya DUP- [L] Xi -(y)ar

'he shot repeatedly' REP- shoot -DUR

d. XaacXaahsa DUP- [L] Xa -hsaY

'it was standing at the edge' SUF- stick up -at the brink [RLc+L]

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

124 Linguistic Theory and Complex Words

Reduplication and Infixation

125

The examples above demonstrate different usages of this pattern. (3a,b) involve suffixes which require the reduplication with insertion of Id, (3b) involving a nasal in the coda of the stem. (3c-e) involve regular reduplication which has the insertion of Id as a phonological concomitant. These will be discussed below in the appropriate sections. Note that in all cases the vowel of the copy is long. 5.1.1.4 CVt Reduplication The second pattern of this type involves the appearance of l\J in the coda of the copy. Examples of this type of reduplication include: witwisksaqXPi DUP- wisksaqX =Pr

'the bad-tempered ones' PL- bad-tempered =DEF

maatmaas DUP- maPas

'the tribes' PL - tribe

haathaak^aX DUP- haak^aaX

'the girls' PL - girl

rhitminuTaqX DUP- minu*q -'aqX

'the disease-throwers' PL - throw disease element xxpert

ifiiitminiiPi DUP- minaaii =?f

'the fishing banks' PL- fishing bank =DEF

Here, the examples involve plural formation. Examples of this pattern, like the previous one, will be discussed more fully further on in this chapter, but the reader should note that (4a,d) involve short root vowels and (4b,c) long ones. This pattern, unlike the one in Section 5.1.1.3, does not stipulate the length of the copy vowel, short vowels remaining short and long vowels, long. Again, note the difference in the effect on the copy between (4d), where the nasal appears in the onset of the stem, and (e) where it occurs in the coda. The length of the copy vowel in the latter is due to the movement of the nasal into the coda via a process of rhythmic reversal and the concomitant lengthening due to a moraicallyheavy root.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

e. PiicPiiqhaCipaX 'relating s.t. about s.o.' r DUP- [L] ?iiqhuk-(y)a -Cf p-'aX PL- tell -DUR-IND1R-NOW

126 Linguistic Theory and Complex Words 5.1.1.5 CVX Reduplication

5.

a. k^iiXk^iiya DUP- k°i -(y)a-

'filing' REP- sharpen -DUR

b. tiiXtiiya DUP- ti -(y)a*

'rubbing' REP- rub -DUR

c. suuXsuuya DUP- suu -(y)a*

'holding' REP- hold -DUR

Note that in these examples, the vowel of the root may be long (5c) or short (5a,b). In all of these cases, the consonant involved is in some way associated with the pattern of reduplication forming the repetitive aspect. The templatic nature of the three preceding patterns will be the topic of discussion in Section 7.2 on fixed segmentism. 5.1.1.6

a Reduplication

Reduplication may also take the form of full syllable reduplication, copying everything within the root syllable. As this pattern suggests, in all such cases an onset must be present, and must obey the constraints on the shape of onsets, but the rhyme must be left underspecified and depends crucially on the shape of the rhyme of the original. Thus, fullsyllable reduplication occurs in the following forms: 6.

a. qaahqaaha yaqCiPathPitq 'killing his neighbours' DUPx-ah-(y)a* yaq^-Ci-'ath-Pf tq REP- dead-DUR neighbour b. naasqnaasqa DUP- nasq -(y)aY

'beating time' REP- beat (time) -DUR

c. dimskdimskquu DUP- cimsk XJUU

'it was always like if whistling' REP- make a whistling sound-3.CND

d. Paaqinhak ditdits 'why do you keep dodging to one side?' ?aqin-[L]-haYk DUP- dit -s why?-2s.lNTER SUF-sidewise-lTER[F] e. AiXxniXxsCiPaXwePin 'bubbling up' DUP- niXx" -s -siX -'aX -we7in SUF- boil -ITER[F]-MOM-NOW-3.QT

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

The final pattern of this type involves /X/ occurring in the coda. Examples of this pattern are provided in (5).

Reduplication and Infixation

127

An interesting example involving full reduplication bears on the nature of word formation in Nuuchahnulth. Examine the following examples of syllable reduplication: a. CiicsCiicsSaqX DUP- Ci -iics -'aqX

'the trailers' REP- pull -go along -expert...-er

b. CiisCiisPa^at haayuuCistuui 'Ten-on-Water was pulled up the beach' DUP-Ci-'is-'ap-'at haayuuCistuui REP- pull -on beach-CAUS-PASS Ten-to-Water-at-Intervals Note that the translation of the first example refers to 'trolling' a fishing technique whereby fishing lines are dragged behind the boat. The gloss however shows us that in fact what is referred to as trolling is a bimorphemic monosyllable, dues, composed of the verb di 'pull' and the verb suffix -iics 'go along'. The second example is similar, involving the same root, but combined this time with the suffix -'is 'on the beach' to create the culturally relevant stem diis 'pull up on the beach.' Note that there are numerous other cases of monosyllabic roots followed by consonants where the syllable reduplication does not copy the following consonants, as in (8): 8.

a. caaXdaaCyakaXuk DUP- da-tLJ-s-yak^-'aX-uk b. suXsuuksiPaX DUP- SUU -S -SiX -'aX

c. fiXfiChtinama Xdup- fi -s -htinama

'... for each wave' REP- flow -iTER[F]-tool-NOW-POSS 'they started taking them' REP- hold -ITER[F] -MOM -NOW

'throwing device' REP- throw -ITER[F] xlevice

The difference between the cases in (8) and those in (7) has to do with lexicalisation: only the cases in (7) involve lexicalised stems, which may then participate in syllable reduplication as a unit. The actual nature of this pattern, and the complexities it involves, will be discussed in greater detail in Chapter 7. 5.1.1.7

0 Reduplication

There is one pattern of reduplication in Nuuchahnulth that involves a partially specified foot, as demonstrated in the examples below. 9.

a. daPaaxdax^ePi DUP- cax10 -(y)aY =?ir

'the spearers' PL- spear -DUR =DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

7.

b. tu?uuxtux<0ePi DUP- tux -(y)a' =?ir

'the jumpers' PL- jump -DUR =DEF

c. waPaacwacaXquu DUP- wac -(y)a--'aXxiUU

'they were wagging their heads' PL- wag head -DUR-N0W-3.CND

d. maPaaimaiawePin DUP- mai -(y)a* -we~?in

'they were wriggling about' PL- move -DUR -3.QT

The readers will observe that all the underlying roots are CVC and that the reduplication involves forming a foot-based copy of the original. Thus, the initial consonant and vowel are copied, followed by a glottal stop, a lengthened copy of the original vowel, and finally the coda of the original root. These are the only clear examples encountered so far, but this seems enough to establish a pattern. All forms involve copying from a CVC root (always containing a short vowel). The interpretation is plural and repetitive. Note that this is similar to repetitive full reduplication, but combined with fixed segment ? and a templatic foot structure (to be discussed in greater detail in Chapter 7). It is not yet clear whether this is some sort of rhetorical lengthening for effect. 5.1.1.8

Vowel-Shortening Reduplication

One type of reduplication that we never encounter in Nuuchahnulth is that where only a single consonant may appear in the coda regardless of the number of consonants in the original coda. This fact supports the claims made by McCarthy & Prince (1998) regarding constraints on the possible forms of reduplication. However, another marginal type of reduplication, attested at present in only a single suffix, is that where the nucleus is constrained to being short. The data are as follows: 10. a. q^aaq^atyakPitq DUP- q^aa -ityak -Prtq b. PuuPutyak DUP- Pu -ityak

'which they who are nerveless ...' SUF- thus -fear [R+S] -3s.REL 'for fear that they ...' SUF- REF -fear [R+S]

d. Pah q^aaq^atyaksiXah. 'that is what I am scared of Pah DUPxf aa-ityak-siX-(m)a-h this SUF-thus-fear[R+S]-MOM-lS.lND d. q^iiq^ityakiik DUP- q<0i(q) -ityak -iik*

'which might frighten you' SUF-that which-fear[R+S]-on head[L]

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

128 Linguistic Theory and Complex Words

Reduplication and Infixation 129

Note that to test this pattern, only stems with underlying long vowels, such as (faa 'thus' or those vowel-final roots which coalesce with a vowel-initial suffix (see Section 2.5.3), such as ?uREF and qmi 'that which', will be useful. With the latter, we would expect a long vowel to appear from the coalescence of the root-final vowel with the suffixinitial vowel, as in the examples in (11) below. 11. a. q^iq^iiyuk^iis DUP- qa,i(q) -ayuk -iis b. q^iiq^iiqXPitq DUP- q
'why I am crying' SUF-that which xry[R+L]-lS.lND.REL 'whatever was inside of him' PL- that which -inside -3s.REL

To date, only one suffix cited by Sapir & Swadesh (1939), -ityak 'fear', appears to exhibit the property of shortening its stem vowel and this is at best a marginal case. Nevertheless, it is an interesting one and one that poses questions for morphological theory. We will now move on to an examination of the various morphological uses of reduplication in Nuuchahnulth, subdividing them into the three categories of inflection, aspect, and derivation. 5.1.2 Inflectional Forms of Reduplication Reduplication is employed in the inflectional system of Nuuchahnulth, primarily to indicate pluralisation or distributivity in certain contexts. We will discuss the two patterns separately in the following two subsections. 5.1.2.1

Pluralisation

Pluralisation is realised in a number of different ways in Nuuchahnulth and the reader is referred to Section 4.3.4 for further details of this category. One means to mark it is by reduplication, although this is only one of a number of possible realisations of the plural.2 Data is presented in (12). 12. a. PePiidim DUP- Piidim

'old people' PL- old person

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

e. hiihityakaah maawicinPi 'Maawitsin had no one to fear' DUP-hi-it^ak-'aah maawicin=Pi* SUF- unable to -fear [R+S] -IRR Maawitsin =DEF

b. taataayi DUP- taayii

'older brothers, seniors' PL- older brother

c. iaaiaak^ahakPi DUP- iaak°ahi -'ak =?r

'his slaves' PL- slave -POSS =DEF

In addition to regular pluralisation, reduplication may also be used to mark a distributive form of reduplication as discussed in the next section. 5.1.2.2 Distributive The distributive implies both a notion of plurality and that of randomness or being scattered about. It is often translated as 'each', 'here and there', or 'all about'. It is represented by the reduplication of the initial CV(V) of the root. With respect to the distributive, Rose (1981:329) observes: Distributive (CV#) reduplication has two main meanings, plurality and spatial distribution, which can, but need not, coexist in a given stem. It denotes plurality for entities which are clearly distributed within some domain of nature, e.g. a forest, beach, or sea, or some other domain such as kinship. Examples of this morpheme include: 13. a. PuuPuucsuXwePin DUP- Pu -iics -awiX -we^Pin b. nunupqimiayiiPat DUP- ftup Xjimi -ayi' -'at

'each took along his (spear)' DISTRIB- REF-take along-MOM-3.QT 'he gave a dollar to each' DISTRIB- one -CLS-give -PASS

c. PaPaXjMwiimaaPaXuksi 'mine were to be in pairs in the house' DUP-PaX-jirT-'ii-maaPaX-uk-si' DISTRIB- two-CLS-in house-intend to -POSS-ls.ABS These two patterns of reduplication, indicating some form of pluralisation, may both be described as inflectional. 5.1.3 Aspectual Forms of Reduplication Reduplication plays a substantial role in aspect marking. Both the repetitive/continuative and the iterative aspect are realised by reduplicative patterns, and both will be discussed in the sections below.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

130 Linguistic Theory and Complex Words

Reduplication and Infixation 131 Repetitive

The repetitive or continuative aspect (REP) is realised by the full reduplication of the root, accompanied by lengthening of the vowels of both copy and base, and the attachment of the durative aspect marking suffix -(y)a\ This may be described as full syllable reduplication, since all verbal roots involved are of only a single syllable and are copied in their entirety. Examples of the usage of this pattern follow. 14. a. ciiqciiqa DUP- ciq -(y)aY b. tuuxtuuxl0a DUP- tux -(y)ac.

fimtiimta3 DUP- fimt -(y)a*

'talking' REP- talk -DUR 'jumping' REP- jump -DUR 'naming' REP- name -DUR

Note here that there is a further modification in this pattern just in case the root to which the pattern is applied consists of a single, open syllable, such as kai 'file'. In such a case, a /X/ is attached to the end of the copy, as shown in the example below. 15. a. k^iiXk^iiya r DUP-X- k*i -(y)a b. qiiCiX qii -siX

tiiXtiiya DUP-X- ti -(y)aY

c. suuXsuuya DUP-X- suu -(y)a*

'filing' REP- file

-DUR

sleep

-MC

'he was rubbing for a long time' long time -MOM REP- rub -DUR 'holding' REP- hold -DUR

Notice that X-insertion is a property of this specific aspectual category. This point will be relevant in later discussion. 5.1.3.2 Iterative Aspect The iterative aspect involves full-root reduplication with the affixation of the suffix -s.4 The length of the root vowel is transferred to the copy vowel. Examples of this aspectual marker include the following (from Rose 1981: 277): 16. a. x^akx^aks DUP- x^ak -s

'it expanded every now and then' SUF- expand -ITER[F]

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

5.1.3.1

b. yimhyimhs DUP- $imh -s

'he became embarrassed e.n.t.' SUF- be embarrassed -ITER[F]

c. maimais DUP- mai -s

'it became cold e.n.t.' SUF- be cold -ITER[F]

When the iterativity is progressive or habitual the copy and the base are both lengthened, as in the examples in (17). 17. a. PuuCqPuuCq DUP- PuCq -[L]

'it's getting foggy again and again' SUF- be foggy -PROG [L]

b. XiihXiih DUP- Xih -[L]

'it repeatedly reddens' SUF- be red -PROG [L]

c. miitxmiitx DUP- mitx -[L]

'he turned repeatedly' SUF- turn -PROG[L] (Rose 1981:277)

The previous two sections have presented cases of aspectual reduplication, which are found to occur together with derivational morphemes in Nuuchahnulth. The following section presents the final category, that of derivational reduplication, a concomitant of the attachment of certain derivational morphemes. 5.1.4 Derivational Forms of Reduplication Reduplication plays an important role in derivational morphology, appearing as a concomitant of many morphemes of different types. In what follows we will examine the various categories involved in these processes. 5.1.4.1 Derivational reduplication A certain subset of the derivational suffixes and bound verbs in Nuuchahnulth require that certain effects on the shape of the root be manifested as a concomitant to their attachment, either length on the stem vowel, reduplication of some portion of the stem, or some combination of the above. The reason for considering these as the 'trigger' for these effects is the fact that whenever they appear in a word, as below (b examples), the appropriate effect also occurs. However, in an identically formed word lacking only the particular suffix, this effect is not found. We can isolate the approximately eighty reduplicative suffixes of this class in the same fashion as illustrated in (18-19) below.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

132 Linguistic Theory and Complex Words

Reduplication and Infixation 13 3 18. a. Xicmis Xic -mis

'white powder, paint, clay' white -thing 'flour' DUP- white -resemble [R]

19. a. siicminfiiPiiePitq siicmin -fiiPii -Pi'tq

'while imitating maggots' maggot -imitate -3s.REL

b. siisiickuk siicmin -kuk

'rice (= looks like maggots)' DUP- maggot -resemble [R]

However one wishes to look at the occurrence of this reduplication, it seems impossible to deny the connection between the morphemes and these effects, in this case the simple reduplication noted as [R]. Simple reduplication and vowel length may also co-occur in a word, as in (2021), indicated as [R+L]. 20. a. FaPak^aCiPeePita F a -Pak^aCiX -'eePita

'it will break (the stakes) to pieces' break -in pieces -3.PURP

b. Fak^aahiPi ^aanaaXa 'the wolf with the broken leg' DUP-kV[L]-hi=Pr cfcanaaXa sUF-break-at limbs[R+L]=DEF wolf 21. a. Puuhtaasa XiCiX qiCinmit Puuhtaasa Xi -siX qiCinmit b. XiXiiyuksiXhak DUP- Xi -ayuk -siX -ha'k

'at once Louse shot' at once shoot -MOM Louse

'have you got angry about the shooting?' SUF- shoot-be angry [R+L]-MOM-2s. INT

Here the suffixes -hi 'at the limbs'5 and -ayuk 'angry at' require that the root vowel be both reduplicated and long. The (a) examples show nonreduplicated forms of the stem. A further theoretical possibility is that a suffix may require reduplication with length, not of the root vowel, but of the copy vowel, annotated as [RL]. This possibility is also manifested in Nuuchahnulth: 22. a. kuhswiiPi CuupafuPi 'through the hole in the dorsal-fin holder' kulf-swi^Pr Cuu$afu=Pr hole -through=DEF dorsal-fin holder=DEF b. kuukuhinqii DUP- kuh -inqii 23. a. Puuhtaasa XiCiX qiCinmit Puuhtaasa Xi-siX qiCinmit

'with a hole in their side' SUF- hole -at the ribs[RL] 'at once Louse shot' at once shoot -MOM Louse

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. XiXickuk Xic -kuk

134 Linguistic Theory and Complex Words b. XiiXics&anup DUP- Xi -swanup

'shoot in the armpit' SUF- shoot -at the armpit [RL]

24. a. hiiahuqsrhinhPi hii -a'hawii-'ahs-rhinh=Pi'

'the ones who sit in front of ...' LOC -in front -in a vessel -PL =DEF

b hiihiiihinminhPi DUP- hii -[L]-hin-rhinh=?r

'both ends there' SUF- LOC -at end [RL+L] -PL =DEF

25. a. yaqCiPathPitq yaqw -Ci -'ath -Prtq b. yaayaaq^hinPitq DUP- yaq* -[L] -hin -Prtq 5.1.4.2

'neighbours (ones who live with one)' REL -along with -live at -3s.REL 'the one on the end' SUF- REL -at end [RL+L] -3S.REL

Action-at-a-Distance

The length and reduplication required by these morphemes can be triggered across intervening, neutral suffixes, as shown in (26): 26. a. taatimisuup DUP- taa -mai -suup 4 I b. dadaafaqXsui DUP- dat [L] -'aqX -sui 4 I

'he rubbed them at the eyes' SUF-poke-move about-at eye[C][R+L] 'with markings about his eyes' SUF-stripe-endowed with-at eye[R+L]

Thus we can see that we have here what appear to be discontinuous morphemes, that is a morpheme which consists of two parts, one a reduplicative copy of part of the root and the other, a suffix with overt phonological form.6 This is certainly true of any theoretical approach, such as that of Marantz (1982), that considers reduplication to constitute a morpheme by itself. This concludes our overview of the types of reduplication that occur in Nuuchahnulth. In the following section we will examine the classes of morphemes that employ the different types of reduplication.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

The fourth logical possibility is that of a suffix requiring reduplication of the root, and length on the vowels of both the root and the copy ([RL+L]):

Reduplication and Infixation

135

The following sections present categories of suffixes that trigger various forms of reduplication on the stems to which they attach. They are provided in this manner for those wishing to investigate possible explanations for these particular groups of suffixes co-occurring with reduplication. 5.1.4.3.1 Somatic Suffixes. Somatic suffixes that require reduplication are typically those that involve paired bodyparts such as legs, arms, eyes, etc. Nevertheless, the suffix may refer to only a single member of the pair and would still induce reduplication, as noted above for (20b). The following is a list of those somatic derivational suffixes that trigger reduplication that are encountered in the data: 27.

-'akXi -akui •as -hi -hta -inqii -kuuPas -nuk

'at the heel' 'at the earlobe' 'at the cheek' 'at the limbs, under the knees' 'on the foot' 'at the side, ribs' 'at side of head' 'in the hand'

-0iC -jMdas -sitk" -sta-qs -sawii -swini -wiihtui -yimi

'at the ankle' 'at the ankle' 'between the fingers' 'at the hips [R], at the side' 'at the eye' 'in the armpit' 'on the shoulder' 'on the shoulder'

It should be noted that, in addition to these, there are the various aspectual combinations of these with Causative, Momentaneous, etc. 5.1.4.3.2 Activity-Related Suffixes. Reduplication associated with activity typically involves actions that are continuous or repetitive, harkening back to the discussion of aspectual distinctions in Chapter 4. However, it should be borne in mind that not all such activities associated with derivational suffixes demand reduplication. Suffixes that do require it include approximately 30 different forms, including -ayuk 'cry for [R+L],' -cama 'go towards [R],' -dii 'name [R]', etc. 5.1.4.3.3 Adjectives. There are only a small number of adjectival suffixes that require reduplication as a concomitant, including the six in (28).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

5.1.4.3 Categories Requiring Reduplication

136 Linguistic Theory and Complex Words

-Ciiqa

'at fault [R+L]' 'along with, attached to [RL]' ' impeded by [R+L]'

-£ap -daq -k^in

'sore in [R]' 'sore at [R]' 'toy, young [RL]'

5.1.4.3.4 Adverbials. There are very few adverbial suffixes to begin with, and only a handful of adverbial suffixes that require the reduplication of the root, as demonstrated below: 29.

-a*pi -Cim

'too [R+S]' 'on account of, for the sake of [R]'

-ink*

'in conversation [RL]'

5.1.4.3.5 Locatives. There also appear to be a handful of locative suffixes which co-occur with reduplication although, given the large number of locative suffixes, this is a small percentage of the total: 30.

-aC

'at the shore, at the margin [RL]'

-hin -hsa*

'at the end [RL]' ' at the brink [RLC+L] '

-adas

'in front of, at the margin [RL]'

-hwak0

'all around [R]'

5.1.4.3.6 Other types. There are one or two cases requiring reduplication that do not fit into the categories already discussed. These include -r 'thing...-ed [R]' and -'in 'sound of [R]'. 5.1.5 Co-occurrence of Reduplication-triggers The above patterns of reduplication present a further interesting theoretical possibility: the co-occurrence of multiple triggers of reduplication on the same form. In this section we will examine what happens when more than one reduplication trigger appears in a form. Recall that a number of derivational suffixes require the concomitant appearance of a copy of the stem when they are attached. The theoretically interesting question then is: what if more than one of these suffixes is attached to the same stem? Should we expect multiple copies? In a constraint-based account there would presumably be only a single copy satisfying the requirement of reduplication. This possibility of multiple triggers of reduplication is illustrated by the following examples, where more than one of these reduplication-

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

28. -aida -Si

Reduplication and Infixation 137 triggers appears on a stem. The examples in (31) are from Rose (1981: 342).

t

'his legs are really big' broad -at leg [R+L]-really [RL+L]

I

b. puucpuumaisuiap DUP- pumai -(c)sui -apa

l

'he has really itchy eyes' itchy -at the eye [Rc+L] -very[RL+L]

i

l c. maamaaiPasap DUP- rhai -'as -apa

i

'he has really cold wrists' cold -at the wrist [RL] -really[RL+L]

i

t d. maamaaiyimiap DUP- rhai -yimi -apa

t

'he's really cold in the shoulders' cold -at shoulder [R]-really [RL+L]

i

e. hiihinPasdpayk DUP- hin -'as -dap -ayuk

t

i

i

'he was crying due to his sore wrist' there -at the wrist[RL]-sore in[R]xry

I

I

Notice that in these examples, where there is co-occurrence of reduplication-triggers, there is only one copy but other effects such as vowel length may also appear, according to the patterns of the suffixes. Furthermore, the effects on the final form are those required by all the triggers, with the proviso that only a single copy occurs. In addition to cases of the co-occurrence of two or more reduplicationtriggers resulting in only a single copy, it is also possible to find cases of these triggers co-occurring with aspectual reduplication, as shown in (32) below. 32.

k^iiXk^iiyaqhPaX DUP- k°i -iyaqh -'aX

*t

diimaakPi diimaak

=?f

i'

REP/SUF- file -sing a song [R] -NOW mussel shell knife =DEF 'he kept on filing with his mussel-shell knife while singing.' Again, we find only a single copy satisfying the requirements of all morphemes. It is important to note that the only way to confirm from the surface shape that repetitive aspect reduplication has occurred in this example is from the fact that the root consists of a single, open syllable,

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

31. a. XuuXuuk°aniap DUP- Xuk -a"nui -apa

138 Linguistic Theory and Complex Words and therefore requires the insertion of /X/, as discussed in Section 5.I.3.I.7 The reduplication-triggering status of -iyaqh 'sing a song' is readily established by data such as the following:

b. PaPaXjMiiyaqh DUP- PaX -jiri" -iyaqh

'they sing songs' SUF- REF -sing a song[R] 'they sang two songs' SUF- two -CLS -sing a song[R]

c. qwaqwarhiyaqh?aXqun 'we sing all these songs' DUP- q^a -ma--iyaqh-'aXxjun SUF-thus-much-sing [R]-NOW-lp.CND d. PuPuyaqhXaa Pah DUP- Pu -iyaqh=Xaa Pah

'they also sang this' SUF- REF -sing a song[R]=too this

Rule-based accounts associating a suffix directly with the reduplication would seriously overgenerate copies and conceal the nature of the operations involved: accounts which treat reduplication as a morpheme lose the relationship between the copy and the suffix, and treatments which consider reduplication to be simply affixation have no answer for the failure of the suffixes to require multiple copies when the environment for it exists. The conclusion to be drawn from the above facts is that reduplication must be treated as a constraint on stem shape in Nuuchahnulth. Regardless of the number of suffixes requiring reduplication that are attached, a single copy satisfies the requirement. 5.1.6 Double Reduplication There have been cases discussed in the literature of double reduplication, i.e. the occurrence of two copies attached to a stem, for instance Broselow (1982, 1983) and Urbanczyk (1995) for Salish, Buckley (1998) for Manam, Stonham (1990b, 1994b) for Ditidaht, etc. In the previous section we saw how Nuuchahnulth appears to avoid double reduplications, opting for a constraint that is satisfied by a single copy, regardless of the number of morphemes requiring reduplication. However, this is not the whole story: there are, in fact, cases of double reduplication in Nuuchahnulth, such as the following: 34. a. q^aq^aq^arhasCi haa\ViihaX 'half of the young men to a side' DUP-DUPxfaa-rhar-a-sCi haawiihaX DISTRIB-SUF- thus-as far as -in one group [R] young men

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

33. a. PuPuyaqh DUP- Pu -iyaqh

139

b. yayayaqhiPi DUP- DUP- ya-q -hi =Pr

'the long-limbed ones' PL-SUF- long -at the limbs[R+L]=DEF

c. k^ak^ak^aXiih DUP- DUP- k^aX -'hh

'sea-otter hunters' PL- SUF- sea-otter -look for [R]

With respect to such forms, it should be noted that, rather than involving two derivational suffixes triggering reduplication, or a derivational suffix and an aspectual morpheme, all of these cases involve either a derivational or an aspectual suffix and either a plain or distributive plural morpheme, the latter two being inflectional categories. This suggests a solution to Nuuchahnulth double reduplication in terms of a stratal model of the morphology and this issue will be explored in greater detail in the next chapter. 5.2

Infixation

We now move on to the second topic of this chapter, infixation. Unlike reduplication, infixation in Nuuchahnulth is quite restricted in its distribution. There are no more than a handful of different patterns of infixation, including both morphological and phonological cases. In what follows, each of these patterns will be presented and analysed as to its relationship to the structure of the word, beginning with the most obvious example of morphological infixation, the plural. 5.2.1 Morphologically-conditioned Infixation In the formation of the plural there are at least two allomorphs that appear to involve infixation, the -t- infix plural and the -y- templatic plural. 5.2.1.1

-t-Plural

This is a variant of the plural that appears in certain forms in the language. It is not the default form, but does occur in a dozen or so cases to mark pluralisation. In these cases, -t- is attached to the coda of the first syllable of the word, immediately before the onset of the second syllable. The first syllable always contains a long vowel, so that any root with an inherently short vowel must be lengthened to fit the pattern. Examples of this operation are presented below: 35. a. naatwaayasPi nawaYs -t- -'as =?r

'those who were sitting outside watching' sit idly chatting -PL- -outside =DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Reduplication and Infixation

b. t'aatnePis fana -t- -Pis

'several children' child - PL- -DIM

c. haatPum haPum -t-

'(every kind of) fish' food - PL-

d. haatwiiqXhqa hawiiqX -t- -(q)h -qtf

'they are eating hungrily' hungry -PL- -MW-3.SUB

e. qaatyapfatqin qayapfa -t- -'at x-in

'the legs' limb -PL--INAL-1P.REL

f. tuutpaati tupaati -t-

'these tupaatis' ceremonial privilege - PL-

g. naatyaqak nayaqak -t-

'the infant children' baby - PL-

h. muutwadaqsup muwaC -t- -'asqsup

'the Muwachath women' Muwachath -PL- woman

Note that for the most part, this pattern occurs when the first syllable is open and the second syllable begins with a sonorant. The reader should note that this may only occur in stems of at least two syllables, ruling out a large number of possible forms for this type of plural. This may be confirmed by examining the examples above, which constitute virtually the entirety of cases of this form of the plural. Given that the vast majority of roots are monosyllabic, the possible number of such cases is, in fact, quite limited. 5.2.1.2

-y-Plural

The infix -y- operates in much the same fashion as the -/- infix discussed above. It also indicates the plural, is in complementary distribution with -t-, and involves certain additional constraints on the root to which it attaches. 36. a. fiyiiqwiiPi fief -j- -'ii =?r

'those sitting on the floor' sit -PL- -on the floor =DEF

b. siyaaXaqa siiX -y- -'aqa

'you're all moving' move house -PL- -several doing

c. dayaapac dapac -y-

'canoes' canoe -PL-

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

140 Linguistic Theory and Complex Words

d. cayaaxuk dapac caaxuk -y- dapac

'swift canoes' swift -PL- canoe

e. huyuuhtakPi huhtak0 -y- =Pi*

'certain ones were known' know -PL- =DEF

f. iuyuuCnakPi iuucsma -y- -na^k0 =?i*

'the married men' woman -PL- have =DEF

g. mayaasCim masCim -y-

'commoners' commoner -PL-

141

We can see from the data above that the use of-y- is quite similar to that of -t-, but at the same time there are distinct differences. First of all, the y- infix appears between vowels, and therefore will appear in the onset of the second syllable rather than in the coda of the first. Furthermore, the consonant after the nucleus of the first syllable is invariably an obstruent. These are the two basically morphological infixes that are found in Nuuchahnulth. Other infix-like elements found in the language include the consonants associated with certain derivational suffixes, as discussed in Section 4.5.1.4, and certain phonologically-conditioned infixes, which will be discussed in the following two sections. 5.2.2 Phonologically-conditioned Infixation Within the domain of phonologically-conditioned infixation we may recognise several phenomena associated with syllable weight and reduplication. This is perhaps better described as morpho-phonological, since it is associated with a particular morphological category, although it is not associated with the marking of any particular morpheme. For a more detailed account of this phenomenon the reader is referred to Kim (2003c). The different forms of this infixation are discussed in the following sections. 5.2.2.1 -X- Insertion -X- insertion is found to occur where the root is a monosyllabic form without a coda. This type of reduplication should be seen as basically phonological in nature fulfilling phonological requirements on the shape of the copy based on the shape of the original. This should be distinguished from the morphological requirements on the shape of the copy

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Reduplication and Infixation

142 Linguistic Theory and Complex Words

imposed by certain suffixes, which have already been discussed above. Examples of the pattern include: 'filing' REP- file -DUR

b. paaXpaaya DUP- j5a -(y)aY

'potlatching' REP- give potlatch gift -DUR

c. tiiXtiiya DUP- ti -(y)a*

'rubbing' REP- rub -DUR

d. CiiXCiiya DUP- Cii -(y)a-

'pulling' REP- pull -DUR

This is the most common pattern of phonological infixation, and is only found in the reduplication of open monosyllables in the language. 5.2.2.2 -c- Insertion -c- insertion appears to perform a similar function to -X- insertion, but there are only a few examples of this possibility in the data. It is still unclear what the conditioning factors governing the choice between -Xand -c- are. Examples of this include: 38. a. XiicXiiya DUP- Xi -(y)aY

'shooting' REP- shoot -DUR

b. XaacXaaya DUP- Xa -(y)aY

'driving wedges' REP- drive wedge -DUR

c. haachuuia8 DUP- hawii -(y)ar

'displaying wealth' REP- display wealth -DUR

Notice particularly the final example, which has undergone further phonological reduction. 5.2.3

Conclusions

Infixation can be seen to be a fairly marginal operation in the morphology of Nuuchahnulth, restricted to a few limited patterns of pluralisation and to the expression of a few morphophonological and phonological contexts. Nevertheless, it poses some challenging questions for morphological theory, including the nature of the templates required to account for it.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

37. a. k^iiXk^iiya DUP- k°i -(y)a*

Reduplication and Infixation

Summary

Among the patterns of reduplication and infixation that are found to occur, it is fairly clear that these are more generally associated with a templatic morphology, requiring the specification of additional properties such as syllable structure and reduplication. The issue of templatic morphology in Nuuchahnulth will be more fully discussed in Chapter 7.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

5.3

143

Stratal Segregation in the Nuuchahnulth Lexicon

One of the primary contributions of Lexical Phonology and Morphology (henceforth LPM) to linguistic theory is in the introduction of stratal segregation to explain behaviour shared by only a subset of the elements of the lexicon within a language. The ordering of strata allows linguists to apply phonological rules to the result of combining one group of morphemes with another, while excluding other, non-participating groups of morphemes. Perhaps the best-known example of this property is that of the application of the English stress rules at only the first stratum, discussed, inter alia, by Kiparsky (1982, 1985) and Mohanan (1981, 1986), but for every language we might expect to encounter particular phenomena indicative of such stratal segregation. More recently, a version of Optimality Theory (OT) advocating stratal segregation and the concomitant feeding of the output of one stratum into the input of the following one has been advocated by a number of linguists, including Bermudez-Otero (1999, 2003), and Kiparsky (2000, 2003). The benefits of this move are obvious and some of them will be discussed in this chapter as they pertain to Nuuchahnulth morphology. For any given language, the number of strata will depend on factors such as the complexity and interaction of the phonology and morphology, but a guiding principle should be that the minimum number of levels necessary for explaining the facts is to be preferred. Assuming this principle and the fundamental tenets of LPM, we will now proceed to an examination of various phenomena in Nuuchahnulth that argue for the necessity of separate stem- and word-level strata. There are a number of areas of Nuuchahnulth grammar where there appear to be distinct differences in the application of rules. Some of these include:

144 10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

6

1. 2. 3. 4. 5. 6. 7.

145

Glottalisation effects Lenition effects Delabialisation P-deletion Stress assignment Bound Root/Free Form Distinctions Suffix Combining-Forms

In what follows we will consider each of these cases in turn, paying particular attention to the possible arguments for different domains of application for phonological and morphological rules in Nuuchahnulth and, in the end, arguing for a two-strata model of Nuuchahnulth morphology. Finally, given these assumptions about the lexicon, we will discuss the theoretically challenging nature of one specific morphological phenomenon in Nuuchahnulth, that of double reduplication, as introduced in Chapter 5. 6.1 Derivation/Aspect vs. Inflection There are strong arguments for dividing the morphology of Nuuchahnulth into two sets of elements, as first suggested by Sapir & Swadesh (1939): (i) derivational and aspectual affixes which were labelled 'formative'; and (ii) inflectional, or 'incremental'. In fact, no great weight should be placed on the significance of these basically descriptive terms, but the segregation of morphemes into different groups is the crucial issue, as will be demonstrated in what follows. It should also be noted that, while in large part the analysis of morpheme membership in one group or the other that is presented here will be the same as that of Sapir & Swadesh (1939), there will also be some differences based on more recent examination and analysis of the data. The distinction between derivation on the one hand and inflection on the other is quite robust in Nuuchahnulth. The distribution is marked most clearly in the 600-700 suffixes of the language, where the order of derivational suffixes is relatively free, being based on the semantics of the construction, but the inflectional suffixes are rigidly ordered, along the lines of the discussion provided in Chapter 4. Additional criteria for the division include the category-changing nature and lexicality of derivational suffixes and the paradigmaticity of most inflectional morphemes. As for linear order, derivational suffixes

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Stratal Segregation

146 Linguistic Theory and Complex Words

1.

PaPaPaXqimihtimyiiminhPaaqXePicuu -mai -'ii ||-minh-PaaqX-(m)eYPicuu DUP-DUP-PaXx]imi-hta PL- SUF- two-CLS-on foot [R]-move-in house|| -PL-INTENT-2P.IND 'you (pi.) will carry two dollars on your feet'

As can be seen from this example with four derivational and three inflectional suffixes, affixation is very productive in Nuuchahnulth. Reduplication, as discussed in Chapter 5, is also common, as demonstrated by the two partial copies of the root in this example. One complication arises with respect to the treatment of aspectual markers, which typically appear between derivation and inflection, as in the following examples: 2.

a. CuCuuk^afinmasnaksiPaXma 'they had them go about inviting' CuCuuk^afinma-'as-na^-siX-'aX-ma- invite-in village-have-MOM -NOW-3S.IND b. haaskwaisiXitwePin haask°ai -siX -(m)it -we'Pin

'it was nearly ...' nearly -MOM -PAST -3S.QT

However, further evidence demonstrates that it is possible for such aspect markers to occur intertwined with the derivational suffixes, although never with inflection, as in the examples below: 3.

a. hissiXmapt his -siX -mapt

'Black Raspberry (bleed-plant)' bleed -MOM -plant

b. iuCnaaksiXmaTiiqstuPaXah 'I want to get married' iuCnatkt0-siX-maS'iiqstuX-'aX-(m)aYh marry-MOM-want-NOW-1 s.IND It is possible to have more than one aspect occurring on a single form and in such cases it is always the outermost aspect marker that has scope over the entire form. Examples of this are provided in (4). 4.

a. wee?iCup§iPaXat wePiC -[L]-Vp -siX-'aX-'at

'... making him fall asleep' sleep -PL -MC -MOM -NOW -PASS

b. fafaapathiCiPaXwePin 'he began to think it over' DUP-fapata-hi'-siX-'aX-we7in REP- think -DUR-MOM-NOW-3.QT c. PuuPaakXinisapatqa Pu -[L]-'akXi-ini-sap-'at xjaY

'he got ahead of him each time' REF -last -ITER -GRD.C[L]-PASS-3.SUB

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

invariably appear closer to the root. As an example of the productivity of affixation in Nuuchahnulth note the following:1

147

d. huuiiPiXsiPaX hui -[L] -iPiX -siX -'aX

'they started dancing into the house' dance -into the house[MC]-MOM-NOW

e.

'they all began crying' cry -DUR -several...ing -INC

fihaaqiiCiX fihak -(y)a' -'aqa -iiCiX

This evidence supports Sapir & Swadesh (1939)'s decision to combine derivation and aspect into a single class of 'formative' morphemes and will be used here to support the decision to combine these elements on a single, stem-level stratum. 6.2 Arguments for Stratal Separation In what follows, various phenomena will be examined that serve to bolster the argument for stratal segregation. Some of these cases are phonological in nature while others may more accurately be described as morphological. Some indicate a level of inflectional morphology, while others provide evidence for a separate level of derivation/aspect. All go to support the notion that Nuuchahnulth morphology must be divided into two distinct strata. 6.2.1

Glottalisation

Glottalisation involves the interaction of a specific subset of suffixes with a preceding segment. The outcome will vary depending on the segment involved, but typically results in the change of stops to glottalised stops, fricatives to laryngealised glides, and sonorants to their laryngealised counterparts.2 The following chart illustrates the various changes that may occur: 5.

Plain Pt cxek r CO

qq sis r x" )f hw m nw y x x h

Glottalised —»

-> —> —> —>

i>i6Xt£F T y w mnwy

As noted in Chapter 2, glottalisation is 'triggered' by certain suffixes, represented by a glottalising mark /'/ at the beginning of the appropriate suffix. All such suffixes begin with a vowel, although not all vowelinitial suffixes trigger glottalisation. The means to differentiate these

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Stratal Segregation

suffixes from those that do not require glottalisation is perhaps an issue for further debate elsewhere.3 However one does it, it will obviously be necessary to make some distinction between the two categories of suffixes and for the purposes of this book the convention used in Sapir & Swadesh (1939) will be employed. Examples of suffixes that trigger glottalisation (6a) and those which do not (6b) are provided below. 6.

a. -'urn -'iX -'aX

'on the rocks' 'invite, go for' NOW

b. -im -iX -ac

'thing' MOM.CAUS 'receptacle'

Examples of how these suffixes combine with preceding elements are provided below. The suffixes in (7a) are all glottalising suffixes, while those in (7b) are not. 7.

a. wiskum wisk -'urn

'angry on the rocks' angry -on the rocks

hiitahfiXsiPaX hita [L] -ht -'iX -siX -'aX

'they started out of the woods' LOC-exit woods-go for[L]-MOM-NOW

qahnaakaX qah -na^k0 -'aX

'he was dead now' dead -have -NOW

b. titinkum DUP- ti -nuk° -im

'handwiper' SUF- wipe -in the hand [R] -thing

PaayimkiXqas Paya -miik -iX -qaYs

'may I be a getter of many ...' many -getter of-MOM.CAUS -ls.SUB

k^anux^aX k^anux*0 -ac4

'chamber-pot' urinate -receptacle

As can be seen from these examples, the result of the combination of a glottalising suffix with a preceding stop/affricate-final base is the creation of a glottalised version of the consonant, as presented in (5) above. Where there is no available appropriate version, as in the case of fricatives, which are [+continuant], and have no glottalised counterparts in Nuuchahnulth, the nearest possible continuant consonant is provided, i.e., for labialised fricatives, a laryngealised labial glide, AW, and for

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

148 Linguistic Theory and Complex Words

Stratal Segregation

149

coronal fricatives, a laryngealised coronal glide, lyl. For fricatives that are neither labial nor coronal, there is no application of glottalisation, as shown in (5).5 The data follow the pattern in (8). a. XiiwinPap Xiix40 -'in -'ap

'he laughed at him' laugh -sound of -CAUS

b. ?atuyis?a^aanitin 'they gave us deer meat to eat' ?aatus-'rs-'ap-'at-(m)it-(m)in deer -eat -CAUS -PASS -PAST-IP.IND c. maaXiicuPayin 'he had been seen wearing it around his head' maX [L]-(w)iic-(y)u?ai-'in tied-around head[L]-perceive-treated as d. sawaaPaX sax40 -'a?a -'aX

'they fled' flee -do for...sake -NOW

When the base ends in a consonant which is not a possible candidate for glottalisation, such as /x x h/, then no change occurs to the fricative and glottalisation is realised as an inserted /?/. 9.

a. cicihPaqXmapt DUP- cih -'aqX-mapt

'crab-apple wood' PL- sour -inside -plant

b. CiyaanuxPath Ciyaanux -'ath

'the Cheanuh Tribe' Cheanuh -tribe

Where the base ends in a vowel, the result is similar to that above, with the insertion of a glottal stop between the two vowels, preventing coalescence while indicating the presence of a glottalising morpheme, as in (10). 10. a. PinFaPap Pink*0 -'a*?a -'ap b. takCiPinim tak -Ci -'fnim

fire

'he made a fire' -in hearth -CAUS 'let us each do ...' each -along with -lp>3.FUT.lMP

With a non-glottalising suffix, coalescence may occur, as in (11). 11.

wePiiqqim we?irq -qr -im

'with a tufted ornament on his head' wear tufts -on head -thing

The most interesting aspect of this process with respect to stratal separation arises with regard to the fricatives. When a fricative constitutes part of an inflectional morpheme, then the anticipated concomitant glottalisation does not occur. At the level of inflectional morphology, there

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.

150 Linguistic Theory and Complex Words

... while glottalizing formative suffixes ... change voiceless fricatives to glottalized semivowels, glottalizing incremental suffixes affect only non-glottalized stops [.] Thus, the rule applies in the same fashion to stops and affricates at either level, but with fricatives there is a distinct difference. Rather than suggest that there are such systematic exceptions to the application of the rule, we may capture the regularities of the process by the introduction of phonological differences associated with stratal membership tied to the stem/word-level dichotomy: in short, there are two phonological rules or constraints with different inputs, one at the stem-level and the other at the word-level. As a demonstration of this distinction, examine the results of the application of the suffix -'aX NOW, a clear example of an inflectional suffix, compared with -'a^a 'on the rocks,' an obviously derivational morpheme. 12. a. hiiPaXwePin haak°aaX?i 'the girl was there' hii -'aX-we"?in haak°aaX=?r LOC -NOW -3.QT girl =DEF b. hiyaaPaqhPaaqXPitq hii -'a?a -(q)h -PaaqX-Pi'tq

'to the rocky point where they were' LOC -on rocks -MW -INTENT -3S.REL

From this evidence we can clearly see the distinction between the treatment of fricatives within the derivational/aspectual component and in the inflectional component. This suggests the need for two strata within the morphology of the language, with two independent rules of glottalisation, Glottalisation 1 and Glottalisation 2. Kim (2003c) provides a detailed account of glottalisation within a mono-stratal OT framework, employing stem and word domains on constraints on a single level to distinguish the two forms of glottalisation. This is a reasonable account given the constraints of mono-stratal OT, but the multi-stratal account seems preferable, explaining as it does not only glottalisation but the other phenomena to be discussed in this chapter. We will examine the merits of the different OT approaches later on.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

will be no change to produce laryngealised sonorants. Sapir & Swadesh (1939: 236) observe that:

Stratal Segregation

151

There is a distinction between those glottalising suffixes that cause preceding fricatives to become the corresponding laryngealised continuant and those which do not. The distinction is made on the basis of whether or not the suffix causing the process is a derivational suffix or an inflectional suffix, as illustrated below: 13. a. CiiwahsuiPap Cii -wahsui -'ap

'he pulled it out' pull -exit -CAUS

b. hitahfasPap hita -ht -'as -'ap

'they brought it out of the woods' LOC -exit woods -go to -CAUS

c. q^isPap q^is -'ap

'he did thus' thus -CAUS

d. wikapwePin PanicdaksiX 'it had not been dark long' wik-'ap-we*?in Panic-da*k-siX not -CAUS-3.QT thus -late -MOM The suffix -'ap CAUS is a causative marker and thus an inflectional suffix, placing it in the group of glottalising suffixes which do not affect preceding fricatives. Note that (13d) demonstrates that -'ap is a glottalising suffix. If we compare this with the derivational suffix -'in 'sound of, it will be quite clear how the level of affixation, and thus the domain, results in different outcomes for glottalisation: 14. a. XiiwinPap Xiix10 -'in -'ap

'he laughed at him' laugh -sound of -CAUS

b. maaXiicuPayin 'had been seen wearing it around his head' maX [L]-(w)iic-(y)u?ai-'in tied-around head [L]-perceive-treated as In this case the suffix triggers the change of fricatives to corresponding laryngealised continuants as expected. In fact, - 'in is not one but several possible suffixes, meaning 'sound of or 'treated as'. Nevertheless, all of these suffixes are clearly derivational in nature and therefore appear inside the inflectional suffixes such as -'ap CAUS. 6.2.2

Lenition

A parallel process to glottalisation is that of lenition, which affects fricatives, converting them to glides as in (15).6

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

6.2.1.1 Glottalisation and Level Distinctions

152 Linguistic Theory and Complex Words 15.

Segments sis

Lenition —>

y

—>

w

r x* x" h*

This process is much more restricted in terms of its frequency of occurrence, appearing with only a limited number of suffixes of which the following represent nearly an exhaustive list: 16.

-'aCiX -'is -'ii -'(i)saX

MOM

'on the beach' 'in the house' 'on the beach (MOM)'

It will be noted that all of the morphemes triggering lenition are stemlevel, either derivational or aspectual. No inflectional morphemes cause this effect and therefore one may assume that lenition is a stem-level process. The following examples illustrate the differences between the suffixes that require lenition (17a) and those that trigger glottalisation (17b), specifically with respect to fricatives. 17. a. caPuwis daPui" -'is

'waves rolling on the beach' waves roll -on the beach

daPuyis daPui -'is

'waves rolling on the beach' waves rolling -on the beach

XaPujMwisaPaXquu7 XaPu~ -prT-'(i)saX-'aX-quu

'when there is another (low tide)' another -CLS-on beach[M]-NOW-CND

yaayis yaai -'is

'there on the beach' there -on the beach

k°isiiyis?i k°is -'rs -'is =?r

'eating on a different beach' different -eat -on the beach =DEF

b. niiwiisitafrPat nixtin -'rs -i'ta -'iX -'at

'he was invited to eat salmon roe' salmon roe -eat -apt to -invite-PASS

qawayis qawas -'i*s

'eating salmonberries' salmonberry -eat

CaamayisnahaXma Camas -'rs -naYh -'aX -ma*

'they are about to eat s.t. sweet' sweet -eat -seek -NOW -3S.IND

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

xxhh

Stratal Segregation

153

caqiicqimyisuk^apatsi 'it cost me twenty dollars' caqix-qimi-'rs-uk-'ap-'at-si* 20-CLS-cost-POSS-CAUS-PASS-ls.ABS

The only argument to be made concerning lenition and its relationship to level ordering is domain of occurrence. As has been noted above, lenition appears to be a property of only the stem-level, since all morphemes exhibiting this property belong to the stem-level. Nevertheless, this suggests that the phonological property of lenition associated with these morphemes belongs to the phonological component of the stem-level. 6.2.3

Delabialisation

As discussed in Stonham (1999b), and briefly in Section 2.5.9, one of the conditioning factors for the loss of labialisation of rounded velars and uvulars is the boundary between derivation and inflection, at which point delabialisation occurs. In Nuuchahnulth there is an underlying distinction between round and unround velar and uvular obstruents as shown below: 18.

ka kii qi Xax Taak

'measure' 'lift up canoe' 'shift position' 'vertically flat' 'dog-salmon'

k°aa k°ii q^i Xax 10 < Tak <0

'move backwards' 'blow spray' 'that which' 'carved' 'whittle'

This distinction is neutralised in several contexts, but the most relevant for us here is at the boundary of stem versus word. In this situation, final rounded consonants will be delabialised at the level of inflectional morphology, but not within the stem level. This is illustrated by the examples in (19) and (20). 19. a. qahnaakaX qah -naV -'aX b. hisiikatwePin his(t) -i'k° -'at -we'Pin

'he was dead now' dead -COMPLETE -NOW 'they passed by' LOC -go along -PASS -3s.QT

Compare the examples above with those in (20), in which the labialisation remains intact. 20. a. Saafrmtnak^as DUP- frmtaa -na'k0 -'as

'you are going to have names' PL- name -have -go to

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

6.2.2.1 Lenition and Level Distinctions

154 Linguistic Theory and Complex Words

'someone going along ...' there -go along -outside

In all of the cases with labialisation, the immediately following suffix is derivational (20), while in those in (19) it is inflectional, indicating that the latter is the level at which the rule applies. This distinction is similar to that encountered with the English rule that deletes final In/ in words such as damn, hymn, solemn, etc. at the end of Stratum 1, resulting in subsequent forms such as damning [daemirj] or hymning [himirj]. In English, further affixation at Stratum 1 allows for the appearance of the underlying Inl, resulting in examples such as damfnjation, hymfnjal, solemfnjification, etc. Just as this argument has been brought to bear in support of stratal segregation in English, so the parallel argument involving delabialisation may be used as a test of stratal separation in Nuuchahnulth. 6.2.4

?-Deletion

The deletion of Rl occurs when a base ending with a fricative comes in contact with a suffix beginning with a glottal stop Rl. Such suffixes should not be confused with those which trigger glottalisation, represented by an initial /'/. In the present case the glottal stop must be viewed as being associated with a C position, as in (21a), as opposed to glottalisation, represented simply as a feature in (21b), as suggested by Zoll (1996), Shaw (1989) for Dakota, and Kim (2003c) for Nuuchahnulth. 21.

a.

a /

/

o

o

I [4-Glottal] In the case of ?-initial suffixes, when a base ending with a fricative comes in contact with the suffix, there are two possible outcomes that depend on the category of suffix involved. If the ?-initial suffix is derivational, then the glottal stop will disappear. If, on the other hand, the suffix is inflectional, then the glottal stop will remain following the fricative. Thus, this phenomenon too is subject to certain domain restrictions.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. yaasiik^as yaai -rkw -'as

Stratal Segregation

155

22. a. k^isitPaqsui k^ist -it -Paqsui PapPaqsui Pap -Paqsui b. caahaqsui daah -Paqsui hasaqsui has -Paqsui

'other side of the mouth' different -at side -at the mouth 'mouth' LOC -at the mouth 'Tough-mouthed (man's name)' tough -at the mouth 'loud voice' loud -at the mouth

Compare this with the results of an inflectional suffix, -?aaqX INTENT, in similar environments and we can see that the results are the same in the case of stops (23a), but quite different with respect to the fricativefinal bases, as in (23b). 23. a. dawaakPaaqXqa dawa* -PaaqX -qaY

'one would ...' one -INTENT -3.SUB

hiniipPaaqXqa hina -iiyip -PaaqX -qa*

'that they find ...' LOC-obtain -INTENT-3.SUB

muu^itPaaqXin muu -$it -PaaqX -(m)in

'we would sing four times' four -times -INTENT -1P.IND

b. crtasPaaqXqa ciTas -PaaqX -qaY diiCmuqwiiiPaaqXqa diCrmr -iii -PaaqX -qa* hiihPaaqXPitq hii -(q)h -PaaqX -Prtq

'he was going to talk marriage' WOO -INTENT 3.SUB 'he made himself ready to exact sacrifices' scapegoat -make -INTENT -3.SUB 'the place he was ...' LOC -MW -INTENT -3S.REL

As can be seen by this data, there is no ?-deletion at the level of inflection. Thus, this rule of ?-deletion operates only at the earliest level of the morphology, the stem level, but not at the word level, further supporting

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Take, for instance, the case of the derivational suffix -?aqsul 'at the mouth'. In the following examples one can see the outcome of this derivational suffix combining with base-final obstruents and fricatives. In the case of stops (22a), the Rl remains, whereas in the case of fricatives (22b), it disappears.

156 Linguistic Theory and Complex Words

the argument for stratal separation and an exclusively stem-level phonological rule of ?-deletion.

In Nuuchahnulth, the domain of primary stress is the first two syllables of the word, as discussed in Section 2.4. Weight, in the form of vowel quantity distinctions or coda sonorants, is a determining factor in stress assignment, but closed syllables containing a short vowel followed by one or more obstruent consonants still count as light for the purposes of stress assignment. Stress always appears on either the first or second vowel of the word, regardless of the presence of long vowels further on in the word. The question arises as to what happens if the word contains no long vowels: in this case, stress is assigned to the leftmost vowel. Note that this result obtains even in the case where there is a longer vowel in the third or later syllable of the word.8 Much work remains to be done in explaining the intricacies of Nuuchahnulth stress assignment and the interaction of the various factors that determine it. One interesting fact about the operation of stress is that stress assignment appears to be limited to the realm of stem-level phonology, as we see in (24). . a. duusuW'iX dusuk -[L] -siX

'begin to suspect' suspect -GRAD -MOM

b. qahnaakaX qah -na-k10 || -'aX

'someone now died'

c. hiiyiih his [L] -'iih

'be after blood' blood -try to get [L]

d. nayaqak

'baby'

e. hisfmyuup his -imi -Vp

'gather together' all -in group -MC

f. Cimsmiit Cims -rhrt

'Bear-son' black bear -Son

g. tanakmiiCiPaX tanakma -iiCiX || -'aX

'turned now into mosquitoes' mosquito -INC -NOW

h. huPacaCiXPaaqXma huPa -ca -SiX || -PaaqX -ma*

'he will come back' back -go to -MOM -INTENT -3S.IND

die -COMPLETE -NOW

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

6.2.5 The Domain of Nuuchahnulth Stress

In the examples of stress assignment discussed in (24) above, stress appears on either the root or on a derivational suffix immediately following it. It is worth asking the question: what happens when an inflectional suffix is in competition for stress assignment? Naturally, this is not a common occurrence given the prolific nature of Nuuchahnulth derivational affixation, but it occasionally arises that a free root with a short vowel may be followed directly by one of those inflectional suffixes described in Chapter 4. Examples below illustrate this situation. 25. a. Puhqaa Puh || -qa*

'it might...' be -3.SUB

b. hfihPiitq hii || -(q)h -Prtq

'where he was' LOC -MW -REL

c. wikquu wik || -quu

'it would not...' not -3.CND

d. Piihmaa Puh || -ma*

'it is...' be-3.lND

In all these examples, the stress appears on an initial, short vowel, in spite of the fact that a long vowel is available in the second syllable. The only apparent difference between these cases and those in (24) is the existence of a word-level boundary that immediately follows the first syllable. These examples suggest that inflection may play no role in stress assignment in Nuuchahnulth, supporting the proposal that stress assignment occurs before the attachment of inflectional affixes, i.e., at the stem level. If stress is already assigned at the stem level, the weight of the vowels of an inflectional suffix will have no bearing on the outcome of stress placement. It should however be noted that a definitive treatment of Nuuchahnulth stress awaits completion and there are some seemingly contradictory examples in the original data gathered by Sapir. Nevertheless, the assignment of stress would appear to support the separation of Nuuchahnulth morphology into two distinct strata, much as the same rule in English supports a similar distinction.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Stratal Segregation 15 7

158 Linguistic Theory and Complex Words

A further important distinction encountered in the data is the morphological phenomenon of root allomorphy. In Nuuchahnulth, there is a high degree of root allomorphy that distinguishes between stem- and wordlevel affixation. One form is chosen when the suffix is derivational or aspectual and another one when the suffix is inflectional. The distinction is made on the basis of whether or not the suffix attaching to it is derivational or inflectional and it is quite pervasive in the grammar. In (26), the first column provides the free form that occurs with inflectional suffixes and on its own. The second column indicates the basic word class of the root and the third column provides a rough gloss. The final column shows the form of the root that is found in combination with derivational and aspectual suffixes. The examples in (26d) are of borrowed words, mainly from English. 26.

Free Form caapin dePinwa da^uunis hax^inmac

Cl.

kiianuus Xuukwaana

N

makit q^aahin

V

tiik*in

V

c.

puS"ai

ADJ

d.

Cikinis misin tiipin Peepinis

a.

b.

N N N N

Meaning 'sawbill duck' ' Sheep's-Feet' 'edible berry species' 'wren'

Bound Form capatq dePitq da^utq hax^itq

'furseal' 'Wolf Ritual'

kiiatq Xuk°atq

'play with shells, dolls' 'fish with prong-spear' 'bake in sand, ashes'

makitq

'sleepy'

pufrtq

N

'chicken'

Cikitq

N

'boarding school'

N

'table'

misitq tipatq

N

'apple'

Pepitq

N

V

q^aahitq tik°atq

The point here is that the choice of bound vs. free form is not a property of any particular word class, native/borrowed distinction, or even a historical remnant. It is an active productive process applied even to appropriate borrowed words.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

6.2.6 Root/Free Form Distinctions

Stratal Segregation 159

27. a. capitqinakyak caapin -rnak10 -yak caapin caapin K Xuk^atquwis Xuuk°aana -uiw -'is Xuuk°aanaX Xuuk°aana -'aX . XaatmaqanuiPi Xatmapt [L] -a*nui =?i* XatmaptukPi Xatmapt -uk =?f

'(I) have a sawbill dance' sawbill duck-imitate in dance-device 'sawbill duck' sawbill duck 'Wolf Ritual Beach' Wolf Ritual -place -on the beach '(I) had given a Wolf Ritual' Wolf Ritual -NOW 'the one of yew' yew -along [L] =DEF

'his yew tree' yew -POSS =DEF

These are not the only instances of bound roots, however. Bound roots may occur in a variety of shapes, including examples such as the following, where the first instance in each pair is the free form and the second is the bound form. Note that in some cases the only difference between the forms is vowel quantity. 28.

diCmuu / diCmuq a. diCmuuPaaqXap diCrmf -PaaqX -'ap diCmuuPi diCrfur =PiY b. diCmuqck^i diCrhuY -ckT diCmuqiii diCrhu* -iii

'scapegoat' 'he was forced to make sacrifices' scapegoat -INTENT -CAUS 'the scapegoat' scapegoat =DEF 'what is sacrificed' scapegoat -remains of 'make a scapegoat' scapegoat -make

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

The allomorph of the nasal-final stem with /tq/ is found to exist only at the level of derivational/aspectual morphology and never at the later level of inflection, suggesting that here again a distinction exists between the two domains. Take, for instance, the following examples:

160 Linguistic Theory and Complex Words

hamuut / hamut a. hamuutPi hamuut =?f

'the bones' bone =DEF

hamuutuk Xuunim hamuut -uk Xuunim b. hamutck°i Piihtuup hamuut -ckwr Piihtuup hamuthtin hamuut -htin 30.

'bone'

difuuiPi difuui =Pr

maht'ii / mahfiq a. mahfiimitPi maht'ii -(m)it =PiY mahfiirhinhPi maht'ii -rhinh =?iY

'an elk's bone' bone -POSS elk 'remains of whale bones' bone -remains of whale 'clubs were made of bone' bone -made of warclub =DEF 'house' 'the former house' house -PAST=DEF 'the houses' house -PL =DEF

b. mahfiqapuiPisPi mahfii -apui -Pis =Pr

'the little pretended house' house -imitate -DIM =DEF

mahfiqcktoak mahfii -ck°r -'ak

'the remains of his house' house -remains of -POSS

As can be seen from the examples above, where the (a) forms involve free roots combining with word-level inflectional suffixes and the (b) forms involve bound roots combining with stem-level derivational and aspectual suffixes, there are a variety of forms that bound roots may take. These may diverge from the free forms to a greater or lesser extent, but the suffixes found to occur with bound forms are consistent throughout the paradigm, suggesting some close affiliation. Such suffixes include derivational and aspectual suffixes, but no members of the previously presented inflectional category (see Chapter 4). What best accounts for this affiliation of bound forms with stem-level suffixes and free forms with word-level morphology is the principle of stratal membership, bound root allomorphs being restricted to Stratum 1, while free forms occur at Stratum 2, again supporting this distinction within Nuuchahnulth grammar.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

29.

Stratal Segregation

Suffix Combining Forms

Akin to the bound roots of the previous section is a set of suffix forms that occur in a special shape only in combination with derivational and aspectual suffixes but never with inflectional morphemes. Since 'bound suffix' would be a somewhat incongruous term, the term 'combining form', first introduced by Sapir & Swadesh (1939), will be employed to refer to this class of stem-level affixes. 31.

-syaqsti / -syaqstiq a. Puksyaqsti Pu -syaqsti yaqsyaqstiPitqPai yaq40 -syaqsti -Prtq =?ai b. Puksyaqstiqistasi Pu -syaqsti -ista -sr

'main or leading' 'the leaders were ...' REF -leading 'the one who was the head' REL -leading -3s.REL =PL ARG T was chief of the crew' REF -lead -..persons in canoe-Is.ABS

'the most valuable thing they had got' yaqsyaqstiqiiyipPitqPai 0 yaq* -syaqsti-iiyip-Prtq =?ai REL -main -obtain -3s.REL =PL 32.

-api / -apiq a. kiicapiPi kicuk -[L] -api =?v Xiikapiminh Xik -[L] -api -rhinh b. diisapiqsiPaX dis -[L] -api -siX -'aX

'up in the air' 'fallen tree that slants upward' log -in the air [L] =DEF

'they were placed stretching upward' have hands in position -in air [L]-PL 'send a telegram' rope, line -in air [L] -MOM-NOW

'they got five (birds) in the air' suudaqimimaiapiqiiyip suda -[L]-qimi-mai-api-iiyip five-CLS-move about-in air[L]-obtain The (a) examples above show what happens when the suffix completes the Stratum 1 stem, whether or not further affixation at later levels occurs while the (b) examples demonstrate the results within the stem level, where a 'combining' form is necessary for further derivational suffixation. Again, the suffixes that concatenate with these combining forms are the same ones that occur with the bound roots discussed in Section 3.6. As such, it seems obvious that there is a clear distinction made in the

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

6.2.7

161

162 Linguistic Theory and Complex Words

grammar between stem-level combining forms of certain suffixes and a stem-final form, which enters the word level. Summary

The seven preceding sections have presented arguments for separating the Nuuchahnulth lexicon into two levels: a stem level and a word level. The division of labour between these levels is as shown in (33) below. 33.

Morphology

Phonology

Stem-level

Bound roots Derivational/aspectual affixes Affix combining forms

Stress Assignment ?-Deletion Glottalisation 1 Lenition

Word-level

Free roots and stems Inflectional suffixes Affix free forms

Delabialisation Glottalisation 2

As (33) shows, both morphological and phonological elements must be separated into distinct levels, a stem level and a word level. We will now proceed to an examination of the importance of this stratal segregation for two different incarnations of one particular theory of grammar, Optimality Theory. 6.3 Double Reduplication in Southern Wakashan There has been previous work on the phenomenon of double reduplication by Broselow (1983) and Urbanczyk (1995) for Salish and Buckley (1998) for Manam. With respect to Southern Wakashan, Stonham (1990) addressed the topic briefly in a discussion of Ditidaht reduplication, but there has been little else on the subject. In what follows, we will examine this phenomenon in more detail, first for Tsishaath Nuuchahnulth, and then for several other varieties of Southern Wakashan. The discussion of reduplication in Chapter 5, and specifically Section 5.1.5 on the 'Co-occurrence of reduplication-triggers,' demonstrated that no matter how many reduplication-triggering suffixes occur in a stem, there will only be one copy on the surface, thereby supporting a constraint-based approach to the phenomenon: see the examples in (34), drawn from Chapter 5.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

6.2.8

Stratal Segregation 163 34. a. XuuXuuk^aniap 'his legs are really big' DUP- [L] Xuk [L] -a-nui -apa broad -at leg [R+L] -really [RL+L]

I

I

t

i

c. maamaaiPasap DUP- [L] mai [L] -'as -apa

t

[RL+L]

i

'he has really cold wrists' cold -at the wrist[RL] -really [RL+L]

I

I

Such an account, employing constraints on reduplication, predicts that, at most, a single copy would occur on a word, no matter how many demands for reduplication are placed on the form. While in general this is true for Nuuchahnulth, in what follows we shall see that it is, in fact, not always the case. Examine the data below: 35. a. q^aq^aq^arhasCi 'half of them to a side' DUP-DUP- q^aa -rhaY -a"sCi DISTRIB-SUF- thus-much-in one group[R] b. yayayaqhiPi DUP- DUP- ya-q -hi =PiT

'the long-limbed ones' PL-SUF- long -at the limbs [R]=DEF

c. PuPuPutah DUP- DUP- Pu -atah

'whalers' PL- SUF- REF -go after [R]

d. FaFaFaXiih DUP- DUP- k^aX -'iih

'sea-otter hunters' PL- SUF- sea-otter -look for [R]

e. XaXaXaPukuk^aX DUP-DUP- XaPu- -kuk -'aX

'each resembled someone else' DISTRIB-SUF- other-resemble[R]-NOW

The above are some examples of double reduplications found in Sapir's Tsishaath texts (Sapir ms). Note that in each case there is a reduplicationtriggering morpheme and either a plural or distributive morpheme associated with the forms, as described in Chapter 5. Bases with double reduplications appear to arise in Nuuchahnulth only in the context where there is a mixture of derivational/aspectual reduplication, i.e., stem-level morphology, with inflectional reduplication, i.e., word-level morphology. This suggests an explanation for the occurrence of double reduplication, which we will now explore.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

t

b. puucpuumaisuiap 'he has really itchy eyes' DUP- [L] pumai [L]-(c)sui-apa itchy -at the eye[Rc+L]-very

164 Linguistic Theory and Complex Words

36.

Morphology

Phonology

Stem-level

Bound roots Derivational/aspectual affixes Affix combining forms Affix-triggered Reduplication Aspectual Reduplication

Stress Assignment P-Deletion Glottalisation 1 Lenition

Word-level

Free roots and stems Inflectional suffixes Affix free forms Distributive Reduplication Plural Reduplication

Delabialisation Glottalisation 2

The issues surrounding double reduplication would appear to be: (i) how to account for the existence of double reduplications in Nuuchahnulth, and (ii) how to account for the absence of double reduplications within a specific morphological domain, while allowing double reduplication on the surface. Given this distinction between levels in the morphology, and some notion of bracket erasure that removes information about the internal structure of a form from the previous stratum, one might expect that it might be possible to apply various rules to the output of the stem level that results in new, reduplicated forms at the word level. A further prediction of such an operation is that the outermost reduplicative copy will be based on the output of the stem level and will not be sensitive to the form of the root, unlike certain stem-level operations that appear to

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

6.3.1 A Stratal Model of Nuuchahnulth Reduplication We have previously provided a number of criteria for distinguishing between stem-level and word-level morphology in Nuuchahnulth. These include both phonological (domain of application of glottalisation, stress, lenition, etc.) and morphological (bound vs. free stems, etc.) conditions. One of the crucial aspects of this analysis is the membership of different types of reduplication to different levels in the grammar. For instance, all derivational and aspectual reduplication belongs to stem-level morphology. On the other hand, plural and distributive plural patterns of reduplication belong with other types of inflectional morphemes at the word level. Schema (36) below, extended from (33) above, presents some of the morphological and phonological properties that distinguish the two levels.

Stratal Segregation 165

37. a.

[x"ak]R0OT [DUP- [x"ak] -s] [x'ak-tx^ak] -s] [x akx aksJSTEM

[x akx aKsJwoRD x^akx^aks b.

[x ak]R00T [xwak] -s] [x aksJSTEM [jfakS [xfflak§]] [x aksx aks]STEM

[x aksx aks]W0RD * x^aksx^aks

Non-derived Root Stem-level Affixation Stem-level Reduplication Bracket Erasure Word-level Morphology Bracket Erasure Non-derived Root Stem-level Affixation Bracket Erasure Stem-level Reduplication Bracket Erasure Word-level morphology Bracket Erasure

In (37a) we can see that the attachment of the aspectual suffix -siTER, triggers reduplication of the entire root, but does not include the suffix itself, as shown in (37b). Even when the trigger is at a distance from the root, only the root is copied, as in (38). 38. a. PuXPucaC DUP- Pu -ca

'he would go there e.n.t.' SUF- REF -go to -ITER[F]

t b. q^iXq^icaCPitq DUP- qtoi(q) -ca -s -Prtq 4 I

'went to the place where ...' SUF- that which-go tO-ITER[F]-3s.REL

c. hiXhinats DUP- hina -at -s 4 I

'(I shall always) be coming' SUF- LOC -come -ITER[F]

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

create a copy of the full root. Note that any operation sensitive to the structure of the root, such as reduplication, must have access to the internal bracketing available on the stem level, before bracket erasure obscures the structure. This prediction appears to be correct:

166 Linguistic Theory and Complex Words

39.

a.

Non-derived Root Stem-level Affixation Stem-level Reduplication (ITER) Bracket Erasure

[PU]ROOT

[[?u] -ca] [PuX [[Pu] -ca] -s] [[PuXPuca]C]STE

Word-level Morphology Bracket Erasure

[PuX?ucaC]w0RD PuXPucaC

b.

Non-derived Root Stem-level Affixation Bracket Erasure Stem-level Reduplication Bracket Erasure

LAU]ROOT

[[[Pu] *a] -s] [[Puca] -s]STEM [Puca [?uca]s] [PucaPucas]STEM

— Word-level Morphology [?uca?uca§]W0RD Bracket Erasure * PucaPucas In addition, within a constraint-based approach to reduplication, two stem-level triggers for reduplication, since they are on the same stratum, will still only result in one copy on the surface, as shown in (40). 40.

[mai]ROOT [[DUP- [mai] -'as] -a]

i

|_

r.

[maa- [rhaai] -'as] -apa] [maamaai?asapa] STEM

[maamaai?asap]W0RD * maamaamaaiPasap

Non-derived Root Stem-level Affixation Constraint unifying copies(*DupDup) Stem-level Reduplication/Length Bracket Erasure Word-level Morphology/Phonology Bracket Erasure

The output of stem-level reduplication will then be available for wordlevel morphology, including further reduplication as shown in (41).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Examples such as these are rather rare, since aspect typically occurs adjacent to the root, but it is possible to find cases where a derivational suffix intercedes. The representation of the form appears as in (39).

41.

[FaX^T [DUP- [raX] -'iih] [Fa- [FaX] -'iih] [FaFaXiih] STEM

Non-derived Root Stem-level Affixation Stem-level Reduplication Bracket Erasure

[DUP- [raraXiih]] [Fa- [raraXiih]] [raFaFaXiihJwofiD

Word-level Affixation Word-level Reduplication Bracket Erasure

167

Thus, a distinction between stem- and word-level morphology, in conjunction with a principle of bracket erasure at the end of each level and a constraint-based approach to copying, will allow us to explain both the absence of double reduplications due to two stem-level reduplication processes and the appearance of double reduplication across two levels. 6.3.2 Variation in Southern Wakashan The study of double reduplication in Nuuchahnulth does not end here. The data involving double reduplication so far in this chapter are drawn from the Tsishaath dialect as spoken in the early part of the 20th century (Sapir ms.). In these texts multiple reduplications are naturally rare, since the confluence of requirements for their appearance are fairly exclusive, but there are upwards of 50 instances, involving a number of different roots and reduplication-triggering suffixes. The existence of this phenomenon seems beyond question. Various authors writing on different varieties of Southern Wakashan have noted differences in the reduplicative patterns found to arise. In what follows we will examine the variation in reduplicative patterns in these different varieties. 6.3.2.1 Kyuquot Rose (1981:340), with reference to the Kyuquot variety, notes that: There is only one pair of reduplicative morphemes which can be present at once in a surface stem, the CV# distributive and the CVC# iterative. 42.

mimiitxmitx9 'they were turning repeatedly here and there' > DUP- DUP- [L] mitX° -(y)a DISTRIB- ITER- turn-GRAD[L]-DUR

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Stratal Segregation

Thus, multiple reduplications may occur in Kyuquot, but Rose notes that they do not involve reduplication-triggering suffixes. However, as noted above, aspect marking is a property of stem-level morphology in Nuuchahnulth, whereas the distributive is a property of the word level, so the generalisation made above still holds. The analysis of such forms proceeds in the following fashion, analogous to the example in (41). 43.

[mitx^RooT [DUP- [mitx*] -(y)a] [mitx^-tmitx10] -(y)a] [miitxmitxa] STEM

Non-derived Root Stem-level Affixation Stem-level Reduplication Bracket Erasure

[DUP-[miitxmitxa] STEM] [mi- [miitxmitxa]] [mimiitxmitx]W0RD

Word-level Affixation Word-level Reduplication Bracket Erasure / Phonology

Thus, double reduplications in Kyuquot still involve distinctions between stem-level and word-level morphology, although the sources of such co-occurrences are more restricted. 6.3.2.2

Ditidaht

Stonham (1994) notes that double reduplication also occurs in Ditidaht. When this occurs, it involves a mixture of stem-level reduplication with inflectional reduplication. Example (44) below illustrates the process. 44.

kakakawadatax DUP- DUP- ka\Vad -atax

'hunting killer whales here and there' DISTRIB- killer whale -hunt [R]

Again, this example may be explained by means of a two-level morphology, where bracket erasure obscures the earlier reduplication, allowing the application of a further copy at the word level, as shown in (45). 45.

[kawad]ROOT [DUP- [kawad] -atax] [ka- [kawad] -atax] [kakawadatax] STEM

Non-derived Root Stem-level Affixation Stem-level Reduplication (ITER) Bracket Erasure

[DUP-[ kakawadatax] STEM] [ka- [kakawadatax]] [kakakawadatax] WORD

Word-level Affixation Word-level Reduplication(DlSTRlB/PL) Bracket Erasure

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

168 Linguistic Theory and Complex Words

Stratal Segregation

169

An interesting complexity in this particular case is the fact that the free form of 'killer whale' is kaka&ad, with reduplication.10 Nevertheless, the form for 'hunt killer whale' does not trigger a further reduplication, attesting to the morphological complexity of the word. kakawadatax * kakakawadatax

'hunt killer whale'

Thus, kaka&ad, a native construction, is seen as an inherently reduplicated form, preventing the further reduplication at stem level, in spite of the fact that a further reduplication at word level is still possible. Contrast this with the word muusmus 'cow', an indirect borrowing, possibly from Cree mustus, via Chinook Jargon. The word for 'cow' is not a native Wakashan form and as such is opaque as far as this apparent reduplication is concerned. Therefore, there appears to be double reduplication even though, in fact, there is only one. 47.

muumuusmusatax

48.

[muusmus]ROoT [DUP- [muusmus] -atax] [muu- [muusmus] -atax] [muumuusmusatax] ST

[muumuusmusatax] WORD

'hunting cows' Non-derived Root Stem-level Affixation Stem-level Reduplication Bracket Erasure

(ITER)

Word-level Affixation Bracket Erasure

The question of what form the distributive of this word might take will have to remain for further research. From the discussion above, we may conclude that Ditidaht permits double reduplication only in those instances where the copies appear as the result of processes occurring at different levels of the grammar, much as in the previous cases from Tsishaath and Kyuquot. 6.3.2.3 Makah Davidson (2002: 211-212) provides a description and an example (49) of double reduplication in Makah, the southernmost variety of the family: Plural reduplication can co-occur with reduplication induced by affix-associated CV templates producing doubly reduplicated words. ... The Makah word is reduplicated according to

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

46.

170 Linguistic Theory and Complex Words the requirements of -wat [LR] 'friend of ...' and re-reduplicated to indicate simple plural. PuPuuPuwafida DUP- DUP- [L] Pu -wat -'ida

'friends' PL- REF -friend of [RL] -treated as

Makah appears to behave just as Tsishaath, Kyuquot, and Ditidaht do with respect to double reduplications, requiring that one copy arises from a reduplication-triggering suffix at the stem level and that the other arises from a plural morpheme at the word level. Bracket erasure eliminates any indication of the application of reduplication on the stem level, allowing for a further copy at the word level. 50.

[PU]ROOT [DUP- [Pu] -wat] -'ida] [?uu- [Pu] -wat] -'ida] [PuuPuwafida] STEM

Non-derived Root Stem-level Affixation Stem-level Reduplication Bracket Erasure

[DUP- [PuuPuwafida] STEM] [Pu- [PuuPuwafida]] [PuPuuPuwafida] W0RD

Word-level Affixation Word-level Reduplication Bracket Erasure

Up to this point, all the varieties, from the most northerly Kyuquot to the most southerly Makah, appear to behave in more or less the same way with respect to double reduplications, distinguishing between stemlevel and word-level morphological operations. 6.3.2.4

Ahousaht

Our final variety of Southern Wakashan for examination is Ahousaht, a Central variety. Kim (2003a,c) observes that, in fact, none of the double reduplications discussed for Tsishaath or the other varieties are possible in Ahousaht and it would appear that there are no contexts where double reduplication is found to occur. This presents an interesting contrast to the other Southern Wakashan varieties under discussion and must be explained. Note that the previous analyses all depended on a two-level grammar in order to explain the occurrence of two copies. One possibility with respect to the Ahousaht case is that, due to language change in progress, there has been a levelling of the grammar, resulting in a single stratum for all morphology. In addition to the absence of double reduplication, there appear to be additional differences between Ahousaht and the other varieties of

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

49.

171

Southern Wakashan. For instance, the glottalisation of nasals, which is a property of stem-level morphology in both Kyuquot and Tsishaath, is at best lexicalised in Ahousaht. Furthermore, the bound vs. free root allomorph distinction, one of the arguments for stratal segregation proposed in Stonham (2003), appears to be absent or, at most, vestigial in Ahousaht (Kim 2003c). Similar cases of such levelling in other languages have been noted and this is highly suggestive.11 Bermudez-Otero (in prep.) cites the case of Kaska, discussed by Kaisse (1993), in which a lexical rule of intervocalic I si deletion has expanded its domain, resulting in a much wider field of application. This is attributed to the movement of the rule to the wordlevel. Bermudez-Otero & Hogg (2003) further observe that a similar movement has occurred in Old English and suggest that if, as a result of a historical change, a phonological process previously associated with a stem-level operation is lost altogether, either of two things will happen: (1) the morphological operation associated with it may become unproductive, with the constructions in which it appears becoming lexicalised or (2) it may be moved back into the word level, becoming fully regularised. This would appear to be the case in Ahousaht.12 Obviously, more work is needed to determine whether such a monostratal analysis of Ahousaht is consistent with the other tests for stratal membership, but given the geographically widespread nature of the level distinctions in Southern Wakashan, it seems reasonable to conclude that Ahousaht is the innovator in this case. Alternatively, given the observations for Kyuquot made above, we might expect that there have been further restrictions imposed on the grammar of Ahousaht that conspire to eliminate the environments for double reduplications. One possibility is that bracket erasure only occurs at the end of the lexicon rather than at the end of each level, allowing the existence of the copy to be observed throughout the grammar. 6.3.3

Summary

Double reduplications can be found in most varieties of Southern Wakashan, and in fact, even in Northern Wakashan according to this quote from Boas on Kwak'wala (1900: 711-712): Other changes of the stem are due to reduplication, the method of which varies according to the grammatical function it performs. Double or even triple reduplication may

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Stratal Segregation

172 Linguistic Theory and Complex Words

They present a challenge to any theory of morphology and particularly for mono-stratal models, which have no way to account for the failure of double reduplication in certain cases such as multiple stem-level triggers and, on the other hand, the necessity of double reduplication in others. 6.4 Mono- versus Multi-Stratal OT The model of Optimality Theory was first proposed in several seminal papers by John McCarthy, Alan Prince, and Paul Smolensky (McCarthy & Prince 1993, Prince & Smolensky 1993). The model proposed was mono-stratal and concerned principally with the phonological component of a grammar. Since this early period, there has been a considerable amount of work done on the theory and on issues of both phonology and morphology. Nevertheless, there would appear to remain some troubling issues to deal with. One reaction to some of these issues has been the fairly recent proposal to divide the lexicon into a stem level and a word level, resulting in a multi-stratal model of Optimality Theory (see for example, BermudezOtero 1999, Kiparsky 2000, etc.). In this section we will not delve into all the intricacies of a multi-stratal versus mono-stratal model of Optimality Theory, but will focus on a single issue, the preferability of allowing alternate rankings of constraints with respect to distinct levels within the grammar. Focussing on the issue of double reduplication discussed above, we may propose a constraint of the following type to ensure that multiple copies are disallowed, as required in the case of multiple stem-level triggers to reduplication, as presented in (51). 51.

*DupDup: disallow multiple copies

For the purposes of this discussion a generic Base-Reduplicant faithfulness constraint that requires exact correspondence will be proposed. We will refer to this constraint as MaxBR but, given the current controversy surrounding the nature of base-reduplicant relationships, no theoretical significance should be associated with this proposal beyond

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

occur in the same word; for instance, from the stem bEgwman [bag 40 -], we have babagum boy [babagum], and babEbagum boys [babQbagum].

Stratal Segregation 173

52. *DupDup » MaxBR (Language X) *DupDup *|

Red- Red- ABC a. ABABABC13 *& b. ABABC

MaxBR

V. r\^-;_;> -\\ .*.

. r '•* * ^

^

If, on the other hand, we reverse the ranking, then double reduplication is ruled out, as in (53). 53. MaxBR » *DupDup (Language Y) MaxBR

Red- Red- ABC i®3 a. ABABABC b. ABABC

*!

*DupDup ;

V* " ^ - - ^-- ^ :

These two constraints will serve us well in ensuring that multiple stemlevel triggers do not result in multiple copies appearing on the surface. However, standard OT is mono-stratal, and therefore all constraints must be applied simultaneously to the input. This means that a ranking such as that in (52) will prevent the possibility of a second copy of the base arising at the word level, an undesirable result when accounting for double reduplication. To take a practical example of this, (54) involves the co-occurrence of multiple derivational reduplicative morphemes (from Ch. 5, eg. 30) 54. a. XuuXuuk°aniap DUP- Xuk -ariui -apa

•

i

'his legs are really big' broad -at the leg [R+L]-really[RL+L]

i

b.*XuuXuuXuuk°aniap As can be seen, the result of attaching two derivational suffixes that trigger reduplication is a single copy, as in (54a). (54b) demonstrates the ungrammatical result of applying the copying twice. Example (55) below involves one derivational and one aspectual reduplicative morpheme, both requiring copying (from Ch. 5, eg. 31).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

the point at hand. We take the further, perhaps controversial, step of defining the base as the initial core syllable of the word (McCarthy & Prince 1998). If we rank *DupDup above MaxBR, double reduplication is ruled out as in (52).

174 Linguistic Theory and Complex Words

55. a. k^iiXk'iiyaqhPaX

'he kept on filing while singing' REP/SUF- file -sing a song [R]-N0W

DUP- kwi -iyaqh -'aX

n

i

J

As in the previous case, the grammatical output consists of a single copy representing the demands of both derivational and aspectual morphemes. Example (55b) is ungrammatical. The best way to represent both cases is as in (56) below. 56. a.

Red- Red- Xuk -a-Aui -apa

= (54a)

to

a. XuuXuuXuuk ariiap *& b. XuuXuuk°ariiap

*DupDup *!

MaxBR ^^^^v^s^ ' ^ V .

'

•'"• -^ffJi-

b.

Red- Red- Wi -iyaqh -'aX

*DupDup

= (55a)

a. k°iik°iiXk^iyaqh?aX *& b. k"iiXk"iiyaqh?aX

*!

MaxBR ";l:j!-f% • •'.v,-«v*;

Up to this point, mono-stratal OT has no difficulty in making the correct predictions about the data. It is only when data of the type in (57) appear that this approach encounters a problem: 57.

TararaXiih DUP- DUP- kwaX -'iih

'sea-otter hunters' PL- SUF- sea-otter -look for [R]

In this case, the data involve one derivational and one inflectional reduplicative morpheme, and the result is two copies appearing on the base. In a mono-stratal approach, the preferred ordering of (56) produces the wrong result, (58b), since the higher ranking of *DupDup prevents double reduplication from appearing on the surface: 58.

Red- Red- TaX -'iih

a. TararaXiih © *& b. raraXiih

*DupDup

MaxBR

*!

Re-ranking the constraints produces the desired output (59a), but poses problems for (56), as shown in (60).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. *k0)iiktoiiXktoiiyaqhPaX

Stratal Segregation

MaxBR

•a- a. rararaXiih b. raraXiih

:fyi^ii^

*!

Red- Red- Xuk -arnui -apa

60.

*DupDup

MaxBR

*DupDup

' ^t*Z ;

(55a) © i®* a. XuuXuuX'uuraniap b. X'uuXuuraniap

-^-\

*!

s

V'J ,

It would seem that whichever ranking is chosen, one set of grammatical data will be ruled out, an undesirable result. Short of introducing new, parochial constraints into the analysis, there seems to be no way for mono-stratal OT to deal with this data and so we will examine how a multi-stratal OT approach to double reduplication might fare. In this case, there is no need to propose new constraints since the problem is obvious: it is related to stratal segregation, involving mismatches of constraint ranking, not constraint formulation. If we admit the possibility of two separate levels of grammar, both employing the exact same set of constraints, but necessarily requiring the re-ranking of at least one pair of constraints, then we can explain the seeming paradox of Nuuchahnulth double reduplication. The analysis proceeds in the following fashion: (i) we employ a stemlevel ranking as in (56) above; (ii) the output of this level will serve as the input to the word level, which employs the ranking in (59); (iii) the final output produces the correct results in both cases. 61. a. Stem Level

Red- R"aX -'iih

FaithIO

a. rararaXiih •sr b. raraXiih c. raXiih

*DupDup

MaxBR

*! *!

Ji b. Word Level

Red- raraXiih FaithIO «• a. rararaXiih b. raraXiih c. TararaTaXiih

MaxBR *!

*DupDup ••II1III1I1I1

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

*!

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Red- Red- TaX -'iih

59.

175

176 Linguistic Theory and Complex Words

*DupDup *!

MaxBR /'< '" *!

Ji b.

XuuXuuraniap

a. XuuXuuXuuraniap Word Level fc^ b. XuuXuuraniap c. Xuuk^aniap

FaithIO *! *!

MaxBR

*DupDup

- i : —\v?

.^Sv-:^-

V ' - Y : - /;••-

A multi-stratal approach to OT (e.g., Bermudez-Otero 1999, 2003, Kiparsky 2003) accounts for the failure of double reduplication within a level while allowing the possibility of multiple copies across levels. In Nuuchahnulth, there are a number of areas of the grammar where a distinction between stem and word level is necessary and in all such cases, a multi-stratal approach will provide a better analysis, without the need to propagate parochial, domain-based constraints to account for the alternations. 6.5

Conclusions

In summary, there are a number of arguments for separating Nuuchahnulth morphology into two strata. Both phonological and morphological processes provide evidence for making such a distinction. On this view, phonological rules such as stress assignment and glottal stop deletion would appear at Stratum 1. There would be two glottalisation rules, one at Stratum 1 affecting all eligible consonants and another at Stratum 2 restricted to just the stops and affricates. Furthermore, Stratum 1 would be the domain of both bound roots and suffix combining forms. Only free roots and stems would be allowed to enter Stratum 2. At Stratum 2 would also be the phonological rule of delabialisation. Adopting this approach to Nuuchahnulth morphology allows us to capture the special properties of the morphology in a concise and straightforward fashion. Certain phenomena are associated with Stratum 1 and others are associated with Stratum 2. The modules of the lexical morphology of Nuuchahnulth can be said to operate as in (63).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Red-Red- Xuk -aYnui -ap FaithIO 62. a. a. XuuXuuXuuTaniap Stem 5 Level PS b. XuuXuuraAiap c. Xuuraniap

Morphology

Phonology

Stem-level

Bound roots Derivational/aspectual affixes Affix combining forms Affix-triggered Reduplication Aspectual Reduplication

Stress Assignment P-Deletion Glottalisation 1 Lenition

Word-level

Free roots and stems Inflectional suffixes Affix free forms Distributive Reduplication Plural Reduplication

Delabialisation Glottalisation 2

Post-Lexical

Clitics

Aspiration Echo-vowels

63.

177

For the time being, a two-strata model appears to be capable of dealing with the various issues encountered in Nuuchahnulth morphology and phonology. This may be encoded in a constraint-based approach by employing multi-stratal OT to account for the stem- versus word-domain variation in the phenomena discussed in this chapter.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Stratal Segregation

7

In this chapter we will examine the various forms of templatic morphology (McCarthy 1981, Marantz 1982, McCarthy & Prince 1995) that are encountered in Nuuchahnulth. Templates are required in a number of areas of the grammar, including the numerous reduplication patterns, both with and without fixed segmentism (Alderete et al. 1999), discussed in Chapter 5, hypocoristic formation (Stonham 1991), vowel lengthening and shortening (Chapter 4), infixation (Chapter 5), and the somewhat more extralinguistic patterns of abnormal speech (Sapir 1915, Stonham 1999). We will begin with an examination of the patterns of reduplication and their treatment within templatic morphology. 7.1

Reduplication and Templatic Morphology

The types of reduplication and their various uses have already been discussed in Chapter 6, so we will focus here mainly on the templatic nature of the process and how to account for it, referring, when necessary, to the data and discussion in Chapter 6. 7.1.1

Onset + Nucleus Reduplication

This is perhaps the most common pattern of reduplication, involving the copying of the onset and nucleus of the stem. In this case, no stipulations about the internal structure of the nucleus need be made, the result is simply a copy of the structure of the original nucleus, whether long or short, following the representation in (1). 1.

a

A O

R N 178

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Templatic Morphology

Templatic Morphology 179 Examples of the implementation of this copying are provided in (2).

A

o

Long Nucleus a

A

O

R

R Copying N

N + cuqhta

A

O

+ taaqhtinXwePin

A

R

o

R

N

I

N

i i

c c.

u

Concatenation

A

+ cuqhta

t

a a + taaqhtinXwePin

cucuqhta

taataaqhtinXwePin Tier Conflation This pattern permits either a long or a short vowel within the nucleus, depending on the weight of the original root vowel. 7.1.2

Onset + Long Nucleus Reduplication

The difference between this pattern and that described above is that this one requires that the copy be long, regardless of the length of the original vowel, whereas in the previous case the copy vowel will only be long if the original vowel is long. This pattern may be represented by (3) and examples follow in (4). 3.

A O

R N

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

a. Short Nucleus a

180 Linguistic Theory and Complex Words a. Short Nucleus

Long Nucleus a

u

A

O

A

R

N

+ suuhinqh

+ PumiiPak a

A

A

O ! !

1

!

1

?

Copying

N

N / \ u u + PumiiPak

1

s

PuuPumiiPak

R 1 N

Concatenation

/ \ u u + suuhinqh

suusuuhinqh

Tier Conflation

Diagram (4a) demonstrates the application of the template to roots with both a short nucleus and a long one. Diagram (4b) shows the template, after copying, being concatenated with the original base from which it was copied. Finally, (4c) is the result, after tier conflation/ bracket erasure. In both the case of the short root vowel and of the long root vowel, both copies end up having long vowels. 7.1.3

Full Syllable Reduplication

The final form which reduplication may take is that of full syllable reduplication, a pattern applying to certain aspectual categories such as repetitive and iterative (for further discussion of the aspectual categories see Section 5.1.3), involving a copy of all the material in the first syllable of the root, as presented below:

A

A

N

C

In all such cases an onset must be present, and must obey the constraints on the shape of onsets, but the rhyme must be left underspecified and

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

4.

Templatic Morphology 181

6.

a. citcits DUP-cit-s

'dodging to one side' SUF-sidewise-lTER [F]

b. paaXkpaaXka PinkPii 'the fire was glowing' DUP- paaXk -(y)aY ?inkw =Pi* REP- glow-DUR fire =DEF c. (Mmskcimskquu DUP- cimsk -quu

'it was always like if whistling' REP- make a whistling sound -3.CND

The application of the syllable template to the data in (6) will result in the forms in (7-9).

7. a.

°

A O

R / \ N C

Copying cit(s)

b O

A A R

N

a.

C

i i

i i

c i

t

+ Sits Tier Conflation

citcits

c. 8.

Concatenation

a

A

0

Copying

R

A

+ paaXk(a)

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

depends crucially on the shape of the rhyme of the original. The weight of the nucleus will depend on that of the nucleus of the underlying form. Examples of this pattern of full reduplication are provided in the following.

82 Linguistic Theory and Complex Words b.

a

A

0

Concatenation

R

1

A

A

pa a Xk paaXkpaaXka

c. K

+ paaXka

a.

Tier Conflation

a

A

0

Copying

R

A

+ cimsk(quu)

A O

R

A

6 i msk c.

dimskc'imskquu

Concatenation

+ dimskquu Tier Conflation

Note that (9) pushes the possibilities for consonants in the coda to its full limit, given the structure of the syllable discussed in Chapter 2. One class of case that has not been presented above is that of an open syllable root, which is not uncommon in Nuuchahnulth. The characterisation of such cases poses certain complexities for this pattern and its implications will be discussed in the following section. 7.1.4

Remaining Issues

Up to this point we have maintained that this pattern of reduplication is a full syllable, but on most interpretations of 'full syllable' this would not be exactly accurate. Note that, in the event of an open syllable root, the pattern employed will be one of those discussed more fully in Section 7.2 on fixed segmentism, illustrated in (10).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

!A

Templatic Morphology 183 10.

A A R

N

C

A

I x

So any instance of full syllable reduplication applied to an open syllable root will result in the phonological infixation of -X-, creating a coda. A necessary concomitant of this, presumably associated with the 'lightness' of the syllable, is the lengthening of the root vowel, as shown by the branching in (10). The actual mechanisms concerning this will be discussed in Section 7.2.2, but the point here is that the template is actually more constrained than the typical expectation for full syllable (see, for instance, McCarthy & Prince (1995). In this template the coda must necessarily be filled, and so this must be stipulated. Previous characterisations of this pattern are rather vague and in certain cases equivocal. To begin with, Rose (1981: 271) has the following to say with respect to this pattern: The principle [sic] iterative allomorphs are distributed as follows. One syllable Vj, V2, Ai, A2, and T roots which can be affixed by -si(X) MOM and are not otherwise reduplicated (i.e. due to presence of derivational suffix inducing reduplication) are affixed by a CVj * 'h# reduplicative morpheme (abbreviated as CVC#) plus the suffix -s SPOR. The X in the reduplicative affix precedes a reduplicated CV- root. This characterisation is inaccurate in several respects. First, the pattern of Jc(C)l CVj \# is insufficient to explain all of the cases, in particular, those involving three consonants in the coda, as in (6c). Secondly, the stipulation concerning the appearance of X fails to capture the true nature of the template, which, as we have seen, is one that requires a filled coda. Thus the generalisation concerning this pattern is lost. Nakayama (1997a: 29), in attempting a description of this pattern, states:

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

O

184 Linguistic Theory and Complex Words

This characterisation again fails to note the effects arising with an open syllable root, referring only to the 'whole root', although it does account for the other cases, by virtue of its vagueness. Finally, Davidson (2002:210) has only this to say about this pattern: We also find types of reduplication in plural marking that do not occur as affix-associated templates, full reduplication (symbolized here as [FR]) and full reduplication with lengthening of the original initial vowel (symbolized [FR+L]). Full reduplication appears to occur only with a few selected monosyllabic roots. The statement that this pattern only occurs with a few selected monosyllabic roots is inaccurate: it occurs with all applicable roots. Davidson does not provide a characterisation of his [FR], and so it is difficult to say more about this description of the facts. The only examples he provides are of CVC roots. Clearly, there has been no accurate characterisation of this template, and it constitutes a rather troubling challenge to prosodic morphology, since it does not easily fit within any straightforward prosodic category, given the requirement that open syllables be modified to fit the syllablebased template. The actual template representing this pattern must be as in (11), where at least one branch must appear in the coda, rather than that in (5), which understates the importance of the coda. This characterisation results in a rather odd template for a syllable-based pattern. 11.

a

A

A

N

C

I Another important point to note, as discussed in Chapter 2, is that the only moraic consonants in Nuuchahnulth are sonorants, much as in the

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

The REPETITIVE morpheme -(y)a (lyl occurs following vowels) is accompanied by reduplication of the whole root and GRADUATIVE aspect (i.e., lengthening effect on the root vowel).

Northern Wakashan Kwak'ala (Zee 1988), and even moraic consonants do not participate in reduplicative templates as weight units, such as the CVV pattern discussed in Section 5.1.1.2. For this reason, the template cannot be explained by saying it is bimoraic, a^. Furthermore, this pattern is no more easily characterised by a CV template, as employed by Rose above, for instance. The fact is, this template is very special and deserves the attention of those interested in reduplicative patterns and the nature of templates. Another point of concern for the model of Prosodic Morphology is the necessity of stipulating that there be at least one consonant in the coda. Note the following quote from McCarthy & Prince (1998:319): Under the Prosodic Morphology Hypothesis, templates can impose prosodic conditions, but not ordinary phonological ones — for example, they can require that the plural affix be a heavy syllable, but not that it have the shape vCv, because vCv is not a prosodically-definable unit. This quote highlights the issue that arises here: not only does the Nuuchahnulth template require a prosodic unit, a, but it also makes use of a further stipulation that the coda of a contain a consonant, i.e., C, an ordinary phonological condition in McCarthy & Prince's terms. This type of structure therefore poses a real challenge to the characterisation of prosodic conditions in Prosodic Morphology. We will now move on to the topic of fixed segmentism, which to some extent overlaps with regular reduplication, as we have just seen, although in other cases this may occur without reduplication. 7.2

Fixed Segmentism

In the following sections we will examine those aspects of templatic morphology that require additional templatic specification, involving what is often referred to as fixed segmentism (see Alderete et al. 1997, Kim & Picanco 2003, for a general discussion of fixed segmentism and Kim 2003c for further discussion of this process in Nuuchahnulth). Templates in this class will have the representation in (12), with the coda position filled in by one of the consonants stipulated in the coda position.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Templatic Morphology 185

186 Linguistic Theory and Complex Words

The nature of the choice of specified material will be discussed in the following sections. 7.2.1

CVc

The first case of fixed segmentism that we will examine is that of the copying of onset and nucleus with the consonant Id prespecified in the coda of the reduplicative template.1 13.

cr

A A

O

R

N

C

A I Notice that the nucleus is prespecified as branching, requiring that whatever the length of the original nucleus, the output of this template will require a long vowel. In what follows we will examine cases of both short and long original nuclei. Examples of this from the data include the following: 14. a. Short Nucleus

Long Nucleus

o

a

oA y\^ N

Ay \ ^ C

N

A I c

Copying

C

A I Xiya

c

PiiqhaSi^aX

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

12.

Templatic Morphology 187

b.

a

A O

A

N Xi c.

Concatenation

N

C

A I

R /\ C

A I

+ Xiiya

1 c

?i

XiicXiiya

i

c

+ Piiqha£i£aX

As is clear from this example, the PiicPiiqhaCi^aX copy template must Id preTier have Conflation associated with the coda position of the template, which is then preposed onto the base. 7.2.2

CVX

The next type of coda-prespecification involves the use of /X/, and may be represented by the following templatic structure: 15.

A

A

N

C

A I x Again we can see that the nucleus is stipulated to be branching in this case. The implementation of this type of reduplication follows the procedure in (16), with examples of both long and short underlying nuclei. Long Nucleus

16. a. Short Nucleus

a

a

A

A A

O

O

R

N

Copying

A

N

C

A J

R

1 (0*

k iya

A

suuya

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

A

188 Linguistic Theory and Complex Words

a

A A

O

A

R

N

0

c.

k"i

!A !A 1

C

A I

Concatenation

R

+ riiya

suu

i X

X

suuXsuuya

+ suuya Tier Conflation

Note that k"iiXk"iiya in this case, as with 7.2.1, the copy vowel is always long. 7.2.3

CVt

This instance of fixed segmentism may be classified into two sub-types: (a) with reduplication; and (b) without reduplication. We will examine the (a) case here and the (b) case in the later section on infixation. The (a) class, with reduplication and coda-pre-specification of Itl, may be represented by the following templatic structure:

a

17.

A A

O

R

N

C

Examples of this type of reduplication that, unlike the two preceding cases, does not require lengthening of the nucleus, as indicated in the template, include the following: Long Nucleus

18. a. Short Nucleus a

a

A A

O

A A

R

N

O

C *

R

N + nusPik

Copying

C t

+ haaraaX

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b.

Templatic Morphology 189 a

a

A A

O

N

A

N

C

n u c.

A

R

+ nusPik

nutnusPik

ha a

Concatenation

C t

+ haaraX

haathaaraX

Tier Conflation

Note that this pattern is directly related to the pattern of full a reduplication discussed above. It applies in those cases where the basic form contains an open syllable, the input of which is insufficient to satisfy the requirements of full a reduplication. The repair for this violation is to employ a default consonant in the required coda position. Normally, one would expect ? as the default, but [+ Glottal] segments are prohibited in the coda, as discussed in Chapter 2. Therefore, an alternative, in the form of a +ant obstruent is employed. -cont 7.2.4

Foot Reduplication with Fixed Segmentism

There is a rather interesting plural template, quite limited in its distribution and perhaps only vestigial given the number of examples extant in the data, which involves the filling of a foot-based template with material from the original monosyllabic root, together with a fixed segment associated with the onset of the second syllable of the foot. It should be noted that there are other foot-based templates however, particularly one to be discussed in Section 7.6. Examples of this template are provided in (19). 19. a. caPaaxcax^ePi DUP- dax" -(y)aY =?r

'the spearers' PL- spear -DUR =DEF

b. tuPuuxtux^ePi DUP- tux -(y)aY =?iY

'the jumpers' PL- jump -DUR =DEF

c. waPaacwacaXquu DUP- wac -(y)aY-'aX -quu

'they were wagging their heads' PL- wag the head -DUR-NOW-3.CND

d. maPaaimaiawePin DUP- mai -(y)a* -we'Pin

'they were wriggling about' PL-move-DUR-3.QT

All the forms above involve copying from a CVC root; in the data there is always a short vowel. Both the consonants and the vowels vary,

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b.

although no examples with l\l have been found to occur so far. The interpretation of the result of this pattern seems to be both plural and, perhaps, repetitive, the output being either nominal or verbal. Note that this is similar to repetitive full reduplication, but combined with fixed segment ? and a templatic foot structure. The template appears to be a full foot, with prespecified nucleus weight and ? in the onset of the second syllable: 20.

*

The matching of the vowel quality in the two nuclei may be attributed to the transparency of ? in such contexts and concomitant spreading of vowel features across it. The implementation of this pattern proceeds as follows: 21. a.

Copying

+ cax<0(ePi)

Concatenation

x10 + dax^ePi

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

190 Linguistic Theory and Complex Words

Templatic Morphology 191

caPaaxdax^ePi

Tier Conflation

There are a limited number of cases employing this pattern but there are sufficient to warrant further investigation and this initial proposal for their treatment. In fact, there is a further foot-based pattern, or perhaps sub-pattern, attested with only a single, high-frequency root, yaqm, the relative pronoun. Examples of this pattern are given in (22). 22. a. yaPaayaqwatqas DUP- yaq* -wa*t -qa*s

'... who were my relatives' PL- REL -related to -ls.REL

b. yaPaayaqhyuqas iuucsaamiih 'my women relatives' DUP-yaq^-hyu^-qa^s iuucsaamiih PL- REL -related to-ls.SUB women c. yaPaayaqPaaqXPitq DUP- yaqw -PaaqX -Prtq

'those who would go along with him' PL-REL -INTENT-3S.REL

d. diiqaaPaX yaPaayaqqas diiqaa -'aX DUP- yaq0* -qa*s

'those with me in the crew chanted' sing secret chant-NOW PL-REL-ls.SUB

It has the following reduplicated plural foot template: 23.

O

Given that this particular pattern is only seen with one root, and that it appears to be in complementary distribution with the preceding one, it seems fair to conclude that it is a sub-pattern of that one, perhaps representing a transition or idiosyncrasy. We will forego the presentation of the implementation of this template, assuming the same procedure as for the preceding cases. As is obvious from these templates, it is important to make available the full range of possible templates, including those with fixed segments and those beyond the level of the syllable, in dealing with Nuuchahnulth templatic morphology.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

c.

192 Linguistic Theory and Complex Words

Infixation and Templatic Morphology

Another morphological process requiring the use of a template involves infixation, as discussed in Broselow & McCarthy (1984), for example. In Nuuchahnulth, it is primarily employed in several patterns marking the plural. The template is necessary to ensure the accurate placement of the infix. We will examine each of these patterns in turn. 7.3.1

The -t- Plural Infix

The -t- infix template is identical to the reduplicative pattern with fixed -/- discussed earlier, but does not involve reduplication, simply the modification of the existing root by means of a III inserted in the coda of the first syllable of the foot. There are numerous examples of this pattern, which serves to indicate the plural in these forms, as shown below. 24. a. t'aatnePis fana -t- -Pis

'several children' child -PL- -DIM

b. haatPum haPum -t-

'every kind offish' food -PL-

c. qaatyapt'atqin qayapt'a -t- -'at -qin

'our legs' limb -PL- -INAL -lP.REL

d. tuutpaati tupaati -t-

'tupaatis' ceremonial privilege -PL-

e. naatyaqak nayaqak -t-

'infant children' baby - PL-

The representation of this morpheme necessarily requires a template, since it involves not only the positioning of the -t-, but also the concomitant appearance of length on the nucleus, so that all forms will have a long vowel in the nucleus, whether originally short or long. If one compares this template with the one for reduplication with the fixed segment /t/, as discussed above in Section 7.2.3, one can see that the reduplicative template in (17) does not stipulate a long vowel, whereas the template in (25), representing the present case, does.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

7.3

Templatic Morphology 193

Examples of the application of this template to the base form are provided in (26) below. 26. a.

°

Alignment h a

t

+ ?um

b.

Concatenation h a

a

t

+ Pum

haatPum

c.

Tier Conflation

There are numerous examples of this pattern, discussed in Chapter 5, and it is a clear example of the need for a fixed segment template. 7.3.2

The -£- Plural Infix

The -y- infix template does not involve reduplication, but a rather complex modification of the existing mono-syllabic root to a bi-syllabic foot structure with -y- in the onset of the second syllable of the foot. This is reminiscent of the foot-based templates discussed above in Section 7.2.4. There are numerous examples of this pattern, which serves to indicate the plural in these forms, as shown below: 27. a. cayaaxuk -y- caaxuk

'swift ones' PL- swift

b. cuyuusuk -y- dusuk

new ones PL- new

c. cayaapac -y- dapac

'canoes' PL-canoe

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

25.

194 Linguistic Theory and Complex Words

d. SayaakPi -y- fak" =Pr

Note that regardless of the length of the original nucleus, the result is always the same: the template for representing these forms requires a short vowel in the first syllable and a long, identical vowel in the second syllable of the foot, as in (28). 28.

*

Notice the similarity between this template and the foot-based template described earlier. However, there is one difference: the coda consonant is not specified in this template, and as one can see below, its attachment depends on the presence of a coda in the original form. Examples of the application of this template are provided in (29) and (30). O 29. a.

Alignment

dapac b.

Concatenation

+ pac c.

dayaapac

'canoes'

Tier Conflation

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

'the knives' PL- knife =DEF

Templatic Morphology 195 30. a.

Tak01 + Pr

Concatenation

r c.

TayaakPi 'the knives'

+ Pr Tier Conflation

The example in (29) provides an example of a form without a coda, while the case in (30) demonstrates the case where there is an existing coda in the form. In both cases the original syllabification is maintained. This form of infixation goes beyond the usual expectation of an infix, involving as it does a foot-based template, containing the infix, which transfers the melody of the original base to the template and then realigns it with the circumscribed remainder of the base. 7.4

Hypocoristic Formation

In Nuuchahnulth, there is a strategy for forming hypocoristic names, or terms of endearment, from the regular form of a name by a combination of truncation, vowel mutation and affixation.3 The nature of this formation is highly suggestive of the type of morphology described by many linguists, for instance Anderson (1988), as subtractive. In this section, however, we will show that what actually occurs is a pattern of templatefilling based on the prosodic structure of the language. We will argue that the building of hypocoristic forms is, in fact, highly reminiscent of

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Alignment

reduplicative strategies employed in this language, as argued for above. The difference is that reduplication subsequently concatenates with the structure from which it has drawn, while Nuuchahnulth hypocoristic formation, henceforth NHF abandons the remainder of the original structure, retaining only the copied portion required for the template. As already discussed in Section 3.6, names in Nuuchahnulth are morphologically complex, containing a single root as in all words in the language, and from one to theoretically any number of suffixes. Examples are as in (31) below. 31. a. tux^siX tux10 -siX b. PanimiyisPaqs Pani -mai -'is -Paqs

'He-jumps' jump-MOM 'Woman-going-alone-on-the-beach' alone -move -on the beach -woman

As can be seen from the examples above, names are usually analysable by speakers and the meanings are typically transparent. In what follows we will not analyse the names except when it is necessary for the argument at hand. In forming the hypocoristic, normal names are transformed in a predictable fashion into hypocoristic forms such as in (32) below. 32. Hypocoristic Form heepPis k°eec?is hoohPis XeehPis

Regular Name hapuui k°aacaapi huuhink°ap XihaPa

There are several changes that forms undergo in order to arrive at the final output: names are shortened, suffixed by -?is, and the root vowels are mutated and, if not already long, lengthened. On examining further instances of NHF it is possible to arrive at a tentative description of the facts such as the following: 33. Hypocoristic Formation: Truncate the name immediately before the second vowel, changing the remaining vowel to mid, either front or back depending on the roundness of the original vowel, lengthen the vowel if not already long, and affix -?is. This.characterisation accounts for all of the following instances of NHF, including those with one (34a), two (34b), and three (34c) consonants in the coda of the derived form. Notice that these examples include a case

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

196 Linguistic Theory and Complex Words

Templatic Morphology 197

34. a. PookPis hooyPis moowPis

Puuk^iinuPa huuyaaimisPaqs muwadath

b. k^eespPis meetk^is meeksPis

k^iispisiis maatk^aayapsiii maaksisaftap

c. Teemt.sPis weestqPis

fimt.siPat wastqaPa

d. seepPis tee^is

sapir David

Note that this process of truncation does not appeal to morpheme boundaries at all, as can be seen by the following examples where either more than a single morpheme is involved (35a), or less than a single morpheme (35b), or an unanalysable form, as for 'Sapir' in (34d) above: 35. a. weestqPis < was-tqa-?a ioo£h?is < Xu£-ha-?aqs b. t'eetPis < t'at'atwaq-'iih moowPis < muwac-'ath

(hiPiiXiik) is coiled-under-on rocks wife-buy -woman speckled trout-hunter of deer -tribe

A description such as that in (33) accounts for a large number of the cases involved, but it also misses a number of cases as well as a generalisation about the nature of the entity that represents the hypocoristic. In Swadesh's unpublished account, he divides the cases up into a number of types. The criteria he chose were: (i) simple name versus reduplicated name; (ii) suffix /-Pis/ or /-is/; and (iii) first medial cluster ending in /m,n,y,w/. We will see shortly that all of these criteria are a result of the template that is used to create the hypocoristic and there is actually no need to mark various forms on the basis of their shape or environment. Examples of cases that do not fit the description in (33) include those of the following groups: 36. a. deejMs t'ooFis keeXkis

di^uuqshwiiim t'uFaaPaqsup kaXkinuPul

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

(34d) of presumably unanalysable foreign names, 'Sapir' and 'David', which illustrates the productivity of the process.

b. hoofrs

huufriPaqsup

c. dooknis feemyis

dukna fimyaaPa

d. tboxPis

tuuxmis

The cases in (36a) above contain glottalised stops or affricates as the final member of the derived form. Remember that [+ Glottal] consonants may not occur in the syllable coda according to the discussion in Chapter 2. Note that when the affix -?is is attached, only part of it, /is/, appears. Likewise, in (36b) for the laryngealised glottal stop which is also considered [+ glottal] and in (36c) for the laryngealised sonorants. In (36d), not all of the consonant cluster after the first vowel is copied, /ml being [+son] and therefore not allowed after 1x1 in the coda of the same syllable. There is another set of examples that appear to constitute exceptions to the general case. All of these cases involve name forms that are reduplicated in their normal shape, as in (37). 37. a. feesPis deestPis

YaTacsat'aqs dadastufrih

b. feeyis cooPis

t'at'ayuuPas £u£uuPup

c. t'eetPis

fat'atwafrih

d. keetxPis

katxkaatxilwat

In the first group, the hypocoristic form patterns after the regular case with the exception that it does not involve the prefixed copy of the name. What we would expect to appear would be something like that given in (38), but that is not the case. 38. a. * £ooc?is b. * deedis

cucuuPup dadastufrih

The second case (38b) patterns after the cases in (36a-c) above, which have only /-is/ attached, with the further consideration that the reduplicative prefix is not included, just as in (37). In (37c) we have another example where all of the cluster is not copied, as in (36d), and again the copy is not included, and (37d) is an example of a different type of reduplication where again the hypocoristic form does not appear to be

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

198 Linguistic Theory and Complex Words

based on the full name form but rather on the unreduplicated form. We maintain this based on the fact that the hypocoristic for (37d) does not appear as in (39), which it should if based on the entire name. Note that there is no apparent reason why this form could not arise, given the proper context, on analogy with forms such as those in (36a), but this is not the case for reduplicated forms. We would, however, expect a form such as (39) if there existed a normal name form such as e.g. katxka, where there was no reduplication. 39.

* keetxkis

katxkaatxilwat

This exhausts the various types of seemingly irregular hypocoristic forms. We will now show that the explanation for these widely divergent forms of the hypocoristic rests with the inherent prosodic structure of the language and the constraints that hold in it, rather than with any rampant allomorphy obtaining in the language. 7.4.1

A Unified Account

Taking the previous discussion and our understanding of the syllable structure of Nuuchahnulth into account, let us look at a unified treatment of NHF in terms of the imposition of a template on a melody containing the entire name. We must link, as we would for reduplicative copies, from left to right, one-to-one until all available slots are filled, as in (40). O

40.

O

r\ R

O

N

N

A

h lo arnd

R

1

[+G1]

What this template represents is a maximal syllable, containing an onset, a partially-specified nucleus, and a coda containing as many consonants as available, disregarding morpheme boundaries, but respecting the phonotactic requirements of the language. This is followed by a second syllable whose nucleus and coda are fully specified but whose onset is only specified for [+ Glottal].4 As the vowel of the template is specified

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Templatic Morphology 199

to be long, and since there are no diphthongs in the language, the vowel slot of the original must link to both slots of the template. No consonants will be allowed in the coda if they would violate any of the restrictions discussed previously, for example a coda containing a glottalised stop as in (36a), a glottal consonant as in (36b), or a sequence such as l-xml or /-tw/ as in (36d) and (37c) respectively. Descriptively, there are two means to resolving this possible conflict. One possible strategy is to place the offending member in the onset of the following syllable. This is the strategy employed to repair the forms in (36a-c) and (37b), where the phonemes are [+ Glottal] and therefore may serve in the [+ Glottal] onset of the suffix -?is. The other choice is to drop the offending member of the cluster. This is the strategy invoked to handle (36d) and (37c), where it is plain resonants that are disallowed in the coda due to sonority violations. The dropping of these resonants repairs the sonority violation, producing outputs that do not contain the sonorants. While these two strategies would seem at first glance to be unrelated, when we turn to the theoretical characterisation that involves the linking of melody to template, we can see more clearly what is actually going on here. In the most straightforward case, as in (41), linking is left-to-right, until we arrive at the second vowel of the melody, which has no available slot within the template. The partially-specified C-slot, stipulated as [+Glottal], is filled in with the default [+ Glottal] consonant, Rl, unspecified for place of articulation, and we arrive at the final form, weestq?is. 41. a.

o

O

r\ R

O

R N

N

A

1

hi lo

[+G1]

ocrnd wastqaPa

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

200 Linguistic Theory and Complex Words

Templatic Morphology 201
b.

R

N

N

A w e e

c.

R

O

s t q

f+Gll

weestqPis

This will be the same for all the regular cases such as those in (34). But what about the other cases? First, let's take a look at those cases which involve glottalised obstruents, such as those in (36a). Linking proceeds left-to-right until we reach the glottalised consonant. It is not allowed to link to the coda of the first syllable since all these slots are specified to be [- Glottal], as noted in Section 2.3. But it can link to the onset of the following syllable, which is specified to be [+ Glottal].5 This fulfils the templatic requirements, giving us the final output in (42). 42. a.

arnd dipuqhwiiim b.

c.

dee^is

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

O

Now what about the cases involving plain sonorants following obstruents in the coda, such as in (36d), t'uuxmisl We proceed as in all the other cases, linking from left to right until we arrive at the coda position where the Sonority Hierarchy rules out sonorants following obstruents within the coda.6 The next position, i.e. the onset of the following syllable does not prohibit sonorants but it specifies that whatever links to it must be [+Glottal] and /ml is not. Therefore, the form appears as t'oox?is. 43. a.

.*

C V V

tuuxmis b.

t' o o c.

t'ooxPis

This template would predict that it is permissible to have a laryngealised sonorant occur in this position, since it would be [+ Glottal], and thus, not subject to the same constraints as a non-laryngealised sonorant. We can test such cases, using data such as that in (36c). As can be seen in (44), the laryngealised sonorant, which follows an obstruent, may be syllabified into the onset of the second syllable, since it meets the requirement of being [+ Glottal], Thus, there is no need for the default insertion of Rl, since the position is filled with an existing consonant, as shown in (44b).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

202 Linguistic Theory and Complex Words

Templatic Morphology 203

C V V otrnd dukna b.

c.

dooknis

This mechanism explains the apparent disappearance of the glottal stop in the suffix -?is in cases like those in (36) and (37): the position is licensed by an already existing [+Glottal] consonant that links to it, giving the effect of coalescence or deletion. The morph /Pis/ which appears in the right branch of the template for the hypocoristic form, is identical to a regular morpheme of the language, -?is DIM (cf. Section 4.3.8). Schematically, -?is DIM is an inflectional morpheme which appears after all derivation, explaining why no further derivational morphology may occur after the affixation of -?is. It may be that these two forms are one and the same. 7.4.1.1

Reduplication and Hypocoristic Forms

Hypocoristics formed from names that involve reduplication have an interesting shape: they contain no reduplication. This is understandable given that the template only contains room for one unspecified syllable, but one might imagine a different outcome. For instance, the name could be based on the melody of the entire name, having a form such as that in (38) or (39). But these are ungrammatical.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

44. a.

In order to arrive at the proper output for the hypocoristic forms of reduplicated names the copy must be made off the melody of the base and not off the copy or the sequence of copy + root. This can be seen in the examples of reduplicated roots given in (37). Because of this, we argue for the derivation of hypocoristics directly from the melody of the original name, rather than off the already syllabified structure of the word. Thus, the input to the template will be just the original melody, as in (45). 45. a.

(da) dastuTiih b.

c.

deestPis

Similarly, the other reduplicants will be formed off their underlying melody. There are, however, two forms in the data that pose further difficulties. These are the structurally similar forms 6eeX?is < caXdaaya?uX and j3eeA?is
10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

204 Linguistic Theory and Complex Words

extends to the formation of the hypocoristic form, which is not reduplicated. In conclusion, we have seen in this section that it is possible to have a formation that operates surprisingly like reduplication yet with a very different outcome: only the copy is saved and the remainder, including the original form, is discarded. This linking of melody to template follows regular patterns seen in many languages, including Nuuchahnulth. 7.5

Vowel Length Patterns

Patterns of vowel lengthening and vowel shortening have been described in previous chapters, and are associated with various morphological requirements. These patterns may be represented by templates that constrain the output to a certain shape. Thus, patterns requiring long vowels specifically may be characterised by the following template: a

46.

R N / \ This would ensure that, irrespective of the weight of the original nucleus, the output will be a long vowel. This template would be associated with the left edge of the word and would apply to the first root vowel. Templates requiring a short vowel would apply in like fashion, ensuring that the root vowel is non-branching, as in (47). 47.

o R N

I Obviously a fully fleshed out template might contain additional information, but the above is sufficient to illustrate the point at hand: the use of templates to ensure the requirements of length imposed by the morphology.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Templatic Morphology 205

206 Linguistic Theory and Complex Words

Abnormal Speech

It seems only fitting to include a discussion here of the various special speech types encountered in Tsishaath, a phenomenon first noted by Edward Sapir in a short monograph entitled 'Abnormal Speech Types in Nootka' (1915). This phenomenon is not unique to Nuuchahnulth, but is special enough to warrant some discussion and exposition. It involves various permutations of conventional phonemes in order to convey certain, extra-linguistic information to the listener.7 This phenomenon appears in several different environments, notably: (i) as an indicator of the speech of certain myth characters; (ii) as a marker of certain characteristics of humans; and (iii) as a marker of different tribal groups within the Nuuchahnulth people. We will examine only one instance of this pattern in this section. There are a number of mythical characters in Nootkan culture, each with their own special characteristics. In addition, many of them have their own personal speech characteristics, realised by the application of various phonological rules of a basically transformational nature. These mutations are used to represent the speech of such characters in storytelling, where the skilled narrator will provide the appropriate changes in order to indicate which character is speaking at any given time. One example of this class of cases is that of Pitch Woman, a mythical character who lives in the woods and steals children away from their families. Her characteristic speech form involves the quantitative and qualitative ablaut and concomitant breaking of the vowel of the initial syllable of words. For example, in (48) below, the word Xul 'good' is changed to Xeewui, with the insertion of a long [e:] followed by a continuant that is homorganic with the following original vowel, while haakwaaX is changed to heeyikaaX in a similar fashion. 48.

Pee XeewuisaTas heeyik°aX 'Oh, I seem to be a beautiful girl' Pee Xui -saY
Further examples in (49) confirm this strategy for representing the speech of Pitch Woman. 49.

XeewuiePic Xui -rrie-Pic

'you are handsome' attractive -2S.IND

heeyita?atuPi hita -Ca-Patu -'r

LOc

' come down!' -move down -2s.lMP

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

7.6

heeyiwiiaX haawiiaX q^eeyiPajtat q^aa -'ap -'at

young man young man 'treated that way' thus -CAUS -PASS

weeyikiiah wik -ii -marh

'I didn't do it' not-do-ls.lND

PeeyiqinPa^aanithak Paqin -'ap -'at -(m)it -ha*k

'how was it done to you?' what? -CAUS-PASS -PAST-2S.INT

On occasion, part of the procedure is omitted, either intentionally or accidentally, as in the following case where, perhaps due to the reduplication, the ablaut has been forgone. 50.

XaXawuiuk°ah Pe?eeyi§dip 'I have nice gum' DUP- Xui-uk-maYh Pisciip PL-good-NOM-ls.lND gum

This kind of template, provided in (51), occurs in a number of cases. 51.

cons

The nucleus of the first syllable is always a long front mid vowel, /ee/, and the onset of the second syllable is always a glide that agrees with the following nucleus in rounding, thus either /w/ or lyl. O 52.

cons

XuisaaTas

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Templatic Morphology 207

208 Linguistic Theory and Complex Words O

O

R

R

I

N

N

A

u

-cons + hi + rnd

e e

c.

i

(saa^as)

XeewuisaS'as

A similar analysis could be applied to the speech of other characters with various abnormal speech patterns, as discussed in Sapir (1915) and Stonham (1999b). 7.7

A Final Word about Foot-based Templates

In the preceding sections we have encountered several foot-based templates: (i) a fixed segment plural template (53a); (ii) a -/- plural infix (53b); (iii) a hypocoristic template (53c); and (iv) the abnormal speech of Pitch Woman (53d). 53. a. (=20)

b. (=28)

O

a O

R I N

O

N

A

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

o

Templatic Morphology 209 O

c. (=40)

R I N

A

[+G1]

d. (=51)

As can be seen from these four templates, there is much in common: (i) three of them require a [+Glottal] element in the onset of the second syllable; (ii) all of them have a light/heavy alternation, although (53a,b) are iambic and (53c,d) are trochaic; (iii) all of them lack a coda in the first syllable. Furthermore, two of them require a coda in the second syllable (53a,c). These foot-based patterns have not previously been noted in the literature and they present new and interesting data supporting the importance of the foot in Nuuchahnulth grammar.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

O

8

In this chapter we will investigate those phenomena that deal with the interaction between morphology and syntax. This includes the topics of (i) clitics, those phonologically-dependent morphemes which are attached in the syntax, (ii) incorporation, which involves the movement of the object into the verb, (iii) compounding, which, in Nuuchahnulth occurs only post-lexically, (iv) classifiers, which have a direct bearing on the syntax, and (v) arguments for an inflectional clitic cluster. We begin with an examination of the traditionally accepted clitics. 8.1

Clitics

Clitics are generally assumed to be morphemes that straddle the boundary between morphology and syntax. They are phonologically dependent but syntactically independent elements. Spencer (1991:350) has the following to say about the nature of clitics: Clitics are elements which share certain properties of fully fledged words, but which lack the independence usually associated with words. In particular, they can't stand alone, but have to be attached phonologically to a host. For Nuuchahnulth, Sapir & Swadesh (1939:241) present three clitics, without further discussion, described by them as 'enclitic suffixes'; they are illustrated in (1). 1.

Xaa ?ai Paaia

'again' PL argument 'always'

However, from an analysis of the phonological properties of a number of suffixal morphemes, it would appear that there are in fact additional

210 10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Morphosyntax

clitic-like elements. Swadesh (1933:19) mentions the same three morphemes as above, but alludes to the existence of 'a few others whose definition presents some complexities.' Foremost among the additional clitics would be the definite marker, ?i\ which occurs quite pervasively in the language. This would give us at least the following examples of clitics. 2.

?ai Xaa Paaia Pi-

PL argument 'again' 'always' DEF

We will examine tests for establishing clitic-hood in the next section. 8.1.1 Tests for Clitic-hood The proof of the clitic status of these elements rests with the standard tests for clitic-hood as described in the literature. With respect to this, Zwicky & Pullum (1983:503-4) is particularly relevant. They propose the following criteria: 3.

a. Clitics can exhibit a low degree of selection with respect to their hosts, while affixes exhibit a high degree of selection with respect to their stems. b. Arbitrary gaps in the set of combinations are more characteristic of affixed words than of clitic groups. c. Morphological idiosyncrasies are more characteristic of affixed words than of clitic groups. d. Semantic idiosyncrasies are more characteristic of affixed words than of clitic groups. e. Syntactic rules can affect affixed words, but cannot affect clitic groups. f. Clitics can attach to material already containing clitics, but affixes cannot.

We will now examine Nuuchahnulth clitics with regard to these properties. The first property can be demonstrated by the placement of the definite marker, ?r, which can be shown to attach to virtually any category of syntactic word, as shown in (4). 4.

a. nayaqakPi nayaqak =Pi*

'the baby' baby =DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Morphosyntax 211

b. XaahmaiPi nayaqak Xahw [L]-mai =?iY nayaqak

'the newborn baby' recently -born [L] =DEF baby

c. PiihPii k°atyiik muksyi Piih0* =?i* k°atyiik muksyi

'the very heavy stone' big =DEF heavy stone

d. PaXePi jMsaq maatmaas PaXa =Pr jMsaq maatmaas

'the two bad tribes' two =DEF bad tribe

e. naacukPi naacuk =?r

'the one who was looking for him' look for =DEF

The example in (4a) shows ?v attached to a noun, (4b) to an adjective, (4c) to an adverb, (4d) to a numeral, and (4e) to a verb, resulting in a relative clause. These data clearly demonstrate the category-insensitivity of?r. With regards to (3b), there are no gaps in the usage of ?r, nor does it exhibit any morphological (3c) or semantic (3d) idiosyncrasies. Syntactic movement rules may affect words, but never move only the ?i\ With respect to attachment facts, the only thing which can attach to the outside of ?r, is one of the other elements listed in (1), for instance, the morpheme ?aaia 'always' which consistently occurs outside of ?v, demonstrating its clitic status, as the following examples: 5.

a. PaPayapaiatPiPaaia '...who has many always doing ...' DUP-?aya-pai-'at=?r=Paaia SUF-many-several..[R]-PASS=DEF=always b. wawee?in?i?aaia waa -we^Pin =Pr =?aaia

'one about whom it is always said...' say -3s.QT =DEF =always

'the ones who were always drumming' c. nasqaamitPiPaaia nasq -(y)av-(m)it =Pr=?aaia beat (time) -DUR-PAST=DEF=always Another property of ?aaia is its rather non-inflectional contribution to the semantics of the phrase within which it occurs. The morpheme ?al PL is interesting because of the nature of the relationship that it bears to elements in the syntax. It does not appear to mark as plural the element to which it is attached, but rather the agentsubject that acts upon that element. It also attaches outside of ?i\ 6.

a. qaiaatikPisukPiPai qaiaatik -Pis -uk =?i* =Pai

'they ... to their younger brother' younger brother -DIM -POSS =DEF=PL

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

212 Linguistic Theory and Complex Words

Morphosyntax 213 b. AuwiiqsakPiPai nuwiiqsu -'ak =Pi* =Pai

'they ... to their father' father -POSS =DEF =PL

Finally, the morpheme Xaa 'again' also makes a special contribution to the meaning of the syntactic whole rather than the element to which it is attached. For instance, when attached to a member of a noun phrase it affects the meaning of the noun phrase, when attached to a verb it affects the meaning of the sentence. As can be seen in (7), it attaches outside of 7.

a. Piihtuup XaPuuPiXaa tupaati 'still another tuupati was a whale' Piihtuup Xa?u*=?i-=Xaa tupaati whale other=DEF=again privilege b. hasaqsuiPiXaa has -Paqsui =?f =Xaa

'who has a strong voice again' loud -at the mouth =DEF =again

8.1.2 Order of Clitics With respect to the order of these clitics, no more than two clitics have been found in a sequence in the data, but all clitics can be found ordered with respect to all others, resulting in the following order: 8.

BASE = ?f > Xaa > Paaia >Pai

Quite clearly all of these elements behave in clitic-like fashion. They are frequently semantically tied to a different target from their phonological one, are categorically insensitive, and always occur outside of paradigmatic inflection, as discussed in Chapter 4. We will encounter arguments later in this chapter for admitting a larger set of morphemes to clitic status, but regardless of one's views on that argument, it seems clear that one must at least admit the morphemes in this section to clitichood. 8.2

Incorporation

Another area where morphology and syntax interact in Nuuchahnulth is incorporation. Incorporation in Nuuchahnulth involves the combination of a root with the bound verb that acts as the head of V. Given the prohibition on compounding in the language (to be discussed in Section 8.3), this means that only such bound verbal roots, as discussed in Chapter 3,

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

c. iuu£muupuktoit?i?ai 'they...w.r.t. their younger sister' iuucmuup -uk-(m)it =Pr =Pai sister -POSS -PAST =DEF =PL

may be involved in incorporation. Bound verbal roots typically have free alternants that appear in non-incorporated elements. Some alternants are syncretic, that is, without obvious overt phonological relationship to the bound form, and others involve the use of the free root ?u, which acts as a semantically empty expletive base to which to attach the bound verbal morpheme. This distinction has been discussed in more detail in Section 3.4.4. Example (9a) demonstrates the use of the expletive base, ?u, in combination with the bound verb, -na"k°i, to form the verbal construct 'have'. (9b) shows the incorporated form without an expletive base. 9.

a. PunaaksiPaX t'anePis Pu-na-k^-siX-'aX t'ana-Pis

'and then she got children' REF-have-MOM-NOW child-DIM

b.faAanak

J'he/she had a child'I [ she gave birth J child -have

t'ana -navk°

Incorporation of the heads of various constituents is a benchmark for deciding on the polysynthetic status of a language. In this respect, Nuuchahnulth bears all of the expected properties of such a system. 1. Presence of object within matrix verb 2. Generic reading of incorporated object 3. Change in transitivity of verb For the most part, this description may be applied ceteris paribus to Nuuchahnulth incorporation. Take, for example, the following instances: 10. a. PuusimC PaniXii Tuyiiyip PuPakXiqiiyi^aX Pu -sim£ PaniXii Suyi -iiyip Pu -PakXi -iiyip -'aX REF -train for [L] so that medicine-obtain REF-at last-obtain-NOW 'he had prayed to get medicine that he at last got it' b. kamatuk°ahXaa Siiuyisimcyak kamatuk -(m)ah =Xaa Tuyi -sim£ -yak*0 know -ls.lND =also medicine -train for [L] -tool 'there is also known to me (a ritual) for training for medicine' c. huucuqXisPathsimda^aXat huu£uqXis -Path -simc -'ap -'aX -'at Uchucklesit -tribe -train for [L] -CAUS -NOW -PASS 'they had him do ritual for getting Uchucklesits'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

214 Linguistic Theory and Complex Words

As can be seen from these examples, incorporation is an optional process that, in its standard form, may be realised in Nuuchahnulth by a lexical incorporation of the object noun, resulting in a generic, non-specific conjunction of verb and object. However, this is not the only form that incorporation may take in the language. In many more cases, an element is extracted from its position within the object phrase and is incorporated into the verbal construct. The fact that previous accounts of incorporation have found that the result of incorporation is a generic object without modifiers, for instance, may be explained by a crucial distinction between two very different processes, one morphological and one syntactic: (1) lexical incorporation, which bears many similarities to compounding, occurs within the lexicon and produces bare, generic words containing verb + argument; (2) syntactic incorporation, involving movement of an element of the object noun phrase into the verbal complex. What we intend to demonstrate here is that, in fact, both types of incorporation may co-exist within the same language. This is not the first such proposal, for instance Baker (1995) discusses Mohawk in this light, but Nuuchahnulth provides a particularly strong case for such a separation. The different types of incorporation involve different results for properties one might expect of such a process, as in (11). 11. a. Is there any coreference to elements outside the predicate? b. Is the semantics transparent or opaque? c. Is the valency of the verb reduced? d. Are subcategorisation restrictions observed? e. Do the verb and incorporee form a coherent semantic unit? In Nuuchahnulth, both of these processes are attested, and the crucial differences between them are clear. The following sections will discuss the various ramifications of these processes in greater detail, in particular, focusing on the syntactic incorporation of elements. 8.1.1 Lexical Incorporation Based on the absence of compounding (for which see Section 8.3) and the assumption that the existence of bound verbal roots allows for the construction of verbal constructs with properties similar to those of compounding in other languages, we propose that Nuuchahnulth makes use

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Morphosyntax 215

216 Linguistic Theory and Complex Words

When the syntax requires the placing of one word inside another the result is a complex word containing two words. This is a kind of compounding. Words produced by incorporation are very similar to compounds produced by derivational morphology. Perhaps lexical incorporation is the closest Nuuchahnulth equivalent to lexical compounding, given the particularly strong polysynthetic nature of the language. 8.1.1.1 Properties Lexical incorporation involves words built within the lexicon and demonstrates properties one might expect of such a process, like the ones in (11) above. In the following sections we will examine each of these properties to clarify the behaviour of these lexical forms of incorporation. 8.1.1.1.1 Coreference. With respect to reference outside the verb, lexical incorporation does not permit such relationships. A form constructed in the lexicon will not be co-indexed with any modifiers in the syntax, as dictated by the Lexical Integrity Hypothesis (Chomsky 1970, etc.). This will not be the case for syntactic incorporation, which will turn out to allow coreference, and this property is therefore significant. 8.1.1.1.2 Idiosyncrasies. Idiosyncrasy is a lexical property found in all languages. Nuuchahnulth is no exception and the productivity of morphological composition allows the construction of many derivatives both transparent and opaque. As Swadesh (1933: 54) states: Although the synthetically constructed words of Nootka are naturally of wide applicability, there are any number of derivatives that have become restricted in application to some special instance of the notional complexes they would theoretically represent.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

of a process of lexical incorporation. Note the following quote from Katamba (1993:282):

Morphosyntax 217

Thus, lexical compositions often have idiosyncratic meanings not decomposable from the parts. Take, for example, the cases in (12), which involve elements other than bound verbs: 'red-headed woodpecker' red -thing

b. Xisai Xis -ai

'blanket' white -on fabric-like surface

c. Piihtuup Piih40 -tup

'whale' big -creature

As you can see, these forms are constructed from the combination of a root plus a nominal or locative suffix, although these are not the only possibilities. Notice that they have acquired idiosyncratic meanings, 'redheaded woodpecker' from 'red-thing', 'blanket' from 'white fabric-like surface', and 'whale' from 'big creature'. Similarly, for combinations of bound verb with object noun, we encounter idiosyncratic meanings in some combinations, as shown in (13):1 13. a. iuchaa iucsma -haY

'take in marriage (for men)' woman -buy

b. capxnaak capxw -na-k0

'take in marriage (for women)' man -have

c. fananak t'ana -na-k0

'give birth' child -have

d. SuSuuk^afin £uk°aa -Tin

'invite' come! -make the sound of

Thus, 'woman-buy' means 'take a woman in marriage' and 'man-have' means 'take a man in marriage', for instance. (13c) is interesting because it actually can have two meanings, an idiosyncratic one 'give birth' and a compositional one 'have (or possess) a child'. The idiosyncrasy of (13d) should be obvious. Examples of the usage of these are provided in context in (14a,b): 14. a. cuu iu£haaXin kuPaiik £uu iucsma -ha- -'aX-(m)in kuPai -(y)ik very well! woman -buy-NOW-lP.iND morning -FUT 'well, we will go for marriage in the morning'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

:. a. Xihma Xih -ma

218 Linguistic Theory and Complex Words

8.2.1.1.3 Valency Reduction. A separate issue in the combination of root with bound verbal morpheme is that of valency. (15a) demonstrates the typical transitive properties of the suffix, -inak 'dance X', in this case, 'dance a Thunderbird dance'. (15b) demonstrates the reduction in valency concomitant with the incorporation of the object into the verb. 15. a. PuunakaXsi t'iickin. Pu -i*nak° -'aX -sit'iickin REF- dance -NOW-lS.ABS Thunderbird 'I performed a Thunderbird dance' (Transitive) b. t'ickatqinak hicstiii hiPiXiik t'iickin -rnak0 hi -stri hiPiXiik 'Thunderbird -dance slither-thru roof Lightning Serpent 'there is a Thunderbird dance with the lightning serpent crawling in through the roof (Intransitive) The incorporation, within the lexicon, of an object into a transitive verb, as in (15b), reduces the verb to an intransitive verb. Thus, we can see that lexical incorporation involves a reduction in valency resulting from the absorption of the argument of the verb. 8.2.1.1.4 Subcategorisation Restrictions. A further issue, again related, is that of subcategorisation restrictions that appear on the verb. In lexical incorporation, we only encounter combinations, typically N + V, indicated by the subcategorisation restrictions imposed by the verbal head.2 For instance, (16a), requiring that the verb take a noun as object, is attested in (13a) above, whereas (16a') is unattested. Similarly for (16b) versus (16b'). 16. a. 7

a. b.

-haY : -M 0

[N _

]v

: * [V _

]v

-na-k :

[N _

]v

b'. -na k° : * [A _

]v

v

Lexical incorporation observes the regular requirements of the head on subcategorisation. If the verb subcategorises for a noun object, then that

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. cuSuuk^afrnmePiXXaa Pamii&PaXPitq DUP- £uuk°a -Tin -ma-'iX =Xaa Parhii5iPaXPitq REP- come!-make..sound-go to[R]-on ground-again the next day 'the next day they again went around inviting'

is the only element that may be incorporated lexically. This point will be significant in our discussion of syntactic incorporation in the next section. One further point about lexical incorporation, illustrated by the word for 'marry (of males)', ludhaa, is that the root morpheme that combines with the bound verb in this case is in its bound form, which, as discussed in Chapter 6, only occurs at Stratum 1 within the lexicon, further evidence that this must be a lexical form of incorporation. 8.2.1.1.5 Coherence of Unit. In lexical incorporation, the incorporee and the bound verb into which it is incorporated form a close semantic unit, which may be more or less lexicalised, as demonstrated in Section 8.2.1.1.2. They also form a unit for purposes of valency and subcategorisation, as discussed in Section 8.2.1.1.3-4. Again, this is not necessarily the case, as we will see now in the discussion of syntactic incorporation. We will now examine the other type of incorporation, which is fully syntactic in nature. 8.2.2 Syntactic Incorporation Syntactic incorporation involves words completed in the syntax. In traditional accounts, these typically involve the head noun of the object being incorporated (Baker 1988), as suggested by the examples in (17), where (17a) shows the unincorporated form of the bound verb -ityaap 'bring as a gift', combining with the expletive ?u, and (17b) demonstrates the incorporated version, with the object inside the verb: 17. a. Puutyaa^aX [[ suuhaa ] N ]Np Pu -ityaap -'aX suuhaa REF -bring...as gift -NOW silver spring salmon 'he brought a gift of silver spring salmon' b. suuhiiitya^at 4

[[ tj ] N ] N P I

suuhaa -ityaap -'at silver spring salmon -bring...as gift -PASS 'he brought a gift of silver spring salmon' In such a case the results are more or less indistinct from lexical incorporation, but there are a few differences, which we will attempt to tease out in this section. Before we move on, we will briefly examine Nuuchahnulth phrase structure, to make what follows more meaningful.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Morphosyntax 219

220 Linguistic Theory and Complex Words Preliminaries

Before moving on to the central issues it will be useful to take a very brief look at those aspects of the syntax which may bear directly on the issues to be discussed herein. Within the sentence, the most obvious factor is the head-initial nature of the language. Verbs occur at the beginning of the sentence and head nouns precede their relative clause modifiers and other complements. This is not to say that nothing may ever precede the verb in the sentence, but those things that do are typically of either a restricted class of function words or are in a special position, i.e. of focus or stylistic fronting: 18. a. [ nuuiimstaqimii ] N P [ Pukiaani ] v [ es ] N P nuuiim -taqimi Pu -kia~ -ni Nuuthlim -band REF -name -lp.ABS 'we were called the Nuuthlim Band' b. [ Taanusmitj ] N P [ PaakuuPatwePin ] v [ ej ]NP Taanus -rhrt PaakuuX -'at -werPin crane -son of borrow [M] -PASS -3.QT 'it was Crane they borrowed it from' [ XaasmaptPi ] N P c. [ Xaqmisj ] NP [ hismkaX ] v Xaq -mis his -ink" -'aX Xaasmapt =?iT oil -thing both -along with -NOW sapling =DEF 'there was oil along with the sapling'

[ ej ] N P

Since the treatment of focus is not directly germane to the issues here, we will not explore it any further, simply assuming that this is a possible outcome in the syntax under the appropriate conditions, to be dealt with by some sort of fronting process. As for the location of the subject and object noun phrases, there is considerable fluctuation in their ordering with respect to each other, so much so that it is difficult to state which is the basic order, VSO or VOS. In fact, either order is commonly found to occur, and this often poses problems for the analysis of the sentence: 19. a. PunaakwePin [ nisrha ] N P [ hawiiuk q^ayadiik ] N P Pu -naYk° -we^Pin nisrha hawii -uk q^ayaciik REF-have-3.QT land chief-POSS wolf

V

o

s

'the chief of the Wolves owned land'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.2.2.1

Morphosyntax 221

k°atyaat [ PaXa haathaak0>aX ]NP k^atyaat PaXa DUP- haak°aX Kwatyat two PL-woman S O sight of two young women.'

Notice that the two sentences (19a) and (19b) provide evidence of the two competing orders for arguments, VOS in (19a) and VSO in (19b). Obviously this issue will be important to an overall understanding of Nuuchahnulth grammar, but the focus here will be on the bare noun phrase, in whatever position it occurs, and its various collocations. The issue of argument ordering, while an important and fascinating aspect of Nuuchahnulth syntax, will necessarily be given little further investigation in this book other than where it is directly relevant to noun-noun collocations. The internal structure of the phrase is somewhat more rigid than that of the arguments of the verb within the sentence. Within the NP, the noun has a more or less central position, being followed by relative clauses and other sentential complements and preceded by adjectives, which may, in turn, be preceded by various quantifiers, including numerals. Adverbs modifying adjectives precede them and the layout of the entire NP is as follows: 20

-

t {QNUMT} [ A D J [ N

[RELATIVE CLAUSE] ] N , ] N , ] N P

The head is the only obligatory element of the NP and many noun phrases consist of only the bare head noun, an important property in what will follow. Note that the position of the head places it away from the edges of the phrase when any of the other elements are present. One other issue of a syntactic nature which will be important because of its bearing on the boundaries of noun phrases will be the status of the definite article or specifier, ?i\ As discussed in Section 8.1, the definite article in Nuuchahnulth is a clitic that moves in the syntax. This property may be utilised as one clear indicator of the left boundary of the noun phrase within which the clitic moves. One final point to be made concerns the status of third person-marking and number-marking. In Nuuchahnulth, there is no third person pronoun, nor is there inflectional marking for third person in many paradigms. In addition, number-marking is optional, and infrequent. Thus, it is possible

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. PuyuPaiwePin Pu -u?ai -we^Pin REF-see 3.QT V 'Kwatyat caught

222 Linguistic Theory and Complex Words

to have sentences without any overt arguments or clear indication of the number of participants, as in the following: qahsaajfoX qah -sa*p -'aX

'they killed them' die -MC -NOW

In this particular context, the sentence is glossed as 'they killed them', but it could just as easily mean: 'he killed them', 'they killed her', 'she killed them', etc. This highlights the potential difficulties that may occur in determining the syntactic arguments in a sentence, particularly in the case of third person participants. 8.2.2.2 Properties With respect to the properties presented in (11), we will see that syntactic incorporation behaves in a very different manner from lexical incorporation, with a much greater involvement with the syntax. 8.1.1.1.1 Coreference. A clear example of the syntactic nature of this type of incorporation is (22), where the head of a relative clause, hatiii 'chief, is incorporated into the verb -narkw 'have', yet is still coreferent with the relative clause to which it is bound. Note the order of head and dependent relative clause provided in (20), indicating the original site of extraction: 22.

hawiijnaksiX

quuPas [[ ti]N [yaTjatPitq

PuPaaiuFat ]] Np

hawii-naYk°-siX quuPas yaq40 -'at -Prtq Pu -Paaiuk -'at chief-have-MOM person REL-PASS-3s.REL REF-look after-PASS 'People have as chief the one who looks after them.' As you can see from this example, the relative clause is bound to its head by a chain linking the incorporated hatiii 'chief with its trace and thereby the relative pronoun. Examples such as this clearly demonstrate the syntactic nature of this type of incorporation in Nuuchahnulth. 8.2.2.2.2 Semantic Transparency. Note the semantic transparency of the predicate, ha&ilnaksiX 'have a chief, in (22) above, compared, for example, with the cases of lexical incorporation in (13). This demonstrates another property of syntactic incorporation, that of the transparency of the relationship between the bound verb and its incorporated element.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

21.

Morphosyntax 223

8.2.2.2.3 No Reduction in Valency. Valency is a significant issue from the point of view of proper government and its concomitants. There can be no reduction in valency upon incorporation of these elements, since the head of the NP may remain in situ and thus must be properly governed (Baker 1988). This must be accomplished by assuming that the verb remains capable of governing its argument and therefore may not be reduced in valency. Consider the following examples: 23. a. PunitPaaqXuk mftaat 'his was to be stocked with sockeye' Pu -nit -PaaqX -uk mftaat REF-stocked with-lNTENT-POSS sockeye b. Payint'aX daastimc Paya -nit -'aX daastimc c. Xuinit Xui -nit

Xudim Xudim

'it is now stocked with many minks' many -stocked with -NOW mink 'which is stocked with good mussels' good -stocked with mussel

Examples such as these will be discussed in greater detail later in this chapter, but for now note the relationship between the bound verb and the element incorporated into it. In (23a) we see the case of the bound verb with the expletive base, ?u, and an unincorporated object, mfiaat 'sockeye salmon'. In (23b) we have an example of a quantifier, Paya 'many', bound to the external argument daastimc 'mink', to give 'many mink', but incorporated into the bound verb -nit 'stocked with'. There can be no direct relationship between verb and quantifier here. In (23c) we have the case of an adjective, Xui 'good', modifying the external argument Xudim 'mussel', giving 'good mussels'. Again, the adjective is incorporated into the bound verb, -nit 'stocked with' and again there can be no direct relationship between verb and adjective.3 Since the verb has no direct relationship with the elements incorporated into it, the verb cannot be reduced in valency. 8.2.2.2.4 Does Not Obey Subcategorisation Restrictions. Another thing to note about the examples in (23) above is that syntactic incorporation must be completely insensitive to subcategorisation restrictions imposed by the incorporating verb, as shown in (23b,c). In these cases, the bound verb -nit 'stocked with' can occur in syntactic combination with a quantifier (23b) or an adjective, as in (23c). As shown by the subcategorisation

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

The semantics of such syntactic constructions are directly decomposable from the parts of the construction.

224 Linguistic Theory and Complex Words

frame for -nit in (24) (compare with (16) above), this should be a violation of strict subcategorisation.

b.

[ N _ ] * [Q _

v

]v

* [ADJ _ ]

v

The examples in (23b,c) are clearly instances of the syntactic incorporation of an element extracted from the object NP into the verb, producing rather novel semantic constructions, such as 'many-stocked with' or 'good-stocked with'. The syntactic structures for the cases in (23) are represented in (25a-c), where the relationships are illustrated by traces and links. V

25. a. (= 23a) V

NP

Pu -nitPaaqXuk REF stocked with

mfiaat sockeye

V'

b. (=23b)

V ADJ

NP V

ADJ 1 Pay ai -infaX t many stocked with

t

N | 1 daastimc mink

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

24. a. -nit

Morphosyntax 225 V'

c. (=23c) V V

ADJP 1

I

Xuij -nit tj good stocked with |

N

.! Xudim mussel

As can be seen from these structures, this is a purely syntactic operation of extraction and movement into a bound V. 8.2.2.2.5 Verb Forms No Semantic Unit with the Incorporee. The final point to be made concerning syntactic incorporation is that the verb forms no semantic unit with the incorporee. Combinations such as 'many-stocked with' or 'good-stocked with' are really not 'nameable' in the sense of Bauer (1983) or 'nameworthy' (Hale & Keyser 1993). In fact, they tend to show little semantic relationship with each other directly, the incorporee having a much more obvious semantic relationship with the rest of the object noun phrase from which it has come. On the other hand, nameability in complex words must also be subject to more general rules governing the permissibility of lexical items, .... Chomsky postulates that such restrictions are universal, and they must apply as filters on complex as well as simplex lexemes. (Bauer 1983:87) Independent of whether the relationships which can be expressed in derivation are universal, it is clear that new derivational markings can be introduced in a language to express new meanings if the need is present. ... To this extent at least, whether or not a thing is nameable must be considered a pragmatic feature. (Hale & Keyser 1993) These collocations are loose, syntactic mergings of elements that are not directly related semantically, the result of the process of syntactic incorporation.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

ADJ

NP

The preceding account has attempted to illustrate the fact that Nuuchahnulth maintains a distinction between two types of incorporation, lexical and syntactic. These two types of incorporation share superficial similarities but are fundamentally distinct in the way they operate within the grammar. We will now examine in more detail the curious behaviour of syntactic incorporation in Nuuchahnulth. 8.2.2.3 Incorporation of Non-head Elements In what follows we will examine the properties of syntactic incorporations such as those introduced in (23b,c) above. We will see to what extent elements may be extracted from the external argument into the main verb. As indicated by the structure presented in (20), the first element of the object NP which one encounters is the Quantifier. This is demonstrated by the form ?aX$iil?i 'the two...', combined with the classifier for songs, in (26a) which is part of the noun phrase 'the two songs'. What is interesting about this construction is the fact that it is possible to incorporate this quantifier, along with its classifier, into the main verb, as is shown in (26b), where 'two' has moved from the object NP into the verb, creating a verbal complex 'two-have'. 26. a. PunaaksapaXah [ PaXjMiiPi nuuk]NP Pu -na'k° -savp -'aX -(m)ah PaXa -jifT =?i* nuuk REF-have -MC-NOW-IS.IND two -songs =DEF song 'I now give (the child) the two songs' b. PaXpiiinaksapat 4

[

Pastimxyak]NP I

PaXa -pi'i10 -naYk° -sa*p -'at Pastimx -yak0 two -songs -have -MC -PASS sing lullaby -song 'he now gives him two lullabies' This process occurs similarly with other quantifiers, as in the example in (23b), and also with complex numerals, which will be discussed further on in this section. Notice that the classifier is incorporated along with the numeral, which is the usual result when there is a classifier. The example in (27) demonstrates the position of adjectives modifying the object noun, that is, they precede the head noun, as shown in (20).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

226 Linguistic Theory and Complex Words

Morphosyntax 227

PuyuPaiwePin Pani hii.qaa [ Xui nisrha]Np Pu -u?ai -we^Pin Pani hii.qaa Xui nisrha REF-perceive-3s.QT that LOC-3.SUB good land '... and saw that it was fine country there'

However, as shown by (23c) and (28a-e) below, adjectives may be incorporated into the main verb, just as with quantifiers. 28. a. qahak°in;yapaX cakupuk°itPi 'she left her dead husband behind' qahak°-inyu-'ap-'aX £akup-uk-(m)it=?i- dead -abandon-CAUS -NOW husband -POSS -PAST =DEF b. wikyuu jMsiip tePiiPi iuucsma 'they never got a badly sick woman' wikyuu jMsaq-iiyip te?ii=PiY iuucsma never bad-obtain sick-DEF woman t'ick^aX c. PuusaTaX Puusa-k -caX fickaa -TaX

'it made a loud thundering sound' loud-make sound thunder -sound

d. XuichinX Xui -chi -inawiX

'he had married beautiful (women)' attractive -marry -GRAD[L+S]

e. taihsaa?aXk°a?at tai -hsa- -'aX -k*a?at

'they were to eat fresh (fish)' fresh -long to eat -NOW -for

Note that in examples (28a, b), the object is marked by the specifier ?i\ Examples such as (28d,e) show an ellipsed object head noun. Clearly, these examples demonstrate the syntactic incorporation of adjectives into the matrix verb. Furthermore, in syntactic incorporation it is not just the bare root that is incorporated into the verb, but it may involve the stem plus the various derivational suffixes that accompany it. Thus: 29.

huu^uqXisPathsimda^aXat 'they had him train to get Uchucklesits' huucuqXis-?ath-sim£-'ap-'aX-'at Uchucklesit-tribe-train for-CAUS -NOW-PASS

In this example, the entire construction 'Uchucklesit tribe', consisting of the placename, huuduqXis, combined with the suffix -?ath 'tribe', forms an object of the verb meaning to 'train for power against something or someone'. Clearly, examples such as this demonstrate that syntactic incorporation is quite distinct from lexical incorporation in all of the properties presented in (11). We will now move on to examine the most

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

27.

228 Linguistic Theory and Complex Words complex case of incorporation in Nuuchahnulth, that of numeral incorporation.

The number system as described in Section 3.5 is quite complex, but the fact that Nuuchahnulth allows incorporation of numerals presents yet a further and more challenging level of complexity that we shall now explore. The issue of coordination is interesting because of its interaction with numerals and incorporation, and this will be the focus of this section. The crucial examples are the following: 30. a. hayuciisiPaX Pis qa£da hayu -cri-siX-'aX Pis qacda

'he did it for 13 days' ten -days -MOM -NOW and three

b. haayumiikuk PuhPis muu Piihtuup 'he captured fourteen whales' hayu-mrk^-uk PuhPis muu Piihtuup 10-capture-POSS and four whale c. hayuuTisuksi Pis sudiiq 'I spent 300 (dollars) on him' hayu-iiq-'rs-uk-si- Pis suda-iiq ten-score-spend-NOW-ls.ABS and five-score d. caqiicqimijayiiPatma

Xapisim [ tj Pis sudaqumi taanaa]

caqrc-qimi -ayi* -'at -ma* Xapisim Pis suda -qimi taanaa twenty-CLS -give-PASS-3s.iND Raccoon and five-CLS dollar 'Raccoon was given twenty-five dollars' Note that in such examples, there is a complex numeral which is separated into two parts, one part incorporated into the verb, the other appearing in situ, preceded by the coordinator, ?is or ?uh?is 'and'. The mechanics of complex numerals have already been discussed in Section 3.5, to which the reader is referred. Such examples pose an interesting problem for lexical treatments of incorporation, since the two parts of what would appear to be a single, complex numeral are located at a distance from one another. This is the typical form in which syntactic incorporation is performed on such elements, and it is always the first element of the complex numeral that is incorporated. This fact suggests a more articulated shape for the phrase containing a numeral. The idea here is that the phrase acting as object is basically a quantifier phrase in which the head is the numeral that delimits the possible range of numbers, much as we talk about units, tens, hundreds, etc.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.2.2.4 Numeral Incorporation

Morphosyntax 229

Should the number involved be complex, then we must have some way to distinguish between the 'head' and the 'complement' members of the number, so that only the head may be incorporated, along the following lines:

Q'

NP

Q b.

We can now explain how incorporation works in these cases. Firstly, it operates on the head of the Object argument, whether this is an NP or a QP. If it is an NP, then the N is incorporated, as one would expect. But if the noun object is actually within a quantifier phrase, then the head of the QP is the object of incorporation, not the head noun. A question that arises here is: is it really quantifier phrases, or just numerals that operate in this fashion? The following examples demonstrate that this process is not limited to numerals, although they pose the most interesting problem for this process.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

31. a.

230 Linguistic Theory and Complex Words

32. a. Payajsiik [ej ciihati] 0 diihati Paya -siik

'he made a lot of arrows' many -make arrow

c. hePiissiX nuusciiPi ha -'rs -siX nuusCii =Pi-

'he ate up all the sea eggs' complete -eat -MOM sea egg =DEF

As you can see from these examples, the quantifiers ?aya 'many' and ha 'all' can appear in a similar position to the numerals, preceding the noun, or it can be incorporated into the verb, just as the head numeral can. Also similarly, when it occurs in conjunction with a noun or noun phrase, it is always the quantifier that is incorporated, not the noun. In addition, the clitic definite article, ?i\ behaves as it does with the numbers, attaching to the quantifier rather than the noun, as in the following example: 33. a. XawaaPaX [PayePi siiXuk] 'the many movers approached' Xawa* -'aX Paya =?r siiXuk near -NOW many=DEF move house b. hiiPiitq [PayePi quuPas] hii -Prtq Paya =?i* quuPas

'... where there are many people' LOC -3s.REL many =DEF person

All of this suggests that, like the numbers, other quantifiers occur within a domain containing both the quantifiers and the NP, and that incorporation is sensitive to the head of the larger phrase, whether it be NP or QP, within the sentence. 8.2.2.5 Incorporation and Unaccusatives Before leaving the topic of incorporation we should examine one last class of cases: the intransitives. It may be surprising to consider the possibility that intransitive verbs may also incorporate, but there exists a class of cases of intransitives that do participate in incorporation. Examples include the following: 34. a. PayasuuX namintPath Paya -sawiX namint -'ath

'many Namint people died' many -die[M] Namint -tribe

b. haayiiPiX qicinstaqimi ha -ayiiPiX qicin -taqimi

'all the Louse people entered' all -enter house louse -tribe

c.

'there's a dog that's dying' dog -die[M] (Kyu)

fiiSsuuX Tiic -sawiX

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. PaayiiXPaaqX maatmaas 'I was going to invite many tribes' Paya-[L]-'iX-PaaqX maatmaas many -invite [L] -INTENT tribe

Morphosyntax 231

The Unaccusative Hypothesis was first advanced within the theory of Relational Grammar .... This hypothesis states that there are two types of intransitive verb. The first has a subject perceived as actively initiating or actively responsible for the action of the verb, such as run, talk, resign. These are known as unergative verbs. It is usually assumed that subjects are assigned an external argument by such verbs, namely the thematic role of Agent. The second type has subjects which lack this active participation, and they include verbs such as arrive, die, fall. These verbs are the unaccusative verbs and in many languages they are distinguished from unergative verbs syntactically or morphologically. ... At the theoretical level it is generally assumed that unergatives have an underlying subject but no object, while unaccusatives have an underlying object (which becomes a surface subject later in the derivation of the sentence,) but no underlying subject. Spencer provides the following phrase structure trees, where (35a) represents the unergative verb and (35b) the unaccusative. 35. a.

S

S

InfT

NP Infl

Tom

b.

VP

ran

arrived

Tom

Haegemann (1994:330) states that '[t]hough the class of unaccusative verbs is not easily defined ... it appears that it contains primarily verbs of movement and verbs that indicate some state or a change of state.' The verbs in (34) would certainly fit into this category. So how are such verbs

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

What all of these cases show is that there may be incorporation within intransitive predicates, under certain conditions. As a working hypothesis, it seems reasonable to propose that this form of incorporation may occur with the class of intransitive predicates known as 'unaccusative'. Spencer (1991:260) states the following concerning unaccusative verbs:

232 Linguistic Theory and Complex Words

s

36.

V 0

NP

-suuX Adj

N

Paya quuPas If this is more or less correct, then one can see how such verbs may participate in incorporation, since the surface subject actually originates in an object position, from which some element may be incorporated, as below. 37.

V

NP

i

0

/\

suuX Adj <\

N

Pava quuPas

The eventual site of the subject would be accomplished by the movement of the V upward. Thus, it is possible to encounter certain cases of what

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

to be characterised? The most important point for us here is in the placement of the argument in unaccusative verb structures. As suggested by (35b), the argument originates in object position (36), but without object case assignment, and is therefore required to move into the empty subject position, which has no assigned 9-role.

Morphosyntax 233

appear to be subject incorporation in Nuuchahnulth, the result of incorporating into an unaccusative verb.

The proposal here maintains that it is necessary to subsume the object NP within a larger QP when it is associated with any quantifier, and that in such a situation incorporation must necessarily involve syntactic movement from the position of head of QP. This, then, would seem to suggest that incorporation in general points to a syntactic movement rule or its equivalent, rather than a lexical rule generating the object of incorporation in situ. The result of numeral incorporation in Nuuchahnulth is a discontinuous element, part of which is found at the head of the clause, and the remainder of which is associated with the syntactic object of that clause. It seems that such a situation would be rare in the languages of the world, since it requires a language to have several quite special properties: object incorporation, syntactically complex numerals, and a QP/NP alternation for syntactic objects. 8.2.3

Summary

The differences between the two types of incorporation, lexical and syntactic, fit into the categories in the following table: Lexical

Syntactic

a. No coreference outside the verb.

Coreference outside the verb possible,

b. The head is incorporated.

Non-head elements may be incorporated.

c. Semantics may be idiosyncratic and opaque.

Semantics are compositional and transparent.

d. Reduces valency of the verb.

Does not affect valency.

e. Observes subcategorisation restrictions.

Subcategorisation restrictions are ignored.

f. The verb and incorporee form a coherent semantic unit.

The verb and incorporee do not necessarily form a coherent semantic unit.

Table 8.1 Lexical vs. Syntactic Incorporation

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.2.2.6 Conclusion

Incorporation in Nuuchahnulth goes beyond the standard treatment of the phenomenon proposed by Baker (1988). It is not restricted to the incorporation of the head noun of the object phrase but involves various X elements in the initial position within the object phrase. It is sensitive to the linear precedence/immediate dominance of the elements within the object. 8.3

Compounding

Nuuchahnulth has been described as a language that has one root per word, combining with any number of suffixes and possibly containing one or more reduplicative prefixes but never a second root. Thus, regular compounding has been ruled out as a process in the language, placing the language in a very special place among the languages of the world. For the most part, this is an accurate depiction of the facts. One will never encounter more than a single root within a word in Nuuchahnulth. However, there are situations where two or more nouns are brought together in a relationship that is highly suggestive of compounding in other languages, such as the examples in (38). 38. a. [PiihPii [ yafrnuut Xaqmis]N]NP 'the large oil-bladder' Piihco=?r ya£aa-mu*t Xaq-mis big =DEF dogfish-leftover oil-thing b. [[ muunaa niiqniiqayak ]N]NP 'sewing-machine' muunaa DUP- niq-(y)a--yak0) machine REP-sew-DUR -tool Examples such as these, and other patterns with which they may be confounded, will be the major focus of the discussion in this section. 8.3.1 Types of Noun-Noun Collocations In the following sections we will examine the various conditions under which two nouns (possibly more in some cases) may occur adjacent to each other. In many cases this will in fact be across some kind of major syntactic boundary, but nevertheless the positing of the syntactic boundaries actually presupposes the ability to distinguish these collocations and therefore there is a danger of circularity in any attempt to use this as the sole criterion for distinguishing such cases. Keenan & Comrie (1977:63) state with reference to their study of relative clauses that:

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

234 Linguistic Theory and Complex Words

We are attempting to determine the universal properties of relative clauses by comparing their syntactic form in a large number of languages. To do this it is necessary to have a largely syntax-free way of identifying RC's in an arbitrary language. Our solution to this problem is to use an essentially semantically-based definition of RC. In order to avoid circularity in the isolating of syntactic arguments we will resort to a similar strategy in attempting to provide tests for the determination of the status of noun-noun collocations in Nuuchahnulth. These will include arguments from both semantics and, where appropriate, morphology and syntax. The reader is urged to bear this in mind as we proceed. 8.3.1.1 Adjacent Arguments of the Verb One possible collocation of two nouns is the situation whereby two arguments of the verb appear as bare nouns one following the other, e.g., [ V [N]NP [N]Np]s. * n tri e following examples, one can see the flexibility of word order inherent in the arguments of the verb as well as the relatively common occurrence of bare nouns in such positions. 39. a. suuweePin [ hupinwasukPi ] N P [ Taatusmit ] N P suu -we-Pin hupinwas -uk =?iSaatus-mrt hold-3.QT small canoe-POSS=DEF deer-son O S V 'Deer kept his little canoe' b. suk^iPaX [ Saatusmit ] NP [ diimaakPi ] NP suu -siX -'aX Taatus -mrt diimaak =?iv hold deer -son mussel shell knife=DEF V S O 'Deer took his knife' Notice that in the first case (39a), the subject follows the object, whereas in the second case (39b), it precedes. This is another example of the relatively free word order described previously. Determining the syntactic status of the nouns in such cases may be quite challenging, and often depends on context and semantics for deciding subject- or object-hood. Simply put, in a sentence with two bare nouns adjacent to each other, if the verb is transitive, each noun may

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Morphosyntax 235

serve as one argument, but which argument is subject and which is object will often depend on factors other than syntactic ones. However, in most cases, it is clear from the context and the thematic roles of the noun constituents which one serves as subject and which as object. In fact, things may not always be so simple, but for the present purposes, we will rely on these diagnostics. 8.3.1.2

Head-Possessor

Nuuchahnulth is a head-marking language, in which the head in a headpossessor relationship occurs before the possessor. It is frequently the case that the noun phrase consists of nothing more than two nouns in collocation, one acting as the head and the other as its possessor. Take the following examples: 40. a. iiikaaPaqhak [ supicmisuk rhaPaquuPa ]NP siikaaPaq -ha'k su^ic-mis -uk rhaPaquuPa my -2s.EXCL sand -thing-POSS Maakoa 'you look like the sand of Maakoa!' b. PuuPiihsiXci^aX [ qasiiPat frniiXPi ] N P Pu -'iih-siX-cip-'aX qasii -'at frniiX =?v REF -get-MOM-INDIR-NOW eye -INAL dog =DEF 'they took out the eyes of the dog' Notice here that in (40a) the head noun has a rather loose relationship to the possessor and is consequently marked by the alienable possessive suffix -uk. However, in (40b) the relationship is one of part/whole and is thus marked by the inalienable possessive suffix -'at, discussed in Chapter 4. In either case, the suffix indicating possession serves to indicate the relationship of the two nouns involved. It should be remarked that the pronominal referent in the possessive is encoded by the addition of inflectional marking after the possessive suffix which, in the case of the first person singular, is -ukqas, e.g. dakupukqas 'my husband'. This contains the -uk possessive suffix followed by the 1st person singular subordinating suffix -qas (see Chapter 4 for a discussion of this aspect of the grammar). This is one of the few cases where we can see that casemarking exists in the language, making it perhaps the most straightforward of the noun-noun collocations. One further observation to make about the case in (40b) is that the definite article is found to occur on the second member of the noun

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

236 Linguistic Theory and Complex Words

phrase, a distinctly odd situation. In a normal noun phrase the definite article, modifying the head, would appear on the first element of the noun phrase, as discussed above. This supports the view that in fact this cannot be either a right-headed or a double-headed noun phrase. Thus, when two nouns are found adjacent to each other, the initial member bearing either of the possessive suffixes, it is the first one that will be the head noun and the second its complement. As a complement, the second noun may or may not be marked as specific, independent of the head. 8.3.1.3 Co-ordination without Conjunctions Co-ordination in Nuuchahnulth may occur at various levels of the syntax, and quite commonly within the noun phrase. It is often accompanied by the use of the conjunction ?is or ?uh?i§ 'and', as demonstrated in the following sentences: 41. a. Paanasaqh Xihas [ saasinmit ?i§ sisitinaakXi]Np ?ana-sasa-(q)h Xih-as saasin-mrt Pis sisitinaakXi only-just[L]-MW paddle-reach Hummingbird-son and Long-Tailed 'only Hummingbird and Long-tailed nearly caught up with him' b. yaksiX [ k°atyaat PuhPis fiixfiyapixin ] N P yak -siX k^atyaat PuhPis t'iixt'iyapixin come in view-MOM Kwatyat and Tihtiyapihin 'Kwatyat and Tihtipihin came into view (from inside the monster)' While the use of these conjunctions is quite common, it is also possible to conjoin elements without the use of any overt conjunction, by simply adjoining the two elements in a noun-noun collocation, as below: 42. a. nucPuuPaXwePin [ hiiyiPi waafrtPi ] N P nuc -'awiX -'aX-we^Pin hiiyi =?iY waaS'it=?r steam-on rocks-NOW-3.QT snake=DEF frog =DEF 'the snakes and the frogs began to steam' b. PiinaxaX hacatakaX suu [ muustati diihati ]Np Piinaxi -'aX hacatak-'aX suu muustati diihati ready-NOW all -NOW hold bow arrow 'all were ready, holding bow and arrow'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Morphosyntax 237

238 Linguistic Theory and Complex Words

In such cases one can see that both syntactically and semantically neither element can be seen to act as the head, either one being interchangeable with the other, just as in co-ordinate structures in other languages. Thus, there must be a difference between a co-ordinate structure and a headcomplement construction. Again, we find that we must resort to semantics to provide the determination of the roles in the sentence. Once the determination is made, we can represent such constructions in a standard phrase structure format, allowing for the presence or absence of a conjunction. 43.

NP / \ NP NP

NP OR NP

^ f CONJ

\ NP

Whether this occurs at N' or at N level is open to further debate, but does not directly bear on the situation with regard to bare nouns, where it is indeterminate. 8.3.1.4 Apposition Apposition involves the non-restrictive modification of one noun by the addition of another one, providing further information. Cases in English may involve constructions such as 'my best friend, John' or 'the winners of the Stanley Cup, the Dallas Stars'. They act as non-restrictive modifiers of the head noun and, as such, are less intimately connected with the head than its restrictive modifiers, including pre-head modifiers such as adjectives, and post-head modifiers such as restrictive relative clauses. This formation is also found in Nuuchahnulth, and provides a further class of cases involving noun-noun collocations, such as the following examples: 44. a. yaacsaapat [ [ k^a^aXas ] N/N ' [ XayaxPi ]NP]NP yaac-sa-p-'at k^aTaXas ] N/N ' Xayax=Pr step-MC-PASS Sound-of-a-stick-breaking fast =DEF 'they sent out Sound-of-a-stick-breaking, the fast runner'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

c. minkaaPaXukquuwePin [ Xaqmis Xaaqaast ] N P mink -'aX-uk-quu-we^Pin Xaq -mis Xaaq -a-§t round-NOW-POSS-3.CND-3.QT oil-thing blubber-dried 'he would have oil and dried blubber'

Morphosyntax 239

c. PahPaaPaXwePin hafiissiPaX [[muuPii quu?as]NP[kwiisahiminh?i]NP]NP PahPaaPaX-we'Pin hafiis-siX-'aX muu=?r quuPas k^iisahi-rhinh =?r and then -3.QT bathe-MOM-NOW four=DEF person hunter-PL=DEF 'then the four people, the hunters, started to bathe' Sentence (44a) for instance, provides the further information about Sound-of-a-stick-breaking-on-the-ground, that he is a fast runner. It further qualifies the noun phrase consisting of his name by providing this additional information. Sentence (44b) does much the same thing, except in this instance the further information is the actual name of the slave, 'Kwatyat'. Note that sentence (44c) cannot mean either that 'the hunters bathed the four people' nor that 'the four people bathed the hunters', since these two possibilities would both require transitivising markers on the verb 'bathe', since in its current form it acts as an intransitive, monadic verb. In order for it to act as a dyadic verb, the causative suffix, -'ap, would be added to the verb, allowing for this transitive reading, as illustrated in the examples in (45a) and (45b). 45. a. hat'iisPa^aXquuwePin haak°aaX?i faatnePisukPi hafiis-'ap-'aX-quu-we^Pin haak°aaX=?i- -t- fana -Pis -uk =?i* bathe-CAUS-NOW-CND-3.QT woman =DEF child-PL-DIM-POSS =DEF 'the young woman had her children bathe' b. haahuupsiPaX hat'iisPajMcim waaPaX waa-'aX haahuup -siX-'aX hat'iis-'ap-'iCim advise -MOM-NOW bathe -CAUS-2>3.IMP say-NOW 'he advised him saying "you will bathe him'" Thus, in cases such as those in (45) we must regard the relationship between the two nouns as one of apposition, the first acting as the head noun of the noun phrase and the second serving as a non-restrictive modifier of the head. Notice that one clear indicator of this is the ability of the appositive to bear the definite clitic, ?i\ as demonstrated in (45a) and (45c) above, which should not be possible if the modifier were directly adjacent to the

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. suk^iX [ [ quuiukPi ]N/N» [ k°atyaat ] N P ] N P suu -siX quui-uk =?iY k°atyaat take-MOM slave-POSS=DEF Kwatyat 'hej took hisj slave, Kwatyat'

head, within the N', since the clitic will always appear on the first element of the NP, as discussed above. As regards the head and its arguments, note that in (45a) it is impossible to decide the level at which the appositive attaches, since we are dealing with a personal name; (45b) provides more indication that the constituent is N', but only if we assume that there is a 3rd person pronominal element following the word for 'slave'. This would make this a possessive structure, analogous to those discussed above. The final example, (45c), provides the clearest evidence that this must be at least N', since it contains a numeral in addition to the head noun and appositive. One further interesting point about (45b) and (45c) is that in both cases there is a definite article on the leftmost member of the noun phrase, suggesting that this is in fact a full noun phrase. Thus, syntactically, the structure of (45c) would be better described as in (46) below: 46. SPEC

muu

quuPas k°iisahiminh?i

Huddleston (1984:266) states that '[n]on-restrictive dependents are peripheral in that they can be omitted with relatively little loss - never loss of grammaticality, and loss only of supplementary information as opposed to information integral to the main message.' Thus, the assignment of appositives in Nuuchahnulth, by definition always non-restrictive, to the X' level as sister to the X'-level head seems appropriate in this case, differentiating it from a restrictive relative clause, and also helps to explain the position of the clitic definite article, which appears on the first, or only element within its domain, i.e. the NP. 8.3.1.5 Headless Relative Clauses A further possibility in Nuuchahnulth is that of the collocation of a headless relative clause co-occurring with its object, as in 'the one fishing for halibut' or 'the one killing people', which may be realised in

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

240 Linguistic Theory and Complex Words

Nuuchahnulth as a collocation of a relativised verbal element, treated as a nominal followed by its object. Relative clauses in Nuuchahnulth are typically headed by a relative pronoun yacf and furthermore often contain inflectional suffixation indicating the relational nature of the phrase, especially the inflectional marker -Prtq, as illustrated below. 47. a. qahsiX PuusaahaX [Xacaqstim [yafiis&XPitq ha?uksiX]RC]NP qah-siX Puusaahi-'aX Xac-'aqstim yaq^-'i-s-SiX-Prtq ha?uk-siX die-mom because-NOW fat-at abdomen REL-eat-MOM-3.REL eat-MOM 'he died because of the abdominal fat which he had eaten' b. [iuucsmePi [yaq^acPitq t'ana]RC]NP iuucsma=?i* yaq*0 -ac-?f tq tana woman=DEF REL-belong to-3.REL child 'the woman whose child he was' [ quuPasPi [ yaqPiitq c. huptsiPaX hupt-siX-'aX quu?as=?i* yaq" -Prtq hide-MOM-NOW person=DEF REL-3.REL 'the man who had seen him hid'

nacuPai ] R C ]NP nacuPai see

This is the typical form of a relative clause in Nuuchahnulth, but it is also possible to form a headless relative clause without the use of either the relative pronoun, ya(f, or the relative inflectional marker -Prtq. In such cases, what appears on the surface is a verbal noun, usually with the specifier clitic attached. Examples of this include: 48. a. faaquk^iPatwePin hiniiPasPaX [ [ naaCukPi ]RC]NP faaquk -siX -'at -we^Pin hin -iiPas -'aX naa£uk =Pi* believe-MOM-PASS-3.QT LOC -go outside-NOW look for=DEF 'he was believed, and the one who was looking for (Wolf) left' b. nacuPaiPaX PuyuPaiPaX [ [ $imiiqsuu£iXuk?i ]RC]NP na£-u?ai-'aX Pu -uPai -'aX yimiiqsu -iiCiX -uk =?r see-perceive-NOW REF-perceive-NOW brother-in-law-iNC-POSS=DEF 'he saw the one who had become his brother-in-law' The cases above are of simple constructions with only a single element within the headless relative clause, but this is not the only possibility. One may also encounter an object of the relativized transitive verb that

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Morphosyntax 241

242 Linguistic Theory and Complex Words

follows it, as would be expected in a head-initial language. Thus, there are examples such as the following:

b. PucaCiXitPi maCiaath 'the one who had gone to the Machlath' ?u-ca-siX-(m)it=Pf ma£iaath REF-go-MOM-PAST=DEF Machhla-tribe Such cases pose particular problems in that they behave very much as if there was a head noun followed by an appositive or other complement. Naturally, the second noun only appears in conjunction with a dyadic verb, and so this may serve as an indicator of the relative clause status of the construction. Thus, in most cases a combination of the argument structure of the original verb combined with the semantics of the construction suggest the alternative analysis of a headless relative clause. Syntactically, the representation of such structures should follow the lines of (50). 50.

DP D'

SPEC D

RC

I

I

?i

VP V

DP

The difference between headed and headless relatives is the presence or absence of the external head, and the nature of REL, whether it contains an overt pronoun or relative inflectional marking. Note that the difference between restrictive relative clauses and appositives rests in the level at which they attach to the noun phrase: in the case of the former, it is as a sister of N, whereas for the latter it is as a sister of N'.

8.3.1.6

Pseudo-Compounding

Our final case of noun-noun collocations is what may be called pseudocompounding. The notion of pseudo-compounding is related to what has

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

49. a. hamat'apPi muxmis 'the one who understands stealing-detection' hamat-'ap=Pr mux-mis know-CAUS=DEF stealing detection-thing

Morphosyntax 243

In stark contrast to the productivity of derivational affixation, the possibilities for compounding more than one independent lexical stem are extremely limited. The only such constructions are 'quasi-compounds' consisting of two nominals in apposition within a tightly bound noun phrase where internal sandhi or deletion has united them into a whole treated as an inflectional unit: they were rare in the older language. Examples from Greenlandic provided by Fortescue (ibid) include: 51.

unnuk 'night' + manna 'this' —> unnumanna 'tonight' atsip 'sleeve'+ paa 'its entrance'—> atsipaa 'cuff (of sleeve)' ullup 'day' + qiqqa 'its middle' —> ulluqiqqa 'midday'

However, in Nuuchahnulth the distance in phonological terms between the elements combined is even greater than in Greenlandic, and for this reason it seems preferable to refer to the process as pseudo-compounding. The cases we are referring to in Nuuchahnulth involve the juxtaposition of two nouns, the first of which acts as the head of the construction, the second delimiting the range of the first, i.e., acting as a restrictive modifier. 52. a. Puktqasanap [[ muunaa niiqniiqayak ]N]NP Pu -tqa -sanap muunaa DUP- niq -(y)a* -yak10 REF-underneath-on beach machine REP- sew -DUR -tool 'a sewing-machine was set under-them-on-the-beach' b. hiiweePin PuPucama [ [ YuditPas daPak ] N ] NP hii-we^in DUP- Pu -cama TuditPas daPak LOC -3.QT SUF- REF -go to flow in ground water 'it led to a water-hole' c. hiiweePin [ [ Xaaq hii -weY?in Xaaq there -3.QT fat 'it was whale-blubber'

Piihtuup ] N ] NP Piihtuup whale

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

been referred to in Greenlandic as 'quasi-compounding' (Fortescue 1984:330):

These examples demonstrate the compound-like structure of these forms. In all of these examples, the second noun acts as an adjunct to the first, consistent with the head-initial nature of Nuuchahnulth. The noun may be simplex, as in (52b,c), or it may contain a root plus one or more derivational suffixes (52a). These pseudo-compounds exhibit a common property of compounds, semantic opacity. For example^ (50c) does not mean a hole in the water but a hole filled with water, (50d) refers specifically to a kind of fat associated with a whale, and (50e) indicates a generic type of night, a winter night. It is still unclear exactly what the status of these compound forms is, but it is clear that they are different from reduced relatives in not allowing definiteness to be encoded on the second noun. This also distinguishes them from appositives. Semantically, they are clearly not co-ordinated nouns, and thus we must conclude that they constitute an independent category which, for the time being, we will represent by the term pseudocompound. Comparing this structure with that of the head-possessor construction in (48), we see that the difference between the two is the presence vs. absence of the possessive marker and the inability of the second member of a pseudo-compound to bear definiteness. Their syntactic description, since they consist solely of bare nouns, may be as Ns branching under another N, as described below. 53.

N N

N

The following examples provide cases of compounds that are modified by either adjectives (54a,b) or quantifiers (54c,d). It is clear from the context that the modifier has scope over the entire compound and not just the nearest branch. 54. a. [PiihPii [ yacmuut Xaqmis]N]NP 'the large oil-bladder' Piih10 =?r yad-muut Xaq-mis big=DEF dogfish-remains oil-thing b. [yaaqPii [Pathii cuPich]N]NP ya*q =?i- Pathii duPiSh

'the long winter night' long =DEF night winter

c. [PayePi [hamutck^i Piihtuup]N]Np 'the many whale bones' Paya=Pf hamut-ckV Piihtuup many=DEF bone-remains whale

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

244 Linguistic Theory and Complex Words

Morphosyntax 245

d. [PayePi [ quuPas niitiinaPath]N]Np'the many Ditidaht men' Paya =?f quuPas niitiina-Path many=DEF people Ditidaht-tribe

55.

N' ADJ

N N

N

The structure may, in fact, be more complex than this, but it will surely allow for the placement of the compound in a position subordinate to whatever modifier it co-occurs with. 8.3.2 Implications for Syntactic Theory One clear point that arises from the preceding exposition is the necessity for a clear set of principles for dealing with noun-noun collocations in languages with properties akin to those of Nuuchahnulth. One cannot rely strictly on syntactic structure, as this carries the inherent danger of circularity and the consequent potential for misanalysis. One must resort therefore to a combination of strategies involving morphological, syntactic and semantic markers of class membership and syntactic category. The list below illustrates the various properties involved in the different kinds of noun-noun collocations in Nuuchahnulth. Type of Collocation Adjacent Arguments

Head-Possessor

Co-ordination

Properties 1. Governed by argument structure of verb 2. Determined by thematic relations 3. Mobility of individual arguments context-dependent 1. Head-marking for possession 2. Head precedes possessor 3. Possessor may bear independent definiteness marking 1. Interchangeable 2. Semantically of same status 3. Syntactically, neither may constitute the head by itself

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Building on our description of the phrase structure in (53), it would seem that these forms have the following structure:

4

Apposition

5

Reduced Relatives

6

Pseudo-compounds

1. Left-headed 2. Appositive is non-restrictive 3. Appositive may bear independent definiteness 1. No external head 2. Involves nominalised verb, which may take arguments 3. Usually marked as definite 1. Two nouns combine, the first as head, the second as modifier 2. Compound-like semantic irregularities 3. Similar structure to the head-possessor construction, but second member cannot bear independent definiteness

Table 8.2 Noun-Noun Collocations in Nuuchahnulth With a clear understanding of these properties we can more accurately determine the exact syntactic structure of a sequence of nouns when they are encountered in the sentence in Nuuchahnulth or any other language with similar properties. 8.3.3

Summary

Due to the nature of Nuuchahnulth grammar, there are many opportunities for nouns to appear next to other nouns and so it becomes necessary to derive heuristics for deciding the nature of the relationship between the two (or possibly more) nouns. We have examined six different types of this and found that for the most part, they can be distinguished by a combination of strategies drawn from the morphological, syntactic, and semantic properties of the language. All of these must be brought to bear in some cases in order to clarify the relationships between bare adjacent nouns in Nuuchahnulth. Finally, we see that there is a type of compounding which occurs in Nuuchahnulth, even though there is a prohibition on the co-occurrence of roots within a single word in the language. 8.4

Classifiers

In this section, we will discuss the use of classifiers in Nuuchahnulth. Following Greenberg (1972) and Allan (1977), classifiers are clearly seen to operate in the syntax in the usual case and, while Nuuchahnulth

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

246 Linguistic Theory and Complex Words

Morphosyntax 247

classifiers have always been subsumed under the class of lexical suffix, this language is no exception to the rule. This section intends to provide a tentative proposal for the syntactic structure of the classifier phrase (C1P) and introduce one puzzle concerning such phrases.

Previous investigations of Nuuchahnulth have briefly discussed what might be referred to as the classifier system of the language. Swadesh (1939), discussing the Tsishaath variety of Nuuchahnulth, mentions a set of morphemes, which he calls 'numerate suffixes' that act as classifiers in the language. He states (ibid: 88): '[t]here is a group of suffixes used almost exclusively with numeral stems. These suffixes, which may be called numerate suffixes, are all governing suffixes.'4 Rose (1981) briefly discusses classifiers in the Kyuquot variety, but does not go into great detail on the subject. She states (Rose 1981:321) that 'there are form-shape classifiers, constituting in Nootka part of a counter affix system.' She goes on to say: The second major type of classifier G(overning)S(uffix) is the counter type, which affixes to a quantity base. Counter classifiers render a stem nominal, locative, temporal, or quantity .... As a general rule, a quantity modifying an inanimate nominal in such a construction must be specified by a governing nominal counter classifier, (ibid) Overall, there has been no detailed account of the mechanics of the classifier system and how it interacts with the rest of the system in any variety of Nuuchahnulth. 8.4.2 The Nature of Classifiers The standard description of classifiers involves a class of morphemes that interact with the quantifier system of a language, serving to mediate between the quantifier and its referring noun within an argument phrase. Greenberg (1972:1) defines classifiers as involving 'the overt expression of one kind of quantification, namely, counting by units.' Allan (1977: 285) states that: ... classifiers are defined on two criteria: (a) they occur as morphemes in surface structures under specifiable conditions;

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.4.1 Preliminaries

248 Linguistic Theory and Complex Words

In the following section we will examine some of the properties employed to decide whether a system makes use of classifiers. 8.4.3 Some Properties of Classifier Systems There are a number of properties of classifier systems that help to distinguish them. These include: 1. Arbitrariness - while the class of nouns delimited by a classifier may often be described by certain properties they share, there are often outliers that don't obviously fit under that description. This arbitrariness makes precise definitions of classifiers extremely difficult. 2.

Semantic opacity - the classifier is a lexicalised item listed in the mental lexicon whose meaning may be rather opaque and difficult to describe, e.g. Sapir & Swadesh's (1939:321) gloss for the Nuuchahnulth classifier -qimi: 'chunk-shaped, (i.e. round, roundish, square, squarish, spherical, spheroid, cubical, etc.) rounded, chunkshaped, rounded object made of...'.

3.

Feature agreement - the referring noun must agree with the classifier with which it is bound, otherwise the result is ungrammatical. All countable nouns must be assigned to a specific classifier.5

4.

Co-occurrence restrictions - subcategorisation is highly restricted most of the time in that the classifier only attaches to numerals and quantifiers but not to other categories.

5.

Complementarity - two classifiers cannot co-occur in the same structure, since their featural properties would conflict.

6.

Pronominal force - a classifier may act as a pronoun without the presence of the coreferential nominal.

With respect to these properties, Nuuchahnulth seems to be a prototypical case.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

(b) they have meaning, in the sense that a classifier denotes some salient perceived or imputed characteristic of the entity to which an associated noun refers (or may refer).

Morphosyntax 249

8.4.4 Categories of Classification

Classifiers are an important element in the syntax of nouns. Each noun is assigned a particular classifier, much as nouns are assigned genders in many European languages. While gender is loosely based on sex, classifiers are based on distinctive features of shape, natural kind and function. ... It is useful to distinguish two types of classifier: (a) measure or mensural classifiers, which denote quantities of an item... (b) type or sortal classifiers which belong with the noun and classify it in terms of some intrinsic features of the noun it takes. Allan (1977) provides a more elaborated set of seven categories for the classification of classifiers. These are: (i) material, (ii) shape, (iii) consistency, (iv) size, (v) location, (vi) arrangement, and (vii) quanta. He goes on to state that the first five constitute inherent properties of classifier languages. Nuuchahnulth clearly evinces certain of these properties, placing it squarely within the category of classifier language. The property of arrangement, demonstrated by morphemes marking concepts such as 'heap, bunch, clump' is not restricted to classifier languages but is more universal in scope. Similarly, the property of quanta, involving units such as 'basketful, handful, etc.,' as well as number names such as dozen, score, etc. and measure categories, including dimension, weight, volume, and time are universal in their distribution. As Allan states (1977:306) 'measurements based on universal human properties and experience recur in many languages.' Such elements do not make a contribution to the distinction of classifier vs. non-classifier languages and therefore will not form part of our discussion here.6 Based on Allan's classification, Nuuchahnulth is clearly a classifier language. What we will focus on in the remainder of this section are those classifiers that satisfy Allan's categories (1-5). 8.4.5 The Classifiers of Nuuchahnulth Swadesh (1939) provides a list of so-called numerate suffixes in Nuuchahnulth. For the most part these agree with the interpretation of classifiers here, but there is at least one case that does not fit. This is the case

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Matthews & Yip (1994:92) discuss the distinction between two types of classifiers, mensural and sortal, with respect to Cantonese:

250 Linguistic Theory and Complex Words

56. a. PaXiiq PaXa -iiq"

'forty' two -score

b. sudiiq suda -iiq"

'hundred' five -score

As noted by Allan (1977) above, number names are universal and for this reason we do not consider this morpheme as part of the intrinsic classifier system, since it is neither arbitrary nor semantically opaque. Furthermore, the base to which it attaches is limited to the simplex numerals, not any quantifier. Most importantly, it does not delimit the class of entities that act as its referents. An additional reason for ruling this out is the fact that -iiqw violates one of our properties for classifier-hood, that of complementarity, as in (57). 57. a. PaXiiqciqitPi jiinwai 'the forty whaling canoes' ?aXa-iiqC0-diq-(m)it=?i- £in\Vai two -score-long objects-PAST=DEF whaling canoe b. PaXiiqhtak kwaaiuk PaXa -iiq-hta-k° k^aaiuk

'forty bundles of rubbing medicine' two-score-bundles rubbing medicine

Note that in both of the cases in (57) there is a classifier: -ciq 'long objects' in (57a) and -htaV 'bundles' in (57b). If -iiq0) 'score' were also a classifier, then we would have a clash of features, and consequent ungrammaticality. Since such cases are not infrequent in the data, we must conclude that -Ha" is, in fact, not a classifier, but a number name a la Allan. This may then combine with a classifier, resulting in forms such as those in (57). Rose (1981) provides a list of suffixes that includes elements we would refer to as classifiers. She does not, however, employ the term classifier in our sense, but rather divides the various morphemes we would label classifiers into several groups based on their affixal properties. Her category of 'Quantity Base' comes closest to our interpretation of intrinsic classifier, but still contains extrinsic classifiers, for example, morphemes for measurement such as -pit 'handspans', and -yai 'fathoms'. She also refers to 'governing nominal counter' classifiers such as -6vi 'days', -qP'v£h 'years,' but does not list them as a separate category.7

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

of -iiq0 'score', which is used to augment the numeral system, creating the numbers for forty, sixty, eighty, etc., as in (56). Note that this is similar to the English use of the term 'score' as in 'four score (= eighty)'.

In addition, her categories of 'Governing Temporal Affixes' and 'Governing Quantity Affixes' are, in fact, not intrinsic classifiers in the strictest sense as (i) they do not delimit the set of entities they take, and (ii) they do not exhibit a bound relation with its complement. These will therefore not feature in our discussion. A comprehensive and yet not exhaustive list of classifiers in Nuuchahnulth is provided in the Appendix, with examples of their usage.8 We will now move on to examine some exceptional cases in the Nuuchahnulth classifier system. 8.4.6

Exceptions

There are exceptions to the typical case of classifiers with nominals. In certain situations, a quantifier may occur without a classifier. One class of such cases concerns nominals referring to humans, where no classifier is typically employed, resulting in a bare numeral preceding the nominal, as in (58). 58. a. cawaak t'anePis dawaak fanePis

'one child' one child -DIM

b. PaXa haawiihaX PaXa haawiihaX

'two young men' two young men

c. PaXa haathaak^aX PaXa DUP- haak^aaX

'two girls' two PL- young woman

d. hayu Pis muu quuPas hayu Pis muu quuPas

'fourteen people' ten and four person

One might suggest that there may be an animacy feature [+human] that explains the absence of an overt classifier with nominals involving humans. However, in addition to these cases, there is a consistent pattern of words for 'blanket' that never seem to occur with an intrinsic classifier. Note that this cannot simply be that 'blankets' are somehow considered to be mass nouns, without intrinsic classifiers, since they do cooccur with numerals, as shown in (59). 59. a. PuyiiPatsi ha\Vii?i suda Xisai Pu -ayr -'at -si* hawii =?i* suda Xisai REF-give-PASS-ls.ABS chief =DEF five blanket 'I was given five blankets by the chief

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Morphosyntax 251

b. PaXa du£maqai PaXa dudmaqai

'two mountain goat blankets' two mountain goat blanket

c. Paya ducmaqai Paya ducmaqai

'many mountain goat blankets' many mountain goat blanket

There may remain other cases not yet discussed and there is clearly further work to be done in explaining such exceptional cases. We should note that this has been known at least since Knipe (1868:13) who states: A very curious feature of the numeration is that while when applied to certain objects the numerals are used simply and without any addition, there are other classes of substantives the individuals of which are apparently of a most incongruous nature, with which the numeral is only used with a particular suffix. Man (ko-us), woman (hahquatl, klootsmah), salmon (tsoowit, hissit), frog (wah-it), with many others, take only the simple numeral, and never noop, but only tsowwauk for one. Many other words use only noop for one, and with them every numeral takes the addition kamilh [-qimi}. Perhaps this is the most numerous class, including all sorts of money, clothing, birds, and beasts, as well as houses, stones, guns, paddles, months, and many more. We will now move on to address the issue of the relationship between classifiers and their governing category. 8.4,7 Quantifiers and their Relationship to Classifiers Classifiers appear conjoined with a certain subset of bases in Nuuchahnulth, the numerals and various other quantifiers. We have already discussed numerals in Section 3.5 and so will limit our discussion here accordingly. When numerals combine with the intrinsic classifiers, the results take three possible forms: (i) simplex number plus classifier (60); (ii) complex number, e.g. PaXiiq 'forty' plus classifier (61); and (iii) the most complicated case, that where the numeral is composed of several independent bases, typically with the classifier attached to the first member (62). 60. a. PaXqimi muksyi PaX -qimi muksyi

'two stones' two -CLS stone

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

252 Linguistic Theory and Complex Words

Morphosyntax 253

b. muupiii nuuk 10 muu -j^ri nuuk

'four songs' four -CLS song

b. PaXiiqhtak Faaiuk PaXa-iiq-hta-k° Faaiuk 62. a. hayuqumi Pis suda taanaa hayu-qimi Pis suda taanaa b. hayusta hayu-ista

Pis suda Pis suda

'40 bundles of rubbing medicine' two-score-bundles rubbing medicine 'fifteen dollars' ten -CLS and five

dollar

'fifteen people in a canoe' ten -CLS and five

The cases in (60) and (61) are quite straightforward, the classifier attaching to the end of the numeral in both cases, but the cases in example (62) are somewhat more complicated, due to the complex nature of compound numerals, which employ a conjunction Pis. As can be seen from these examples, the usual result is that the classifier appears on the first element of the numeral. This is the usual pattern of attachment. An alternative pattern marks both members of a compound numeral with the classifier, in which case it must always be the same classifier, as in (63) below. 63.

caqiicqimi Pis sudaqimi kuukuhwisa 'twenty-five seals' caqix-qimi Pis suda-qimi kuukuhwisa 20-CLS and 5-CLS seal

Such cases, while relatively rare, do occur and must be treated as cases of double-marking. Besides the numerals there is a small class of quantifiers that may cooccur with classifiers, as shown in (64). Paya Panah Pana q^amaa his hiyiq XaPu duu£k tak'a

'many' 'few' 'how many?' 'thus many' 'all' 'all sorts of 'another' 'all' 'each'

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

61. a. PaXiiqdiqitPi jMnwai 'the forty whaling canoes' <0 PaX-iiq -diq-(m)it=?r jMnwai two-score-CLS-PAST=DEF whaling canoe

254 Linguistic Theory and Complex Words

These quantifiers can be seen in combination with various classifiers in the following examples.

b. hissaathPi his -sa*th =?f

maatmaas maatmaas

c. XaPuu^iiPi XaPu* -jM'i10 =Pi*

hiina hiina

'the other sea lion' other -CLS =DEF sea lion 'both the tribes' both -CLS =DEF tribe 'the other Quartz Being' other -CLS =DEF Quartz Being

d. wik Puh Payadiq cisaaPath 'the Tsishaa didn't have many(canoes)' wik Puh Paya-diq disaY-'ath not is many-CLS Tsishaa -tribe e. q^amaahtakuk qaPuuc qtoama--hta-k-uk qaPuuc

'her many baskets' thus many-CLS-POSS packbasket

This concludes the discussion of the quantifiers that may be found to co-occur with classifiers. The following section will examine the internal syntactic structure of the combinations of quantifier, classifier, and nominal that occurs in Nuuchahnulth. 8.4.8 The Syntax of Classifiers Having demonstrated the existence and nature of Nuuchahnulth classifiers, we will now move on to a discussion of how the classifiers operate within the syntactic system of the language. 8.4.8.1 The Pronom inal Use of Classifiers In languages that employ classifiers, they may act as pronouns in combination with numerals or other quantifiers. The following examples, along with (65d), illustrate this property in Nuuchahnulth. 66. a. XawiiciPaX XaPuktaqimiPi 'he approached the other (pair)' XawiiCiX-'aX XaPu'-taqimi-?i- approach-NOW other-CLS =DEF b. Pah PaXpiiiPi Pah PaXa-prT =Pr

'these two (songs)' this two-CLS =DEF

c. maaXsiPaXXaa PaXjMiiPi 'again they tied up the two (whales)' maX-[L]-siX-'aX=Xaa ?aXa-j}rT=?r tie -GRAD -MOM -NOW =again tWO -CLS=DEF

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

65. a. XaPuqumiPi tukuuk Y Xa?u -qimi =?i* tukuuk

Morphosyntax 255

Note that this property is entirely in keeping with our expectations for a classifier language.

In any construction employing a classifier, the classifier predetermines the range of possible nouns that must agree with it. Thus the classifier has scope over its referent. (67) demonstrates a case where the numeral applies to all of the conjoined members of the object phrase. 67.

hayuqumiPaXukquuwePin £adak FaXaq k^ak^aX -a?aq hayu-qimi-'aX-uk -quu-we~?in jiadak ten -CLS-NOW-POSS-CND-3.QT potlatch gift sea-otter-skin ducmaqai hiscuq^ati ditiiPasim hiix^aa di^uuqs. ducmaqai hiscuq^ati ditrPasim hiix'V dijkrqs goat-hair blanket abalone copper dentalia brass 'he used as gifts ten sea-otter skins, goatskin blankets, abalone, copper, dentalia, brass'

If the referents do not agree in terms of the classifier they take, as shown in (68), then the numeral must be repeated with each classifier in the sentence. 68.

PaXciq iuuTaminh suu^iisyak Pis PaXpiii PaX -diq iuu^a-minh suup-'rs-yak* Pis PaX -$rT two-cls tray -PL soup -eat-tool and two-CLS 'two soup plates and two lengths of calico'

qicaaiminh. qicaai-rhinh calico -PL

In the cases below, the quantifier has scope over all the referents in the phrase. Furthermore, definiteness is marked on the quantifier in (69a), again having scope over the entire phrase. 69. a. hisimyuupaX PayePi quuPas Pis iuucsaamiih XiiPiiPaX his -imi -'u-p-'aX Paya =Pi* quuPas PiS iuucsaamiih XiiPii-'aX all-in group-MC -NOW many=DEF men and women feast-NOW 'he assembled the many men and women and gave a feast' b. cawaak quuPas Pis iuucsaamiih 'each man and woman' cawaak quuPas Pis iuucsaamiih one man and women The above cases clearly suggest a situation where the numeral is intimately connected to the classifier, with scope relations based on the

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.4.8.2 The Scope of Classifiers

256 Linguistic Theory and Complex Words

combination having scope over any nominal argument. This will figure centrally in our discussion of the syntactic structure of the quantifier phrase in the next section.

With regard to the previous discussion, we conclude that the numeral/quantifier, the intrinsic classifier, and its referent (the NP) within a nominal expression take the following order (70). Note that the classifier may be absent under the conditions discussed in Section 8.4.6. 70.

Num/Quantifier — (CLS) — NP

Since classifier always requires a governing category which is always the numeral or quantifier, we will assume that the quantifier appears as the head of a Q(uantifier) P(hrase), which then takes the classifier as its complement. This suggests a structure as in (66), where the Cl(assifier) P(hrase) is the sister of Q. 71.

QP SPEC

Q' C1P SPEC

Unlike English, in Nuuchahnulth, a numeral or quantifier cannot combine directly with a noun even if the noun is countable except, perhaps, in those exceptional cases of 'blanket' and human nominals discussed in Section 8.4.6. We would argue that there is a functional layer of C1P above the NP, taking NP as its complement with the evidence from the feature agreement and subcategorisation of the classifier, which acts as the head. Given the fact that the classifier subcategorises for a certain class of nominals, it should act as head of a phrase taking a nominal complement; for example, a subset of entities bearing the semantic feature of -diq 'long object', say 'canoe', are restricted in their place of occurrence, as

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.4.8.3 The Syntactic Structure of the QP

the physical properties stipulated cannot be taken by nominals requiring a classifier which specifies the feature of 'chunk-shape' such as -qimi. In such an eventuality, there will be a feature clash and the nominal phrase will not be read off properly. Therefore, this shows that Cl, as a functional head, is needed to carry the functional information that contributes directly to semantic interpretation, although such semantic interpretation is not overt lexically. An example of this is provided in (72). In (72a), the entity subcategorised by the classifier, mfiaat 'sockeye salmon,' shares the same feature, while (72b) is ungrammatical, as the classifier requires an entity with the feature of the same sort but the NP nuutyak 'hoops' is of another feature which does not comply with the classifier. 72. a.

CIP

b.

* CIP

Cl [Feature a]

Cl [Feature a] -diq ^SJ^ [Feature a] mifaat

-flq

" ^ ^ [Feature (3] nuutyak

Further evidence to support the presence of such a functional layer above NP within the nominal phrase is that the classifier may act as a pronominal, without the accompanying nominal, as in (73). 73.

Q' CIP

I PaX

'the two ...'

Cl -qimi

As to the position of numerals and quantifiers with respect to the CIP proposed above, we argue that they constitute the head of QP taking the CIP as its complement, given the fact that intrinsic classifiers never occur without the presence of a quantifier in Nuuchahnulth and that they are a dependent element and must attach to the quantifier.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Morphosyntax 257

258 Linguistic Theory and Complex Words

The relationship between quantifier, classifier and its referent is manifested in the phrase structure provided in (69), exemplifying the phrase muufiiii nuuk 'four songs' (from (60b).

SPEC

muu

8.4.9

Summary

What the preceding has shown is that classifiers in Nuuchahnulth, while obviously combined in the morphology with numerals and quantifiers, have a special status in the syntax. They demand a specific set of nodes in the phrase structure, CIP, dominating NP, since the classifier delimits the class of nouns that may appear under NP. Similarly, CIP is dominated by the QP, which specifies the numeral or quantifier. 8.5 Inflection as Clisis All previous accounts of Southern Wakashan morphology recognise the existence of clitics in the grammar. However, the extent of clisis in the accounts varies widely. Sapir & Swadesh (1939) recognise only three clitics in Nuuchahnulth, Klokeid (1976, 1978) extends this number for Ditidaht, and Davidson (2002), discussing both Nuuchahnulth and Makah, considers the entire inflectional morphology to consist of clitics. In this section we will investigate the arguments for and against such extensive use of clisis, and examine its potential utility in explaining the phenomenon of incorporation in Nuuchahnulth. Zwicky & Pullum (1983:502) have the following to say about clitics.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

74.

Morphosyntax 259

They go on to discuss various properties shared by clitics, as presented in Section 8.1.1. We have already discussed the position of Sapir & Swadesh (1939) on the subject in that earlier section, and so we will move on to the views of Klokeid (1976, 1978) on the subject of cliticisation in the closely-related Southern Wakashan language, Ditidaht. Klokeid's position is that much of the inflection involves cliticisation, and he makes the following observation: Haas and Swadesh (1933:198) use the term 'word suffix' for a set of morphemes that is nearly, but not completely, coterminous with the set of enclitics. Sapir and Swadesh (1939: 236-41) in dealing with the more northerly West Coast languages, especially Tseshaht, call the corresponding set 'incremental suffixes.' These scholars based their distinction on purely phonological criteria, while the present study is concerned with the suffix as well. In addition to the enclitics, the 'incremental' or 'word suffixes' include certain true suffixes such as the passive -Pit. [my emphasis] (Klokeid 1978: 175, fn 3) As stated in the above quote, certain inflectional morphemes, including the Passive and Causative markers, are treated as suffixes by Klokeid, so that inflection is divided up between suffixes and clitics. Klokeid furthermore claims that the sentence clitics attach to the first word in the sentence, but this is not necessarily the case, at least in Nuuchahnulth, as shown by the example in (75). 75.

Piiwaanux Pukiaamita Puyi PahPaa Piiwaanux Pu -kia* -(m)it -ma* Puyi PahPaa Iiwaanuh REF -name-PAST-3s.lND when that way 'he was called Iiwaanuh at that time'

Here, the fronted element does not bear the inflection, which appears on the immediately following V. This form of fronting is not rare in Nuuchahnulth, particularly with the verb Pukiaa 'name'.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Two types of bound morphemes are found attached to (free) words in many languages: clitics and affixes, in particular inflectional affixes.

260 Linguistic Theory and Complex Words

76.

musPasimPi PukiaaPaX paak^atqsPatu ihus -'as -im =?i* Pu -kia- -'aX paak°in -sPatu close-on ground-thing =DEF REF-name-NOW skate -at door [L] 'the door is called Skate-at-door'

Klokeid (1978:172) notes this and adds the condition: 'No nominal that has its own enclitics can function as topic, unless the sentence is in narrative style.' Unfortunately, he provides no definition of 'narrative style' with which to test this claim. So we are left with a number of questions about the claims made by Klokeid, but it seems clear that he never intended for all inflection to be part of the clitic set. Renker (1988) examines the behaviour of AUX elements in Makah, based on Steele et al. (1981) and proposes that 'a clitic sequence', marking tense and modality, occurs as the second sentential element, the usual Wackernagel position. However, Renker provides no examination of topicalisation or fronting to accompany her claims. The clitic sequence in Makah with the form: (INDEFINITE TENSE) (TEMPORAL PROXIMITY) (DEFINITE TENSE) (MODALITY) SUBJECT MARKER (OBJECT MARKER) (SPEAKER INTENTION) may be termed AUXM (Renker 1987:134) She goes on to propose that a subset of inflection be placed in AUX including tense and mood markers but not causative, applicative, passive, etc. i.e. those that interact with predicate argument structure and perhaps some others. This then agrees with Klokeid's position on the matter. It is Davidson (2002:94) who makes the most radical claim regarding cliticisation in Southern Wakashan, in the following statement: Though all the aforementioned authors refer to these formatives as suffixes of one type or another, their status as such is actually rather ambivalent. They differ from lexical and aspect suffixes on a number of phonological, morphological, and semantic points. Some researchers (e.g. Klokeid 1976,

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Klokeid (1978:172) observes that '[n]o nominal that has its own enclitics can function as topic' but again it is possible to have a nominal plus a specifier, which is a member of the standardly accepted group of clitics discussed in Section 8.1, in topic position, as in (76).

Morphosyntax 261

Davidson goes on to propose that all inflection be treated as cliticisation. He proposes a number of criteria to demonstrate that this is the case and we will examine these criteria in the following two sections, beginning with the phonological differences. The discussion of word-level phenomena from Chapter 6 will be relevant to this investigation and the reader is referred to that chapter for further information. 8.5.1 Phonological Differences Davidson provides a number of phonological differences between suffixes and clitics, and we will examine each of them in turn. 8.5.1.1

Delabialisation

Davidson's first putative difference is that '[l]abialized velar and uvular consonants alternate with their homorganic non-labialized counterparts before clitics just as they do before word boundaries' (Davidson 2002:252). As we have already seen in Chapter 6, delabialisation occurs at the word level. When a form occurs word-finally, it is delabialised as in (77b). Similarly, when it is followed by a clitic, all of which are consonant-initial, it will be delabialised. 77. a. ha\Viiuk°itPi hawii -uk -(m)it =?r b. hawiiuk hawii-uk

q^ayaciik q^ayaciik

c. hawiiukPi q^ayaciik ha\Vii -uk =?r q^ayaciik

'their former chief chief-POSS -PAST =DEF 'chief of the Wolves' chief-POSS wolf 'the chief of the Wolves' chief -POSS =DEF wolf

Thus, delabialisation does not distinguish these categories. 8.5.1.2 Templatic Effects Davidson's second claim is that '[c]litics are never associated with CV templates.' (ibid:252) It should be noted that for the most part only a restricted set of stem-level morphemes are associated with CV templates,

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

1978, Jacobsen 1979a 151, note 31, Renker 1987) have suggested that these formatives in Ditidaht and Makah are actually clitics (an analysis that extends easily to Nuuchahnulth as well). This idea appears to account for a number of their syntactic properties.

262 Linguistic Theory and Complex Words as in (78) (see the discussion in Chapter 4); the majority of bound morphemes in the language are not. ducuqhtanuPaX \Vasniipmit 'Snipe washed his feet' DUP- duq-hta-nuX-'aX \Vasniip-mrt SUF-wash-on foot[R]-MC-NOW snipe-Son

There is little to suggest that this is more than a coincidental effect. 8.5.1.3 Glottalisation Davidson's third claim is that glottalising clitics do not affect fricatives. To begin with, this same observation may be equally applied if these elements are treated as Stratum 2 inflectional suffixes. Furthermore, none of the standardly accepted clitics in (1) above trigger glottalisation, so these cannot be used as support for this argument. Finally, Rose (1981: 19) has this to say about post-lexical glottalisation: Because glottalization of consonants also occurs when a word-final consonant precedes a word beginning with ?, as in [wik'aya] wik Paya 'not many', one might claim that inflectional suffixes are not suffixes, but are words or clitics beginning with P. However, glottalization over word boundaries has no effect on the uvular stops; nor does an initial glottal stop cause an aspect-final X to delete. It seems clear that phonetic glottalization should be distinguished from morphophonemic glottalization and likely that inflectional suffixes are suffixes rather than more independent morphemes. Here again there seems to be no argument for inflection being cliticisation. 8.5.1.4

Vowel Contraction

Davidson further observes that long vowels in the second syllable of a word do not contract over a glottal stop inserted by a glottalising clitic. Since most words consist of at least a monosyllabic root plus aspectual suffix (consisting of one or more syllables), few opportunities arise for such a generalisation to obtain. The inflectional morpheme, -PaaqX INTENT, which is not a glottalising suffix, contracts across the third and later (79a), but not the second (79b), just as any other morpheme would.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

78.

Morphosyntax 263

79. a. waayaqhaaqXhas waaya -haY -PaaqX -ha*s

'which one will I buy?' which? -buy [M] -INTENT -ls.lNTER

8.5.1.5

Vowel Coalescence

Finally, Davidson claims that '[sequences of vowels that do not normally reduce over a glottal stop do reduce with glottal stops inserted by glottalizing clitics.' (Davidson 2002:252) Actually, the rule of vowel coalescence operates at all levels of the grammar, so this will not distinguish stem-level from word-level from postlexical level. 80.

maaidaas ma -aica -'as

'house against wall on the ground' dwell-at upright surface-on ground

In summary, none of Davidson's proposed phonological criteria distinguish clitics from word-level morphology, nor do they strongly support the argument for inflection as clisis. We will now move on to the morphological criteria. 8.5.2 Morphological Differences Davidson remarks on several morphological differences between suffixes and clitics and we will examine each of these. The first is: '[c]litics may be added only to complete words, that is, words that are capable of independent use in a sentence, while most suffixes are core suffixes that attach to both bound and free bases.' (Davidson 2002:252) Note that Structure Preservation ensures that the output of any level must be a potential word and so this property cannot distinguish word level from post-lexical level. 8.5.2.1 Free vs. Bound Davidson's second observation notes that '[c]losely related to this is the fact that clitics never occur with the combining forms of free roots or suffixes, only with free forms.' (Davidson ibid:253) This is also a property of Structure Preservation and does not distinguish word level from postlexical processes.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

b. PuhaaPaaqXah XuiPii 'I will buy the good one' Pu -haY-?aaqX-(m)a-h Xui=?i* REF -buy -INTENT-Is.IND good-DEF

264 Linguistic Theory and Complex Words

Davidson (ibid) states that: 'In the overall structure of the word, suffixes appear closer to the root than clitics; clitics never precede suffixes in a word.' Again, this argument does not distinguish Stratum 2 affixes from clitics, although once clitic-hood is established, then elements outside the established clitic can be considered also to be clitics, per Zwicky & Pullum's criterion F (Section 8.1.1). 8.5.2.3 Fixed Order The property of fixed order, described by Davidson (ibid) as: '[c]litics must occur in a fixed order relative to each other, while (nuclear) suffixes may often occur in different orders to express different meanings' only applies to a subset of the Stratum 1 derivational suffixes. Inflectional morphemes occur in a fixed order in all languages, so this does not decide between inflectional suffix and clitic. 8.5.2.4 Attachment Restrictions Davidson's final property is described as: '[c]litics attach without formal or semantic irregularity to almost any word in the language according to its syntactic function. Suffixes are collocationally more restricted, and can have idiosyncratic semantic effects or enter into lexicalized combinations with bases.' (ibid) Again, the latter is more a property of stem-level derivation than of inflection. Furthermore, Davidson recognises portmanteau combinations of aspect and causative morphemes, the former said to be a suffix, the latter a clitic (Davidson 2002: 226): Perfective marking has become tied up with causative marking. When the perfective suffix is not present (with no explicit aspect marking, or with the continuative, repetitive, etc.) causative is expressed by the clitic M, N -'ap. With the perfective suffix, however, it is normally expressed by a causative perfective portmanteau suffix. In summary, none of the morphological criteria provided by Davidson distinguish word-level suffixes from clitics, and so there is still no strong motivation for supporting the claim that all inflection is clisis.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.5.2.2 Positional Effects

Morphosyntax 265

In this section we will examine arguments against the proposal that inflection is actually clisis, beginning with a brief discussion of Rose's (1981) position. Rose's position is two-edged, arguing first that inflection must be affixal because of its contextualised interpretation, as in the first quote below, and secondly, that certain phonological effects distinguish inflection from clisis, as in the second quote. A second reason why inflectional affixes are assumed to be affixes is that at least some of them depend, for their interpretation, on the particular stem they are affixed to. The past morpheme -int can be interpreted as 'deceased' when attached to non-predicative animate nominals. The passive -'at can be interpreted as 'possessor of when affixed to a body part nominal or to a limited set of intransitive verbal predicates. The plural -rhinh, normally translated as 'a bunch of, can be interpreted as 'each' when affixed to quantity stems (in addition to having its regular collective interpretation); for example, 'two each (of each type of object)' as well as 'two (of them)'. Finally, the diminutive -Pis, although normally interpreted as 'small, endearing, young', obligatorily accompanies certain adjectival stems such as 'narrow', 'only' etc. whose object of attribution can be neither small, endearing, nor young. (Rose 1981:255) Rose goes on to discuss phonological criteria which distinguish between lexical and syntactic effects. Because glottalization of consonants also occurs when a word-final consonant precedes a word beginning with ?, as in [wik'aya] wik Paya 'not many', one might claim that inflectional suffixes are not suffixes, but are words or clitics beginning with P. However, glottalization over word boundaries has no effect on the uvular stops; nor does an initial glottal stop cause an aspect-final X to delete. It seems clear that phonetic glottalization should be distinguished from morphophonemic glottalization and likely that inflectional suffixes are suffixes rather than more independent morphemes. (Rose 1981:19)

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.5.3 Arguments against Inflection as Clisis

266 Linguistic Theory and Complex Words Rose clearly supports the locus of inflection as being within the lexicon rather than the view that it involves cliticisation.

Our first argument against inflection as clisis comes from the morphosyntactic interaction between elements. To begin with, there is clearly the potential for phonological interaction between a base and an adjacent clitic, as demonstrated by the following example from English. 81. a. the dog's paw [z]

b. the cat's paw [s]

c. the lynx's paw [iz]

Note that the nature of the preceding phoneme is important for determining the phonological form of the clitic in (81). However, there is an absence of morphosyntactic interaction between the base and an adjacent clitic, as demonstrated in (82). 82. a. the woman who loved him's mother b. the woman's mother c * the woman who loved his mother d. Jim's mother So, the interpretation of (82a) must be (82b) and not (82c) even though there is no phonological clash between the him and 's, as demonstrated by (82d). Clearly there is a syntactic clash. So, if morphemes such as the causative, passive, or applicative were clitics, then we would not expect them to affect the valency of the verb to which they cliticise, since they would not interact with them on that level, just as him does not interact with the 's possessive on that level. However, these morphemes do have this valency-altering effect in Nuuchahnulth, and therefore must be lexical, not postlexical, as shown in (83). 83. a. qahsiX Bill qah-siX Bill b. qahsaap Bill Joe qah -sa*p Bill Joe

'Bill dies (Intrans)' die-MOM Bill 'Bill kills Joe (Causative)' die -MC Bill Joe

c. qahsaapat Bill PuhPat Joe 'Bill got killed (by Joe) (Passive)' qah -sa-p -'at Bill PuhPat Joe die -MC -pass Bill by Joe

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.5.3.1 Morphosyntactic Interaction

Morphosyntax 267 Obligatoriness

Verbal inflection is part of virtually every predicate in Nuuchahnulth, since, as Carstairs (1981) puts it 'inflectional affixes are also "members of a relatively small closed system, one of whose members must always appear at the relevant place in structure.'" So this inflection is obligatory, but clitics, at least the standard class discussed in Section 8.1 are not obligatory. They provide additional information associated with a predicate, but this is not required for grammaticality, unlike person/number marking, for example. For instance: (i) plural is an optional category in the grammar (84a); (ii) =Paaia 'always' has the semantic counterpart sadinkm 'always,' which is a free-standing adverb (84b); and (iii) =Xaa 'again' has the semantic counterpart huPa 'back, again,' also a free adverb (84c). 84. a. waaPaXwePinXaa waa -'aX -we^Pin =Xaa

'they said it again' say -NOW -3.QT =again

b. dawaakPicaXquuwePin saadinkPis TihakPis finiiXPisPi dawaak -Pis-'aX-quu-we^Pin saadink-Pis frhak-Pis frniiX -Pis =Pr one -DIM-NOW-CND-3.QT always-DlM cry-DlM dog -DIM=DEF 'now there was one little dog who was always crying' c. huPaasPaX huPa -'as -'aX

'again (Skate) did as before' again -go to -NOW

While the clitics discussed above are very common in Nuuchahnulth, there are clearly alternative means to achieving the same semantic result, and so the clitics cannot be considered obligatory morphemes, as one would claim for inflectional morphemes. 8.5.3.3

Morphotactics

Perhaps the strongest argument for inflectional morphemes not being clitics comes from morphotactics, as discussed in Chapter 3. Note the examples of inflectional morphemes below that contain initial consonant clusters (more examples are available in Sections 4.2 and 4.3 on inflection): 85. a. -£kar
2nd sing. Inferential mood 2nd sing. Relative Dubitative mood Comparative

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.5.3.2

As affixes within the lexicon, these forms have the opportunity to re-syllabify the egregious clusters into the preceding coda, but if these were syntactic words, then they would violate the morphotactic constraints on words which demand that there be one and only one wordinitial consonant, as discussed in Chapter 3. Note also that weight-wise, these constitute sufficiently heavy morphemes to be words (compare ciyas 'swan', for example). 8.5.4 Clitic or Suffix? In deciding whether or not Nuuchahnulth inflection should be treated as cliticisation, there are at least three possible positions that one might take. Each of these will be discussed below. 8.5.4.1 Position 1: All Inflection is Cliticsation If one were to adopt the position that all inflection is cliticisation, then it would be necessary to allow operations such as causativisation and passivisation to take place in the syntax. One would also have to account for the portmanteau aspect/inflectional morphs and how they arise (see Section 4.4.3). Obviously, this would also have severe consequences for the claims of word-level morphology and would complicate the issue of double reduplication enormously (for which, Section 6.3). 8.5.4.2 Position 2: All Inflection is Affixation Under this interpretation, all inflection would occur in the lexicon, and would appear as word-level morphology in a stratal-ordered morphology. It would, however, pose problems for the explanation of the position of paradigmatic inflection in the syntax. This issue will be discussed in Section 8.5.5. 8.5.4.3 Position 3: Some Inflection is Word-level Morphology and Some is Cliticisation The final position is the one that divides inflection between the morphology and the syntax. Under this account, any inflection that affects predicate argument structure or other aspects of the base would occur within the lexicon. Inflection that interacts directly with the syntax, i.e., mood/person marking, would be assembled as a clitic cluster and lexically inserted at a functional head (I or D).

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

268 Linguistic Theory and Complex Words

Morphosyntax 269

This analysis would answer many questions regarding the grammar of Nuuchahnulth and the appearance of inflection in unexpected positions and this proposal will be pursued further in the following section.

At this point we move on to investigate the status of inflection from a syntactic point of view. This may assist us in explaining some confusing facts about Nuuchahnulth grammar. We have presented arguments in Section 8.2.2 for the phenomenon of fronting, and its role in the grammar of Nuuchahnulth. We have also noted that inflection never appears on such fronted elements, suggesting that the placement of inflection must be syntactically governed, i.e., that it does not simply attach to the first available element of the sentence, but to a specific syntactic node. If we consider I to be the site of verbal inflection and D to be the site of nominal inflection, as in (86), then we can view the position of inflectional morphology on certain elements as being due to adjunction of the lexical material to I or D via head movement from a lower head (87). 86. a.

b.

IP

SPEC

SPEC I Inflectional Cluster 87. a.

b.

IP

SPEC

D Inflectional Cluster

VP

r

DP SPEC

D' D

s£ • n A \ \\ Inflectional / / Cluster IJ

D

NP

/£ n

A \ Inflectional / / II \ \ Cluster

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

8.5.5 Phrase Structure

The heads from which the lexical material may be raised are straightforward in the case of the verb, but in the case of the noun, as noted in Section 8.2.4, they may vary within certain parameters, allowing for raising of a quantifier, adjective, or noun. If all (or part of) inflection is lexically inserted at I or D, then in a head-initial language like Nuuchahnulth it would give the effect of being a second-position clitic. One significant difference between I and D in this respect is that raising is obligatory with the DP, but optional within IP, so that verbal inflection may actually cliticise to elements such as conjunctions, which are located in SPEC,IP, as shown by the example below. For more discussion of this point, see Yiu & Stonham (2004). 88. a. PahPaaPaXma cnqSiPaX haakwaaXPi PahPaaPaX-ma* diiq -siX -'aX haakwaaX =?iY and then-3s.lND sing chant -MOM-NOW young woman=DEF 'and then the young woman started singing' b. PahPaaPaX diiqsiPaXma iuucsaamiiriPi PahPaaPaX diiq -siX -'aX -maiuucsaamiih =Pi* and then sing chant -MOM-NOW-3s.lND women =DEF 'and then the woman chanted' This is where confusion might arise regarding the nature of verbs in Nuuchahnulth. If one observes verbal inflection on a conjunction such as PahPaaPaX 'and then' one might be led to assume that there are no category distinctions in the language. This, however, is far from the truth and the explanation for the position of inflection is that it is located in I, but is phonologically dependent on what precedes it, either a raised V or an element in its own SPEC, if there is one. Since inflection never appears on those elements that have been fronted, one must assume that such stylistic fronting is a late, phonological rule, and unavailable for cliticisation. 8.5.6 Summary Clearly this exposition of the facts of Nuuchahnulth inflection has only scratched the surface, and there is room for much further work in this area. Nevertheless, it seems clear that one must assume that there is a clitic cluster containing a subset of the inflectional morphemes which are base-generated in the functional heads of IP and DP and which require

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

270 Linguistic Theory and Complex Words

the raising of material into an adjunction position or, in the case of IP, the cliticisation to SPEC, IP. In the case of IP, verbs, either free or bound, are base-generated in V and then may be raised into I. Free verbs do not incorporate, but a bound verb may incorporate, in which case the highest object head is adjoined to V and then optionally raised to I. A consequence of this analysis is the prediction that inflection may appear on one of four different categories of elements: (i) lexical verbs with or without an incorporated element; (ii) auxiliaries, such as the negative wik, etc. which are obligatorily raised to I, (iii) wh-words which are obligatorily incorporated if possible and always raised to COMP, and (iv) adverbs, conjunctions, etc. which appear in SPEC, IP position.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Morphosyntax 271

9

In this concluding chapter, we will review the discussion in the previous chapters, focussing on important issues and insights. We will also consider some of the key issues in Nuuchahnulth morphology that remain to be addressed. 9.1

Review

In the preceding chapters, we have examined Nuuchahnulth morphology in some detail. The nature of the word is brought to the forefront in this language where even the status of the word itself is called into question. We have witnessed the importance of suffixation for the grammar and have seen that the processes of vowel lengthening, shortening, infixation, and, most notably, reduplication figure prominently in the morphology. We have considered theoretical approaches to dealing with the complexities of the grammar, including multi-stratal treatments and the use of templatic morphology. The array of templates necessary for dealing with the morphology is large and, in some cases, complex. Clitics play a role in the grammar and we have seen that a clitic-based approach to inflectional morphology can yield interesting results in explaining the complexities of morphosyntax. Finally, the nature of incorporation and related processes in Nuuchahnulth, and in all Wakashan languages, is both complex and theoretically challenging. Distinctions must be made between lexical and syntactic incorporation and, furthermore, between incorporation and the cliticisation of inflection. It is hoped that the presentation of Nuuchahnulth morphology in this book will prompt others to delve further into the complexities presented here and to find better accounts of the various processes discussed here.

272

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Conclusions

Conclusions 273

A number of theoretical issues have arisen from the discussion in this book and they have important consequences for the formulation of new theories and for the viability of existing theories of grammar. Hopefully, there is ample data to illustrate the problems and, in most cases, there is additional data available in the published sources discussed in Chapter 1. The following may be seen as some of the key areas worthy of further investigation and (re-)analysis. 9.2.1

The Status of Word Categories

The discussion of category neutrality and the status of word categories in Nuuchahnulth remains an open question. The ability to determine word class is important for assigning phrase structure, and thus for determining order of incorporation, as discussed in Chapter 8. Further tests must be developed for probing the relationships that exist between words. Furthermore, what is the status of aspect vis-a-vis verbs: what is inherent aspect, as first mentioned in Sapir & Swadesh (1939:240)? What are the class(es) of verbs exhibiting certain aspects inherently? 9.2.2 The Complexity of the Lexicon Following on from the previous section, much work has yet to be done on subcategorisation and the nature of predicate argument structure in Nuuchahnulth. Swadesh (1939) used the term 'word-internal syntax' to refer to morphological structure and it would be interesting to see how far the Mirror Principle holds in Nuuchahnulth. In addition, how do the various valency-altering morphemes operate in the grammar and what is their relationship to arguments in the syntax? And the nature of - 'at PASS remains unclear, in spite of all the work that has been done on it. 9.2.3 Treatment of Reduplication Nuuchahnulth reduplication poses challenges for morphological theories such as Templatic Morphology and Optimality Theory. The wide array of structures, both syllable- and foot-based, and the complexities they pose are important and intricate. In particular, the issue of full syllable reduplication and the foot-based templates present crucial topics for further study.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

9.2 Theoretical Issues

274 Linguistic Theory and Complex Words

Furthermore, double reduplication raises issues about constraint-based accounts and level-ordering. The absence of double reduplication in Ahousaht suggests that a thorough dialect survey would be of value in establishing the extent of this phenomenon. Suffix-triggered Stem Changes

The suffix-triggered stem changes, such as lengthening, shortening, infixation, and reduplication, present an interesting argument for discontinuous morphemes, since these elements obligatorily co-occur in a form, and yet they are never adjacent. How do morphological theories explain this kind of behaviour without appealing to discontinuous morphemes, a rather undesirable construct. A further question involves the origin of the process. Are there clear semantic factors influencing the nature of the trigger? Does the triggering arise from any particular historical source? 9.2.5

Action-at-a-distance

The fact that suffix-triggered changes may be effected across intervening suffixes, provides crucial evidence of action-at-a-distance. Effects upon the root, including vowel lengthening and shortening, infixation, and reduplication, may be induced by morphemes attached at some distance from the root. This raises questions regarding cyclicity, bracket erasure, and the nature of circumfixes. It furthermore calls into question the notion that reduplication may be treated as a morpheme itself (e.g., Marantz 1982), since in this situation that is clearly not the case. 9.2.6

Phonologically-conditioned Infixation

Infixation caused by obligatory requirements on the shape of the base constitutes another interesting phenomenon found in Nuuchahnulth. The fact that the inserted elements do not augment the weight of the base, since they are not moraic, presents further concerns. And the variation between -X- and -c- in this process raises additional questions, since one would expect only one default consonant for insertion, and although both of these are coronal, they are far from proto-typical default consonants. 9.2.7

Incorporation

Obviously, the facts of incorporation in Nuuchahnulth pose many questions for morphologists and syntacticians. The existence of the lexical

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

9.2.4

275

form of incorporation, akin to compounding in other languages, allows for the construction of idiosyncratic combinations that may then operate as lexical items. Syntactic incorporation poses different and extremely interesting problems, involving as it does many different elements of the object phrase. The fact that incorporation may involve members of the object other than the head noun, combined with the observation that these are strictly ordered and incorporated in this fashion, and the further condition that certain elements, notably deictics may not be incorporated, suggests a rather complex process of incorporation in Nuuchahnulth. Incorporation into certain intransitive verbs suggests that this process must apply at an underlying level of the grammar and the fact that it may be an adjective or quantifier that is incorporated indicates that it cannot be lexical. Incorporation of quantifiers, adjectives and other elements of the object in general raises serious questions not easily explained. 9.2.8

Compounding vs. Pseudo-Compounding

The absence of regular, lexical compounding in Nuuchahnulth marks the language as typologically special, if not unique. Its polysynthetic nature makes the situation even more curious. Nevertheless, the language employs a syntactic strategy for the creation of compounds, providing an open-ended source for new lexical items by means of the collocation of nouns, along similar lines to the English constructions such as kick the bucket, keep tabs on, etc. Questions arise over the extent of compounding of this type: can it involve adjectives? verbs? Furthermore, how open is the process in forming new lexical items. Can we find any evidence of lexicalisation, resulting in lexical compounds? 9.2.9 Hypocoristic Formation and Templatic Morphology The templatic nature of hypocoristic formation, basically a foot-based strategy, provides another example of the complex types of template utilised in the language. Superficially subtractive in nature, closer examination reveals the use of prosodic circumscription in conjunction with partial prespecification of the template. It is worth investigating the possibility of other templatic effects of this type, some of which have been briefly examined in preceding chapters.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Conclusions

276 Linguistic Theory and Complex Words

The issue of stratal segregation and the multi-stratal nature of the Nuuchahnulth lexicon seems beyond question. This separation of stem and word levels explains a number of difficult issues in the grammar, including the variable application of glottalisation and the existence of double reduplication. However, this separation of word and stem levels poses a major challenge to mono-stratal approaches, such as standard Optimality Theory (McCarthy & Prince 1993, etc.), and it will be important to exhaust all avenues of investigation in determining whether alternative means of analysis for the issues discussed in Chapter 6 exist. 9.2.11 Special Speech Forms We have examined only one form of special or abnormal speech, that of Pitch Woman, which involves a rather complex, foot-based template with prespecification. However, there are numerous other forms of special speech, involving various mutations of the base (Sapir 1915), and it may be worth exploring the nature and extent of this process in Nuuchahnulth and the other Wakashan languages. 9.2.12 Inflection as Clisis The proposal for placing large portions of the inflectional morphology into the syntax as cliticisation opens the doors for further investigation into the relationship between syntax and morphology in Nuuchahnulth. There has been some initial work on the symmetry between IP and DP (Yiu & Stonham 2004), but there remains much work to be carried out in this area of inquiry. Current Minimalist approaches to syntax would appear to be conducive to such a clitic-based analysis of inflection, especially with respect to the status of functional heads, and this may prove a fruitful avenue of investigation for anyone interested in this area. 9.2.13 Dialectal Variation In this account, data have been provided from three dialects of Nuuchahnulth, Kyuquot, Ahousaht, and Tsishaath. However, as discussed in Chapter 1, there are a number of other dialects and there may be more variation than is evident from the discussion here. Already we have seen the variation that exists in double reduplication, discussed in Chapter 6, but this may extend to other phenomena discussed in this book.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

9.2.10 Stratal Distinctiveness

Nuuchahnulth Classifiers (i) -C*iq

iong objects'

a. nupdiqPi nup -diq =?r

mftaat miTaat

'one sockeye salmon' one -long objects -DEF sockeye salmon

b. muudiq hiinaanuhsim muu -diq hiinaanuhsim

'four whaling spears' four -long objects whaling spear

c. PaXdiqPisuk diihati PaXa -diq -Pis -uk diihati

'his two arrows' two -long objects -DIM -POSS arrow

(ii) -btalf

'containers'

a. hiiaasPaX XapaatPi PaXhtaak 'on it were two storage baskets' hii-aVaX Xapa"t =?r PaXa-hta-k01 LOC-on-NOW basket=DEF two-container b. nuphtaak nup -hta^k"

fapqimi t'apqimi

'... for each bale' one -container bale

c. £a£iXsi hayuhtak XiXickuk 'I potlatched ten sacks of flour' fja-siX-sr hayu-hta-k01 XiXickuk potlatch -MOM-Is.ABS ten-containers flour (iii) -frUvyuk

'strings'

a. suudahtayuk hiixwaa suda -htaryuk hiix^aa

'five strings of dentalia' five -strings [L] dentalia

b. caaqiichtayuk caqix -[L] -hta-yuk

'twenty lengths (of dentalia)' twenty -strings [L]

(iv) -firi*

'long objects'

a. PaXrMiiPi PaXa -$ri" =?i

ficsyin Ticsyin

'the two spines' two -long objs=DEF spine

b. muu^iiiuk Xama muu -$'v¥° -uk Xama

'his four house posts' four -long objs -POSS house post

c. PaXjMiiPi PaXa -^fT =?r

'the two cannons' two -long objs =DEF cannon

(v) -firi"

saasaqi saasaqi

'songs'

a. PaXjMiiPi nuukrfiinh PaXa -£ri" =Pir nuuk -rhinh

'the two songs' two -songs-DEF song -PL

277

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Appendix

278 Appendix b. nupfjiii nup -r>rT

'one spirit song' one -songs spirit song

'unit(s)'

a. hayuqumiPi hayu -qimi =?r

nuutyak nuutyak

b. hayuqumi mucmuhaq hayu -qimi mucmuhaq (vii) sa^tak"

'the ten hoops' ten -unit =DEF hoops 'ten bear skins' ten -unit bear skin

'kind(s)'

a. muusaahtak Suyi muu -sa'htak Tuyi

'four medicines' four-kinds medicine

b. muusaahtak muu -sa'htak

'four ceremonial privileges' four-kinds ceremonial privilege

(viii) -sartb

tupaati tupaati

'tribe(s)'

a. PaXsaathPi maatmaas PaXa -saYth -?'v maatmaas

'two tribes' two -tribes=DEF

b. hii§saath?iX hi§ -[L] -sa'th -'iX

'he invited all the tribes' all -tribes -invite [L]

(ix) -taqak

tribe

'unit(s)'

a. hayustaqisuk paacsadim hayu -taqak -'is -uk rjaacsadim

'his ten potlatch platforms' ten-units-on beach-poss potlatch platform

b. nuptaqakukwah siiyassa nup-taqak-uk-(m)a"h siya-sasa

'I have one thing of my very own' one -units-POSS-ls.iND I-only[L]

(x) -taqimi

'bundle(s)'

a. hayu§taqimi maatmaas hayu -taqimi maatmaas

'ten tribes' ten -bundle

tribe

b. PaXtaqimi huux^apiih distuup 'two balls of rope' PaXa -taqimi huux^apiih distuup two -bundle ball rope (xi) -ya*sda a. PaaXyaasda PaXa -ya'sda

'roll(s), line(s)' ^aaXyaqanui SaaXyaqanui

b. PaaXyaasdaqiics PaXa -yarsda -iics

£iics Ciics

'two rolls of cedar-branch rope' two -rolls [L] cedar-branch-rope 'they trolled with two lines' two -rolls -go along troll

c. PaayasdaTaqstumah disaiaPap 'I have tied on many lines' Paya-yaYsda-'aqstuX-(m)a-h dis-ai- ap many -lines -inside [M] -IS.IND in line -attach to -CAUS

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

(vi) -qimi

didihinkyak didihinkyak

Chapter 1 1. For a recent treatment of the morphology of one of the Northern languages, see Howe (2000) on Oowekeno. 2. It would seem that the discussion of Nuuchahnulth in Sproat (1868) is more or less directly drawn from Knipe (1868), as was first pointed out in Boas (1890:31): 'the linguistic part of his [i.e. Sproat's] book is taken almost bodily from an anonymous work by a Catholic Missionary, named Knipping [sic], 'Some Account of the Tahkaht Language as spoken by several tribes in the Western Coast of Vancouver Island' (London, 1868), which latter book has remained almost unknown. The power of observation exhibited in the descriptions of the author, however, is not to be depreciated'. 3. See for example Thomas & Arima (1968), Powell (1994). 4. See the article in Hashithlsa, (1978) the news organ of the Nuuchahnulth Tribal Council. 5. There are various reasons for disprefering the term 'Nootka', although there is no pejorative connotation to the word. 6. Oxford English Dictionary, (2nd ed). 7. A further curious usage that has arisen in this context is the inclusion of hyphens, 'Nuu-chah-nulth', presumably under the misguided notion that this will somehow facilitate the pronunciation of the term. 8. The term 'West Coast people' is first attested in the work of Terry Klokeid (ms). Chapter 2 1. Abbreviations for the heading of the table are: Alv. Alveolar, Lat. Lateral, Pharyn. Pharyngeal. iw and hw are both morphophonemes, relevant for certain processes in the language but never appearing on the surface, x and x" are extremely rare and are usually either borrowed or appear in personal names. 2. The only sonorants that appear in the coda on the surface are nasals. See Stonham (1999b) for further discussion. 3. Klokeid (ms.) suggests that there may be some dialectal variation on the placement of stress, possibly necessitating a revision of this characterisation, but there is insufficient data to confirm this at the present. 4. See Kim (2001, 2003c) for a detailed treatment of this phenomenon in Nuuchahnulth. 5. Note another morphophonological rule of Nuuchahnulth at work here: the labialisation on the morpheme -narka 'have,' disappears when it occurs im-

279

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Notes

6.

7.

8. 9. 10.

11.

mediately before an inflectional suffix. For more discussion of this rule, see Section 2.5.7. The symbols [L] and [R] represent lengthening and reduplication triggered by the suffix, respectively. These and other diacritics will be discussed in greater detail in Chapter 4 on suffixation. These two examples demonstrate the interesting property of instability in the underlying representation of a single morpheme. Thanks to Stephen Matthews (p.c.) for suggesting the notion of 'instability' based on similar experiences working with the Daghestanian language, Tsez. But see Kim (2003, Section 3.2.5) for a recent account. See Section 5.1.6 for a discussion of double reduplication in Nuuchahnulth. It should be noted that i" and hm must be considered morphophonemes that do not appear on the surface, but do play a role in various processes, including glottalisation and lenition. This example and the next one have undergone the further process of "V—HI, itself a result of the labialisation of the /q/.

Chapter 3 1. See Section 8.3 for a discussion of phrase-level compounding in Nuuchahnulth. 2. ?u and the locative bases will be discussed in greater detail in section 3.4.3. 3. The only real complication to this generalisation is the existence of a small number of apparent root morphemes involving the repetition of identical consonants and vowels, e.g. C7C7CJ 'tooth', ftjft/Su 'chicken', etc. 4. This issue will be explored further in Chapter 6 on levels within the grammar. 5. hinaq is the bound form of hiina and -apiq is the combining form of -api. 6. In fact, ?u has all the markings of a prefix, similar to the semantically obscure prefixes in English forms such as deceive, infer, resist, etc. 7. See Fortescue (1985), for example. 8. Emmon Bach, p.c. 9. Mithun (1984:888) cites spurious examples of what she calls 'etymological relationships between stems and suffixes'. She provides the examples Punaak and -nak (sic), which the reader will now recognise as involving the same bound verb, -narkm 'have', combined with ?u REF in the former case. 10. The root nu(p)- 'one' alternates with the root cawar 'one' in a number of different contexts. 11. In all cases encountered in the data, the combination of a number + -fit 'times (= multiplier)' combined with hayuuq results in the preceding number also bearing the suffix -uuq, which is otherwise unattested. In the case of hayuuq, it is analysed as hayu + -iiq 'score'. 12. Note that the asterisk is used here and elsewhere to indicate unattested forms and readings in the large database of Nuuchahnulth textual material that serves as the source for much of the data in this book.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

280 Notes

281

13. Jacobsen (1979b) cites one case of an apparent counterexample to the generalisations stated in this section but we have been unable to find any comparable examples in our extensive data on Nuuchahnulth. It may be that the element outside of the verb is in fact an adjunct, rather than part of the argument of the verb. Alternatively, Jacobsen notes that there are differences in a number of areas between Nuuchahnulth and the more southerly Ditidaht and Makah. 14. An alternative view is that there is raising into an INFL position. Chapter 4 1. In the following discussion paradigms and examples from each major branch, i.e., Southern (Tsishaath), Central (Ahousaht), and Northern (Kyuquot) Nuuchahnulth are provided. Examples marked (Tsish) are drawn from Sapir's materials, those marked (Ahous) from Nakayama (1997a), and those marked (Kyu) from Rose (1981) unless otherwise noted. 2. For discussion of the inferential in a related language, see Carlson & Thomas (1979). 3. Note the 1st person plural, which seems a more likely underlying representation for the form than the Tsishaath -qun. 4. See Lecarme (to appear), which discusses other cases of tense marking on nominals, in particular in East Cushitic languages but also in Wakashan (Kwak'wala (Boas 1947)) and Salish (St'at'imcets (Lillooet) (Demirdache 1997), and Halkomelem (Burton 1997, Wiltschko 2001)). 5. Note the application here of the assimilation rule discussed in Section 4.3.1.1. 6. Thanks to Ben Braithwaite for pointing this example out. 7. Actually, the coreference of 'his' in 'his head' is ambiguous in this case, since the head was originally part of one individual^ but is now a trophy of the other individual^ 8. The symbol /=/ will be employed to indicate a clitic boundary. 9. For a detailed discussion of clitics, see Section 8.1. 10. It is not clear when the allomorph -?ic of the diminutive surfaces, but this is one example of it. 11. See Sapir (1915), Stonham (1999b), and Chapter 7 for further discussion of this aspect of Nuuchahnulth grammar. 12. Davidson (2002:303) notes for a similar example that '[d]iscussing the text later, the speaker indicated that inclusion of the diminutive in the greeting has the force of an insult — such a greeting would normally only be given to a child or slave.' 13. For a more detailed discussion of classifiers in Nuuchahnulth, and a comparison with another classifier language, Cantonese, see Yiu & Stonham (2002). 14. Classifiers are regular because the same classifier is always employed in the same context, but they are arbitrary since there is no obvious semantic or other determiner for which classifier is used.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Notes

282 Notes

1. In fact, this is the traditional account. There are a number of cases where it would make more sense to refer to the stem rather than the root, for example diicstiicsfaqX 'expert troller' < DUP- diics 'troll' -'aqX 'expert', fries < 6ii 'pull' + -iics 'go along'. One might argue for the opacity of the stem, although both morphemes are easily found in many other environments. 2. In addition to reduplication, the plural may be marked by infixation, suffixation, a combination of these, or may be unmarked. 3. The absence of length here is due to the moraic nature of coda nasals discussed in Chapter 2. 4. There are further complications here depending on the prosodic structure of the root, but these are not relevant to the point at hand. 5. Note here that any explanation in of the reduplication in terms of a pairing of body parts falls short of the mark, since the case in (19a) does not involve the wolf breaking both legs. The pairing of body parts may have played some role in the origin of the requirement of reduplication, but it does not appear to play any current role. 6. I will argue against this analysis on the basis of data introduced further on in this chapter. 7. A further important point to note here is that bracket erasure must occur no earlier than the end of the stem level, given that the rule inserting /X/ must have access to the information that the root consists of a single open syllable, which is not the case at the end of the stem level, thus: [[DUP- [k^i] -iyaqhlsiEM -PaX]WORD 'filing while singing' vs. [[DUP- [qah] -a]STEM]WoRD 'killing' DUP- qah -(y)a-

REP- kill -DUR

=> qaahqaaha NOT: *qaaXqaaha 8. The vowel in hatiii becomes Am/ through a rule of coalescence discussed in more detail in Stonham (1999b). Chapter 6 1. The symbol || represents the boundary of stem- and word-level morphology here. 2. Refer to Chapter 2 for a more detailed discussion of this and any other basically phonological processes discussed in this chapter. 3. See Kim (2003c) for one theoretical treatment of this phenomenon. 4. The /X/ in the surface form here is the result of the application of a special speech form representing the speech of Mink. 5. It would appear that continuancy is an important factor in the process of glottalisation, dividing up the consonants into two groups, stops/affricates versus continuants.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Chapter 5

283

6. Kim (2003c) notes that lenition in Ahousaht applies only to alveolar fricatives, not to palatal or back fricatives. This suggests a radical restriction of the rule in this variety of Nuuchahnulth. 7. The /X/ in the morpheme -'(i)saA disappears due to a deletion rule discussed in Chapter 2. 8. Note that this is quite different from the facts for Kwak'wala, a Northern Wakashan language, as discussed in Zee (1988). 9. Phonologically, there are several additional processes operating in this case, including the reduction of the aspectual lengthening of the root vowel, indicating that the lengthening appears to have effect only in the first or second syllable, reminiscent of the restrictions on variable-length vowels. We will not be able to pursue this issue here. 10. The morphemes involved are: ka- 'object protrudes' and -wad [R] 'at the middle', referring to the upright dorsal fin of the killer whale. The Nuuchahnulth equivalent is kakawin. 11. Thanks to Ricardo Bermudez-Otero for pointing out this research. 12. An alternative solution might be that there have been further restrictions imposed on the grammar of Ahousaht that conspire to eliminate the environments for double reduplications. One possibility is that bracket erasure only occurs at the end of the lexicon rather than at the end of each level, allowing the existence of the copy to be observed throughout the grammar. Theoretically, this seems rather less desirable. 13. The underlined portion constitutes the base for purposes of this section.

Chapter 7 1. From the available examples, it appears that the nucleus of the copy is always long and will be represented this way in the template. 2. Note the shortening of the final vowel here, indicating that it is underlyingly variable in length. 3. For a more detailed exposition of this phenomenon, see Stonham (1994b). 4. It should be noted that /h/ is a possible [+Glottal] element, but it will never occur in this formation since, as discussed in Chapter 2, it only occurs wordinitially and in the copy of a word-initial /h/. 5. This strategy suggests a melody-driven mechanism of linking. 6. On the Sonority Hierarchy see, for example, Selkirk (1984). 7. See Stonham (1999b) for a more detailed discussion of the phonetics and phonology of this process. Chapter 8 1. Arguments concerning the exact translation of a morpheme, while interesting for other reasons, will turn out not to be fruitful in resolving the differences discussed here. Separate lexical items for women marrying and men marrying are common in many languages, in Cantonese, for instance.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Notes

2. Rose (1981) provides an appendix outlining the requirements of different bound verbs. 3. It is important to note here that (23c) cannot mean 'well-stocked with', a meaning more closely related to (23b). 4. Swadesh includes the element -iiq0 'score' in his list, but we will argue against this morpheme as a classifier later in this section. It should be noted that Swadesh makes no claim regarding the status of these elements as classifiers. 5. We will appeal to the notion of 'feature' throughout this paper without actually committing to a specific set of features. We leave any such determination of featural specifications to a more in-depth, cross-linguistic survey of classifier systems. 6. A (non-exhaustive) list of Nuuchahnulth members of these non-intrinsic classifier classes is, however, provided in an appendix at the end of the monograph for the reader's information. 7. Rose's typology merges the categories of intrinsic and extrinsic classifiers. For more information on her system of categorization, cf. Rose (1981). 8. There are certain ambiguous cases that we have not included in this list, for instance, -ista 'person(s) in canoe', in the example: XiihsiPaX PaXistaX fiik 'now the two brothers paddled off in their canoe' Xih-siX-'aX PaX-ista-'aX ?iik paddle-MOM-NOW two -in canoe-Now brothers This would seem to be an instance of an intrinsic classifier. The reason for not considering this to be one is that in most other cases it behaves more as a locative than a classifier, as in the example below: haawiihaXista haawiihaX -ista

'the young men in the canoe' young men -persons in canoe

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

284 Notes

Alderete, J., J. Beckman, L. Benua, A. Gnanadesikan, J J . McCarthy & S. Urbanczyk. 1999. 'Reduplication with Fixed Segmentism.' Linguistic Inquiry 30: 3, 327-364. Allan, Keith. 1977. 'Classifiers.' Language, 53(2): 285-311. Anderson, S.R. 1988. 'Morphological Theory.' Linguistics: The Cambridge Survey, vol. 1. Cambridge: Cambridge University Press. Badecker, W. & A. Caramazza. 1989. 'A Lexical Distinction Between Inflection and Derivation.' Linguistic Inquiry, vol. 20(1), 108-116. Baker, Mark. 1988. Incorporation: A theory of grammatical function changing. Chicago: University of Chicago Press. Baker, Mark. 1995. 'Lexical and Nonlexical Noun Incorporation.' U. Egli, P. Pause, C. Schwarze, A. von Stechow, & G. Wienold (eds) Lexical Knowledge in the Organization of Language, John Benjamins, Amsterdam, pp. 3-33. Bauer, Laurie. 1983. English Word Formation. Cambridge: Cambridge University Press. Bermudez-Otero, R. 1999. Constraint Interaction In Language Change. Ph.D. dissertation, University of Manchester/University de Santiago de Compostela. Bermudez-Otero, R. 2003. 'The Acquisition of Phonological Opacity.' Rutgers Optimality Archive, ROA 593-0403. Bermudez-Otero, R. in prep. The Lifecycle of Constraint Rankings: Studies in early English morphophonology. Bermudez-Otero, R. & R.M. Hogg. 2003. 'The Actuation Problem in Optimality Theory: Phonologization, rule inversion, and rule loss.' D. Eric Holt (ed.), Optimality Theory And Language Change, 1-40. Netherlands: Kluwer Academic. Bickel, Balthasar & Johanna Nichols. Forthcoming. 'Inflectional Morphology.' In T. Shopen (ed.) Language Typology and Syntactic Description. (2nd edn.). Cambridge: Cambridge University Press Bloomfield, Leonard. 1933. Language. Holt, Rinehart & Winston: New York. Boas, Franz. 1890. 'The Nootka.' Report of the 60th Meeting of the British Association for the Advancement of Science (BAAS-R 60), 582-604. Boas, Franz. 1900. 'Sketch of the Kwakiutl Language.' American Anthropologist, n.s. 2: 708-721. Boas, Franz. 1911. 'The Kwakiutl Language.' Handbook of American Indian Languages, vol. 1: 708-721. Boas, Franz. 1947. 'Kwakiutl Grammar.' Transactions of the American Philosophical Society, n.s. vol. 37.

285

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Bibliography

Broselow, E. 1982. 'Interactions of Reduplication Processes: Evidence from Salish.' Texas Linguistic Forum, vol. 21, pp: 21-46. Broselow, E. 1983. 'Salish Double Reduplications, Subjacency in Morphology.' Natural Language and Linguistic Theory, vol.1, no. 3: 317-346. Broselow, E. & J. McCarthy. 1984. 'A Theory of Internal Reduplication.' The Linguistic Review, vol. 3, pp: 25-88. Buckley, Eugene. 1998. 'Integrity and Correspondence in Manam Double Reduplication.' Proceedings of the Northeast Linguistic Society (NELS) 28. Burton, Strang. 1987. 'Past Tense on Nouns as Death, Destruction, and Loss.' Kiyomi Kusumoto, (ed.), Proceedings of the Northeast Linguistic Society (NELS) 27. Carlson, Barry F. & John Thomas. 1979. 'The Nitinaht Inferential.' Anthropological Linguistics 21, Oct. 317-27. Carstairs, A. 1987. Allomorphy in Inflexion. Beckenham: Croom Helm. Chomsky, N. 1970. 'Remarks on Nominalizations.' R. Jacobs & P. Rosenbaum (eds), Readings in Transformational Grammar. Waltham, MA: Blaisdell. Comrie, Bernard. 1981. Language Universals and Linguistic Typology. Chicago: University of Chicago Press. Comrie, Bernard. 1985. Tense. Cambridge: Cambridge University Press. Davidson, Matthew. 2002. Studies in Southern Wakashan (Nootkan) Grammar. Doctoral dissertation, State University of New York at Buffalo. de Saussure, F. 1974. Course in General Linguistics. New York: Fontana. Demirdache, Hamida. 1997. 'A Crosslinguistic Asymmetry in the Temporal Interpretation of DPs.' Anthony Green & Virginia Motapanyane, (eds), Proceedings of the Thirteenth Eastern States Conference on Linguistics (ESCOL). Ithaca, NY: Cornell Linguistic Circle. Drucker, Philip. 1951. The Northern and Central Nootkan Tribes. Washington, D.C.: Bureau of American Ethnology Bulletin 144: 1-480. Ellis, David W. & Luke Swan. 1981. Teachings of the Tides: Uses of marine invertebrates by the Manhousat People. Nanaimo: Theytus Books. Emanatian. Michele. 1988. 'The Nootka Passive Revisited.' In William Shipley (ed.), In Honor of Mary Haas. Berlin: Mouton de Gruyter, pp. 265-291. Fortescue, Michael. 1984. West Greenlandic. Amsterdam: Croom Helm. Gleason, H.A, Jr. 1961. An Introduction to Descriptive Linguistics, rev. edn. Toronto: Holt, Rinehart & Winston. Greenberg, Joseph H. 1972. 'Numeral Classifiers and Substantival Number: Problems in the genesis of a linguistic type.' Working Papers on Language Universals 9:1-39, Stanford: Department of Linguistics. Haas, Mary R. & Morris Swadesh. 1932. 'A Visit to the Other World, a Nitinat Text (with Translation and Grammatical Analysis).' International Journal of American Linguistics 7:195-208.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

286 Bibliography

287

Haas, Mary R. 1969a. 'Internal Reconstruction of the Nootka-Nitinat Pronominal Suffixes.' International Journal of American Linguistics 35(2): 108-124. Haas, Mary R. 1972. 'The Structure of Stems and Roots in Nootka-Nitinat.' International Journal of American Linguistics 38(2): 83-92. Haegemann, L. 1994. Introduction to Government & Binding Theory. 2 nd edn. Oxford: Blackwell. Hale, K. & S.J. Keyser. 1993. 'On Argument Structure and the Lexical Expression of Semantic Relations.' K. Hale & S.J. Keyser (eds) The View from Building 20, Cambridge, Mass.: MIT Press. Ha-Shilth-Sa Newspaper [haSiisa]. 1978. 'Nuu-Chah-Nulth-Aht not Nootka.' Port Alberni, 10 Nov. 1978, p. 4. Hess, T.M. 1990. 'A Note on Nitinaht Numerals.' International Journal of American Linguistics, 56, 3: 427-431. Howe, Darin. 2000. Oowekyala Segmental Phonology. Ph.D. thesis. University of British Columbia. Huddleston, R. 1984. Introduction to the Grammar of English. Cambridge: Cambridge University Press. Jacobsen, William H., Jr. 1979a. 'Wakashan Comparative Studies.' Lyle Campbell & Marianne Mithun (eds), The Languages of Native America. Austin, Texas: University of Texas Press, pp. 766-791. Jacobsen, William H., Jr. 1979b. 'Noun and Verb in Nootkan.' Barbara S. Efrat (ed.), The Victoria Conference on Northwestern Languages, 83-155. Victoria: British Columbia Provincial Museum Heritage Record, no. 4. Jacobsen, William H., Jr. 1994. 'Nootkan Vocative Vocalism and Its Implications.' Leanne Hinton, Johanna Nichols, & John J. Ohala, (eds), Sound Symbolism, Cambridge, England: Cambridge University Press, 1994, pp 23-39. Kaisse, E.M. 1993. 'Rule Reordering and Rule Generalization in Lexical Phonology: A reconsideration.' S. Hargus & E.M. Kaisse (eds), Studies in Lexical Phonology. San Diego: Academic Press, pp. 343-64. Keenan, E. & B. Comrie. 1977. 'Noun Phrase Accessibility and Universal Grammar.' Linguistic Inquiry, 8: 62-100. Kim, Eun-Sook. 1999a. 'Hypocoristics of Nuu-chah-nulth.' M. Caldecott, S. Gessner, & E.-S. Kim, (eds.) University of British Columbia Working Papers in Linguistics, vol. 1: 125-138. Kim, Eun-Sook. 1999b. 'Glottalization in Nuu-chah-nulth in Optimality Theory.' John Kyle (ed.), Kansas Working Papers in Linguistics, vol. 24: 2, 85-100. Kim, Eun-Sook. 2000a. 'Glottalization in Nuu-chah-nulth (Nootka): A module interaction case.' S. Gessner, S.-Y. Oh, & K. Shiobara (eds), Papers from the 35* International Conference on Salish and Neighbouring Languages, UBCWPL, vol. 3: 107-117.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Bibliography

Kim, Eun-Sook. 2000b. 'The Morphology and Syntax of -?at in Nuu-chah-nulth in the Minimalist Theory.' H.-K. Moon (ed.), In Honour of Professor YoungJae Lee, pp. 237-281. Korea University Press. Kim, Eun-Sook. 2001. 'Glottalization in Southern Wakashan.' L. Ba-rel, L. Watt, & I. Wilson (eds), Papers from the 36 International Conference on Salish and Neighbouring Languages, UBCWPL, vol 6: 175-188. University of British Columbia. Kim, Eun-Sook. 2002. 'The Consequences of Richness of the Base for Glottalization of Nuu-chah-nulth Obstruents.' Proceedings of WECOL 2001. University of California, Fresno. Kim, Eun-Sook. 2003a. 'Patterns of Reduplication in Nuu-chah-nulth.' M. Kadowaki, & S. Kawahara (eds.), Proceedings of NELS 33, 127-146. University of Massachusetts, Amherst. Kim, Eun-Sook. 2003b. 'Delabialisation in Nuu-chah-nulth.' J. Brown & M. Kalmar (eds), Papers from the 38th International Conference on Salish and Neighbouring Languages, UBCWPL. Kim, Eun-Sook. 2003c. Theoretical Issues in Nuu-chah-nulth Phonology and Morphology. Ph.D. dissertation, University of British Columbia. Kim, Eun-Sook & Gessiane Picanco. 2003. 'Reconsidering Fixed Segmentism in Reduplication.' Paper delivered at the GLOW colloquium 2003. 9-11 April. Lund, Sweden: University of Lund. Kim, Eun-Sook & Rachel Wojdak. 2002. 'A Survey of Nuu-chah-nulth Reduplication.' C. Gillon, N. Sawai, & R. Wojdak (eds.), The 37th International Conference on Salish and Neighbouring Languages. University of British Columbia. 189-202. Kiparsky, Paul. 1982. 'Lexical Morphology and Phonology.' I.-S. Yang (ed.) Linguistics in the Morning Calm, 3-91, Seoul: Hanshin. Kiparsky, Paul. 1985. 'Some Consequences of Lexical Phonology.' C.J. Ewen & J.M. Anderson (eds.) Phonology Yearbook, 2.85-138. Kiparsky, Paul. 2000. 'Opacity and Cyclicity.' The Linguistic Review 17: 351365. Kiparsky, Paul. 2003. 'Diachronic Evidence for Stratal OT.' Talk given at the Manchester Phonology Conference, May 2003. Klokeid, Terry J. ms. [1972] An Introduction to the West Coast [Nootka] Language of Vancouver Island. Ottawa: Canadian Museum of Civilization. Klokeid, Terry J. 1976a. 'Encliticization in Nitinaht.' 11th International Conference on Salish Languages, Seattle: University of Washington, 211-46. Klokeid, Terry J. 1978. 'Surface Structure Constraints and Nitinaht Enclitics.' E.-D. Cook and J. Kaye (eds), Linguistic Studies of Native Canada, 157-76. Vancouver: University of British Columbia Press. Knipe, C. 1868. Some Account of the Tahkaht Language, as Spoken by Several Tribes on the Western Coast of Vancouver Island. London: Hatchard.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

288 Bibliography

289

Kroeber, A.L. 1910. 'Noun Incorporation in American Languages.' Verhandlungen des XVI. Internationalen Amerikanisten-Kongresses, Wien 9. bis 14. September 1908, Zweite Halfte, pp. 569-76. Wien and Leipzig. [Reprinted in Sapir, Collected Works, V, 1, 541-546.] Kroeber, A.L. 1911. 'Incorporation as a Linguistic Process.' American Anthropologist, n.s., vol. 13, no. 4, pp. 577-84. [Reprinted in Sapir, Collected Works V, 1,547-557.] Lecarme, Jacqueline, (to appear). 'Tense in Nominals.' J. Gueron & J. Lecarme, eds., The Syntax of Time, MIT Press. Marantz, A. 1982. 'Re reduplication' Linguistic Inquiry, vol. 13.3: 435-482. Matthews, S. & V. Yip. 1994. Cantonese: A Comprehensive Grammar. London: Routledge. McCarthy, John J. 1981. 'A Prosodic Theory of Non-concatenative Morphology.' Linguistic Inquiry, 12: 373-418. McCarthy, John & Alan Prince. [1993]. Prosodic Morphology I: Constraint interaction and Satisfaction. U Mass and Rutgers. McCarthy, John & Alan S. Prince. 1995. 'Faithfulness and Reduplicative Identity.' UMass Occasional Papers 18, 249-384. McCarthy, John & Alan S. Prince. 1998. 'Prosodic Morphology.' Andrew Spencer & Arnold M. Zwicky, (eds). The Handbook of Morphology. Oxford: Basil Blackwell. Mithun, Marianne. 1984. 'The Evolution of Noun Incorporation.' Language 60, pp. 847-94. Mohanan, K.P. 1981. Lexical Phonology. MIT doctoral dissertation. Mohanan, K.P. 1986. The Theory of Lexical Phonology. Dordrecht: D. Reidel. 219 pp. Nakayama, Toshihide. 1997a. Discourse-pragmatic Dynamism in Nuu-chahnulth (Nootka) Morphosyntax. Ph.D. dissertation, University of California at Santa Barbara. Nakayama, Toshihide. 1997b. 'Functions of the Nootka (Nuu-chah-nulth) "Passive" Suffix.' International Journal of American Linguistics 63.3: 412-31. Nakayama, Toshihide. 2001. Nuuchahnulth (Nootka) Morphosyntax. University of California Publications in Linguistics, 134. Berkeley: University of California Press. Paik, Tai Shung. 1968. Clayoquot Phonology. M.A. thesis, University of Washington. Powell, J. 1991. Our World, Our Ways: T'aataaqsapa cultural dictionary. Nuuchahnulth Tribal Council. Prince, Alan & Paul Smolensky. ms [1993]. Optimality Theory: Constraint interaction in generative grammar. Rutgers University & University of Colorado.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Bibliography

Renker, Ann M. 1988. Rethinking Noun and Verb: An investigation of AUX in a Southern Wakashan language. American University doctoral dissertation, Washington, D.C. Rose, Suzanne M. 1976. Lenition and Glottalization in Nootka. M.A. thesis, University of Victoria. Rose, Suzanne M. 1981. Kyuquot Grammar. Ph.D. dissertation, University of Victoria. Rose, Suzanne M. & Barry F. Carlson. 1984. 'The Nootka-Nitinaht Passive.' Anthropological Linguistics 26(1): 1-12. Sapir, Edward, n.d. Fieldnotes on Nootka. In Boas Collection of the American Philosophical Society [W2a.l8]. Sapir, Edward. 1911. 'The Problem of Noun Incorporation in American Languages.' American Anthropologist 13: 250-282. Sapir, Edward. 1915. 'Abnormal Types of Speech in Nootka.' Canada Department of Mines, Geological Survey, Memoir 62 (Anthropological Series No. 5). Ottawa: Government Printing BureaUniversity Sapir, Edward. 1921. Language: An introduction to the study of speech. New York: Harcourt Brace. Sapir, Edward. 1924. 'The Rival Whalers, a Nitinat Story. (Nootka text with translation and grammatical analysis).' International Journal of American Linguistics, 3:76-102. Sapir, Edward & Morris Swadesh. 1939. Nootka Texts, Tales and Ethnological Narratives, with Grammatical Notes and Lexical Materials. Philadelphia & Baltimore, Md.: Linguistic Society of America. Sapir, Edward & Morris Swadesh. 1955. Native Accounts of Nootka Ethnography. Bloomington, Ind.: Indiana University Research Center in Anthropology, Folklore and Linguistics. Scalise, S. 1988. 'Inflection and Derivation.' Linguistics 26: 561-581. Selkirk, E. O. 1984. 'On the Major Class Features and Syllable Theory.' M. Aronoff & R.T. Oerhle (eds), Language Sound Structure: Studies in phonology presented to Morris Halle by his teacher and students. Cambridge, Mass.: MIT Press. 107-136. Shaw, Patricia. 1989. 'The Complex Status of Complex Segments in Dakota.' Donna Gerdts & Karin Michelson (eds.), Theoretical Perspectives on Native American Languages. New York: SUNY Press. 3-37. Spencer, Andrew. 1991. Morphological Theory. Oxford: Basil Blackwell, 1998. Spencer, Andrew & Arnold M. Zwicky, (eds). 1998. The Handbook of Morphology. Oxford: Basil Blackwell, 1998. Sproat, Malcom Gilbert. 1868. Scenes and Studies of Savage Zife, Smith, Elder, and Co: London. [Reprinted 1987]

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

290 Bibliography

291

Steele, S., A. Akmajian, R. Demers, E. Jelinek, C. Kitagawa, R. Oehrle, & T. Wasow, (eds). 1981. An Encyclopedia of AUX: A study in cross-linguistic equivalence. Cambridge, Mass.: MIT Press. Stonham, John. 1990a. 'Variable-length Vowels in Nootka: The long and the short of it.' Annual Meeting of the Canadian Linguistic Association, May 20, 1990. Stonham, John. 1990b. Current Issues in Morphological Theory. Stanford University doctoral dissertation, 266 pp. Stonham, John. 1991. 'Hypocoristic Formation in Nootka.' J. Myers & P.E. Perez, (eds), University of Arizona Coyote Papers: Arizona Phonology Conference, Volume 3, Tucson: University of Arizona, pp. 119-134. Stonham, John. 1994a. 'All Moras are not Created Equal.' Cahiers Linguistiques d'Ottawa, vol. 21: 1-25. Stonham, John. 1994b. Combinatorial Morphology. Amsterdam: John Benjamins, 206 pp. Stonham, John. 1998. 'Numerals and Incorporation in Nootka.' D. Bates (ed.), Papers for the 33rd International Conference on Salish and Neighbouring Languages, Seattle: University of Washington, 384-394. Stonham, John. 1999a. 'Noun Collocation in Nootka.' M.Q. Hinkson & M. Ignace (eds.), Papers for the 34th International Conference on Salish and Neighbouring Languages. Kamloops, B.C.: Simon Fraser University, pp. 231250. Stonham, John. 1999b. Tsishaath Nootka Phonetics & Phonology. Munich: LINCOM Europa; LINCOM Studies in Native American Linguistics 32. xii, 155 pp. Stonham, John. 2003. 'Southern Wakashan Double Reduplication.' J. Brown, M. Kalmar & T. Peterson, (eds), Papers for the 38th International Conference on Salish and Neighbouring Languages. Vancouver: University of British Columbia, pp. 237-251. Stonham, John. 2004 (in press). 'Level Ordering in Nootka.' K. Hanson & S. Inkelas (eds), The Nature of the Word: Essays in honor of Paul Kiparsky. Cambridge MA: MIT Press. Stonham, John. 2004 (to appear). Nuuchahnulth Dictionary. Edwin Mellen Press. Stonham, John & Yiu Sze Man. 2000a. 'Woman-buy vs. Two-have: Two types of incorporation in Nootka.' Annual Meeting of the Linguistic Society of America, Jan. 8, 2000. Stonham, John & Yiu Sze Man. 2002a. 'The Word in Nootka.' Annual Meeting of theCanadian Linguistic Association, May 24, 2002. Swadesh, Morris. 1931. The Nootka Aspect System. M.A. thesis, University of Chicago. Swadesh, Morris. 1933. The Internal Economy of the Nootka Word. Yale University Ph.D. dissertation.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Bibliography

Bibliography

Swadesh, Morris. [1937]. 'Nootka Phonology and Morphology.' Unpublished manuscript no. 30(W2a.l0) in the Boas Collection of the American Philosophical Society. Swadesh, M.H. 1939. 'Nootka Internal Syntax.' International Journal of American Linguistics 9.77-102. Thomas, Alex & Eugene Y. Arima. 1970. ta:ta:qsapa: A practical orthography for Nootka. Ottawa: National Museum of Man, Publications in Ethnology 1. Turner, Nancy J. & B.S. Efrat. 1982. Ethnobotany of the Hesquiat Indians of Vancouver Island. Victoria: British Columbia Provincial Museum. Urbanczyk, Suzanne. 1995. 'Double Reduplications in Parallel.' University of Massachusetts Occasional Papers 18, 499-531. Whistler, Kenneth W. 1985. 'Focus, Perspective, and Inverse Person Marking in Nootkan.' J. Nichols & A. C. Woodbury (eds), Grammar Inside and Outside the Clause', Cambridge: Cambridge University Press, pp. 227-265. Wilson, S.R. 1985. Stress Rules in Wakashan. M.A. thesis, University of California at Berkeley. Wiltschko, Martina. 2001. 'Tense on D and (the Lack of) Nominative Case.' M.J. Kim & Uri Strauss, eds., Proceedings of the Northeast Linguistics Society 31, Georgetown University. Wojdak, Rachel. 2000. 'Nuuchahnulth Modification: Syntactic Evidence Against Category Neutrality,' International Conference on Salishan and Neighboring Languages 35, pp. 269-281. (University of British Columbia Working Papers in Linguistics, vol. 3). Wojdak, Rachel. 2001. 'An Argument for Category Neutrality?' K. Megerdoomian & L. A. Bar-el (eds), WCCFL 20 Proceedings, pp. 621-634. Somerville, MA: Cascadilla Press. Yiu Sze Man & John Stonham. 2000. 'Good-stocked with Mussels: Incorporation on the edge.' Annual Meeting of the Linguistic Society of America, 2000. Yiu Sze Man. 2001. 'Incorporation in Nuuchahnulth' University of Manchester MA thesis. Yiu Sze Man, Winnie & John Stonham. 2002. 'Classifiers in Nuuchahnulth.' 37th International Conference on Salishan and Neighboring Languages. Yiu Sze Man, Winnie & John Stonham. 2004. 'Symmetry in Nuuchahnulth Syntax: The structure of the IP and DP.' Proceedings of the Canadian Linguistic Association, pp. 308-319. Zee, Draga. 1988. Sonority Constraints on Prosodic Structure. Ph.D. dissertation, Stanford University. Zoll, Cheryl. 1996. Parsing Below the Segment in a Constraint Based Framework. Ph.D. dissertation. University of California, Berkeley. Zwicky, A.M. & G.K. Pullum. 1983. 'Cliticization vs. Inflection: English n't.' Language, vol. 59, no. 3.

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

292

ablaut, 206-207 abnormal speech, 3, 7, 100-101, 178, 208, 277 absolutive mood, 68 action-at-a-distance, 2, 274 actor, 79 actualis, 101 adjective, 57,59-60,62,212,223,275 adjunction, 269, 271 adverb, 116, 136 suffixes, 116, 136 affix, 9, 35, 49, 67, 88, 99, 103, 183, 185, 196, 198,247 derivational, 32, 50, 155, 243 affix-triggered reduplication, 2, 7 agent, 78-79 agentive, 115 Ahousaht, 4-5, 8, 10, 14, 17, 66, 7576, 122, 170-171, 274, 276, 281, 283 alienable possession, 94 allomorph, 9-10, 39, 82, 89, 91, 95, 104, 158-159 applicative, 260, 266 apposition, 239, 243 appositive, 239-240, 242 argument phrase, 58, 62, 247 argument structure, 93, 242, 245, 260, 268, 273 aspect, 2, 9, 44-46, 55-57, 64, 89, 91, 100, 106, 108, 119, 129-131, 146147, 166, 184, 260, 264, 268, 273, 283 imperfective, 72 inceptive, 106-109 iterative, 130-131 markers, 44, 89, 106, 146 repetitive, 56, 126, 131, 137

aspectual reduplication, 132, 137, 162 assimilation, 26-27, 84, 281 augmentative, 82, 100-101 Baker, M., 215, 219, 223, 234 base, 19-20, 29, 32, 35, 43, 46-50, 52, 104, 115, 120-121, 131-132, 148149, 154, 172, 180, 187, 193, 195, 204,213-214,250,266,268 expletive, 48-49, 214, 223 Base-Reduplicant Faithfulness, 172 Bauer, L., 225 Bermudez-Otero, R., 3, 144, 171-172 Boas, F., 3, 5, 7, 17, 34, 54-55, 279, 281 borrowed words, 38, 158 bound root, 38, 41, 43-44, 159-160 bound verb, 2, 46-48, 50-52, 114-115, 132, 213-215, 217-219, 222-223, 271,280,284 bracket erasure, 164-165, 167, 170, 180,274,282-283 Broselow, E., 138, 162, 192 Buckley, E., 138, 162 Cantonese, 60, 249, 281, 283 Carlson, B.F., 7, 89, 281 case-marking, 236 category distinctions, 57, 270 category neutrality, 2, 61-63, 86, 273 causative, 91, 108-109, 151, 239, 260, 264, 266 inceptive, 109 Central Nuuchahnulth, 4 Chiclesit, 4 Chinese, 11,60 Chinook Jargon, 169 Chomsky, N , 216, 225 classifier, 116, 210, 226, 246-258, 277,281,284

293

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Index

classifier, intrinsic, 250-252, 256-257 phrase (CIP), 247 clause, relative, 75-76, 212, 220-222, 234-235, 238, 240-242 clause, subordinate, 74-77 Clayoquot, 4-5 clisis/clitics, 66, 81, 85, 97, 102, 210213, 221, 230, 239-241, 258, 260266,268,270,276,281 clitichood, 264 cliticisation, 3, 259-261, 266, 268, 270-271,276 coalescence, 14, 20-24, 28-29, 94, 129, 149, 203 collocation, 48, 235-236, 240, 275 complement, 60, 229, 237, 242, 256257 complementary distribution, 75, 83, 93, 140, 191 complex numeral, 226, 228, 233, 252 compounding, 3, 8, 32, 34, 210, 213, 215-216,234,243,246,275 pseudo-, 242-243, 246, 275 quasi-, 243 Comrie, B., 82, 106,234 conditional mood, 74 conjunction, 53, 112, 215, 230, 237, 242, 253, 270 constituent, 59, 88, 214, 236, 240 constraints, 23, 32, 35-36, 111, 122, 138,150,172-173,204 MaxBR, 172-174 ranking, 173 semantic, 47 contemporaneity, 99 continuant, 148, 206 coordination, 50, 110, 228 copying, 36, 122, 124, 126, 128, 173, 178-180, 186, 189 coreference, 215-216, 233 coronals, 19, 149,274 counterfactual, 98 Cree, 169 Dakota, 154

Davidson, M., 5, 8, 26, 85-86, 169, 184,258,260-264,281 definite article, 221, 230, 236, 239240 definiteness, 102, 244-246, 255 delabialisation, 29-31,153-154,261 derivation, 2, 31, 64, 66, 113, 129, 145-147,152,203-204,225 affix, 32, 50, 157,243 morphology, 20, 27, 29, 132, 203 suffix, 44-45, 54, 110, 112, 114115, 119-120, 132, 135-136, 139, 141, 145-146, 150-151, 154, 157, 160, 164, 183, 227, 244, 264 diminutive, 100-101,265 discontinuous morpheme, 134, 274 distributional properties, 57 distributive, 130, 163-164, 168 Ditidaht, 3, 5, 7, 11, 122, 138, 162, 168-169, 245, 258-259, 261, 281 ditransitive verb, 56 domain, 15-16, 93, 100, 141, 150-152, 154-156,230 double reduplication, 3, 8, 25, 121122, 138-139, 161-163, 165-171, 173-174,268,274,276 Drucker, P., 4, 10 dubitative mood, 72, 77 dyadic verb, 242 Efrat, B., 5 Emanatian, M., 7, 89 English, 11, 33, 38, 45-46, 53-55, 58, 61,86,88,91, 144, 154, 158,238, 250, 256, 266, 275 expletive base, 48-49, 214, 223 faithfulness Base-Reduplicant, 172 fixed segmentism, 126, 178, 182, 185186, 188 foot, 123, 127-128, 189-194,209 first, 16,22-24,28 full, 190 foot-based templates, 189, 193195,208,273,276 foreign borrowing, 13-14

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

294 Index

formative suffix, 20, 150 Fortescue, M., 49, 243, 280 free root, 41, 157, 159, 169, 214, 263 fricative, 29, 149, 154-155 fronting, 220, 259-260, 269-270 full syllable, 122, 126, 131, 180, 182183 function word, 220 functional head, 257, 268, 270, 276 future, hypothetical, 85 future imperative mood, 79, 105 glide, 8, 19,26, 148,207 glottal stop, 20, 24, 26-29, 94, 128, 149, 154-155, 198, 203, 262-263, 265 glottalic action, 14 glottalisation, 7, 14, 16-18, 20-21, 145, 147-152, 162, 262, 276, 280, 282 glottalising suffix, 9, 19, 148, 150151,262 governing suffix, 110, 247 government, proper, 223 grammaticality, 59, 240, 267 Greenlandic, 243 West, 49 Haas, M.R., 7, 259 habituality, 74 Haisla, 3, 50 Hale, K., 225 harmony, 100 head, 3, 47-48, 86, 101, 103, 114, 213, 218-223, 226-230, 233-234, 236240, 242-243, 245-246, 256-257, 269,271,275 functional, 257, 268, 270, 276 noun, 3, 101, 219-221, 226-227, 229, 234, 236-240, 242, 275 head-marking, 93, 236 headless relative, 240-242 Heiltsuk, 3 Hesquiaht, 4-5 Hess, T.M., 5 Hogg, R., 171 hortative, 72, 78 Huddleston, R., 240 Huiath, 2, 5

295

Hupachasath, 2, 5 hypocoristics, 3, 7, 54, 178, 195-199, 203-204, 208, 275 formation, 3, 71, 96, 178, 196, 199,276 hypothetical future, 85 imperative mood, 79, 80 imperfective aspect, 72, 106 inalienable possession, 93 inceptive aspect, 106, 108-109 inceptive causative, 109 incorporation, 3, 7, 34, 47, 50, 52, 114, 210, 213-216, 218-219, 222223, 226, 228-234, 258, 272-275 lexical, 215-216, 218-219, 222, 227, 272, 275 numeral, 8, 228, 233 syntactic, 215-216, 219, 222-228, 272,275 incorporee, 50, 215, 219, 225, 233 incremental suffix, 20,145,150,259 indefinite marker, 103 indefinite relative mood, 76, 105 indicative mood, 66, 68-69, 75, 83 inferential mood, 73-74 infix, 139-141, 192-193, 195,208 infixation, 2-3, 97, 121, 139, 141-143, 178, 183, 188, 192, 195, 204, 272, 274, 282 inflection, 2-3, 7, 31, 39, 46, 64, 66, 119, 129, 145-146, 152, 155, 157158, 259-260, 262, 264-272, 276 non-paradigmatic, 66, 82, 103, 105 paradigmatic, 66, 97, 105, 213, 268 insertion, 100, 112-113, 124-125, 149, 202, 206, 274 instability, 280 instrumental, 49, 115 intransitive verb, 56, 96, 114, 218, 230-231,239,265,275 intrinsic classifier, 250-252, 256-257 IPA, 8 irrealis, 84, 98 iterative aspect, 130-131

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Index

Jacobsen, W.H., Jr., 4, 6-7, 56-57, 60, 261,281 Kaisse, E., 171 Kaska, 171 Katamba, F., 216 Keenan, E., 234 Keyser, S.J., 225 Kim, E.-S., 5, 8, 17, 24, 26, 28, 31, 89, 113, 122, 141, 145, 150, 154, 170, 185,279-280,282-283 Kiparsky, P., 3, 144, 172, 176 Klokeid, T., 4, 14, 258-260, 279 Knipe, C , 6-7, 252, 279 Kyuquot, 4, 7, 10, 17, 36, 75-76, 83, 85, 88, 96, 121, 167-170, 247, 276, 281 labialisation, 14, 27-28, 31, 153, 279280 laryngealisation, 8, 17, 19-20, 147151,202 Lemmens, N , 5 lengthening, 2, 34, 110, 112-113, 120, 122, 124-125, 128, 131, 178, 183184, 188, 205, 272, 274, 280, 283 lenition, 16-17, 20-21, 145, 151-152, 162,280,283 level, see stratum level, stem, 147, 150, 152, 155-156, 160-163, 165-171, 173, 263-264, 276, 282 level, word, 144, 150, 155, 160, 162163, 165-171, 173, 261, 263-264, 268, 276, 282 level ordering, 20-21, 28-29, 153 Lexical Phonology and Morphology (LPM), 144 lexical incorporation, 215-216, 218219,222,227,275 lexicon, 46-47, 63, 119, 144-145, 162, 171-172, 215-216, 218-219, 248, 266, 268, 276 locative, 33, 49, 103, 115, 118, 136, 217,247 suffix, 49, 115, 118, 136,217

long vowel, 15-16, 23, 111, 124-125, 129, 156-157, 180, 186, 192, 205, 262 Makah, 3, 56, 60, 169-170, 258, 260261,281 Manam, 138, 162 Manhousaht, 4-5 matrix verb, 78, 227 Matthews, S., 249, 280 MaxBR, 171-174 McCarthy, J., 122, 128, 172-173, 178, 183, 185, 192,276 merger, 21-22, 36, 115 metrical structure, 16 minimal triplets, 25 modification, 62, 113, 116, 131, 192193,238 modifier, 56, 59, 101, 103, 239, 243246 non-restrictive, 238-240, 246 Mohanan, K.P., 144 Mohawk, 215 momentaneous causative, 92 monostratal OT, 172 monosyllabic root, 127, 184, 189,262 mood, 66, 68-72, 75-79, 83, 108, 119, 260,268 absolutive, 68 conditional, 74 dubitative, 72, 77 future imperative, 79, 105 imperative, 79-80 indefinite relative, 76, 105 indicative, 66, 68-69, 75, 83 inferential, 73-74 irrealis future, 84 irrealis, 84, 98 present imperative, 78-79, 105 purposive, 71-72, 105 quotative, 69-70, 98 moras, 16, 122 morphophonology, 2, 21 morphosyntax, 66, 272 morphotactics, 267 movement, 48, 115, 118, 125, 210, 212,215,225,231-233,269

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

296 Index

Mowachath, 4, 11 mythical characters, 206 Nakayama, T., 5, 8, 10, 68, 89, 183, 281 nameability, 225 nominal, 37, 46, 56-57, 61, 87-88, 103, 116, 190, 217, 241, 247-248, 250-251, 254, 256-257, 260, 265, 269 nominalisation, 60 non-paradigmatic inflection, 66, 82, 103, 105 non-restrictive modifier, 238-240 Northern Nuuchahnulth, 4, 66 Northern Wakashan, 16, 50, 54, 185, 283 noun phrase, 56, 82, 213, 215, 220221, 225-226, 230, 236-237, 239240, 242-243 noun, head, 3, 101, 219-221, 226-227, 229, 234, 237-240, 242, 275 noun-noun collocation, 236-238, 245 number marking, 221 number system, 52-53, 228 number, complex, 226, 228, 233, 252 numeral incorporation, 8, 228, 233 Nuuchahnulth Ahousaht, 4-5, 8, 10, 14, 17, 66, 75-76, 122, 145, 170-171, 276, 281,283 Central, 4 Chiclesit, 4 Clayoquot, 4-5 Hesquiaht, 4-5 Huiath, 2, 5 Hupachasath, 2, 5 Kyuquot, 4, 7, 10, 17, 36, 75-76, 83, 85, 88, 96, 121, 167, 168170,247,276,281 Manhousaht, 4-5 Mowachath, 4, 11 Northern, 4, 66 Uchucklesit, 5,214,227 Ucluelet, 2, 4-5

297

Tsishaath, 2, 5, 10-11, 14, 17, 25, 75-76, 85, 162-163, 167, 169, 206,247,281 Old English, 171 onset, 14, 36, 122, 124-126, 139, 141, 178-180, 186, 189, 193, 199-202, 207, 209 Oowekeno, 3, 279 opacity, 56, 248, 282 semantic, 244, 248 open syllable, 131, 137, 182-184, 189, 204 Optimality Theory (OT), 3, 8, 144, 162, 172-173,273,276 monostratal, 172 Paik, T.-S., 5 paradigmatic inflection, 66, 97, 105, 213,268 particularising marker, 102 passive, 83, 89-90, 96, 259-260, 265266 past tense, 9, 65, 82-83, 85-87 patient, 79 phonological rule, 16, 144, 150, 156, 206, 270 phonological-conditioning, 141 phrase structure, 219, 231, 238, 245, 258, 273 phrase, quantifier (QP), 228-229, 256 Picanco, G., 185 Pitch Woman, 206, 208, 277 pluralisation, 95, 129, 130, 142 plurality, 96-98, 130 Port Alberni, 2, 5, 70 portmanteau, 67, 80-81, 83, 89, 108109,115,119,264,268 possession, 82, 93-94, 236, 245 alienable, 94 inalienable, 93 possessive, 83, 95, 98, 236-237, 240, 244, 266 marker, 95, 244 possessor, 93, 98, 236, 244-245, 265 postlexical process, 262-263, 266

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Index

predicate, 47, 59-60, 62, 70, 82, 89, 93, 99, 103-104, 114, 116, 215, 222, 260, 267-268, 273 present imperative mood, 78-79, 105 prespecification, 275-276 primary stress, 15, 156 Prince, A., 122, 128, 172-173, 178, 183,185,276 pronominal, 7, 56, 70, 236, 240, 257 proper government, 223 Prosodic Morphology, 185 prosodic circumscription, 275 prosodic structure, 195, 199, 282 pseudo-compounding, 242-243, 246, 275 Pullum,G., 211,258, 264 purposive mood, 71-72, 105 qualitative ablaut, 206 quantifier, 37, 59, 116, 223, 226, 228230, 233, 247, 250-251, 254-258, 270, 275 phrase (QP), 228-229, 256 quantity base, 247 quantity coalescence, 22 quasi-compounding, 243 quotative mood, 69-70, 98 ranking, constraint, 173 reduplication, 2-3, 7, 14, 34, 42, 97, 111-113, 120-139, 141-143, 162171, 178, 180-185, 187-190, 192193, 196, 198, 203-205, 207, 272, 274, 276, 280, 282-283 affix-triggered, 2, 7, 164, 177 aspectual, 132, 137, 162 double, 3, 8, 25, 121-122, 138-139, 161-163, 165-171, 173-174, 268, 274, 276, 280, 283 multiple, 167 template, 185-187, 192, 204, 207, 275 relative clause, 75-76, 212, 220-222, 234-235, 238, 240-242 headless, 240-242 Renker, A., 7, 260-261 repetitive aspect, 56, 126, 137 restrictive suffix, 109-110

rhetorical modifications, 13 rhyme, 126, 180 root, 2, 7, 9, 14, 24-25, 32-35, 37-38, 44-45, 53-54, 95, 101, 110-112, 122-128, 130-134, 136-137, 139141, 146, 157-158, 164-167, 179180, 182-184, 189, 191-193, 196, 204-205, 213, 217-219, 227, 234, 264, 274, 280, 282-283 bound, 38,41,43-44, 158-160 free, 41, 157, 159, 169,214,263 Rose, S., 4, 7, 10, 17, 35-37, 68-70, 72, 79, 83, 85, 88-89, 96, 99, 101103, 121, 130-132, 137, 167, 183, 185, 247, 250, 262, 265-266, 281, 284 rule, phonological, 16, 144, 150, 156, 206, 270 Salish, 118, 138,162,281 Sapir, E., 2, 5-7, 10, 12, 18, 20, 24-25, 33, 43, 55, 58, 64, 66, 82, 86, 8990, 92, 96-97, 100-102, 106, 108109, 113, 121, 129, 145, 147-148, 157, 161, 167, 178, 197,206,208, 210, 248, 258-259, 273, 277, 281 Scalise, S., 64 scope, 45, 100, 146, 244, 249, 255 semantic constraints, 47 semantic transparency, 222 semantics, 34, 48, 53, 119, 145, 212, 215,223,235,238,242 Shaw, P., 154 shortening, 2, 34, 38, 101, 113, 120, 122,129, 178,205,272,274,283 Smolensky, P., 172 song texts, 13 sonorant, 16, 124, 140,202 Sonority Hierarchy, 202, 283 sonority violation, 200 Southern Wakashan, 3, 6-11, 56-57, 121, 162, 167, 170-171,258-260 specificity, 102 Spencer, A., 210, 231 spreading, 100, 190 Steele, S., 260

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

298 Index

stem, 3, 32-33, 39, 41-42, 44-46, 55, 88, 103, 109-110, 113-114, 119, 123-125, 127, 129-130, 132-133, 136-138, 144, 150, 152-153, 155158, 160-171, 178, 227, 243, 247, 265, 274, 276, 280, 282 stem level, 144, 147, 150-153, 155158, 160-171, 173, 175-176, 263264, 268, 276, 282 Stonham, J., 5-8, 13, 15, 17, 26, 28, 122, 138, 153, 162, 168, 171, 208, 270,276,279,281-283 stratal segregation, 20, 28, 66, 122, 144, 147, 149, 154, 171,175,276 stratum, 144-145, 150, 155-158, 160161, 170-171, 173-174, 219, 262, 268 stress, 15-16, 26, 124, 144, 155-156, 162, 174,279 assignment, 16, 124, 145, 155-156, 162 placement, 26, 156,279 primary, 15, 156 system, 26 subcategorisation, 215, 218, 223, 233, 248, 256, 273 restrictions, 215, 218, 223, 233 subject, 55, 64, 69, 90-91, 97, 202, 220,231-232,235-236 subordinate clause, 74-77 suffix, adjectival, 135 adverbial, 136 aspectual, 139 causative, 239 combining form, 161 derivational, 44-45, 54, 110, 112, 114-115, 119-120, 132, 135136, 139, 141, 145-146, 150151, 154, 157, 160, 164, 183, 227, 244, 264 formative, 20, 150 glottalising, 9, 19, 148, 150-151, 262 governing, 110, 247 incremental, 20, 145, 150, 259

299

locative, 49, 115, 118, 136, 217 restrictive, 109-110 verb, 114, 127 suffixation, 32, 64, 106, 122, 241, 272, 280, 282 Swadesh, M., 5-7, 18, 20, 24-25, 43, 47, 55, 58, 60, 66, 89-90, 92, 97, 106, 108-109, 113, 121, 129, 145, 147-148, 197, 210-211, 216, 247249, 258-259, 273, 284 switch reference, 89 syllable, 14-16, 23-25, 28, 122, 124, 126-127, 131, 139-141, 143, 156, 180-185, 189-194, 198-203, 206207, 209, 262, 273, 282-283 full, 122, 126, 131, 180, 182-183 open, 131,137,182-184,189,204 structure, 15,24, 143, 199 template, 181 syntactic category, 61, 245 syntactic incorporation, 215-216, 219, 222-228, 272, 275 syntax, 3, 56, 210, 212-213, 216, 219222, 235, 237, 246, 249, 258, 268, 273,276 template, 3, 178, 180, 183-197, 199200, 202-205, 207-208, 276, 283 foot-based, 189, 193-195, 208, 276 reduplicative, 185-186, 192 syllable, 181 Templatic Morphology, 3, 121, 178, 192,275 temporal, 37, 99, 101, 103-104, 106, 115-116,247 predicate, 101 tense, 56-57, 65, 82-83, 85-89, 260, 281 past, 9, 65, 82-83, 85-87, 89 theme, 47, 78, 109 tier conflation, 180 transitive, 47, 51, 56, 58, 218, 235, 239,241 Tsishaath, 2, 5, 10-11, 14, 17, 25, 7576, 85, 162-163, 167, 169, 206, 247,281

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

Index

Turner, N., 5 Uchucklesit, 5, 214, 227 Ucluelet, 2, 4-5 Urbanczyk, S., 138, 162 valency, 47, 89, 91, 215, 218-219, 223,233, 266 variable length vowel, 22-26 verb, 7, 46, 50, 52, 55-57, 61, 65, 8182, 84, 86,91, 101, 106, 114,210, 212-216, 218-228, 230-233, 235, 239, 241-242, 245-246, 259, 266, 270 bound, 2, 46-48, 50-52, 114-115, 132, 213-215, 217-219, 222223,271,280,284 ditransitive, 56 dyadic, 242 intransitive, 56, 96, 114, 218, 230231,239,265,275 matrix, 78, 227 suffix, 114, 127 vocative, 7 VOS word order, 220-221 vowel, variable-length, 22-26 vowel coalescence, 263

vowel merger, 22 vowel quality, 21, 190 vowel quantity, 21, 39, 156, 159 VSO word order, 220-221 Wakashan, 2-3, 9, 16, 55, 169, 272, 276-277, 282 Northern, 16, 50, 54, 171, 185, 283 Southern, 3, 6-11, 56-57, 121, 162, 167, 170-171,258-260 weight, 111, 123, 141, 157, 179, 181, 185, 190,249,274 West Greenlandic, 49 Whistler, K., 7, 89 Wojdak, R., 5, 8, 61-62 word classes, 55-58, 60-61 word level, 144, 150, 155, 160, 162163, 165-171, 173, 261, 263-264, 268, 276 word order, VOS, 220-221 VSO, 220-221 Yip, V., 249 Yiu, S.M., 8, 270, 276,281 Zoll,C, 154 Zwicky, A.M., 211,258, 264

10.1057/9780230505551 - Linguistic Theory and Complex Words, John Stonham

Copyright material from www.palgraveconnect.com - licensed to Universitetsbiblioteket i Tromsoe - PalgraveConnect - 2011-03-11

300 Index