The Squamish language: Grammar, texts, dictionary [Reprint 2015 ed.] 9783111358024, 9783111001074

Citation preview

THE SQUAMISH LANGUAGE

JANUA LINGUARUM STUDIA MEMORIAE NICOLAI VAN WIJK DEDICATA edenda curai

C. H. VAN SCHOONEVELD I N D I A N A UNIVERSITY

S E R I E S PRACTICA LXXIII

1967

MOUTON & CO. THE HAGUE • PARIS

THE SQUAMISH LANGUAGE GRAMMAR, TEXTS, D I C T I O N A R Y

by

AERT H. KUIPERS U N I V E R S I T Y OF LEIDEN

1967

MOUTON & CO. T H E H A G U E • PARIS

© Copyright 1967 by Mouton & Co., Publishers, The Hague, The Netherlands. No part of this book may be translated or reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publishers.

Printed in The Netherlands by Mouton & Co., Printers, The Hague.

To

LOUIS MIRANDA and to the memory of

ISAAC JACOB and

ALEC PETER

M A P OF THE SALISH LANGUAGE AREA

PREFACE

The Squamish tribe, a subdivision of the Coast Salish stock of North America, at one time inhabited the area around Howe Sound and along the banks of the lower Squamish River and Cheakamus Creek in southern British Columbia. More recently they spread to the shores of English Bay, False Creek and Burrard Inlet as far as Seymour Creek on the northern and Coal Harbour on the southern bank. 1 Their neighbors were the Sechelt in the west, the Lillooet in the north and northeast, and Halcomelem-speaking tribes in the south and southeast. Their first contact with white men occurred in the summer of 1792 when Captain Vancouver sailed into English Bay and Howe Sound, an event still retained by popular memory (see text IV). Early white settlers estimated the number of the Squamish at several thousands.2 The tribe was decimated by a succession of smallpox epidemics, and was reduced to a few hundred at the end of the 19th century. Though since then the population has increased,3 the language is now close to extinction: it is spoken only by older people (number unknown) most of whom are living in the Vancouver area. A classification of the Salish languages was made by Swadesh on the basis of a lexicostatistical analysis of comparative vocabularies in the Boas Collection of the American Philosophical Society.4 This classification, which progresses from Divisions to Branches, Groups and Languages, is quoted below with some changes and additions. A few alternative designations are given in parentheses. Where Swadesh's groups encompass more than one language, his group-names may cause confusion, being identical with group-member names. In these cases they are replaced by other names where such have been proposed; otherwise they are left out and the group is identified by an Arabic numeral only. In some groups a number of languagenames have been added for practical reasons, and the languages within each group are listed without further bracketing. In the case of the Halcomelem (II B2) as well 1

Hill-Tout 1900:473-475; Barnett 1938:140. » HiU-Tout 1900:473. ' Verma 1954:11. 4 Swadesh 1950.

8

PREFACE

as of the Lkungen group (II B3) the members are mutually intelligible dialects of languages referred to with one and the same name by all speakers.8 This may be the case in one or more of the other groups too. For the geographical distribution of the languages see the map on p. 6. I. Bella Coola II. Coast Division A. North Georgia Branch 1) Comox (Sliammon) 2) Sechelt (Sisiatl) 3) Pentlatch B. South Georgia Branch 1) Squamish 2) Halcomelem: Nanaimo, Cowichan, Chemainus, Musqueam, Kwantlen, Katzie, Sumas, Chehalis, Chilliwack, Tait (the latter seven referred to as "Lower Fraser") 3) Lkungen (Straits): Clallam, Songish, Saanich, Semiahmoo, Lummi, Samish 4) Nootsack C. Puget Sound Branch: Skagit, Duwamish, Snoqualmie, Snohomish, Nisqualli D. Hood Canal Branch: Twana E. Olympic Branch 1) Cowlitz, Upper Chehalis 2) Satsop 3) Lower Chehalis, Quinault III. Oregon Division: Tillamook, Siletz IV. Interior Division 1) Lillooet 2) Thompson, Shuswap 3) Okanagon, Colville, Sanpoil, Lake, Spokane, Kalispel, Pend d'Oreille, Flathead, Sematuse, Plains Salish 4) Wenatchi, Columbia, Peskwaus 5) Coeur d'Alene Sapir hypothetically combined Salish with Chemakuan (Chemakum-Quileute) and Wakashan (Kwakiutl-Nootka) into a "Mosan" super-stock9 and suggested a rela6

Elmendorf-Suttles 1960:3. • The name is based on the numeral 'four', in many Salish languages mus, Chemakuan maPyas, Wakashan (Kwakiutl) mu.

PREFACE

9

tionship of this stock to Kutenai and Algonquian. 7 No thorough testing of these hypotheses is possible because of insufficient data. 8 Of the Salish languages, only a few have received more or less extensive treatments (overall or partial) in recent times: Halcomelem (Elmendorf-Suttles 1960: Vocabulary), Snoqualmie-Duwamish (Tweddell 1950: Phonology, Morphology), Upper Chehalis (Kinkade 1963-64: Phonology, Morphology), Kalispel (Vogt 1940: Grammar, Texts, Dictionary), Coeur d'Alene (Reichard 1938: Grammar; 1939: Stem-List). For others only sketchy and/or older data are available (see Bibliography). A sketch of Squamish was published by Hill-Tout as part of an interesting ethnographical description.9 This material has to be used with caution because of an inadequate and inconsistent notation and various other defects. Nevertheless, it has not lost its value: Hill-Tout's word-list proved useful as a starting-point in obtaining lexical items, his phrases contain valuable syntactic material, and in individual cases his data make it possible to establish an older, fuller word-form than was recorded half a century later. The material on which the present description is based was collected in the years 1951-54 and in the fall of 1956. — During the first of these periods I was on the staff of the Department of Slavonic Studies of the University of British Columbia in Vancouver. My informants were first Mr. Isaac Jacob (1883-1956): in the text referred to as "IJ") of Capilano Reserve, and later Mr. Alex Peters, known as "Alec Peter" (1881-1964; referred to as "AP") of Musqueam Reserve. Mr. Peters knew Musqueam besides Squamish. The working-sessions were held at the homes of the informants, and occasionally Mrs. Jacob and Mrs. Peters would make comments or volunteer information. During this period a fairly large number of words, short phrases and sentences were collected, and outlines of phonology and morphology were worked out. Much remained problematic; also, I was unable to obtain longer connected texts. — In the fall of 1956 I returned to Vancouver for a short period, in which I had the good fortune of being able to engage the services of Mr. Louis Miranda (born 1892; referred to as "LM") of Mission Indian Reserve. Mr. Miranda, an exceptionally alert informant, was of great assistance in checking and clearing up a number of difficult points. He was furthermore able to dictate texts, which in turn provided additional lexical and grammatical material. Translations and a rapid analysis of these texts were made while I could still dispose of Mr. Miranda's invaluable cooperation. The material thus obtained was further analysed and classified after my return to Holland. To many questions which arose during the work answers could be found in the available corpus. In a few cases alternative solutions had to be given (the choice depending on Squamish forms not represented in the material). Some questions had to remain open. In such cases I have preferred stating the problem '

See Swadesh 1953:28. See Appendix, p. 401. » HillsTout 1900:495-518 section "Linguistics".

8

10

PREFACE

in detail to "streamlining" the description. The present book contains all the material of which I dispose; on matters that are not treated (e.g. sentence-intonation, the expression of ordinal numbers) data are lacking. The categorization of the material is based primarily on formal criteria. Except in trivial cases, all available examples of a given category are quoted (e.g. of a particular type of transitive verb or subordinate clause). In the Dictionary internal and external etymologies are given, the latter take into account only those languages for which more or less extensive word- or stem-lists are available (for details see p. 245f.). A number of uncertain comparisons are included (marked "Cf.?"); here additional data may show whether the question-mark or the comparison itself has to be dropped. In the course of my work on Squamish I had the pleasure of incurring debts of gratitude to several persons and institutions. First of all I want to thank Wayne Suttles, who introduced me both to Salish linguistics and to all three of my informants. In the difficult initial stages of the work it was a great advantage to be able to discuss problems with a fellow-linguist interested in the same field. I have also greatly profited from Wayne Suttles' expert ethnographical knowledge of the Coast Salish. I gratefully acknowledge the help and encouragement given to me by Harry B. Hawthorn. Thanks are also due to the Canadian Social Science Research Council, which, through the good offices of Harry Hawthorn, defrayed my expenses for travel and informant-fees. I thank Mrs. Ingeborg Houwen-van Driel, Miss Vera Litwinzeff and Miss Sietske Visser for their conscientious assistance in the preparation of the various drafts of the manuscript. I am indebted to the Board of Curators of the University of Leiden for providing secretarial help. The greatest debt of all I owe to my informants. I dedicate this book to the memory of Isaac Jacob and Alec Peter, who patiently gave their best efforts in what to them was an unfamiliar and difficult job, and to Louis Miranda, whose alert and tireless assistance yielded so much both in material and in insight.

TABLE OF CONTENTS

Preface

7

Symbols, Abbreviations, References

18

GRAMMAR P A R T i : PHONOLOGY

I. Phonemic Data 1-3 1. Phonemes. 2. Stress. 5. Juncture.

21

II. Phonetic Data 4-38 24 A. Consonants 4-7 24 4. Place of Articulation. 5. Character of Plosives. 6. Labialization. 7. Laryngeal Articulation. B. Sonants 8 25 C. Stressed Vowels 9-13 25 9. General. 10.1a/. 11.1u/. 12. ß/. 13. M. D. Vowel and Diphthong 114-19 28 14. General. 15. /aj/ and /i/. 16. Note. 17. /a«/ and /u/. 18. /ah/ and /a/. 19. Conclusion. E. Glottalization 20-30 30 20. /?V, ?R, V?/. 21. /i?/. 22. /!?/. 23. /R?/. 24. /-m?, -n?/. 25. /-I?/. 26. /-i?, -U?/. 27. /-h?/. 2«./-R?-/. 29./-i?u-/. 50. Note. F. Vowel and Diphthong II 34 5/. General;/si? = i?/. 52. /a«?/ and /u?/. 55./ah?/and/a?/. 54. Distribution. G. Unstressed Vowels 55-55 35 55. General. 56. Reduction. 57. Progressive Assimilation. 55. Transcription. III. Special Questions 39-56 A. Juncture 39-40 39. /C.R/. 40. /R.?/. B. Consonants 41-46

38 38 39

12

TABLE OF CONTENTS

41. Coalescing. 42. /s/ or /c/. 43. Loss of /h-/. Loss of /?-/• 4.5./?K/and /(?)K7- 45. Geminates. C. Svarabhakti-Vowels 47-49 47. The type [CVRaC]. 48. The Types [naCu-] and [nsaK-]. 49. Other Cases. D . Exceptional Vowel-Types 50-51 50. Long Diphthongs. 51. Ultrashort Vowels. E. Clitics 52-56 52. Introductory; Transcription. 53. Juncture. 54. Initial /?-/. 55. VowelAssimilation. 56. Moiphophonemic- and Reduction-Rules. IV. Structure of Morphemes 57-66 A. Morphemes /CVC/ 57-64 57. General. The Type /CVh/ 58-60 58. /Csh/. 59. /Cuh, Cih/. 60. Reduction. The Type /CVR(?)/ with /h, u , \ j 61-64 61. General. 62. /CVR?/ and /CA?/. 63. Reduction. 64. Special Cases. B. Longer Morphemes 65 C. Affixes 66 V. Questions of Word-Phonology 67-78 A. The Occurrence of /?/ 67-73 67. General. 68. Morphemic Glottalization. 69. Junctural Glottalization with Vocalic Suffixes. 70. Idem with Consonantal Suffixes. 71. Reduplicative Glottalization. 72. Suffixal Glottalization. 73. Conclusion. B. The Occurrence of /a/ 74-75 74. Rules. 75. Problem of Elimination. C. Vowel and Sonant 76 D. Consonant-Groups 77 E. Alternations 78

41 42 42

45 45 46 47 49 50 51 51

56 58 58 59

PART II: MORPHOLOGY

VI. Introduction 79 VII. The Nominalizer /s-/ 80-81 80. General. 81. In Factual Forms. V n i . Transitivizers and Intransitivizers 82-84 82. General. 83. Meanings of Transitivizers. 84. Examples of Forms. IX. Transitivizers 85-106 85. Introductory; Recessive Transitivizers. A. Transitive /-t/ Suffixes 86-90 86. General. 87. Zero-Grade Stems. 88. Stems with /a/. 89. Stems with /i/. 90. Stems with /u/. B. Causative /-s/ 91 C. Transitive /-n/ Suffixes 92-101 92. General.

63 66 68 71 71 73 74

TABLE OF CONTENTS

The Type /CV'G-n/ 93-96 93. Zero-Grade Stems. 94. Stems with /u/. 95. Stems with /a/. 96. Stems with /i/. The Type /CA 'C-An/ 97-99 P7. Stems with /a/. PS. Stems with /i/. PP. Stems with /u/. The Type /CVCC-a'n/ 700 Longer Units 101 D. Transitive /-nax 0 / 702 E. Complex Transitivizers 103-106 103. General. 104. /-Sit/. 705. /-nit/. 106. l-min?/. X. Intransitivizers 107-122 707. Introductory. A. Plain Intransitivizers 108-118 Simple Types 108-114 108. /Ca'C-m/. 109. /CVm?/. 770. /CA'C-m/. 777. /CA'C-Am/. 112. /CC-Vm(?)/. 113. /CV'CC-m/. 774. /CVCC-a'm/. Lo/rger l/nfty 775-77« 775. General. 77ic,aip>-an/ 285-290

189

285. Introductory. 286. Future. 287. Present Conditional. 288. Indirect Questions. 289. Alternative Questions (Pres., Fut.). 290. In Negative Sentences.

Type II: /fi-n^q^c'ic'a'p'/ 291-293

191

291. Sub-Types. 292. Ila, b : Past Irreal Conditional. 293. lie: Potential.

Type III: /q^Pan^'ic'a'pV 294-301

192

294. General. 295. Conditional-Temporal. Conditional. 297. With Verba Sentiendi. 298. Alternative Questions (Past). 299. Sub-Types in Negative Sentences. 300. Ilia: Past; Present. 301. Illb: Future; Imperative.

Type IV: / q a ^ _ c ' i c V p 7 302 Type V: /hay-q/ 303 XXVII. /¿./-Clauses 304-312

194 195 196

304. Classified.

Type I: /X.w?an/ 305 Type II: /?nA/ 306-312

196 196

306. Introductory. 307. Ila (/ya/-Type) in Causal Clauses. 308. Ila in Local Causes. 309. Ila in Modal Clauses. 310. Ila in Special Expressions. 311. l i b (Zero-Type). 312. He (/k ,c a/-Type).

xxvra.

Special Topics 313-327 A. Negative /k°i/-Clauses 313-316

199 199

313. Introductory. 314. Type /q/ Illb. 315. Type /q/ IV. 316. Other Types.

B. Factual Forms Without /k°i/ 317-318

199

317. After Local Predicates. 318. As Predicates.

C. Zero-Clauses 319-322

200

319. Introductory. 320. Conditional-Temporal. 321. With Verba Declarandi. 322. Other Cases.

D. Nominal Verb-Forms 323 E. Ad-hoc Nominalizations 324-327

201 202

324. Introductory. 325. Intrans. Stems. 326. Trans. Stems. 327. Pass. Stems.

XXIX. Predicative Clitics and Suffixes 328-356 General Data on Order 328-336

206 206

328. /na, «a/. 329. It/. 330. /?u/. 331. /ua, 5-/. 332. /na, £-/. 333. /na, »a,«-/. 334. /nam?, m?i/. 335. /?aq', Pit'/. 336. /k°/.

The Clitic /Pi/ 337-339

200

337. General. 338. With /tmta'm, fia'X/. 339. In Subordinate Predicates.

The Clitic /maty 340

206

TABLE OF CONTENTS

17

The Clitic Imn/ 341-348 209 341. General. 342. In Factual Forms (Past). 343. Idem (Fut.). 344. Emphatic Alternant /mani'ty. 345. After /¡aty. 346. After /mnwja'ty. 347. After /¡a'mn/. 348. In Other Surroundings. Suffixes 349-350 211 349.1-6', -x7. 350. /-ka/. Combinations of\i\\ and /k°j 351-356 212 351. Classified. 352. la: /?i_k°_ò-n/. 353. Ib: /?i_k 0 w ?i_a-n/. 354. II: /?i_e-n_k°/. 355. Ill: /?i_6-nw?iwk°/. 356. IV:/?i_k°_i-n_mn/. XXX. Coordination 357-365 214 Coordination of Predicates 357-364 214 357. Introductory. 358. /«aty. 359. /¡a'mn/and/?i...k°/. 360. Juxtaposition of Finite Forms. 361. /huj/. 362. /¡aX/. 363. /Pi/. 364. Juxtaposition of Factual Forms. Coordination of Relata 365 215 TEXTS Introductory Note I. Qa'lqaliX, Analysis II. Sm3iq>0a'? III. Siinukqai? and X°8'5ta'l? IV. ?3?u'? x°ali'tn 'The First White Men' V. A Hunting Incident

219 219 221 229 230 236 240

DICTIONARY Introduction to Dictionary Organization Etymologies Dictionary

245 245 245 248

Appendix

401

Bibliography

406

SYMBOLS, ABBREVIATIONS, REFERENCES

A = full vowel (vowel other than /a/) AP See p. 9 C ~ consonant C» = non-uvular cons. C u = uvular cons. C,, C, = 1st, 2nd cons, of root IJ Seep. 9 K = plosive or fricative LM See p. 9 N — nominal predicate N a = noun in absolutive case O = direct object

R (Phonology:) = resonant (Syntax:) = secondary relatum (complement other than S or O) R, — initiator of action expressed by V p R, = R other than R, S = subject T - sentence-adjunct V (Phonology:) = vowel (Syntax:) = verb Vi = intransitive verb Vp = passive verb Vt = transitive verb

ab. = about act. = active approx. = approximately borr. fr. = borrowed from caus. = causative CdA. = Coeur d'Alene Ch. = Chilliwack cons. = consonant Cw. = Cowichan dimin. = diminutive fem. = feminine geogr. = geographical term Hale. = Halcomelem itr. = intransitive Kal. = Kalispel lge. = language masc. = masculine Ms. = Musqueam

obj. = object orig. = original(ly) pass. — passive pers. = person pi. = plural poss. = possible, -ly pref. =- prefix prob. = probable, -ly rec. = recorded red. = reduplication, -ted refl. = reflexive Sh. = Shuswap Sq. = Squamish subj. = subject suff. = suffix tr. = transitive unstr. = unstressed zool. = zoological term

[...] phonetic notation /.../ phonemic notation 0 zero element

* reconstructed form; form not recorded t conjectured form V root (in phonemic notation)

Numbers in italics refer to sections. Roman numbers followed by Arabic ones in Roman type refer to texts and numbered sentences thereof. References to literature are made by name of author, year of publication and page (see Bibliography). For the alphabetic order used in Grammar and Dictionary see p. 245.

PART ONE

PHONOLOGY

I PHONEMIC DATA

I. Phonemes.— The Squamish sound-system (see chart on p. 22) comprises 29 consonants, 4 vowels, and a feature of glottalization (glottal stop). The consonants (symbolized "C") are subdivided into 23 consonants proper (Gerauschlaute, symbolized "K") and 6 resonants ("sonants" for short, symbolized "R"). The vowels ("V") fall into 3 full vowels ("A") and a schwa. The consonants proper fall into the following series as regards place of articulation: labial /p, p7, dental /t, t \ c, c \ s/, palatal /£, 5', s/, lateral ¡X, V/, velar /k, k°, k', k'°, x0/ and uvular /q, q°, q\ qJ°, X, i°/. As for mode of articulation, the fricatives /s, x°, i, are opposed to the plosives of the respective series. In the dental series, the affricates /c, c'/ are opposed to the stops It, t7- Labialization plays a role in the velar and uvular series, where /k°, k>0, q°, q>0, are opposed to plain consonants, while the velar fricative /x°/ lacks a plain counterpart. The plosives of all series, and also the laterals (where a distinction plos. - fric. is lacking), oppose glottalic members / p \ t \ c', 5', X', k \ k'°, q \ q>0/ to plain ones. The sonants comprise labial and dental nasals /m, n/, a lateral liquid ¡1/ and laryngeal, labiovelar and palatal glides /h, u, j/. The vowel-system opposes open/a/to close front /i/ and back-rounded/u/ ;in addition, there is a mid-central vowel /a/. For a particular type of /i/ written /i/ see 22, 62. The consonants /k, k'/ (but not their labialized counterparts) are of markedly rare occurrence in the lexicon (cf. the lack of */x/ vs. the presence of /x°/). It would be wrong to regard /p\ q'% etc., as sequences */p + ?/, */q + ? + u/, etc. In the first place, glottalic and labialized consonants are as frequent in the lexicon as plain ones (/k°, k'°/ are many times more frequent than /k, k'/, while /c'/ is about twice as frequent as /c/). In the second place, in case of so-called partial reduplication (in general: Q V C j QV-CjiYJCa), the glottalic and labialized consonants are reduplicated as a whole. In the third place, such an analysis would entail quite arbitrary decisions as to the order of the elements of oral closure, glottalization and labialization: */q?u, qu?, ?qu/, etc.

22

PHONOLOGY CONSONANTS PROPER FRONT

BACK

RESONANTS

ñas. liqu.

Plos. Fric. Plos. Fric. uvulars labials

1 s° Ì1

q q> ¡ * 1

P P'

velars k°k >0 k k'

§ dentals

s ««•I 1 ia ^i i 1 2

m

VOWELS

glide high low

high

GLOTT.

low

u

y

x° h

1

laterals

j

a—

X X» »

c c'

palatals

n

i

i(î) 1

fi 5' 1 ä i

Only in ft?/ (see 22, 62). Combination of /?/ and /-/ (see below).

Other symbols: /'/ /./ /w/ /-/

Stress (written after stressed vowel: /a1/). Special type of juncture (see 5); combined with /-/ as /»/. Connects clitics and full words (see 2). Separates morphemes (no phonetic value); combined with /?/ as /*/ and with /./ as /»/.

Vowels do not occur initially; if no (con)sonant precedes them, they are preglottalized: /?a-, ?u-, ?i-, ?a-/. The vowel /u/ is much less frequent in the lexicon than the others. This is to a large extent due to the fact that the distinction between /u/ and /ay/ has been preserved better than that between /a/ and /ah/, /i/ and /ai/, see 74-/9, 37-54. Within the limits of the morpheme, the vowels /a, u, i/ cannot be directly opposed to the sonants /h, u, j/ (see 76). Furthermore, the distinction of /a, u, i/ on the one hand and /ah, au, aj/ on the other hand is precarious (see 74-79, 31-34). The presence or absence of /a/ is to a large extent predictable if the morphological structure of a word is given (see 74). The presence vs. absence of /?/ can distinguish free forms ending in a vowel or sonant, cf. /tai?/ 'that one' vi. /taj/ 'canoe-race', /lam?/ 'house' vs. /lamI 'whiskey, liquor'; but morphemes of this type can lose or acquire /?/ as a

PHONEMIC DATA

23

result of morphological processes (see 67-73) and the relation of I?I to semantic content is not on a level with that of the other phonemes. 2. Stress.— There is a strong expiratory stress, written /'/ after a vowel. Its position is not restricted to any particular part of a word, cf. /Ca'jai?/ 'widow(er) of sibling' vs. /C'auPa'i?/ 'wooden spoon'; /WX,'isi/ 'have jaundice' vs. /sisili'us/ 'get scared'. A number of words regularly occur without stress; in transcription, these clitics are connected with each other and with the word with which they form a stress-unit by means of the symbol /. ./, e.g. /5n w .ua w c J ic > a'p > / 'I am (was) working'. Longer words may have two stresses, e.g. /sma'q^a 1 ?/ 'crane' or may allow alternative stresses, e.g. /Xi'cqn, Xicqa'n?/ 'fell (a tree); tr.\ or may be pronounced with one or two stresses, e.g. /k°a l iucmix°n, k°a'iucmi l x°n/ 'to murder' (the stress closest to the end of the word usually being the facultative one). A stress-shift towards a later syllable, or the occurrence of a second stress, are particularly common when a word is extended with suffixes, or is followed by a clitic: one says /si'XaPn/ or /siX,a?a'n/ 'buy; tr.\ but normally /siXaPa'n^Cx 0 / '(you) buy it!', cf. also /na^S'a'uatas/ 'he helped him' vs. /na^CVyatasuit/ or /na w £'a l yata l suit/ 'they helped him'. Elements which normally occur as clitics can occasionally receive a stress. This is the case (a) in emphatic statements, e.g. / n a w y a w ? m ? u ' t / 'he is at home' vs. /na'P ua^PmPu't/ '(yes,) he is at home'; (b) when the clitic is provided with an adverbial suffix, e.g. /na w ua w ?3sXi'c/ 'he is lying down' vs. /na'?-x° ua_?3sXi'c/ 'he is still /-x°/ lying down'; (c) facultatively in some cases where a clitic follows the word of which it is a satellite, e.g. /ha'?X. w 6n/ or /ha'?X, w 5a l n/ 'I am good' (these cases are probably of the same type as the facultative second stress in certain longer words, and I maintain the symbol / w / here). 3. Juncture.— In those cases where two types of juncture are opposed to each other, the symbol /./ is employed to indicate the less frequent type. The cases in question are (a) syllabic vs. non-syllabic /./ contact between consonants, see 8, 39, 48, 52; (b) close vs. open /./ contact between consonants, see 46, or between a sonant and a following /?/, see 28, 40. Reading rule: to /C.C/ there corresponds phonetically [CC], i.e. consonants are pronounced without intervening [a] and without coalescing.

II PHONETIC DATA

A. CONSONANTS

4. Place of Articulation.— The labials are bilabial. The dental closure is formed with the tongue-tip against the back of the front teeth, but with a larger area of the tongue touching the alveoles than is the case, for instance, in French, so that /n/ sounds slightly more palatal than French /n/, though it is closer to the latter than to French /fi/. The palatals are closer to Russian "soft" /5/ than to "hard" /§, 2/. The laterals are bilateral; in /X, X'l the tongue is released approximately from the second molars on, and their acoustic effect is closer to "tl" than to "kl". The sonant /I/ is in between the "mid" variety (as in Dutch) and the velarized type (as in English); it is most strongly velar after a vowel before cons, or at the end of a word, and least so in word-initial position. Of the velars, /k, k5/ are slightly more front than /k°, k >0 , x°/; they are articulated at the same place as English /k/ in "cat", while the labialized velars are farther back than Engl, /k/ in "cool". The uvulars are quite close to the labialized velars, so that /q°, q >0 , i°/ are hard to distinguish from /k°, k >0 , x°/, though there is no doubt that they constitute an independent series, cf. /q°aX/ 'drift ashore' vs. /k°M/ 'split open; itr.', /t'a'q^an/ 'cut, break; tr.' vs. / t W a n / 'dig; tr.\ / p a ^ n / 'spit at; tr.' VJ. /spax0/ 'animal stomach, tripe'. 5. Character of Plosives.— The dental affricate /c/ differs from a sequence /ts/ by the shorter duration of its fricative element and by the lesser energy of its explosive component. The difference was observed only immediately before a stressed vowel: /ci?/ [ce?] 'there is' vs. /tsi?/ [tse?] 'feel cold', /sca'mjias/ [sca-'mies] 'Tuesday' vs. /stsasI [stsa's] 'poor'. In the sequence /tn/ the /t/ normally has a velic release [T], e.g. /Aa'C'tn/ [Xa- '5'Tij] 'knife'. Of the other plosives, the labials are stops, the palatals are affricates, of the laterals, /A/ varies from an affricate with a weak plosive element to a pure fricative [A,], whereas Ik'/ is always an affricate (the two are transcribed [A, X'] henceforth). The non-glottalic velars and uvulars are stops, while the glottalic ones are often affricated, especially before another consonant. 6. Labialization.— In the labialized consonants, the element of rounding is present during the whole articulation of the sound. It starts with the implosion (so that it

25

PHONETIC DATA

can affect a preceding vowel) and it persists after the release, where it is audible as a short u-glide before vowels other than /u/. Before /u/ plain uvulars become automatically labialized. 7. Laryngeal Articulation.— The non-glottalic plosives are facultatively aspirated before stressed vowels. In the glottalic consonants, the oral and glottal releases are simultaneous. The sonants are phonetically voiced. B. SONANTS

8. The sonants /m, n, 1/ when preceded by a consonant are pronounced as syllabic [iji, 9, J] or as [am, an, al] (instead of [a] also [e, ae, u, o] depending on the preceding cons.), e.g. /tmta'm/ [tipta-'m], etc., 'when?', /tmi'x 0 / [tijie* 'x°] 'earth, ground', /qli'm/ [qjc'm] 'weak'. Where the pronunciation is non-syllabic (i.e. primarily explosive), the symbol /./ is written before the sonant, e.g. /mu's.mas/ [mc'smas] 'cow', /k >0 a'5.nax7 [ k ' V ^nax 0 ] 'to see; tr.'. Compare: /t'lma'j?/ [t'alama-'i?1] 'wild cherry tree' and /st'l.mu'Pt/ [st'almo'Pt] 'old person'. The sonant /h/ when preceded and not followed by a vowel is produced as a lengthening* (and facultatively opening) of a preceding /i/ or /u/ and as a lengthening and opening of a preceding /a/ (after /a/ the sonant /h/ does not occur). In other words: the sequences /ih, uh, ah/ before cons, or at the end of a word are produced as long vowels [e:, e:], [a:, o:] and [a:]. In the case of /uh/ and /ah/ there may be a schwacolored offglide [*] before a consonant. Examples: /5'ihn/ [C'ein, 5'e:n] 'lift, raise; tr.', /puht/ [po:'t, po:'t] 'blow; tr.', /c'ah/ [c'a:] 'be hit, bump', /sahc/ [sax] 'be left over'. C. STRESSED VOWELS

9. The Squamish vowels — especially /a/ — show more variation than the consonants, and they vary mainly in function of the latter. The following symbols are used for the different timbres (those in parentheses occur in unstressed syllables only): Front

Central

plain

plain

Low

a

L.-Mid

X

Mid

8

H.-Mid High

(i)

Back

half-rounded

rounded a

A

ö

a

0

0

(ü)

00

* Three degrees of vowel-length are rendered as follows: short (no indication), half-long [a-], ong [a:].

26

PHONOLOGY

10. Stressed /a/ is a somewhat front [a-], in between the vowels in German ja and French pâte: /man/ [ma'n] 'father', /qaX,'/ [qa'V] 'clouds', /staq°/ [sta-q°] 'water', /X°ai/ [X°a-i] 'be senseless'. In the two words where /a/ occurs after /k/, it has a more palatal timbre [ae ]: /kay/ [kae-y] 'descend', /kat/ [kae't] 'ascend'. In both words, the vowel can be narrowed to [e] and shortened : [keu, ket] ; these must be regarded as phonemically alternative forms /kay, kat/. There are no other words with either /ka1-/ or /ka1-/; in /k'a'k'ltX/ [kVk'eltX] 'gamble' there is the usual [a-], different from /a/ in / k W a ? / [ k ' e ^ a ? ] 'lacrosse'. Before a final consonant-group, /a/ is shortened from its usual half-length: /na*6/ [naXfi] 'hand', hard to distinguish from the vowel in /c'ait/ [c'aeXt] 'gravel beach'. 11. Stressed /u/ is pronounced [o ] as in German Sohtt: /cut/ [co't] 'say', /t'uk' 0 / [tVk > 0 ] 'go home', /nuq 0 / [noq°] 'noon', /q>0uj/ [q>0o"i] 'to die'. Occasionally one hears more open variants, but not regularly, e.g. in /lu'lum/ [lo-llom, kr'lom] 'sing' (where the medial [1] may be responsible). After /X, X'/ the vowel /u/ varies with a diphthong, e.g. Aup/ [Xo'p, Xôup] 'be out of reach', /Xus/ [tars, X.ous] 'slide down' ; there are intermediate variants with a diphthong [oo]. 12. The main variants of stressed /i/ are a rather close [e'] as in German Weh, [£•] as in English there and [ej[] as in Dutch ei. The variant [e-] can be a slightly diphthongal [e'i], approximately as in Dutch zee, especially at the end of a word and before dentals and palatals. The variant [e ] may have an open offglide [e®], in which case it closely resembles the vocalic part of British English there. The choice of variants depends on the surrounding consonants in terms of uvulars and non-initial /l/ (symbolized C u ) versus all others (symbolized Ct) as follows: (a) /CtiC,/ [c-] (b) /CiCu/ [e-] (C = any cons.) (c) /C„iC t /[ei] Examples: (a) /ti/ [te-] 'this', /Pi/ [?c] 'be here', /cix0/ [ce'x0] 'reach, get there; itr.', /tiX/ [Se-Jl] 'top', /k>0in/ [k>0e-n] 'few'. (b) A'iqI [X'e-q, X.Je»q] 'arrive', /t'iq 0 / [t'e-q0] 'cold', /uilq ,0 t/ [ue lq>0t] 'ask (a person); tr.', /q'i'Xi?/ [q'e-'Xei?] 'become black', /Xiq/ [Xe'q] 'scratch; itr.', /*°iq>0/ [X°e-q>0] 'be arrested', /nq'i'lus/ [naq'e^los] 'wise'. (c) /q'it/ [q'eit] 'be morning', /XipV [fejp'] 'get touched, nipped', /*i 'c'im/ [Xej 'c'em] 'itch', /x°aliitn/ [x^leiTç] 'white man' (but cf. after initial /I-/: /lix°/ [le'x 0 ] 'fall'). In a number of words diphthongal variants [ei] were recorded as alternatives of [e-], e.g. in /pi'ô'm/ [pe-'ô'ip, pe'iô'ip] 'to spark', /5isk°/ [5e-sk°, ôei§k°] 'recede, ebb'. See 16. The above delimitation of variants is not absolute. A non-adjacent uvular may

27

PHONETIC DATA

influence a vowel in neutral immediate surroundings, e.g. /Xicq/ 'fallen tree, timber' is pronounced [Xejcq] or [Xfi'cq], cf. also /ui'taX/ [ye 'tasX] 'type of canoe', /cWaX 0 / [c'e-'t'aeX0] 'diving duck'. On the other hand, in reduplicated forms vowels in different syllables may influence each other: the word /X>i'X,5iX°ai/ 'brook trout' is pronounced [X'e''X'efai], with an "irregular" [£•] in the first syllable under the influence of the uvular in the next one (see above), or [X'e* lA,>ei°ej[], with an "irregular" [e] before uvular in the second syllable due to assimilation by the "regular" [e ] in the first. 13. The vowel /a/ stands out by being shorter and more variable than /a, u, i/. Its timbre centers around [A] as in British English but, and the deviations from this center are determined by the surrounding consonants. The variants are tabulated below, as found in contact with the following consonants, each symbolized by a single phoneme: /[/ (stands only for itself); /£/ palatals; /t/ dentals, laterals and preceding /h-/; /p/ labials; /q/ plain uvulars and following /-l?/; /k°/ labialized velars and /u/; /q°/ labialized uvulars and following /-uC/ (C = any cons.). l-il

/-ö/

l-tl

/-p/

/-q/

/-k°/

li-l

e

£

£

A

A, ae

E, Ö

ae, ö ,

a

/E-/

e, e

I

E

A

A, ae

Ö

ae, ö ,

a

M

e, e

e, A

A

A

A, ae

Ö

ö,

a a

/-q 7

/P-/

e, e

A

A

A

A

Ö

ö,

/q-/

e , ae

A, ae

A, ae

A, ae

A, a

as, ö

a

/k°-/

o , Ö, e

0 , Ö, E

0 , Ö, A

0, Ö, A

o,

o , Ö, e

o, ö, e

0 , Ö, A

0, Ö, A

o, a o, a

0, ö

/q°-/

0, ö

o, a

a

The timbre [e] is heard before / i f , which has a strong palatalizing effect: /moj/ [mej, me-i], also [me - ] 'sink; itr.', A'ajn/ [XV'¡in, X'e'ifl] 'stop; tr.', /jp'jaPt/ [ie'jaPt, ie - '¡a?t] 'vomit continuously'. The timbre [e] is found in the neighborhood of /i/ and other palatals: /6'atx 0 / [6'etx 0 ] 'carve; itr.', [Xe§] 'tug-of-war', /qaj/ [qej[] 'bad', cf. also /maj/ above, but otherwise than before /¿/ an adjacent labial requires [A], an adjacent uvular requires [A, ae] (see below). The timbre [i] is heard between palatals (other than /i/): /¿a'sa?/ [Ci'se?] 'mother', /Ca'Sn/ [iiigp] 'send; tr.'; also after /k, k'/ : /k'a'x^?/ [k>ilx°a?, k ' e ^ a ? ] 'lacrosse', /skamPc/ [ski m ?c, ske m ?c] 'small clam'. The timbre [ae] occurs in the neighborhood of plain uvulars, especially if the other adjacent cons, is a dental or palatal: /saqV [saeq', sAq'] 'crack, split; itr.', /ssq/ [Saeq, SAq] 'be finished', /qaC/ [qaeS, qAC] 'be full', [PaeX, ?aX] 'wild goose'; [ae] is particularly common before /%/, where it is heard also if another uvular precedes: /qaX/ [qaeX] 'much', /Xa'Xn?/ [Xae'Xffi11?] 'frost'; a following /-I?/ has the same effect as a uvular: /telPt/ [tae11?^ 'learn; tr.'. The timbre [a] is heard especially before labialized uvulars, also before plain ones

28

PHONOLOGY

if another uvular or labialized cons, precedes: /t'aq10/ [t'aq >0 ] 'break; itr.', /taX0/ [taX°] 'be settled', /k>0aq/ [k J °aq, k>0oq] 'be split'; also between a labiovelar or uvular and a following /-uC/: /XauX'/ [XauX,', XaeuA,'] 'break; itr.', /sx°a'uqn/ [sx'Vuqij] 'white swan', /q'a'um/[q'a'uip, qV^um] 'howl, whistle'; also after /?/: /PaX/ [PaX, ?xi] 'wild goose', /Paq°/ [?aq°, ?oq°] 'fall out (ab. hair)'. The timbre [a] is heard after labialized consonants. Between labialized and plain cons, /a/ is a diphthong in which either the initial [a]-element predominates, or the delabialized final part; this may be as much a matter of impression as of actual differences in articulation. The word /x°at/ 'wren' is heard as [x°oet, x°ot, x°At]; the diphthongal articulation creates even the impression /x°et]. The same variation in /x°aj/ [x°oj, x°5j, x°ej[] 'appear'. Before and after uvulars the range of timbres is less wide: /k'°aq/ [k>0oq, k'°aq] 'be split', / X ° a s / [ X ° o s , X°AS] 'fat', /q°aq°tq/ [q°oq°tq] 'pass by; itr.'. The timbre [o] is heard before labialized velars and /u/: /tak'°/ [tok >0 ] 'tight', /spax°/ [spox°] 'animal stomach, tripe', /nau/ [nou] 'thou', /sa'uaj/ [so'uai, so'uaej] 'grow'; also between /¿/ and a labialized cons.: /iaq0/ [ioq°] 'clothes', and as an intermediate stage between the variants [a] and [A, e] after labialized consonants, cf. /x°at/ above. The timbre [A] is heard between labials, dentals and laterals: /pac/ [pAc] 'bend, fold; itr.', /paA,c/ [pAta] 'thick-lipped', /sap/ [sAp] 'stiff', /tamty [tAmk] 'red paint', /c'als/ [C'AIS] 'shiny, cleaned', /tas/ [X.AS] 'bitter, sour'; also between a palatal and a labial: /5amX/ [CAmX] 'pitch', /ma'Sn/ [mA'Cjj] 'black louse'; also as an alternative for other timbres, cf. /saq', saq, qa5/ under [ae] and /x°at, X°as/ under [a] above. D.

VOWEL AND DIPHTHONG I

14. A distinction between /aj, ay, ah/ on the one hand and /i, u, a/ on the other hand exists, but certain reservations must be made. In the first place, the two types easily merge, especially in more rapid speech. In the second place, the distinction is in many cases a matter of a wider vs. a narrower range of variants, the former including the latter. For instance, /maj/ 'sink; itr.' is pronounced [mej] or [me-], whereas /mi/ 'come' is pronounced [me-] (and not *[mei]). In the third place, the distinction is in some cases a matter of presence vs. absence of a morpheme-border, cf. /pa'j-m/ [pe- 'iim, pe'itp] 'fall overboard' vs. /Paci'm/ [Pace- 'm] 'small'. In the fourth place, the distinction is limited to certain positions: it is not made before a vowel, and, particularly in the case of /ai/ vs. /i/, it was not observed in several other positions (see below). Etymologically, /ai, au, ah/ and /i, u, a/ are in many cases identical. 15. A distinction between /ai/ and /i/ is made consistently only before uvulars, cf. /XajX/ [Xei*, XsiiX] 'war', /X.'ajqV [^'eiq'] 'get trapped', /saiq'qs/ [sejq'qs] 'pass across a flat surface' versus /X,iq'/ [^e a q'] 'always', /siq5/ [seaq£] 'shingles'. The distinction can furthermore be made under the stress at the end of a word, cf. /maj/ [mei, me-] 'sink; itr.' vs. /mi/ [me - ] 'come', where a diphthongal variant is lacking; likewise in /ti/

PHONETIC DATA

29

'this', /Pi/ 'be here'. Where a uvular or labialized cons, precedes, a diphthongal pronunciation is regular at the end of a word, and I write / s j j : /qaj/ [qej] 'bad', /q°aiI [q°ojL q°ej] 'be at lowest ebb', /x°3j[/ [x°oi, x°ei] 'appear'. In the third place, the distinction is made before /-n/ or /-m/ not followed by a vowel (morphologically, these cases concern morphemes /Cai/ followed by the transitivizer /-n/ or the intransitivizer /-m/, as opposed to single morphemes /Cin, Cim/), cf. /Xa'jn/ [Xe-|iln, Xe'ig] 'stop (people) from fighting; tr.', A'a'in/ [ A V i i n , X'ejjj] 'stop; tr.', /pa'jm/ [pe -| ilm, pejip] 'fall overboard', with a disyllabic pronunciation different from e.g. monosyllabic /k'°in/ [k > 0 e - n] 'how much?', though in rapid speech the former type tends to merge with the latter. In all other positions no distinction between / and /i/ is maintained, and only /i/ is written, so that morphemes /mai, qaj/, etc., are written /mi-, qi-/, etc., when they are unstressed, or, if stressed, when not followed by / - n # , - m # / ( # = cons, or zero), e.g. /qsj/ 'bad' but /qi'uPc/ [qe'joPc] 'blunt' ('bad-edged'), /qis/ [qeis] 'dislike' ('consider bad') /qii'?/ [qeie1?, qiie1?] 'become bad'; /maj/ 'sink' but /mi'mi/ [me*'me] 'drown' (reduplication). 16. Note.— Otherwise than before uvulars the distinction of /ajj and /i/ is facultative. L M , with whom the matter was gone into in some detail, did not allow a diphthongal pronunciation in the words /ti, ?i, mi/ quoted above, and in careful pronunciation distinguished /maj/ from /mi/. It is worthy of note that /ti, ?i, mi/ all three occur as clitics (the latter, also pronounced /m?i/, is a reduction of /hsmPi1/ 'come'); these words do not belong to the "lexical" part of the morpheme-inventory (neither does the article /k°i/, but this does not occur stressed). The simplest favorite form of lexical morphemes is CVC, to this type /maj/ belongs, and obviously also /qsj/ 'bad', etc., where therefore /aj/ is written even though after uvulars the diphthongs [ei, oj] cannot be opposed to plain vowels. — I J and AP used diphthongal variants in more positions than LM, but not consistently, e.g. in the stressed vowel of /pi'S'm/ 'spark; itr.', /p'icV 'get squeezed', /5isk°/ 'recede, ebb', /5'i67 'twist; itr.', /sik 0 / 'clam', /X-i'c'm/ 'spark; itr.', /sX.i'sm/ 'cedar-leaf rope', /Pis/ 'have a good time'; on the other hand, in the material recorded from them I find only [e - ] in /ns?i/ [nsPe1] 'loud', /hii 1 / [hiie -1 ] 'big' (once pronounced with emphasis with a long-drawn final [i:], so that an interpretation /hii'h/ is possible). The absence of diphthongal variants in words recorded from I J and AP may in individual instances be accidental; for instance, the words /Pis/ and /nsPi/ quoted above may well contain an etymologically identical element */PaU 'good', though only o f the first a diphthongal variant was recorded. My transcription reflects the pronunciation of L M , but where diphthongs were recorded from I J or AP this is indicated in the Dictionary by [aj] after the word concerned. The non-glottalized variant of the inchoative suffix /-i(P)/ (see 186:34) allows under the stress a diphthongal pronunciation (LM), e.g. /p*si(?)/ [p'seP, p'sejp, p'se - , p'sej] 'go to shore' (derived from /p'ss/ 'k/.'; for the pronunciation of /i?/ see 21).

30

PHONOLOGY

17. The distinction of /au/ vs. /u/ is phonetically much more clearcut than that between /aj/ and /i/. Only after / X , X,1/ there is free variation of a vowel and a diphthong (see 11), and here I write /u/. Otherwise, /ay/ and /u/ are written as recorded, though it must be noted that the functional yield of the opposition is minimal. Before vowels no distinction is made; I write — conform pronunciation — /a'u/ in stressed and /u/ in unstressed position, e.g. /sa'uaj/ [So'uaei] 'grow', /sua'X,/ [Sua'X] 'trail, door'. In stressed word-final position /ay/ was recorded in /na'u/ 'thou', /u/ in the interjection /?u/ 'allright' and in the borrowings /k°a§u'/ 'pig' (French cochon) and /kapu 1 / 'coat' (Fr. capot). Stressed before cons, there is /ay/ in I tank3/ 'break; itr.', /sx°a'uqn/ 'white swan', /q'a'um/ 'howl, whistle' ([q'a'uip, q'a'tfum], morpholog. /q'ay-m/, parallel to /paji-m/ in 14, 15), while there is /u/ e.g. in /¿uX/ 'burn; itr.', /sXVp'aPn/ 'women's neckwear', /sum?/ 'smell; tr.\ 18. The distinction between /ah/ and /a/, in those positions where they are directly opposed to each other, is phonetically a matter of vowel-length. In stressed word-final position /ah/ fa:] was recorded in /c'a'h/ 'get hit, bump', /q°a'h/ 'be perforated', /sna'h/ 'name', / k ' V h / 'maybe so' (interjection expressing doubt); /a/ [a] was recorded only in /na1/ 'there you are!' (interjection used when handing over something). The latter word is etymologically identical with the clitic /na/ 'there', so that the distinction of /ah/ vs. /a/ parallels that of /aj/ vs. /i/ as regards its distribution in the lexicon (full stem vs. clitic). The distinction /ah/ vs. /a/ is also made under the stress before a consonant, cf. /ta'hm/ [ta':m] 'to leak' (morphologically parallel to /pa'im, q'a'um/ above) vs. /sta'm/ [sta -l m] 'what?', further in /Xa'hm/ 'to cry' (also /Xam/), /sahc/ 'be left over', /sp'ahc'/ 'blackberries', /6'ahX'am/ 'hunt; itr.'. In productive combinations of roots /Cah/ with consonantal suffixes the root is usually reduced to /Ca-/, e.g. /q°ah-n, q°a-n/ 'perforate; tr.', /s-q°a-qs/ 'hole through the nose' (cf. /q°ah/ quoted above; the roots of /ta'h-m, xa'h-m/ are not found in isolation). — Before a vowel, only [-aeh-] is found (to the exclusion of [-a(:)-, -h-]) and I write /ah/ (note that the transcription /a/ or /h/ would also be unambiguous): /spahi'm?/ [spaehe-lm?] 'wind', /iaXaha'm/ [xaeXaeha-'m] 'to be crying' (reduplication of /Xahm/). 19. The positions where /aj, au, ah/ and /i, u, a/ are and are not distinguished, and the transcriptions chosen in the latter case are summed up in a chart on p. 31. These elements usually occur after consonants, and they are quoted in combinations /Caj/, etc. E. GLOTTALIZATION

20. The glottal stop occurs only in contact with a vowel (V) or a sonant (R); for purposes of phonetic description the possibilities /?V, ?R, V?, R?/ must be distinguished. The case /?V/ calls for few special remarks: the glottal stop has no particular effect

PHONETIC DATA Before Vowel

Stressed

Before Consonant

word-finally

uvular (C u ) j - n # , - m # j others (C t )

Ca'i

CajCu

\

Cajn#

j

Ci'

CiC„

|

Cin

{

Ca'*

CauC

Cu'

CuC

Ca'h

CahC

Ca1

CaC

31

CiC t

stressed Ci'V Ca'uV

J unstressed 1 j CiV> 1 1 j CuV1 1

(n° 1 examples) j

CahV'

Chart Section 19

on a following /a, u, i/, and /i/ has the variant required by a following cons.: /Pi1 tut/ [?e - 'tot] 'sleep', /?iq'7 [?e-q'°, ?£ a q >0 ] 'be rubbed'. The variants of/a/ before uvulars tend to be more open and back after /?/ than after other consonants, cf. /PaX/ [ ? « £ , PaX] 'wild goose' vj. /saq1/ [saeq\ sAq'] 'crack, split; itr.', /qaX/ [q»X] 'much'; this must be due to the absence of an oral closure or constriction in /?/. The case /?R/ parallels /CR/; particularly, /?m, ?n, ?1/ are pronounced [Pam, ?ip], etc. (see 8). The case /V?/ requires little comment as far as /a?, u?/ are concerned. The vowel is somewhat shorter than before cons. Examples: /na?/ [na?] 'be on, at', /haPty [haPX] 'good' Aa'Pn/tXa 1 ??] 'touch; tr.\ /c'uP/tc'oP] 'be pulled out', /k°u?s/ [k°o?s] 'spring salmon' /c'u'Pn/ [c'o'Pp] 'pull out; tr.'; a more open variant was recorded in /mu?n/ [mo 1 ?^] 'drop; tr.'. Occasionally an echo-vowel is heard after the glottal stop. In the case o f /a?/ such a variant was recorded in reduplicative formations of the type /QaPCjCa/ (see 158): /sca'Pcqai/[sca'Pcqaj, sca'P^qaj] 'salmonberry', /X,a?X5/[XaPXC, Xae?»X6] 'be on top', /£i*..na'?n£7 [Cekna 1 ?^, aeX.nse 1 ?«^'] 'chief's daughter, girl o f upper class'. In the case of /u?/ such a variant was recorded only in /q^u'PnauPa'sn/ [q^o'pnaupya'sg, q'VponaPua'sij, q'VPonPua'sp] 'assemble; tr.'. 21. The sequence /i?/ offers a more varied picture than /a?, u?/, undoubtedly because it is not distinct from */si?/, whereas /a?, u?/ can be opposed to /ah?, au?/ (see 32,33). What we transcribe /i?/ is pronounced [e?] after neutral, [ejP] after plain uvular and [oj?, ojp, ej?] after labialized consonants. The variant [e?] may have a brief reflex [*] after the glottal closure; in addition, in word-final position it may be opened to [e?]. All three variants were recorded in the root A'i?/ 'dear, expensive, dangerous' and its nominalized form /sX'i?/ 'desire, thing wanted': [X'eP, X'eP, sX'e?, sX.'e?1] (the last variant may be considered to represent the articulation /sVaj?/, see 31. Word-final examples after uvular and labialized cons.: /sq'i?/ [sq'ei?] 'slices o f dried salmon', /x°i?/ [x°oi?, x°Ej?] 'be lost'. Medial examples: /ni?5/ [ne?5, neP'C] 'high seas', /k°ci?c/ [k°ce?c, k°ce?'c] 'person with magic power', /q'i'gn/ [q'e'jPSn] 'small mat used in canoe', /nq°i'Pstn/ [naq^'iPsTp, naq°e l j[?sTo] 'cooking pot',

32

PHONOLOGY

/k >0 i'?qtn/ [k > °o l i?qTQ, k^s'iPqTp] 'fur, skin'. As the last example shows, a following uvular does not affect /i?/ the way it does /i/; it may have an opening influence on the initial part of the articulation, however, so that the variant [ei] rather than [e] is heard after a plain cons.: /t'i'Pqi/ [t'e'jPqei] 'soak dried salmon in water', /t'i l ?q°m/ [t'e l i?q°om] 'spark', /hi?q°i'n/ [hePqV'n, h q P q V n ] 'lamp'. Before a vowel, /i?/ has facultative variants with a glottalized glide written [i?i] (see 28), e.g. /ji'Puty ¿e'i?ioX,] 'fire', /q , 0 i'?us/ [q >0 E'i?os, q 10 e'i?ios] 'spring (season)'. 22. The rare complex /i?/ (to which the occurrence of /i/ is limited) is pronounced [£•?], occasionally[r?]. Examples: /q'iPXja/ [q'e- i?*ie] 'Negro', /Pi?X°/ [?E-?X°, PE- '?3X°] 'all', /k°i?k°ty [k°r?k°X.] 'small canoe'. Occasionally a variant [e? E ] is heard, e.g. /p'i'Pnax 0 / [p's-iPnax 0 , p V P i j a x 0 , p'e'Penax0] 'to hold; tr.', /sqi'PmPq0/ [sqe'?em?q 0 ] ' devilfish' (IJ) (according to AP, who pronounced [sqe - 'Pipaq0], this word is Musqueam and Cowichan, the Squamish equivalent being /st'alXV/). 23. It is characteristic of combinations /R?/ that the glottal closure may either fall at the beginning of the articulation of the resonant, or interrupt it briefly at any point between im- and explosion, or come at the end. This variation is in principle free, but the preferred variants depend on the position in the word in terms o f final, preconsonantal or prevocalic, and also on the particular sonant involved. In words like /lam?/ 'house', /nam?/ 'go', where the glottal closure tends to fall at the beginning of the articulation of the sonant, any clear nasal resonance may be lacking, but they remain distinct from words ending in /a?/, because the color o f the vowel is affected by the implosion of the labial. This variant, with or without clear nasal resonance, is symbolized /a -m P], etc. I f a consonant follows, e.g. in /lamps/ 'his house', the explosive moment o f the labial becomes relatively prominent after the silence of [?] and is audible as a weak [P] which explodes into the fricative: [la -m ?Ps]. This explosive element is of necessity unobservalbe when a homorganic plosive follows, e.g. in /sq°in?c/ [sq°ei n ?c] 'beard'. The extreme variants of the abovementioned words can easily be mistaken for *[la?ps, sq°Ej[?c]. In the case of /I?, j?, u?/ such extreme variants do not occur. This is not necessarily due to a different timing of the glottal closure, but rather to the fact that a preceding vowel is more strongly affected by a lateral or glide than by a nasal. An "l"-element is always audible in /l?/, though it may be of shorter duration than in /I/, cf. /X°il?/ [ X V 1 ? ] 'come out, off' vs. /slil/ [sls'i] 'bunch of blankets'. In final or preconsonantal /j?, u?/, when the glottal closure falls near the beginning, a brief vocalic reflex u ] is heard after its release, sometimes whispered (in case the vocal chords do not start vibrating again after the release of /?/). As the preference for a particular variant is not the same in all cases, the nasals, liquid, glides and /h/ are treated separately, first in word-final and preconsonantal position, and then intervocalically.

PHONETIC DATA

33

24. In the case of /m?, n?/ the preferred variants finally or before cons, have the glottal closure near the beginning: /lam?/ [la -m ?] 'house', /Xim?/ [Xe"m?] 'be accepted', /hin?/ [he-n?] 'be a long time', /sum?/ [so"mP] 'smell; itr.% /man?/ [ m A n ? ] 'child, offspring'. The same variants before cons.: /lam?s/ [la -m ?s] 'his house', /sq°in?c/ [sq°ein?c] 'beard', /k°um?5/ [k°o' m ?£] 'go over a hill', /st'amPq 0 / [st'Am?q°] 'scalped'. 25. In final /-IP/ the liquid is clearly audible and I write [1?]: /q'al?/ [qVl?] 'consent', /X°il?/ [X°E'l?] 'come out, come off' (note the difference from the type / si 'PI/ [se'Pal]), /ial?/ [XslP] 'write; itr.'. Except in /-ilP-/ the preconsonantal variant is the same: /?al?s/ [Pa'IPs] 'to pity' /XalPt/ [XslPt] 'write; tr.', though sometimes an "l"-resonance is heard after [?] so that the impression is almost [?a-l?Js, Xae'PJt]. The sequence /il?C/ is always pronounced [e'PelC], e.g. /X°il?t/ [X°E'?elt] 'take out, take off' (a transitive derivation of /X°il?/ quoted above), ji°V\?mj [XVPElip] 'rope'. This sequence is common in reduplicative formations characterized by glottalization: /Pas.li 'l?x°/ [?3slei?Elx°] 'lying down' (cf. / l i x ° / [k'x°] 'fall down'), /lil?s/ [lei?els] 'be under, below' (cf. /las/ [Lvs] 'bottom'), /sX-Wi'lPqm/ [sA.'eX.'s'Pelqip] 'bug' (a diminutive reduplication of /sX'a'lqm/ 'monster'). 26. Final /-j?/ and /-u?/ are not parallel phonetically. In /-i?/ the resonant is usually interrupted near the beginning, and a brief ['] is heard after the release of the glottal stop: /tai?/ [ta-i?1, ta*?1] 'that one', /slai?/ [sla-i?1, sla"?1] 'fir bark'. In /-u?/, on the other hand, the glottal stop usually falls quite at the end of the resonant: /§au?/ [sa'u?] 'bone' Aau?/ [Xa-uP] 'recover; itr.', /c'iu?/ [c'e-u?]'heal, close'(ab. wound); itr.', /sXau?/ [sXau?] 'salmon backbone'. Both types of variants are possible in both cases, however. Before cons, both resonants have a brief reflex after [P]: /?ai?X/ [Pat»)?1*] 'crab', /hui?s/ [ h o - ^ s ] 'eat;tr.\ /XauPs/ [Xa-(U)?us] 'new', /?au?t/ [?a"(V)?»t 'future'. Before a group of two consonants /-ay?-/ is usually pronounced [a?o], i.e. the postglottal reflex has become dominant: /sau?cq/ [§a'?ocq] 'bony-faced' (cf. /Say?/ quoted above), /stau?x°A,/ [sta'?ox°X,] 'children'. 27. Final /-h?/ occurs only in /tah?/ [ta? a ] 'undergo; be located' (a homophonous root means 'mother' (address), cf. also /hata'h?/ [hata'? a ] 'aunty'), further in /Xah?/ [Xa?a] 'be touched'. In medial position /ah?/ is not distinguished from /a?/. 28. Medially before a vowel or syllabic /m, n, 1/ variants with glottal closure near the beginning of the sonant are usual in all cases: /na'm?n/ [na" lm ?m$] 'go and get someone; tr.' (a derivative of /nam?/ [na' m P] 'go'), /nSa'uPus/ [n5a-|w?uos] 'skinny' (a derivative of /§au?/ [Sa'u?] 'bone'), /si'uPaC/ [se-l»?uae£] 'foreleg', /t^a'jPus/ [t'*a -| i?ios] 'lightning'. Deviations from this norm in one direction may create a trisyllabic impression [na'Pipy], while in the other direction the sonant may be articulated as a whole before the glottalization, e.g. [t>X°a* 'i?os]. However, in these cases the pronunciation of /R?V/ overlaps that of sequences which are phonemically

34

PHONOLOGY

distinct from it, namely on the one hand that of /?RV/ as in /na'Pnat/ [na'Pgast] 'evening', /li'Plam/ [le'PJam] 'little house', /§a'?iu/ [Sa'P^o] 'corpse', and on the other hand that of /R/ + /?V/, written /R.?V/ as in /tl.?a 'c.na5/ [tJpAcniei] 'east wind' (see 40). 29. In a few words sequences [-i?u-, -u?i-, -j?i-] were recorded, and since these cannot be opposed to any other type of glottalized sequence of two sonants [i, u], it is in itself immaterial whether one transcribes /-i?u-/, etc., or /-¿u?-/, etc. The former alternative is clearly indicated by those cases of [-i?i-, -y?i-] which involve stems ending in /-¿?, -u?/ followed by the 2nd pers. plur. possessive suffix /-iap/, e.g. /sia'i?/ 'friend', /Pa-sia'iP-jap/ 'your pi. friend', and I choose this alternative in all cases, e.g. /ja'iPuas/ 'bed platform', /sua'i?ui/ (geogr. name). — It is probable that a number of cases recorded [-e»?u], transcribed /-iu?-/, go back to */-ai?u-/ (e.g. in the suffixes /-iy?iX, -iy?as/). See also 70. 30. Note.— LM claimed on one occasion that there is a difference in pronunciation between the final part of /lam?/ 'house', /sum?/ 'smell' on the one hand, and that of /ham?/ 'be covered', /pum?/ 'swell' on the other hand, pronouncing the former with [ m ?] and the latter with [m?]. IJ pronounced [ha -m ?, po , m ?] however, and for LM the medial pronunciation in /ha'm?n/ 'cover; tr.' was the same as that of other medial cases of /-m?-/. When a number of cases were checked, LM pronounced [m?] (rather than [ m ?]) only in /sx°im?/ 'discarded object'. Having mentioned the pertinent cases, I further disregard this distinction, which in any case is not made by all speakers. F. VOWEL AND DIPHTHONG H

31. The distinction between /ai?, au?, ah?/ and /i?, u?, a?/ — like that between their non-glottalic counterparts (see 14-19) — has a low functional yield, and here, too, there often is an etymological equivalence between the two series. It is again the cases with /i, i/ where the distinction is weakest, in this case to the extent of being non-existent: /ai?/ and /i?/ are either in free variation or in complementary distribution, and only /i?/ is written (see 21; the variants [e?1, ej?, oj?] may be considered to represent /ai?/, the variants [e?, E?] representing /i?/). 32. /au?/ and /u?/ are in several cases kept apart, cf. /na'u?n/ [no'W?uij] 'put in; tr.' vs. /c'u'Pn/ [c'o'?9] 'pull out; tr.', /mu'?n/ [mo 1 ??] 'drop; tr.'; however, the root of the first example appears with /u?/ in /tx°.nu?/ [tx°no'?] '(land) into'. Stressed final /au?/ was recorded in /?asia'u?/ [PasiS'u?] 'seer' and /sXau?/ [sXay?] 'salmon backbone', stressed final /u?/ in /c'u?/ [c'o?] 'come out (being pulled)' and /tx°.nu?/ (see above). In unstressed syllables /ay?/ is found before vowels and syllabic sonants, /u?/ elsewhere. In a few cases the two are in free variation: /Pasq'aq'a'y?, Pasq'aq^u1?/ '(be) together with', /x 0 au?a'x°, x 0 u?a'x7 'not yet'. 33. /ah?/ and /a?/ are distinguished in stressed word-final position only, cf. /tah?,

PHONETIC DATA

35

(ha)t3'h?, A,ah?/ quoted in 27, pronounced with [a'? a ] as opposed to [a1?] in /na?/ 'be on, at', /tx°ta1?/ '(land) onto'. 34. It is worthy of note that /na?, tx°ta?/ and also /tx°.nu?, Pasq'aq'V?/ quoted in 32 belong to the category of so-called "relator-verbs" (corresponding to English prepositions, see 219, 220). This category has /A?/ (A = full vowel) in two cases where related "plain" lexical items have the shape /aR?/, namely in /tx°.nu?, tx°ta?/ vs. /na'u?n, tah?/, and if /Pasq'aq'V?/ has /Pasq'aqVu?/ besides it, this is because the plain uvular apparently resists the reduction of a following /au?/ to /u?/, before which it becomes automatically labialized (5). The tendency of relator-verbs to reduce /aR?/ to /A?/ may be due to a less prominent stress of such words in the sentence as a whole. However, a few "plain" lexical items were recorded with stressed final /A?/: /¿a?/ 'be tight', /c'u?/ 'come out', cf. also—with two stresses—/sma l q >0 a l ?/[sipo l q >0 a 1 ?] 'crane'. G. UNSTRESSED VOWELS

35. In the unstressed part of a word the consonants do not differ from those in stressed syllables. Of the full vowels, /i/ and /u/ in preaccentual position allow a closer pronunciation than under the stress (extreme variants [i] and [u]), e.g. /hiPa'm?/ [he?a ,|m ?, hi?a ,|m ?] 'come home', /tuta'u?/ [tota-'u?, tuta-'u?] 'bright'. Unstressed /a/ has the variants [a, ae, e, o] depending on the consonantal environment in terms of neutral [ae], palatal [e] and labialized [o] consonants, the timbre [a] being heard most often before glides and /I/, in the neighborhood of /?/ and, to a lesser extent, of plain uvulars, and also at the end of a word. Unstressed /a/, and also the facultative vowel is unstressed /Cm, Cn, CI/ (see 8) is a schwa-type vowel [a], strongly colored by neighboring palatals [i] and uvulars [a, se], by preceding labialized consonants, velar [u] or uvular [u, o] and by following labialized consonants, velar [ii] and uvular [ii, o]. As these timbres show, uvulars tend to open /a/, resulting in a pronunciation which largely coincides with that of /a/ (see 38). 36. Vowel-Reduction.— A striking feature in the pronunciation of longer words is the strong tendency to reduce unstressed vowels to [a] and its variants, or even to zero. Here follow some examples (optimal and reduced forms are given in this order): /sta'uaqin/ [sto'yaqeyi, sto'yaeqen, sto'uaqan, -qg] 'graveyard', /ma'lalus/ [mA'lalos, mA'lalas] 'raccoon', /?u'x 0 umix°/ [?o-'x°omex°, ?o -l x°umux 0 ] 'village'; a particular type of reduction in /aft?)/, e.g. /X'VX'iX"^ [Ve- 'X»e*°ai, X l e ,| reX 0 ei, -X°e] 'brook trout', /sli1 *°ai?§n/ [sis''i c ai?§ij, sic'Xcoi?sn] 'foot of mountain', where /aj(?)/ tends to merge with /i(?)/. In the neighborhood of uvulars, reduced vowels tend to be more open, and particularly /i/ is often reduced to [e] in this position (cf. /sta'yaqin/ quoted above); the same goes for /i(?)/ as a reduction of /aj(?)/, e.g. /si'uai?q/ [se- 'yai?q, se- 'ye?q] 'pants'. The unstressed suffixes /-yiX, -yas/ reduce to /-vik, -us/,

36

PHONOLOGY

e.g. /5a'?tuity [5a'?tuek, -uiX, - o l ] 'build a canoe', /nq>0i 'q>0lxjas/ [naq'V 'q>0Juaes, -os] 'stick for holding salmon above the fire'. This tendency to reduce unstressed vowels has had phonemic consequences: many words contain affixes in reduced form even in the most careful pronunciation, so that alternative forms of the affix must be recognized: /-qin, -qn/, /-mix°, -max0/, ¡-uik, -vX/, etc. As an example may serve the suffix /-qin, -qn/ 'head; hair; throat; language' which occurs stressed in /ns-qi'n-m/ [nsqe'jnip] 'rub oil /ns-/ in one's hair', and unstressed in reduced form in /s-na's-qn/ [snA'sqjj] 'hair-oil', in non-reduced form in /6>ix°-i'-qin/ [S>ex°e' 'qeyi] 'be thirsty', lit. 'have a dry /C'ix0-/ throat' (of course, the latter word allows also reduced forms with [-qen, in the same way as the example /sta'uaqin/ quoted above, but /sna'sqn/ does not allow a form with [-qejn] — it is as if, in English, both "forehead" and "bulkhead" had a form with [-id], which only in the latter case allowed a fuller pronunciation with [-hed]). 37. Progressive vowel-assimilation.— The vowel in a stressed syllable exerts an assimilating influence on the vowel(s) following it. For instance, after a stressed /a1/ an /a/ in the next syllable is often pronounced [a] rather than [ae], e.g. /t'a't'aX/ [t'a - 't'aeA,, t V t ' a X ] 'loom' versus /mi'Xa[me^'XaeX] 'black bear'. This tendency to progressive vowel assimilation has had phonemic consequences: certain suffixes often occur with the same vowel as is found in the stem to which they are added, e.g. the transitivizer /-(V)n/, cf. /ma's-n/ 'put, stick together; tr.', /c'a'q'-an/ 'hit; tr.', /mu'j-un/ 'soak; tr.', /Xi 'k>0-in/ 'hook, hang up; tr.' (see sections 86,92,107). (This is a tendency and not a strict rule, cf. /na'm?-n/ 'go and get someone; tr.). For a case of progressive vowel-assimilation involving clitics see 55. 38. Phonemic transcription of unstressed vowels.— The possible reduction of all unstressed vowels to schwa and the possible opening of schwa-type vowels to [ae, o], etc., in the neighborhood of uvulars, combined with the fact that reduced forms in many cases have become standardized, causes certain difficulties with regard to the phonemic interpretation of individual recordings. The following principles have been adhered to: (a) Close vowels [a, I, u, ii] in their regular positions, of which no open alternative [ae, e, o, o] was recorded in the same word or grammatical (not lexical!) affix, are transcribed /a/ (resp. zero before /m, n, 1/, see 8). (b) Open vowels [a, ae, e, o, o] in their regular positions are transcribed /a/, except that /a/ is written in the neighborhood of uvulars in those cases where in an identical morphological formation /a/ rather than /a/ is found in non-uvular surroundings. For instance, the first vowel in [staeqtaqe '¿u] 'horses' is transcribed with /a/: /staqtaqi 'u/ (morphologically /s-taq-taqi'u/) because the reduplicated plural has close vowels, i.e. /a/, in neutral surroundings, cf. /s-mnTma'nit/ [smanma - 'net] 'mountains'. (c) In cases where full and reduced forms were recorded and where the reduced form is very much more common than the full one, or where it was used in deliberately

PHONETIC DATA

37

careful pronunciation, alternative forms are recognized (thus in the case of /nqJ0i 'q>0lus/ quoted in 36, which is much more common than the form with /-uas/). As far as unstressed vowels are concerned, the possibility of errors in individual cases must especially be taken into account in the case of the following pairs: the variant [u] of /s/ and the variant [u] of /u/; the variant [i] of /a/ and the variant [i] of /i/; the variant [ae] of /a/ and the variant [E] of /i/; in addition, certain cases with /a/ may allow full forms which were not recorded (see 97, 111). The account of unstressed vowels given here is certainly incomplete: a special investigation would probably reveal a complex system of neutralizations. For instance, in postaccentual position before /%, X'/ only open vowels were recorded, and according to (b) above I write /a/, e.g., in /mi'Xaty [me-'XasX,] 'black bear', /sta'jak/ [sta-'jeX] 'brother's or sister's child', though a transcription with /a/ would be just as unambiguous here.

ni SPECIAL QUESTIONS

A. JUNCTURE

39. The distinction between syllabic and non-syllabic juncture between a cons, and a following /m, n, 1/ (see 8) is made more or less consistently only between syllables, i.e. in the case /CVC.RVC/ versus /CVCRVC/ (where VC may be replaced by syllabic /m, n, 1/, as in the examples /st'l.mu'Pt, t'lma'i?/ quoted in 8), and here /./ is written. In word-initial position there is perhaps a tendency to non-syllabic contact in combinations of the nominalizing prefix /s-/ and stems /RVC.../ when R is represented by /I/, or when the syllable is unstressed. For instance, one normally pronounces [sipa-1-, s^a-1-] in /s-ma'nit/ 'mountain', /s-nat/ 'night'; on the other hand I recorded [sla-1-] rather than [sja-'-] in /slaut'/ 'herring', and [sman-, smQ-] rather than [sipan-, sipp-] in /s-mnTma'nit/ 'mountains'. But this pattern is not rigidly adhered to; I recorded [sgox0-] in /s-n3xci'X/ 'canoe' (reduplicated plural /s-n3x0»nax°i'X/). A distinction between /CiV1-, CuY1-/ and /QV 1 -, CuV1-/ is not made at all, and I use the latter transcription only where C is represented by/s-/, e.g. in /si»'cm/ [siie'cip] 'news', cf. /ia'cm/ [ie'crp] 'tell, report' (an initial /s-/ so frequently represents the nominalizing prefix (80) that in all units /sC.../ where there is no morphological evidence to the contrary it may be regarded as such). When a vocalic prefix precedes a nominalized unit /s-RVC.../ the tendency to non-syllabic contact when R is represented by /m, n, 1/ is somewhat stronger, and here /./ is written, e.g. /ti-STmaq'rma'q'-m/ [tesmAq'-] 'overeat'. A distinction of juncture-types is also problematic in sequences /-V?R-/: in /p'i'Pnax0/ 'have seized' one hears syllabic variants o f / n / much more often than in the morphologically parallel case /k'VC.nax 0 / 'see', lit. 'have caught sight o f ' (cf. also, for an extreme case, the last variant of /q^u'PnauPa'sn/ quoted in 20, end). 40. In a number of words medial sequences o f / R / and /?/ were recorded which differ from those treated in 28 in that the glottal stop clearly falls after the complete articulation of the sonant. In some cases these may be accidental variants, e.g. in /hiuPi'qn/, recorded [heuPe-'qp] 'those sitting in front', where the root itself is

SPECIAL QUESTIONS

39

facultatively glottalized (/hiu(?)/ 'be upstream'). In others there is a morphemeborder between /R/ and /?/, and the second morpheme is found independently with /?/; in these cases I write /R.?/ or, where morphemes are separated, /R»?/, e.g. /?mr?i 'mac/ 'grandchildren' (reduplicated plural of /Pi'mac/),/tUPa'cTnaC/ [tJpA'cneC] 'east wind' (cf. /Pa'crnaC/ 'territory of Indian Reserve no. 3'). Such cases were sometimes also recorded where junctural glottalization (69, 70) is involved, but in most examples of this category one hears the usual variants of /R?/, so that at most the existence of a tendency can be registered; I do not mark these cases with /./. B. CONSONANTS

41. In several cases a sequence of two identical consonants regularly merges into a single one. For instance, the 3rd. pers. possessive suffix /-s/ is inaudible after stems ending in /s/: the form [taesPa- 'cos] means 'the face' as well as 'his face', and in the latter case I transcribe /ta w s?a'cus(-s)/, on the analogy of /ta^la'mP-s/ 'his house', etc. In other cases such a merger is frequent, but a pronunciation as geminates is possible. For instance, the relative-case prefix /t-/ often merges with a following article or demonstrative beginning with /t/, e.g. in /timPa1 t-ta'j?/ 'like that' one hears sometimes [tta*1?1] with two separate explosions, but also [ta-1?1]; in the latter case I write /(t-)ta'i?/ in recorded sentences. 42. After /n(P), 1(P)/ a distinction of /c/ and /s/ is problematic. I have recorded /Pans/ 'I', though comparative evidence (Cw. a'nGa) points to /c/, and /xcilPq°l?s/ 'steamboat', though a final /-c/ would allow a likely etymology 'fiery-mouthed' (see Dictionary), and so in other cases. A similar problem exists with regard to the distinction of non-initial /c/, /ts/ and /tc/. I have recorded /p'i'Ptcka/ [p'e- 'Pt'cke] 'grab me!', with a slightly affricated variant of /t/ followed by a full affricate (morphologically the word consists of the root /p'i?-/, the transitivizer /-t/, the 1 sg. object-suff. /-c/ and the imperative suff./-ka/). In /p'i'Ptcax0/ 'your /-ax°/ grabbing me /-c-/' the same sequence /-t-c-/ is morphophonemically present, but I doubt if before a vowel /tc/ is distinct from /c/. Words containing the reflexive suffix /-sut/ or /-cut/ were always recorded with [cot], and here the pronunciation was checked carefully: it is identical with that of /cut/ [co't] 'to say' where initial /c-/ is distinct from /ts-/ (see 5). But since trans, verb-stems which take this reflexive suffix have a final /-t-/ before all other personal suffixes (also before passivizing /-m/, see 82), and particularly since the reflexive form of verbs with the transitivizers /-at, -ut, -it/ has /-acut, -ucut, -icut/, these forms end morphophonemically in /-(V)t-cut/ or /-(V)t-sut/. There is nothing in Squamish itself that points to the former alternative, and for analytical purposes I write /-(V)t-sut/, for practical purposes (texts) /-(V)cut/. 43. A morpheme-initial /h/ is dropped after a consonantal prefix, e.g. /hiy/ [he-u] 'be upstream', /s-(h)iu/ [se'u] 'upstream region', /tx°-(h)iu/ [tx°e-y] 'move upstream'; /hii1/ [hiie-1] 'big', /nax0-(h)iiLqs/ [nux°iie- >qs] 'big-nosed' (suff. /-qs/ 'nose').

40

PHONOLOGY

44. In unstressed syllables a morpheme-initial /?/ is sometimes dropped after a consonant, e.g. /s-(?)3q>0i'?tl/ 'siblings and cousins' (also /?3q>0i'?tl/). I write /(?)/ only in cases where /?/ is usually or always absent. The /?/ in the 1st pers. sing, possessive prefix /?n-/ is always dropped after a clitic connected with the word of which it is a part: /Pn-sja1!?/ '(he is) my friend' versus /ta, .n-sja'j?/ 'my friend'; it is usually absent in all other sentence-medial positions. This case is very frequent, and instead of /(?)n-/ I write simply /n-/ here. The 2nd pers. possessive prefix /?©-/ is normally dropped in the 2nd pers. subject forms of the factual paradigm (see 133) after the article /k°i/, and here I write /(?s-)/; the use of the prefix may well be artificial. 45. In a few cases glottalic consonants seem to have lost their glottal feature when preceded by /?/. The root-final glottalic cons, in /siS'/ 'be all around', /SiC-a'n/ 'circle around; tr.' was recorded /£/ in /§i'-§i?2/ 'round', and /t/ was recorded in /mu L m?t-m/ 'blue grouse', but cf. Cw. mrt' 'id.'. In these cases /?/ is characteristic of reduplicative formations (see 71). Cf. also /Pas-iu'Pk0/ 'stingy' vs. Cw. ewk>0 'wealth, property'. On the other hand, an unexpected /k >0 / was recorded in /s-?9-k>0al?-k°la§/ 'wound', probably < */s-?a?-k°al?-k°la§/, and in any case a reduplicative derivative of /kVlaS/ 'shoot, sting'. To sum up, in two cases /?K/ seems to be a secondary development of /?K'/, in one / K 7 certainly represents an original /?K/, and for /?3siu?k°/ vs. Cw. ewk'° both interpretations are possible, though Sq. /n3X°-iuk°-a1?min/ 'what belongs to a stingy person', CdA. du?uk° 'be stingy', Kal. yayu'k°e? 'stingy' suggest the second one for Cw. ewk'°. — But cases of /?K7 were also recorded, e.g. /sc'i'uPq'/ 'elderberry'. It is not impossible that /?K/ and /?K'/ are in free variation in Squamish. 46. In a small number of words geminated or long consonants occur. Where these are found in medial position they are combined with a double stress if one of the adjacent vowels is /a/, and this /a/ is in several cases irregular in comparison to nongeminated counterparts of these words (see below). I write these geminates /C.C/, where /./ fulfills the double function of indicating that the consonants do not merge, and that the second one is not syllabic (in the cases with /n.n/). Such geminates occur without double stess in /s-?a't.tam/ 'things, stuff, belongings' (probably a derivative of /s-tam/ 'what?') and in /PaC.fca'uam/ 'to spawn', /(n-)s-?a5.£a'uam/ 'conger eel'. With double stress they occur in four of the numerals referring to days: /t , 3 l k >0 .k >0 usa l £/ '7 days', with a surprising /a/ in the first syllable, as compared to / t ' a k ' W C / '7', /ta'q.qa'5/ '8 days', cf. /tqa8/ '8', /Pu'p.pa'n/ '10 days', cf. /Pu'pn/ '10' and /ci'ja'Cis = ca'i.ja'iis/ [-e - 'ja - '] '5 days', /s-ci'ja'Cis/ 'Friday', cf. /ci'afcis/ 'five'; without double stress in the numeral /?a'n.n?us/ '2 animals' (possibly a misrecording for */?a'?n?us/, see 213). Further in two non-numerical words: /nax'-ua'u.ya'nm/ [-ua'u^ua -1 -] 'be surprised', /tina'n.na'Pt/ in the expression /\}a^n.X w tina , n.na l ?t/ [uAnXtene'nna'Pt] 'my deceased father', lit. 'the past one I am from', cf. /tina1?/ 'be from' (where /ti-/ means 'from' and /na?/ '(be) at').

41

SPECIAL QUESTIONS

In a few cases geminates occur word-initially, e.g. in the phrase /ua^,X»sw u a w n.na'?/ 'where they (persons) were'; here /n.na1?/ is a form of the morpheme /na?/ 'at' mentioned above. Since reference to persons in several cases entails reduplication in Squamish, gemination is considered to be a special type of reduplication (see 162). In the medial cases, the reduplication is the primary phenomenon, from which the double stress and the appearance of /s/ in unexpected places result. C. SVARABHAKH VOWELS

47. In a few words of the shape /CVR(?)C/ variants with a svarabhakti-vowel between /R/ and /C/ and without glottalization were recorded. The instances are the following: [CVR«C] [Xe'jpX]

[CVR?C] [Xei?X]

[tt'i'q1] [scV 'uSq'] [U'm^x 0 ]

[sc'e-uPq*] [Um?x°]

[CVRC] [XejX]

/Xi?X, XajX/ 'war'

[sc'e-uq'-] —

A'aiqV 'get trapped' /s-c>iu(?)q>/ 'elderberry' /tamPx0/ 'rain'

The variants with svarabhakti-vowel resemble the Halcomelem forms (Cowichan Xa'ylaX, ?'i 'w3q\ s-ta'max"); since in Squamish a type /CVRaC/ is very rare while /CVRPC, CVRC/ are common (see 65), I regard the latter forms as optimal, and in particular the form with glottalization, of which that without may be regarded as a reduction. In the word 'war' the form /Xaj[X/ is the most frequently used, and alternative forms are recognized; note that /Xi?X = XajPX/. Of A'ajq'/ no alternative with glottalization was recorded, but a glottal stop may easily have escaped notice before glottalic /q'/, see 45. With 'elderberry' a form CVRC was recorded only in the derivative /s-c'i'uq'-aj?/ 'elderberry-tree'. In all three cases with glottalization, the variant with [?] after the full sonant was recorded. 48. In word-initial position a sequence of /n/ plus uvular is pronounced with an anaptyctic vowel [a, ae] in between. All instances concern the prefix /n-/, alternant of /nax0-/ (locational), see 169ff. When a clitic ending in a vowel precedes, the anaptyctic vowel is facultative. Examples: /n-q'aX/ [naqVX, naeqVX] 'be in the way', /n-q°u'?us/ [n3q°o'?os] 'tears' (cf. before a non-uvular cons. /n-ti'q>0us/ [nte -l q >0 os] 'bump one's head'). Where such a vowel remains absent between /n-/ and a following uvular, the initial /n-/ is secondary and results from /?n-/; these cases all concern the 1st pers. sing, possessive prefix /?n-/, where /?/ is often dropped, see 44, and where this prefix in its form /n-/ occurs before uvulars I write /n»/ (where /./ indicates the absence of an anaptyctic vowel), e.g. /ta w nTq'a l i?q/ 'my chair'. An anaptyctic vowel is also heard between the combination /n-s-/ (consisting of the abovementioned prefix /n(ax°)-/ and the nominalizer /s-f) and a following consonant proper; the only instances are /n-s-t'q'i'us/ 'side of body' and /n-s-Xa'?Xm/ 'wild pigeon', pronounced with an initial [nsat'-, nsaeX-]. Such a vowel remains

42

PHONOLOGY

absent in /ns-qi'n-m/ 'rub oil in one's hair', where /ns-/ is a root (cf. /s-na's-qn/ 'hair-oil'), and also in all cases where the 1st pers. sing, possessive prefix /(?)n-/ is added to stems beginning in /sK/. Since the latter cases are very common, whereas the svarabhakti-vowel was recorded in two words only, I write only /nsK.../, giving a reference to pronunciation [nsaK] where necessary (technically, of course, the distinction should be expressed /nsK/ with schwa, and /ns.K/, resp. /(?)nsK/ without). 49. There are a few other cases where svarabhakti-vowels — mostly facultative — occur before uvulars, e.g. in /t(3)X°l?a'm?/ 'truly, completely', /Pi (?)&"/ 'all'. Not facultative is the schwa in /q>°aj?q>0ali?3X/ [-a- 'iPjeX] 'shy, wild' and in the suffix /-ai?3q% reduced /-iPaq0/ [-aipjoq0, -U?ioq°, -iPaq0] 'top of head'. I do not write /-aX, -aq°/ in these cases because the second one clearly represents a combination of the connective element /-aj-/ and the suffix /-q°/ 'head' (184) with so-called junctural glottalization (see 70); on the other hand, these combinations differ markedly from the final part of /Pa1!?*/ [Pa-'*??1*] 'crab' and from the medial part of /x°a'i?q°m?/ (recorded [-a- 'i?q°-]) 'wash one's head'. I think a particular combination of glottalizing and juncture-factors is responsible here (for types of glottalization see 67-73), and I write /a/ (see 75). D. EXCEPTIONAL VOWEL-TYPES

50. In a few words a long diphthongal [e'j] occurs, in phonemic transcription /ijj. In some of the instances morphological evidence shows that reduplications of forms with /i, i/ are involved (see the plural forms below, and cf. 146) and this is probably the case in the remaining instances too. The cases are the following: /si 'ia[?/ 'friends' (reduplicated plural of/s-ja'i?/ 'friend'), /si'i?ulx0a/ [se-'i?icr'x°a] 'old people' (red. pi. of /s-iu'Piux°a/ 'old person'), /si'jPam?/ 'chiefs' (red. pi. of /s-(h)i?a'm?/ 'chief'); the remaining instances are /ti'jPuiX/ [te- 'j?ueX] 'be protected' (cf. /ti'uPi^-n/ [te-'wPueXg] 'protect; tr.') and /Pi'ici/ 'those' (alternative form of /Pi'ci/). Here also belongs the case /ci'ia'Cis/ already discussed in 46. 51. In three sound-imitative words stressed short vowels were recorded which deviate from the normal variant-pattern of /a/. They are /k°3'm-i?n/ [k°ulme?u] 'make a thumping noise' and its reduplication /k°3f-k°m2m/ [k°u'k0-], and /q0a'-q0aX,q-m/ [q0u'q0-], a reduplication of /q°3'X,q-m/ 'rap, knock', itself recorded with a normal variant of /a/. No special notational device is adopted for these words; in the Dictionary their pronunciation is indicated by [u1]. E. CLITICS

52. Clitics occur very frequently in texts, and in the large majority of cases they form a stress-unit with the word that follows them. Often strings of three or four of them

SPECIAL QUESTIONS

43

are found together. In principle, the rules for unstressed vowels (35-38) apply in the same way to clitics, except that the tendency to reduction is even stronger here than elsewhere, and sometimes groups of clitics tend to merge into simplified forms. For reasons of expediency, in monosyllabic clitics with schwa and which consist of a combination of a morpheme /C-/ with a suffix /-C/ the symbol /a/ is not written. The cases in question are the combinations of the auxiliary /5-/ with the 2 sing, and 1 plur. subject-suffixes /-x°, -t/: /£-x°, 6-t/; the clitic /s-s/ (see 135), and the combination of the relative clitic fk/ with the 3rd pers. possessive suffix /-s/: A-s/. These units are pronounced [Cax0, 5at, sas, tas], with the appropriate, but strongly reduced variants of schwa. In the single case where a clitic CC is pronounced without possible schwa, the symbol /./ is used, viz. in the clitic /n.X,/ (morphologically /n*Xf) meaning approximately 'which I, where I, why I . . . ' ; the /n-/ here is the 1st pers. sing, possessive prefix already discussed in 44, 48. The symbol /./ is also used in the comparatively rare cases where a clitic /CC/ (see above) is extended with a suffix /-C/, always without intervening schwa, e.g. /5x°.57 (morpholog. /5-x°»6')/ [53x°S']. 53. After a clitic ending in (or consisting of) a consonant a stem-initial /m, n, 1/ is not syllabic, e.g. / n a w k 0 w n u ' q 7 [nok°no-lq°] 'it is noontime now'. The same rule holds between clitics, e.g. the combination / s s ^ m n w / 'then he ...' is pronounced [sasmg]. A clitic consisting of a single consonant does not normally merge with an identical initial stem-consonant, e.g. /na w k°_.k°3 l mi?n/ [nok°k°ulme?5t] 'they made a thumping noise' (I, 14). In the case of longer clitics such coalescing is common, and between clitics it is the rule, e.g. the combination / m n w n a m w / is usually pronounced [mpaem]. 54. The initial /?/ of a clitic may be dropped after a consonant proper, but it is facultatively preserved, and this distinguishes clitics from suffixes (see 69). In particular, this characteristic distinguishes the personal clitics 1 sg. /Pan/, 1 pi. /?at/, 2 sg. /?ax% 2 pi. /?a(ia)p/, 3 sg. (pi.) /Pas(-uit)/ from the otherwise identical suffixes /-an, -at/, etc. 55. The clitic /s-s/ (nominalizer /s-/ combined with 3rd pers. possessive sufF. /-s/) when preceded by the article /k°i/ [k°e] is pronounced /sis/ [ses], an instance of progressive vowel assimilation. Where morphemes are separated I write /k°i w s-i-s/ in order to avoid introducing an allomorph and having to assign /i/ arbitrarily to either affix. A similar assimilation occurs in the combination /k°i w s-ua-s w / [k°esyaes], also /k°i w s-yi-s w / [k°esues], and in /k°i w s-ni-s w / [k°esnes] (see 135); these combinations contain the same morphological elements as /k°i w s-i-s_/ except that the zero-element between /s-/ and /-s/ is replaced by the clitics /ua/ 'continuously', resp. /na/ 'there'. Besides progressive vowel-assimilation, analogy to /k°i w s-i-s w / may therefore play a role here.

44

PHONOLOGY

56. Here follows an account of the reductions and mergers which occur in strings of clitics. A few morphophonemic rules to which clitics are subject are given first. (1) The clitics /ua/ 'continuous' and /na/ 'there' have the forms /u, n/ when immediately followed by one of the personal clitics (summed up in 54): / u _ ? a n , n w ? a t / , etc. See also (6) below. For the alternations /ua, -ui-/ and /na, -ni-/ see 55. (2) The clitic /q/ 'if' has the form /qa/ when immediately followed by the relative clitic /X/ or by the 2nd pers. possessive prefix /?a-/: / q a w A , w / , /qa^,?3-.../. (3) The clitic /?u/ [?o] 'if, whether' in combination with a preceding clitic /na/ or /ua/ results in a unique vowel [0-], the combinations being pronounced [n0-, U0-], the former sometimes faintly sounding like [naua]. I write these combinations / n _ 3 u , u w 9 u / . The combination / n _ a u / can be used in isolation with an emphatic stress (meaning approximately 'did he really?') and is then pronounced [ n 0 - 1 ] . (4) The complex clitic /£-x°/, consisting of the clitic /5-/ 'be, do' and the 2nd pers. subject-suffix /-x°/ usually changes to [5ae] when immediately followed by the clitic /q/ 'if': /2(-x°)^_,q/ [Caeq] (perhaps only artificially [53x°q]). This reduction is reminiscent of that of the prefix /nax°-/ to /n-/ in certain complexes, see 48 and 169ff. (5) The clitic-combination / m n w u a w / tends to merge into [mae], e.g. / n w 3 u ^ 5 - x ° w m n w u a ^ _ h a l ? X / 'do you feel good?', in distinct pronunciation [n0'C3x°mijU£eha tends to be pronounced [n0£x°maeha'?A,] in more careless speech. (6) The clitic-combinations / q w u ^ , ? a n w , q w u ^ , ? a x ° w / , etc. (the third clitic one of the personal elements, before which /ua/ has the alternant /u/ acc. to (1) above) tend to merge into either [q>0£en, q>0ox°] or [q°aen, q°ox°] (glottalization facultative, see 54). This combination is common in negative sentences, where it is preceded by the predicate-word /ha'u/ 'it is not the case'. In these sentences the final /u/ of /hau/ tends to be lost, or rather, it merges with the labialization in [q'°, q°], e.g. /ha'u q ^ u ^ P a n ^ P i ' X n / [ha -l u qu(?)aen?e-'X,ij, haqC)°aen?e-'X,n] 'I am not going to eat'. (7) The clitic-combinations /q^_,?an^, q w ? a x ° w / , etc., tend to merge into [q'aen, q'ox 0 ], etc. when the glottalization is preserved (otherwise [qaen, qox°], etc.). Before these combinations, too, /hau/ tends to lose its final /u/, possibly on the analogy of the cases mentioned under (6), e.g. /ha'u q ^ P a n ^ P a ' n u X / [ha - '(u) qC)®n?a-'noX] 'I don't agree'. The reductions (5-7), being facultative, are not reflected in phonemic notation, whereas the alternations (1-3) are. In case (4) I write /S(-x°) w q/.

IV STRUCTURE O F MORPHEMES

A. MORPHEMES CVC

57. The favorite type of root-morpheme (i.e. morpheme other than clitic, affix or reduplication-syllable) has the form Q V Q , where either C x or C 2 may also be represented by /?/, and C 2 also by /R?/. Henceforth, when "morphemes CVC" are discussed, the types /?VC, CV?, CVR?, ?VR?/ will be understood to be included. When V in CVC is represented by A ( = full vowel), there is often a corresponding form with /a/, e.g. /sa'q'-an/ 'split; tr.' (transitive suff. /-an/) and /saq'/ 'split; itr.'; /mui/ 'submerge' and /maj/ 'sink'. Under certain phonological conditions a morpheme /CaC/ alternates with /CC/, cf. /saq'/ just quoted and /sq'-i'm?/ 'split; act.-itr.'. Where forms /CAC/ and /C(a)C/ occur side by side, we speak of full and /a/- or zero-grade. From the morpheme-type CVC are excluded: (a) Cases with identical C x and C 2 . In view of the very extensive use of reduplication in Squamish, such formations are considered reduplicative. They are summed up in section 161. (b) The types */Cah(?), Cuu(?), Cij(?)/; in other words, no morpheme has a full vowel followed by the corresponding sonant. (c) The type */Ca?/. (d) The type */?3R?/. Subdividing the type C a VC 2 into classes according to the nature of C x and C 2 in terms o f / K , ?, R, R?/ (the latter for C 2 only) and of V in terms of/A, a/, 3 X 2 X 4 = 24 possible sub-types are obtained, of which 5 are excluded (to wit */?A.?, ?a?, Ka?, Ra?, ?aR?/). The occurrence of /?/ after a sonant is often due to morphological factors (see 69-72); furthermore, the types /CVR/ and /CVR?/ are sometimes in free variation even in the most careful pronunciation. In giving examples, I quote only such cases as are regularly pronounced without, resp. with /?/ when used in isolation or in formations where /?/ is minimally suspect of being due to morphology. The types /Ca?, Cu?, Ci?/ are etymologically equivalent to /Cah?, Cay?, Cai?/, and these six are treated

46

PHONOLOGY

separately in 61-64, so that from the examples given below the types /KA?, RA?/ are excluded as a whole, while of /KaR?, RaR?/ only examples with /m?, n?, 1?/ are quoted. Of the very numerous types /KAK, KaK/ three examples each are quoted; of the types ending in /R(?)/ examples with each of the sonants /m, n, 1, h, u, jj (in this order) are given (in so far as they are available). When possible, free forms CVC are quoted; next in line are forms with the nominalizing prefix /s-/ or with trans, and intrans. suffixes. Where no such cases are available, the root is quoted as such. /KAK/: /taq°/ 'drink', Auc'/ 'be scabby', /fiiX/ 'top'; /KAR/: /s-tam/ 'what?', /k J °in/ 'how many?', /5'i'h-n/ 'lift; tr.\ /kau/ 'descend', /tui/ 'go across'; /KAR?: /q >0 um?/ 'kelp', /qan?/ 'steal', /X°il?/ 'come off', /p'i?-t = p'ih?-t/ 'seize; tr.' (see 62), /Say?/ 'bone', /k >0 aj?/ 'be hungry'; /RAK/: /maqV 'thick, coarse', /nuq°/ 'be noontime', /lix°/ 'fall down', /hiq/ 'be under', /uaS/ 'be removed from the fire', /¿iq/ 'to snow'; /RAR/: /lam/ 'alcoholic beverage', /man/ 'father', /hi 'l-it/ 'roll; tr.', /juh/ 'be careful', /hay/ 'be not the case', /mui/ 'submerge'; I RAR?/: /lam?/ 'house', /hin?/ 'be a long time', V n i\J? 'be inserted'; /?AK/: V?ac 'surface, front', /?u'X°-un?/ 'cut, carve; tr.', /?iq'°/ 'be rubbed'; /?AR/: /?an/ 'very', \/?il (meaning uncertain) in /s-?il-tx°/ 'roof, Indian house', etc., V?ai 'be inside'; /?AR?/: /?im?/ 'little boy', /?al?-s/ 'pity; tr.'; /KaK/: /qap>/ 'close', /pac/ 'bend, fold', /tax 0 / 'be settled'; /KaR/: /X°am/ 'swift (ab. running water)', /k°an/ 'start', /q>0al/ 'be cooked, ripe', /c'ah/ 'be hit', /q'a'u-m/ 'howl, whistle; itr.', /qaj/ 'bad'; /KaR?/ with /m?, n?, 1?/: /5'am?/ 'bite', /s-tan?/ 'woven material', /Xal?/ 'write'; /RaK/: /maq'/ 'be full (from eating)', /naik°-n/ 'move; tr.', /las/ 'bottom', V ^ 'good', /uaX,/ 'wild', /s-jiac/ 'realistic story'; /RaR/: /iam/ 'cranky', /yal/ 'unidentified plant' (prob. 'tule'), /nay/ 'thou', /maj/ 'sink'; /RaR?/ with /m?, n?, 1?/: Vhsm? 'be covered', /man?/ 'child, offspring'; /?aK/: /?aX/ 'wild goose', /?aq°/ 'fall out (ab. hair)'; /?aR/: V?®11 (meaning uncertain) in /?aln-yiX/ 'be in the center' and derivatives; historically in /s-?a'li/ 'dream', etc.; V? 3 i 'good'. The Type /CVh/ 58. In 8 the long vowels [a:, o:, e:] were interpreted phonemically as /ah, uh, ih/. This interpretation is based on the following considerations. Given a root ending in [a:], we find an alternant with [aeh] before a stressed vowel: /q°ah/ [q°a:] 'be pierced', /q°ah-a'ia?n-i?n/ [q°oh-, q°aeh-] 'pierce someone's ears /-aia?n/; tr.'; /?ah/ [?a:] 'hurt', /?ah-a'num/ [?ah-, ?aeh-] 'be painfully hungry'; /c'ah/ [c'a:] 'be hit', /c'ah-i'm?/ [c'seh-] 'throw punches; act.-itr.'. The interpretation of these forms as

STRUCTURE OF MORPHEMES

47

roots ending in /h/ is phonetically and structurally obvious. Since the vowel [a, o, a ] before /h/ is not opposed to any other, the interpretations /a/ and /a/ are both possible. Now, act.-itr. derivatives in /-im?/ of roots KVK either have a full vowel and rootstress, or zero-grade and stress on the suffix, e.g. /p'a'c'-im?/ 'sew' vs. /t'k'M'm?/ 'dig' (full grade of root in /t'a'k^-an/ 'id.; tr.'). The stress in /c'ah-i'm?/ points to a root with zero-grade; in fact, the transcription */c>hilm?/ would do as well, but since the choice between the root-shapes /CaC/ and /CC/ is largely automatic, and the use of the symbol /a/ is in principle a concession to phonetics (see 74, 75), we select the transcription /Cah-/. As far as the non-suffixed forms are concerned, these could still be interpreted as the corresponding full grades */Cah/, but this would be a needless complication, and the evidence points to zero-grade. In the first place, where the root of an act.-itr. derivative /CC-im?/ occurs without affixes, it has the form /CaC/, cf. /t'q^-im?/ 'chop' and /t'aq' 0 / 'break, be bruised', /cx°-im?/ 'throw' and /c9x°/ 'get hit', /sq'-im?/ 'split' and /saq'/ 'be split'. In the second place, the reduplication /?as-q0a'-q°a?/ 'perforated', with /a/ in the reduplication-syllable, points to a root /q°oh/ rather than */q°ah/ (cf. /?as-ta LtX0/ 'correct' besides /taX°/ 'be corrected' versus /?as-k°a'-k0ai?/ 'hidden' besides /k°ai/ 'to hide'; see 154-156). Structurally, the interpretation /Cah/ allows the simple rule given in 57 under (b). 59. Of the three roots with final [o:, e:], two have zero-grade alternants with /ah-/, which safeguards the interpretation /uh, ih/, cf. /puh-t/ 'blow' and /s-pah-i'm?/ 'wind' (a nominalization of the act.-itr. form, as in /s-tV-i'm?/ 'pit, dugout' besides / t V - i ' m ? / 'dig; act.-itr.', full grade in /t'a'k^-an/ 'dig; tr.'), /5'i'h-n/ 'lift, raise' and /d'ah-i'm?/ Hd.; act.-itr.'. The third case, /juh/ 'be careful' appears with zero-grade in /ia-n?/ 'to warn', with the usual shortening of /ah/ to /a/ when non-vocalic suffixes are added (cf. /q°ah/ 'be pierced' and /q°a-n/ 'pierce; tr.'). In this way, the alternation of /u, i/ with /a/ is indicative of an original root ending in /h/ even in cases where the vowels have been shortened. For instance, the word /¿u'jaPk0/ 'wave, current, rapids' is recognizable as an instance of the reduplication-type /CAC-CaC/ (see 144) if an original form */iuh-iah-k°/ is posited. 60. Roots /CAh/ are seldom preserved in their original form: they are usually shortened to /CA/, e.g. /5'ih-/ appears as /5'i-/ in /S'i'-Sit/ 'lift up for someone', /5'i'-q-am?/ 'rise from one's seat', lit. 'raise one's behind; itr.'. In the case of morphemes which are encountered only in the shapes /Cu-, Ci-/ it is impossible to say — in the absence of other data — whether they represent reductions of /Cuh, Cih/ or of /Cay, Caj/. Such is the case, for instance, with /cu-/ in /cu-t/ 'say', /cu-n/ 'tell, order; tr.', /cu-cin/ 'mouth', and with V^'i- 'near' which occurs only with various petrified extensions (but cf. CdA. 5'ih 'approach, get near'). The Type /CVR(?)/ with /h, y, ]J

61. The morpheme-type CVR with /h, u, j/ as R falls into 9 groups according to the nature of V and R as follows:

48

PHONOLOGY

V V \

i

h 1. Cau

a

2. Cai

u

3. Cuh

4. Cui

i

5. Cih

6. Ciy

—

a

7. Cah

8. Cau

9. Cai

These types will now be exemplified together with their glottalized counterparts /CVR?/, except for the types (3) and (5), the instances of which were already quoted in 57 (for glottalic counterparts see next section). In anticipation of a separate discussion of the occurrence of /?/ (see 67-73) it is necessary to point out here that plain and glottalized forms can be distinguished clearly only for morphemes used in isolation or with prefixes. In many reduplicative and suffixed formations plain forms become glottalized, and for morphemes whose occurrence is limited to such formations the "independent" shape in terms of/CVR/ or /CVR?/ cannot be established. (1) /Cau(?)/ Plain: /kau/ 'descend', /q'a'u-at/ 'pay someone; tr.', /hau/ 'be not the case'. Glottalized: /Sau?/ 'bone', /Xau?/ 'recover'. (2) /Cai(?)/ Plain: /taj/ 'canoe-race', /q'aj/ 'be high up', /Vaj/ 'be as previously'. Glottalized: /tai?/ 'that one', /k>0ai?/ 'be hungry', /tfaj?/ 'be sheltered'. (4) /Cui(?)/ Plain: /mujj 'submerge', /tui/ 'go across', /q >0 uj/ 'die', /huj/ 'be finished'. Glottalized: no independent examples. In /huj?s/, causative derivative of /hui/, glottalization occurs, whereas otherwise this formation lacks it, cf. /kau-s/ 'bring down', /hau-s/ 'refuse (see examples under 1.). The morphemic status of final /t/ in /t'ui?t/ 'medicine' is uncertain; the element /t'ui(?)-/ corresponds to Cw. -t'al-. (6) /Ciu(?)/ Plain: no independent examples. Glottalized: /c'iu?/ 'healing up', A'iu?/ 'run away (ab. a captive)', -v/niy? in /nax^ni'uP-it/ 'instruct'. Free variation in /hiu(?)/ 'be upstream'. (7) /Cah(?)/ Plain: /c'ah/ 'be hit', /q°ah/ 'be pierced', /?ah/ 'hurt'. Glottalized: /tah?/ 'undergo', 'be located', /A,ah?/ 'be touched'. (8) /Cau(?)/ Plain: /nau/ 'thou'. Glottalized: /s-Xau?/ 'dried salmon backbone', /na'u?-n/ 'put in; tr.', ' n /Pas-ia'u?/ 'seer'. (9) /Coi(?)/ Plain: /mai/ 'sink', /qsj/ 'bad', /pa'i-m/ 'fall overboard'. Glottalized: /X'i? = X,5ai?/ 'dear', / « ? = S'aj?/ 'to fast', /x°i? = x°ai?/ 'be lost'. 52. As it was pointed out in 31-34, morphemes /Cah?, Cau?/ can have alternants /Ca?, Cu?/, while /Cai?/ and /Ci?/ are not distinguished at all. The distribution of the alternants is partly correlated to word-classes, cf. /tah?/ 'be located', /na'u?-n/ 'put in; tr.' and the relator-verbs /tx°-ta?/ '(move) onto', /tx0Tnu?/ '(move) into', partly it is automatic, namely in the case of /Cah?/, which appers as /Ca?-/ when followed by a suffix. But besides the type /Cau?, Cu?/ there are morphemes which appear only in the form /Cu?/, cf. /c'u?/ 'come out (being pulled)' (not a relator-verb!)

STRUCTURE OF MORPHEMES

49

and the trans, derivatives /c'u'P-n/ 'pull out', /mu'?-n/ 'drop', which behave differently from /na'uP-n/. Besides the type /Caj[? = Ci?/ there is the type /Ci?/ (see 22), represented by the roots of /p'i?-t/ 'seize; tr.' and /si?-n/ 'wipe; tr.'. Structurally, the hypothesis seems justified that the type /Cu?/ not alternating with /Cau?/, and the type /Ci?/ represent the "missing" glottalized types /Cuh?/ and /Cih?/. This interpretation can be corroborated only in the case of /p'i?-/ which has an alternant /p'a?-/ in a number of derivatives, e.g., in act.-itr. /(Pi-jp'aP-i'm?/ 'hold', where the stress points to zero-grade of the root, so that /p'a?-/ may well represent a medial reduction of "/p'ah?/. 63. The V p ' ^ occurs in a further stage of reduction in /Pi'-p'i-s/ 'take, hold; tr.' (causative suff. /-s/) and in /p'ic7 'get trapped, squeezed'. These words have no direct synchronic connection with Vp'i? (except for a possible "symbolic" value of /p'-/), but they show that /Ci?/, like /Cih/, can be reduced to /Ci-/. I have no corresponding examples involving /Cu?/ < */Cuh?/. 64. Glottalized forms easily lose the the feature /?/ in rapid speech, and where forms were recorded with and without /?/ I consider that with /?/ as the "optimal" one, except in cases where even in careful pronunciation both forms are used, as in /hiu(?)/ 'be upstream'. However, certain unexplained cases remain. The root which otherwise occurs as /p'i?-, p'a?/ was recorded as /p'ah-/ in /n-p'ah-a '6/ 'palm of hand' (but cf., with the same suffix /-a£/ 'hand', /p'a?-a'6-i?n/ 'grab someone by the hand; tr.'). The root /tah?/ 'undergo' has the act.-itr. derivative /tah-i'm?/ or /taPh-i'm?/ 'make'. The cases /np'aha'C, tahi'm?/ may be instances of loss of glottalization in careless speech; the form /taPhi'm?/ may be a contamination of /tah-/ and /ta?-/ (cf. the parallel case /(?i-)pJa?-i'm?/ 'hold'; a form */ta?i'm?/ was not recorded). But /taPhi'm?/ may also be regarded as a parallel to the variant [ei?jV] of /i?V/ (see 21), in which case a better transcription would be /taPi'm?/ or /tahPi'm?/, depending on whether or not /(Pi)p'aPi'm?/ allows a similar variant. The root meaning 'name' appears in the usual forms /nah, na-/ in /s-nah/ 'name', /na-n/ 'call, name; tr.' but was also recorded glottalized in /s-na?/ 'name' (a variant */nah?/ was not recorded). B. LONGER MORPHEMES

65. The most frequent type of root-morpheme larger than the type QVC, is the type QVCjCa, where Q and C a stand for the same sounds as in QVCa (see 57), and where C 3 cannot be /h, u, \ j and is rarely represented by /n, 1/ (as for a final /-m/, this is so often the itr. suffix /-m/ that all units /CxVCgm/ are suspect of being bimorphemic). The most common representative of QVCjCj is therefore Q V Q K , here subdivided into /CVKK, CVRK, CVRPK, CVPK/. Examples: /CVKK/: /nai5/ 'hand', /k^uX'k 0 / 'salt water', /5iSk°/ 'ebb', /c'aXt/ 'gravel beach'.

50

PHONOLOGY

/CVRK/: /t'amk'7 'salmon eggs', /c'uis/ 'crazy', /sink1/ 'grandparent', /tamty 'red paint', /Pans/ 'I' (only ex. with /-an-/), /palq0/ 'kneel', /s-p'ahc'/ 'blackberries', /iau^V 'break'. /CVRPK/: /?ai?X/ 'crab', /t'ujPt/ 'medicine', /s-c'iuPq'/ 'elderberry', /s-tomPx"/ 'rain', /s-CanPq/ 'Gibson's Landing' (geogr.). /CVPK/: /sa?X/ 'parent- or child-in-law', /k°u?s/ 'spring-salmon', /ni?5/ 'high seas'. Less common is the type CjVCaR with /(m), n, 1/, e.g. /t'a'lm/ 'wild cherry' (bimorphemic?), /s-q'a'ml/ 'paddle', /Pi'Xn/ 'eat', /si'PI/ 'grandparent'. The type /C1VCgA(?)/ may be considered to represent the "missing" cases with final /h, i, u/; it is much more common with /-A?/ than with /-A/: /ma'qa?/ 'snow', /Xa'Su?/ 'lake', /p'a'li?/ 'bark of tree', /s-fii'Pi/ 'strawberry'. On the other hand, final /-A/ is the norm in /CVPRA/, e.g. /Xa'Plu/ 'spoon made of horn', /Sa'Pju/ 'corpse', 'screech-owl', 'co-wife', /si'Pla/ 'granny' (cf. /si'PI/ 'grandparent'). Morphemes beginning with a group of two consonants are fairly common, and besides many of the types treated so far examples can be found with KK- instead of Clt e.g. KKVK: /t'q'aX/ 'fall backward', /s-£'q>0ap/ 'feather in hair'; KKVR: /c'&l/ 'kingfisher', /Xcam/ 'cedar box', /q>0X°ui-/ 'nail'; KKVKK: / c \ ° a U / 'marshy land'; KKVRK /tails/ 'stand up', KKVRPK: /qVa'yPty 'West Coast canoe', etc. Cases with initial KR-are rarer, e.g. KRVK in /tmi'x0/ 'land, earth', KRVRP in /qlu'm?/ 'eye', etc. Longer unanalyzable units follow no definite pattern except that the only group KKK encountered is that in /Xsq'iPs/ 'know; tr.', and this group is unstable and varies with /X,q'-, sXq'-, sq'-/. Many morphemes more complex than CVC can be shown to be petrified affixed forms or compounds. For instance, /naX5/ 'hand, arm' contains an element /-aX-/ which recurs in /ta 'Xn-tn/ '(upper) arm' and in the suffix /-ai2aX-a?n/ 'arm'; an identical element /-i?q0-/ occurs in /hi?q°i'n/ 'light, torch', /t^i'?q°iri/ 'spark', /x°i'?q°l?s/ 'steamboat' (cf. also /i3'q°16p/ 'put wood on the fire'). More generally speaking, the morpheme-types /CVRK, CVRPK/ are formally identical with combinations /CVR(P)/ + /-K/, morphemes /KKVK/ with zero-alternants /KK/ + /-VK/ or with prefixed combinations /K-/ + /KVK/, etc.

C. AFFIXES

66. The few prefixes that occur in Squamish are limited to the shapes C-, CC-, CV-, CVC-. The numerous suffixes show a large variety of forms. The most common types are -C, -CC, -CVC, -VC, -VCVC, -VCCV. Already here, some of the longer affixes are clearly original complexes, cf. /nax0-/ (location) and /tx°-/ (direction); the type -VCCV is represented by /-aX,Xa/ 'neck, throat' and /-aX.q°u/ 'water' with a probably identical formative /-aX-/ (cf. for the last example •s/q°u 'water' in /n-q°u'*us/ •tear', with suff. /-us/ 'face').

y QUESTIONS OF WORD-PHONOLOGY

A. THE OCCURRENCE OF /?/

67. Glottalization is easily dropped in Squamish, especially in combinations /R?/, which are almost always in free — or at least stylistic — variation with /R/ (the latter variant is characteristic of rapid speech, but by no means limited to it). On the other hand, there are many cases of /R/ not in free variation with /R?/, so that the two must be distinguished. I write /R?/ always in all words where this pronunciation was used regularly in careful speech; in the Dictionary alternative forms are given when glottalized and plain forms were both common in careful speech, and where such words occur in sentences I write /R/ or /R?/ as recorded in each particular case. My material undoubtedly contains cases recorded only with /R/ but where /R?/ is possible, and therefore in discussing the occurrence of /?/ I shall in several cases have to register tendencies where a more detailed investigation may establish rules. One generalization that can be made about the occurrence of /R?/ is that it is very much more common in stressed syllables than in unstressed ones, and in the part of the word following the stress than in that preceding it. In the occurrences of /?/ two main groups of cases must be distinguished. In the first of these, /?/ is a part of a morpheme as such; these cases concern morphemes /?VC, CV?, CVR?/ and longer types such as /PVCC, CV?C, CVRPC/, etc. In the second, the occurrence of /?/ is characteristic of certain morphological complexes as such, and not of the morphemes entering into such complexes. We shall speak of morphemic and morphological glottalization respectively. An example of morphemic glottalization is /nam?/ 'come', of morphological glottalization /£i£a 'j?aqm/ 'follow continuously; itr.\ cf. the non-reduplicated /ia'jaqm/ 'follow; itr.' without glottalization (and many similar pairs /CV-CVR?.../ vs. /CVR.../), cf. also the trans, derivative /5a'i-nI 'follow; tr.' as opposed to /na'mP-n/ 'go and get someone; tr.'. When morphemes are separated, I write /?-/ if the morpheme preceding /-/ was recorded with/?/ in isolation, cf. /nam?/ and /na'm?-n/ above. Otherwise, /?/ and /-/ are combined as /*/. The transcriptions /R?-, V?-/ and /R2, V2/ do not imply a phonemic difference. Nor is the distinction correlated to morphemic and morpho-

52

PHONOLOGY

logical glottalization: though all cases written /?-/ are morphemic, not all cases of/ 2 / are necessarily morphological. 68. Morphemic Glottalization.— The cases of morphemic glottalization can be subdivided into those where /?/ takes the position of Cx and those where it does not; the former type is referred to as "preglottalization". In the position of C 1; /?/ is reduplicated like any other consonant, e.g. /Pu'miC/ 'go upstream', /?u'-?um?i£/ 'be going upstream'. Preglottalizing /?-/ can be dropped after a consonantal prefix, e.g. /s-(?)ac'a'm?3s/ 'women's shoulder blanket', but it is not dropped as regularly as /h-/ in the same position (see 44). A morpheme-final /?/ can be dropped in certain morphological formations, e.g. in the unstressed syllable of a plural reduplication: /lam?, lmrla'm?/ 'house(s)\ /man?, ma'nPrmn/ 'child(ren)'. As was mentioned in 57, root-morphemes */Ca?/ do not occur; to this must be added here that the sequence /a?/ is very rare in general: it was recorded in /sX°laua1?/ 'turnip', /q0ux0q°ux°jpl?qm/ (besides /-ja'Pqm/) 'slide down', in the reduplications /sa'PsqV 'chips of wood' (cf. /saq7 'split'), /qa'Pqi?/ 'soft' and /ua'Pu/ 'continue' (cf. the clitic /ua/ 'continuously'), and unstressed in a few cases where /a/ probably is a svarabhakti vowel, e.g. /taPu'sm/ 'look up', which could be unambiguously written /tPu'sm/, with /t?/ [to?] different from /t'/. 69. Junctural Glottalization.— Glottalization may occur at the border between root and suffix, or between suffixes, when the preceding one of these elements ends in a vowel or sonant; in these cases we speak of junctural glottalization. There are two groups of cases, depending on whether the following element begins in a vowel or a consonant. In the case of vocalic suffixes the phenomenon of junctural glottalization is simply explained by the fact that such a suffix is potentially preglottalized. This preglottalization is maintained after a vowel and dropped after a cons, proper and /?/; after a sonant it is usually absent, but it tends to be maintained in a particular group of cases (see below). For instance, the combination of the suffix /-ulX/ 'young specimen' with /sui'Pqa/ 'man', /puS/ 'cat', /sXa'naj?/ 'woman', /staqi'u/ 'horse' results in /sui?qa 2 u'R/ 'boy', /pu§-u'lty 'kitten', /sXniP-u'lX./ 'girl' (see 36), /staqiu-u'lty 'colt'. Other examples with stem-final sonants: /i°a ! j-a5/ 'have a paralyzed arm' (/-a6/ 'hand, arm'), /t'lm-a'i?/ 'wild cherry tree' (/-ai?/ 'bush, tree'), /?upn-u'i?s/ 'ten large pieces' (/-ui?s/ 'large piece'), /c'ah-u's/ 'be hit in the face' (/-us/ 'face'). The cases where glottalization is maintained most often after sonants are the following: (a) With the pronominal stems /tam/ 'what?', /Pin-/ '(the) one, (the) other', e.g. /s-tam2a'i?/ 'what kind of tree?', /ti w n-s-?i'n 2 a£/ 'my /n-/ one (other) hand' (/s-/ nominalizer; /ti/ article). (b) After connective /-aj[-/ (see 184), e.g. in /-aj2us/ 'eye', /-aj2a?n/ 'ear'. (c) In a number of cases which make the impression of being recent formations,

QUESTIONS OF WORD-PHONOLOGY

53

e.g. /suk>0-um2a'u?tx°/ 'bath-house', /c'aX-ti^a'n?/ 'to poison; tr.'; here the suffixes /-au?tx°/ 'house', /-an?/ 'transitive' are added to stems already containing suffixes, namely the intransitivizer /-um/ and the implement-suffix /-tn/ (/§u'k>0-um/ 'bathe; itr.', /c'a'X-tn/ 'poison'), and such combinations are unusual. In response to a question a propos of /hi-i'k°up/ 'big fire', IJ produced the artificial combination /?3ci'm2ik°up/ 'small fire' (/Pad'm/ 'small'). (d) After roots /CAh/, with loss of the final sonant, cf. /iuh/ 'be careful', /ju^an/ 'my hypothetical being careful', /iu l2 as/ 'his id.'; since such roots are often reduced to /CA/ (60) this is but a special instance of preglottalization maintained after a vowel (see above). 70. With consonantal suffixes junctural glottalization occurs most often in the case of stems ending in a sonant other than /h/ followed by a suffix consisting of a single consonant. Examples: /x°i2q° = x°3\lq°l 'be uncovered, exposed' (cf. /x°aj/ 'appear, become visible'), /s-q°in*c/ 'beard' (/-c/ 'mouth'), /s-q°in26/ 'hair on the back' (cf. /s-q°i'n-a5/ 'hair on the arm'), on the other hand /X°a'i-sn/ 'have a paralyzed leg' (cf. /Xaj/ 'be senseless'). After a vowel: /hiil2c/ 'big mouth' (cf. /hii1/ 'big'), on the other hand /n3x°-(h)iil-qs/ 'big-nosed'; however, glottalization is lacking in the petrified combination /hii'-q 0 / 'headdress used by medicine-man'. It is possible that this type of glottalization is due to the loss of /a/ between stem and suffix, cf. the variation of /CVRaK/ and /CVR?K/ discussed in 47. In a few instances junctural glottalization occurs when the plural suffix /-uit/ is added to stems ending in a vowel. Though these combinations are pronounced /Vu?it/, I transcribe /V2uit/, e.g. in /kVci^uit/ [k06'ceu?uet] 'those (absent)', cf. /k°a'ci/ 'that one (absent)'. Compare for these cases section 29. 71. Reduplicative Glottalization.— Roots containing at least one sonant can become glottalized in certain types of reduplicative formations. The most evident cases of reduplicative glottalization are those where a root-initial resonant becomes glottalized, e.g. /?3s»m3'-m?k°/ 'lumpy' as compared to /s-mak°/ 'lump', or where a final resonant becomes glottalized at the end of a word, e.g. /?3s-k°al-k°aj[?/ 'hidden' as compared to /k°ai/ 'to hide'. The first of these two types of cases is evident even if the root was not recorded in isolation (since roots do not begin with */R?-/); the second is not, because the glottalization may be morphemic, as it is, for instance, in the verifiable case /k>0a'-k>0ai?/ 'be very hungry', cf. /k>0aji?/ 'be hungry'. Again, as evident may be considered those cases where a longer unit (analyzable or not) is unglottalized as a simplex and glottalized as a reduplication, e.g. /hi'l-it/ 'roll; tr.', /hi'-hil2it/ 'roll continuously; tr.', /Pi'mas/ 'walk', /?i'-?im?a§/ 'be walking around'. Due to the proliferation of reduplicative forms in Squamish, many morphemes are found in such forms exclusively, and in these cases glottalization, if present, is always doubtfully morphemic, and sometimes it could also be junctural. On the other hand, absence of glottalization in a particular reduplicative form can always be counted as

54

PHONOLOGY

an instance of lack of glottalization in the type of formation concerned, regardless of whether or not the root was recorded in isolation. Therefore, if one counts the unambiguous cases, one counts all those without, and only part of those with reduplicative glottalization. The fact that probably a number of words recorded with /R/ allow alternative forms with /R?/ (see 67) further tends to discriminate against the number of cases where reduplicative glottalization occurs. It is possible that a full account of reduplicative glottalization will have to distinguish not only various formal types of reduplicative formations but also ad-hoc vs. petrified cases. However, formal distinctions alone can be carried through so far that the number of examples available for each category becomes too small to establish firm rules. In the survey that follows, only the large-scale formal distinctions of sections 141-162 are made. Even so, reduplicative glottalization is clearly characteristic of certain overall types as such, while for other types only a tendency to glottalization can be established. Glottalization occurs:1 (1) Regularly in the type / Q A - Q Q / (755) when Q is a sonant: /Pas^li'-l?x°/ 'lying down' (/lix°/ 'fall down'), /li-l?s/ 'be below' (/las/ 'bottom'), /li-l?q/ 'cheap'(*), /ua-u?S/ 'keep away from the fire' (/ua'S-an?/ 'take away from the fire; tr.'), /ma'-mPli/ (geogr.)(*), /mu'-mPtm/ 'blue grouse'(*), possibly also /s-ui'-uPlus/ 'youth'. No examples without glottalization. (2) Regularly in the type /Qa-QCVOCy (156) when Q is a sonant and C a is not: /n3x°-s»n9l-n?p/ 'follow a trail'(*), /ti-la'-lPs/ 'from below' (/las/ 'bottom'), /la'-IPC'/ 'bright yellow'(*), /Pasrina'-mPk0/ 'lumpy' (/s-mak0/ 'lump'), m/ 'be boiling; itr.' (/u31Xq°-m/ 'boil; itr.'), /iiPqi1 = ia-i?qi'/ 'be creeping' (/jaqi1/ 'creep'), /?as-iji?X° = -i9'-i?*7 'free, loose'(*), /haPX = ha-h?ty 'good' (/haX-i1/ 'become good'). No examples without glottalization. (3) Frequently in the same type /Qa-QOOCg/ when C a is a sonant (and in all cases where simple roots CVR are involved): /pa-p^a't-uX = pa-pajV 'be hunting in canoe' (/pi'atuX/), /ma'-mi?/ 'be sinking' (JmsU), /5a-£au2a't-ai?/ 'help each other' (/5'a'u-at/ 'help; tr.'), /sa'-SauPai/ 'be growing' (/Sa'uaj/), /5a-§m?a'n-mrnauPas/ 'be enemies' (/Sma'n/ 'enemy'), / k ^ ' - k ^ m / 'be thumping' (/k^'m-iPn/ 'thump'), /qa-qi^i1?/ 'become bad' (/qi-i1?/), /?as-q°3 !-q°a?/ < */-q°ah?/ 'pierced' (/q°3h/ 'be pierced'), /ja'-ilPX-t/ 'be looking for; tr.' (/¿alX-t/), in addition in unverifyable cases: /pa'-piPs-m/ 'bubble; itr.', /ta'-tiPX0/ 'be busy', etc., but since a complex root-type /CAPC/ exists, this glottalization may be junctural in origin (as in /s-Xam2xc/ 'rain', see Dictionary), or an original complex /CaRP-C/ may be involved, in which case the glottalization is morphemic. In a minority of cases no glottalization was recorded, e.g. in /XaX-5a'-5mX/ 'chew resin' (/Sami/ 'resin'), /qa-qli'm-i?/ 'become soft' (/qli'm-i?/). (4) Regularly in diminutive reduplications (see 159) when the root begins in a sonant: /li^litam/ 'little table' (/lata'm/, French la table), /li^lam?/ 'little house' (/lam?/), /s-nil*nix°ity 'little canoe' (/s-nsx°VXf). 1

Where no simplex was recorded this is indicated by (*).

QUESTIONS OF WORD-PHONOLOGY

55

(5) Frequently in the type /C1A1-C1A1Ca/ (see 154): /p'a-p'a'iPaq-n/ 'cure; tr.' (/p > a l iaq-n/),/k°ai-k°ai?l/'to dawn' (/s-k°ail/ 'day'), /?3s-k0al-k°aj?/ 'hidden' (/k°aj/ 'hide'), /hi'-hil2it/ 'roll; tr.' (/hi'l-it/), /?i'-?i m ? a g/ 'walk' (/Pi'maS/), /?u'-?um?at/ 'be lazy' (/Pu'mat/), /?u'-?um?i£/ 'move upstream' (/tx0-?u'mi5/ lid.'). No glottalization in a minority of the cases where Simplicia were also recorded, e.g. in /t'a'-t'ajaq'/ 'be angry', /Pa'-Pa^nax°/ 'be alive' (these may well be rapid-speech forms) and in a few petrified cases, e.g. /q'a'-q'au/ 'skate', /s-na-na'^max°/ 'Nanaimo' (geogr.). (6) Occasionally in other types, e.g. in /k°a'i*k°ai/ 'play hide and seek' (/k°ai/ 'hide'), /sa'l'sl-tn/ 'implement for spinning thread' (/n-sa'l-us/ 'spin thread'), /s-X.aLXn2ai/ 'female' (/s-Xa'n-aj?/ 'woman'). 72. Suffixal Glottalization.— The transitivizer /-Vn/ and the intransitivizer /-Vm/ have glottalized alternants /-Vn?, -Vm?/. The choice between these alternants depends on the stem to which they are added, cf. on the one hand /k°a'i-an/ 'hide; tr.', /ii'k >0 -in/ 'munch; tr.', /mu'j-un/ 'soak; tr.' and on the other hand /q'a'i-an?/ 'put high up; tr.', /x0i'k°-in?/ 'brush; tr.', /c'u'i-un?/ 'peel; tr.'; with stressed suffix /q'lq'-a'n/ 'wind around; tr.' (cf. /q'alq'/ 'be wound around') versus /jaX'q'-a'n?/ 'paint, rub; tr.' (cf. /¿a'Vq'-tn/ 'paint (noun)'; implement-suff. /-tn/). In the case of ad hoc combinations of roots with lexical suffixes, the choice of alternant in the (in)transitivizer is determined by the suffix, e.g. /p'iP-qs-a'n, p'iP-qs-a'm/ 'grab someone by the nose', 'grab one's nose', /siP-qs-a'm/ 'wipe one's nose', /q°a-qs-aln/ 'perforate someone's nose', etc. versus /p>i?-q°-aln?, -a'm?/ 'grab someone by the head', 'grab one's head', /tmX,-q°-alm?/ 'paint one's hair'. Suffixes requiring the glottalized alternants are /-c/ 'mouth', /-5q/ 'hip', /-5n/ 'foot, leg', /-q/ 'behind', /-q 0 / 'head' (in a few words plain alternants were recorded, undoubtedly rapid-speech forms). The plain form is required by /-qs/ 'nose' (no exceptions). 73. Conclusion.— The material quoted in the preceding sections shows that glottalization is definitely characteristic of certain morphological formations as such. On the other hand, there is not a single case of glottalization in the whole category of the reduplicated plural, and morphemic glottalization is lost in the unstressed syllable of these forms, cf. /lam?, lmrla'm?/ 'house, houses', /s-manPa'X, s-mnrmanPa'V 'person(s) of high class', /s-tau?x°X, s-tu-ta'y?x°X,/ 'children', /man?, ma'nPrmn/ 'offspring', /s-Xan?, s-Xa'nP-xn/ 'leg'. This may be regarded as a regularizaron of a tendency to drop /?/ in unstressed syllables, cf. /?3s-k°u'i?c/ 'joking, talking funny' vs. /k°ui-k°uic-alm/ 'joke, talk funny', /Pas-iu'Pk0/ 'stingy' vs. /nax°-iuk°-al?min/ 'treasure' /Xa'mPx0/ 'rain' vs. /XmxSni't-m/ 'be caught by rain', /ó'i'Px0/ 'dry' ví. /£'ix°-i'?/ 'get dry', /iicq-a'n?/ or /Xi'cq-n/ 'fell (tree)', but in cases other than the red. plur. there are exceptions, e.g. the glottalization in /n-q0i'?s-tn/ 'cooking pot' is preserved in /q°i?s-a'n/ 'boil; tr.' (as opposed to /ó'ix^a'n?/ 'to dry; tr.' besides /ó'i'Px0/). In several cases where glottalization is absent it is probable that reduced forms —

56

PHONOLOGY

or at least free variants of glottalized ones — were recorded. For instance, if of /q°ah/ 'be pierced' and /k°aj/ 'hide' glottalized forms were recorded in /?as-q°3l-q°a?/ 'perforated', /?as-k°a'-k°ai?/ 'hidden', it is probable that /Pas-q'a'-q'i/ 'being on top' is a reduction of "/Pas-q'a'-q'ai?/ (see 36). On the other hand, I cannot explain the presence of glottalization in a reduplication like /mi 2 ma'i?/ 'lose one's way' besides /majj 'forget', /ma^nax 0 / 'have forgotten'. — If in /min?.min?u'lX/ (recorded [me n ?men?o' 'R]) 'young of animal' glottalization is present before the stressed syllable (a position where it is particularly subject to elimination), this must be due to the fact that it is morphemic (cf. /man?/ 'offspring'), but morphemic glottalization is often lost, and regularly so in the reduplicated plural (see above); I write /min?» minP-u'lX/, though the 2nd instance of/?/ could also be junctural. — In/n-XatXi'tPajus/ 'far-sighted' (cf. /Xa'ta/ 'far', /-ai?us, -ajus/ 'eye') there is a unique occurrence of /?/ at the border between a stem ending in a cons, proper and a vocalic suffix; possibly the full form of the simplex is */ia'ta?/, here reduced to /tot?-/ (cf. also /£m?a'?/ 'carry a load' and /£a'm?atn/ 'shoulder strap' (implement-suffix /-tn/), where stemfinal /-a?/, here preceded by /m?/, is reduced in a different way, resulting in an unstressed occurrence of /a/). — The /n?/ in /?as-n?i'u?/ 'be inside' (Vniu?, nau?, nu?) is also hard to account for; possibly /-n?iu/ goes back to */?ni'u?/, reduplicated */?3-?ni'\i?/ >/?n?i'u?/, as in /Pmu't/ 'assume a sitting position', red. /?m?u't/ < */?a-?mu't/ 'sit; be at home'. The initial /?/ of */?n?i'u?/ could then be lost after /?as-/, see 44. In this way, the occurrence of /?/ leaves a number of problems which require a special investigation. B. THE OCCURRENCE OF /a/

74. The overwhelming majority of the occurrences of /a/ are predictable on the basis of the following rules: (1) In the absence of a full vowel, a morpheme /CVC/ has the form /CC/ when unstressed and preceded or followed by a vowel; otherwise it has the form /CaC/. (2) In the absence of a full vowel, a partial-reduplication syllable has the form /C x a-/. (3) A morpheme-initial sonant is non-syllabic in the morphological environment /...CVC + R(V)C.../ (phonemically /...CVC.R(V)C.../, cf. 39; note that /./ indicates the absence of /a/) except if the stress falls "between" the constituents (phonemically /...CVCa'RVC.../). With regard to rule (1) account must be taken of the automatic alternations of morphemes /Caj, Cau, Cah/ discussed in 14-19; in particular, the forms of these morphemes corresponding to /CC/ can be /Ci, Cu, Ca/. Examples with the zero-grades of Vt'ak >0 , t'ak>0 'dig', V t ' a q " , t'aq' 0 'break', Vt'aq', t'aq' 'across' (used in the numeral /t'a'q'-ad/ 'six', lit. 'across-hand', and derivatives) and with -v/t>3* 'open': (A) Without the conditions mentioned under (1): /t'a'q' 0 / 'break', /t'aq'^mi'n?/

QUESTIONS OF WORD-PHONOLOGY

57

'broken-off half' (cf. rule 3), /s-t>alq>0-t'aq>o-s/ (geogr.), lit. 'cut up in chunks', /t'aq'-t'a'q'-aC/ 'six (persons)', /t'a'X-t'aX-c/ 'have the mouth open'. (B) Under the conditions of (1) / t V - i ' m ? / 'dig; act.-itr.', /t'q^-i'm?/ 'chop; act.-itr.', /n-s-t'q'-i'us/ 'side of body', /t'X-a'iPus-m/ 'open one's eyes', /t'aX-t'X-a'iPus/ 'have the eyes open' (2nd root-syllable), /Pas-t'a'-t'k'0/ 'pit, dugout' (cf. rule 2; an unstressed reduplication-syllable in /ta-ts-i'P/ 'feel cold'), /Pas-t'a'-t'q'/ 'lying across'. (C) Examples illustrating rule (3): /s-mnrma'nit/ 'mountains', /mUmi'li?/ 'get mixed up', /PnTma'n?/ 'my /?n-/ child'; the stress falls "between" constituents in /q>cq°a'mn/ 'ax', cf. -vAfuq 0 , q>03q° 'beat, strike' and the unique constituent /-mn/. 75. If morpheme-borders are indicated, the examples given in the preceding section could be unambiguously written without /a/: an automatic schwa is present between consonants (a) under the stress, (b) in the environment / Q - Q / , and (c) in the environment l#CC#l (where # denotes absence of a vowel). The elimination of /a/ would have the advantage that an automatic schwa between two morphemes would not have to be assigned arbitrarily to either the preceding or the following one, cf. /q>0q°-a'mn/ or /q>0q°a'-mn/ quoted above, and other cases such as /taia'uiX/ 'racecanoe' (/taiI 'race in canoe', /-uity 'container, canoe'), /qp'at/ 'close; tr.' (/qapV 'close; itr.', /-t/ transitivizer). The difference between the trans, derivatives of /qap'/ and of/q >0 al/ 'be done, ripe, etc.', viz. /qp'at/ vs. /q'°alt/ 'cook; tr.', could be treated as a matter of stress: /CC'C/ vs. /C'CC/. Furthermore, a more elegant notation would be achieved, e.g., in the case of prefixes such as /?n-/ 'my' and /Pas-/ (participial prefix), which could be uniformly transcribed /PC-/. Of the difficulties entailed by an elimination of /a/, some could be solved easily. Prefixes with the shape /KK-/ are pronounced without schwa (the only instances are /tx0-/ (directional) and /sx°-/ 'step (-relative)'), and the syllabic structure of a word like /sx°Tman/ 'stepfather' differs from that of /saq'rmi'n/ 'cuttings, waste wood', where /saq'/ is a root CVC. The distinction could be expressed either by using a special symbol for the border between prefix and stem, or by regarding these prefixes as bimorphemic (besides /tx°-/ there is /nax°-/ (locational), and /sx0-/ can be said to contain the nominalizer /s-[), and by stating a rule for the pronunciation of initial /C-C-C.../. — It would be necessary to distinguish, in the notation, reduplicative formations from "accidental" sequences / Q - Q . . . / , cf. /ta-tauP-i'P/ 'become bright' vs. /t-ta'i?/ 'that one (relative case)'. A sophisticated morphophonemic notation would go a long way towards eliminating the non-automatic occurrences of /a/, while an indication of immediate constituents would largely remove the need for the symbol /./. There are certain factors, however, which stand in the way of an elimination of /a/ and /./ from the notation. In the first place, /a/ was sometimes recorded in unstressed syllables of longer, to me unanalyzable units, e.g. /s-Pac'a'mPas/ 'women's shoulder blanket'. It was even recorded in a few cases which violate rule (1) in 74, e.g. /Xa'-Xa5/ 'remember', /qa'-qaC/ 'be increasing (ab. moon)'. Very probably, either incidental svarabhakti-vowels or reduced full vowels are involved here; it is also possible that

58

PHONOLOGY

the stress was misrecorded. In the second place, complex statements would be necessary to distinguish, for instance, the 1st pers. sing, object suffix /-m§/ from the in some words habitually reduced form of the formative /-mi§, -ma3/. I also hesitate to remove the schwa from the notation in such cases as /q ,0 aj?q ,0 a 'iPaX/ 'shy, wild', /-ai?aq°, -iPaq0/ 'top of head' (as compared to /?aj?X/ 'crab'), see 49, from /ta?a'u?n/ 'expect', etc. (as compared to /tPa'X-i/ 'that (female) one (relative case)'), see 68, end, from /Ca'mPatn/ 'shoulder strap', see 73, etc. On the other hand, the elimination of/a/ before unstressed /m, n, 1/ already necessitates the use of /./ in a few cases which from a synchronic point of view are monomorphemic, e.g. in /s-t'l.mu'Pt/ 'old person', /-X.n-I 'throat'; when /a/ is eliminated altogether, this number increases accordingly. Nevertheless, the status of /a/ is not on a level with that of the full vowels. In more than one respect, /a/ has the character of a juncture-feature rather than that of a phoneme.

C. VOWEL AND SONANT

76. If only free morphemes CVC are considered, our transcription contains a double redundancy, viz. the use of the symbol /a/ and the distinction of vowel /a, u, i/ and sonant /h, y, j/. Since in our transcription initial vowels are excluded, it would be possible to have */hC, uC, iC/ represent /haC, uaC, iaC/, while */Ch, Cu, Ci/ could stand for /Cah, Cau, Cai/. Also, */hhi, iuh, uhi/, etc. would be unambiguous renderings of /hiy, j[uh, uai/, etc. Going on to longer — but still monosyllabic — morphemes, the difference between /Xiq'/ and /Vaiq'/ (see IS) could be viewed as a matter of stress: q'/ vi. */X,>liq>/, cf. the difference between /Xca'm/ 'cedar box' and /c'a'Xt/ 'gravel beach', in a notation without schwa /Xc'm/ vs. /cJ|Xt/. The interpretation of these distinctions as a matter of stress would account for the near-absence of a distinction /aR/ vs. /A/ in unstressed position (see 19, 32, 33); it would account for the fact that the non-affixal morpheme-type CA is limited to clitics, which are habitually unstressed, and even where clitics have stressed counterparts such as /ti1/ 'this', they would still be distinguished from the type /ma'i/ 'sink', the latter being transcribed */m'i/. However, in longer words similar difficulties are encountered as with the elimination of /a/. Also, a transcription without the distinction vowelsonant is hard to read even in unambiguous cases, cf. */uhi/ = /uai/. One could, of course, write */uai/ or */whj/, but cf. the transcriptions of/Paha'num/ 'be painfully hungry', /huia1?/ 'depart', which would be either "/Phh'nwm, hwjh1?/ or •/Paa'num, auia1?/. A notation with vowel- and sonant-symbols is much more convenient; one regrets only the fact that /a/ is not a familiar symbol for /h/. Even though — apart from practical considerations — a notation without the distinction vowel-sonant does not suffice for certain longer words, the absence of such an opposition in the simpler morpheme-types is significant and constitutes one of the characteristic features of Squamish phonology.

QUESTIONS OF WORD-PHONOLOGY

59

D. GROUPS OF CONSONANTS PROPER 77. The zero-alternant of a morpheme K V K can occur word-initially (e.g. / t V - i ' m ? / 'dig; act.-itr.', of -v/t'ak' 0 , t'ak' 0 ), so that in principle any group K K can occur in this position. There is therefore no need to sum u p the ca. 90 initial groups K K which occur in my material: further additions to the lexicon will add to them in a random way. The groups are spread evenly over the grid of possibilities, except that groups beginning in /s-/ are particularly numerous due to the frequent use of the nominalizer /s-/. Initial groups K K K involve either this nominalizer or one of the prefixes /tx 0 -/ (directional), /sx°-/ 'step (-relative)', e.g. /sx°-5i'Sa?/ 'step-mother' (here also belongs /X(s)q'i?s/ 'to know', see 65). Medial and final groups are also determined morphologically. Some represent zero-alternants, e.g. /q 0 3'-q°tq/ 'be passing by', a reduplication of /q°3tq/; others result from suffixation of morphemes /-K, -KK/, e.g. /Pas-Xa'uV-Cq/ 'lame'. A special type of medial group results from the reduplication-type / K ^ V K ^ K J i J , e.g. /qa'p'-qp'-us/ 'ly facing downward', lit. 'close /qapV face /-us/', /Pas-t'aX-t'X-a'Cx0/ 'branchy, many-limbed', cf. -y/t'sX 'open', suif. /-a6x°/ 'branches, foliage'. The only consonant-sequences that are systematically excluded are groups of /s/ + /£, 5', S/. Where such a group would be expected morphologically, /§/ is found instead of /s/. We transcribe /s/ in all cases where there is no morphological evidence to the contrary (and in the large majority of these cases, there is positive evidence for /s/). The nominalizer /s-/ may serve as an example: /s-£iX-c/ [s£-] 'upper lip', /s-d'aq/ [§£'-] 'dirt'. The nominalizer is assimilated even before the zero-form of a root with a palatal C s , e.g. /s-q >0 §3'cut/ [sq'°s-] 'toy' (cf. /q >0 sa-t, q'^a'cut/ 'play; tr. and refl.'). E. ALTERNATIONS 78. Except for the partly automatic alternations to which roots ending in a sonant are subject (see 14-19, 31-34), for the loss of an initial /h/ or /?/ (43, 44) and for the alternation of full and /a/- or zero-grade (57), Squamish has but a few instances of alternating forms. Those concerning clitics were already mentioned in 56. All other cases are incidental; they are summed u p here according to the type of alternation concerned. (1) Free variation of forms with /t'/ and /X,'/ in /t'aX.na'im, X'aLna'im/ 'go on a prolonged hunting-trip', /st'l.mu'Pt, sX'l.mu't/ 'old person' (the absence of /?/ in the second form is probably incidental). (2) Free variation of forms with /X'a?-/ and ¡V-l in /X'aPa'Sn, k'a'Sn/ '(invite to) feast, party, dance (not potlatch)', /X'aPa'nq, Vanq/ 'potlatch' (cf. also /X'aPi'min/ 'sinew' and Cw. X'i 'man 'tendon'). (3) Loss of final vowel before vocalic suffixes in /hii 1 , hi-/ 'big', /?ina, Pin-/ 'the one, the other'. (4) Roots with any vowel can have an alternant with /i/ in diminutive forms. This

60

PHONOLOGY

happens regularly in diminutive reduplications of polysyllabic stems, and occasionally in other cases, see 159. In a few instances, roots with /a/ appear with /i/ in related words without obvious diminutive meaning, cf. /malq"/ 'be mixed' and /mlrmi'lC'/ 'get mixed up', /A.a'C'-tn/'knife' and /Xi5*/ 'be cut', the unique pair /sui'uPlus/ 'young man' and /sua'uPlus/ 'id., plur.'. The difference /a/ w. /i/ is accompanied by other phonetic differences in /XauPs/ 'new, fresh' and /Xis-i?/ (stress unknown) 'be fresh', /X,au?/ 'recover; run away (ab. a captive)' and /X,'iu?/ 'run away (ab. a captive)'. (4) An alternation /a/-/u/ occurs in the pairs /?acq/ 'the outdoors, outside' and /?ucq/ 'go outside', \/? a i(?) 'inside' (/?aj[-s/ 'be inside', /?ai*tx°/ 'be at home', /?i-?a'i2uX,/ 'be aboard') and V? u i 'enter' (/?ui-s/ 'enter', /Pu'j-uX,/ 'go aboard'). (5) For related words with initial /[-/ and /hi-/ see Dictionary under Vi-1» VfaK /iauPa'n?/.

PART II

MORPHOLOGY

VI INTRODUCTION

79. Squamish words fall into two main classes: full words and clitics. Full words may consist of a single morpheme, of a reduplication of such a morpheme, or of either of these two combined with one or more affixes. Productive compounds of free forms do not occur. The affixes fall into five main categories: (1) personal affixes, (2) modal and adverbial affixes, (3) relative case prefix, (4) plain nominalizer and verbalizers, (5) others. This last category is not homogeneous. It contains elements which — besides having a lexical meaning — nominalize or verbalize a unit to which they are added, and elements which do not. It contains elements with a clear and concrete lexical meaning, elements which might be called grammatical, and elements to which no clear meaning can be ascribed (formatives). A unit which contains no other affixes than those mentioned under (4) and (5) above is called a stem. Since the affixes (1-3) are added at the beginning or end of a word, a stem is a word minus its personal, modal, adverbial and case-affixes (if any). Disregarding for the moment the occasional nominalizing or verbalizing role of the affixes (5), we may divide the Squamish stems on the basis of the affixes (4) into three formal categories: (a) nominalized stems, (b) verbalized stems and (c) unmarked stems. Nominalized stems contain as their initial part the nominalizer /s-/, e.g. /s-taq°/ 'water', /s-t'iq 0 / 'cold (noun), cold weather', /s-pcj 'realistic story', /s-qa?q/ 'younger sibling'. Verbalized stems contain as their final part an intransitive suffix /-(V)m(?)/ or one of the transitive suffixes /-(V)t, -s, -(V)n(?), -nsx0/, e.g. /jp'c-m/ 'tell, report; itr.', /ta'q°-an/ 'drink; tr.\ Unmarked stems are defined by the absence of any of these affixes. If a stem contains both /s-/ and one of the verbalizers, it is nominal, i.e. the IC's are /s-/ and the remaining part. Particularly common are nominalized derivatives with itr. /-(V)m(?)/, e.g. /s-js'c-m/ 'news, report'. Nominal stems can be combined with possessive affixes, e.g. /?n-sqa'?q/ 'my /?n-/ younger sibling', verbal stems cannot. Nominal stems are frequently combined with identificatory clitics (articles) or demonstratives, e.g. /ta w sta'q°/ 'the water', while verbal stems are frequently combined with predicative clitics, e.g. / n a ^ y a ^ i o ' c m /

64

MORPHOLOGY

'he is telling, giving a report'; however, both types of clitics are possible with both types of stems, and this difference in syntactic status is statistical rather than absolute. Of the stems belonging to the unmarked class, some have the abovementioned peculiarities of nominal, others those of verbal stems. Among the nominal stems there are only a small number that represent the simplest lexical morpheme-type CVC: /man/ 'father', /tai/ 'canoe-race', /5iX/ 'top', /las/ 'bottom', /k ,0 a57 'dogfish', /x°at/ 'wren', /ya 1/ 'unidentified plant', borrowings /puS/ 'cat', /lam/ 'whiskey' (Engl, rum), /ua51 'watch'; with final glottalized sonants /man?/ 'child, offspring', /5'i?/ 'person who is fasting', /sau?/ 'bone', /lam?/ 'house', /k>0al?/ 'stomach', /?im?/ 'small boy'. In the overwhelming majority of cases nominal simplicia are represented by longer forms, e.g. /naiC/ 'hand, arm', /ma'qsn/ 'nose', /Xa'Su?/ 'lake', /Xqaj£7 'moon', /ki'lala/ 'butterfly'. The remaining unmarked stems are verbal in character; they are intransitive and occur in general in the same environments as stems marked by intrans. /-(V)m(?)/. These verbal stems correspond to English verbs, adjectives, adverbs, numerals and prepositions; they include the large majority of the morpheme-type CVC, and longer forms are rare here. As examples may serve the stems with /t, t'/ as first and /k°, k'°, q°, q J 7 as second consonant: /tek'7 'be tight', /taq°/ 'drink', /tiq°/ 'be muddy', /tiq'7 'bump, run aground', /t'ak0/ 'get stuck in the mud, get mired', Vt'ak' 0 , t'3k>0 'dig', Vt'aq", t'aq' 0 'break', /t'uk'7 'go home', /t'iq0/ 'be cold', /-t'uq0-/ in /s-t>uq°-£-us/ 'forehead' (the suffix-combination /-i-us/, lit. 'back-face', means by itself 'forehead'; the presence of the nominalizer /s-/ in this combination implies the verbal nature of /-t'uq0-/). Within this group only one nominal stem was recorded, namely the reduplicative formation /t3q°-tu'q°/ 'red codfish, red snapper'. The meanings of unmarked verbs are given as active or passive according as they occur in recorded phrases or were given by the informants, e.g. /5-n w ta'q7 means 'I /6-nI drink', /5-n_c J 3'h/ means 'I am hit'. There are cases where both are possible (cf. English "he reads the book" and "the book reads easily"), e.g. /x°i?/ 'be lost', 'be the loser (in a game)', /653tx°/ 'be decorated with carving', 'carve; itr.'. In some cases where an unmarked stem was recorded with an English translation but without examples of use, it is uncertain whether the translation as verb or noun reflects the true nature of the stem. For example, the meaning of /Xal/ was recorded as 'a pulling-game with a rope, to see who is strongest'; the word may mean 'tug-ofwar' or 'play tug-of-war', or both. Unmarked stems containing affixes of category (5) above may be nominal or verbal, in the same way as simple stems. When their character in terms of noun or verb is clearly due to the affix, the latter is classed as nominalizing, resp. verbalizing. Most affixes are neutral in this respect. For details see 163. Within the class of verbal stems, two categories are set off by certain syntactic and semantic (in part also morphological) peculiarities: these are the numerals and the relator-verbs, the latter corresponding to English prepositions.

INTRODUCTION

65

There is, in addition, a class of substitutes which includes nominal as well as verbal stems; this class corresponds to English pronouns and pronominal adverbs. Finally, a very small number of stems is set apart as residual; this class contains a few words corresponding to English adverbs or adverbial phrases, conjunctions and interjections. In addition to full words, Squamish employs a number of clitics. These fall into two main classes: nominal (identificatory) clitics or articles, and predicative clitics (a better term might be "factual clitics"), the latter indicating person, tense, mood, aspect and other features of predicative expressions (main and subordinate). The modal and adverbial suffixes (category (2) above) can be attached to predicative clitics and are often found within strings of the latter, so that it is convenient to treat these two categories together. In describing Squamish morphology, we treat first the various types of affixation (nominalizer, verbalizers, personal affixes, non-personal affixes) and reduplication. The distinction noun-verb is made, where necessary, within each affixal or reduplicative category. For reasons of expediency, reduplication is treated before non-personal affixation; this requires fewer references ahead. — Thereafter are treated the case-forms, the articles, the substitutes, the special verb-classes (numerals, relator-verbs), the predicative clitics together with the modal and adverbial suffixes, and the residual words.

vn THE NOMINALIZER /a-/

80. The nominalizer /s-/ occurs in a large number of fixed combinations with unmarked stems as well as with intransitive (seldom transitive) derivatives. Examples: /taq°/ /t'iq0/ /XucV A'i?/ /c'ic'a'p7 /tan?-t/ /t'mS-a'n/

'to drink' 'be cold' 'be scabby' 'be dear' 'to work' 'weave; tr.' 'braid; tr.'

/s-taq°/ /s-t'iq0/ /s-XucV /s-X'i?/ /s-c'ic'a'pV /s-Xan?/ /s-t'amS/

'water' '(the) cold' 'scabies' 'desire, thing wanted' '(the) work, job' 'mat, woven fabric' '(a) braid'

/s-ia'c-m/

'news'

Examples with intransitive derivatives: /ia'c-m/ 'to tell* (cf. /s-jac/ 'story') /pah-i'm?/ 'to blow' (wind) (cf. /puh-t/ 'blow; tr.) /tV-im?/ 'dig; itr.' (cf. /t'a'k^-an/ 'dig; tr.')

/s-pah-i'm?/ '(the) wind' /s-t'k'Mm?/ 'pit, dugout'

In a number of cases the root was not recorded without intransitive suffix: /ni'Sim/ /Xi'c'mI /p'u'X'amI /t'lq^'m?/

'speak; itr.' 'spark; itr.' 'smoke; itr.' 'nurse, take the breast'

/s-ni'Cim/ /s-Xi'c'm/ /s-p'u'X'am/ /s-t'lq^'m?/

'speech, word, language' '(a) spark' '(the) smoke' 'breast'

A fixed nominalization of a trans, derivative (root not recorded without tr. suffix): /5a'§-nI

'send; tr.'

/s-Sa'S-n/

'messenger'

In a number of cases no counterparts without /s-/ were recorded, e.g. /smic/

THE NOMINALIZER /s-/

67

meat', /sX,iq°/ 'flesh', /sqalX/ 'stick for digging clams', /sC'alq's/ 'slingshot', /sX,i'Sm/ 'twisted cedar-leaves of which rope is made'. In general, any stem of the type /sCVC.../ may be regarded as an /s-/ derivative of a stem /CVC.../, except, of course, for cases where /s/ belongs to the root and a suffix /-VC.../ is involved, as in /sq'-im?/ 'split wood; itr.', where the root is /sq'-/> cf. /saq'/ 'crack, split lengthwise; itr.', in this example combined with the active-intrans. suffix /-im?/. In a few cases forms with and without /s-/ occur side by side, e.g. /(s-)tqa'ia?/ 'wolf', /(s-)c'na'i?/ 'bullhead', /(s-)5'iu'i/ 'twins'. The words /s-taj?6/ 'back' and /s-?a'cus/ 'face' may be used with and without /s-/ when they refer to the back, resp. front, of a house, and not to parts of the body. Note /k°ci?c/ 'person with magic power' versus /s-k°ci?c/ 'relative'. 81. Besides occurring in fixed combinations referring to objects, i.e. in nouns, the nominalizer /s-/ may as a result of a living grammatical procedure be combined with any verbal stem (and also with certain predicative clitics) to refer to a fact. In the most frequent type of case, such a nominalized verb-form is accompanied by a possessive affix, and the whole expression takes the indefinite article /k°i/, e.g. /k°iwn-STnil5im/ 'the fact of my /n-/ speaking', 'that I speak', /k^^n-Srna'm?/ 'the fact of my going', 'that I go'. Note that there is a noun /s-ni'Cim/ 'word, speech, language' (/ta^n-Srni'iim/ 'my word, speech, language', with the definite article /ta/); on the other hand, there is no noun */s-nam?/. For this use of /s-/ see 133, 135.

Vili TRANSITIVIZERS A N D INTRANSITIVIZERS

FORMS WITH / - ( V ) m /

Unmarked form

Plain itr. /-(V)m/ Active-itr. /-im?/

ITIVE

FORMS WITHOUT / - m /

Active TRANSITIVE

IN-

TRANS-

82. Unmarked stems are transitivized by the addition of suffixes characterized by a final /-t/, /-s/, /-n/ or /-nax 0 /. Both unmarked and transitivized stems may be followed by an intransitive suffix characterized by a final /-m/. An unmarked stem with an intrans. suffix yields an intransitive (often reflexive) verb. A transitive stem followed by an intrans. suffix yields a passive form. The suffixes with /-n, -m/ also appear with /-n?, -m?/, depending on the stem t o which they are added (see 72). Of the abovementioned suffixes, those with /-t, -n, -m/ may either consist of a single consonant or have the shape /-VC/, i.e. they may include a vowel. This vowel is often — but by no means always — identical with the (last) vowel of the stem t o which the suffix is added. There is, in addition, an intrans. suffix /-im?/, always with /i/ and glottal stop, which carries the meaning 'active, non-reflexive intransitive' ("be busy cleaning", " b e throwing punches", etc.). The suffixes /-s/ and /-(V)n/ are extended with an element /-t-/ if other suffixes (for object or subject, or itr. /-m/) are added. The suffix /-nax°/ has the form /-n-/ when an object-suffix or itr. /-m/ follows (it remains /-nsx 0 / before a subject-suffix). T h e above information on the transitive and intransitive suffixes is summed u p here:

final /t/-form /s/-form /n/-form /nax°/-form

j

medial

/-(V) t/ H ! l-s-t-l /-(V)n/ | /-(V)n-t-/ /-nax 0 / j /-nax 0 -/ /-n-/

Passive /-(V)t-m/ /-s-t-m/ /-(V) n-t-m/ /-n-m/

TRANSITIVIZERS AND INTRANSITTVIZERS

69

In the few cases where /t/- and /n/-forms occur of the same root, the former refers to an action which affects its object as a whole, the latter to one which affects it partially or superficially, cf. /Xi°-\xt/ 'spit (something) out' vs. /ta'jf-n/ 'spit (at someone)', /q°a-t/ 'save, rescue (someone)', lit. 'pull through', w. /q°a-n/ 'perforate (something)'; originally also /cu-t/ 'say (something)' vj. /cu-n/ 'tell, order (someone) (for /cut/ see 85). The distinction is not productive; from a synchronic point of view the suffixes /-(V)t/ and /-(V)n/ may be regarded as non-automatic allomorphs of one transitivizer. The /s/-form is causative; the suffix /-s/ does not occur with all stems, but within certain groups of cases it is productive (see 91). The /n3x°/-form refers to a process viewed as independent of the will of the actor; it either denotes an act which is accidental, or it refers to a state of affairs which results from an action (voluntary or not) and the existence or maintenance of which does not require an act of the will. Examples: unmarked form /i-n^ca'x 0 / 'I /6-n/ am (have been) hit (with something thrown)', /n/-form /5-n w c3 l x°-n/ '1 hit him (with something thrown)' (on purpose), /n3x°/-form /5-n w c3 l x°Tn3x7 'I hit him (accidentally)', 'I have hit him'; /t/-form /5-n w p'i'?-t/ 'I grabbed it', /n 3 x 0 /-form /{-n^p'i 1 ?nax°/ 'I am holding it', 'I have caught it', 'I got it' (note that the seizure was an act of the will, the resulting holding or having is not); /fc-n^q^u'P-t/ 'I included him, put him into the group (on purpose)', versus /£-n w ,q > °u l ?-nax7 'I got him into it (accidentally)', 'I have included him'. One says /5-n w ua w 5'a l u-at-umi/ 'I am helping you /-umi/' (/ua/ 'imperfective-continuous') but / n w a u w 6 - n w u a ^ 5 > a l u » n umi/ 'am I helping you?', 'am I of (any) assistance to you?' ( / n w a u w / introduces the question, /-nax0/ has the alternant /-n-/ before the object-suffix /-umi/); the speaker may have the best intentions, but the answer to his question is independent of these intentions. The /n3x°/-form is a productive non-volitional counterpart of the volitional /t/and /n/-transitives; it occurs also as a counterpart of a few causative verbs (see 102). 84. To demonstrate final and medial forms, here follow examples of /-t, -s, -n, -nax0/forms (1) without suffix, (2) with 2nd pers. sing, object-suffix /-umi/, (3) with 3rd pers. sing, subject-suffix /-as/ and (4) with intrans. suffix /-m/ which here yields passive forms. Except for (3), forms are given with 1st pers. sing, subject expressed by the clitic-combination /5-n/; in (3) the clitic /na/ 'there, then, at' is employed. With all but the /-n3x°/-forms, translations are given in the past tense, which is not expressed as such in Squamish. /caxc/ 'be hit (by something thrown)'

V/p'i?-/ 'grab, seize' 1. 2. 3. 4.

/£-n_p'i'?-t/ /5-n^p'i'P-t-umi/ /na^p'i'P-t-as/ /6-n w p > i'?-t-m/

'I grabbed him' 'I grabbed you' 'he grabbed him' 'I was seized'

1. 2. 3. 4.

/i-n_a>'x°-n/ /5-nv_calx°-n-t-umi/ /na w ca'x°-n-t-as/ /£-n^_,c3 'x°-n-t-m/

'I hit him' 'I hit you' 'he hit him' 'I was hit'

70

MORPHOLOGY

/x°i?/ 'disappear, get lost'

V/P^? - / 'grab, seize'

1. 2. 3. 4.

1. 2. 3. 4.

/6-n_x°ii?-s/ 'I discarded it' /6-n w x°i l ?-s-t-umi/'I discarded you' /na_x°i'?-s-t-as/ 'he discarded it' /S-n^xI'P-s-t-m/ 'I was discarded'

/C-n^p^P-nax 0 / /S-n^p'i'P-n-umi/ /na w p > ! l ?-n3x°-as/ /S-n^p'i'P-n-m/

'I hold it' 'I hold you' 'he holds it' 'I am held'

IX TRANSITIVIZERS

85. In the following sections the transitive verbs are summed up, subdivided on the basis of the transitivizers they contain. A number of verbs referring to the production of sounds are provided with trans, suffixes but are treated as itr. verbs (i.e. the 3 sing, predicative form lacks the subjectsuffix/-as/). They are the following: With trans, suffix /-t/: /cu-t/ 'say', /?i'na-t/ 'say what?', /q'ai-t/ 'holler'. With /-at/: /A,a'q-at/ 'whisper'. With /-n/: /cVm*n/ 'whistle', /Su'p-n/ 'whisper', /X°u'q,0-n/ 'snore', /uu^n/ 'bark', /Pa'guP-n/ 'sneeze'. Note that all cases with /-n/ involve stems with /u/. A special group of these "recessive" tr. verbs are those in /-i?n/; these include, besides verbs referring to sounds, a few verbs referring to light-flashes and sudden movements. The following examples were recorded: /q°3'Xq-i?n/ 'rap, knock' (cf. /qV^q-m, q V - q ^ q - m / 'id.'), /XVS-iPn/ 'splash' (cf. /XVS-m, XV-X^-m/ 'id.'), /p3'k>0-i?n/ 'puff, splash' (cf. /s-pa'k^-m/ 'dust', /pa'-pk^-m/ 'puff clouds (of smoke), be dusty (e.g. about a road)'), /pi'SMPn/ 'flash' (cf. /pi'C'-m/ 'sparkle', /pa'-piC-m/ '/¿'), /k°3'm-i?n/ 'thump' (cf. / k ^ ^ k V m / '/¿.'), /ci'k>0-i?n/ 'start, twitch (of fright)', (cf. /c3-ck>0-i't/ 'frighten; tr.'), /^iiik^n-i'Pn/ 'gnash one's teeth', /sa^iPn/ 'be audible', /Pa iX°-i?n/ 'cough'. With /-s/: /j|ulia?k0m-s/ 'make waves' (ab. boat). A. TRANSITIVE /-t/ SUFFIXES

55. The trans, /-t/ suffixes occur in the forms /-t/ (stressed /-s't/), /-at/, /-ut/, /-it/. Their selection is in part determined by the (last) vowel of the stem, as follows: /-t/ /-at/ /-ut/ /-it/

with with with with

stem-vowel stem-vowel stem-vowel stem-vowel

/a, a, i, u/ /a, a^ /a, u/ /i/

57. Zero-grade stems.—In the large majority of cases the suffix is /-t, (-3 't)/. The stress

72

MORPHOLOGY

is on the root C jVCgj Cj VC2C3 if C 2 is a sonant (including the case /Ca? = *Cah?/): /pan-t/ 'bury', /tol?-t/ 'know', /S'amP-t/ 'bite', /Xan?-t/ 'weave', /Xal?-t/ 'write', /q>03l-t/ 'cook', /ialX-t/ 'look for', /n-talq-t/ 'answer', /Xi-Xi '-t/ 'laugh at', /q°3h-t, q°a-t/ 'save, rescue', /t'aP-t/ 'try, taste', /ja?-t/ 'vomit', /5a?-t/ 'make, build', possibly also /ci-cut = ci'-t-sut/'paint oneself', with a \ / c 3 i not recorded otherwise than in this reflexive form (the root may alternatively be */cih/, in which case the word belongs in 89). Otherwise the stress is on the suffix: /cX-at/ 'push', /X\°-9t/ 'win, master', /q°5-at/ 'spit (blood)', /q>0s-3t/ 'play with', /q5-at-/ 'fill' (recorded only in /qC-a 't-m/ 'suck blood (ab. mosquitos)', litt. 'fill itself', cf. /qaS/ 'be full (ab. the moon)'), /&c°-3t-/ only in /5x°-9't-m/ 'swell', /(n-)qp'-3t/ 'close'. In /Pa'mPaq-t/ 'deliver', only example of a longer stem with /a/ in the last syllable, the stress remains on the stem, likewise in /k>0a'-k>05-t/ 'be looking at', a reduplication of/k'°a5-t/ 'take a look at'. In a small number of cases stems with /a/ take the suffix /-at/: /k>0X,-at/ 'pour out, spill' (also in /k'^a'cut = k'^-a't-sut/ 'capsize', lit. 'pour oneself out', cf. /kJ°aty 'spill; itr.'), /mi'acut = ma'j-at-sut/ 'go down, sink', lit. 'immerse oneself' (cf. /maj/ 'sink; itr.'), /q'i-a't/ 'consider too weak', /pi'-at-uty 'hunt in canoe /-uX/' (Vpsj). /qVq'X-a 't-aj?/ 'argue, debate' (reduplicated form of a root not recorded by itself; reciprocal suff. /-aj?/), possibly also /tqat/ 'ask' (see Dictionary). There are only two instances of zero-stems with /-ut/, both involving a labialized stem-final consonant: /£'x°-ut/ 'increase' and /X.X°-ut/ 'spit out' (but cf., also with lab. cons., /XV-at, 5x°-st/ above). For /c3-ck>0-i't/ see 59. Stems with /a/.— Of the 12 monosyllabic cases, 8 have /-at/ and 4 /-t/. The stress is always on the stem: /t'a'm-at/ 'guess', /S'a'u-at/ 'help', /Xa'u?-at/ 'heal', /X.a'q-at/ 'whisper' (see 55), /ka'u-at/ 'take, bring down', /k^a'j-at/ 'purify', /q'a'u-at/ 'pay (somebody)', /q'a'nacut = q'a'n-at-sut/ 'return' (reflexive suff. /-sut/); /pal-t/ 'skim off', /k>0a£-t/ 'take a look at', /malq°-t/ 'mix', /q^aj-t/ 'holler' (see 55). Longer stems take /-t/: /kVlaS-t/ 'shoot', /?3'x°a?-t/ 'give', /?iina-t/ 'say what?' (see 55), /na'uPnaC-t/ 'pay (somebody)', possibly /tm 2 a'i?-t/ 'get hungry for' (see Dictionary). 89. Stems with /i/.— Of the 15 examples, 11 have /-it/ and 4 /-1/. Stress on the monosyllabic stem: /ci'x°-it/ 'reach', /si'n-it/ 'move', /X.i'm?-it/ 'accept', /Xi't'-it/ 'spend gifts, distribute', /W^'-it/ 'cut, shear', /XW-it/ 'stalk (game)', /hi'l-it/ 'roll', /ii'C'-it/ 'fill', /n3x°Tni'u?-it/ 'teach, instruct' /q'i'uPicut = q'i'u^it-sut/ 'sail around a promontory' (reflexive suff. /-sut/); /qi'X-it-m, qi'Xicut = qi'X-it-sut/ 'slip, slide' (itr. and reflex, suff.); /lix°-t/ 'put down', /X°il?-t/ 'take off', /p'x?-t/ 'grab', /ui'lq>0-t/ 'ask somebody'. An exceptional case is /c3-ck>0-i't/ 'frighten' (cf. /ci'k>0-i?n/ 'be frightened, startled'). For /ci'cut/ see 87. 90. Stems with /u/.— Of the 9 cases, 4 have /-ut/ and 5 /-t/; the stress is on the stem: /Su'k'°-ut/ 'bathe', /q ,0 u'q°-ut/ 'club, hit', /q ,0 u'i-ut/ 'beat', /hu'i-ut/ 'prepare'; /puh-t/ 'blow', /q>0u?-t/ 'include', /cu-t/ 'say' (see 55), /?u-t/ 'call, invite'. Note that

TRANSinVIZERS

73

the cases with /-t/ are limited to roots with final /-h/ or /-?/, or without final consonant (the roots of the last two words are not found without suffixes; they may result either from */cau-, ?au-/ or from */cuh-, ?uh-/). This rule does not hold for /X0u'q>0cut = X°ulq>0-t-sut/ 'pole up (in canoe)' (reflex, suff. /-sut/), besides which no non-reflexive */X°uq>0-t/ was recorded.

B. CAUSATIVE /-s/

91. The causative suffix /-s/ is found almost exclusively with stems not containing other verbalizers; in only three examples it is added to stems with itr. /-m/ suffixes. With stems referring to motion or location it conveys the meaning 'take, bring'; with evaluative stems it gives the meaning 'consider as'; with a personal name 'X' it means 'call someone X' (with the implication that 'X' is not the person's original name); with stems referring to objects or objective qualities it means in general 'cause to be', with stems with verbal meanings 'cause to act'. The borders between these categories — especially the latter two — are not rigid. In some cases /-s/ has coalesced with stems no longer found by themselves. The following examples were recorded (the meanings of the Simplicia are added in parentheses): /mi-s/ 'bring' (come), /nam?-s/ 'take' (go), /X'iq-s/ 'bring' (arrive), /kau-s, kau-s/ 'take (bring) down' (also /ka'u-at/; cf. /kau, kau/ 'descend'), /t'uk >0 -s/ 'take home' (go home), /k°um?-s/ 'take ashore' (go ashore), /x°i?-s/ 'throw away, chase away, send home (one's wife)' (be loser, get lost); /?i-s/ 'keep' (be here), /SiX-s/ 'raise, put on top' (top), /Pa'nui^-s/ 'put in the center' (be in the center), /?a(s)Xi'c-s/ 'lay down' (lie down), /6m?a'?-s/ 'carry on one's back' (also /5m?a'?-n/; for /5m?a'?/ the meaning 'carry a load' was given); /ha?X-s/ 'like' (good), /qi-s/ 'dislike' (bad), /ViP-s/ 'like, love' (dear), /uana'x°-s/ 'respect' (true), /siPa'mP-s/ 'respect as a chief' (chief); /tx°n53?a'm?-s/ 'ask how much money for' (lit. 'cause to be how much'), /hau-s/ 'refuse' (be not), /?i?X°-s/ 'finish' (all), /s-Pa'li-s/ 'see in dream' (/s-Pa'li/ 'dream, vision, guardian spirit'; /Pl-Pa'li/ 'to dream'), /n-q'i'-q'l-s/ 'fool, outsmart' (cf. /n-q'i'l-us/ 'smart, clever'), /x°ak°-s/ 'use' (be used), /hui 2 s/ 'eat' (be ready, finished), /ia'P-ja-s/ 'put on tight, tie tightly' (/¿a?/ 'tight, tightly shut'), /Pip'a'q'^aX-s/ 'frighten' (be afraid), /?i-s/ 'make merry, have a good time' (V?3l 'good'); /ta?-s/ 'make, build' (/tah?/ 'undergo'), /cax°-s/ 'throw' (hit; itr.), /n-qVX-s/ 'hit', lit. 'cause to be in the way' (cf. /n-q'a'X,/ 'be in the way', /n-q'a'k-n/ 'put in the way'), /Pi'X-n-s/ 'feed' (eat), /q'a'uaXn-s/ 'punish' (expiate, atone), /k°u'^n-s/ 'lend to' (borrow), /ia-s/ 'do what with', lit. 'cause (it) to do what?' (see Dictionary under /6a-/), /timPa'-s/ 'cause to be like' (be like), /ua'Pu-s/ 'keep at' (continue); /5i-fii§k°-s-t-m/ 'fall behind (in a canoe race)' presupposes /£i§k°-s/ 'leave behind' (cf. /5i5k°/ 'recede, ebb'). The causative suffix is added to itr. derivatives in /Xa'i-m-s/ 'cause to laugh' (laugh), /t>lq°-ilm?-s/ 'suckle; tr.' (id., itr.) and /iu'jaPk^m-s/ 'make waves' (undulate). A case apart is /k>0al5rmix°-s/ 'cause to look' (y7k'°a5-/ 'look, see'), of which no simplex was recorded. The relation between /t'3k°s/ 'explode'

74

MORPHOLOGY

(/na_t'ak°s ta^naxMa'mai/ 'the bottle exploded') and /tJ3k°/ 'get stuck in the mud, mire' is not quite clear. There remain a number of cases where the bare stem was not recorded by itself or where its meaning is uncertain: /X,(s)q*i?-s/ 'know', /han-s/ 'tell (somebody)', /?al?-s/ 'feel sorry (for somebody)' (cf. /Pah/ 'hurt'), /6i-5aj2s/ 'follow, chase' (cf. /Ca'i-n/ 'id.', /Sa'iaq-m/ 'id.; itr.'), /Pi-Pa'j^s/ 'wait for', /XiP-Xi'-s/ 'make a feigning movement at', /ii'c'-ic'-s/ 'step on', /Pi'p'i-s/ 'hold, carry' (V/p'i?-/ 'grab, hold'). I am not sure of the morphological status of final /s/ in /sua'ls/ 'scatter'. C. TRANSITIVE /-n/ SUFFIXES P2. The trans. /n/-suffixes occur in the forms /-n/, /-an/, /-un/, /-in/ (also /-n?, -an?, etc./). In the case of stems CVC, the suffix /-n/ is found with any V; the vocalized alternants of the transitivizer follow the stem-vowel: /pVc'-an/ 'sew', /p'ic'-in/ 'squeeze', /c'up^-un/ 'connect', etc. (except that /-a'n/ is found in a few cases with stem-vowel /i/, which clearly represents /ai/ here, see stems CVCC below). The stress is on the stem. Of the type /CV'C-n/ there were recorded 26 cases with /a/, 10 with /u/, 6 with /a/ and 4 with /i/ as stem-vowel, of the type /CA'C-An/ 26 with /a/, 17 with /i/ and 11 with/u/. Stems CVCC (23 cases with /a/, plus 3 cases of /CiC = CajC/; 2 cases with /i/ and one with /u/ and /a/ each) take /-a'n/, with the stress on the suffix; this goes for unanalyzable units as well as for derivations CVC-C (the latter not included in the above count). For longer units see 101. The Type /CV'C-n/ 93. Zero-grade stems: /pa'c-n/ 'bend, fold', /pa'X°-n/ 'spit at', /ma's-n/ 'put together, cause to stick together', /c3'k 0 -n/ 'tug, pull', /ca'x°-n/ 'throw at', /c'a'k^-n-t-m/ 'be worm-infested' (see 121), /na'k°-n/ 'move, shake, stir', /na'u?-n/ 'put in', /5a'§-n/ 'send (a person)', /ta'm-n/ 'pick berries', /X.9'it0-n/ 'spit at', /Va'i-n/ 'stop', /k >0 a'c-n/ 'pluck', /qa'x°-n/ 'gather, pick' /q'a'm-n/ 'swallow', /q >0 a'c-n/ 'make wet', /Xa'm-n/ 'make heavy', /Xa'i-n/ 'stop (people) from fighting' /?a l q°-n/ 'remove hair from (hide)' (/?a'q°ncut/ 'moult'), /ca'h-n, ca-n/ 'punch, hit', /na'-n/ 'call' (cf. /s-nah/ 'name'), /q°a-n, q°a'h-n/ 'pierce', /ia-n, ja'h-n/ 'warn', /Xa'P-n = Xa'h?-n/ 'touch', /ia'?-n = ia'h?-n/ 'tie tightly shut' (cf. /¿a?-s/ 'put, tie on tightly'). 94. Stems with /u/: /mu'?-n/ 'drop', /tu'j-n/ 'leave, abandon', /cu-n/ 'tell, order', /c'u'm'n/ 'hiss', /c'u'P-n/ 'pull out', /§u'p-n/ 'whistle', /Xu'p-n/ 'put away', /x 0 u'k ,0 -n/ 'pull, drag', /X°u'q>0-n/ 'snore', /uu^n/ 'bark'. For /c'u'mPn, gu'pn, r u ' q ^ n , uu'Pn/ see 0-an/ 'slap', /ua's-an?/ 'remove from the fire', /ua'X-an/ 'chase away', /ja'q'-an/ 1. 'cause to fall down', 2. 'sharpen (by filing)', /ia'q°-an/ 'extinguish'. 95. Stems wj'i/j /£/; /p'i'c'-in/ 'squeeze', /mi'k>0-in/ 'wash', /mi'q'-in/ 'press down', /ci'q-in?/ 'stab', /C'i'm-incut = -in-t-sut/ 'approach' (reflex, suff. /-sut/), /Wt'-in/ 'scatter', /Xi'k>0-in/ 'hang, hook up', /(nOk'i's-in/ (also /(n^q'i's-in/) 'tie, knot', / k ' W - i n / 'clean (fish, etc.)', /x°iik°-in?/ 'brush', /Xi'p'-in/ 'scratch', /Si'm-in?/ 'pull someone's hair', /Xi'X'-in/ 'chop (wood)', /Xi'k>0-in/ 'munch, chew', /Xi'q'-in/ 'scratch', /X0i'q>0-inI 'tie up, connect', /?i'q>0-in/ 'rub off'. 99. Stems with /u/: /mu'i-un/ 'soak', /c'u'p'-un/ 'connect', /c'u'i-un?/ 'peel', /X,u'k°-un/ 'push out of the way', /nu'q>0-un/ 'poke', /XVc'-un?/ 'lever, block up', /Pu's-un?/ 'instruct, advise', /?u'X°-un?/ 'plane, dress, cut', /ju'c'-un?/ 'nudge, shove, push aside', /ju'X-un/ 'burn'; alternative stresses in /tu'j^un, tui2u'n/ 'lean'(cf.?/tui/ 'goacross'). The Type /CVCC-a'n/ 100. There are 29 examples (not counting the cases where the stem-final C represents a productive suffix); of these, 21 concern stems /CaCC/: /pJ9sk>0-a'n/ 'squeeze', /tmX-a'n/ 'paint', /t'mS-a'n/ 'braid', /clq'°-ain/ 'kick', /c'mq'-aty 'cut', /c'ls-a'n/ 'scour', /XmS'-a'n/ 'chip', /tojq'-ah?/ 'pinch', /&>k°X-ain/ 'fry', /&>X,q°-a'n/ 'shove

76

MORPHOLOGY

through a hole', /5'atq^-a'n?/ 'crumble', /qms-a'n/ 'pack', /qasn-a'n?/ 'miss', /q'lq'-a'n/ 'wind around', /XauX.'-a'n/ 'break', /Xlq'-a'n/ 'knock off, roll down', /q°3tq-a'n?/ 'pass by', /u3X.q°-a'n/ 'boil', /uac'q'-a'n?/ 'spread', /yaXq-a'n?/ 'chase away', /jpX'q'-a'n/ 'paint'. Here also belong /C'auPa'n?/ 'recite, perform', where /u?/ is treated as /CC/ (cf. /na'uPn/, where it is not). In /S'iS'-a'n/ 'twist' we have the root /5'i5\ CajC'/ 'id.; itr.' (see 16); the root of /siS'-a'n/ 'circle around' goes back to •/SajgV or */&>i5/ (see 45; cf. /sif-§i?5/ 'all around'); in /