UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Lexical phonology of Chilcotin Andrews, Christina 1988

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1988_A8 A52.pdf [ 4.32MB ]
Metadata
JSON: 831-1.0097945.json
JSON-LD: 831-1.0097945-ld.json
RDF/XML (Pretty): 831-1.0097945-rdf.xml
RDF/JSON: 831-1.0097945-rdf.json
Turtle: 831-1.0097945-turtle.txt
N-Triples: 831-1.0097945-rdf-ntriples.txt
Original Record: 831-1.0097945-source.json
Full Text
831-1.0097945-fulltext.txt
Citation
831-1.0097945.ris

Full Text

LEXICAL PHONOLOGY OF CHILCOTIN by Christina Andrews B.A. The University of Minnesota, 1980 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES (Department of L i n g u i s t i c s ) We accept t h i s thesis as conforming to the required standard THE UNIVERSITY OF June Christina BRITISH COLUMBIA 1988 Andrews, 1988 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Linguistics The University of British Columbia Vancouver, Canada DE-6 (2/88) 1 1 ABSTRACT This thesis analyses the native Indian language C h i l c o t i n through the use of the l e x i c a l phonology model. Data were col l e c t e d from fiv e speakers varying in age, d i a l e c t and sex. Chapter 1 discusses the segemental, tonal and s y l l a b i c systems in C h i l c o t i n . Chapter 2 i s a discussion of the vowel harmony process, f l a t t e n i n g . Chapter 3 i s an analysis of morphological rule formation and Chapters 4 through 7 present a discussion of the l e x i c a l and p o s t - l e x i c a l l e v e l s . C h i l c o t i n was found to be composed of three l e x i c a l levels and one p o s t - l e x i c a l l e v e l . 1 i i TABLE OF CONTENTS CHAPTER 1: INTRODUCTION 1.0 Introduction 1 1.1 Speakers Consulted ...2 1.2 The C h i l c o t i n Segmental System , 5 1.2.1 Consonants... 5 1.2.2 Vowels 9 1.2.3 The Syllable 10 1.2.4 Tone 10 1.3 Verbal Morphology 11 Footnotes for Chapter 1 14 CHAPTER 2: FLATTENING 2.0 Introduction 15 2.1 Flattening 15 2.2 Velar Harmony 18 2.3 Alveolar Harmony 23 2.3.1 Rightward Alveolar Harmony.. 23 2.3.2 Leftward Alveolar Harmony. 26 2.3.3 Summary 27 2.4 Theoretical Aspects of Harmony 28 i v 2.4.1 Rightward Alveolar Harmony 29 2.4.2 Leftward Alveolar Harmony 31 2.4.3 Velar Harmony 32 2.5 Summary 33 2.6 Comparison to Cook (1987) 35 Footnotes for Chapter 2 38 CHAPTER 3: LEXICAL PHONOLOGY 3.0 Introduction 40 3.1 The Lexical Phonology Model .40 3.2 Evidence for LP in Athapaskan 43 3.3 Morphological Rule Formation.... 46 3.4 Summary 58 Footnotes for Chapter 3 59 CHAPTER 4: LEVEL 1 4.1 Morphology 60 4.1.1 Stems 60 4.1.2 C l a s s i f i e r s 61 4.1.3 Person Markers 62 4.1.4 Conjunct.. 65 4.1.5 Mode 66 4.1.6 Derivative.. 72 V 4.1.7 Summary 75 4.2 Level 1 Phonology 76 4.2.1 Rules That Create a Level 1 D i s t i n c t i o n 76 Vowel Deletion I 76 i-Lower ing .. . . . 82 4.2.2 Rules That Apply at Levels 1 and 2 87 F r i c a t i v e Voicing 88 4.2.3 Rules That Only Apply at Level 1 90 D-ef feet 91 Continuant Coalescence ..94 Continuant Deletion 95 4.2.4 Vowel Rounding 100 4.3 Summary 101 Footnotes for Chapter 4 102 CHAPTER 5: LEVEL 2 5.1 Morphology....... 108 5.1.1 Derivative Affixes 108 5.1.2 Obviative 115 5.1.3 Subject Affixes ...116 5.1.4 Object A f f i x e s . . 118 5.2 Level 2 Phonology 119 5.2.1 Vowel Deletion II 119 5.2.2 Onset Formation 121 v i 5.2.3 F r i c a t i v e Voicing 123 5.2.4 Diphthongization 124 5.2.5 e-Raising 127 5.3 Conclusion... — 130 Footnote for Chapter 5 131 CHAPTER 6: LEVEL 3 6.1 Morphology 132 6.1.1 Durative 132 6.1.2 Stems 133 6.1.3 Adverbs. . . 133 6.1.4 Postpositions 134 6.2 Level 3 Phonology 136 6.2.1 Vowel Deletion II 136 6.2.2 Onset Formation 137 6.3 Summary 137 CHAPTER 7:P0ST-LEXICAL PHONOLOGY 7.0 Introduction 139 7.1 Nasalization 139 7.2 /gh/ Contraction 141 7.3 Velar Deletion 142 y i i 7.4 Onset Default 1 4 3 7.5 Summary 1 4 4 7.6 Summary of Thesis 1 4 4 LIST OF ABBREVIATIONS 1 4 7 REFERENCES 148 v i 1 i LIST OF TABLES Table I: Speaker Information P- 4 Table I I : C h i l c o t i n Consonants P« 6 ACKNOWLEDGEMENT I would l i k e to thank my commitee members: Dr. David Ingram, Dr. M. Dale Klnkade, and Dr. P a t r i c i a Shaw. I would also l i k e to thank the many students and v i s i t i n g f a c u l t y that lent a helping hand throughout the writing of t h i s t h e s i s : Bruce Bagemihl, Henry Davis, Heather Goad, Kathy Hunt, Helen L i s t , Diane Massam, Yves Roberge, Diane Rodgers, and Linda Walsh (for help s p e c i f i c a l l y on the morphology section). Last but c e r t a i n l y not least I would l i k e to thank my family for making this a l l possible. 1 CHAPTER 1: INTRODUCTION 1.0 Introduction This thesis presents a l e x i c a l phonology analysis of C h i l c o t i n , a language spoken in the southwestern area of B r i t i s h Columbia, Canada. It proposes that the theory of l e x i c a l phonology (Kiparsky 1982, 1983), hereafter LP, provides an i n s i g h t f u l description of selected complex phonological and morphological processes found in C h i l c o t i n . The present chapter begins the presentation with an introduction to C h i l c o t i n . It provides an overview of the native speakers consulted, and the C h i l c o t i n segmental system, tones, s y l l a b l e structure, and verbal morphology. Chapter 2 discusses the phenomenon of f l a t t e n i n g , a vowel harmony process ( c . f . Cook 1983), and provides a t h e o r e t i c a l account for i t within LP. Although t h i s chapter deals with a subset of the ove r a l l phonology, I have ordered i t before the broader phonological discussion since an understanding of fl a t t e n i n g is necessary for the l a t t e r . Chapter 3 begins the analysis by giving an overview of LP. This is followed by the main body of the thesis in Chapters 4 through 7. These outline how the rules of C h i l c o t i n phonology operate at each of three l e v e l s , and p o s t l e x i c a l l y . 2 1.1 Speakers Consulted C h i l c o t i n is an Athapaskan language^ which is is a member of the larger Na-Dene language family. As B r i t i s h Columbia's southernmost Athapaskan language, C h i l c o t i n is surrounded primarily by Salishan languages and is bordered by L i l l o o e t and Shuswap. Its only neighboring Athapaskan language is Car r i e r . The data were collected in Vancouver during 1984 and 1985 from five native speakers who participated in varying degrees in a f i e l d methods class on C h i l c o t i n . (Here each of the fi v e speakers is assigned a l e t t e r A-E.) Additional data were subsequently obtained from two of the speakers (A and B). The speakers ranged in age from a young man in his teens (C) to a grandfather in his 50's (D). Speakers C, D and E had only recently l e f t the reserve, while speakers A and B had l i v e d away from the reserve for quite some time. Speaker A claimed to be also fluent in Carrier as well as having some knowledge of French and German. He had worked as a translator in Canadian legal courts for monolingual C h i l c o t i n s , and also had some previous experience teaching C h i l c o t i n to an English speaker. With the exception of the eldest man (D), a l l of the speakers had some t r a i n i n g at school in reading and writing C h i l c o t i n . In fact, speakers A and C often commented on the phonetic t r a n s c r i p t i o n . A l l of the speakers were fluent in English. Due to the differences in age, area, and time away from the reserve there is v a r i a b i l i t y among the speakers for many forms. 3 T h i s i s p a r t i c u l a r i l y n o t i c i a b l e i n the f l a t t e n i n g p r o c e s s . There a r e many i n s t a n c e s where s p e a k e r s do not f l a t t e n vowels t h a t ' f a l l w i t h i n the scope of the f l a t t e n i n g r u l e s (see Chapter 2 f o r r u l e f o r m a t i o n ) . These examples w i l l be marked (FL. IRR.). T a b l e I p r o v i d e s a summary of i n f o r m a t i o n about the f i v e s p e a k e r s c o n s u l t e d . 4 Table I. Speaker Information Speaker Sex Age Area Languages Known M Late 20*s Anaham English,Carr i e r , C h i l c o t i n , some German and French B F Early 20*s Nemiah Valley English, C h i l c o t i n M Late teens Nemiah Valley English, C h i l c o t i n M Late 50's Alexis Creek English, C h i l c o t i n F Early 20's Nemiah Valley English, C h i l c o t i n 5 1.2 The C h i l c o t i n Segmental System In t h i s section, I w i l l discuss the segmental system of C h i l c o t i n (consonants, vowels, and tones), and give a b r i e f overview of the C h i l c o t i n s y l l a b l e . 1.2.1 Consonants. C h i l c o t i n has a r i c h consonant system, as shown in Table I I . The inventory i s comprised of stops, f r i c a t i v e s , a f f r i c a t e s , nasals, l i q u i d s ^ , and g l i d e s . The stops and a f f r i c a t e s appear in three orders--plain, aspirated, and g l o t t a l i z e d , and w i l l be transcribed following t r a d i t i o n a l Athapaskanist orthographic practices. The p l a i n unaspirated series is represented by the symbol usually reserved for voiced consonants. The aspirated series i s represented by those usually reserved for voiceless segments, while the ejectives are transcribed in conformity to the International Phonetic Alphabet. This practice i s exemplified below for the dento-alveolar s e r i e s : Phonetic Orthographic t d t h t f t' The C h i l c o t i n consonant system exhibits f i v e places of a r t i c u l a t i o n : l a b i a l , dento-alveolar, alveopalatal, velar and g l o t t a l . Both the dento-alveolar and velar series contain 6 T a b l e I I . C h i l c o t i n c o n s o n a n t s . F l a t consonants are i n d i c a t e d by a r a i s e d ~, and f o l l o w t h e i r s h a r p c o u n t e r p a r t s L a b i a l Dento- A l v e o - V e l a r G l o t t a l a l v e o l a r p a l a t a l Unround Round Stops p l a i n a s p i r a t e d g l o t t a l i z e d A f f r i c a t e s t ' 9/9 k,k gW / g W kw, kw k'^k' k'w p l a i n a s p i r a t e d g l o t t a l i z e d F r i c a t i v e s v o i c e l e s s v o i c e d N a s a l s G l i d e s m d z , d z , d l t s , t s , 1 4 s, s z, z 3 n d3 t f t s ^ t i ^ t * ' t j ' gh w fW L a t e r a l s v o i c e l e s s v o i c e d 4 1 7 further subdivisions determined by t h e i r type of release. In the dento-alveolar series, a f f r i c a t e s are divided into two groups determined by t h e i r type of r e l e a s e — l a t e r a l or s i b i l a n t : (1) Lateral Release S i b i l a n t Release a. ?isdlos 'sleigh' E c. bidzax 'spittoon' A b. t^owesan 'snake' E d. t s a i 'head' A The velar consonants are also subdivided into two groups-rounded and unrounded: (2) Rounded Unrounded a. g_wih ' s h i r t ' E c. gaendzaz 'elbow' A b. k_wik_wih 'Canada Jay' E d. kantsai 'basket' E One of the most s a l i e n t subdivisions in Table 2 is the flat-sharp d i s t i n c t i o n found in the dento-alveolar and velar s e r i e s . In the dento-alveolar s e r i e s , the d i s t i n c t i o n i s r e s t r i c t e d to s i b i l a n t s and a f f r i c a t e s with s i b i l a n t release. The f l a t consonants are characterized by a s l i g h t l y retracted a r t i c u l a t i o n . Latimer (1978) has found through spectrographic analysis that the f l a t series exhibits a locus increase of 500 Hz. Some examples in phonetic t r a n s r i p t i o n are: 8 (3) Sharp F l a t /s/ sae 'sun' B /s/ tas 'bed' B /z/ d iz ' s i s t e r ' A /z/ zai 'belt' C / t s / di4tsaen 'blue' E / t s / tsa 'beaver' E / t s ' / t s ' i boat /dz/ dz^ndi ' a l l day' E / t s ' / ts'ay 'plate' A /dz/ dzat 'tobacco' A These differences, however, are d i f f i c u l t even for "a seasoned Athapaskanist to d i s t i n g u i s h " (Cook 1983). However s l i g h t the phonetic difference may be, I w i l l argue in Chapter 2 that the phonology of C h i l c o t i n d e f i n i t e l y manifests a vowel harmony process of f l a t t e n i n g triggered by f l a t consonants. The velar stops, divided into rounded and unrounded, are further subdivided into sharp and f l a t , where f l a t velars are marked phonetically by a heavy f r i c a t i v e o f f g l i d e and a s l i g h t l y uvular a r t i c u l a t i o n , as seen in the following data: (4) Sharp /k/ tselEk^ 'coyote' E /RW/ kwikwih 'Canada Jay' E /g/ 9eg_uh 'when' A /gw/ tsainaeg_wi 'hat' B Fl a t /£/ sa gXai 'my foot* C /fc"w/ hunitt'awgxwa 'why' D /g/ nesgey 'I walked* E /g w/ naeg_wayltaeno 'did i t rain?' E The f l a t velars are a l l marked by a heavy o f f g l i d e , and the rig h t hand examples in (4) have a retracted uvular 9 a r t i c u l a t i o n . 1.2.2 Vowels. The vowel inventory is comprised of six underlying sharp vowels and s i x allophonic f l a t vowels ( i . e . retracted). The presence of the f l a t vowels is e n t i r e l y predictable in that they occur only when a f l a t consonant i s present within the word (as w i l l be discussed in further d e t a i l in Chapter 2). The vowel q u a l i t i e s with respect to height and backness are sketched below: Sharp F l a t i v u e 6 o i ai ae The sharp-flat correspondances are as follows: i > e or a i , i and e > a, ae > a, u > o, u > o. ae,a,u,o,e, and i are non-central and a,B,v,6 and i are ce n t r a l i z e d . Cook (1986) labels these l a t t e r two groups as f u l l and reduced respectively. Both sharp and f l a t vowels have nasal vowels as allophones: (5) a. / i n 4 i / [ i 4 i ] 'one' E b. /gunzun/ [guzun] ' i t is good' A c. /kongh/ [ke>] 'house' E d. /naen4cngh/ [naiEy] 'horse' E 10 1.2.3 The Sy l l a b l e • I follow Cook (1986) in assuming there are three core s y l l a b l e types: CV, CVC, and CVC where C is any consonant, V i s a f u l l vowel, and V is a reduced vowel. In the examples below, 1.' stands for a s y l l a b l e boundary, and the relevant s y l l a b l e is underlined. (6) CV ? i n . t s i 'grandfather' E tsae 'beaver' E se.k'i •cow* E CVC kWin.kWjh 'bluejay' E yet 'breastbone' A dsas 'fishhook' A CVC may 'berries' A 1 in 'dog' B ?as.Kay ' c h i l d ' A Complex onsets occur in my data very infrequently. The only example I have i s [sdley] 'pants'. 1.2.4 Tone. High ('), low (*)> and mid tones (which w i l l be unmarked) are found in C h i l c o t i n , as in tko] 'house* E, [d l i k ] ' s q u i r r e l ' E, and [yat] 'breastbone' E. Mid tones are derivable (as in Roberge 1985) from the (underlying) high and low tones. A phonemic high tone lowers to a mid tone when i t occurs to the right of a voiced consonant as shown in the pairs below. 11 (7) a. i . / d i / --> t d i ] 'horn' E 11. / k ' i / — > [ k ' i ] ' w i l l o w ' E b. i . / z i / ~ > [ z a i l 'mouth 1 E i i . / s i / — > [ s a i l ' b e l t ' E A phonemic low tone r a i s e s t o a mid tone when i t o c c u r s a f t e r a v o i c e l e s s consonant. (8) a. i . / d e - p a e t / --> t d e - p a e t ] ' i t i s f l a t t e n e d ' E i i . [ b a e t ] ' m i t t e n s ' E b. i . / J e n / — > t f e n ] 'song' E i i . [nen] 'back' E Of the c o n t o u r t o n e s , r i s i n g tone (mid t o h i g h ) i s found o n l y i n q u e s t i o n s , and f a l l i n g tone (mid t o low) i s found i n s y l l a b l e s t h a t c o n t a i n f u l l v o w e l s , n a s a l c o n s o n a n t s , and g l i d e s . I w i l l not be i n d i c a t i n g tone i n t h i s t h e s i s as i t s i n t e r a c t i o n s w i t h the segmental phonology i s not f u l l y u n d e r s t o o d . See Roberge (1985) f o r f u r t h e r a n a l y s i s . 1.3 V e r b a l Morphology T h i s t h e s i s d e a l s m a i n l y w i t h the phonology which a f f e c t s v e r b stems and t h e i r v e r b a l morphology. The m i n i m a l C h i l c o t i n 12 sentence contains one word which is composed of a series of prefixes b u i l t up on a verbal stem. A template of these prefixes i s as follows (also see chart at the end of the thesis for a summary of the abbreviations used): post - adv - stem - dur # obj - subj - oby - derv % derv - mode I 2 3 4 5 6 7 8 9 1 0 conj - pm - c l s f - stem II 12 13 14 (# and % indicate d i f f e r e n t boundaries between a f f i x e s which w i l l be discussed in d e t a i l in Chapter 3). As we can see in the s t r i n g above, these a f f i x e s provide information regarding subject (6,12), object (5,7), and mode and aspect (4,10,11). There are also a f f i x e s of an adverbial (2,8,9) and postpositional (1) nature. Some examples of th e i r occurrence are given below. ( A l l of these examples are from speaker A). (9) a. t a - zai - s - t a e l 'I kicked' derv mode pm stem b. naen - d3e - t i - 1 - bin 'they swam away from you 1 post subj derv c l s f stem 13 d3l - yo - ghe - n - ta e l 'they kicked i t (many times)' subj obv mode conj stem ho - ghe - n - t s i 'you shot i t (many times)' derv mode pm stem na - gha - h - d - zun 'you (pi) are good again' adv mode pm c l s f stem 14 Footnotes for Chapter 1 1. Also spelled Athabaskan. 2. Phonologically, the l a t e r a l s /4/ and / l / class themselves with the f r i c a t i v e s in regard to the f r i c a t i v e voicing rule whereby voiceless continuants become voiced i n t e r v o c a l i c a l l y . Here 4 > 1. 3. There i s a variant of z, which occurs in s y l l a b l e f i n a l p o sition, that has some 1 q u a l i t y to i t . This is probably a borrowing of a similar sound in Thompson. 15 CHAPTER 2: FLATTENING 2.0 Introduction This chapter deals with the vowel harmony process of f l a t t e n i n g . It precedes the general chapter on phonology, as an understanding of the f l a t t e n i n g process i s necessary to the discussion of the vowel mutation rules found in Chapters 4-7. Latimer (1978) argues that the t r a d i t i o n a l term f l a t t e n i n g should be replaced by a more phonetically descriptive term. He suggests the name 'retracted tongue root' or RTR. Here I w i l l r e t a i n the t r a d i t i o n a l term as a general name, but w i l l refer to RTR in the more formal description of the r u l e . 2.1 Flattening Flattening i s a harmony process triggered by the f l a t consonants-*- (see Table 2). As a f i r s t approximation, we can say that sharp vowels are retracted to their f l a t allophones when a f l a t consonant i s present in the word. (Flattening does not extend beyond word boundaries). The allophonic vowel variations are repeated below as follows: 16 (1) Sharp F l a t i a i ,e 2 E a z a ae a u o u v o Before discussing the d e t a i l s of f l a t t e n i n g , I would l i k e to discuss how examples w i l l be presented in the chapter. Words w i l l be shown in t h e i r phonetic forms, followed by t h e i r underlying representations between slant l i n e s . In phonetic forms, a l l f l a t consonants and flattened vowels w i l l be underlined. In the underlying forms, however, I w i l l only underline those vowels and consonants which are involved in the aspect of f l a t t e n i n g being discussed. In each of the examples below, (i) gives an example of a flattened vowel and ( i i ) an example with the corresponding non-flat allophone. The f l a t vowel occurs in each case because of the presence of a f l a t consonantal trigger within the word. The type of trigger and consequent spread of harmony w i l l be discussed s h o r t l y . 17 (2) Some Examples of Flattening a. i —> ai i . [ s a i - t a i n ] / s i - t i n / 'he i s sleeping' C perf stem 10 14 i i . [ta-gha-ti4] /tae-ghe-t14/ 'we are going to sleep' B derv inc stem 8 10 14 b. i --> e i . t t s ' i de-nazl / t s 1 i di_-nez/ 'the boat i s long* A 'boat' derv stem 9 14 i i . [ t s ' i di_-n-di] / t s ' i d i ^ n - d i / 'the boat is short' A 'boat' derv conj stem 9 11 14 c. e--> a i . [na4ey sa-s - t a i ] /na4ey si-s-t£4/ 'I kicked the horse* B 'horse* perf pm stem 10 12 14 i i . [ta-ghi-n-te4] /tae-ghi-n-te4/ 'he kicked into a group'B derv inc conj stem 8 10 11 14 d. ae --> a i . [ta-za-h-tal1 / t e - s i - e h - t a e l / 'you (pi) kicked* A derv perf pm stem 9 10 12 14 i i . [ta-zai-n-tael1 / t e - s i - i n - t a e l / 'you kicked' A derv perf pm stem 9 10 12 14 18 e. u --> o i . [go-ze-t'in] / g u - s i - i d - t ' i n / 'we want to s i n g 1 B derv perf pm stem 8 10 12 14 i i . [4aen gu-lin] /4-aen g u - l i n / 'there i s a l o t ' D 'a l o t ' derv stem 8 14 f. — > o i . I - d l o x ] / d l u x / •to laugh .3 stem 14 2.2 Velar Harmony Vowels may be flattened by the following velar consonants: / § / 9W/ £ w / Jc', x, xw, gh/. The extent to which the f l a t t e n i n g harmony spreads is dependent upon the type of trigger involved. An alveolar t r i g g e r , subject to certain r e s t r i c t i o n s to be discussed below, can a f f e c t vowels up to two s y l l a b l e s away. In [na-ta-kasj ' i t i s slowly turning' B, a l l three vowels have succumbed to the f l a t t e n i n g process. In (2) a . i , b . i , c . i , and d . i above, a l l of the vowels in the word have also flattened. The velar t r i g g e r , however, can a f f e c t only vowels in adjacent s y l l a b l e s . This can be seen by contrasting (2)b.i and a . i i above. In example (2)b.i, the dento-alveo /z/ has flattened two vowels to i t s l e f t : / t s i di-nez/ > [ t s ' i de- nazl 'boat is long'. However, the velar t r i g g e r , tghj, in ( 2 ) a . i i can only f l a t t e n the immediately 19 adjacent vowels— [a] to the l e f t and le] to the r i g h t , r e s u l t i n g in [ta-gha-ti4-] . We can see the process at work also in example (2)c. The velar trigger in ( 2 ) c . i i has flattened only the adjacent vowel t a e l , leaving the stem vowel as sharp [e], whereas in (2) c . i t h i s vowel has flattened to [al because the trigger was from the dento-alveolar s e r i e s . Other examples of the limited spread from velars are given below. In each case only vowels in immediately adjacent n u c l e i ! have flattened. In f.and g. the f l a t t e n i n g process has spread leftward over the neutral consonant [4>] and the [ -RTR] [ s ] . (3)a. [Jcanxh] /Jcaenxh/ 'spoon' E b. Iho-ghe-n-tsi1 /hu-ghi-n-tsi/ 'he shot i t repeatedly' B derv ser conj stem 8 10 11 14 c. [hu-ta-gha-tsax] /hu-tae-ghc-tsaex/ 'we are going at i t ' B derv derv inc stem 8 8 10 14 d. [ x w a - *?intsi] / x ^ e - ^ i n t s i / 'our grandfather' E poss stem e. Cne-s-gey] /ne-s-gi/ 'I walked' E perf pm stem 10 12 14 da-ne-4-jcat 1 / d e - n i - l - l c E t / 'he broke the chair in h a l f A derv perf c l s f stem 8 10 13 14 f. [betJet chair 20 g. [ b E t f c t da-na-s-£at] /de-ni-s-4-lcet/ 'I broke the chair in h a l f A chair derv perf pm stem 8 10 12 14 h. [midoj /midugh/ 'whitertian' E i . [ b i l o ] /bilugh/ 'knife' E j . [i4a] /in4aengh/ 'one' E k. [diya] /diyaengh/ 'man' E In the l a s t four examples, additional rules have applied to give the correct surface form, a nasalized vowel. See section 7.1 for further discussion. If we compare the data above to the following data, we can see that i t i s indeed the distance from the trigger to the p o t e n t i a l vowels which prohibits the f l a t t e n i n g , rather than the manner of a r t i c u l a t i o n of the velar t r i g g e r . The examples below indicate the a b i l i t y of each type of velar consonant .(stop, f r i c a t i v e , and labiovelar) to trigger f l a t t e n i n g in adjacent vowels. The only occasional exception to the harmony process i s the vowel [ i l . This could be due to the fact that [ i ] i s a front high vowel and in order to a r t i c u l a t e i t the tongue root has to return to a more neutral position thus stopping the r e t r a c t i n g process. In example (4)a.i, the f i r s t t i ] does not f l a t t e n to [e] or [ai] although i t i s within the domain of velar harmony (see also Latimer 1978 and Cook 1983). 21 (4) Stops a. i — > a i i . [ni-jcain] /ni-£in/ 'we paddled' E perf stem 10 14 i i . [?as£aiU /es£i/ ' c h i l d ' E i i i . [sa-gaij / s t h / 'my foot' C i v . [yajcaik] /yaekik/ ' b a l l ' B b. ae —>a, u—> o i . [sakot] /saefcut/ 'shadow' A i i . [yakox] /yaekux/ ' r i v e r ' A (5) F r i c a t i v e s a. ae —>a i". [ nae-ta-gjva-gh was ] /nae-tae-ghe-s-gh wes/ obj der inc pm stem 5 8 10 14 'I am going to t i c k l e you' A b. e --> a i . [gax] /gex/ 'rabbit' A (6) Labiovelars a. e —> a 22 i . [nae-ta-gha-s-gh was ] /nae-tae-ghe-s-g_h_wes/ obj derv inc pm stem 5 8 10 12 14 •I am going to t i c k l e you' A b. i —> e i . [bi-ta-s-k wey / b i - t e - s - k ^ i y / post derv pm stem 1 9 12 14 'I am vomitting' A In t h i s l a s t example i t i s interesting to note that velar f l a t t e n i n g , a l o c a l process, has spread over the sharp /s/ To describe the facts of velar f l a t t e n i n g , a l l that i s needed is a l o c a l rule of the type found in Anderson (1974), where the percentage mark can be interpreted "on either side of" the environment. (An autosegmental analysis w i l l be discussed in section 2.4.3). I Velar Flattening V --> [+RTR] % C f-anterior] [-coronal] [+RTR] Vowels become f l a t when they are in a nucleus that i s on either side, and immediately adjacent to, f l a t (or (+RTR1) velar consonants. 23 2.3 Alveolar Harmony As has been noted elsewhere (Latimer 1978, Cook 1983, Krauss 1975), alveolar-triggered harmony consists of two d i s t i n c t processes. One process f l a t t e n s to the right and is subject to several blocking r e s t r i c t i o n s . The other process f l a t t e n s in a leftward d i r e c t i o n and i s unbounded in i t s spread. I w i l l f i r s t discuss the rightward process, as i t is more l i m i t e d . 2.3.1 Rightward Alveolar Harmony. Rightward harmony does not spread beyond the vowel [ i ] (although i t does i t s e l f f l a t t e n ) , and does not extend beyond sharp consonants. (Sharp designates those consonants that have a f l a t counterpart: k, g, k', k w, g w, k , w , s, z, t s , dz, t s ' ) . The set of data below i l l u s t r a t e s the i n a b i l i t y of rightward harmony to f l a t t e n beyond the vowel [ i ] . In the examples in 7, the [ i ] has flattened to [ ai ] or [e), but has blocked any further spread of the harmony. (7) a. [go-ze-t 1 in] /gu-si^-id-t • i n / 'we want' A derv perf stem 8 i 10 14 b. [sai-daeh] /si-daeh/ 'I am s i t t i n g ' C perf stem 10 14 c. [sai-deh] / s i - d E h / 'I am l i v i n g (with him)' B perf stem 10 14 24 d. [na-sai-dsit) /nae-s^-dsi t / 'I crawled' B dur perf stem 4 10 14 e. [ta-zai-n-tael] / t e - s i - i n - t a e l / 'you kicked' A derv perf pm stem 9 10 12 14 f. [na-se-bin] /nae-si-bin/ 'I swam' A dur perf stem 4 10 14 (I cannot account for the missing person markers in f and d). A l l of the above data have the perfective marker / s i / as the blocking element. I have no other examples of a f l a t /%>/ or a sharp / i / in a preverbal a f f i x because of the limited type of a f f i x e s a v a i l a b l e . However, the incorporated stem (position 3) is a possible source of an additional /s/. The following data i l l u s t r a t e the blocking e f f e c t of sharp consonants in rightward harmony. (Here both the f l a t consonant which causes f l a t t e n i n g and the sharp one which blocks i t are underlined in the underlying forms). (8) a. [ba-zak'i] /be-sek«i/ 'his cow' E poss stem b. [^a-ta-za-s-ts'i] / ^ e - t e - s i - s - t s * i / 'I shot' obj derv perf pm stem 5 9 10 12 14 c. (nae-g_wa-da-z-k'aen] /nae-gwE-di-£-k_*aen/ adv obj derv perf stem 2 5 8 10 14 In the data above, tk'J and [s] have blocked the harmony, so that the vowels occurring to the right of the sharp 25 consonants remain sharp. In (8)b and c, a rule for i-lowering (p.80) has f i r s t applied, lowering the / i / to [e] before the consonant I s ] . Since the / i / has lowered, i t does not block the f l a t t e n i n g process. (The d e t a i l s of i-Lowering w i l l be discussed in section 4.2.1). Another s i t u a t i o n where / i / does not block f l a t t e n i n g is when i t is deleted before f l a t t e n i n g occurs. The following examples i l l u s t r a t e t h i s , where / i / deletes due to the existence of the following /e/ (see section 4.2.1 for further discussion). The f l a t t e n i n g process in these examples is allowed to spread to the end of the word. This shows that i t is indeed the presence of t h i s vowel that blocks rightward spread of f l a t t e n i n g . (The / i / s which delete are underlined in the underlying forms). (9) a. [ go-za-h-t' en ] /gu-s i_-eh-t * i n / 'you (pi.) want' A derv perf pm stem 8 10 12 14 b. [ ta-z-ah-ta_l ] /te-s i^-ch-tael/ 'you (pi.) kicked' A derv perf pm stem 8 10 12 14 Due to the nature of rightward harmony, a segmental rule s i m i l a r to Velar Flattening in section 2.2 would be very encumbered. It would have to state that the harmony proceeded unless either a [-RTR] consonant or the vowel [ i ] was present, in which case the vowel would become [+RTR]. 26 2.3.2 Leftward Alveolar Harmony. Leftward harmony is unbounded, capable of spreading beyond both the vowel [ i ] and sharp consonants. The following data i l l u s t r a t e that leftward harmony spreads as far as possible within the word—even up to three s y l l a b l e s away from the t r i g g e r . (10) a. [ d a - t a - z a - s - t s ' E t ] /dae-te-zi-s-ts«et/ 'I f e l l down' derv derv perf pm stem 8 8 10 12 14 (this [dae] a f f i x could be either derv or adv) b. NP [da-na-l-daz] /dE-nae-l-dez/ 'ground meat' A derv derv c l s f stem 8 8 13 14 c. i . [na-na-ta-zai-tin 1 /nae-nae - t E-s^i-tin/ adv dur derv perf stem 2 4 8 10 14 'I started dreaming again' B i i . [nae-na-ta-zai - tin 1 (same as above, spoken slowly) (note: f i r s t [nae] did not flatten) Leftward harmony i s not blocked by the vowel / i / , as the following data i l l u s t r a t e , where the f l a t t e n i n g process has spread beyond / i / . (The underlying flattener and the / i / to i t s l e f t are underlined). (11) a. [sa-deyaz] /se-di_yEz/ 'my male' A poss stem b. [na-de-naz] /nae-di_-nez/ ' i t i s long again' A adv derv stem 2 8 14 27 c. [na-de-s-k 'aen] /nae-di^-s_-k *aen/ ' i t is burning again' E adv derv perf stem 2 8 10 14 d. [ya-te-z-l-tjTos ] /ye-t i_-Z-l-t Jus/ 'he i s holding i t ' obv derv perf c l s f stem (from Cook 1983) 7 8 10 13 14 e. tna-ne-£-tl'on J /nae-ni^-s-tl' un/ 'fence' 4 Nor i s leftward harmony blocked by sharp consonants, as shown in (11). (11) f. [na-ta-k'a£] /nae-tae-k'ES/ ' i t i s slowly turning' B dur derv stem 4 8 14 g. [tanantilk'az] /taenentilk'ez/ 'water s t a r t s to get c o l d ' 5 In both these cases, the harmony has spread over a sharp [k'J. Examples ( l l ) a and (10)c.i show that the harmony may spread as far as the adverbial a f f i x (position 2) and the durative a f f i x (position 4). As seen in ( 1 0 ) c . i i , however, the process tends to weaken as i t moves further and further away from the tr i g g e r , leaving the leftmost vowel as sharp when the word is spoken slowly. 2.3.3 Summary. Alveolar harmony has two d i s t i n c t d i r e c t i o n a l processes. When the harmony spreads to the right , i t i s blocked by sharp consonants and the vowel / i / , although / i / w i l l i t s e l f f l a t t e n to [ai] or [e]. Leftward alveolar harmony, on the other hand, is not blocked by either sharp 28 consonants or the vowel / i / . / i / , however, i s somewhat re s i s t a n t to r e t r a c t i o n . 2.4 Theoretical Aspects of Harmony Having described the b i d i r e c t i o n a l harmony process in C h i l c o t i n , I now turn to the t h e o r e t i c a l mechanisms available to describe alveolar harmony. Kiparsky (1985) suggests that a l l harmony systems can and should be handled by the autosegmental framework f i r s t developed by Goldsmith (1976). Kiparsky claims that 'the autosegmental theory is correct and requires no metrical supplementation for harmony processes' (1985: 68). He wants to r e s t r i c t the use of the metrical framework, as developed in Hayes (1980), to stress rules, because the h i e r a r c h i c a l metrical trees imply 'the capacity for global large-scale transmissions of phonological information' (1985: 2). Kiparsky further claims that implementing autosegmental harmony rules within the l e x i c a l phonology model, which we have independent motivation for in the grammar of English, e n t i r e l y does away with the necessity of using both metrical and autosegmental frameworks to analyse what had been e a r l i e r described as dominant versus d i r e c t i o n a l harmony (Kaye 1982; Halle and Vergnaud 1981). By using autosegmental rules within the l e x i c a l phonology model, we no longer need the powerful mechanism of metrical structure to describe harmony systems. Such a decision w i l l 29 s i m p l i f y the phonological component of many languages. As LP ( l e x i c a l phonology) is independently motivated for C h i l c o t i n (as w i l l be shown in Chapter 4), I follow Kiparsky in implementing an autosegmental approach to vowel harmony within the LP model. 2.4.1 Rightward Alveolar Harmony. As noted e a r l i e r , the vowel / i / and a l l sharp consonants block the spread of rightward harmony. I assume that they are linked to a t-RTR] (retracted tongue root) autosegment. The trigger of the harmony process is a f l a t consonant which has attached to i t a [+RTR] autosegment as shown below.^ (12) a. [+RTR][-RTR] [na-se-bin] /nae-s-i-bin/ 'I swam' A b. [+RTR] [-RTR] I I [go-z-ah-ten1 /gu-z-eh-tin/ 'you (pi.) want i t ' A c. [+RTR] [-RTR] [ 9 a - t a - z a - s - t s • i ] / ? e - t e - z e - s - t s • i / 'I shot' B (I assume that the Obligatory Contour P r i n c i p l e w i l l collapse the three t-RTR] features on the [ s i , [ts] and [ i l into one.) The feature [+RTR] i s allowed to spread f r e e l y under the r e s t r i c t i o n of Goldsmith's Well Formedness Condition (WFC), which states that there may not be any crossing autosegmental 30 l i n e s . This produces the phonetic forms shown in (12). The spread of f+RTR] i s blocked by the [-RTR] autosegments. If the [+RTR] were to spread beyond the linked [-RTR], crossing l i n e s would be created, thus v i o l a t i n g Goldsmith's WFC. I assume that i n the case of / i / —> [ai] the [+RTR] feature attaches to the prelinked [-RTR] of / i / and creates a branching RTR feature node as shown below. 7 [+RTR] [-RTR] i The [+RTR] i s r e a l i z e d as the onglide [a) in [ a i ] . I assume there is an optional low le v e l phonetic rule that w i l l change [ai] to [e]. The following rule can thus be written for rightward harmony. Rightward Harmony Spread [+RTR] to the r i g h t . For those vowels that remain unlinked, I assume a default rule that inserts the feature [-RTR1 as in (13). 31 (13) a. I-RTR) / - > [ g u w e n ] / g u w E n / •after' A [ ] [-RTR] b. /t/uh/ > t t f u h ] 'and' D [-RTR] / \ c. /be - yet/ —> [be - yet] 'his breastbone' E Vowels (except / i / ) are unmarked for the feature [RTR] on the segmental t i e r and w i l l receive t h i s feature only through the default r u l e . 2.4.2 Leftward Alveolar Harmony. Leftward alveolar harmony, as shown e a r l i e r , i s not blocked by the [-RTR] autosegments of [ i ] or sharp consonants. To prevent crossing autosegmental l i n e s , I assume the following context sensitive delinking rule (as suggested by Dr. P a t r i c i a Shaw). [-RTR] Delinking x x [-RTR] [+RTR] This accounts for the fact that the l i g h t diphthong [ai] (which has a [-RTR] feature attached) is never found to the l e f t of a f l a t t r i g g e r . 32 Leftward Harmony Spread t+RTR] on the vowel t i e r Here are some examples of how t h i s rule operates : (14) [-RTR] [+RTR] f+RTR] a. /nae-tae-k'es/ —> nae-tae-k'aes —> [na-ta-k'as] 'It i s slowly turning' B (-RTR] t+RTR] [+RTRJ b. /taenentilk•es/ --> taene'ntilk ' es —> [ tanantelk ' as 1 •water s t a r t s to get cold' (Latimer 1978: 9) [-RTR][+RTR] [+RTR] c. /nae-di-nez/ —> nae-di-nez/ —> [na-de-naz3 ' i t i s long again' A 2.4.3 Velar Harmony. Velar Harmony can also be written as an autosegmental rule, with the r e s t r i c t i o n that i t spreads only to immediately adjacent vowels. Velar Harmony 1) Spread [+RTR] b i d i r e c t i o n a l l y to a maximum of one nucleus in either d i r e c t i o n . 33 Some examples are given in (15). (15) t+RTR] [+RTR] I l \ a. /fcaenxh/ —> Jc aenih —> [kanih] 'spoon' E [+RTR] [+RTR] I / |\ b. /de-ni-<i-k Net/ — > de-ni-4-k* et — > tde-ne-Rat] •he broke i t i n h a l f A [+RTR] [+RTR] I / | \ c. /se-saeRut/ --> se-saek" ut — > [se-salcot] 'my shadow 1 A 2.5 Summary The f l a t t e n i n g phenomena are summarized below. General Description Velar Harmony - a f l a t velar consonant f l a t t e n s immediately adjacent vowels. Alveolar Harmony Rightward - a f l a t alveolar consonant f l a t t e n s vowels to the r i g h t but i s blocked by t i l and sharp consonants. Leftward - a f l a t alveolar consonant f l a t t e n s vowels to the l e f t . This process is unbounded. 34 Autosegmental Representation 1) A l l f l a t consonants are linked to a [+RTR] autosegment. 2) The vowel [ i j and a l l sharp consonants are linked to a [-RTR] autosegment. 3) There i s a context sensitive delinking rule whereby the sequence [-RTR] [+RTR] is not allowed and the [-RTR] delinks. Rule Formation** Velar Harmony Rule Spread [+RTR] b i d i r e c t i o n a l l y to a maximum of one nucleus in either d i r e c t i o n . [RTR] Delinkink x x [-RTR] [+RTR] Alveolar Harmony Rule (subject to the WFC) Spread [+RTR] 35 Domain of RTR Harmony None of these rules apply across word boundaries as shown below. (16) a. [se tsi^ £ Wat) »my knee' A b. [ba d3a| t l ' u l ] 'his f i s h hook1 A c. [ni tsi_ zaz] 'deer skin' A Therefore a l l three harmony rules are assigned to l e v e l 1 (Strong Domain Hypothesis) and do not shut off u n t i l the p o s t - l e x i c a l l e v e l is reached. 2.6 Compar ison to Cook (1987) The main arguements in t h i s section were written before the publication of Cook's 1987 a r t i c l e . In t h i s section I w i l l discuss the differences between my data and that in Cook (1987) . There are two main differences. The f i r s t has already been mentioned. In my data both Iai] and [e] occur to the rig h t of a tr i g g e r , as shown below for each speaker (there was i n s u f f i c i e n t data for speaker C). 36 (17) a i . [ k , w a s se_-lin] 'I have a cold' A cold perf stem 10 14 i i . [sa-zai] 'my mouth' A poss mouth b i . [yo-ghe-n-tsi] 'he shot i t (many times) ' B obv perf conj stem 7 10 11 14 i i . [sai-daehl 'I am s i t t i n g 1 B perf stem 10 14 c i . [ghe_-li] ' i t was' D perf stem 10 14 i i . [nae-sai-tin] 'I dreamt' D (FL. IRR.) dur perf stem 4 10 14 37 d i . [del sa-leyn] ' i t was bloody' E blood perf stem 10 14 i i . ( a s k a U ' c h i l d ' E Cook reports only [ai] to the r i g h t of the trigger and [e] to the l e f t (I also have only [e] to the l e f t ) . The second main difference i s Cook's account of s i b i l a n t harmony whereby "non-neutral s i b i l a n t s assimilate to the rightmost non-neutral s i b i l a n t " (pg.56). This serves to block leftward harmony. Cook thus analyses the [+RTR] as a f l o a t i n g feature. I do not have any data that exemplify this,however. There are many basic s i m i l a r i t e s in the analyses. Cook has 2 main processes; alveolar and velar. He uses [+RTR] as the spreading segment and [-RTRJ as the blocking feature. Neutral consonants are attached to neither. Velar harmony i s much more limited in scope and can f l a t t e n only neighboring vowels. 38 Footnotes for Chapter 2 1. This type of vowel harmony, where the trigger i s a consonant, has also been documented for the African language Tigre by Vergnaud (1985). * 2. Latimer (1978: 20) states that only [ai] occurs to the right of the f l a t consonant (and [e] occurs only when i t is to the l e f t ) . My data do not suggest t h i s I have examples from four of my speakers where both [ai] and fe] ocurr to the right of the t r i g g e r . 3. This example i s from Cook (1976: 21). 4. I have l e f t the a f f i x e s undefined since I am unsure of their meaning. This i s a verb s t r i n g that functions as a noun (a zero derived noun). 5. This example i s from Latimer (1978). He does not provide an explanation of the morphemes present in this s t r i n g . Latimer states that in leftward harmony [ i l sometimes does not f l a t t e n , although i t does not block the further spread of f l a t t e n i n g . 6. Having both features present on the autosegmental t i e r i s not e n t i r e l y unprecedented. Kiparsky uses both + and - nasal for nasal harmony i n Guarani, and both + and - for Akan tongue 39 root harmony (Kiparsky 1983). 7. I am indebted to Dr. P a t r i c i a Shaw for t h i s suggestion. 8. Ideally there should only be one fl a t t e n i n g rule since i t is the same feature being spread i n each case. It has been suggested to me by Dr. Shaw that the in velar harmony process both the velar trigger and the RTR vowels may be linked to a dorsal node on their feature trees (dorsal refers to the tongue position) which would prohibit further spread. t+RTRJ would not be able to spread beyond the dorsal node. CHAPTER 3: LEXICAL PHONOLOGY 3.0 Introduction As noted i n section 1.3 of Chapter 1, the minimum sentence in C h i l c o t i n consists of a sel e c t i o n of elements from the following verbal s t r i n g : post adv stem dur # obj subj obv derv % derv mode conj pm c l s f stem The a f f i x a t i o n of these various morphemes res u l t s in a number of complex phonological changes. Analyses of these w i l l be presented in Chapters 4 through 7, using Lexical Phonology (LP) as the t h e o r e t i c a l model. This chapter gives an overview of the general c h a r a c t e r i s t i c s of the model and i t s appropriateness for the analysis of C h i l c o t i n . 3.1 The Lex i c a l Phonology Model LP inter-orders morphology and phonology within the lexicon. The lexicon consists of ordered levels (or strata) and each morphological a f f i x a t i o n process takes place at a pa r t i c u l a r l e v e l . "The rules of phonology interact with the morphology in that the phonology rules are assigned s p e c i f i c morphological s t r a t a as the i r domain and a given 41 phonological rule applies only at the stratum that is assigned to i t . " (Halle & Mohanan 1985) These phonological rules are termed ' l e x i c a l ' , since they apply in the lexicon after each application of a morphological rule, and are opposed to ' p o s t - l e x i c a l ' rules, which apply after the syntax.1 Besides d i f f e r i n g in their domain of a p p l i c a t i o n , l e x i c a l and p o s t - l e x i c a l rules may also have d i f f e r e n t properties (Kiparsky 1985). For one, l e x i c a l rules are usually boundary sensit i v e and may have id i o s y n c r a t i c exceptions, whereas p o s t - l e x i c a l rules apply 'across the board' wherever their environment is met. P o s t - l e x i c a l rules may also be optional, as in the case of leftward harmony in C h i l c o t i n . Another difference is that l e x i c a l rules usually produce strings that are structure preserving in the sense that they do not introduce nondistinctive features such as nasalization in English. The introduction of nondistinctive features i s the domain of p o s t - l e x i c a l rules. Another d i s t i n c t i o n between l e x i c a l and p o s t - l e x i c a l rules is in t h e i r mode of a p p l i c a t i o n . Lexical rules must obey the S t r i c t Cycle Condition or SCC (Kiparsky 1985), while p o s t - l e x i c a l rules do not. The SCC states that l e x i c a l phonological rules can only apply to derived environments. A derived environment is defined as a s t r i n g to which either a morphological or phonological rule of the same l e v e l has 42 applied. Thus a l e v e l 2 rule could not apply to a form that has had neither a l e v e l 2 morphological or phonological rule already apply. Structure building rules, e.g. those which add prosodic structure or f i l l in previously unspecified feature values, are considered by Kiparsky (1985) to be an exception to the SCC. As we w i l l see in 5.2.2, however, at least one structure building rule in C h i l c o t i n must obey the SCC. Kiparsky (1985: 1) claims that the d i s t i n c t i o n between l e x i c a l and p o s t - l e x i c a l rules is important due to i t s implications for l e a r n a b i l i t y . He states that "...the learner does not have to f i x the domain of these rules by checking the i r ordering or other properties". If a rule is boundary sen s i t i v e and structure preserving, then i t must be l e x i c a l . On the other hand, i f a rule applies whenever i t s environment is met and i s not structure preserving, then i t w i l l necessarily be p o s t - l e x i c a l . There are two opposing hypotheses regarding rule ordering within the lexicon. Kiparsky (1984:5) states that: "...the grammar may s t i p u l a t e merely where a rule ceases to a p p l y . . . A l l rules are p o t e n t i a l l y applicable at the f i r s t l e v e l of the lexicon, and apply there provided that the p r i n c i p l e s of grammar permit i t ; at lower levels of the lexicon and in the post l e x i c a l l e v e l phonology rules may be 'turned o f f but no new ones added (The Strong Domain Hypothesis)." 43 An a l t e r n a t i v e hypothesis for the assignment of rule domain is given in Halle & Mohanan (1985). They claim that the p o s t - l e x i c a l is the most unmarked l e v e l and rules are assigned to the l a t e s t l e v e l possible rather than the e a r l i e s t . Both Kiparsky (1984) and Halle & Mohanan (1985) claim that t h e i r model i s best suited to a theory of l e a r n a b i l i t y . In Kiparsky's model, i t is easier to learn where a rule shuts off, while in Halle and Mohanan's, i t i s easier to learn where a rule s t a r t s . I agree with Hargus (1985) that neither claim has any empirical evidence to support i t . I have, therefore, chosen Kiparsky's model because i t explains the h i s t o r i c a l process of the l e x i c a l i z a t i o n of rules (Kiparsky 1982). However, as we w i l l see in Chapter 4, the Strong Domain Hypothesis must be weakened s l i g h t l y to allow some new rules to be added at n o n - i n i t i a l levels of the lexicon. If LP i s the correct model to use to describe the interplay of C h i l c o t i n morphology and phonology, we should expect to find data that would indicate the existence of l e x i c a l and p o s t - l e x i c a l l e v e l s . Such data have been found in e a r l i e r generative studies of Athapaskan, as I w i l l now discuss. 3.2 Evidence for LP in Athapaskan E a r l i e r generative studies noted a common phonological d i v i s i o n among a f f i x e s b u i l t up on the Athapaskan stem. There 44 seemed to be a c l u s t e r i n g e f f e c t at work, di v i d i n g the s t r i n g of a f f i x e s into two phonological domains. For descriptive purposes I have represented t h i s word internal d i v i s i o n with a boundary symbol of the type developed by Chomsky & Halle (1968). The f i r s t published reference to boundary-related phenomena i s found in L i (1946:409) where i t is reported for Chipewyan that... "...there are two classes of prefixes, the conjunctive and the d i s j u n c t i v e . The conjunctive prefixes occur immediately before the stem and after the pronominal objective prefixes... There are also frequent contractions of these prefixes when they come together. The disjunctive prefixes occur before the pronominal objects and are less connected with the stem...[they] do not as a rule contract with the conjunctive prefixes." Similar observations are made in Kari's study (Kari 1975) of the disjunct boundary in Navaho and Tanaina. He states (p. 331) that t h i s "word internal boundary... plays a s i g n i f i c a n t role in the phonologies", despite th e i r being geographically distant from each other. This disjunct boundary creates a d i v i s i o n in the prefixes as shown below. Here the # marks the disjunct boundary where the prefixes to i t s l e f t are disjunct, and those to i t s right are conjunct. 45 Morphological Structure of Navaho and Tanaina adverbial - i t e r a t i v e - p l u r a l # d i r e c t object - d e i c t i c -1 2 3 4 5 aspect - mode - perfective - subject - c l a s s i f i e r - stem 6 7 8 9 10 11 Kari (p.333) adds the following comments about th i s boundary: " I t seems clear that the secondary [or disjunct] prefixes are probably late incorporations of independent prefixes as data from the archaic Eyak* language show that Eyak verbs begin with the d i r e c t object position and what are disjunct prefixes in Athapaskan are preverbal elements in Eyak." This secondary or disjunct boundary of Athapaskan plays a large role in my analysis of C h i l c o t i n as well. It conforms to the d i v i s i o n between the durative and d i r e c t object a f f i x e s in C h i l c o t i n . It i s also a major d i v i s i o n in the phonology of Sekani (Hargus 1985). Instead of having to encode t h i s d i s t i n c t i o n with an ad hoc boundary, LP captures t h i s c l u s t e r i n g e f f e c t by proposing morphological s t r a t a which phonological rules are assigned to. The fact that c e r t a i n phonological rules only apply to some a f f i x e s (or morphemes) and not to others is expected in a theory which groups a f f i x e s into separate classes or s t r a t a . In th i s case, the primary or conjunct a f f i x e s are assigned to one le v e l and the secondary or 46 disjunct ones to another. While Kari discusses two types of a f f i x e s , other research on Athapaskan languages has suggested even more d i s t i n c t i o n s . L i (1946) has evidence for three types of a f f i x e s in Chipewyan --disjunct, pronominal subject and object, and conjunct. Hargus (1985) has reported the existence of four lev e l s in Sekani. As w i l l be shown below, there i s evidence for three morphological levels in C h i l c o t i n . Thus although the t h e o r e t i c a l framework is d i f f e r e n t the current analysis w i l l be s i m i l a r in content to that described by L i for Chipewyan. Before presenting my analysis of the phonological and morphological lev e l s of C h i l c o t i n , I w i l l begin with a discussion of the nature of morphological rules. This i s because the t h e o r e t i c a l assumptions which I am making are d i f f e r e n t from those usually used in LP"*. 3 . 3 Morphological Rule Formation Athapaskan languages present three ordering problems for any theory of morphology. F i r s t , the verbal a f f i x e s occur in a s t r i c t order. A theory of a f f i x a t i o n w i l l have to ensure that t h i s order is preserved. Second, the d i r e c t i o n in which the a f f i x s t r i n g i s b u i l t i s problematic. If i t i s done ri g h t to l e f t , i t v i o l a t e s Anderson's (1982) claim that the a f f i x a t i o n of derivation should precede that of i n f l e c t i o n . This i s because C h i l c o t i n has i n f l e c t i o n a l a f f i x e s in positions 5 (object), 6 (subject), 7 (obviative), and 12 (person marker). 47 ( I n f l e c t i o n i s defined here as a f f i x e s which specify either person or number of the subject or object.) Third, some of the a f f i x e s in positions 2 (adverb), 8 and 9 (both derivative) form discontinuous stems with the stem of position 14. Some examples are [tae...daen] 'to d r i n k 1 , [yae...tek] 'to speak', [ye...zun] 'to think', and [dae...le4J 'to f l o a t ' . In these cases both the stem and the derivative a f f i x are needed to convey the f u l l meaning of the verb. The theory of a f f i x a t i o n w i l l need to allow for the intervention of other a f f i x e s between these discontinuous elements. The method of a f f i x a t i o n which I present below provides a s o l u t i o n to a l l three of the above problems. I w i l l f i r s t present arguments in favor of a right to l e f t a f f i x a t i o n process, and then turn to the d e t a i l s of the theory that ensure a f f i x order and allow the presence of discontinuous stems. As Rice (1985) has pointed out, i f derivational a f f i x e s are attached before i n f l e c t i o n a l ones, i t is necessary to p r e f i x the d e r i v a t i o n a l morphology onto the stem, and then i n f i x a l l of the i n f l e c t i o n a l morphology. This i s objectionable on two counts. F i r s t , i t is counterintuitive, since the l i n e a r order would not r e f l e c t the true ordering of the a f f i x a t i o n process. Second, i t i s impossible to formulate phonologically. As she comments (p. 159), "Once the entire d e r i v a t i o n a l s t r i n g i s present, there is no phonological context on which i n f i x a t i o n (of i n f l e c t i o n ) can be defined". One would have to resort to an i n f i x a t i o n process defined by 4 8 morphological p o s i t i o n . Rice argues that t h i s is equally unacceptable: " I t might be possible to formulate a complex i n f i x a t i o n rule making c r u c i a l reference to various morphological boundary types but i f Anderson's theory were to permit resorting to such devices, his claim that a l l i n f l e c t i o n a l a f f i x e s must be outside derivational a f f i x e s would be rendered v i r t u a l l y u n f a l s i f i a b l e " . (Rice 1985:160) Furthermore, in LP an i n f i x a t i o n process defined by either morphological position or boundary type i s impossible, since r e f e r r i n g to morphological position is not allowed and boundaries have been done away with in favor of levels (although reference to morphological bracketing is in r e s t r i c t e d conditions available to phonological r u l e s ) . I follow Rice in assuming that i n f l e c t i o n a l a f f i x a t i o n in C h i l c o t i n must be allowed to precede derivational a f f i x a t i o n , creating a uniform right to l e f t process. The theory of a f f i x a t i o n which I adopt follows proposals made for English by Aronoff (1974), Lieber (1980), and Selkirk (1976). A f f i x a t i o n takes places through the insertion of morphemes into branching trees. In t h i s system, a f f i x e s have subcategorization frames which ensure the co-occurrence of discontinuous morphemes while generating the correct surface order in a uniform right to l e f t process. 49 Lieber (1980) proposes a system of morphological a f f i x a t i o n that maintains a single context-free rule which generates unlabelled binary branching tree structures. Stems are f i r s t inserted into the tree structure, followed by a f f i x e s , based upon their subcategorization frames. For example, the English adjective 'happy' i s f i r s t inserted into the tree, since i t is a stem, e.g. A / happy A Then the s u f f i x '-ness* i s inserted, subject to i t s subcategorization frame which states that i t can only attach to adjectives: A / \ / \ happyA n e s s N Each a f f i x i s l i s t e d in the lexicon with a subcategorization frame which states i t s co-occurrence r e s t r i c t i o n s . For English, for example, the suffixes '-ness' and '-i z e ' would have the following subcategor iz a t ion frames: [adj ]JJ and [N_ ] v Thus '-ness' attaches to the ri g h t of adjectives and creates nouns such as 'happiness', and '-i z e ' attaches to the right of nouns to create verbs such as 'standardize'. I can now show how t h i s a f f i x a t i o n process operates in C h i l c o t i n . Due to the many types of a f f i x e s found in the verb 50 s t r i n g I propose that the relevant features for subcategorization are those of the a f f i x categories, i . e . +stem, f c l a s s i f i e r , ±conjunct, ±mode, ±derlvative, etc. rather than simply Noun and Verb. This i s not very r a d i c a l i f we look at what Chomsky (1965:199) o r i g i n a l l y said about the use of N and V. " I t i s quite possible that the categories noun, verb, adjective are the r e f l e c t i o n of a deeper feature structure, each being a combination of features of a more abstract sort". Chomsky leaves the exact nature of the features an open question and continues to use the abbreviations N, V, and A, since they do function as discrete classes in English. The use of features other than N, V, and A is not e n t i r e l y unprecedented. Aronoff (1974) uses abstract morphological features such as ' l a t i n a t e ' , 'Greek', 'Romance*, and 'native' for English to ensure that the a f f i x e s such as ' - i t y ' and '-hood' attach to the correct roots, e.g. ' i t y ' is [+latinate] and '-hood' is t+native]. He states (p.19): "The most important thing to be noted about a feature l i k e l a t i n a t e i s that i t i s abstract... there i s good evidence that the feature l a t i n a t e is a property of morphemes. Further evidence... of the abstract and a r b i t r a r y nature of the feature l a t i n a t e i s that monomorphemic (and t r u l y l a t i n a t e ) words tend to move into the native c l a s s i f i c a t i o n for purposes of a f f i x a t i o n as in priesthood and statehood." 51 Selkirk (1982) mentions the p o s s i b i l i t y of languages in which i t might be necessary to specify such features as gender, p l u r a l and case in order to derive correct a f f i x order. The features I propose for C h i l c o t i n are used in the subcategorization frames of a f f i x e s to ensure that they attach in the correct order to the correct stems. C r u c i a l l y , I assume that morphemes attach when only the features in their subcategorization frames are present and no other features are present. To exemplify how correct a f f i x order i s obtained, I w i l l use the mode a f f i x /ghe/ as an example. The mode a f f i x e s (position 10) a l l have the features (t+conj]), [+pm], and t+clsf] in t h e i r subcategorization frames to ensure that they are added only after the c l a s s i f i e r (position 11), person marker (position 12), and conjunct a f f i x e s (position 13), and no others, have been attached. (The conjunct a f f i x is optional and i s added only i f the verb is marked for t h i r d person). I assume that the features of each morpheme w i l l percolate up the binary branching tree under the following Feature Percolation Conventions, as defined by Lieber (1980): Feature Percolation Conventions I: A l l features of a stem morpheme, i . e . a morpheme lacking a subcategorization frame, label the f i r s t non-branching node dominating i t ; 52 I I : A l l features of an a f f i x percolate to the f i r s t branching node; I I I : If a branching node f a l l s to obtain any features v i a II, features from the next lowest node percolate up to i t , i.e. i f i t f a i l s to obtain any features that are not mutually exclus ive.^ We can see how these operate by looking at the l a b e l l i n g of the tree for •happiness'. By Convention I, the feature f+A] of the stem 'happy' percolates to the f i r s t non-branching node dominating i t : /\ / \ [+A] \ I \ 'happy' '-ness' By Convention I I , the feature I+N] of the a f f i x '-ness' percolates to the f i r s t branching node: [+N] /\ / \ t+A] \ I \ 'happy' '-ness' Convention III i s ruled out since the branching node i s already s p e c i f i e d for a category. The operation of Convention III can be shown by giving Lieber's example from Latin of 'dix-era-mus'. /dix/ i s 53 [-present], /era/ [+perfect], and /mus/ [+first person]. These features percolate up to the uppermost branching node via Convention I I I , since t h i s node has f a i l e d to receive any s p e c i f i c a t i o n s for these features v i a Convention I I : / t+V] [-present] [+perfect] [+ 1 person] /\ / \ •dix' A / \ 'era' 'mus' To exemplify these conventions for C h i l c o t i n a f f i x a t i o n , I w i l l use the example /tae-s-daen/ 'I'm drinking'. /daen/, with i t s features [+stem, +drink], i s f r e e l y inserted into a binary tree. By Convention I, the features [+drinkj, [+stem] percolate up: [+drink] [+stemJ I 'drink' [+drink] [+stem] This produces a node which the c l a s s i f i e r can attach to. The c l a s s i f i e r i s only subcategorized for stems and does not require any other features to be present. 54 /\ / \ •1' \ [ + c l s f ] \ \ [ + d r i n k ] [+stem] • d r i n k ' t+drink J [+stem] At t h i s s t a g e , the f e a t u r e f + c l s f ] w i l l p e r c o l a t e up t o the f i r s t b r a n c h i n g node v i a C o n v e n t i o n I I : t + c l s f ] /\ / \ '1' \ [ + c l s f ] \ \ f + d r i n k ] [ tstetn] ' d r i n k ' [ + d r i n k ] t + s t e m j Now, by C o n v e n t i o n I I I , the f e a t u r e [ + d r i n k ] i s a l s o a b l e t o p e r c o l a t e up t o the b r a n c h i n g node, s i n c e t h i s node i s not s p e c i f i e d f o r t h a t f e a t u r e : 55 t + c l s f ] [ + d r i n k ] A / \ •1' \ [+clsf] \ \ [ + d r i n k ] f +stem] • d r i n k ' [ + d r i n k ] [+stem] P e r s o n markers can now a t t a c h , s i n c e t h e y r e q u i r e o n l y the f e a t u r e [ + c l s f ] i n t h e i r s u b c a t e g o r i z a t i o n frames: A / \ / \ ' s ' l + c l s f ] t+pm) [ + d r i n k ] /\ / \ '1' \ [ + c l s f j \ \ [ + d r i n k ] [+steml I ' d r i n k ' [ + d r i n k ] [+stem] N e x t , the p e r s o n marker f e a t u r e p e r c o l a t e s t o t h e b r a n c h i n g node v i a C o n v e n t i o n I I . The f e a t u r e s [ + c l s f ] and [ + d r i n k j may a l s o p e r c o l a t e up v i a C o n v e n t i o n I I I : 56 t +pm] [ + c l s f ] t + d r i n k ] /\ / \ / \ • s ' [ + c l s f ] [+pml [ + d r i n k ] /\ / \ •1' \ [ + c l s f ] \ \ [ + d r i n k ] (tstem} I • d r i n k • [+drink J [ t s t e m ] The p h o n o l o g i c a l l y n u l l i m p e r f e c t marker can now a t t a c h , as i t s s u b c a t e g o r i z a t i o n frame r e q u i r e s o n l y the presence of a p e r s o n marker (where C o n v e n t i o n I I I has a l s o a p p l i e d ) : 57 t+imp] [+pm] [ + c l s f ] [ • d r i n k ] /\ / \ / t+pm] 0 [ + c l s f ] [+imp] [ + d r i n k ] /\ / \ / \ 's' [ + c l s f ] [+pm] [ + d r i n k ] /\ / \ •1' \ [ + c l s f ] \ \ [ + d r i n k ] [+stem] • d r i n k ' [ + d r i n k ] [+stem] The t r e e a t t h i s p o i n t w i l l now e n t e r l e v e l 2, whereupon a l l i n t e r n a l m o r p h o l o g i c a l t r e e s t r u c t u r e s a r e e r a s e d , l e a v i n g : [+imp] T+pm] f + c l s f ] [ • d r i n k ] [0 - s - 1 - daen] Now the l e v e l 2 a f f i x / t a e / , which has the s u b c a t e g o r i z a t i o n frame [ + d r i n k ] , i s a l l o w e d t o a t t a c h , g i v i n g the t r e e : 58 A / \ / \ 'tae* [+imp] [+pm] [+clsf] I+drink] I [0- s - 1 - daen] A phonological rule later deletes / l / , r e s u l t i n g i n [tae-s-daenl. 3.4 Summary We have seen above how the problematic C h i l c o t i n word formation process can be s i m p l i f i e d by using a few r e l a t i v e l y straightforward mechanisms that have been independently motivated for English a f f i x a t i o n (Kipasky 1982). This has been done by making the following general assumptions: (1) a f f i x e s have subcategorization frames; (2) the relevant features for C h i l c o t i n consist of several a f f i x categories of the type [ c l s f ] , [mode], [person], etc., and of abstract semantic features, e.g. [drink], [ t a l k ] , etc. ( c . f . Aronoff 1974, Lieber 1980); (3) word formation i s in the form of binary branching trees into which a f f i x e s are inserted, subject to the i r subcategorization frames; and (4) there are three Feature Percolation Conventions. Discontinuous morphemes are no longer a problem, nor i s a contiguous r i g h t to l e f t a f f i x a t i o n process. 59 Footnotes for Chapter 3 1. I assume that after the l a s t l e x i c a l l e v e l the morphemes enter the syntax where l e x i c a l i nsertion takes place. See Chomsky (1981) for d e t a i l s . 2. Eyak i s a s i s t e r language to the Athapaskan languages which, according to Krauss, retains many elements of an Eyak-Athapaskan language ancestor. 3. See Speas (1984) for an alternative type of a f f i x a t i o n device for Navaho. 4. There i s which I have a fourth convention that deals not included here as i t is not with compounds relevant. 60 CHAPTER 4: THE LEXICAL PHONOLOGY OF CHILCOTIN: LEVEL 1 As mentioned in the previous chapter, my analysis of C h i l c o t i n proposes three levels to the morphology. Each level w i l l be presented in a separate chapter, with the current chapter dealing with Level 1. Within each of these chapters, I w i l l f i r s t discuss the morphological classes which are associated with i t , followed by a description of the rules which apply. 4.1 Morphology 4.1.1 Stems (Position 14). The C h i l c o t i n verb stems vary in accordance with the mode and aspect of the entire s t r i n g . Both the stem vowel and the f i n a l consonant may change, as can be seen in the following examples:* (1) a. [tae-zul] 'he's going to s t a r t being good' A derv stem 9 14 b. [na-gha-d-zun] 'he i s good again' A derv mode els stem 9 10 13 14 c. [na-gha-d-zun] 'he often becomes good' A dur mode els stem 4 10 13 14 61 d. [nae-tae-d-zul] 'he is going to be good again' A adv derv e l s stem 2 8 13 14 e. [la h i - z u l ] 'he i s not good (for the job)' A neg derv stem 9 14 f. [la hi-zuh] 'he's not good (healthy)' A neg derv stem 9 14 g. [ghe-n-zu] 'he was once good' A mode conj stem 10 11 14 4.1.2 C l a s s i f i e r s (Position 13). There are 4 C h i l c o t i n c l a s s i f i e r s : /d,1,4,*/ ^. When productive, / 4 / generally marks t r a n s i t i v i t y as seen in the data below: (2) a. i . [da-z-*-k'an] ' i t i s burning' E derv perf els stem 9 10 13 14 i i . [di-4-k'aen] 'I burned i t ' E derv els stem 9 13 14 b. i . [g wa-da-gha-s-0-mal] 'I r o l l e d ' E obj derv perf pm els stem 5 8 10 12 13 14 62 i i . [g we-day-4--mal] 'I r o l l e d i t ' E obj d e r v e l s stem 5 9 13 14 c i . [ t a e - ^ - s a x ] 'he i s g o i n g t o s p i t ' A d e r v e l s stem 9 13 14 i i . Tbs-dzax b i - t a e - 4 — s a x ] 'he i s g o i n g t o s p i t h i s h i s gum adv d e r v c l s f stem gum 1 A I n some paradigms, the c l a s s i f i e r i s not needed t o mark t r a n s i t i v i t y . I n the f o l l o w i n g t r a n s i t i v e v e r b s , the c l a s s i f i e r i s /#/. (3) a. [ ho- g h e - n - t * - t s i J 'you s h o t i t (many t i m e s ) ' B d e r v p e r f pm e l s stem 8 10 12 13 14 b. [ys-0-daen] 'he i s d r i n k i n g i t 1 C obv e l s stem 7 13 14 c. 'Mary' [ s a i - n - t f - t a 4 ] 'you k i c k e d Mary' B p e r f pm e l s stem 10 12 13 14 I n the m a j o r i t y of c a s e s , the c l a s s i f i e r s have become i d i o s y n c r a t i c , and i t i s s i m p l e s t t o l i s t the stems t o which t h e y a t t a c h i n t h e i r s u b c a t e g o r i z a t i o n frames. 4.1.3 P e r s o n Markers ( P o s i t i o n 1 2 ) . These a f f i x e s mark 63 the person and number of the subject. The inventory is l i s t e d below with examples following: Person F i r s t Second Third S ingular s in 0 (4) a. [ h e - s - d 3 e n ] 'I sing' E derv pm stem 9 12 14 P l u r a l id ch 0 (marked by d 3 E in position 6) b i . [ h - i n - d 3 E n ] 'you s i n g 1 E derv pm stem The vowel here must be / i / rather than /e/ (which would not raise to [ i ] before [n]) due to the following data where / E / does indeed occur before [n]. i i . [ne-nsz] ' i t i s long' A derv stem 9 14 i i i . [te-na-d3i-ya-gha-l-tJut] 'they caught i t again' A adv dur subj obv perf c l s f stem 2 4 6 7 10 13 14 iv . [g WE-ni-nk'aez] 'water is cold' (Cook 1987) 64 v. [nae-se-nae-ghe-ne-1-t.fens ] 'you are h i t t i n g me' (Cook 1987) c. [ h e - * - d 3 e n ] 'he sings' E derv pm stem d. [ h - i - d 3 e n ] 'we sing' E derv pm stem e i . [ h - e h - d 3 e n ] 'you (pi.) sing' E derv pm stem This a f f i x must be vowel i n i t i a l due to the following examples. e i i . [gax de-n_i-4-te4-] »he shot the rabbit' B rabbit derv perf c l s f stem 8 10 13 14 i i i . [gax de-n-e-4—te4] 'you (pi.) shot the rabbit' B rabbit derv perf pm c l s f stem 8 10 12 13 14 The perfective a f f i x i s / n i / as shown in example i i . In example i i i , i t i s the vowel [e] that surfaces. This can only be due to an / e / - i n i t i a l person marker. The /h/ of the person marker would have coalesced with the following /4/ c l a s s i f i e r (see 4.2.3) before the / n i / perfective was added. The /h/ would thus not be available to lower the i of / n i / to [e]. Therefore t h i s person marker must be /eh/. (I also assume that t h i s a f f i x i s vowel i n i t i a l based on h i s t o r i c a l evidence. Krauss (1964) has reconstructed t h i s a f f i x as *ax for Na-Dene. 65 f. [d.3£-d3En] 'they sing' E subj stem 4.1.4 Conjunct (Position 11). C h i l c o t i n has one conjunct a f f i x /n/ . It occurs only in the t h i r d person forms of strings that contain either the /ghi/ s e r i a t i v e or / n i / perfective a f f i x e s . The function of the conjunct in C h i l c o t i n i s not c l e a r . In Sekani the conjunct a f f i x marks conjugation classes of modes. (Hargus 1985). (5) /ghi/ Se r i a t i v e a. [yo-ghe-n-tael] 'he kicked i t ' (seriat.) A obv perf conj stem 7 10 11 14 b. [yo-ghe-n-tsi ] 'he shot i t ' (seriat.) B obv perf conj stem 7 10 11 14 c. [d3i-yo-ghe-n-tael] 'they kicked i t ' (seriat.) B subj obv perf conj stem 6 7 10 11 14 d. [ s i t na-ghi-n-tsun] 'he kissed me' B (FL. IRR.) pron dur perf conj stem 4 10 11 14 (6) / n i / Perfective a. [dsi d3e-ni-n-dil] 'they arrived here' A •here' subj perf conj stem 6 10 11 14 66 b. 'Mary' [ni-n-di] 'Mary was short' A noun perf conj stem 10 11 14 c. [na4ey x wa-ta-ni-n-tael] noun obj derv perf conj stem 5 9 10 11 14 •the horse was kicking into us' B As can be seen in the data below, conjunct [n] occurs only in the t h i r d person. (7) a. tyo-ghe-n-tael] 'he kicked i t ' A obv perf conj stem 7 10 11 14 b. [ho-gha-s-tael] 'I kicked i t ' A derv perf pm stem 8 10 12 14 c. [ho-gha-h-tael] 'you (pi.) kicked i t ' A derv perf pm stem 8 10 12 14 d. [ho-ghe-tael] 'we kicked i t ' A derv perf stem 8 10 14 4.1.5 Mode (Position 10). These a f f i x e s ; perfective, imperfective, optative, inceptive, and s e r i a t i v e — mark the modal mood of the verb s t r i n g as described in the sections 67 below. Perfec t i v e : /ni,si,ghe/. This a f f i x indicates a completed action.' 4 The semantic d i s t i n c t i o n s among these three a f f i x e s are not clear from my data. Stems with which each has been found are l i s t e d below. / s i / / n i / /ghe/ 'arrive * •crawl * •shake' NP •turn' NP 's p i t ' NP •shatter' NP 'glance at' NP •push' NP •poke' NP 'swim' 'want (to sing)' 'cut' NP •dream' ' f a l l down' • s i t ' 'get bloody' ( S E ) 'sleep' ( S E ) •shoot* unspecified NP 'break in h a l f ' shoot' 'sing' 'smell' NP 'scratch' NP • cry' 68 I have c l a s s i f i e d the /se/ a f f i x in 'get bloody' and 'sleep' as a v a r i a t i o n of the perfective / s i / a f f i x rather than as a conjunct a f f i x due to i t s semantic function- a marker of completed action. / s i / perfective also has / z i / as a variant which cannot always be accounted for by i n t e r v o c a l i c voicing of the /s/. Imperfective; /*/. Phonologically n u l l , t h i s a f f i x indicates actions that are incomplete. Both examples are from .speaker A. (8) a. [na4-yae-0-s-ta*] 'I am speaking with you' post derv imp pm stem 1 8 10 12 14 b. [sak'i ? e - d 3 e - t E - 0 - t s i h ] noun obj subj derv imp stem 5 6 9 10 14 'they are shooting into a group of cows' Optative /gh we/. Both phonologically and semantically, t h i s i s a very inte r e s t i n g a f f i x . Phonologically i t a f f e c t s c e r t a i n vowels to the r i g h t , rounding and lowering them. (This w i l l be further discussed on pages 96-7). Semantically, t h i s mode translates as a desire or wish of limited sorts as in the following: 6 9 (9) a. [gh wo-d3En] ' l e t ' s sing* B opt stem 10 14 b. [wo-n-d3en] '(they asked) you to sing' B opt pm stem 10 12 14 c. [gha-su] 'I'm thinking about being good' A opt stem 10 14 d. [wa-s-t/ah] 'I'm thinking about being big' A opt pm stem 10 12 14 (i n b. and d. the velar [gh] was so weakly a r t i c u l a t e d that i t is not represented in the phonetic transcription) Inceptive: /tae. . .ghe/. /tae/ in position 8 plus t h i s a f f i x , /ghe/, indicate the i n c e p t i v e — a n event or action that w i l l soon s t a r t . In the f i r s t person singular and the t h i r d person singular and p l u r a l , the perfective /ghe/ often does not occur. This i s shown below in the (i) examples. (10) a. i . [ tae-s-"?il ] 'I'm going to look at i t ' A derv pm stem 8 12 14 70 i i . [ t a - g h a - l - ^ i l ] 'we are going to look at i t ' A derv inc els stem 8 10 13 14 b. i . [d3e-tae-dil] 'they are going to drink' A subj derv stem 6 8 14 i i . [ta-gha-dil] 'you (pi.) are going to drink' C derv inc stem 8 10 14 c. i . [tae-tsat] 'he i s going to swing' E derv stem 8 14 i i . [ta-gha-n-tsat] 'you are going to swing' E derv inc pm stem 8 10 12 14 The a f f i x a t i o n of /ghe/ i s therefore optional for t h i r d person singular and p l u r a l and the first.person singular. (See section 7.2 for [gh] contraction in fast speech). Ser i a t i v e /ghi/. As mentioned e a r l i e r , the s e r i a t i v e mode denotes usually a singular action that has been repeated many times, as in the following: (11) a. i . [yo-ghe-n-tael] 'he kicked i t many times' A obv ser conj stem 7 10 11 14 71 i i . [ya-z-ta4] 'he kicked i t once obv perf stem 7 10 14 b. i . [x wa-d3o-ghi-n-tael] 'they kicked us many times' A (FL. IRR.) obj subj ser conj stem 5 6 10 13 14 i i . [d3e-ta-z-ta4-] 'they kicked once' A subj derv perf stem 6 9 10 14 c. i . [so-ghi-n-tal] 'he kicked me many times' A (FL. IRR.) obj ser conj stem 7 10 11 14 i i . tsa-z-ta4J 'he kicked me once* A obj perf stem 5 10 14 d. i . Cho-ghe-n-tsi) 'you shot i t many times' A derv ser conj stem 9 10 11 14 i i . 'Mary* [hu-ni-n-tsax] 'you shoot Mary' A noun derv perf conj stem 8 10 11 14 Of phonological interest i s the vowel to the l e f t of the s e r i a t i v e a f f i x . In every case, i t i s the [+RTR] mid back round vowel [o], often accompanied by a [w] g l i d e . This phenomenon w i l l be further discussed in the section on l e v e l 2 phonology. 72 4.1.6 Derivative (Position 9). These a f f i x e s are generally quite i d i o s y n c r a t i c and many must simply have their subcategorization features l i s t e d in the lexicon. At best, I can only state semantic or paradigmatic tendencies among these a f f i x e s . The a f f i x e s are l i s t e d below with examples of each and a semantic explanation when possible. / t c / . This a f f i x i s found in verb strings that have an understood goal. (12) a. [ta - z - t'as] 'he cut i t ' A derv perf stem 9 10 14 [d5E-te-ta4] 'they are kicking' B subj derv stem 6 9 14 c. [?a-ta-za-si] 'I shot' A obj der perf stem 5 9 10 14 d. [ta-ze-4-kat] 'he broke' A derv perf e l s stem 9 10 13 14 73 e. NP [bi-tg-4-zaxl 'he usually s p i t s NP1 A adv derv c l s f stem 2 9 13 14 f. [da-da-ta-za-4-vjn] 'you (pi.) glanced at i t ' A derv derv derv perf c l s f stem 8 8 9 10 13 14 g. [bi-n-ta-za-4-qeyl 'you (pi.) vomited' A post adv derv perf c l s f stem 1 2 9 10 13 14 h. [tu ta-za-len] 'water i s flowing' A noun derv perf stem 9 10 14 i . [da-te-za-d-ts'at] 'I f e l l down' E derv derv perf c l s f stem 8 9 10 13 14 j . [naen-ta-za-s-bin1 'I started to swim away from you' A post derv perf pm stem 1 9 10 12 14 Idiosyncratic Derivatives. /ne/ and / d i / are i d i o s y n c r a t i c a l l y linked to a stem. /ne/ occurs with *to think' as shown below. (13) a. tne-s-at] 'I am thinking' A derv pm stem 9 12 14 74 b. [n-j^-zat] 'you are thinking' A derv pm stem c. [ne-zat] 'he i s thinking' A derv stem d. tn-id-zat] 'we are thinking' A derv pm stem e. [n-eh-zat] 'you (pi.) are thinking' A derv pm stem (In b. and d., / E / has deleted before the vowel / i / of the person marker see Vowel Deletion I, section 4.2.1.) / d i / co-occurs with the stem 'to say'. (14) a. [hae-dc-s-den] 'I said' D derv derv pm stem 9 8 12 14 b. [hae-dl-n-dEh] 'you s a i d 1 D derv derv pm stem 9 8 12 14 c. [hae-di-dEh] •he sa i d ' D derv derv stem 9 8 14 75 d. [hae-di-deh] 'we said' derv derv stem 9 8 14 e. [hae-de-h-deh] 'you (pi.) said' derv derv pm stem 9 8 12 14 (In a., / i / has lowered to te] due to the following [ s ] , and in e., the / i / has deleted before the person marker vowel / E / , see section 4.2.1 for both rules.) In conclusion, the derivative a f f i x e s of l e v e l one are /te/, which attaches to verbs with a goal, and /ne/ and / d i / , which are completely i d i o s y n c r a t i c in their attachment. 4.1.7 Summary of Sect ion Level 1 contains the following morphological a f f i x e s : derv mode conj pm c l s f stem  Per te n i n s 0 ne s i in 4-di ghe 0 1 imp- 0 id d opt- gh we eh inc- tae. . .ghe ser- ghi 76 4 . 2 Level 1^  Phonology This section i s divided into 3 main subsections: (1) rules that create a l e v e l 1 d i s t i n c t i o n , i . e . t h e i r environment i s met at leve l s 1, 2, and/or 3, but they apply only to level 1 a f f i x e s ) ; (2) rules that apply f r e e l y to a f f i x e s of levels 1 and 2, but that "shut-off" after l e v e l 2; and (3) rules that apply only to l e v e l 1 a f f i x e s but that do not create a l e v e l 1 d i s t i n c t i o n , i . e . their environment i s met only at l e v e l 1. 4.2.1 Rules that Create a Level iL D i s t i n c t i o n Vowel Deletion I_. Three of the person marker a f f i x e s are vowel i n i t i a l - - / i n / (second person singular), / i d / ( f i r s t person p l u r a l ) , and /eh/ (second person p l u r a l ) . When an a f f i x of the form CV 5 i s added to a s t r i n g beginning with one of the VC person markers, a W sequence is created. This 1CVVVC.stem1 sequence is not allowable. There are various options available to s y l l a b i f y such a sequence; a consonant may be inserted between the two V s (creating 'CV.CVC. stem'), one of the vowels could change into a glide or one of the vowels may delete (creating •CVC.stem'). C h i l c o t i n uses the la s t option of vowel d e l e t i o n . This creates the l e v e l 1-2 d i s t i n c t i o n at l e v e l 1 that the leftmost vowel deletes (at l e v e l 2, the rightmost vowel deletes). This i s i l l u s t r a t e d in the data below for the l e v e l 1 a f f i x e s /te/ ' d e r i v a t i v e 1 , and / s i / 77 •perfective•. (15) a. [t-in-te4] /te-In-te*/ 'you are kicking' A derv pm stem 9 12 14 b. [t-id-te4] /te-id-te4/ 'we are kicking' A derv pm stem 9 12 14 c. [te-te4] / t e - t c V 'he i s kicking' A derv stem 9 14. (16) a. NP [bi-t-i_-4-saxl NP /bi-te-in-4-sax/ adv derv pm c l s f stem 2 9 12 13 14 •you are s p i t t i n g NP' A b. NP [ b i - t - i - l - s a x ] NP /bi-te-id-4-sax/ adv derv pm c l s f stem 2 9 12 13 14 •we are s p i t t i n g NP1 A 6 c. NP tbi-tE-4-sax] NP /bi-te-4-sax/ adv derv c l s f stem 2 9 13 14 •he i s s p i t t i n g NP• A 78 (17) a. [na-s-ah-bin] /nae-si-eh-bin/ dur perf pm stem 4 10 12 14 •you (pi.) swam1 (A) (here the / e / of the person marker has flattened to [ a ] ) b. [na-se-bin] /nae-si-s-bin/ 'I swam' A dur perf stem 4 10 , 14 (the / i / of the perfective marker has flattened to te]; I cannot account for the missing /s/ person marker). c. [na-se-bin] / n a - s i - i n - b i n / 'you swam* A dur perf stem 4 12 14 (here the / i / o f the / s i / perfective has flattened t o [e]. I can not account for the missing /n/.) (18) a. [ s - a h - g w a t ] / s i - e h - g W e t / *you (pi.) poked' A perf pm stem 10 13 14 (here the /e/ of the / e h / person marker has flattened to [ a ] ) b. [ s a i - g W a t ] / s i - s - g W e t / 'I poked' A perf stem 10 14 (the / i / of the perfective marker has flattenend to [ a i ) ; I cannot account for the missing /s/ person marker) 79 (19) a. lgh-i_-l-gwat] /ghe-in-l-gWet/ 'you crawled' B perf pm c l s f stem 10 12 13 14 [gh-i_-l-gwat] /ghe-id-l-gw et/ 'we crawled 1 B perf pm c l s f stem 10 12 13 14 c. [gha-l-gWat] /ghe-l-gWet/ 'he crawled' B perf c l s f stem 10 13 14 (in a. and b., [ i ] has f a i l e d to f l a t t e n . This was common in the speech of speaker B) To account for these data the following rule can be written: Vowel Deletion 1^  V --> 0 / V (I assume that these are V s l o t s on the s k e l e t a l t i e r ) The Domain of Vowel Deletion I_. The following data reveal that Vowel Deletion I must precede RTR Velar and Alveolar Harmony. In (20)a, Vowel Deletion I has applied f i r s t , deleting the I i ] of the s e r i a t i v e marker (/ghi/), then RTR Velar Harmony applies f l a t t e n i n g / E / to [ a l . Example (20)b '80 reveals the underlying form of the s e r i a t i v e a f f i x . (20) a. [ho-gh-ah-tael] 'you (pi.) kicked i t (many times) derv ser pm stem 8 10 12 14 b. [yo-ghe-n-tael] 'he kicked i t (many times) obv ser conj stem 7 10 11 14 (the vowel of the s e r i a t i v e a f f i x has flattened to [e] here) If Velar Harmony had applied f i r s t , the following ungrammatical form would be produced: (21) a. /hu-ghi-eh-tael/ RTR Velar Harmony ho-ghe-eh-tael Vowel Deletion I [ho-gh-eh-tael] The same incorrect ordering can be seen in the following sentence: b. /gax ho-ghi-eh-tsi/ 'you (pi.) shoot many rabbits' A RTR Velar harmony gax ho-ghe-eh-tsi Vowel Deletion I *[gax ho-gh-eh-tsi] Vowel Deletion I must also apply before RTR Alveolar Harmony. As shown in the data below, the [ i ] of the perfective marker / s i / deletes (via Vowel Deletion I) afte r being affixed to the [e] person marker, leaving [S E] (or [Z E] i f i t i s in 81 i n t e r - v o c a l i c p o s i t i o n ) . (22) a. / t e - s i - e h - z i / 'you (pi.) s p i t ' A Vowel Deletion I tE-z -eh-zi RTR Harmony [ta-z -ah-zl] b. /tE-si-Eh-Jcet/ 'you (pi.) broke NP' Vowel Deletion I tE-z -eh-JtEt RTR Harmony [ta-z -ah-kat] c. / s i - E h - b i n / 'you (pi.) swam' Vowel Deletion I s -Eh-bin RTR Harmony [s -ah-bin] ( A l l examples are from speaker A) Thus vowel Deletion I must apply before both Velar and Alveolar Harmony. Level Domain of Vowel Deletion I_. Vowel Deletion I does not apply at l e v e l s 2 or 3. In examples (23) and (24), i t i s the rightmost vowel that deletes rather than the leftmost. 82 (23) Level 2 Affixes a. /dae-eh-dil/ —> [dae-h-dil] 'you (pi.) a r r i v e d ' A derv pm stem 8 12 14 b. /hu-id-tsax/ —> [hu-tsax] 'we shoot NP' B derv pm stem 8 12 14 c. /ni-eh-4-vaeh/--> [ni-4-?aeh] 'you (pi.) look at i t ' A derv pm c l s f stem 8 12 13 14 (24) Level 3_ A f f i x e s 7 a. /nae-in-d3it/—> [nae-d3it] 'you are crawling around' adv pm stem 2 12 14 (I cannot account for the missing In].) b. /nae-id-l-d3it/--> [nae-l-d3it] 'we are crawling around 1 adv pm c l s f stem 2 12 13 14 c. /nae-Eh-l-d3it/--> [nae-4-d3it] 'you (pi.) are crawling around' adv pm c l s f stem 2 12 13 14 ( a l l examples are from speaker B) In every case, [ae] has remained and the person marker vowel has deleted. i-Lowering. Another rule that provides evidence for level B3 1 i s i-Lowering, where / i / lowers to [E] before [s] or [ z ] . This can be seen in the following data. A l l the ( i i ) example in (25) reveal the underlying form of the a f f i x that contains the / i / . (25) a. i . [ n e - s - E h ] /ni-s-ysh/ 'I arrived here 1 A perf pm stem 10 12 14 i i . [ni-yeh] /ni-yeh/ 'he arrived here' A perf stem 10 14 (this [ni] i s the perfective marker as in [ni-4—tE4-] 'he shot rabbit' A and [hu-nae-ni-tsin] 'you kissed him' B b. i . [ n a e - b l - t a - z i - 4 - t s E t ] / n a e - b l - t i - s - 4 - t S E t / adv adv derv pm c l s f stem 1 1 8 12 13 14 'I pushed you away' B i i . [ d a e - b i - y u - t i ^ - 4 - t s E t ] /dae-bi-yE-hu-t i_ - 4-tset/ adv adv obv derv c l s f stem 1 1 7 8 13 14 'he pushed him over' B c. i . [nae-ns-s-dsit] /nae-ni-s-d3it/ *I crawled across 1 adv perf pm stem 2 10 12 14 i i . [nae-ni-dsit] /nae-ni-dsit/ 'he crawled across' B adv perf stem 2 10 14 84 a. i . [ne-s-icat] / n i - s - l - k e t / 'I broke NP' A perf pm c l s f stem (the [1] c l a s s i f i e r has deleted; see Continuant Deletion below) i i . I n i - l - k a t ] / n i - l - k e t / 'he broke NP' A To account for these data the following rule can be wr i t t e n : i-Lower ing Domain of i-Lower ing. i-Lowering must be ordered before Velar Harmony. As shown below in (26), the s e r i a t i v e a f f i x / g h i / surfaces as [gha]. F i r s t , i-Lowering applies, creating Ighe], then Velar Harmony flattens [ghe] to [gha]. (26) a. /hu-ghi-s-tael/ 'I kicked i t (many times)' perf c l s f stem 10 13 14 V [+ high] [- back] — > V [-high] i-Lower ing hu-ghe-s-tael Velar Harmony [ho-gha-s-tael] derv ser pm stem 8 10 12 14 85 b. /hu-ghi-s-tsi/ 'I shot i t (many times) i-Lowering hu-ghe-s-tsi Velar Harmony [ho-gha-s-tsi] derv ser pm stem 8 10 12 14 (both examples are from speaker A) i-Lowering and Alveolar Harmony. i-Lowering must also apply before Alveolar Harmony. In the examples below the perfective marker / s i / has become [S E] due to the following person marker /s/. The / E / has then retracted to [a] due to Alveolar Harmony r e s u l t i n g in [sa] or [zaj. Alveolar Harmony ordered before i-Lowering would produce [se - (stem)] or * [ s a i - (stem)]. (27) a. [na-sa-s-bin] 'I swam' A dur perf pm stem 4 10 12 14 b. [?a-ta-za-s-i1 *I shot' A obj derv perf pm stem 5 9 10 12 14 c. [na-za-s-d3it] 'I crawled across' B adv perf pm stem 2 10 12 14 86 d. [ta-za-s-tsat] 'I f e l l down1 E derv perf pm stem 9 10 12 14 e. [ta-za-s-Rat] 'I broke NP' A derv perf pm stem 9 10 12 14 f . [?a-ta-|a-s-tael] 'I kicked NP' obj derv perf pm stem 5 9 10 12 14 Level Domain of i-Lower ing. i-Lowering does not apply at l e v e l 2. When the / i / of a le v e l 2 a f f i x occurs to the l e f t of an [s] or [ z ] , i t does not lower to [e] as shown in the examples below. (28) a. [d5i-zuhl 'they are not good (healthy)' A subj stem 6 14 b. [dsi-zu] 'they are not good (for the job)' A subj stem 6 14 c. I t i - z a h l 'he i s gone' A derv stem 8 14 87 d. [nl-s-'Paeh] 'I look at i t 1 A derv pm stem 8 10 14 e. [ghe-zu] 'he was once good 1 A derv stem 8 14 f. [may ghe-zun] 'berries are good' A 'berry' derv stem 8 14 g. [na-ne-s-tIon] 'fence' adv derv perf ? 2 8 10 14 i - Lowering also does not apply at l e v e l 3. In example (29), the [ i ] of the l e v e l 3 adverb / b i / has not lowered to [e] although there i s an Is) to the immediate r i g h t . (29) [nae-bi-si-4-tsat ] 'I pushed you around' (FL. IRR.) adv adv perf c l s f stem 2 2 10 13 14 4.2.2 Rules That Apply at Levels 1 and 2 Having established the existence of l e v e l 1, I w i l l now turn to other rules that apply to l e v e l 1 a f f i x e s , but are not r e s t r i c t e d to l e v e l 1. That i s , they also apply at l e v e l 2. 86 F r i c a t i v e Voicing. The following data i l l u s t r a t e the phenomenon of f r i c a t i v e voicing, whereby f r i c a t i v e s become voiced in i n t e r v o c a l i c p o s i t ion. In the examples below i t i s the /s/ of the perfective marker / s i / (as in [nae-sai-tin] 'I dreamt 1 B and [del sa-leyn] ' i t got bloody' E) which becomes voiced (at le v e l s 1 and 2). (30) a. [g a-dae-zai-ta41 'I kicked i t ' B obj derv perf stem 5 8 10 14 b. [tu ta-za-len] 'water is flowing* A noun derv perf stem 9 10 14 c. [ta-zai-4-kat] 'he broke NP' A derv perf c l s f stem 9 10 13 14 d. [ta-zi_-4-tsat ] 'he pushed NP' E derv perf c l s f stem 9 10 13 14 e. I?a-ta-za-tsi] 'I shot' A obj derv perf stem 5 9 10 14 These data can be accounted for by the following r u l e . 89 F r i c a t i v e Voicing 10 [+vce] [+vce ] V C [+cont] V Domain of F r i c a t i v e Voicing. This rule also applies at l e v e l 2, as can be seen in the following example where the a f f i x that contributes to the environment of the ru l e is a l e v e l 2 a f f i x . (31) [nae-j[zai-ta41] 'I kicked you 1 B obj perf stem 5 10 14 At l e v e l 3, however, i n t e r - v o c a l i c f r i c a t i v e s remain voiceless. (32) a. [nae- 2[sc-s-d3it]] 'I crawled' B (FL. IRR.) adv perf pm stem 2 10 12 14 b. [nae~2[sai-tin]1 'I dreamt' C (FL. IRR.) dur perf stem 4 10 14 90 c. [nae-yi-b-ig[sai-4-tsat1] 'I pushed him around' B (FL. IRR.) adv adv adv perf c l s f stem 2 2 2 10 12 14 This rule also does not apply p o s t - l e x i c a l l y to underived items or across word boundaries. (33) a. [^eii] 'confluence' A b. [t4owesan] 'snake' C c. [ i 4 i ] 'one' E d. [ 4uwi 4aen] 'a l o t of f i s h ' D ' f i s h ' 'a l o t ' As shown above, F r i c a t i v e Voicing applies at le v e l s 1 and 2 1 1 . I follow Kiparsky's Strong Domain Hypothesis (1984) and assign i t to both l e v e l s . 4.2.3 Rules That Only Apply at Level 1^  Three other rules that apply to l e v e l 1 a f f i x e s , but do not provide evidence for a l e v e l d i s t i n c t i o n , are D-effect (a common phenomenon in Athapaskan), Continuant Coalescence, and Continuant Deletion. These three rules provide d i f f e r e n t strategies for getting r i d of consonant clusters in the verb s t r i n g . Continuant Coalescence and D-effect are a c t u a l l y rather s i m i l a r in that they collapse two adjacent consonants 91 into one, whereas in Continuant Deletion a consonant deletes rather than coalesces. These rules w i l l be discussed in the following order: D-effect, Continuant Coalescence, and Continuant Deletion. D-effect. D-effect, as i t has been t r a d i t i o n a l l y referred to, involves the [d] c l a s s i f i e r (or the /d/ of the f i r s t person p l u r a l /id/) and the stem i n i t i a l consonants as described below 1 2 (see Howren 1971, Krauss 1969 for further d e t a i l s ) . (34) d + f d + z d + 9 d + 4 d + gh Some examples are as follows: (35) d + T --> d3 a. Ine-s-d3en] /ne-s-d-JTen/ 'I sang' E perf pm c l s f stem b. [sE-fen] 'my song' E poss stem ( A l l modal variations of 'sing' have the Id] c l a s s i f i e r and are thus tds] i n i t i a l . I assume 'sing' is derived from 'song'.) — > d3 — > dz — > t' --> r (a flapped [1]) 92 (36) d + z — > dz a. [ n a - n e - n - i d - z a t ] / n a - n i - n i - i d - z e t / 'we a r e t h i n k i n g ' A dur ? d e r v pm stem 4 9 10 14 b. f n a - n e - n - z a t ) / n a e - n i - n i - z e t / 'he i s t h i n k i n g ' A dur ? d e r v stem 4 9 14 (37) d + 9 --> t ' a. [k'aen za d g i - d a e - t ' a e z ] /k'aen za d 3 i - d a e - d - v a e z / ' j u s t now p l a c e ' s u b j d e r v pm stem 6 8 12 14 'we (2) j u s t a r r i v e d (somewhere)* A b. [k'aen za d s i - d a e - h - ^ a e z ] /k'aen za d s i - d a e - e h - ^ a e z / ' j u s t now p l a c e ' s u b j d e r v pm stem 7 8 10 14 'you (2) j u s t a r r i v e d (somewhere)' A (38) d + 4 -> r a. i . [ ta-ze-T-Jcat J / t i - s i - i d - 4 — R e t / 'we broke i t ' A d e r v p e r f pm c l s f stem 9 10 12 13 14 i i . [ta-za -4;-k*at] / t i - s i - 4 ~ k * e t / 'he broke i t ' A d e r v p e r f c l s f stem 9 10 13 14 ( / t i / l o w e r s t o [ t e ] t h e n f l a t t e n s t o l a ) ) . 93 b. i . [ni-l-vaeh] /ni-id-l-?aeh/ 'we looked at NP' A perf pm c l s f stem 10 12 13 14 i i . [ye-ni-4-vaeh] /ye-ni-4-9aeh/ 'he looks at NP' A obv perf c l s f stem 7 10 13 14 To account for the behavior of t h i s c l a s s i f i e r I assume, following Wright (1984) and Speas (1984), that i t i s f l o a t i n g ( i . e . i t i s not attached to i t s own s k e l e t a l point) and w i l l l i n k up to the s k e l e t a l s l o t of the preceding stem consonant. In examples 34, 35, and 36 th i s w i l l create an a f f r i c a t e or e j e c t i v e , but in the case of [1] < /d-1/ and [g] < /d-gh/ the [-cont] feature of [d] takes precedence over the [+cont] of the stem consonant, as the sequence /d-gh/ cannot be s y l l a b i f i e d into an a f f r i c a t e (although [dl] i s an acceptable a f f r i c a t e in C h i l c o t i n the combination /d-1/ results in [1].) D-effect . C [-cont] [+cont] [+vce] Domain of D-effeet. This rule applies exc l u s i v e l y to l e v e l 1 a f f i x e s , as t h i s i s the only place at which the environment of the rule is met. D-effect i s thus assigned to l e v e l 1. — > [-cont] [+vce] [+cont] 94 Continuant Coalescence. Whenever [hi and [1] are a f f i x e d adjacent to one another, the strategy used to eliminate t h i s continuant c l u s t e r i s coalescence, as in (38). In each verb, the [h] i s from the second person p l u r a l person marker /eh/ and the / l / i s a c l a s s i f i e r . The i i . examples reveal the underlying [1] c l a s s i f i e r . (39) a. i . [gha-na-gu-4-yax] 'you (pi.) were playing b a l l ' A adv dur derv c l s f stem 2 4 8 13 14 (FL. IRR.) i i . [gha-na-gu-^-yax) 'he was playing b a l l 1 A (FL. IRR.) adv dur derv c l s f stem 2 4 8 13 14 b. i . [ gha-4_-gwat ] 'you (pi.) were crawling' B perf c l s f stem 10 13 14 i i . [gha-l_-g wat] 'he was crawling' B perf c l s f stem 10 13 14 c. i . tbaen-ta-za-*-*?!! ] 'you (pi) swam away from i t 1 A post derv perf c l s f stem 1 9 10 13 14 (FL. IRR.) 95 i i . [ baen-ta-za-l-'?il ] 'we swam away from i t ' A post derv perf c l s f stem 1 9 10 13 14 (FL. IRR.) These data can be accounted for by the following r u l e . Continuant Coalescence C .C C C I [-vce] [+cont] [+lat ] — > t-vce] t +cont] [+lat] [-vce1 Domain of Continuant Coalescence. Again, following Kiparsky's Strong Domain Hypothesis, Continuant Coalescence i s assigned to l e v e l 1 exclusively since t h i s i s the only l e v e l where i t s environment is met. Continuant Deletion. In the examples below, the strategy used to eliminate clusters of continuants i s deletion rather than coalesccence. When two continuants occur adjacent to each other, one of them always deletes. This includes a f f r i c a t e s whose second member i s [+continuant]. The following data are organized by consonant type. In each section, a l l i . examples are the resul t of the a f f i x a t i o n of the alveolar continuant /s/ ( f i r s t person singular person marker). A l l i i . examples reveal the underlying form of the s t r i n g . In the i . examples the second continuant ( i . e . the 96 rightmost) deletes. (40) /s/ and /z/ a. i . [tae-s_-ul] /tae-s-zul/ 'I am going to s t a r t being good' A derv pm stem 8 12 14 i i . [tae-zul] 'he i s going to s t a r t being good' A derv stem 8 14 b. i . [hE-s-un] /he-s-zun/'I am good' A derv pm stem 8 12 14 i i . [hE-zun] 'he i s good' A derv stem 8 14 c. i . [hE-s-uh] /he-s-zuh/ 'I am not good' A derv pm stem 8 12 14 i i . [he-zuh] 'he i s not good' A derv stem 8 14 (41) /s/ and / t s / a. i . [tae-s_-ih] / t a e - s - t s i h / 'I am shooting' A derv pm stem 8 12 14 i i . [tae-ts_ihl 'he i s shooting' A derv stem 8 14 97 b. i . [?a-ta-za-s-i] / ? e - t e - z a - s - t s i / 'I shoot NP' A obj derv perf pm stem 5 8 10 12 14 i i . [ ?a-ta-zai-n-tsi] »you shoot NP' B obj derv perf pm stem 5 8 10 12 14 c. i . [hu-dae-s-i] /hu-dae-s-tsi/ 'I shot NP' A derv derv pm stem 8 8 12 14 i i . tho w-ah-ts_i ] 'you (pi.) shot NP' A derv pm stem 8 12 14 (42) /s/ and /y/ a. i . [dae-s_-Eh] /dae-s-yEh/ 'I arrived' A derv pm stem 8 12 14 i i . [day-yeh] /dae-in-ysh/ 'you arrived' A derv pm stem 8 12 14 (/n/ deletes a f t e r i t nasalizes the preceding vowel). b. i . [nE-s-aeh] /nE-s-yaeh/ 'I just arrived (somewhere)' perf pm stem 10 12 14 i i . [ni-yaeh] 'he just arrived (somewhere)' A perf stem 10 14 c. i . [tae-s-aeh] /tae-s-yaeh/ 'I am going to arriv e * A derv pm stem 8 12 14 98 i i . ttae-yaeh] 'he is going to a r r i v e 1 A derv stem 8 14 (43) /s/ and / l / a. i . [gha-s-Rey] /gha-s-l-Riy/ 'I vomited' A perf pm c l s f stem 10 12 13 14 i i . [ gha-^-Rey] 'he vomited' A perf c l s f stem 10 13 14 b. i . (gha-s_-gwat) /ghe-s-l-g wet/ 'I was crawling' B perf pm c l s f stem 10 12 13 14 i i . [ ghi-l_-g wat ] 'you were crawling' B perf pm c l s f stem 10 12 13 14 (the /n/ has deleted after nasalizing the preceding vowel.) c. i . [hc-s_-gi] / h e - s - l - g i / 'I run' E derv pm c l s f stem 8 12 13 14 i i . [he-l^-gi] 'you run' E derv pm c l s f stem 8 12 13 14 (44) /s/ and /*/ a. i . [tc-s-ax] /te-s-4--zaex/ 'I am s p i t t i n g ' A derv pm c l s f stem 9 10 13 14 99 I assume an i t e r a t i v e order of application of the deletion rule for t h i s example. F i r s t /z/ w i l l delete then /4/. i i . [te-4-zax] 'he i s s p i t t i n g ' A derv c l s f stem 8 13 14 b. i . [tae-s_-ghas] /tae-s-4-ghes/ 'I am going to t i c k l e NP' A derv pm c l s f stem 8 10 13 14 i i . [tae-4-ghas] 'he i s going to t i c k l e NP' A derv c l s f stem 8 10 14 c. i . [ta-za-s-katI /te-ze-s-^-ket/ 'I broke NP' A derv perf pm c l s f stem 8 10 12 13 14 i i . [ta-za-^-kat] 'he broke NP' A derv perf c l s f stem 8 10 13 14 As shown by the above data, when there i s a series of two continuants present in the s t r i n g the rightmost one deletes, leaving the f i r s t person singular /s/ i n t a c t . 1 4 The following rule accounts for these data. Continuant Deletion 1*- 3 C —> 0 / C [+cont] [+contJ [+cor1 I+cor1 t+ant] 100 Domain of Continuant Deletion. Following the Strong Domain Hypothesis Continuant Deletion can be assigned to l e v e l I since i t s environment i s not met at any other l e v e l . 4.2.4 Vowel Rounding This occurs when the optative marker /gh wc/ i s present in a s t r i n g ; the vowel to the right becomes round (after Vowel Deletion I has applied). I have included the discussion of this rule here and assume that i t applies only at l e v e l 1 because I do not have any examples of the optative mode where prefixes occur to the l e f t of /gh we/. (45) a. / g h w e - i d - d 3 e n / —> [ g h w - o d - d 3 e n ] ' l e t ' s sing' B opt pm stem 10 12 14 b. /gh we-in-d3en/ —> [ghw-on-d3En] '(they asked) you to sing' B c. /gh we-s-tJaeh/ —> [gh wa-s-tJaeh] 'I am thinking about being big' A 1 0 1 To account for these data I propose that the optative a f f i x /gh we/ has attached to i t the autosegmental feature l+round] which w i l l spread to the r i g h t . The following rule can be written. Vowel Rounding Spread the feature [+round] to the r i g h t . Per ivations; [+round] [+round] /gh we-s-zu/ —> [gh wa-s-u] 'I am thinking about being good' (Continuant Deletion has applied deleting the /z/ of the stem) (/e/ —> [a] as /e/ has no [+round] counterpart for these speakers. ) 4.3 Summary In conclusion we have the following rule system: Level 1: 1 ) Vowel Deletion I 2) Vowel Rounding 102 3) Continuant Deletion 4) i-Lowering 5) Continuant Coalescence 6) D-effeet 7) F r i c a t i v e Voicing 8) Velar Harmony 9) Alveolar Harmony Cru c i a l Ordering 4,8; 4,9. FOOTNOTES CHAPTER 4 1. It has been suggested (Kari 1975) that these variations are the vestiges of two e a r l i e r phonological processes—stem s u f f i x a t i o n and stem vowel ablaut. I w i l l not attempt to trace the C h i l c o t i n developments of these processes here. 2. The modern C h i l c o t i n c l a s s i f i e r s (d, 1, 4, and 0) derive from "a far more complex system of c l a s s i f e r s than i s found in the Athapaskan languages today" (Hoijer 1948). See also Krauss (1969) for further discussion of a pre-Athapaskan set of c l a s s i f i e r s . I w i l l not mark the 0 c l a s s i f i e r a f t e r the discussion of i t in thi s section. 103 3. Hargus (1985) has a rule of Conjugation /a/ Deletion for Sekani whereby the /a/ of the conjunct a f f i x (/gha/, /sa/ or /na/) deletes when there are no a f f i x e s intervening between i t and the stem. This appears to be the regular case in my data i f the prefi x i s assumed to be /na/. Conjunct /n/ i s found only in the t h i r d person, which is n u l l phonologically. Thus, as in Sekani, there are no a f f i x e s intervening between the conjunct marker and the stem. A form of the Sekani rule of Conjugation /a/ Deletion seems to apply to the / s i / perfective a f f i x in C h i l c o t i n and delete the vowel / i / . . In the t h i r d person s (or z) i s found rather than the f u l l / s i / (or [ z i ] ) . This can be seen in the data below. [tu ta-|-len] 'water i s flowing' A 'water' derv perf stem 9 10 14 [d3a-ta-z^tael] 'they kicked' B subj derv perf stem 6 9 10 14 [ya_z-g wat] 'he poked somebody' A obv perf stem 7 10 14 Instead of c l a s s i f y i n g t h i s /s/ as a true conjunct a f f i x as does Hargus, however, I w i l l r e s t r i c t i t to the mode set 104 (marking p e r f e c t i v e ) , and assume that t i j deletes in the t h i r d person. The perfective never co-occurs with any other modal a f f i x (though conjunct /n/ does), and functions the same as the regular / s i / perfective marker. I thus have only one conjunct a f f i x , /n/, which co-occurs f r e e l y with s e r i a t i v e /ghi/ and perfective / n i / when the subject is t h i r d person. 4. Hoijer (1948) has noted the following d i s t i n c t i o n s among the perfective a f f i x e s for Athapaskan in general: / n i / : an action that is to a point or completive /ghi/: an action that i s from a point or s t a t i c / s i / : an action that i s s t a t i c 5. I use C's to abbreviate an X s l o t that i s dominated by an onset or coda and V to abbreviate an X s l o t dominated by a nucleus. Also I assume in accordance with the central tenet of autosegmental s t a b i l i t y that deletion on the s k e l e t a l t i e r does not e n t a i l segmental deletion. 6. Evidence that the person marker / i d / was indeed a f f i x e d is found i n i t s behavior with the c l a s s i f i e r /4/. Here /d+4/—> [ 1 ]. 7. I am simply assuming the existence of l e v e l 3 here. Actual evidence for l e v e l 3 w i l l be provided in Chapter 6. 105 8. Further evidence for / n i / belonging to l e v e l 2 is found in the following example: /ni-Eh-l-'Jaeh/ --> [ni-4-"?aeh] 'you (pi.) look at i t ' A derv pm c l s f stem 8 12 13 14 If / n i / were a l e v e l 1 a f f i x , i t would be subject to Vowel Deletion I, whereby the / i / of / n i / would delete upon a f f i x a t i o n to the person marker /eh/ (second person p l u r a l ) . 9. The function of / b i / here i s not e n t i r e l y c l e a r . It occurs in 'push away', 'push into' and 'push over' but not in 'push NP'. If i t were a post-position, i t would occur to the l e f t of the adverb /nae/. See Witherspoon (1977) for a discussion of the / b i / a f f i x in Navaho. 10. The choice i s seemingly a r b i t r a r y which feature spreads. I assume leftward spread in accordance with other rules of leftward spread in my analysis. 11. F r i c a t i v e Voicing also applies to possessive pronoun-noun concatenations as shown below. [ s a i l [sak ' i ] •belt ' cow [sa - zai] 'my b e l t ' [sa - zaJc'i] 'my cow' 106 [411] 'bread' [se - l i l ] 'my bread' [4in] 'dog' [se - l i n ] 'my dog' ( A l l examples are from speaker E.) Possessive pronouns must thus be a f f i x e d at l e v e l 1. 12. This phenomenon i s common in Athapaskan languages See Krauss (1969) and Hargus (1985) for further discussion. 13. I have no examples of /d + gh/. See Krauss (1975:15). 14. There are some examples from speaker A (which I cannot account for) where the /s/ person marker has deleted. [ghi-zun] /ghi-s-zun/ 'I am good* A derv pm stem 8 12 14 [nae-?e-te-tae-tsih] /nae-?e-te-tae-s-tsih/ adv obj derv derv pm stem 1 5 8 8 12 14 'I am going to shoot' A [ni-yeh] /ni-s-yeh/ '1 just arrived (somewhere)' A perf pm stem 10 12 14 15. I am assuming that [+cont] (from the /s/) in the a f f r i c a t e / t s / w i l l p r e v a i l over the {-cont] / t / . However, when the a f f r i c a t e following /s/ i s the derived a f f r i c a t e [dz] (d 107 c l a s s i f i e r plus /z/ stem), the /s/ does not delete. Perhaps at t h i s stage the /d/ c l a s s i f i e r and the [z] i n i t i a l stem have not been s y l l a b i f i e d into an a f f r i c a t e . [na-gha-s-d-zun] 'I am good again' A adv derv pm c l s f stem 2 9 12 13 14 [na-gha-s-d-zuhl 'I often become good' A adv ser pm c l s f stem 2 10 12 13 14 [nae-te-s-d-zun) I am going to be good again' A adv derv pm c l s f stem 2 9 12 13 14 108 CHAPTER 5: THE LEXICAL PHONOLOGY OF CHILCOTIN: LEVEL 2 5.1 Morphology The following l i s t of a f f i x e s comprises l e v e l 2: 5 6 7 8 Object Subject Obviative Derivative Pronouns e (unspecified) dse (3pl.) ye dae hae (areal) ts'e ni se (lsg.) ye ne (2sg.) gu g we (3sg.) de we ( l p l . ) i WE (2pl.) nae g webE (3pl.) ghe ghi hu tae Affixes 5, 6, and 7 are semantically similar in that they a l l express information regarding the subject or object of the verb. The derivative a f f i x e s (8) are mainly i d i o s y n c r a t i c (as were the derivative a f f i x e s of l e v e l 1). 5.1.1 Derivative Affixes Listed below are the stems, with examples, in which the i d i o s y n c r a t i c derivative a f f i x e s occur, followed by the non-idiosyncratic a f f i x e s [hu] and [tae]. 109 (1) /dae/ ' f l o a t ' , 'shake', ' f a l l ' , ' r o l l ' , 'stop', 'follow 1, 'arrive', 'burn' a. [te-dae-s-le4] 'I am f l o a t i n g ' A adv derv pm stem 2 8 12 14 b. [dae-za-s-ta] 'I am shaking' B (FL. IRR.) derv perf pm stem 8 10 12 14 c. [da-ta-za-s-tsat] 'I f e l l down' A derv derv perf pm stem 8 9 10 12 14 d. NP [g we-da-y-4-mell 'I r o l l e d the NP (down the h i l D ' E obj derv derv c l s f stem 5 8 8 13 14 e. NP [se-da-y-4-ti 3 'NP stopped me' E obj derv derv c l s f stem 5 8 8 13 14 f. NP [ d 3 e - d a - y - n - d i l ] •they followed me' A subj derv derv conj stem 6 8 8 11 14 [dse-da-y-n-dil] 'they just a r r i v e d ' A subj derv derv conj stem 6 8 8 11 14 (2) / n i / 'look at', 'fence' a. (g we-ni-4-vaeh] 'you (pi.) look at i t ' A obj derv c l s f stem 5 8 13 14 110 b. [na-ne-s-tlon ] 'fence' (Krauss 1975) (3) /ye/ 'think' a. [dsi-ye-ne-zun] 'they are thinking' A subj derv derv stem 6 8 9 14 (4) /gu/ 'happy', 'want', 'be*, 'hungry' a. [d5£-gu-n-t'in] 'they are happy' D subj derv conj stem 6 8 11 14 b. [ 4-aen g u - l i ] 'there i s a l o t * D a l o t derv stem 8 14 c. [go-ze-n-t'in] 'you want (to sing) B (FL. IRR.) derv perf conj stem 8 10 11 14 d. [go-ta-zo] 'we are going to be hungry' D derv derv stem 8 8 14 (See Hargus pg. I l l regarding occurence of more than one d e r i v a t i o n a l a f f i x of the same class.) (5) /de/ 'grind' a. NP [da-na-l-daz] 'ground NP' A derv derv c l s f stem 8 9 13 14 I l l (6) / i / (This a f f i x always co-occurs with another a f f i x from l e v e l 2) a. NP I 2 i - g u - t ' i n ] »it looks l i k e NP * E derv derv stem 8 8 14 (I assume that t h i s /"?i/ i s the same a f f i x that occurs in /d3e-ye-nae-i-4-9in/ 'they looked at i t * A. It must therefore be a l e v e l 2 and not a l e v e l 3 a f f i x since i t occurs to the ri g h t of the obviative a f f i x . ) In the following examples / i / has changed to [y] due to i t s position in the s y l l a b l e . b. [ha-y-yeh] 'I just a r r i v e d 1 A derv derv stem 8 8 14 c. tda-y-yeh] 'I just arrived (somewhere s p e c i f i c ) 1 A derv derv stem 8 8 14 d. [ko gwe-da-y_-4-k'aen] 'the house burned down' E house obj derv derv c l s f stem 5 8 8 13 14 e. NP [g we-da-£-4-mel] 'I r o l l e d the NP down the h i l l ' E obj derv derv c l s f stem 5 8 8 13 14 (7) /nae/. This a f f i x i s homophonous with the l e v e l 3 durative a f f i x . Due to i t s placement in b. and c. to the right of the 112 obviative and object a f f i x e s , however, i t must be a derivative a f f i x of position 8. a. NP [ g w e - n a e - s - g W a t ] 'I am shaking NP' B obj derv pm stem 5 8 12 14 b. [d5i-ye-nay-4-?in] 'they looked at i t ' A subj obv derv c l s f stem 6 7 8 13 14 In the l a s t example, the /nae/ diphtongizes to [nay] due to the presence of the underlying / i / d e r i v a t i v e . (8) /ghe/ 'frighten' a. [na-gha-ne-d3ut] ' i t frightened you' A obj derv perf stem 5 8 10 14 b. Isa-gha-ne-d3ut] ' i t frightened me' A obj derv perf stem 5 8 10 14 /hu/ and /tae/ are the only a f f i x e s that are not completely i d i o s y n c r a t i c in their a f f i x a t i o n . /hu/ occurs in s e r i a t i v e verbs and t h e i r semelfactive counterparts. (9) /hu/ a. Ihu-s-d3it] 'I am poking her leg' A derv pm stem 8 12 14 113 b. [na4ey hu-ta4] 'the horse i s kicking us' B horse derv stem 8 14 c. (gax nae-hu-dae-tsi] 'he i s shooting many rabbits' A rabbit adv derv derv stem (FL. IRR.) 2 8 8 14 d. [gax hu-dc-ta-n-tsax] rabbit adv derv derv pm stem 2 8 8 12 14 'you are going to shoot a l o t of rabbits' A e. [hu-ni-n-tsi] 'you shot i t once' A derv perf pm stem 8 10 12 14 f. [ho-ghe-n-tsi1 'you kicked i t many times' B derv perf pm stem 8 10 12 14 g. [ho-ghe-tsi] 'we shot i t many times' A derv perf stem 8 10 14 (In f. and g., the /u/ of /hu/ has flattened to /o/.) When a l e v e l 2 a f f i x occurs to the l e f t of /hu/ there is a coalescence. The h of /hu/ and the vowel of the object a f f i x do not surface. (10) a. [Mary s-u-tae4] /se-hu-tae4/ 'Mary i s kicking me1 A obj derv stem 5 8 14 114 b. [Mary n-u-tae4] /ne-hu-taeV 'Mary is kicking you' obj derv stem 5 8 14 c. [Mary £-u-tae4] /ye-hu-tae4/ 'Mary i s kicking him' A obv derv stem 5 8 14 /hu/ also occurs i d i o s y n c r a t i c a l l y in the following paradigms (11) a. [hu-4-nez] 'how t a l l i s she?' A derv c l s f stem 8 13 14 b. 'Heather' [hu-n-draen] 'Heather i s shy' A derv conj stem 8 11 14 Because the /h/ of /hu/ does not surface when another a f f i x of le v e l 2 is present I w i l l assume that the /h/ is f l o a t i n g and is only r e a l i z e d when an onet s l o t i s created (see Onset Formation 5.2.2). /tae/. This a f f i x usually co-occurs with the modal /ghe/ to create the inceptive mode. If /ghe/ i s present, /tae/ always fl a t t e n s to [ t a ] . (12) a. [ta-gha-t'as] 'he i s going' A derv perf stem 8 10 14 115 b. [ta-gha-t111 'we are going to sleep* A derv perf stem 8 10 14 c. [ta-gha-h-zul] 'you (pi.) are going to be good* A derv perf pm stem 8 10 12 14 d. [?e-tae-s-dael] 'I am going hunting' A obj derv pm stem 5 8 12 14 e. [tae-yael] *he is going to go* A derv stem 8 14 5.1.2 Obviative This a f f i x occurs when both subject and object are t h i r d person, as in the following: (13) a. [j[£-tE-l-zax] 'he is s p i t t i n g NP * A obv derv c l s f stem 7 8 13 14 /ye/ becomes [yo] in the s e r i a t i v e /ghi/ mode. b. [dgi-yo-ghi-n-tael] 'they kicked them many times' B subj obv ser conj stem (FL. IRR.) 6 7 10 12 14 116 c. t d 3 i - y o - g h e - n - t s l ] 'they shot i t ' A subj obv perf conj stem 6 7 10 12 14 Although t h i s a f f i x never co-occurs wiht an object a f f i x , i t cannot be classed with the position 5 a f f i x e s as i t always occurs to the ri g h t of the subject marker, position 6. 5.1.3 Subject Affixes This position i s f i l l e d by two a f f i x e s — / d 3 e / t h i r d person p l u r a l , and /ts'e/ fourth person p l u r a l (which corresponds to 1 one• in English). (14) /dse/ a. [ nae-d_3£-t ih ] 'they are dreaming' C dur subj stem 4 6 14 b. [henes bi-nae-d_3e-kayn ] 'they are going in a r a f t ' E r a f t adv dur subj stem 2 4 6 14 The vowel q u a l i t y of / d 3 E / varies in accordance with the aspect of the verb. It takes the form [ d 3 a e ] when the / g h E / mode is part of the paradigm. However, t h i s mode is usually absent in the t h i r d person forms. In the data below, the ( i i ) examples are non-third person and reveal the presence of the / g h E / modal a f f i x for each paradigm. 117 (15) a i . [nae-dsae-d-zun] 'they are good again' A adv subj c l s f stem 2 6 13 14 i i . [na-gha-h-d-zun] 'you (pi.) are good again' A adv perf pm c l s f stem 2 10 12 13 14 b. i . [dsae-l-g wat] 'they are crawling' A subj c l s f stem 6 13 14 i i . [gha-l-g wat] 'he is crawling' B perf c l s f stem 10 13 14 c. i . [dsae-dsen] 'they sang 1 E subj stem 6 14 i i . [ g h - e n - d 5 e n ] 'you sang' E perf pm stem 10 12 14 / d 3 £ / also changes to [dsi] as in [ d 3 i - n - z u n ] 'they are good' A. In the optative, / d 3 e / becomes [dsu]. 118 (16) a. [dsu-dsen] '(they asked us) i f we would sing' B subj stem 6 14 b. [d3u-zu] 'they are thinking about being good 1 B subj stem 6 14 /ts'e/ occurs only in formal speech. (17) [ t s ' e - d e - n i - l - t E l l 'one shot at i t ' B subj derv perf c l s f stem 6 9 10 13 14 5.1.4 Object Affixes This includes the unspecified object / E / and the unspecified areal /hae/, as well as the pronominal d i r e c t objects. (18) / E / a. [ * ? E - d 5 E - t g - t s ih ] 'they are shooting' B obj subj derv stem 5 6 9 14 b. [?a-ta-za-s-tael] .'I kicked ' A obj derv perf pm stem 5 9 10 12 14 (19) /hae/ a. [hae-tae-s-ael] 'I am going to arr i v e (anywhere)' A obj derv pm stem 5 8 12 14 119 b. [hae-dsa-y-n-dll1 'they arrived there" A obj subj derv conj stem 5 6 8 11 14 c. [sae hae-?ael] 'the sun i s up (there)' A sun obj stem 5 14 d. l ^ a e l dzi ha-y-?en] 'the moon i s out (there)' A moon obj derv stem 5 8 14 A e. [ t l ' a da-ha-y-kai 4] 'we are going to s i t down (there)'C there ? obj derv stem 5 8 14 The pronominal d i r e c t objects are l i s t e d below. S ingular P l u r a l se we ne WE „w_. „w g we be 5 . 2 Level 2_ Phonology One phonological process that distinguishes l e v e l 2 from l e v e l 1 i s Vowel Deletion, which occurs (as was discussed e a r l i e r ) when two vowels are af f i x e d adjacent to each other in the verb. 5.2.1 Vowel Deletion II In section 4.7, i t was shown that when a series of vowels was produced by a f f i x a t i o n at level.1 the leftmost vowel 120 deleted (Vowel Deletion I ) . The reverse occurs at l e v e l 2. Here, when there are 2 adjacent vowels in the s t r i n g , the rightmost one deletes. (20) a. /dae-eh-dil/ —> [dae-h-dil] 'you (pi.) arrived (somewhere s p e c i f i c ) ' derv pm stem 8 12 14 b. /hu-in-tsax/ —> [hu-n-tsax] 'you are shooting i t 1 A derv pm stem 8 12 14 c. /ni-eh-l-^aeh/ —> [ni-4-^aeh] derv pm c l s f stem 8 12 13 14 d. /ye-nae-eh-zen/ —> [ye-nae-h-ZEn] obv derv pm stem 7 8 12 14 'you (pi.) are thinking about being good' A 'you (pi.) looked at i f (A) 121 e. / t a e - i n - z u l / —> [tae-n-zul] 'you are going to s t a r t being good' derv pm stem 8 12 14 In each case, the vowel of the person marker (the rightmost vowel) has deleted. To account for these data the following rule can be written: Vowel Deletion II V —> 0 / V (These are V s on the sk e l e t a l t i e r . ) 5.2.2 Onset Formation Another phonological process that distinguishes l e v e l 2 from l e v e l 1 i s Onset Formation. At l e v e l 2, any vowel i n i t i a l a f f i x e s (including the a f f i x /hu/ whose /h/ i s unattached) receive an onset s l o t . The unattached /h/ w i l l then link up to this newly created onset and the a f f i x w i l l be rea l i z e d as [hu]. In a l l other cases the empty onset s l o t w i l l be f i l l e d in p o s t - l e x i c a l l y with the default consonant /*>/. Onset Formation x I [ V x x I I — > [ V 122 Although t h i s rule i s structure-building and not structure-changing, i t does obey the SCC and applies only to str i n g s that have had a previous morphological or phonological rule of l e v e l 2 apply to them. This rules out application of Onset Formation to strings that enter l e v e l 2 with the vowel i n i t i a l a f f i x e s / i d / , / i n / , and /eh/ (12). Although they meet the environmental conditions of Onset Formation, these strings w i l l enter l e v e l 3 unchanged unless they have independently met the conditons for Vowel Deletion. Onset Formation does apply to vowel i n i t i a l a f f i x e s of l e v e l 2 as shown in example (21). (21) a. [7_i_-gu-t1 in) ' i t looks l i k e (something) 1 E derv derv stem 8 8 14 b. (4in ? a - t a - z a - s - t a e l 1 'I kicked the dog' A dog obj derv perf pm stem 5 9 10 12 14 c. [ ? e - d 3 e - t e - t s i n ] 'they are shooting into a group ' B obj subj derv stem 5 6 9 14 When a l e v e l 3 a f f i x i s added, the empty onset s l o t prevents vowel deletion, as shown in example 22. (22) a. tbi-nae-"?a - d 3 a z ] 'something to write with' (pencil) A post adv obj stem 1 2 5 14 123 b. fnae-?E-te-tae-tslh] he is going to shoot' A adv obj derv derv stem 2 5 8 8 14 Onset Formation also applies at l e v e l 3 as well as p o s t - l e x i c a l l y , as shown in (21). (23) a. [?ae-nae-n-t'in] 'you work' E adv dur pm stem 2 4 12 14 b. f2£skai] ' c h i l d ' E c. [ ? i n t s i ] 'grandfather' E d. [?aeldzi] 'moon' A Onset Formation must then "turn on" at l e v e l 2, and continue to apply throughout the l e x i c a l and p o s t - l e x i c a l l e v e l s . Having established the d i s t i n c t i o n between levels 1 and 2 via Vowel Deletion II and Onset Formation, I now turn to a discussion of the phonological processes that create the d i s t i n c t i o n between levels 2 and 3: F r i c a t i v e Voicing, and Diphthongization. 5.2.3 F r i c a t i v e Voicing Another rule that supports the d i s t i n c t i o n between levels 2 and 3 i s F r i c a t i v e Voicing. It was shown in Chapter 4 that alveolar f r i c a t i v e s become voiced i n t e r v o c a l i c a l l y at level 1. The same i s true at l e v e l 2. 124 (24) a. [ g w a - z a i - t a 4 ] /gW e-si-tE4/ 'I kicked i t ' B obj p e r f stem 5 10 14 b. [na4ey d3a-za-ta4] /d3E-§E-te4/ 'they k i c k e d the horse' (B) horse subj perf stem 6 10 14 c. [ g u - z a - s - t ' i n ] / g u - s i - s - t ' i n / 'I want to s i n g ' B derv p e r f pm stem (FL. IRR.) 8 10 12 14 F r i c a t i v e s do not voice at l e v e l 3. (25) a. [na-sa-s-d3itJ 'I crawled' B adv perf pm stem 2 10 12 14 b. [na-s^ai-tin] I dreamt' C dur perf stem 4 10 14 In each case in (25), the perfective marker has remained as A voiceless /s/ although i t is in i n t e r v o c a l i c p o s i t i o n . 5.2.4 Diphthongization The t h i r d process that establishes a d i s t i n c t i o n between levels 2 and 3 i s Diphthongization. When an a f f i x of the form /Cae/ (where C stands for any consonant) i s followed by an / i / - i n i t i a l a f f i x , the r e s u l t is [Cay] as shown in examples (26) through (30). 125 (26) a. [da-y-n-le4] /dae-in-le4/ 'you are f l o a t i n g ' A derv pm stem 8 12 14 b. [da-y_-d-le4] /dae-id-le4/ 'we are f l o a t i n g ' A derv pm stem 8 12 14 c. [dae-le4] /dae-le4/ 'he i s f l o a t i n g ' A derv stem 8 14 d. [d5e-dae-le4] /d3e-dae-le4/ 'they are f l o a t i n g ' A subj derv stem 6 8 14 (27) a. [da-y-yeh] /dae-i-yeh/ 'he arrived' A derv derv stem 8 8 14 b. [d5E-da-y-n-dil] /dge-dae-i-n-dil/ 'they a r r i v e d ' A subj derv derv conj stem 6 8 8 11 14 c. [da-y-yeh] /dae-i-in-yeh/ 'you arrived' A derv derv pm stem 8 8 12 14 (28) a. [na-y-4-?in] /nae-i-4-?in/ 'he looked at i t ' A dur derv c l s f stem 4 8 13 14 b. [ d 3 i - y e - n a - y - 4 - 9 j n ] /d3£-ye-nae_-i_-4-^in/ subj obv derv derv c l s f stem 6 7 8 8 13 14 •they looked at i t ' A 126 c. [na-£-4-'?in] /nae-l-in-4-?in] 'you looked at i t ' A derv derv pm c l s f stem 8 8 12 13 14 (29) [ko g^-da-y-^-k'en] /ko g we-dae-i-4-k'en/ house obj derv derv c l s f stem 5 8 8 13 14 'I burned the house 1 E (30) [detsen g we-da-y-4-mel] /detsen g we-dae-i-4—mel/ log obj derv derv c l s f stem 5 8 8 13 14 •I r o l l e d the log (down the h i l l ) ' E Nuclear Fusion 1 changes the prosodic organization of the sequence / a e - i / by combining them under one nucleus node. Nuclear Fusion N N N N X X X X • « I I I ae i ae i Low l e v e l phonetic rules w i l l change / i / to [y] due to i t s postion in the s y l l a b l e - the second half of a branching nucleus. The Obligatory Contour P r i n c i p l e (OCP) w i l l change /ae-y/ to [ay] since both [ae] and [y] are [-back]. Because [ae] i s the more marked element i t w i l l lose i t s [-back] feature and become [a] . 127 DomaIn of Nuclear Fusion. Nuclear Fusion can be ordered at l e v e l 1, in accordance with the Strong Domain Hypothesis, since there are no instances of the seguence / a e - i / at le v e l 1. It does not, however, apply at l e v e l 3, as shown by the data below. Instead Vowel Deletion II w i l l apply deleting the rightmost vowel. (31) a. [nae-dsit1 /nae-in-dsit/ 'you crawled around' B adv pm stem 2 12 14 b. [nae-d3it] /nae-id-d5it/ 'we crawled around' B adv pm stem 2 12 14 5.2.5 E-Raising Unlike i-Lowering at le v e l 1 where / i / — > [ E] / s, /e/ raises to [ i ] at le v e l 2 when i t occurs before [ y j , as can be seen in the following data where /d3e/ becomes [d3i] when i t is aff i x e d to the l e f t of the [y] i n i t i a l obviative a f f i x . (32) a. [d3i_-v_o-ghe-n-tsi ] 'they shot i t (many times)' A subj obv ser conj stem 6 7 10 11 14 b. td3i_-y_o-ghe-n-tael ] 'they kicked i t (many times)' B subj obv ser conj stem 6 7 10 11 14 c. [na4ey d3^-ya-z-tat] 'they kicked the horse' B horse subj derv perf stem 6 8 10 14 128 I n c o n t r a s t t o t h i s , the f o l l o w i n g examples i l l u s t r a t e t h e u n d e r l y i n g form of the s u b j e c t p r e f i x ( / d j c / ) . (33) a. tnae-d3E-d-zuh] 'they o f t e n become good' A adv s u b j c l s f stem 2 6 13 14 b. [ d 3 E - t i - 4 - k a t ] 'they broke the window' A s u b j d e r v c l s f stem 6 8 13 14 The f o l l o w i n g r u l e a c c o u n t s f o r / E / r a i s i n g t o I i ] . E - R a i s i n g v . c * * • [+cont] " • • - . I [-low] [+high] [-back] Domain of E-Ra i s i n g . I have no e v i d e n c e of t h i s r u l e a p p l y i n g e l s e w h e r e i n my d a t a , s i n c e the sequence / E - y/ i s found o n l y a t l e v e l 2. A l t h o u g h t h i s r u l e a p p l i e s o n l y t o l e v e l 2 a f f i x e s , i t i s a s s i g n e d t o l e v e l 1 f o l l o w i n g K i p a r s k y ( 1 9 84), s i n c e t h i s i s the l e a s t marked l e v e l f o r a s s i g n i n g r u l e s . Vowel D e l e t i o n I I I Repeated below i s example 10 from 5.1.1 The l a b e l l e d 129 brackets refer to the l e v e l at which the a f f i x was added. (Recall that while /h/ is present on the segmental t i e r i t has no corresponding s k e l e t a l point and i s thus i n v i s i b l e to the vowel deletion r u l e s ) . (10) a. [Mary s-u-tae4] 2 [ s e - 2 f h u - 1 [ t a e 4 ] ] 'Mary i s kicking me' A obj derv stem 5 8 14 b. [Mary n-u-tae4] 2 * n E " 2 f h u ~ 1 [ t a e 4 ] ] 'Mary is kicking obj derv stem you' B 5 8 14 c. [Mary y-u-tae4] 2[ye- 2[hu-^ftae4]] 'Mary i s kicking him' B obv derv stem 7 8 14 Vowel Deletion II would delete the rightmost vowel /u/. However, t h i s would give the incorrect r e s u l t . It i s the leftmost vowel /e/ that deletes leaving the /u/ in t a c t . There must then be a t h i r d rule of vowel deletion that w i l l delete the leftmost vowel for the examples above but w i l l leave i t intact in example 20 of section 5.2.1. One difference between the examples in 10 and those in 20 130 i s that a l l of the examples in 10 are the re s u l t of a l e v e l 2 a f f i x added to another l e v e l 2 a f f i x . In the examples in 20 the CV a f f i x e s of l e v e l 2 are added to a l e v e l 1 a f f i x (a person marker). The t h i r d rule of vowel deletion can refer to th i s inner bracketing since the inner brackets are not erased u n t i l the end of a l e v e l (see Kiparsky 1982). Vowel Deletion III 2 t C V 2 t V --> 2 t C 2 [ v 5.3 Conclus i on The following rules apply to le v e l 2 a f f i x e s . I have simply l i s t e d the rules that have been discussed in t h i s Chapter. At the end of Chapter 6 the rules w i l l be re-ordered at the optimal l e v e l at which they can apply. 1. Vowel Deletion II 2. Onset Formation 3. F r i c a t i v e Voicing 4. Nuclear Fusion 5. e-Raising 6. Vowel Deletion III 131 FOOTNOTES CHAPTER 5 1. This was suggested to me by Dr. P a t r i c i a Shaw. 132 CHAPTER 6: THE LEXICAL PHONOLOGY OF CHILCOTIN: LEVEL 3 6.1 Morphology The following a f f i x positions comprise l e v e l 3: 1 2 3 4 postposition adverb stem durative 6.1.1 Durative The durative a f f i x /nae/ occurs in verbs where the action described spreads over a period of time. This i s opposed to verbs that express an instantaneous action (such as 'kick' and •shoot'). (1) a. [nae-sai-tin] 'I dreamt' C dur perf stem 4 10 14 b. Ina-se-bin] 'I swam' A dur perf stem 4 10 14 c. [na-ghi-tsun] 'you kissed him long' B (FL. IRR.) dur perf stem 4 10 14 d. [na-gha-s-d-zun] 'I am good again' A dur perf pm c l s f stem 4 10 12 13 14 133 e. [nae-g wg-d5el] ' i t snowed' E dur obj stem 4 5 14 f. [?ae-nae-tae-s-t'in] 'I am going to work* E adv dur derv pm stem 2 4 8 10 14 6.1.2 Stems I have only one example of an incorporated stem in my data, although incorporation is a productive process in C h i l c o t i n . (2) a. £bi-na-zai-g wa-k'an] 'something to color your mouth with ( l i p s t i c k ) ' A post poss noun obj stem 1 3 5 14 b. [sa-zai] 'my mouth' E my mouth 6.1.3 Adverbs There is a va r i e t y of i d i o s y n c r a t i c l e v e l 3 adverbs. A few examples follow. (3) a. [gha-na-g we-d5E-l-yax] 'they are playing b a l l ' A adv dur obj subj c l s f stem (FL. IRR.) 2 4 5 6 13 14 b. [tae-dsE-daen] 'they are drinking' A adv subj stem 2 6 14 134 c. [ t e - d 5 E - d a e - l c 4 ] 'they are f l o a t i n g ' A adv subj derv stem 2 6 8 14 d. [hae-dge-del] 'they are saying' D adv subj stem 2 6 14 e. [te-nae-dgi-ya-gha-l-tJut] 'they caught i t again' A adv adv subj obv perf c l s f stem 2 2 6 7 10 13 14 f. [ n a e - b i - t i - z i - t s a t ] 'he pushed you around' B adv ? derv perf stem 2 9 10 14 (The speech of B frequently did not manifest flattening.) g. [bi-na-gha-s-Rwey1 'I vomited i t ' A post adv perf pm stem 1 2 10 12 14 h. [?ae-nae-n-t'in] 'you are working' E adv dur pm stem 2 4 12 14 i . f?ae-n-lal 'NP made NP' E (FL. IRR.) adv conj stem 2 11 14 6.1.4 Postpositions There are two types of postpositions. One i s a combination of postposition and pronoun (e.g. 'from'-pronoun). The other i s /bi/, usually translated as 'with' 135 (4) 'from'-pronoun a. [naen-ta-ze-n-bin] 'you swam away from you' A post derv perf pm stem 1 9 10 12 14 b. [saen-ta-ze-n-binl 'you swam away from me' A post derv perf pm stem 1 9 10 12 14 c. [baen-ta-ze-n-bin] 'you swam away from them' A post derv perf pm stem 1 9 10 12 14 (5) / b i / a. [bi-tae-daenl 'something to drink with (cup)' A post adv stem 1 2 14 b. [bi-nae-^a-dgaz] 'something to write with ( p e n c i l ) ' A post adv obj stem 1 2 5 14 fbi-nae-na-ka] 'something to make holes with (sewing machine)' A post adv dur stem 1 2 4 14 [bi-yae-s-tak] 'I am ta l k i n g into something' A post derv pm stem 1 8 12 14 e. [bi-tae-l-daen] 'make her drink' A c. d. post adv c l s f stem 1 2 13 14 136 f . [NP bl-d5e-tae-l-zax) 'they are going to s p i t 1 A post subj derv c l s f stem 1 6 8 13 14 (See Witherspoon (1977) for further discussion of the / b i / a f f i x in Athapaskan.) 6.2 Level 3_ Phonology There are only two rules that apply to l e v e l 3 a f f i x e s -Vowel Deletion II and Onset Formation. No new rules are assigned to l e v e l 3. With the exception of Vowel Deletion II and Onset Formation, a l l other l e v e l 1 and l e v e l 2 rules shut-off. 6.2.1 Vowel Deletion II As was shown in Chapter 5, when a series of 2 vowels occurs the rightmost one deletes at l e v e l 2. This i s also true at l e v e l 3. (6) a. /tae-in-daen/ —> [tae-daen] 'you are drinking' C adv pm stem 2 12 14 b. /tae-ch-daen/ —> [tae-daenJ 'you (pi.) are drinking' C adv pm stem 2 12 14 c. /?ae-nae-in-t'in/ —> [?ae-nae-n-t'in] 'you are working' E adv dur pm stem 2 4 12 14 137 6.2.2 Onset Formation It was also shown in Chapter 5 that Onset Formation applies at l e v e l 3 (as well as word i n t e r n a l l y at l e v e l 2), creating an empty onset s l o t before vowel i n i t i a l a f f i x e s (which w i l l be f i l l e d in by the default consonant /*?/). Example (23)a from Chapter 5 i s repeated below. [?ae-nae-n-t'in] 'you work' E adv dur pm stem 2 4 12 14 6.3 Summary Only 2 phonological rules apply at l e v e l 3- Vowel Deletion II, and Onset Formation. A l l other rules that applied at l e v e l s 1 and 2 have shut-off. Listed below ( i n th e i r optimal order) is the rule system that has been developed in thi s thesis for the l e x i c a l component of C h i l c o t i n . LEVEL 1 1. Vowel Deletion I 2. Vowel Rounding 3. Continuant Deletion 4. i-Lowering 5. d-Effect 6. F r i c a t i v e Voicing 7. e-Raising 8. Continuant Coalescence 138 9. Velar Harmony 10. Alveolar Harmony C r u c i a l l y ordered; 4,9; 4,10 LEVEL 2 Vowel Deletion I shuts off i-Lowering shuts off (I assume that a l l other level 1 rules that have not shut-off may also apply here. I have not ,however, l i s t e d them). 1. Vowel Deletion II 2. Vowel Deletion I I I 3. Onset Formation 4. Nuclear Fusion LEVEL 3 F r i c a t i v e Voicing shuts off Nuclear Fusion shuts o f f 139 CHAPTER 7 : POST-LEXICAL PHONOLOGY 7.0 Introduction P o s t - l e x i c a l rules apply "across the board" wherever th e i r environment i s met (Kiparsky 1982. Nasalization, /ghe/ contraction, /gh/ Deletion, and Onset Default are a l l p o s t - l e x i c a l rules in C h i l c o t i n . 7.1 Nasalization Nasal vowels occur when underlyingly they are followed by a homo-syllabic nasal which is in turn followed by a continuant. There is v a r i a t i o n in my data as to whether or not the following continuant must be homosyllabic with the nasal consonant (see Cook 1986). In example 1 below the continuant i s homosyllabic with the nasal consonant and a nasal vowel r e s u l t s . (1) a. [bi-ta-4-zax] /bi-tae-in-zex/ 'you w i l l s p i t ' A post derv pm stem 1 8 12 14 b. [tajg-4-tak] /tae-in-4-tek/ 'you w i l l shoot' A derv pm c l s f stem 8 12 13 14 c. [ta-ze-l-gey] / t e - s i - i n - l - g i y ] 'you are walking' E derv perf pm c l s f stem 8 10 12 13 14 140 d. [ha-l-gi] / h e - i n - l - g i ) 'you run' E derv pm c l s f stem 8 12 13 14 In example 2 the vowel becomes nasalized although the continuant i s not homosyllabic. (2) a. [da-y-yeh] /dae-in-yeh/ 'you arrived (somewhere)' A derv pm stem 8 12 14 b. [hi-zOh] /he-in-ztfh/ *you are not good 1 A derv pm stem 8 12 14 c. [ i 4 i l / i n 4 i / 'one E Example 3 appears to be counter evidence to example 2. Here the vowel does not nasalize when the following continuant is not homosyllabic. (3) a. [ g h i - n - l i ] / g h i - n - l i / ' i t used to be that way' A mode conj stem 10 11 14 b. [ta-n-zax] /tae-n-zex/ 'he is going to s p i t ' A derv conj stem 8 11 14 c. C?ae-n-la] V?ae-n-laegh/ 'he i s making NP' E adv conj stem 1 11 14 Further research needs to be done on the interaction between tone and nasa l i z a t i o n . In a l l of the examples that did 141 not nasalize the a f f i x involved was the conjunct. I w i l l not attempt to write a rule here as I think the nasalization process i s related to the tone r u l e s . 7.2 /gh/ Contraction In fast speech the perfective marker /ghe/ i s often deleted in i n t e r v o c a l i c position a f t e r having retracted the vowel to i t s l e f t . A l l of the cases of /ghe/ contraction involve the inceptive mode: /tae/ (8) and /ghe/ (10). Without both /tae/ and /ghe/ having been present, the verb cannot be translated as the inceptive. This, and the fact that the vowel to the l e f t of the missing /ghe/ i s [+RTR], rule out the p o s s i b i l t y of /ghe/ never having been added to the s t r i n g at a l l . (4) a. /tae-ghe-t'es/ --> [ta-t'es] 'he is going to cut' A derv perf stem 8 10 14 b. /tae-ghe-in-tse4/ --> [ta-n-tse4] 'you are going to swing' E derv perf pm stem 8 10 12 14 c. /gutja gu-tae-ghe-zu/ —> [gut/a go-ta-zo] insides derv derv perf stem 8 8 10 14 'their insides w i l l be numb (they are going to be hungry)' E This process can be accounted for by the following r u l e . 142 gh Contraction gh —> 0 / V V (the vowel of /ghe/ w i l l delete v i a one of the vowel deletion r u l e s ) . 7.3 Velar Deletion The [+RTR] velar /gh/ also deletes in s y l l a b l e f i n a l p o s i t i o n a f t e r RTR Harmony and nasalization have applied. The only examples of s y l l a b l e f i n a l /gh/ are in non-derived l e x i c a l items. (5) a. /kungh/ --> [ko] 'house' E b. /naengh4-ingh/ —> [na4ey] 'horse' E c. /bilugh/ —> [bilo] 'knife' E d. /midugh/—> [mido] 'whiteman' E. The following rule can account for these data. Velar Deletion r ime I gh — > 0 / 143 This rule must be ordered after RTR Velar Harmony due to the fact that the adjacent vowels are a l l [+RTR]. 7.4 Onset Default It was shown e a r l i e r that vowel i n i t i a l a f f i x e s a l l receive an onset s l o t at levels 2 and 3. This s l o t i s f i l l e d in with /?/ p o s t l e x i c a l l y . This i s i l l u s t r a t e d in example 6. (6) a. C?i-gu-t•in] ' i t looks l i k e (something)' E derv derv stem 8 8 14 b. [nae-^e-te-tae-tsih] 'he is going to shoot' A adv obj derv derv stem 2 5 8 8 14 c. Ibi-nae-^a-dsaz] 'something to write with' A post adv obj stem 1 2 5 14 The following rule accounts for these data. Onset Default onset rime » 1 144 7.5 Summary The following rules apply at the p o s t - l e x i c a l l e v e l . 1. Nasalization 2. gh Contraction 3. Velar Deletion 4. Onset Default 7.6 Summary of Thesis In conclusion, the following rule system has been developed. LEVEL 1 1. Vowel Deletion I 2. Vowel Rounding 3. Continuant Deletion 4. i-Lower ing 5. D-effeet 6. F r i c a t i v e Voicing 7. Continuant Coalescence 8. e-Raising 9. Alveolar Harmony 10. Velar Harmony Cruci a l Orderings; 4,9; 4,10. 145 LEVEL 2 Vowel Deletion I shuts off i-Lowering shuts off 1. Onset Formation 2. Vowel Deletion II 3. Vowel Deletion III 4. Nuclear Fusion LEVEL 3_ F r i c a t i v e Voicing shuts off Nuclear Fusion shuts off POST LEXICAL 1. Velar Deletion 2. gh Contraction 3. Onset Default It has been shown that some complex issues of C h i l c o t i n morphology and phonology become clearer when the theory of l e x i c a l phonology i s used to analyse them. The c l u s t e r i n g of rules around groups of a f f i x e s is a consequence of phonological 146 rules being assigned to morphological l e v e l s . What seems to be an i n t r i c a t e system of rules becomes somewhat simpler where rule a p p l i c a t i o n i s determined by conditions such as the S t r i c t Cycle Condition (Kiparsky 1982) (sections 3.1 and 5.2.2), the Strong Domain Hypothesis (Kiparsky 1982) (section 5.5.2) and the Obligatory Contour P r i n c i p l e (section 5.2.5). 147 Abbreviations Used in This Thesis adv Adverb c l s f C l a s s i f i e r conj Conjunct derv Derivative dur Durative FL. IRR. Flattening Irregular imp Imperfective LP Lexical Phonology obj Object obv Obviative OCP Obligatory Contour P r i n c i p l e perf Perfective p i . p l u r a l pm Person Marker post Postposition SCC S t r i c t Cycle Condition SDH Strong Domain Hypothesis ser s e r i a t i v e sg. singular subj Subject WFC Well Formedness Condition 148 REFERENCES Anderson, Stephen R. (1974). The Organization of Phonology, New York, Academic Press. Anderson, Stephen R. (1982). "Where's Morphology?" L i n g u i s t i c  Inquiry, 13:571-612. Aronoff, Mark (1974). Word Formation in Generative Grammar, L i n g u i s t i c Inquiry Monograph 1, Cambridge, MA, MIT Press Chomsky, Noam (1965). Aspects of the Theory of Syntax. Cambridge, MA, MIT Press. Chomsky, Noam (1970). "Remarks on Nominalization". In R. Jacobs and P. Rosenbaum, eds. Readings on English  Transformational Grammar, Waltham, MA, Ginn and Co., 184-221. Chomsky, Noam (1981). Lectures of Government and Binding, Dordrecht, Foris Publicatons. Chomsky, Noam and Morris Halle (1968). The Sound Pattern of  English, New York, Harper and Row. Cook, E.D. (1976). "Flattening and Rounding in Velars". Unpublished Manuscript. Cook, E.D. (1983). " C h i l c o t i n Flattening", Canadian Journal of  L i n g u i s t i c s 28.2:123-132. Cook, E.D. (1986). "Ambisyllabicity and Nasalization in C h i l c o t i n " , Work ing Papers for the 21st Conference on Sal i s h and Neighboring Languages, University of Washington, Seattle, 1-6. Cook (1987). "An Autosesgmental Analysis of C h i l c o t i n Flattening", in Bosch et a l . (eds.) Proceedings of the  Parasess ion On Autosegmental and Metr i c a l Phonology, Chicago L i n g u i s t i c Society, 23.2. 149 Halle, Morris and Jean-Roger Vergnaud (1981). "Harmony Processes", In W. Klein and W. Levelt eds., Cross ing the  Boundaries in L i n g u i s t i c s , Dordrecht, Reidcl, 1-23. Hargus, Sharon (1985). The Lexical Phonology of Sekani. Unpublished doctoral d i s s e r t a t i o n , UCLA. Hayes, Bruce (1980). A Metrical Theory of Stress Rules. Unpublished doctoral d i s s e r t a t i o n , MIT, Cambridge,MA. Hoijer,H. (1948) "The Apachean Verb, Part IV: Major Form Classes", International Journal of American L i n g u i s t i c s 14:247-259. Howren, Robert (1971). "A Formalization of the Athabaskan *D-Effect'." International Journal of American  L i n g u i s t i c s 37:96-113. K a r i , James (1975). "The Disjunct Boundary in the Navajo and Tanaina Verb Prefix Complexes." International Journal of  American L i n g u i s t i c s 41: 330-345. K a r i , James (1976). Navaho Verb Prefix Morphology. New York, Garland. Kaye, Jonathan (1982). "Harmony Processes in Vata", in H. Van der Hulst and Smith, eds., The Structure of Phonological  Representations Part II, Dordrecht, Foris Publications, 1-88. Kiparsky, Paul (1984). "on the Lexical Phonology of Icelandic". In C. E l e r t et a l . eds., Nordic Prosody III  Papers From a Symposium, Stockholm, Almquist and Wiskell, 135-162. Kiparsky, Paul (1985). "Some Consequences of Lexical Phonology". Phonology Yearbook 2_, Cambridge University Press, 85-138. Krauss, Michael (1964). "Proto-Athapaskan-Eyak and the problem of Na-Dene: the Phonology." International Journal of  Amerlean L i n g u i s t i c s 30:118-131. 150 Krauss, Michael (1969). "On the C l a s s i f i c a t i o n in the Athapaskan, Eyak and T l i n g i t Verb." Indiana  Univers i t y Publications in Anthropology and L i n g u i s t i c s , Memoir 24:49-83. Krauss, Michael (1975). " C h i l c o t i n Phonology, a Descriptive and H i s t o r i c a l Report with Recommendations for a C h i l c o t i n Orthography" Unpublished Manuscript, Alaska Native Language Center. Latimer, Richard (1978). "A Study of C h i l c o t i n F l a t t e n i n g " . M.A. thesis, University of Calgary, Calgary. Leer, J e f f (1979). "Proto-Athabaskan Verb Stem Variation I: Phonology". Alaska Native Language Center Research  Papers, 1. Fairbanks. L i , Fang-Kuei (1946). "Chipewyan." In H. Hoijer et a l . eds., Viking Fund Publications in Anthropology, 6, New York, 398-423. Leiber, Rochelle (1980). On the Organization of the Lexicon Unpublished doctoral d i s s e r t a t i o n , MIT, Cambridge, MA. Rice, Keren (1985). "On the Placement of I n f l e c t i o n " , L i n g u i s t i c Inquiry 16:155-161. Roberge, Yves (1984). "Tone in C h i l c o t i n " , Unpublished Manuscript. University of B r i t i s h Columbia. Selkirk, Elizabeth (1982). The Syntax of Words, L i n g u i s t i c Inquiry Monograph 7, Cambridge, MA, MIT Press. Siegel, Dorothy (1974). Topics in English Morphology, Unpublished doctoral d i s s e r t a t i o n , MIT. Speas, Margaret (1984). "Navaho Prefixes and Word Structure Typology", in M. Speas and R. Sproat eds., MIT Working  Papers in L i n g u i s t i c s Vol. 7. 151 Vergnaud, Jean-Roger (1985). Lecture at University of Quebec at Montreal. Witherspoon, Gary (1977). Language and Art in the Navajo  Universe, University of Michigan Press, Ann Arbour, MI. Wright, Martha (1984). "The CV Skeleton and Mapping i n Navajo Verb Phonology". In proceedings of NELS 14, January 1984, Amherst, MA. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0097945/manifest

Comment

Related Items