Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The grammar and acquisition of Secwepemctsín independent pronouns Lai, I-Ju Sandra 1998

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

Item Metadata


831-ubc_1998-0507.pdf [ 6.25MB ]
JSON: 831-1.0088602.json
JSON-LD: 831-1.0088602-ld.json
RDF/XML (Pretty): 831-1.0088602-rdf.xml
RDF/JSON: 831-1.0088602-rdf.json
Turtle: 831-1.0088602-turtle.txt
N-Triples: 831-1.0088602-rdf-ntriples.txt
Original Record: 831-1.0088602-source.json
Full Text

Full Text

THE GRAMMAR AND ACQUISITION OF SECWEPEMCTSIN INDEPENDENT PRONOUNS by I-JU SANDRA LAI B. A , The University of British Columbia, 1996 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES Department of Linguistics We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA October 1998 © I-Ju Sandra Lai, 1998 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. 1 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 The University of British Columbia Vancouver, Canada (drfoUr ill* DE-6 (2/88) ABSTRACT Secwepemctsin, also known as Shuswap, is an endangered language spoken in the interior of British Columbia. No research dedicated to the study of Secwepemctsin pronouns is currently available. This thesis examines the independent pronouns of this language. Secwepemctsin is a radical head-marking language, and its independent pronouns function very differently from its bound pronominal clitics/affixes. This thesis provides a detailed description of the internal and external syntax of Secwepemctsin independent pronouns. They are analyzed as maximal projections that can occupy predicate positions as well as adjoin to DPs in argument positions. Binding effects follow from this analysis. Secwepemctsin independent pronouns show a strong subject orientation in third person contexts; it is the combined result of the independent pronouns' sensitivity to discourse and a subject-object asymmetry in the language: discourse familiarity is associated with syntactic positions via a mapping principle, yielding the Independent Pronoun Restriction. A semantic account of independent pronouns is provided. Secwepemctsin independent pronouns are shown to be contrastive focus when in predicate position, and contrastive topic when in argument position. Their behaviour is analyzed according to an alternative-based view of semantic theory. A case study of a child's acquisition of Secwepemctsin independent pronouns is documented, and the characteristics of her pronominal system are compared to those of an adult fluent speaker. Issues regarding language acquisition are discussed. It is found that although the child gets both Secwepemctsin and English input, her performance of Secwepemctsin is also influenced by Secwepemctsin-external and English-external factors. While this thesis investigates the technical details of Secwepemctsin syntax, it also places the study in the social context in which the language is in the process of being revived. A chapter on implications discusses the importance of providing learners of endangered languages with a positive environment in which to use the language. This chapter also points out the direction that endangered aboriginal languages may be heading, and stresses the importance of using language creatively. 11 TABLE OF CONTENTS Abstract 1 1 Table of Contents i» Acknowledgments vi CHAPTER 1 INTRODUCTION 1 1.0 Introduction 1 1.1 Main Goal of the Thesis 1 1.2 Overview of the Thesis 3 1.3 Current Status of Secwepemctsin 4 1.4 Data, Fieldwork Methodology and Consultant Information 5 CHAPTER 2 INTRODUCTION TO SECWEPEMCTSIN SYNTAX 7 2.0 Introduction 7 2.1 Secwepemctsin as a Radical Head-Marking Language 7 2.2 Arguments versus Predicates 8 2.2.1 Predicates 8 2.2.2 Position of Overt Arguments 10 2.3 Preverbal Positions 11 2.3.1 External Topic Position 11 2.3.2 Wh/Focus Position 12 2.3.3 Internal Topic Position 13 2.4 Binding Asymmetries: Gardiner (1993) 14 2.5 Re-evaluating Binding Asymmetries 18 CHAPTER 3 THE SYNTAX OF SECWEPEMCTSIN INDEPENDENT PRONOUNS 21 3.0 Introduction 21 3.1 Secwepemctsin Bound Pronouns 21 3.2 Secwepemctsin Independent Pronouns - Basic Introduction 22 3.3 Syntactic Atoms 24 3.3.1 Secwepemctsin Independent Pronouns are Syntactic Atoms 24 3.3.2 Halq'emeylem Independent Pronouns are not Syntactic Atoms 26 3.4 Argument and Predicate Positions 27 3.4.1 Independent Pronouns in Predicate Position 27 3.4.2 Independent Pronouns in Argument Position 28 3.4.3 First/Second Persons versus Third Person 29 3.4.4 Binding Properties 30 3.4.5 Independent Pronoun is an Adjunct to DP 32 iii 3.4.6 Comparison with Halq'emeylem 36 Halq'emeylem Independent Pronouns 36 Binding Properties ; 38 3.4.7 The Category of Secwepemctsin Independent Pronouns 40 Independent Pronoun is not N 40 Independent Pronoun is not D 41 Independent Pronoun is not DP 42 Independent Pronoun is of Category X 45 3.5 Subject-Object Asymmetry 46 3.5.1 Asymmetry in Secwepemctsin 46 3.5.2 Asymmetry in Halq' emeylem 48 3.6 Correlations 48 3.6.1 One Nominal Interpretation 49 3.6.2 Clefting of Third Person Subject 49 3.7 Analysis • 51 3.7.1 Familiarity 51 3.7.2 Analysis of the One Nominal Interpretation :. 55 3.7.3 Analysis of Subject Clefts 55 3.7.4 Analysis of the Independent Pronoun Restriction 56 3.7.5 Blocking 59 3.8 Implications: Test for Subjecthood 61 3.9 Remaining Questions ...: 63 3.10Chapter Conclusion 64 3.11 Appendix to Chapter 3: Rejecting Other Hypotheses 65 3.11.1 Ergative/Absolutive Difference 65 3.11.2 Aspectual and Thematic Differences 67 CHAPTER 4 TOWARDS A SEMANTIC ACCOUNT 70 4.0 Introduction 70 4.1 Basic Characteristics 70 4.1.1 Independent Pronoun * Out of the Blue 70 4.1.2 Independent Pronoun * Exhaustive 71 4.1.3 Independent Pronoun * Sentential Adverb 71 4.1.4 Independent Pronouns = Contrastive 72 4.2 Deriving Contrastiveness 74 4.3 Chapter Conclusion 78 CHAPTER 5 ACQUISITION: A CASE STUDY 79 5.0 Introduction ... 79 5.1 Preliminaries 79 5.1.1 Background: The Child Subject and the Family Context 79 5.1.2 Collection and Methodology 83 5.1.3 Transcription and Coding 84 Utterance.. 85 Language 86 iv Completeness 86 Number of Morphemes 87 Yes/No or Non-Yes/No 87 Imitation or Non-Imitation 88 Pronominal Reference 88 Bound Pronouns Used 88 Independent Pronouns Used 89 Additional Notes 89 5.2 General Analysis 89 5.2.1 Results 89 5.2.2 Secwepemctsin Vocabulary and Phrase Structure 93 5.2.3 General Aspects of Syntax: Adult versus Julienne 95 Radical-Head Marking Language 95 Optional DPs and Their Order 96 Summary 97 5.3 A Special Case: The Comparison of Pronominal Systems in Adult Secwepemctsin and Julienne's Secwepemctsin 99 5.3.1 Overregularization of Independent Pronouns: Data 99 Extensive Occurrence , 99 Independent Pronouns versus Bound Clitics/Affixes 101 5.3.2 Adult Independent Pronouns versus Julienne's Independent Pronouns ... 105 5.3.3 Syntactic Analysis of Julienne's Independent Pronouns 107 The Familiarity Hierarchy 108 The Category of Julienne's Pronouns 108 5.4 Remaining Questions 114 5.5 Chapter Conclusion 116 CHAPTER 6 CONCLUSIONS AND IMPLICATIONS 118 6.1 Summary 118 6.2 Looking Forward and Beyond 118 6.2.1 Interesting Differences: Comparative Salish Linguistics 119 6.2.2 To Keep the Language Alive 119 6.2.3 Possible Future Changes 122 6.3 Final Summary and Conclusions 123 References 126 Appendices ..130 Appendix A: Abbreviations 130 Appendix B: Phonetic Inventory of Secwepemctsin 130 Appendix C: Pronominal Paradigms 131 Appendix D: Determiner Paradigm 131 Appendix E: Julienne's Language Samples 132 ACKNOWLEDGMENTS This thesis would not have been possible but for my consultants Mona Jules and Julienne Ignace. You are both amazing teachers! I am also grateful to Florence Simon for her time. Kukwstsetselp! . My sincerest gratitude to the members of my thesis committee, who push me to excel. My mentor in Linguistics, Henry Davis, has been the crucial force driving my progress in the field; thanks for everything, Henry! Marianne Ignace, who for a long time has wanted to see a project like this, has been instrumental in the realization of this thesis; I am extremely appreciative of all the inspiring things that you have done. Thank you to my first Linguistics professor David Ingram for teaching me about language acquisition, and for your expertise in the field. To Lisa Matthewson, who got me started in Kamloops: I'm indebted to you for your diligence, encouragement, and insights; it's a privilege learning elicitation from you! Martina Wiltschko has been vital behind the development of my analysis; a thousand thank you's\ To Hamida Demirdache: thanks for always being so wonderfully dedicated, interested, and supportive; learning from you is a great joy! Many thanks to Dwight Gardiner for lending your knowledge and experience on Secwepemctsin, and for having done all that work on your dissertation! Thanks to Doug Pulleyblank for being such an excellent grad advisor; TA-ing for you has been a great learning experience. I have benefited much from the expertise of Michael Rochemont and Strang Burton. Su Urbanczyk has shown much enthusiasm in my work; thank you. I am also thankful for having been in classes taught by Pat Shaw and Laura Downing. Many thanks to Dr. Kinkade, for teaching the much-anticipated Salish seminar. Thanks to the entire Ignace family, who opened their home to me. Many thanks also go to Elizabeth Currie, my 301 TA, who encouraged me to attend grad school. To Carmen da Silva, the heart of the department: what are we gonna do without you? To fellow grad students, especially those who went through it with me, Maxine Baptiste, Leora Bar-el, and Nicole Horseherder it's all worth it! To Lisa Chang, Carrie Gillon, Ikuyo Kaneko, Eunsook Kim, Cora Li, Yumiko Nakamura, Sunyoung Oh, Cody Shepherd, and Linda Watt: thanks for the positive energy! To Tomio Hirose and Uri Strauss: it's been fun sharing offices with you! Thanks to the support of SSHRCC Grant #410-95-1519 awarded to Henry Davis, I was able to take all the field trips I needed to Kamloops. I would also like to express my gratitude to those who helped shape me and whose friendship I cherish. To'Kai Ch'un, for being my friend since grade one, and for watching over me; to Nina Hsiao, for keeping in touch; to Ms. Kuo, my grade 3 and 4 teacher; to Sylvia E. Chaverri, por los diarios; to Marlene, por las palabras de Dios; to Shiema, for the gossip; to Posheng, for getting me through the highschool years; to M. Potkonjak Sr., for being inspirational; to Eunice, for being a seester; and to Lisa Chen, Jennifer Chen, and Doris Huang, por los anos universitarios. I'm so happy you've all stuck by me. I would not be here today but for the support of my family, especially my grandparents, aunty Irene, uncle Jonsky, and K.C.. Grandpa in Heaven, I hope you're proud of me! Thank you to Richard Lai and Kenny Lai for being my chauffeurs, bodyguards, and most of all, for being my brothers. Gam-Dong! Last but definitely not least, lots of love to Mom and Dad, ^ J|jf ^ "gT f | j ! vi CHAPTER 1. INTRODUCTION 1 1.0 Introduction In Chapter 1, I lay out the aims of the thesis, and describe how the thesis is organized. Background information on the language being studied, including methodology of elicitation, is given. 1.1 Main Goal of Thesis This thesis examines the use of independent pronouns in Secwepemctsin2, also known as Shuswap. No research on Secwepemctsin independent pronouns is currently available, and this thesis contributes tothe study of Secwepemctsin for at least the following two reasons. A. Secwepemctsin is a radical head-marking language where pronominal arguments are generally encoded by null pro and agreement morphology on the predicate. It is very different from English in this respect, since English entirely lacks pro in argument positions. In addition to pros, and pronominal clitics and affixes which license them, there is a set of independent pronouns which function rather differently. This thesis will examine the differences and similarities between these two sets of pronouns in Secwepemctsin, and also explore these pronouns' relation to independent pronouns in non-head-marking languages. 'Secwepemctsin data are primarily elicited from language consultant Mona Jules, to whom I am extremely grateful. Additional data were obtained from Gardiner (1993). For a list of correspondences between the Secwepemctsin orthography used throughout this paper and IP A, and for a list of abbreviations, please refer to the appendices at the end of the thesis. 2Secwepemctsin is a member of the Salish family, belonging to the Northern Interior branch. 1 B. It has been observed by at least one Secwepemctsin teacher that many new speakers of Secwepemctsin use independent pronouns differently from the way they are used by fluent adult speakers3; so in order to understand the discrepancy between the grammar of independent pronouns as employed by native speakers and the way that these pronouns are treated by new speakers, it is necessary to conduct an in-depth study of the pronouns of the Secwepemctsin spoken by native speakers. This syntactic analysis of Secwepemctsin independent pronouns will contribute to the cross-linguistic study of pronouns, as well as to research on a little studied and endangered indigenous language. After the grammatical analysis of independent pronouns, I examine Secwepemctsin language samples from a seven-year-old Secwepemctsin learner, illustrating her overgeneralization in the use of independent pronouns. This case study has implications for child language, bilingual acquisition, and the acquisition of minority languages. When a child receives input from languages as distinct as English and Secwepemctsin, does she and can she clearly separate the two? When one of the languages is used much more extensively than the other in the child's surroundings, but the child consistently receives input from her parents in the non-dominant language, what happens? The present study has direct relevance for many aboriginal communities that find themselves lacking new speakers. If the small number of new speakers acquire a version of the language that is significantly different from that spoken by the prior generations, then the language, although maintained, might lose many of the core characteristics that make it unique. It is therefore important to find out what barriers learners of the language are facing so that changes and improvements can be made in Several adult new speakers consistently use Secwepemctsin independent pronouns in sentences where fluent native speakers view them as unecessary. Students at the Secwepemctsin Immersion Program at Chief Atham (elementary) school in Chase, B.C. have also been observed to use Secwepemctsin independent pronouns extensively (Marianne Ignace, p.c.) 2 language-teaching programs. In this way, the language; may be passed on to future generations without losing the properties that make it unique. . 1.2 Overview of the Thesis The thesis is organized as follows. The rest of this introductory chapter provides information on the current status of Secwepemctsin and information on the fieldwork done for this thesis. An introduction to the syntax of Secwepemctsin is given in Chapter 2. Chapter 3 is devoted to the grammar of Secwepemctsin independent pronouns. A short summary of the 'regular' pronouns (pro and pronominal affixes and clitics) is given in 3.1. In 3.2, a basic introduction to independent pronouns is given. In 3.3 and 3.4, the syntax of independent pronouns is discussed. I show that independent pronouns are adjoined to DP arguments. The binding properties of independent pronouns will be illustrated in 3.4.4, and Gardiner's claim for a configurational view of the language is supported by newly found data. An asymmetry concerning the use of independent pronouns is shown in 3.5. In sections 3.6 and 3.7,1 offer an analysis of the asymmetry, and conclude that this asymmetry emerges from the mapping of discourse relations onto argument positions. I further show that this mapping relation derives the One Nominal Interpretation Law (Gerdts 1988), and a restriction on third person subject clefts, and can be used as a test to show subject status in Secwepemctsin. The semantics of independent pronouns is discussed in Chapter 4. I describe the function of the independent pronouns in Secwepemctsin, the semantic values they provide, and make a tentative proposal on how to derive this semantic value. 3 Having done an analysis on the syntax and semantics of independent pronouns, I examine the acquisition data and find out how the system illustrated in Chapters 2, 3, and 4 is acquired by a new speaker of Secwepemctsin. Chapter 5 is a case study of a young child who is growing up receiving input in both English and Secwepemctsin. An analysis is given of grammatical characteristics that differ between the learner's and the adult's Secwepemctsin. At the end of Chapter 5,1 focus on the child's innovative way of using independent pronouns. Conclusions and implications of the thesis are reported in Chapter 6. Possible future changes in the Secwepemc language and possible future strategies for teaching endangered languages are suggested here. It is hoped that this thesis will contribute to the study of Salish languages in general, as well as to the reintegration of minority languages into the community. 1.3 Current Status of Secwepemctsin Secwepemctsin is spoken by the Secwepemc people, and belongs to the Salish family, which has approximately twenty-three languages (Thompson 1979); unfortunately not all twenty-three are still being spoken. Salish languages are divided into several branches: Secwepemctsin belongs in the Northern Interior Branch, along with Nlaka'pamuxcin (Thompson) and St'at'imcets (Lillooet). Speakers of Secwepemctsin are spread out over the Shuswap Territory in the interior of British Columbia. Two major dialects exist within the Secwepemc language: the western dialect and the eastern dialect (Kuipers 1990), divided geographically by a vertical boundary slightly east of Kamloops, British Columbia. Gardiner (1993) has noted that the primary difference between the dialects is in the phonetic inventory and lexical variations, but that there are no observed syntactic differences. 4 A survey conducted by the Secwepemc Cultural Society has shown that only 3.5% of Secwepemc people consider themselves, or are considered by their peers, to be fluent speakers of Secwepemctsin (Ignace 1995). Czaykowska-Higgins and Kinkade (1998) give the figure of roughly 500 speakers in a separate estimate; however, they note that not all speakers are necessarily fluent. Almost all of those speakers who can converse comfortably in Secwepemctsin with each other are aged 60 and over. Much effort has been directed to creating language programs in schools with the aim of reintegrating the language into the lives of Secwepemc children. The existing Secwepemctsin immersion program at Chief Atham (elementary), school in Chase, British Columbia, is an example, although no formal study of the results is available at this point. With the exception of the family of the young Secwepemctsin speaker studied for this thesis, there is no other known family in which children are being raised with Secwepemctsin as the first language. The following are the main linguistic studies of Secwepemctsin: • Gibson 1973 — a Ph.D. dissertation on the Eastern dialect • Kuipers 1974 ~ a description of the phonological and morphological characteristics of primarily the western dialect, a collection of texts with corresponding English translation, and a Secwepemctsin-English dictionary • Kuipers 1990 — additions and updates of the 1974 grammar; new texts • Gardiner 1993 -- a Ph.D. dissertation focusing on structural asymmetries and preverbal positions in the western dialect • Gardiner 1998 — a paper on preverbal positions in Secwepemctsin I rely heavily on Kuipers 1974 and Gardiner 1993, 1998, as my main sources of information on Secwepemctsin syntactic properties. 1.4 Data, Fieldwork Methodology and Consultant Information 5 Secwepemctsin fieldwork started in September 1996 with my primary language consultant Mona Jules, an elder who speaks the Skeetchestn dialect4 of Secwepemctsin. I have also done elicitations with Florence Simon of Skeetchestn, but these sessions total up to less than 4 hours. The bulk of data used in this thesis were mainly elicited between September 1997 and April 1998, although irregular sessions were set up after April 1998 to double-check existing data and to get new data. The regular elicitation sessions took place in the language consultant's office in Kamloops, British Columbia, and lasted up to an hour each, with a frequency of roughly twice per month. Data consist of utterances translated from English, corrections of Secwepemctsin utterances provided by myself, and forms that the consultant provided voluntarily without being elicited. Transcriptions were made at the time of elicitation, and subsequently checked with audio recordings of the sessions. "Skeetchestn is located on the southwest of the Shuswap territory, and the dialect spoken there belongs to the Western dialect. 6 CHAPTER 2. INTRODUCTION TO SECWEPEMCTSIN SYNTAX 2.0 Introduction In Chapter 2, I give a summary of basic Secwepemctsin syntax, describing the argument, predicate, and preverbal positions. Gardiner's (1993) arguments for structural asymmetry in Secwepemctsin are reviewed, and re-evaluated. 2.1 Secwepemctsin as a Radical Head-Marking Language Secwepemctsin is a radical head-marking language. A radical head-marking language exhibits the following characteristics. 1 Characteristics of radical head-marking languages (a) All arguments of a head are obligatorily marked on the head (typically the verb, or an associated auxiliary) as pronominal clitics or affixes. (b) All overt DP arguments are optional. In Secwepemctsin, the predicate is marked by pronominal clitics and affixes, and no overt arguments are required (See tables in 4-6 below for the paradigms of the pronominal clitics and affixes). These bound pronominals are obligatory; hence, Secwepemctsin is a radical head-marking language. (2a-d) are complete clauses; these show that overt argument DPs are not required. (3 a, b) show overt argument DPs5. 2 a wik-t-0-s see-tr-3sg.obj-3sg.subj He; saw him, Since D is obligatory on all argument positions in Salish, even in clausal arguments, I will assume the DP Hypothesis (Abney 1987) and make the standard assumption (Longobardi 1994, Matthewson 1996) that all arguments are DPs. In the DP Hypothesis, the head of the DP is the functional category D, and it selects NP as its complement, in contrast to the traditional assumption that determiners are dominated by NP. 7 . b Shuswap(l sg.redup)- 1 sg.ind I am Shuswap c pelq'-ilc-0 leave-aut-3sg.ind He left d tsu-n-t-0-em punch-fc-tr-3 sg. obj -pas He was punched 3 a pelq'-ilc-0 re John leave-aut-3sg.ind det John John left b tsu-n-t-0-em re John punch-fc-tr-3sg.obj-pas det John John was punched 4 Intransitive Clitic Paradigm (Kuipers 19.74) 1 sg. 2sg. 3 sg. 1 pi. incl. 1 pi. excl. 2 pi. 3 pi. Indicative -ken -k -0 -kt -kucw -kp -0 Conjunctive -wen -(w)cw -(w)s -(w)t -kucw -(w)p -(w)s Possessive n- -7 -s -kt -kucw -mp -s 5 Transitive Subject Suffix Paradigm (Kuipers 1974) 1 sg. 2sg. 3 sg: 1 pi. incl. 1 pi. incl: 2 pi. 3 pi. -(e> -(e)c -(e)s -(e)t/-(e)m -kucw -(e)p -(e)s Transitive Object Suffix Paradigm (Kuipers 1974) 1 sg. 2sg. 3 sg. 1 pi. incl. 1 pi. excl. 2 pi. 3 pi. -s6m-/-sm- -S1-/-S- 0 -61-/-1- -kucw -ulm-/-lm- 0 2.2 Arguments versus Predicates 2.2.1 Predicates 8 In Secwepemctsin, as in most Salish languages, almost any independent lexical item can function as a predicate. Predicates are characterized by pronominal morphology. (7a) shows an adjectival predicate; (7b) shows a verbal predicate; (7c) shows a nominal predicate; (7d) shows a cardinal adjective, or quantifier, predicate; finally, (7e) shows a deictic predicate. 7 a. [kwalt-0]pred [re skelklets]sUbj green-3sg.ind det grasshopper The grasshopper is green b. [qwetsets-0]pred [re nuxwenxw]sUbj leave-3 sg.ind det woman The woman left c. [Secwepemc-0]pred [re Sam]sUbj Shuswap-3 sg.ind det Sam Sam is Shuswap d. [cw7it-0]pred [re speq.peq]Subj many-3sg.ind det berry(redup) There are many berries / The berries are many e. [yiri7-0]pred [re tewi.w.t]sUbj deic-3sg.ind det grow(redup) The child is there The lexical item that functions as a predicate takes morphological person markings. The examples in (8) show the predicate taking first, second, and third person singular subject markings. 8 a Shuswap(redup)-1 sg. ind I am Shuswap b kwalt-k green-2sg.ind You are green c qwetsets-0 leave-3 sg.ind S/heleft 9 2.2.2 Position of Overt Arguments Just as predicates are not limited to verbal categories, arguments are not limited to nominal categories either. Arguments are identified by their being marked with a determiner; thus even verbal elements can function as arguments via the addition of a determiner6. In the following examples, the arguments are nominalized verbs, which take possessive clitics as subjects. Notice that the nominal predicates here, as well as in (7c) above, cannot take a determiner7. 9 [skelklets-0]pred[re n-s-wi.w.k-emJMgument grasshoppers det lsg.poss-nom-see(redup)-intr The ones I saw were grasshoppers. 10 [tsiqw te speq.peqjpred [re 7-s-7illen-c]AIBument red obi berry(redup) det 2sg.poss-nom-eat-intr The ones you ate were red berries. When optional argument DPs are present in a clause, the most common word order in spoken Secwepemctsin is SVO (11). VOS and VSO are both also possible; thus, the order of post-predicate nominals is quite free (12, 13), although my language consultant generally prefers VOS over VSO. Only the subject, but not the object, can appear to the left of the verb (11) unless an A' movement has taken place (as in clefts or wh-questions; see discussion on the cleft construction in 2.3.2). 11 re Mary wik-t-0-s re John det Mary see-tr-3sg.obj-3sg.subj det John Mary sees John / *John sees Mary 12 wik-t-0-s re John re Mary see-tr-3sg.obj-3sg.subj det John det Mary John saw Mary / Mary saw John "There is controversy on the status of non-nominal DPs in Salish. The dispute is on whether DPs are internally clausal or not. See Jelinek and Demers 1983, Montler 1993, Matthewson and Davis 1995, and Demirdache and Matthewson 1995. 7For a complete list of determiners in Secwepemctsin, please consult the appendix. 10 13 wik-t-0-s re Mary re John see-tr-3sg.obj-3sg.subj det Mary det John John saw Mary / Mary saw John If there is only one post-verbal overt nominal in a transitive clause with two third person arguments, it is associated with the absolutive argument, which is the object in transitive clauses, and subject in intransitive clauses8. This constraint is known as the One Nominal Interpretation Law9 (Gerdts 1988; see also Roberts 1994, Davis 1994) 14 One Nominal Interpretation Law (Gerdts 1988) In the absence of marking for other persons, a single third person nominal is interpreted as the absolutive. 15 wik-t-0-s re John see-tr-3sg.obj-3sg.subj det John S/he saw John / *John saw him/her 2.3 Preverbal Positions Three preverbal positions are available in Secwepemctsin (Gardiner 1998), and all three can be filled at one time. Gardiner calls these positions (a) the external topic position, (b) the wh or focus position, and (c) the internal topic position. I will adopt Gardiner's (1998) analysis for the sake of concreteness. 2.3.1 The External Topic Positions "Secwepemctsin exhibits morphological split ergativity. With third persons, the subject of intransitive clauses and the object of transitive clauses are marked absolutive, while the subject of transitive clauses is marked ergative. With first and second persons, intransitive subject and transitive object are marked by different sets of bound pronominal markers. 9Gerdts (1998) defines the One Nominal Interpretation Law for the single DP within a clause. In Secwepemctsin, the One Nominal Interpretation Law'is only holds for post-predicate DPs. 11 The external topic position is base generated and is adjoined to CP, following Aissen (1992). This is the position for dislocated nominals, which are equivalent in meaning to nominals following "as for..." in English. The external topic is typically followed by a pause, and functions like a switch-topic position, foregrounding a previously mentioned topic. This position cannot host second position clitics, and does not obey adjunct and complex noun phrase constraints. An external topic syntactically occupies the same clause as does its predicate. 2.3.2 The wh/focus Position The wh/focus position is adjoined to CP via A'-extraction, and can be occupied by wh words or clefted elements. The cleft and wh question constructions have been likened to relative clause constructions in that all three have a dependent clause introduced by a determiner (Gardiner 1998). Syntactically, the clefted element is predicative, and is coindexed with an empty operator, Oi, in the dependent clause (16). Semantically, the cleftee is the argument of the dependent clause by virtue of coindexation (Kroeber 1991, Davis et al.. 1993, Gardiner 1998). 16 CP cleftee spec C O, C e; All arguments can be clefted or questioned without any special change in morphological marking on the subordinate predicate. On the other hand, when adjuncts are clefted, v 12 conjunctive clitics or nominalization must be used (Gardiner 1998). When a third person subject in a clause containing two third person arguments is clefted, conjunctive morphology appears on the lower predicate; this suggests that in third person subject extraction, the argument in question is actually of adjunct status (Gardiner 1998) (See also section 3.7.3). Clefted elements and wh words occupying the wh/focus position can be followed by second position clitics, and they are sensitive to the complex noun phrase constraint and the adjunct island condition (Gardiner 1998). 2.3.3 The Internal Topic Position There can be more than one internal topic per sentence; it is adjoined to VP. The internal topic position is the position for preverbal subjects, it is adjoined to VP. Gardiner (1998) has claimed that Secwepemctsin's ability to take multiple internal topics is similar to scrambling. Internal topics are not followed by a pause, as are external topics, nor can they function as switch-topic positions. Internal topics differ from wh/focus elements in that internal topics do not select a subordinate clause headed by a determiner. Internal topics are clause-internal and cannot undergo long distance extraction, therefore they must obey island constraints (Gardiner 1998). The characteristics of the positions are summarized in (17), and these positions are shown on the tree in (18). 17 Preverbal positions External topic Wh/focus Internal topic - base-generated - left-dislocated nominals - adjoined to CP - wh words and clefted nominals - adjoined to V P - scrambled DPs 13 18 Structure of preverbal positions in Secwepemctsin. (Gardiner 1998) E External Topic Internal Topic In addition to his study of pre-verbal positions, Gardiner also argues that Secwepemctsin is a configurational language, showing structural asymmetries. His arguments are reviewed in the next subsection. 2.4 Binding Asymmetries: Gardiner (1993) In this subsection I review Gardiner's (1993) arguments for a configurational analysis of Secwepemctsin. Although the binding evidence that Gardiner uses to show subject-object asymmetry is not strong, his claim is ultimately supported by his Condition on Antecedence, and by newly found binding data, as Chapters 3 and 4 will show. According to the Binding Theory, different types of nominals are constrained by different conditions on antecedence. Thus, Chomsky and Lasnik (1993) identify the following conditions: 19 Binding Conditions Condition A) An anaphor must be bound in a local domain Condition B) A pronoun must be free in a a local domain Condition C) An R-expression must be free 14 The binding conditions make predictions about coreference based on syntactic structures and can therefore be used as a test for syntactic configurationality. Speas (1990) has pointed out that the application of binding conditions yields the predictions in (20) below, for configurational and non-configurational languages. 20 Predictions for configurational anc flat languages Configurational language Flat languages a Maryi likes hen father b Mary'Si father likes her; c Hen father likes Mary; d *She; likes Mary's; father a Maryi likes hen father b *Mary's, father likes her; c Hen father likes Mary, d *Shei likes Mary'Si father The predictions are correct for English, which is a configurational language. Let us examine the Secwepemctsin data. For (20a), the Secwepemctsin equivalent is given in (21). 21 re Mary xwist-0-es re qe7tse-s det Mary like-3sg.obj-3sg.subj det father-3sg.poss Mary, likes hen father (Gardiner 1993) Here, Mary binds the coreferent possessive pronominal in the object, but the pronominal is free within the object DP, so Condition B is satisfied. The comparison between English and Secwepemctsin for (20a) does not show anything, since configurational languages and non-configurational languages behave alike with regard to this sentence type. For (20b), the Secwepemctsin equivalent is given in (22a, b). 22 a re Mary reqe7tse-s xwist-0-es det Mary det father-3sg.poss like-3sg.obj-3sg.subj *Mary'$ father likes hen (Gardiner 1993) 15 b re qe7tse-s re Mary xwist-0-es det father-3sg.poss det Mary like-3sg.obj-3sg.subj *Mary'Si father likes her; (Gardiner 1993) According to Speas, the ungrammatically of (22a, b) above would categorize Secwepemctsin as a flat language, since, in a flat language, both DPs {Mary and her father) would be sisters to the verb, and the R-expression Mary would be bound by the possessive pronoun, violating Condition C. However, (22a, b) are ungrammatical for independent reasons. In Secwepemctsin, no overt subject is possible when the object is null. This is formulated in Gardiner (1993) as (23) and exemplified in (24) (See the One Nominal Interpretation Law in section 2.2.2). 23 Generalized Condition on the Interpretation of Empty Categories: An overt NP or a variable subject cannot occur with a null pronominal object (Gardiner 1993:202) 24 *Mary wik-t-0-s *Mary see-tr-3sg.obj-3sg.subj Mary sees (someone) For (20c), the Secwepemctsin equivalent is shown in (25). 25 re qe7tse-s xwist-0-es re Mary det father-3sg.poss like-3sg.obj-3sg.subj det Mary *Hen father likes Maryi (Gardiner 1993) The ungrammaticality here is not predicted by anything in the binding conditions. Gardiner argues that the explanation lies in another condition on coreference. He argues that (25) is ungrammatical because Mary, the antecedent for the subject pronominal possessor, does not c-command the pronoun. In other words, a DP can only be an antecedent for a pronoun when it is higher in the structure. The generalization is as follows: 26 Condition on Antecedence A pronoun must be c-commanded by its antecedent (Gardiner 1993: 183) 16 In (25), pro is in the subject DP, and is not c-commanded by Mary in object position. This results in ungrammaticality (28). Compare this to (21), whose tree is shown in (27). Here, the subject DP c-commands the coreferent pronoun in object position, and the sentence is grammatical. (Gardiner 1993) pro, re qe7tses xwistes re Mary (Gardiner 1993) Now, for (20d), the Secwepemctsin equivalent is given in (29). 29 a xwist-0-es [re Mary re qe7tse-s] like-3sg.obj-3sg.subj det Mary det father-3sg.poss *Shei likes Mary's; father (Gardiner 1993) b xwist-0-es [re qe7tse-s re Mary] like-3sg.obj-3sg.subj det father-3sg.poss det Mary *She; likes Mary's; father (Gardiner 1993) Since the R-expression Mary is in object position, it is bound by the coreferent pro in subject position, violating Condition C. It is only possible to have an interpretation for (29a, b) above where the subject pro and Mary are disjoint. 17 2.5 Re-evaluating Binding Asymmetries In this section, I will re-evaluate Gardiner's arguments for binding asymmetries in Secwepemctsin, and conclude that although certain of his arguments are not supported, his general conclusions concerning configurationality are correct. First of all, note that (20a, d) given in the previous section are uniformative, because predictions for configurational and non-configurational languages are the same. Furthermore, for the Secwepemctsin equivalent of (20d), which is shown in (29), there is a perfectly grammatical structure with the same meaning, which acts as an interfering factor in the grammaticality judgment. Namely, the nominals in (29a, b), re Mary, and re qe7tse-s, can be interpreted as one single possessive DP Mary's father, or as two separate DPs, Mary and her father, one occupying the subject position and other occupying object position. Therefore, under normal circumstances, a speaker would not interpret the sentences as 'She; likes Mary's; father', but as 'Mary; likes hen father' - this complication makes the sentence type an unreliable test for a Condition C violation. It appears that the ungrammaticality of (22a, b), which are the Secwepemctsin equivalents of (20b), categorizes Secwepemctsin as non-configurational. However, this ungrammaticality is due to the independent reason that the One Nominal Interpretation Law is applying. Therefore, the ungrammaticality of (22) cannot be used as evidence that Secwepemctsin is non-configurational10. '"(22a, b) would be ungrammatical due to the One Nominal Interpretation Law if the nominals re Mary 'Mary' and re qe7tse-s were interpreted as one single nominal modified by a possessor DP. However, the two nominals may be interpreted as separate DPs, both occupying internal topic positions (see section 2.3). Judgments on what interpretations (22a, b) would have given that particular structure are unavailable at this point. 18 The Secwepemctsin equivalent of (20c) is ungrammatical, but this is not predicted by anything in the binding conditions. The ungrammatically is predicted by the Condition on Antecedence, according to Gardiner (1993). This condition may be viewed as a generalized form of weak crossover. Consider the following explanation. In English, there is weak crossover in a sentence like (30a), where the possessor pronoun is not c-commanded by the quantifier 'everyone'. There is no weak crossover when the possessor pronoun is c-commanded by 'everyone', as in (30b). In Secwepemctsin, weak crossover cases are not limited to quantifier constructions; all possessor pros are variables, and must be c-commanded by their antecedent, as suggested by Gardiner's (1993) Condition on Antecedence. This means that (25) is predicted to be ungrammatical because there is a weak crossover violation, while (21) is predicted to be grammatical because there is no weak crossover violation. 30 a *HiSi mother likes everyone; b Everyone, likes his; mother 21 re Mary xwist-0-es re qe7tse-s det Mary like-3sg.obj-3sg.subj det father-3sg.poss Mary, likes hen father (Gardiner 1993) 25 re qe7tse-s xwist-0-es re Mary det father-3sg.poss like-3sg.obj-3sg.subj det Mary *Her; father likes Mary, (Gardiner 1993) Secwepemctsin sentence types are compared to configurational and non-configurational binding predictions in (31). The rightmost column summarizes the effectiveness of each sentence type as a test for configurationality. In (32), the Condition on Antecedence, a generalized form of weak crossover, suggests that Secwepemctsin is configurational. 19 31 Effectiveness of binding as a test for structural asymmetry in Secwepemctsin Language type vs. Sentence type Configur-ational Non-config-urational Secwepemctsin Effectiveness a Mary; likes hen father Irrelevant b Mary'Si father likes hen * * FLAT! However, independent reason: One Nominal Inter-pretation Law c Her, father likes Mary * Unpredicted by binding! Possibly due to weak crossover d She; likes Mary's; father * * * Impossible to test effectively and irrelevant 32 Effectiveness of Condition on Antecedence (weak crossover) as a test for structural asymmetry in Secwepemctsin Language type vs. sentence type Configur-ational Non-config-urational Secwepemctsin Effectiveness Hen father likes Mary * * Configurational! Mary; likes hen father V Irrelevant In conclusion, binding is not a reliable test for structural asymmetries in Secwepemctsin, but Gardiner's (1993) Condition on Antecedence might be, since it suggests that weak crossover applies to possessor pronominals in Secwepemctsin. The weak crossover is evidence that a subject-object asymmetry is present. In Chapters 3 and 4, I will provide further support for Gardiner's (1993) claim that Secwepemctsin is not a non-configurational, but a configurational language. 20 CHAPTER 3. THE SYNTAX OF SECWEPEMCTSIN INDEPENDENT PRONOUNS 3.0 Introduction In this chapter, I will first give an introduction to bound pronominal morphemes in Secwepemctsin, and proceed to give a detailed description of the independent pronouns. The syntactic category, site of occurrence, and binding properties of Secwepemctsin independent pronouns are discussed (sections 3.3 and 3.4) and compared to those of Halq'emeylem independent pronouns. In 3.5 I show that in Secwepemctsin, there is an independent pronoun restriction, which allows only third person independent pronouns to occupy the subject position if the object is also third person. I give an analysis of this restriction in 3.6 and 3.7. The consequences and predictions of this analysis are given in 3.8 and 3.9. 3.1 Secwepemctsin Bound Pronouns The radical head-marking characteristics of Salish languages (section 2.1) has induced Jelinek and Demers (1994) to analyze at least Straits (a Coast Salish language) as a 'pronominal argument language'. In such an analysis, bound pronominal clitics and affixes are considered to be occupying argument positions; in fact, these bound pronominals are considered to be the arguments. Overt DPs, when present, are adjuncts that are coreferent with these pronominal arguments; such an analysis predicts that overt DPs are not subject to A-binding effects, since they are adjuncts. However, arguments have been given that at least some Salish languages are not pronominal argument languages (Davis 1997). I will argue for the. view that Secwepemctsin is not a 21 pronominal argument language, following Gardiner (1993). I will view the radical head-marking characteristics in the following way. Secwepemctsin regular pronouns are pro, licensed by clitics or affixes attached to the predicate of a clause. Following the standard assumption that pronouns are DP projections, I will assume that pro is contained within a DP, which occupies the argument position. Thus, in any given clause, when the arguments are covert, it is because empty DPs (occupied by pro) are in the argument positions; when the arguments are overt, it is because overt DPs (containing a determiner and its complement) are in the argument positions. The full clitic and suffixal paradigms are shown in the following tables. In transitive sentences, the object suffix always precedes the subject suffix. 1 Intransitive Clitic Paradigm (Kuipers1974) 1 sg. 2sg. 3 sg. 1 pi. incl. 1 pi. excl. 2 pi. 3 pi. Indicative -ken -k -0 -kt -kucw -kp -0 Conjunctive -wen -(w)cw -(w)s -(w)t -kucw -(w)p -(w)s Possessive n- -7 -s -kt -kucw -mp -s Transitive Subject Suffix Paradigm (Kuipers 1974) 1 sg. 2 sg. 3 sg. 1 pi. incl. 1 pi. incl. 2 pi. 3 pi. -(e)n -(e)c -(e)s -(e)t/-(e)m -kucw -(e)p -(e)s Transitive Object Suffix Paradigm (Kuipers 1974) •1 sg. 2sg. 3 sg. 1 pi. incl. 1 pi. excl. 2pl. 3 pi. -sem-/-sm- -S1-/-S- 0 -el-/-l- -kucw •ulm-/-lm- 0 3.2 Secwepemctsin Independent Pronouns - Basic Introduction In addition to the clitic and suffixal paradigms presented above, there is a separate set of pronouns in Secwepemctsin. Since this additional set of pronouns are free morphemes and 22 behave like independent lexical items, as opposed to the bound status of the pronominal clitics and suffixes, I refer to them with the analysis-neutral term, "independent pronouns". The full independent pronoun paradigm is given in table (4). 4 Independent pronominal paradigm (adopted from Kuipers 1974)11 1 sg. 2sg. 3 sg. ntsetswe7 7-enwi7 newi7-s 1 pi. incl. 1 pi. excl. 2 pi. 3 pi. wll-enwi7-kt wll-enwi7-s-kucw wll-enwi7-mp wll-enwi7-s These independent pronouns are generally known as emphatic pronouns across Salish languages, signaling their function; in Chapter 4, I analyze this emphatic function as contrastiveness that is inherently contained in Secwepemctsin independent pronouns. The independent pronouns are related to a discourse referent which may be supplied either by prior discourse or by deixis. Syntactically, independent pronouns appear adjacent to and on the left of an overt nominal (5), or without an overt nominal (6a, b). When an independent pronoun does not occur with an overt nominal, I will assume that it is associated with a covert nominal (pro). Regardless of the overtness or covertness of the nominal, the pronominal suffix or clitic on the predicate is always present. In other words, Secwepemctsin independent pronouns supplement rather than replace pronominal clitics or suffixes. 5 [newi7-s12 re Mary] ts'urri-qs-en-0-s re John [3sg.indpr det Mary] lick-nose-fc-3sg.obj-3sg.subj det John MARY kissed John13 1 'Note that independent pronouns appear to have internal structure. See 3.3 for a discussion. 1 2 Although independent pronouns are hyphenated, I will only gloss them as indpr for independent pronoun. Please see section 3.3 for a discussion of the elements that make up the independent pronouns. ' Capitalized (pro)nominals in English translations throughout this paper reflect the emphasis expressed by Secwepemctsin independent pronouns. The function of Secwepemctsin independent pronouns is similar to that of having a contrastively stressed (pro)noun in English, as the capitalized letters show. 23 6 a [newi7-s] ts'um-qs-en-0-s re Mary 3sg.indpr lick-nose-fc-3sg.obj-3sg.subj det Mary HE kissed Mary b tsu.ts'.em-qs-en-tsem-s [newi7-s] lick(redup)-nose-fc-lsg.obj-3sg.subj 3sg.indpr HE kissed me In the rest of this chapter, I will discuss the internal syntax and the external syntax of independent pronouns (sections 3.3 and 3.4). In 3.5,1 will present a subject-object asymmetry that arises from the investigation of independent pronouns in Secwepemctsin. I will give an analysis of the asymmetry in sections 3.6 and 3.7. 3.3 Syntactic Atoms 3.3:1 Secwepemctsin Independent Pronouns are Syntactic Atoms An independent pronoun in Secwepemctsin can function as a unit per se, either as an argument, or as a predicate, but there is morphological evidence that independent pronouns in Secwepemctsin may not be single-morpheme items, and that at least historically they must have had internal structure. Consider the table in (7), which gives the internal morpheme breakdowns of independent pronouns. 7 Independent pronominal paradigm (adopted from Kuipers 1974) 1 sg. 2 sg. 3 sg. n-tse.ts-we7 1 sg. poss-emph(redup)- deic 7-enwi7 2sg.poss-emph newi7-s emph-3poss 1 pi. incl. 1 pi. excl. 2 pi. 3 pi. wllTenwi7-kt pl-emph-lpl.poss wll-enwi7-s-kucw pl-emph-3 poss-excl wll-enwi7-mp pl-emph-2pl.poss wll-enwi7-s pl-emph-3 poss Independent pronouns are all composed of a stem (henceforth "emphatic stem") and the possessive clitic. The emphatic stem for the first singular form is different from that of the 24 rest of the paradigm. According to Newman (1977), the first person singular stem in Secwepemctsin is derived from the proto-Salish first person singular emphatic stem *?sntfd, while the stem for the rest of the Secwepemctsin paradigm is derived from the proto-Salish second person singular stem *ngwi. In fact, the Secwepemctsin independent pronominal system is one of the two most innovative ones in Salish. The other one is Halq'emeylem, represented in (8) by Musqueam (Newman 1977 considers Musqueam a dialect of Halq'emeylem). 8 Salish Independent Pronouns (Newman, 1977)14 Is PS •>3nc4 (1) LIL s-'ince (2) NSH nce-we SSH nce-we' (3) THOM . . . .ncc-we' (4) OKAN . . . .enca COLV inca' (5) 'KAL (q"u-ye'e) SPOK (k"u-ya'e) (6) CA (cm-e) (7) CLM 'incl (8) TIL 'anca (9) UCH 'anca (10) TW dacah (11) SPU aca NPU 'aca (12) CLAL 'ac SON >aS3 (13) MUS enGe (14) SQ 'ans (15) SL1 (cinal) (16) BC he 2s 3s •csnil IP •nimat, 2p •Hp, 'waUp s-nuwe s-nii s-nimul s-nu-l6p '-nwi' (nwi-s) (wl-nwi'-kt) (wl-nwi'-mp) e'-nuwf' (nuwi'-s) ('ul nuwi-kt) (•ul nuwi'-cp) e'-wi cnil-0 n-m-iml p-ye>st a-nwe . cnel-c m-niml-tt (mniml-mp) a-nwi' cnil-c ma-nlml-tat (manlml-amp) a-nwi canil-c (qe'-npali) n-pa-le" ha-nawl' canil-c (qe'-anpate') na-pa-li' (k-uvl-e) cinil-0 (c-lipust) k"u-p-lip-ust 'i-nwi cnil-0 nam-nlmal pl-p-lap-st 'a-nki canil niwal kal-kila nawe cine 'e-n6m ?c-lap-a da'weh cadil dibit ta-wal'-walap dag"e cadet ddbl gwl4p-o' dag"i cadi! dibal g'alap-u nak" nil l-nirjl (nakw) nak"a nil 1-ninai (nakM'Iaye') ta-nawa (tuA'a) ta-l-nimal te-l-walap naw (tiwa) nimal naw-yap negi (ti'ta) nemul nu-wap 'i-nu (tix) l-mil lup 'Replacements of stems are enclosed in brackets. Accretions (additions of new morphemes) are separated from steins by hyphens or spaces. Abbreviations for language names are as follows: LIL= Lillooet; NSH=Northern Shuswap; SSH=Southern Shuswap; THOM=Thompson; OKAN=Okanagan; COLV=Colville; KAL=Kalispel; SPOK=Spokane; CA=Coeur d'Alene; CLM=Columbian; TTL=Tillamook; UCH=Upper Chehalis; TW=Twana; SPU=Southern Puget; NPU=Northern Puget; CLAL=Clallam; SON=Songish; MUS=Musqueam; SQ=Squamish; SLI=Sliammon; BC=Bella Coola. 25 In Secwepemctsin the first person singular form is composed of the first person singular possessive clitic, n-, the first person singular emphatic stem (often appearing in reduplicated form) -tse.ts-, and the deictic marker -we 7, no longer productive (Newman 1977). The rest of the paradigm consists of the emphatic stem -mvi7-, derived from the proto-Salish second person singular stem -nawi-, and the appropriate possessive clitic. Although possessive clitics are used productively in the language, there is no evidence of the occurrence of -nwi7-elsewhere in the language. In addition, the plural independent pronouns contain the affix wll-, 'group of people', 'collective plural'. Since the plural marker wll- and the deictic -we7 are never separated from independent pronouns and are never dropped, they are syntactically invisible, and I will analyze them as being part of the pronoun stem itself. For simplicity, and since there is no syntactic evidence for an atomistic view (Safir 1994) of Secwepemctsin independent pronouns, in which each atom within the pronoun has its own unique function, I will treat the entire independent pronoun as a single syntactic atom. 3.3.2 Halq'emeylem Independent Pronouns are not Syntactic Atoms Compare the Secwepemctsin independent pronouns to Newman's table of Salish independent pronouns (see table in 8). Proto-Salish essentially has a mono-morphemic independent pronoun system, with single stems for each and every independent pronoun. Though there are variations and some accretions (new morphemes added to the stem), the systems of the modern Salish languages are essentially the same as Proto-Salish, with the exception of Secwepemctsin and Halq'emeylem. 26 Halq'emeylem stands out from its sister languages because its independent pronouns are clearly multi-morphemic; morphological evidence for this claim can be seen in Newman's table (8). The main syntactic evidence stems from Wiltschko's (1998) study of independent pronouns in Upriver Halq'emeylem. Wiltschko demonstrates that the elements that make up the independent pronouns are syntactically active (see section 3.4.6). Upriver Halq'emeylem independent pronouns are given in table (9) below. 9 Halq'emeylem independent pronouns (Galloway 1993:403) sg pi 1 te'elthe/te a'elthe telhimelh 2 telewe telhwelep 3 tutl'6/thutl'6 tutl' 6: lem/thutl' 6: lem/yutl' 6: lem Secwepemctsin independent pronouns are also different from those of its sister languages. So far we have seen that Proto-Salish independent pronouns are mono-morphemic, while Halq'emeylem independent pronouns are multi-morphemic. Secwepemctsin independent pronouns are actually in between the Proto-Salish and the Halq'emeylem sets: they appear multi-morphemic, in that they contain separate identifiable morphemes, but they are syntactically simple in behaviour, in that the identifiable morphemes are not active syntactically. 3.4 Arguments and Predicates 3.4.1 Independent Pronouns in Argument Positions Independent pronouns in Secwepemctsin can be associated with argument positions. When an independent pronoun is associated with a DP, the pronoun's inherently contrastive value is added to the modified DP. 27 10 wi.w.k-t-0-en re n-tsetswe7 see(redup)-tr-3 sg. obj-1 sg. subj det 1 sg.indpr I saw him 11 a re John wik-t-0-s det John see-tr-3sg.obj-3sg.subj John saw Mary re Mary det Mary b newi7-s re John wik-t-0-s 3sg.indpr det John see-tr-3sg.obj JOHN saw Mary re Mary 3sg.subj det Mary c newi7-s wik-t-0-s 3 sg.indpr see-tr-3 sg. obj-3 sg. subj HE saw Mary re Mary det Mary Independent pronouns can also be the possessor in an argument DP. Matthewson (1996) has shown that possessors require determiners in Salish; it is shown below that independent pronoun possessors do not require determiners. 3sg.indpr det brother-3sg.poss like-tr-3sg.obj-3sg.subj det Mary HIS brother likes Mary 3.4.2 Independent Pronouns in Predicate Positions Independent pronouns can be predicates (13). In Salish, predicates do not take determiners, and this is illustrated in (14)15 with a regular nominal the man. The same is the case with independent pronouns (13, 15). A clefted predicate takes clausal arguments, which are headed by a determiner. 13 a newi7-s re wik-t-0-m-es 3sg.indpr det see-tr-3sg.obj-pas-3sg.conj It is HIM that saw him/her There may be counterexamples to the claim that predicates never take determiners. In the Southern Interior and in Upper Chehalis the predicate can take something that is homophoneous with a determiner (Matthewson, 1996). In Secwepemctsin, sometimes the first person singular independent pronoun can take the determiner re in predicate position. 12 newi7-s re uq'wi-s xwis-t-0-es re Mary 28 b *renewi7-s re wik-t-0-m-es *det 3sg.indpr det see-tr-3sg.obj-pas-3sg.conj It is HIM that saw him/her 14 a sqelemcw re wik-t-0-m-es man det see-tr-3sg.obj-pas-3sg.conj It is the man that saw him/her b *re sqelemcw re wik-t-0-m-es *det man det see-tr-3sg.6bj-pas-3sg.conj It is the man that saw him/her 15a newi7-s 1 m qwetsets-0 3sg.indpr det past leave-3sg.ind It was HER that left back then b 7-enwi7 1 m qwetsets-k pent'lu7 2sg.indpr det past leave-2sg.ind deic YOU are the one that left back then 3.4.3 First/Second Persons versus Third Person Given the appropriate context, it is possible to have two non-third person independent pronouns occupying the subject and object argument positions in a clause (16). This is not the case with third person independent pronouns. There can be no more than one third person independent pronoun per clause (17a-d). 16 a n-tsetswe7 xwe.xwis-t-si-n 7-enwi7 1 sg.indpr like(redup)-tr-2sg.obj-lsg.subj 2sg.indpr I love YOU b n-tsetswe7 wik-t-s-en 7-enwi7 lsg.indpr see-tr-2sg.obj-lsg.subj 2sg.indpr I saw YOU 17 a re John xwist-0-es re Mary, newi7-s ts'um-qs-0-ens det John like-3sg.obj-3sg.subj det Mary 3sg.indpr lick-nose-3sg.obje-3sg.subj John likes Mary, HE kisses her 29 b *reJohn xwist-0-es re Mary, *detJohn like-3sg.obj-3sg.subj det Mary, ts'um-qs-en-0-s newi7-s newi7-s lick-nose-fc-3sg.obj-3sg.subj 3sg.indpr 3sg.indpr John likes Mary, 3RD PERSON kisses 3RD PERSON (uninterpretable) c *reJohn xwist-0-es re Mary, *detJohn like-3sg.obj-3sg.subj det Mary, newi7-s ts'um-qs-en-0-s newi7-s 3sg.indpr lick-nose-fc-3sg.obj-3sg.subj 3sg.indpr John likes Mary, 3RD PERSON kisses 3RD PERSON (uninterpretable) d n-tsetswe7 wik-t-s-en 7-enwi7, lsg.indpr see-tr-2sg.obj-lsg.subj 2sg.indpr *newi7-s . wik-t-0-s newi7-s *3sg.indpr see-tr-3sg.obj-3sg.subj 3sg.indpr I saw YOU, HE, saw HTMj Another difference between the behaviour of third person independent pronouns versus first and second person forms will be further discussed in 3.5, and an explanation for this difference will be offered in 3.6 and 3.7. 3.4.4 Binding Properties of Independent Pronouns Independent pronouns have no intrinsic binding properties of their own. When they are attached to pro, they are subject to Condition B, like any regular pro. When they are attached to a lexical DP, they are subject to Condition C, like any regular lexical DP1 6. 16The reason for independent pronouns' having no intrinsic binding properties of their own will be clear once the category of independent pronouns is discussed in section 3.4.6. 30 (18) shows that a lexical DP in Secwepemctsin such as John is subject to Condition C, as in English; John cannot be coreferent with pro in the upper clause because it is an R-expression, which cannot be bound. When an independent pronoun is attached to it, the binding effects are the same. The lexical DP cannot be bound, because it is an R-expression. 18 a tsut-0 m qwetsets-0 re John say-3sg.ind past leave-3sg.ind det John Hei said that John*i/j left17 b tsut-0 m qwetsets-0 newi7-s re John say-3sg.ind past leave-3sg.ind 3sg.indpr det John He, said that JOHN*;/, left (19a, b) show that pro is not subject to Condition C, but to Condition B. This is expected of a pronoun. When an independent pronoun is attached to pro, the entire DP is still a pronoun, and is still subject to Condition B. (20), on the other hand, shows that independent pronouns are not subject to Condition C, because newi7-s is coreferent with the upper clause pro. It is impossible to have disjoint readings unless an overt nominal DP is available for the lower argument, which induces a Condition C effect. 19 a pro tsut-0 m qwetsets-0 pro pro say-3sg.ind past leave-3sg.ind pro Hei said that he,/»j left b re John tsut-0 m qwetsets-0 pro det John say-3sg.ind past leave-3sg.ind pro Johni said that he^ left 20 tsut-0 m qwetsets-0 newi7-s say-3sg.ind past leave-3sg.ind 3sg.indpr He, said that H E , left The equivalent of the coreferent interpretation of this sentence in St'at'imcets (Lillooet Salish) is well-formed (Matthewson and Davis 1995). 31 Quantifier binding shows that independent pronouns are variables when they are modifiers of pro. 21a xwexweyt re swet xwis-t-0-es re qe7tse-s all det who like-tr-3sg.obj-3subj det father-3sg.poss Everyone likes his/their own father18 b xwexweyt re swet xwis-t-0-es newi7-s re qe7tse-s all det who like-tr-3sg.obj-3subj 3sg.indpr det father-3sg.poss Everyone likes HIS/THEIR OWN father Independent pronouns do not affect the binding properties of the nominals they are associated with. Therefore, an R-expression remains an R-expression and continues to observe Condition C; a pronoun remains a pronoun, and continues to observe Condition B. 3.4.5 Independent Pronoun is an Adjunct to DP Since independent pronouns do not affect the binding properties of the nominals they are associated with, yet are still related to the nominals, the maximal projection of the argument must be the same as it would be without the independent pronoun. This means that the independent pronoun is contained within an argument DP. Coordination is evidence showing that independent pronouns are contained within DP. Only maximal projections of the same category can be successfully coordinated. The sentences in (22) show that any DP can coordinate with a DP plus an independent pronoun. 22 a [reMary]Dp ell [newi7-s re John]DP qwetsets-0 det Mary and 3sg.indpr det John leave-3sg.ind Mary and JOHN left 1 8 * * It is impossible to tell if the possessor pronoun is a singular or plural variable, since third singular and plural subjects are both null. 32 b [newi7-s re John]Dp ell [re Mary]DP qwetsets-0 3 sg.indpr det John and det Mary leave-3 sg.ind JOHN and Mary left c [reMary]Dp ell [newi7-s]Dp qwetsets-0 det Mary and 3sg.indpr leave-3sg.ind Mary and HTM left d [newi7-s]Dp ell [reMary]op qwetsets-0 3sg.indpr and det Mary leave-3sg.ind HE and Mary left It is likely that an independent pronoun is contained within a DP as a modifier. One type of DP modifier is the adjective. However, independent pronoun modifiers are very different from adjectival modifiers. 23 a wi.w.k-t-0-en [re lecus te nuxwenxw]Dp see(redup)-tr-3sg.obj-3sg.subj det pretty obi woman I saw the pretty woman b wi.w.k-t-0-en [newi7-s re nuxwenxw]oP see(redup)-tr-3sg.obj-3sg.subj 3sg.indpr det woman I saw THE WOMAN In the adjective-modified DP in (23 a), the oblique determiner occurs between the adjective and the noun, and the D head is at the left edge of the DP. In the independent pronoun-modified DP in (23b), the independent pronoun is on the left of the head D. For this reason, the attachment sites of independent pronouns and adjectives within DP must be different. Since modifiers are generally analyzed as adjuncts, I will propose that adjectives in Secwepemctsin are N(P) level adjuncts occurring to the right of the D head (24)19,while independent pronouns are DP level adjuncts (25a, b). ^ will n ° t be specific about the structure of an adjective-modified nominal, and have nothing to say about the syntactic position occupied by the oblique determiner te. The structure of the adjective-modified DP is provided to illustrate the difference between it and the independent pronoun-modified DP. 33 24 re AP det / \ leciis te nuxwenxw pretty obi woman 25 a Adjunction of independent pronoun to a regular DP2 DP XP DP X D NP newi7s re nuxwenxw 3sg.indpr det woman b Adjunction of independent pronoun to a pro DP XP DP I X I newi7s pro 3sg.indpr pro When this lexical independent pronoun is not adjoined to any DP, and is not modifying any DP, it functions as a predicate on its own, like any lexical category can in this language. Data below motivate the constructions proposed. While independent pronouns often appear adjacent to their coreferent overt DP (26a, 27a), they can also occupy the external topic position (26b, 27b), stranding their nominal referent. Evidence that they occupy this position The syntactic category of independent pronouns is not yet known, and is represented by X, which projects to XP, in the trees given. The independent pronouns always have to be coreferent with the DP that they adjoin to; they cannot refer to other nominals. 34 is supported by the pause after the external topic, of'as for...'. and by the consultant's English translation 26 a ts'um-qs-en-0-s re Mary [newi7-s re John] lick-nose-fc-3sg.obj-3sg.subj det Mary 3sg.indpr det John It was John who kissed Mary *It was MARY that John kissed b newi7-s, ts'um-qs-en-0-s re Mary re John 3sg.indpr lick-nose-fc-3sg.obj-3sg.subj det Mary det John As for JOHN, he kissed Mary (literally, as for HIM, John kissed Mary) *As for MARY, John kissed her 27 a m qwetsets-0 newi7-s rekukwpi7 past leave-3sg.ind 3sg.indpr det chief THE CHIEF left b newi7-s, m qwetsets-0 re kukwpi7 3sg.indpr past leave-3sg.ind det chief As for THE CHIEF, he left (literally, as for HIM, the chief left) The opposite situation is also possible. That is, an overt DP can occupy the external topic position while the independent pronoun (along with a pro DP), occurs in the subject argument position21. 28 a re xpe7e, newi7-s pixen-0-s re ts'i det grandfather 3sg:indpr hunt-3sg.obj-3sg.subj det deer As for grandfather, HE hunts deer b re kukwpi7 m qwetsets-0 newi7-s det chief past leave-3sg.ind 3sg.indpr As for the chief, HE left Since the independent pronouns used in (26-28) clearly are related to the overt DP referent, yet can be detached from it as a unit, it must be the case that the independent pronoun itself is also a maximal projection22. The observations above follow from the analysis that independent My description with respect to stranding of independent pronouns of their nominal referents is contra Kayne (1995), who claims that maximal projections can move, but segments of maximal projections cannot. I have shown that non-predicative independent pronouns must be associated with DPs, which are arguments. 35 22-pronouns are adjuncts to DP (whether lexical or covert). An independent pronoun is adjoined to a nominal and its maximal projection (DP), it can freely move as a constituent away from the modified DP, and the intrinsic binding property of the modified DP is not changed in any way. 3.4.6 Comparison with Halq'emeylem Now contrast the binding properties of Secwepemctsin to those of Halq'emeylem. Wiltschko (1998) shows that Halq'emeylem independent pronouns are R-expressions, and are subject to Condition C effects when in argument positions. Halq' emeylem Independent Pronoun * Adjunct to DP This subsection reviews the categorial analysis of Halq'emeylem independent pronouns (Wiltschko 1998). In Halq'emeylem, determiners are syntactically active with independent pronouns. That is, all determiners are productively used on independent pronouns (29), reduplication of lexical elements does not 'see' the determiner on independent pronouns (30), after prepositions, independent pronouns take oblique determiners, as all other nominals do, and where a regular nominal would require a determiner, the independent pronoun would also take a determiner, but no additional determiner is added in addition to the one already on the independent pronoun (31,32). Can independent pronouns be associated with adjuncts or indirect arguments? Independent pronouns do not co-occur with the oblique determiner te (see section 3.4.7); it is not clear whether they do not occur with te because they cannot occur with this determiner, or because they cannot occur with indirect arguments. However, there is one example showing that independent pronouns can follow the locative marker ne- (i), which markes locative adjuncts. Further research is necessary, (i) kec-t-si-s te sulen'sem ne 7-enwi7 give-tr-2sg.obj-3sg.subj obi flowers prep 2sg.indpr He gave some flowers to YOU 36 29 Independent 3rd pronouns with different determiners (Galloway 1993:403) male female human plural singular tu(:)tl'o thu(:)tlo — plural tutl'olem thutl'olem yutl'6(:)lem absent kwthu:tl:6 kwsu:tl'6 kwthu:tl'61em 30 Diminutive reduplication (CV) (Galloway 1993: 377) (Galloway 1980:32) (Wiltschko 1998:430) (Wiltschko 1998:433) the slha:li det woman (Wiltschko 1998.433) 32 a i:mex-tsel te-'a'elthe walk-lsg.s det-lsg.Indep I am walking " (Wiltschko 1998:433) b kw'ets-lexw-es tu-tlo thu-tlo see-tr(-3o)-3s det-3Indep det-fem-3Indep He sees her (Wiltschko 1998:433) A determiner is dropped in the usual syntactically defined environments. In Salish languages nouns in predicate position are not preceded by a determiner (33 a). Independent pronouns, when in predicate positions, do not appear with a determiner (33b). 33 a swiyeqe te i:mex man det walking That's the man that's walking (Wiltschko 1998:434) a sto:lo stotelo river creek b g'a:mi g'aq'emi girl little girl c tutl'6 tu:tl'6trem he little one d thutl'6 thu:tl'6tl'em she little one 31a hikw te swiyeqe big det man The man is big b kw'ets-lexw-es te swiyeqe see-tr(3o)-3s det man The man sees the woman. 37 b elthe te i:mex 1 sglndep det walking That's me that's walking (Wiltschko 1998:435) Wiltschko (1998) claims that independent pronouns in Halq'emeylem are AgrDs that are selected by D°s, and that take NPs as complements. AgrDP accommodates person and number features. The presence of AgrD licenses null NP, explaining why independent pronouns can occur with or without overt referents. 34 Structure of Halq'emeylem DP (Wiltschko 1998:439): DP D° AgrDP AgrD0 NP te tl'6 0 / swiyeqe det ind.pron 0 / man That independent pronouns in Halq'emeylem are not adjuncts to DP, while independent pronoun in Secwepemctsin are, has consequences for the syntactic behaviour of these pronouns, as the next section will show. Binding Properties Since Halq'emeylem independent pronouns in argument positions are their own DP projections, they should behave like full lexical DPs rather than personal pronouns - that is, they should behave like R-expressions with respect to binding. Wiltschko (1998) shows that this is the case. 38 First of all, Halq'emeylem independent pronouns should be subject to Condition C. This is attested in (35). 35 a suq'-t-es te swiyeqe te kopu-s look-for-tr-3s det man det coat-3poss The man; was looking for his; coat (Wiltschko 1998: 444) b siiq'-t-es te swiyeqe te kopu-s tut'16 look-for-tr-3s det man det coat-3poss det-3Indep The man; was looking for his»i/j coat (Wiltschko 1998. 444) Since the only difference between (35a) and (35b) is that in (35b) an independent pronoun is placed in the possessor position of the object DP, Wiltschko (1998) concludes that the ungrammaticality of the coreferent interpretation of (35b) must be due to the fact that independent pronouns in Halq'emeylem behave like R-expressions. Thus, the R-expression tut 'Id is c-commanded by the higher DP the man, and cannot be coreferent with the man. R-expressions cannot usually receive a bound variable interpretation23, and since Halq'emeylem independent pronouns are R-expressions, they are also expected not to receive bound variable interpretation. 36 a mekw' ye sewiyeqe xwoyi:wel lhi-s t'i:telem every .det.pi men happy when-3s singing Every man; is happy when he; sings (Wiltschko 1998:445) b mekw' ye sewiyeqe xwoyi:wel lhi-s t'i:telem tut'ld every men happy when-3s singing 3Indep Every man; is happy when he*i/j sings (Wiltschko 1998:445) In (36a), a bound variable interpretation is possible, but if an independent pronoun is present as the overt argument of sing (36b), then the bound variable reading is impossible to attain. It is possible for R-expressions to receive bound variable interpretations; epithet readings are an example: Every boy 'st mother loves the boy,. 39 Wiltschko thus concludes that the binding properties of Halq'emeylem independent pronouns are evidence that these pronouns are lexical argument DPs, and that they follow from the structure in (34). 3.4.7 The Category of Independent Pronouns Pronouns have been argued to be pro forms of DPs rather than NPs. One common claim is that they are intransitive determiners (Postal 1969), occupying the D° position, licensing an empty complement NP (Abney 1987). On the other hand, Wiltschko (1998) argues that independent pronouns in Halq'emeylem are not Ds per se, but do occur with Ds (See review of Wiltschko (1998) in 3.4.6). It is possible to analyze Secwepemctsin pro as a determiner occupying D° position, projecting up to a DP. However, are independent pronouns in Secwepemctsin also DPs? Do they occur with determiners? What possible categories could independent pronouns be? Since they are associated with nominals, we will look at the following possibilities: (a) Independent pronouns are Ns (b) Independent pronouns are Ds. (c) Independent pronouns are DPs. Independent pronoun * N Independent pronouns cannot be Ns. N(P)s in argument position in Secwepemctsin obligatorily take determiners. Independent pronouns do not take determiners in argument position (37). Secondly, independent pronouns co-occur with Ns when they appear adjacent to an overt referent (38); an independent pronoun occurs to the left of a D while a overt nominal occurs to the right of the D as a complement. (38) tells us that since woman already 40 occupies the complement N position, the independent pronoun cannot occupy the same N position. 37 [(*re) newi7-s] wik-t-0-s re John [(*det) 3sg.indpr] see-tr-3sg.obj-3sg.subj det John SHE saw John 38 [newi7-s reD° [nuxwenxw]N] wik-t-0-s re John [3sg.indpr det woman] see-tr-3sg.obj-3sg.subj det John THE WOMAN saw John Moreover, independent pronouns cannot be heads of complex nominal predicates, which must be nominal; this has been shown independently in Davis, Lai and Matthewson (1997). (39a) shows a noun heading the complex nominal predicate; (39b) shows that the sentence is ill-formed when an independent pronoun replaces the noun. 39 a [yiri7 te [sqelemcw]N]cNP 1 wi.w.k-t-sem-s deic obi man det see(redup)-tr-lsg.obj-3sg.subj That's the man that saw me b *[yiri7 te [newi7-s]*N]cNP 1 wi.w.k-t-sem-s *deic obi 3sg.indpr det see(redup)-tr-lsg.obj-3sg.subj That's HTM that saw me Independent Pronoun * D Independent pronouns can co-occur with D when they appear adjacent to an overt referent (40). 40 [newi7-s reo0 [nuxwenxw]N] wik-t-0-s re John [3sg.indpr det woman] see-tr-3sg.obj-3sg.subj det John THE WOMAN saw John Since re already occupies D°, newi7-s cannot be in the same D°. It is of course plausible to have a D° taking a DP complement, as in (41). 41 41 DP D DP D NP newi7-s 3 sg.indpr re nuxwenxw det woman This would suggest an argument can contain two determiners, which is not unlikely, as Giusti (1994) has reported that there are languages having arguments with two determiners24. The analysis would require independent pronouns to trigger a DP on top of existing DP arguments. However, if independent pronouns were analyzed as D, there is no explanation for why independent pronouns cannot replace other determiners, as shown below. Independent pronouns cannot be in D position, either by being base-generated there, or by being moved there. Independent pronouns cannot replace D (42a, b). 42 a *newi7-sD nuxwenxw b reD nuxwenxw Moreover, D never occurs in predicate position; since independent pronouns can be predicates, they must not be Ds. Independent Pronoun * DP Since independent pronouns are not D heads, they cannot be heads of DPs. I have also established they are not NP complements of Ds. If independent pronouns are not Ds or Ns, however, Giusti (1994) does not analyze such arguments as having 2 DP projections. Rather, she claims that demonstratives occupy the Spec DP position while determiners occupy the D-head position. She points out that the markedness of two determiners per argument may be due to a "doubly-filled DP filter". 42 could they still possibly be DPs? One piece of evidence suggests that they might be. According to Kuipers (1974:59), who did research on the Canim Lake and Alkali Lake dialects of Secwepemctsin, the singular independent pronouns are all recorded with the articles re or te25, but not any other article. This is so even in predicate position (43). 43 re n-tsetswe7 1 wcste-0-c26 det 1 sg.indpr det mention-3sg.obj-2sg.subj I am the one you mentioned Literally: The "I" you mentioned is ME : (Kuipers 1974:117, line 39) However, this thesis concentrates on the dialect of Secwepemctsin spoken by an elder from Skeechestn, in the southern portion of the Shuswap territory. It is not unusual in this dialect to have phonetic deletion of the determiner re21. The language consultant does not have any determiner (either re or te) co-occurring with independent pronouns28, whether they are in argument or predicate position. The only exception to this observation is with the first singular independent pronoun, which can optionally take re in all argument and predicate positions (44), according to my consultant. Such judgments suggest that re may have been merged into the first person singular independent pronoun, and is not syntactically active, which means that an independent pronoun cannot be a DP while being an argument, and a non-DP while being a predicate. ^Kuipers (1974) glosses re as the present absolutive actual determinate article, while te is glossed as the relative actual determinate article. I have consistently glossed re as the (present) determiner and te as the oblique determiner. 2 6 / - is the absent determiner. 2 7This observation arose from my fieldwork on the Skeetchestn and Kamloops dialects of Secwepemctsin and from conversations with teachers and speakers of the language in Kamloops, B.C. (Mona Jules, Marianne Ignace, Lisa Matthewson, p.c). However, I am uncertain whether the re deletion is a case of dialectal difference. According to Kuipers (1974), there are only two dialects of Secwepemctsin: the Eastern, and the Western, although Newman (1977) has distinguished between Northern Shuswap and Southern Shuswap. Data which appeared in Kuipers (1974) belong to the Western dialect, as does the Secwepemctsin spoken in the Skeechestn and Kamloops area. 28According to Newman (1977), Halq'emeylem is the only Salish language where determiners are found on independent pronouns. 43 44 a re n-tsetswe7 re wik-t-0-m-es det lsg.indpr det see-tr-3sg.obj-pas-3sg.conj It is I that saw him b re n-tsetswe7 wi.w.k-t-0-en re Mary det lsg.indpr see(redup)-tr-3sg.obj-lsg.subj det Mary I saw Mary c *re newi7-s wik-t-0-s re Mary *re3sg.indpr see-tr-3sg.obj-3sg.subj det Mary HE saw Mary Although the lack of determiners accompanying independent pronouns in the elicited data may arise from the fact that speakers of this area often delete re [ya] phonologically, I discredit this possibility for the following reasons. A. Even if the direct determiner re is dropped due to phonological reasons, there is no reason for the oblique determiner te to be absent, since there has not been any known case of te deletion in Secwepemctsin anywhere in the territory. In (45a, b) below, the sentences are grammatical in the data cited by Gardiner (1993), but ungrammatical when tested with the language consultant involved in the present project. 45 a */•/ te n-tsetswe7 re wikt-(t)-0-m-es "TV obi lsg.indpr det see-tr-3sg.obj-pas-3sg.conj I'm the one who saw him (It is I that saw him) b */V te 7-enwi7 re wikt-(t)-0-m-es */V obi 2sg.indpr det see-tr-3sg.obj-pas-3sg.conj YOU're the one that saw him (It is YOU that saw him) B. The language consultant offers strong intuitions about the presence or absence of determiners. She is able to distinguish between cases where "re is there, but you just don't hear it", versus cases where re is simply absent. She insists that re is absent on independent pronouns29. 29Marianne Ignace, who teaches Secwepemctsin in Kamloops, B.C., offers equally strong judgments that re 44 Whether independent pronouns can appear with determiners or not, it is a fact that independent pronouns can be in the predicate position, without being different from the form it assumes while in argument position. Since no DP can occupy predicate position, while an independent pronoun can, independent pronouns must not be DPs. Independent Pronoun is of Category X I have so far concluded that independent pronouns are not Ns, Ds, or DPs. It follows that they do not have binding properties of their own, as binding only applies to N/D categories. Since at least person and number features are encoded by independent pronouns, we might consider analyzing them as functional projections (for example, AGR). However, this cannot be correct. Independent pronouns can function as predicates as well as arguments, and since functional projections cannot be predicates30, independent pronouns cannot be functional projections. should be present, and observes that speakers in Williams Lake, B.C. also demonstrate the presence of re on independent pronouns in their speech. Note that if this judgment is correct, the structure of independent pronouns would be quite different from that which is suggested under 3.4.5. If independent pronouns take Ds in argument position, and drop them in predicate position, independent pronouns would behave like nominal categories (i). If independent pronouns always follow D, in predicate as well as in argument position, then this could be a case of incorporation, where a determiner and an independent pronoun are incorporated into a single syntactic atom, whose internal structure is unanalyzable syntactically. This suggests that there would be no syntactically active determiner for independent pronouns. It is still possible for independent pronouns to be DPs (ii) (iii), but this prevents an analysis of predicative independent pronouns in the current model of clefting adopted (see 2.3.2); an actual cleft movement without empty operator would be necessary, (i) DP (ii) / \ D NP I A . _ . . det indpr indpr 0 det+indpr 3 0This claim is disputed. Wiltschko (1998) does have an AgrDP occupying the predicate position. 45 3.5 Subject - Object Asymmetry 3.5.1 Asymmetry in Secwepemctsin Independent pronouns have to be the subject of a clause in contexts where the arguments are all third person. I will call this subject orientation of Secwepemctsin independent pronouns the "Independent Pronoun Restriction", and henceforth abbreviate it IPR. The EPR stands regardless of whether the independent pronoun is preverbal (46) or post-verbal (47). Readings where an independent pronoun is interpreted as the object are not possible. 46 a newi7-s wik-t-0-s re John 3sg.indpr see-tr-3sg.obj-3sg.subj det John FIE saw John / * John saw HTM b *reJohn wik-t-0-s newi7-s *det John see-tr-3sg.obj-3sg.subj 3sg.indpr *HE saw John / *John saw HTM 47 wik-t-0-s re John newi7-s see-tr-3sg.obj-3sg.subj det John 3sg.indpr HE saw John / *John saw HIM The IPR is also true for cases where the independent pronoun appears adjacent to its overt referent. In (48) the pronoun is subject oriented, and in (49) all readings where the pronoun is object oriented are ill-formed. Readings (ii) and (iii) are ill-formed because the independent pronoun is object oriented, and reading (iv) is ill-formed because newi7s re John, the only overt post-verbal nominal, cannot be interpreted as the subject. 48 [newi7-s re Mary] ts'um-qs-en-0-s re John [3sg.indpr det Mary] lick-nose-fc-3sg.obj-3sg.subj det John MARY kissed John 46 49 ts'um-qs-en-0-s newi7-s re John lick-nose-fc-3sg.obj-3sg.subj 3sg.indpr det John (i) SHE kissed John (ii) *John kissed HER (iii) * Someone/he/she kissed JOHN (iv) *JOHN kissed someone (50a) below presents an interesting case. It is marginal because it is caught in a conflict of grammatical constraints. The ONI Law (section 2.2.2) requires that the post-verbal overt nominal (independent pronoun) be the object, while the independent pronoun forces it to be the subject of the clause. The sentence cannot satisfy both restrictions at the same time, and becomes difficult for the consultant to interpret. (50b), on the other hand, satisfies the IPR, and the one nominal constraint does not apply, because the single overt nominal is preverbal, so it is fully grammatical. 50 a ?wik-t-0-s newi7-s ?see-tr-3 sg.obj-3 sg. subj 3 sg.indpr HE saw someone / * Someone saw HIM/HER b newi7-s wik-t-0-s 3sg.indpr see-tr-3sg.obj-3sg.subj HE saw something/somebody Such judgments as those of (50a,b) show that the restriction on subject orientation is stronger than the one nominal interpretation when there is a clash of the two rules. Indeed, Davis (1994) has shown that the one nominal constraint can be overruled if pragmatic factors are strong enough. In sentences with non-third person arguments, there is no subject orientation for independent pronouns. The IPR applies only in cases where both arguments are third person. The sentences below are judged to be grammatical by the language consultant. 47 51 a wi.w.k-t-0-en n-tsetswe7 see(redup)-tr-3 sg.obj-1 sg. subj 1 sg.indpr I saw him b n-tsetswe7 lsg.indpr I saw him wi.w.k-t-0-en see(redup)-tr-3 sg.obj-1 sg. subj c wi.w.k-t-0-en newi7-s see(redup)-tr-3 sg.obj-1 sg. subj 3 sg.indpr I saw HIM d wi.w.k-t-sem-s n-tsetswe7 see(redup)-tr-1 sg.obj-3 sg. subj 1 sg.indpr He saw ME 3.5.2 Asymmetry in Halq'emeylem The asymmetry whereby third person independent pronouns have to be associated with the subject of a transitive clause is absent in Halq'emeylem. It is possible to have two third person arguments in a transitive clause, both arguments surfacing as independent pronouns. 52 kw'ets-lexw-es tii-tlb thu-tlo see-tr(-3o)-3s det-3Indep det-fem-3Indep He sees her (Wiltschko 1998:433) Note that Galloway (1993) has claimed that in a sentence such as the one presented above, the word order is strictly VSO. 3.6 Correlations Section 3.5 shows that there is an asymmetry involved between the subject position and the object position where third person independent pronouns are involved. Namely, in clauses containing two third person arguments, one of which is an independent pronoun, this 48 independent pronoun always has to be the subject, and not the object. Similar subject-object asymmetries have also been observed with respect to the distribution of overt arguments, the One Nominal Interpretation Law (Gerdts 1988) and A' Subject clefts are examples. 3.6.1 One Nominal Interpretation In a clause containing a predicate with two third person argument bound pronominals and an overt post-predicate DP, the DP is always the object. This is the One Nominal Interpretation (Gerdts 1988). 53 One Nominal Interpretation Law (Gerdts 1988) In the absence of marking for other persons, a single third person nominal is interpreted as the absolutive. 54 nik'-n-0-s re nuxwenxw cut-fc-3sg.obj-3sg.subj det woman (i) He cut the woman (ii) *The woman cut him The one nominal interpretation does not apply when a clause does not contain two third person arguments. 55 wi.w.k-t-0-en ntsetsewe7 see(redup)-tr-3sg.obj-lsg.subj lsg.indpr I saw him 3.6.2 Clefting the Third Person Subject In Secwepemctsin, arguments can be extracted via A'-movement to the wh/focus position (see section 2.3.2). All arguments can be realized in the wh/focus position with no change in the lower predicate (56a), with the exception of third person subjects in clauses containing only 49 third person arguments (56b). The said third person subject cannot appear in the wh/focus position unless passive and third' person conjunction morphology have been added to the predicate. 56 a John (re) wik-t-0-s John (det) see-tr-3sg.obj-3sg.subj It's John s/he saw / *It's John who saw him/her b John (re) wik-t-0-m-es John (det) see-tr-3sg.obj-pas-3sg.conj It is John who saw him/her / *It is he who saw John The extracted argument is the cleftee, which occupies the predicate position of the main clause, while the syntactic argument of this. predicate is a dependent clause headed by a determiner. When an argument from a clause that does not contain two third person arguments (57), or an object from a clause with two third person arguments (56a) is A' extracted, no such special morphology appears. 57 Scott re xwe.xwistet-0-en Scott det like(redup)-3sg.obj-lsg.subj It's Scott that I like I summarize in (58) what the One Nominal Interpretation, the Subject Cleft, and the Independent Pronoun Restriction have in common. (58a)(i) shows that a clause can have one single overt DP that is interpreted as the subject if it is either first or second person, (ii) and (iii) show that when a clause contains two third person arguments, if there is only one overt DP and it is post-verbal, it must be the object. (58b)(i) shows that it is grammatical to cleft a third person singular subject when the object is non-third person; when both the subject and object are third person, the clefting of the subject is ungrammatical. (58c)(i) shows that in 50 clauses containing arguments that are non-third person, an independent pronoun can be in the object position. However, when a clause contains arguments that are both third person, an independent pronoun cannot be in the object position. 58 Summary of correlations a. ONI Law31 b. Subject Cleft c. IPR (i) VV-1/2-1/2 1/2 subj (ii) * V-3-3 3 subj (iii) V V-3-3 3 obj (i) V 3 subj 1/2 obj J (ii) * 3 subj 3 obj J (i) V 1/2 subj 3 obj i indpr (ii) * 3 subj 3 obj I indpr The correlations are as follows. The ONI, the Subject Cleft, and the IPR (a) are all limited to clauses containing two third person arguments, and (b) point to an asymmetry between subjects and objects. 3.7 Analysis The ONI, the Subject Cleft, and the IPR are all active and distinct constraints in Secwepemctsin grammar. However, at a certain level, they are all related to the Familiarity Hierarchy (to be discussed below). Note that the constraints discussed here are necessary only when there is ambiguity; that is, only when two third person arguments are involved. 3.7.1 Familiarity 3 1 To illustrate the pronominal morphology on the predicate, I use V for 'verb', and the numbers 1, 2, and 3 to signal first, second, and third person, respectively. 51 I will establish that discourse referents are ranked according to familiarity; that is, according to the amount of familiar-novel information that they contain. The Familiarity Hierarchy is an interpretation of the nominal hierarchy32 (Dixon 1979, 1994), and is comparable to the continuum-based view of referentiality in Anagnostopoulou & Giannakidou (1995). This being a hierarchy, it is necessarily scalar,. despite the fact that a familiar versus novel distinction can be made. The continuum encodes the degree of familiarity or identifiability particular referents have. A familiar referent is anaphoric in a discourse, either by virtue of having a discourse antecedent, or by being present and prominent in the discourse context. A novel referent is new in the discourse and is not easily identified with the information in the discourse common ground; for instance, something mentioned for the first time in a discourse is novel. Familiarity will be defined as follows: 59 Familiarity An item Xi is more familiar than an item x2 if Xi is more anaphoric than x2 I will assume that elements in the discourse structure get mapped onto the syntactic structure via the Familiarity Hierarchy, with the most familiar element mapped to the highest syntactic position (see also Erteshik-Shir 1993). Within a discourse, the referents that are always readily available are the discourse participants: the speaker, and the hearer(s). They are referred to by the first and second person pronouns, and are inherently present in the discourse. In this sense, the participants of a discourse are most familiar. The discourse participants are so omnipresent in any given 3 2The nominal hierarchy has also been referred to as topicality, definiteness, agent, referentiality hierarchy, etc.. The approximate order of the nominal hierarchy is as follows: 1 pronoun > 2 pronoun > 3 pronoun/deictics > proper noun > animate common noun > inanimate common noun ... In subsequent sections evidence will show that the familiarity hierarchy referred to in this thesis is certainly not a definiteness or agent hierarchy. 52 discourse that there is no restriction on where they can appear in the discourse or in a sentence within a discourse. On the other hand, all non-participants in a conversation take grammatical third person; this is the case whether a pronoun or an R-expression is used. There are restrictions on the occurrence of third person arguments, and these restrictions are discourse-oriented. Distinctions within grammatical third person need to be made, since third person is encoded with pronominals as well as R-expressions.' Pronominals are anaphoric and need discourse antecedents. This means that referents of pronouns, when well-formed within a discourse, are not new information within a discourse context; rather, they are already available in the immediate discourse context. This follows from the standard claim that pronouns cannot introduce a discourse referent. Pronouns are thus topical and presuppositional. R-expressions, on the other hand, are generally used when a new referent is brought into the discourse. Matthewson (1996) has claimed that Salish determiners (heads of R-expressions within the DP hypothesis (Abney 1987)) do hot encode definiteness or specificity. This claim entails that R-expressions per se are neither definite nor specific, which means that R-expressions in Salish are non-presuppositional, and are necessarily new information. Since pronominals are contextually not only familiar but specific, they are more familiar than R-expresssions. Next, wh elements in a sentence correspond to wh words in a question; therefore, they are certainly new information, and are more indefinite than R-expressions. We have now arrived at the ranking stated below. 60 Familiarity Ffierarchy (nominals) 3pro > R-expressions > wh 53 Independent pronouns also have to be ranked in terms of familiarity. Independent pronouns in Secwepemctsin are used in contrastive contexts. They contrastively mark the modified pro or lexical DP as the referent, as opposed to other potential referents, x, y, or z. In other words, an independent pronoun is a contrastive modifier. When an independent pronoun is added to pro or lexical DP, the function is to contrastively emphasize. In addition to the antecedent referent necessary for the pro (in the case of lexical DPs the overt DP satisfies as a referent) in question, further information is necessary to identify and set apart contrastively that referent from a set of potential discourse-restricted supersets. A whole set of potential alternatives have to be brought forth in the discourse common ground. Needing a set of antecedents is more anaphoric than needing one single discourse antecedent (see Chapter 4 for a semantic analysis of independent pronouns). For this reason, when a pro is modified by an independent pronoun, it is more familiar than a pro by itself; likewise, a DP with an independent pronoun is also higher than a single DP without the contrastive independent pronoun. 61 Familiarity Hierarchy (contrastiveness) contrastive > non-contrastive The two ranking orders that will explain the ONI, Subject Cleft, and DPR phenomena are re-stated as follows: 62 Familiarity Hiearchy a) nominals: 3pro > R-expressions > wh b) contrastiveness: contrastive > non-contrastive Familiarity in discourse is mapped onto syntactic positions via the principle stated below (see also Erteshik-Shir 1993). 63 Mapping Principle The most familiar referent maps onto'the highest argument position. Let us now explain the data by mapping discourse referents onto syntactic positions. 54 3.7.2 Analysis of the One Nominal Interpretation According to the Familiarity Hierarchy, a lexical DP is less familiar than pro. When there is only one overt DP in a transitive sentence, the second argument is covert, and it is pro. Since the most familiar referent has to be mapped onto the highest argument position, in this case, pro (more familiar than lexical DP) has to be mapped onto the subject position, which is higher than the object position. (64b) shows the Mapping Principle being followed, resulting in a grammatical sentence, while (64a) shows the Mapping Principle being violated, resulting in an unrammatical sentence. 64 nik'-n-0-s re nuxwenxw cut-fc-3sg.obj-3sg.subj det woman (Gardiner 1993) a) *The woman cut pro b) pro cut the woman *DP - pro pro - DP ""subject - object subject - object *less familiar - more familiar more familiar - less familiar 3.7.3 Analysis of Subject Clefts When an argument is clefted and is predicative, it is coindexed with an empty operator that has a wh trace in the lower clause (see section 2.3.2). If the lower clause is transitive, then it would contain a wh-trace (coindexed via empty operator with the cleftee) and either a lexical DP or pro. Since both lexical DPs and pros are both more familiar than wh elements, a lexical DP or pro is mapped onto the subject position while a wh-trace is mapped onto the object position (65a)33. (65b) violates the mapping principle, and is ungrammatical. 3 3 I assume the Copy Theory (Chomsky 1995) so that the tail of a chain is identical to the head of the chain in terms of features. 55 65 re John, [re O* wik-t-0-s t; re Bill] det John [det OPERATOR see-tr-3sg.obj-3sg.Subj TRACE det Bill a) It is John that Bill saw b) *It is John that sees Bill John, see t; Bill *Johni see t, Bill wh - DP *wh- DP obj - subj *subj - obj less - more familiar *less - more familiar In order for the semantic subject of the dependent clause to be clefted, the subject is demoted to adjunct status via passive and conjunctive morphology. Wh elements are low in the Familiarity Flierarchy, and it does not matter that adjunct positions take items of low familiarity. This is why a third person subject is clefted with passive and conjunctive morphology when the object is also third person. 3.7.4 Analysi s of the Independent Pronoun Restriction When a DP (whether lexical or pro) is modified by an independent pronoun, it is contrastive. When such a DP is not modified by an independent pronoun, it is not contrastive. Contrastive elements are more familiar than non-contrastive elements, therefore, contrastive elements get mapped onto the subject position while hon-cbritrastive elements get mapped onto the object position. (66a) is grammatical while (66b) is not because the R-expression modified by the independent pronoun is contrastive, and therefore more familiar, than a plain DP. (67a) is grammatical while (67b) is not because a pro modified by an independent pronoun is contrastive, and therefore more familiar than a plain R-expression. 56 66 wik-t-0-s [newi7-sreMary] re John see-tr-3sg.obj-3sg.subj 3sg.indpr det Mary det John a) MARY saw John b) *John saw MARY contr. - non-contr. *nonr-contr - contr. subj - obj *subj - obj more - less familiar *less - more familiar 67 wik-t-0-s re John newi7-s see-tr-3sg.obj-3sg.subj det John 3sg.indpr a) HE saw John b) *John saw HIM contr. - non-contr. *non-contr - contr. subj - obj *subj - obj more - less familiar *less - more familiar (68) and (69) are plainly ungrammatical because the DV/pro modified by the independent pronoun is contrastive, and more familiar than a pro. 68 wik-t-0-s pro [new'il-spro] see-tr-3sg.obj-3sg.subj pro 3sg.indpr *He saw HTM 69 wik-t-0-s pro [newi7-s re Bill] see-tr-3sg.obj-3sg.subj pro 3sg.indpr det Bill *He saw BILL However, there are also questionable intepretations associated with (68) and (69); these are given in (70) and (71) below. 70 wik-t-0-s pro [newi7-s pro] see-tr-3sg.obj-3sg.subj pro 3sg.indpr ?HE saw him 71 wik-t-0-s pro [newi7-s re Bill] see-tr-3sg.obj-3sg.subj pro 3sg.indpr det Bill ?BILL saw him The reason (71) is questionable but not completely ungrammatical is due to the fact that the Familiarity Hierarchy also requires pro to be more familiar than lexical DPs, preferring pro as 57 the subject over the lexical (although contrastive) DP. This conflict results in the questionable judgment. Note, however, that the interpretation allowing pro to be subject and the contrastive DP to be object is plainly ungrammatical, which shows that it is ultimately more important for the grammar to satisfy Familiarity Hierarchy (contrastiveness) than Familiarity Hierarchy (nominals). I am at a loss to explain the questionable judgment of (70). Since all pros should be equally familiar, and contrastive pros are more familiar than non-contrastive pros, the interpretation in (70) should be perfectly grammatical according to my analysis. Note that if the pro DP modified by an independent pronoun were preverbal, as in (72), the sentence is perfectly well-formed with the independent pronoun interpreted as the subject. 72 [mvAl-sprp] wik-t-0-s pro 3sg.indpr/?ro see-tr-3sg.obj-3sg.subj pro (i) HE saw him (ii) *He saw HTM In summary, I have claimed that pros are more familiar than R-expressions, and that when independent pronouns adjoin to either pros or R-expressions, the modified DPs become more familiar than they already are. Applying the Mapping Principle, I claim that the most familiar DPs map onto the highest argument position (subject), and less familiar DPs are left for lower argument positions (object or oblique). This argument applies when ambiguity is present; that is, when there are two third person arguments within a clause. Although this argument relies on the discourse notion of familiarity, it also crucially relies on the fact that there is something intrinsically different about subjects and objects. If subjects and objects had equal status, then it would not matter whether it is the subject or the object that is familiar. 58 3.7.5 Blocking So far, I have analyzed the ONI, Subject Cleft, and TPR with the presumption that the proposed constraints and mapping principle are needed only when potential ambiguity is present; that is, when both arguments in a clause are third person. If the Blocking Principle (Williams 1997) is applied, this presumption can be eliminated. The Blocking Principle is employed in the following way: When there are two distinct morphological (or syntactic) forms, there must necessarily be two distinct meanings or interpretations associated with them; that is, there must be a one-to-one correspondence between form and meaning. If there are two forms with the same meaning, one of the forms is blocked. The blocked form is always less specific, while the existing form is always more specific. For instance, in English, the form 'happier' blocks the form 'more happy' because 'happier' is specific (marked) while 'more happy' is general (and belongs to the 'elsewhere' cases). For a given clause with two third person arguments, two potential forms are available (73). 73 a) arguments argumentB b) argumentA arguments >!>• ^ . X subject object object subject The Mapping Principle dictates that the most familiar DP must map onto the highest argument position (subject) (see section 3.7.1). A form observing the Mapping Principle is specific, and results in its blocking the alternative form, where the more familiar form is not mapped onto the highest argument position. Thus, if third person arguments were more familiar than third person argumentB, then having arguments map onto the subject position as that shown in 59 (73a) blocks out the form in (73b), where the more familiar form argumentA is mapped onto the object position. The Blocking Principle thus conforms to the analysis already given in this thesis and does not change the results yielded by the existing analysis in any way. However, the virtue of the Blocking Principle lies in its applying blocking only when there are two ambiguous forms -when both subject and object pronominal markers on the predicate are third person. This accounts for why the ONI, Subject Clefts, and IPR phenomena do not occur in clauses that do not contain two third person arguments. When a clause's arguments are different in person, the subject and object persons are encoded via pronominal morphology on the predicate, so no ambiguity can arise. In other words, there is only one single potential form that has one single interpretation. Since blocking does not apply when there is a one-to-one correspondence between form and meaning, it is not necessary to apply the Familiarity Hierarchy and Mapping Principle when a clause contains arguments that are non-third person. The two examples below show that an independent pronoun can be the object when there is no potential ambiguity with respect to the interpretation. 74 a wi.w.k-t-0-en newi7-s see(redup)-tr-3sg. obj-lsg. subj 3sg.indpr Essentially, the Blocking Principle allows for a more formal way of stating the presumption that the Familiarity Hierarchy and the Mapping Principle apply only when there is potential I saw HIM b wik-t-s-en see-tr-2sg. obj-1 sg. subj I saw YOU 7-enwi7 2sg.indpr ambiguity. 60 3.8 Implications: Test for Subjecthood In this section, I will discuss an implication that arises from my analysis of the independent pronouns and of the subject-object asymmetry I have argued for. An implication resulting from the study of Secwepemctsin independent pronouns is that the use of a third person independent pronoun becomes a test for subjecthood. If third person independent pronouns contained within a clause with only third person arguments always have to be interpreted as the subject, then third person independent pronouns can be used as reliable evidence on whether a given position is necessarily a subject position. One possible application, of the test is passivization, as shown in the data presented below. There has been an extensive debate among Salish linguists regarding the passive construction, centering around the question of whether the patient truly achieves syntactic subject status in a passivized sentence (Hukari 1976, Gerdts 1987, 1989, Blake 1997). The use of independent pronouns shows evidence that patients in Secwepemctsin indeed get promoted to subject. (75a, b) are two ordinary transitive sentences, (b) cannot be interpreted as 'Mary kissed him' because the independent pronoun cannot be interpreted as the object. 75 a ts'um-qs-en-0-s re Mary re John kiss-ls-fc-3sg.obj-3sg.subj det Mary det John John kissed Mary / Mary kissed John b ts'um-qs-en-0-s re Mary newi7-s kiss-ls-fc-3sg.obj-3sg.subj det Mary 3sg.indpr HE kissed Mary / *Mary kissed HTM 61 Passivization in Secwepemctsin is achieved by replacing the subject suffix with the passive marker -m on the predicate. The semantic object (theme or experiencer) remains in the clause, as well as the object suffix on the predicate, and the semantic subject (agent) is demoted so that no marking on the predicate is present, and the overt agent DP is either marked with an oblique te or completely disappears (76). (77) shows that the derived subject DP can be replaced by an independent pronoun, evidence for subject status. After the semantic subject (agent) is demoted, it can no longer be an independent pronoun (78). 76 ts'um-qs-en-t-0-em (teJohn) re Mary lick-nose-fc-tr-3sg.obj-pas (obi John) det Mary Mary was kissed (by John) 77 ts'um-qs-en-t-0-em (te John) newi7-s lick-nose-fc-tr-3sg.obj-pas (obi John) 3sg.indpr SHE was kissed (by John) 78 *ts'um-qs-en-t-0-em [te newi7-s] f e Mary lick-nose-fc-tr-3sg.obj-pas [obi 3sg.indpr] det Mary Mary was kissed by HTM Compare the ungrammatical sentence in (78) with the grammatical sentence in (75b). The themantic roles of newi7-s in both sentences are exactly the same; it is the syntactic status that makes them different. The distinction between non-passivized transitive sentences and passivized sentences shown here presents important evidence that promotion of the object does result in a newly derived subject, even though the newly derived subject is being marked by an object suffix on the predicate. I will not attempt an explanation for the object suffix, and will leave this question open34. Passivization in Secwepemctsin indeed involves syntactic promotion of the object to ^Gerdts (1989) demonstrates that in Halq'emeylem, the object suffix appears in passive constructions not because the argument has object status, but because it is morphologically fused with the transitive suffix, and cannot be dropped. 62 subject, and demotion of the agent to oblique status. It is not merely a process of eliminating the ergative argument. This conclusion is achieved because independent pronouns can be used as a test for subjecthood. 3.9 Remaining Questions The Familiarity Hierarchy says nothing about clauses having arguments of the same familiarity 'force'. That is, when a clause has two pro. arguments, or two DP arguments, there is potential interpretive ambiguity; either pro, or either DP, can be interpreted as the subject. 79 wik-t-0-s re Mary re John see-tr-3sg.obj-3sg.subj det Mary det John (i) John sees Mary (ii) Mary sees John With independent pronouns, it is only grammatical to have two non-third person arguments within a clause. It is ungrammatical to have two independent pronouns per clause when at least one of those pronouns is third person. It is perhaps the case that there can only be one contrastive element per clause, in which case the prediction would arise that not only is a clause ill-formed with two contrastive arguments, but a clause is also ill-formed when there is a contrastive argument and a contrastive predicate35. Another question that remains is that we have seen clauses in which the ONI and the IPR both apply. In such cases, the grammar prefers to satisfy the IPR even though this would violate the ONI Law. That the ONI Law is governed by Familiarity Hierarchy (Nominals) and the IPR is governed by Familiarity Hierarchy (Contrastiveness) provides a clue as to why both 3 5The speculation there there can only be one contrastive element would have to hold only for clauses with two third person arguments, since it is possible to have two independent pronouns per clause if they are both non-third person (see section 3.4.3). 63 constraints can be based on the same notion of familiarity, but are distinct and unequal in terms of actual application. I will not venture further into finding out at which level the two constraints interact, and at which level the two constraints are clearly separate. Finally, even though we have said that clauses containing two third person arguments are ambiguous in that either of the arguments can be interpreted as the subject, the definition of 'clauses containing two third person arguments' needs to be further refined. It is really in the clauses containing two post-predicate third person arguments where potential ambiguity can arise. Note that this may follow from Gardiner's (1998) analysis of all preverbal positions as topic positions (see section 2.3). Since a preverbal position is a topic, and therefore is more familiar than a post-verbal argument position, there would be no difficulty in the mapping of a familiar argument to a DP in a preverbal position. 3.10 Chapter Conclusion In this chapter, I have illustrated the internal and external syntactic behaviour that independent pronouns in Secwepemctsin, and I have shown that their binding properties follow from their syntactic category. A subject-object asymmetry emerged from the study of independent pronouns. When a clause has two third person arguments, if one of the arguments contains an independent pronoun, that argument is always the subject. This subject-object asymmetry is paralleled by other asymmetrical grammatical constructions in Secwepemctsin. In 3.7 I proposed an analysis to account for the subject-object asymmetry. This analysis reduces the Independent Pronoun Restriction to the One Nominal Interpretation and the third person Subject Cleft phenomenon. I conclude that Secwepemctsin independent pronouns, which are inherently contrastive, show strong subject orientation. Structural asymmetry is central to the 64 analysis: a subject must be structurally different from an object. The Maping Principle needs to recognize the subject position as being higher than the object position so that it can correctly identify the site for the most familiar discourse referent to map onto. These facts argue for a structurally asymmetric view of subjects versus objects, and hence a structurally configurational view of Secwepemctsin. 3.11 Appendix to Chapter 3: Rejecting Other Hypotheses for the Independent Pronoun Asymmetry The analysis offered in Chapter 3 is the best hypothesis to account for the IPR. Two alternative hypotheses are discussed here: A. The asymmetry is due to the difference between ergative and absolutive arguments. B. The asymmetry is due to thematic differences. I will refute both hypotheses, and conclude that the asymmetry is due to the structural asymmetry found between subjects and object. 2.4.1 The Ergative/Absolutive Difference Given that Secwepemctsin is a morphologically split ergative language in third person contexts, the subject of intransitive clauses and the object of transitive clauses share the absolutive case, while only the subject of transitive clauses are ergative. Ergative arguments are marked by subject suffixes, absolutive arguments in transitive clauses are marked by object suffixes, and absolutive arguments in intransitive clauses are marked by clitics (see appendix C at the end of the thesis). If the subject-object asymmetry of independent pronouns were due to case - that is, if independent pronouns could only replace ergative arguments - then it would be impossible for the subject of intransitive clauses to be replaced by independent 65 pronouns. This is also true vice versa; if independent pronouns could only replace absolutive arguments, then both the subject of an intransitive clause and object of a transitive clause should be able to take independent pronouns. 80 wik-t-0-s re Mary re John see-tr-3sg.obj-3sg.subj det Mary det John Mary saw John / John saw Mary 81 wik-t-0-s re John newi7-s see-tr-3sg.obj-3sg.subj det John 3sg.indpr SHE saw John / * John saw HER 82 a setsinem-0 newi7-s sing-3sg.ind 3sg.indpr It is HER that sings. b newi7-s setsinem-0 3sg.indpr sing-3sg.ind It is HER that sings. 83 a setsinem-0 newi7-s re Mary sing-3sg.ind 3sg.indpr det Mary It is MARY that sings. b newi7-s re Mary setsinem-0 3sg.indpr det Mary sing-3sg.ind It is MARY that sings. (80) shows a transitive sentence with two overt DP arguments. As explained previously, the order of the subject and the object is free post-predicatively, giving the sentence two interpretations. In (81), with the presence of the independent pronoun, there is only one interpretation available: that which has the independent pronoun newi7-s as the ergative argument. Note that the independent pronoun is the ergative argument. (82) and (83) show independent pronouns occurring as the absolutive argument of intransitive clauses. (81, 82a, b) are examples of independent pronouns occurring on their own, and (83a, b) are examples of independent pronouns occurring with an overt referent. These independent pronouns are all 66 associated with the subject position, and since the subject in these sentences is not always associated with a specific case, the ergative/absolutive distinction cannot account for the independent pronoun asymmetry. 2.4.2 Aspectual and Thematic differences Suppose the EPR of Secwepemctsin arose from the thematic role of the argument associated. In that case, we would expect independent pronouns to be assigned a certain thematic role, or to be used only with a certain type of event structure. These possibilities can be ruled out, however, based on the wide range of thematic and event structures that can be associated with independent pronouns, as shown in the following data. The following two sentences are agentive eventive clauses; independent pronouns are well-formed with agentive arguments and eventive structure. 84 a ts'um-qs-en-0-s newi7-s re Mary kiss-ls-fc-3sg.obj-3sg.subj 3sg.indpr re Mary HE kissed Mary / *Mary kissed HTM b newi7-s illen-0-s re s.peq.peq 3sg.indpr eat-3sg.obj-3sg.subj det berry(redup) HE ate berries The following are agentive stative clauses; independent pronouns are well-formed with agentive arguments and stative structures. 85 a newi7-s ec ts'kwenst-0-es-es re st'eqwmeke7 3sg.indpr prog hold-3sg.obj-3sg.subj-3sg.conj det ball HE is holding a ball 67 b newi7-s musen-0-s re skiiyes 3sg.indpr feel-3sg.obj-3sg.subj det child HE feels (his own) child Independent pronouns are well-formed with non-agentive arguments and eventive structures. These are non-agentive eventive clauses: 86 a ciyucwt-0 newi7-s re Mary le m temtew's-es 3sg.indpr det Mary det past riding.horseback-3sg.conj MARY fell off a horse b melcen-t-0-em newi7-s kick-tr-3sg.obj-pas 3sg.indpr .HE got kicked c m qwetseq-0 newi7-s past die-3sg.ind 3sg.indpr HE dies Independent pronouns are well-formed with non-agentive arguments and stative structures. These are non-agentive stative clauses: 87 a lecus-k 7-enwi7 pretty-2sg.ind 2sg.indpr It is YOU who is pretty b 7-enwi7 lecus-k 2sg.indpr pretty-2sg.ind It is YOU who is pretty 88 a setsinem-0 newi7-s sing-3sg.ind 3 sg.indpr It is HER that sings. b newi7-s setsinem-0 3sg.indpr sing-3sg.ind It is HER that sings. 89 ts'ellt-0 newi7-s re sk'wimemelt cold-3sg.ind • 3sg.indpr det child THE CHILD is cold 68 In all cases^ , the behaviour of independent pronouns is identical despite the thematic and aspectual differences. These data suggest that the occurrence of independent pronouns is not related to whether they bear a certain thematic role, or whether they are associated with a certain event structure. Agentive versus non-agentive clauses show that volition does not affect the usage of independent pronouns; independent pronouns are well-formed with predicates like kiss, eat, being pretty, or falling off a horse. Stative versus eventive clauses show that whether a predicate is eventive or not is of no consequence when it comes to independent pronoun usage; again, independent pronouns are well-formed whether they are involved in holding something, being cold, getting kicked, or dying. Independent pronouns are not affected by aspectual or thematic factors. As a result, the possibility that thematic roles or event structure as reasons for the asymmetrical behavior of Secwepemctsin independent pronouns can be rejected. > 69 CHAPTER 4 . TOWARDS A SEMANTIC ACCOUNT 4.0 Introduction In this chapter, I show that independent pronouns in Secwepemctsin are inherently contrastive. I illustrate the derivation of contrastiveness via formal semantics, and conclude that alternative semantics (Rooth 1992, Biiring 1995) is slightly inadequate for the Secwepemctsin data on focus and topic. 4.1 Basic Characteristics 4.1.1 Independent Pronoun * Out of the Blue Secwepemctsin independent pronouns cannot be used in out-of-the-blue contexts, just as pro. It is only possible to use an independent pronoun when the discourse referent is apparent to both the speaker and hearer of a dialogue (usually through a linguistic antecedent in the discourse). (1) below, uttered out of the blue, is considered ill-formed. All the grammatical utterances with independent pronouns used in this thesis are utterances elicited in the appropriate context36. 36Independent pronouns always refer to humans. They cannot refer to inanimates or animals even when they are forced to occupy the subject position along with a non-human overt DP. This is consistent with Cardinaletti and Starke's (1996) claim that when a language has more than one set of pronouns, one of the sets always has a more restricted usage. Languages with more than one set of pronouns distinguish between strong and weak pronouns. The set of pronouns with the limited usage is always the strong set. In Secwepemctsin, the independent pronouns correspond to the strong pronouns, and have to refer to [+human] individuals, (i) re melemst'ye illen-o-s re ts'i det wolf eat-3sg.obj-3sg.subj det deer The wolf ate the deer (ii) *newi7-s illen-e-s re ts'i *3sg.indpr eat-3sg.obj-3sg.subj det deer *IT ate the deer (iii) *newi7-s re melemst'ye illen-o-s re ts'i *3sg.indpr det wolf eat-3sg.obj-3sg.subj det deer *THE WOLF ate the deer 70 1 newi7-s qwetset-0 3sg.indpr leave-3sg.ind *HE left (in out of the blue context) 4.1.2 Independent Pronoun * Exhaustive There is preliminary data suggesting that independent pronouns in Secwepemctsin are not specified for exhaustivity. That is, although contrastive emphasis is present, as in 'MARY but not Sue, Jill, or Nancy, kissed John', it is not necessarily the case that only Mary kissed John. To express exhaustivity, a separate morpheme is used (2). 2 tsukw newi7-s re John wik-t-0-s re Elvis only 3sg.indpr det John see-tr-3sg.obj-3sg.subj det Elvis Only JOHN saw Elvis The Secwepemctsin equivalent of the minimal pair in (3) would be evidence for the claim that independent pronouns indeed are not specified for contrastiveness. 3 a JOHN saw Elvis, and Peter did too. b Only JOHN saw Elvis, *and Peter did too. 4.1.4 Independent Pronoun * Sentential Adverb Although Secwepemctsin independent pronouns have been analyzed as adjuncts that modify DP, they are not to be confused with another type of nominal modifier, the English -self adjuncts, as in (4a). English -self can be used in DP-modifying constructions, as in the contrastive constructions studied in this thesis, but it can also be used adverbially, as in (4b) (Edmondson and Plank 1980). 71 4 a b The King himself [not cook, servant, or his mother] baked a cake. The King baked a cake by himself. (Edmondson and Plank 1980) This is not so in Secwepemctsin. To express an adverbial reading such as that in (4b), a lexical item meaning 'by oneself or 'alone' is used. 5 neqwestsu.ts.t-ken te re newi7-s re kukwpi7 alone(redup)-lsg.ind obi kill(redup)-3sg.obj-lsg.subj det3sg.indpr det chief I killed THE CHIEF by myself 4.1.4 Independent Pronouns = Contrastive Independent pronouns in Secwepemctsin are used to emphasize the reference to a particular nominal in the discourse context. For an utterance like (6), the implication is that it is MARY, and not Sue, Jill, Nancy, or anybody else, that kissed John; that is, MARY is contrasted with all possible kissers of John. 4 [newi7-s re Mary] ts'um'-qs-en-0-s re John [3sg.indpr det Mary] lick-nose-fc-3sg.obj-3sg.subj det John MARY (not Sue, Jill, or Nancy) kissed John It is precisely this kind of contrastive implication that induces the capitalized nominal in the English translation, which is pronounced with contrastive stress. To illustrate this point, consider the following two dialogues. Dialogue 1 A: I heard that the lovebirds were really enjoying themselves at the Christmas party. B: Yeah, Mary kissed John under the mistletoe as soon as they arrived, and they stayed there during the entire party. Dialogue 2 A: I heard that John was taken under the mistletoe. Did his girlfriend kiss him? B: No, MARY kissed him/John. 72 In dialogue 1, the underlined sentence is uttered without any special stress. In dialogue 2, presumably Mary is not John's girlfriend, so MARY is pronounced with contrastive stress. This is the kind of emphasis that an independent pronoun provides to its nominal referent. If appears that the contrasted set has to be available for the discourse participants somehow, but it is not clear whether they have to be identified explicitly in the prior discourse. The function of contrastive stress in English appears to be taken over by independent pronouns in Secwepemctsin. Hence, when an independent pronoun co-occurs with its overt referent in Secwepemctsin, the construction is equivalent to an English stressed DP (see English translation of (6) above). When an independent pronoun co-occurs with a covert DP, i.e., pro, the construction is equivalent to an English stressed pronoun (10). 10 newi7-s setsinem-0 3sg.indpr sing-3sg.ind SHE sings Essentially, independent pronouns in Secwepemctsin do what contrastive stresses in English do37. They induce a set of alternatives for the independent pronoun to be contrasted with38. When an independent pronoun is in predicate position, it is providing contrastiveness to a focus, since predicative independent pronouns occupy wh/focus position. This contrastive focus is different from presentational focus. Contrastive focus is the focus of a single item within a proposition, while presentational focus is focus of an entire proposition. Presentational focus will not be dealt with in this thesis39. 37Note that if independent pronouns were the exact equivalent of English contrastive stress, they should freely associate with anything, and should not be limited to DP-associations only. However, my intuition is that Secwepemctsin independent pronouns can only be associated with noun-like items. The observation that independent pronouns can only be associated with [+human], but not [-human] individuals, partially supports this intuition. 38Note the absence of focus-related stress in Salish languages ODavis, p.c.) 3 9For details on presentational focus in Salish, see Demirdache 1998. 73 When an independent pronoun is associated with an argument DP, it is providing contrastiveness to a topic. Note that even a contrastive lexical DP can be topical because it must be available in the discourse to be contrasted40. The difference between an independent pronoun-modified pro DP and an independent pronoun-modified lexical DP is that the former is an anaphoric contrastive topic while the latter is a non-anaphoric contrastive topic. The contrastive function of independent pronouns can be summarized as follows. A predicative independent pronoun is a contrastive focus; a pro DP modified by an independent pronoun is an anaphoric contrastive topic, and a lexical DP modified by an independent pronoun is a non-anaphoric contrastive topic. 4.2 Deriving Contrastiveness How can the 'contrast' in 'contrastive focus' and 'contrastive topic' be represented formally? One of the ways to derive this contrastiveness is to apply the alternative semantics (Rooth 1992) view adopted in Buring (1995, 1998). In Buring (1995), focus is defined as the word in a proposition corresponding to a wh word in a question; it is accented, and "marks what is new or unexpected in a sentence" (Buring 1995:21). (11), which has BASEball accented, "draws attention to the fact that it is the baseball (as opposed to - say - the vase) that John threw" (Buring 1995:21). II41 a John threw [the BASEballJF (Buring 1995:21) ^It may also be that lexical DPs are contrastive topics, and not foci, because contrastive DPs are associated with subjects. Subjects are generally associated with topics. Also, Gardiner (1998) has shown that all preverbal positions in Secwepemctsin are topic positions, and there is evidence that in Secwepemctsin's sister language, St'at'imcets, out-of-the-blue preverbal subjects are impossible (Davis, p.c). Therefore, the preverbal lexical DP subject has to be a topic. A full investigation of how focus and topic are used in Salish languages is necessary to explain these observations. 4 1The capitalized letters in BASEball indicate the 'accent', while the subscript F marks BASEball as the 74 Thus, in a regular assertion A without any contrastive element, no alternative is necessary for any element in the assertion (12). However, when the focus element of a sentence has A-accent, a semantic value is induced, where the contrastive focus element corresponds to a wh word. This semantic value (shown in italics) is the focus assertion A f (13). 12 A Mary saw him 1342 A Mary saw JOHNA A f Who did Mary see? The contrastiveness of JOHNA is available from the alternatives that can answer the wh word in A f in (13). On the other hand, when a topical element of a sentence is contrastive, a set of topic assertions A' are induced, in addition to the A f. Birring's topics also contain an accent; this is the B-accent. The A' is a set of questions formed by taking the topical item (the item containing the B-accent) in the A f and replacing it with a set of alternatives of the same type. 14 A MaryA saw JOHNB A f Who saw JOHN? A 1 Who saw John? Who saw Bill? Who saw Fred? What we have so far is laid out below. 15 Contrastive focus versus contrastive topic Contrastive focus: A f Contrastive topic: A* English A-accent (focus accent) B-accent (topic accent) Secwepemctsin Predicate independent pronoun DP/pro independent pronoun focused constituent. I take what Birring refers to as the 'accent' on focus to be A-accent (focus accent). Focus is marked by subscript A, for A-accent, while contrastiveness is marked by capital letters, as in the rest of the thesis. I will later mark Topic with a subscript B, for B-accent. 75 A regular assertion without any contrastive element needs no alternative, therefore no A f, and definitely no A1. An assertion with an A-accented focus (or a focus (predicate) independent pronoun) gets an A f semantic value. An assertion with a B-accented topic (or a topics (DP/pro) independent pronoun) gets an A1 semantic value. However, to arrive at A1, there must first be an A f. This is the problem. To derive contrastive focus in Secwepemctsin, A f is formed from A (16). newi7-s re JOHN is the contrastive focus (equivalent of B-accent), inducing an A f, which contains a set of alternatives that can answer 'who'. 16 A It's [newi7-s re J O H N T J A that Mary saw A f Who did Mary see? To derive contrastive topic in Secwepemctsin, consider A in (17). [newi7-s re MARY] is topical, and is the equivalent of Buring's B-accent, so it should ultimately get an A 1 set of values. A1 is only possible if an A f is first available; A f is available for the A in (17) only if John has the equivalent of Buring's B-accent. Suppose this were true, and John had contrastive focus, then the A f and A' in (18) would be derived, and the alternatives necessary for the contrastive topic [newi7-s re MARY] would be derived. 17 A [newi7-s re MARY]B saw John 18 A f [newi7-s re MARY] saw who? A1 [newi7-s re Mary] saw who ? Sue saw who? Jill saw who? Beth saw who? The problem with (18) is that John in (17) is not a contrastive focus (at least it is not a focused independent pronoun). John in (17) is at most a simple focus, and therefore not an 76 AMnducing element. Without A f, there is no A' in (18). This problem can be rephrased in the following way. In order to have A f, we need a contrastive focus. In order to have A1, we need a contrastive topic plus A f. Hence, in order to have A', we need a sentence with a contrastive topic plus a contrastive focus (a sentence with a focus independent pronoun and a topic independent pronoun). There are two ways to solve this problem. The first possible solution is as follows. If any focus (contrastive and non-contrastive) can induce A f, then John in (17) can easily induce A f in (18), and in turn this A f can induce A 1 in (18), which yields the desired set of alternatives for the contrastive topic [newi7-s re MARY]. This solution is inadequate, since it would lose the distinction between a contrastive focus versus a non-contrastive focus, as in (19)43. 19 a I saw the boy b I saw THE BOY The second possible solution is as follows. Although John in (17) is by itself not a contrastive focus, it gains contrastiveness automatically by virtue of being in the same sentence as the contrastive topic [newi7-s re MARY]. This inherited contrastiveness would successfully derive A f and A' in (18), yielding the set of alternatives for the contrastive topic. Claiming that John gains contrastiveness from [newi7-s re MARY] predicts extremely high discourse sensitivity for this type of sentences; it means that a sentence having a contrastive subject automatically makes the object contrastive as well44. An investigation of the issues involved is beyond the scope of this thesis; further research needs to find out precisely what kind of discourse contexts these sentences are used in. 4 3This is precisely the kind of problem that alternative semantics has been criticized for: it cannot distinguish between contrastive focus and presentational focus (Rochemont, p.c). •"Note that in English, the only kind of focus a subject can get is contrastive focus (Davis, p.c). On the other hand, what we need to derive Secwepemctsin contrastive topic is to claim that a Secwepemctsin object is always a contrastive focus if the subject of the clause is a contrastive topic. 77 4.3 Conclusion In Chapter 4, I have determined that Secwepemctsin independent pronouns are contrastive foci when in predicate position, and contrastive topics when modifying DPs in argument position. I have shown that a problem is encountered when using alternative semantics (Rooth 1992, Buring 1995) to derive the contrastiveness. This problem is a general problem that formal semantics has in dealing with the notion of focus. The Secwepemctsin independent pronoun data shows that more intricate tools are necessary to account for them. 78 CHAPTER 5 . ACQUISITION: A CASE STUDY 5.0 Introduction Independent pronouns in Secwepemctsin have been shown to observe strict constraints regarding discourse familiarity and contrastiveness. Their behaviour follows from their syntactic category and position. In Chapter 5,1 will ultimately illustrate that a new speaker of Secwepemctsin does not adhere to the grammatical restrictions of independent pronouns at this stage in her acquisition, and instead uses these pronouns differently from fluent speakers. The differences'follow from: A. an internal categorial reanalysis she has applied to independent pronouns; B. an external syntactic reanalysis she has applied to the site of occurrence of independent pronouns, so that they function like R-expressions. Chapter 5 first gives a general view of the Secwepemctsin language production by a seven-year-old girl who has been exposed to both English and Secwepemctsin. The child's linguistic background, the elicitation process, and the transcription and coding methodology are described in 5.1. In section 5.2, results and analyses* including the MLU of the child's language samples, will be given. A section follows with a general description of the child's Secwepemctsin phrase structure rules and vocabulary, and, finally, section 5.3 is a presentation on the child's use of Secwepemctsin independent pronouns. 5.1 Preliminaries 5.1.1 Background: The Child Subject and The Family Context 79 The child subject, Julienne Ignace45, was born in November, 1990 to a household where both parents speak Secwepemctsin fluently. Her father is a native speaker of Secwepemctsin. He grew up in a multi-generational household, and was raised mainly by his great-grandmother, who was about 60 years old when he was born, and did not tolerate any English from young people. Julienne's father lived in this environment until his mid-teens, at which time he entered residential school for a year, then went to work outside his community. When he was away from his community and had no one to converse with in Secwepemctsin, he tried to keep the language alive by- thinking to himself in Secwepemctsin. Julienne's mother is an anthropologist and linguist from Germany. In addition to her native Dutch-related dialect and German, she learned English, French, and Latin in her childhood. She moved to North America to do research with the Haida people in her 20s, and proceeded to work with the Secwepemc people in 1984. She needed to learn Secwepemctsin to interview elders, record and transcribe narrative stories, and plan language lessons for new speakers of Secwepemctsin. Even before Julienne's mother was completely comfortable with Secwepemctsin, her experience learning and working with languages allowed her to have native-like pronunciation. By 1990 she was able to carry on normal adult conversation in, Secwepemctsin. Julienne has two older sisters, age 9 and 5, and two older brothers, age 9 and 3, at the time of her birth. All four of Julienne's older siblings speak English, and understand only a few Secwepemctsin words. From the time Julienne was born and before she entered daycare when she was 10 months old, Julienne was taken care of exclusively by her mother, who speaks to her only in Secwepemctsin. She was also away from her older siblings at that time. Julienne spent 4 days a week at an English-speaking private daycare starting at age 10 months, for 5 or 6 hours per day. Outside the daycare, she continued to be taken care of and spoken to by her mother 'Julienne is adapted from French to Secwepemctsin, and is pronounced [suli?aen]. 80 virtually at all times. Her father also always speaks to Julienne in Secwepemctsin, but his role in child-rearing is less than that of the mother. Since Julienne's first intelligible words (all Secwepemctsin) emerged at about 1;1 or 1;2, her mother started keeping a record of her language learning. Spontaneous responses in Secwepemctsin were the norm during this period, as revealed by the early tape recordings, logs and notes46. According to this journal, her learning pattern appears to be no different from what is generally known about infant language acquisition (See Lai & Ignace 1998). Her comprehension was more advanced than her performance. She truncated longer words into mono- or bi-syllabic utterances, she reduplicated some mono-syllabic words, and there were cases of overgeneralizations, such as using the word qexe (her simplified form of nts'sqexe 'horse' or sqexe 'dog') for all four-legged animals except cats. At age 2 years 6 months (JI's age will subsequently be noted as year semicolon month: 2;6 or year semicolon month semicolon day 2;6;15), when Julienne's younger sister was born, Julienne started to spend time with her English-speaking older siblings instead of her mother, who had to take care of the new baby. A younger brother was born when Julienne was about 5 years old. Although at this time she had gotten used to using English with her older siblings, she also became aware that her parents only speak in Secwepemctsin to her two younger siblings as they do to her. Her passive Secwepemctsin vocabulary continues to expand, and she begins to use mixed English/Secwepemctsin utterances. Some typical ones of this period are in (1-3). 1 I wanna kwen-t-e 7ene I wanna take-tr-2sg.imp this I wanna take this Correct form: me7 kwekwen yi7ene ^Subsequent data referenced with ages that are before 6 are all cited from the records kept by Julienne's mother. 81 2 I kwe.kwen oats I take(redup) oats / take oats Correct form: kwekwen re "oats" 3 I'm meq' I'm full I'm full Correct form: memq'-ken English became increasingly predominant as Julienne entered preschool and elementary school. Although she continued to hear Secwepemctsin exclusively from both her father and mother, and occasionally uses it with her younger siblings, she would only reply in Secwepemctsin when prompted. At this time, Julienne also started to translate her parents' commands in the form of questions, which showed her comprehension, such as "Why do you want me to get the cup?" Such translations show that in comprehension of Secwepemc speech Julienne can distinguish the full range of intransitive pronominal clitics, transitive subject and object suffixes with accuracy47, but the speech production patterns show already a tendency to use stems without the pronominal morphology on the predicate, despite the caregivers' continued modelling of accurate speech in repetitions and prompting. The replacement of correct suffixes or clitics by second singular imperative forms such as that in (1) can be explained by the fact that in caregiver speech, imperative forms were likely the most frequently used and heard forms, and thus became parts of active vocabulary, although used incorrectly. At a later era, which carried over into the period of acquisition research for this thesis, a feature of Julienne's Secwepemctsin speech became the use of independent pronominal forms, in at least simple sentences making requests and statements, sometimes in addition to the For examples, see utterances 14, 20-29, 220, etc, which are all Julienne's English translations of her mother's Secwepemctsin speech. 82 pronominal clitics and suffixes. This phenomenon will be dealt with in the latter sections of Chapter 5. In summary, up until age 6, Julienne's active use of Secwepemctsin decreased as her exposure to English increased. Although her parents continued to model correct Secwepemctsin with her and her younger siblings, her Secwepemctsin utterances began to incorporate English stems, and she began to use Secwepemctsin independent pronouns. 5.1.2 Collection and Methodology Collection of Julienne's Secwepemctsin data for this case study began early in 1997, when Julienne was 6;3, at which time a small sample of 18 utterances were taken. More systematic visits were made between September 1997, when Julienne was 6;10, and March 1998, age 7;4. During this seven-month period, I visited Julienne's home on average twice a month for 12 visits, resulting in a total of 7 recorded sessions. 4 Utterances recorded Age at session 6;3 6;10;15 6;10;25 6;11;16 7;0;6 7;2;28 7;4;6 Total Utterances 18 27 69 63 11 107 10 305 Due to the relatively small size of samples from each session, the data from all sessions have been collapsed into one database, as if they were collected from the same stage of acquisition, and there will be no discussion in this thesis of whether or not there was any advancement in Julienne's Secwepemctsin learning during the collection period. For the sample collection sessions, Julienne was told that I wanted to learn Secwepemctsin, therefore Julienne should speak in Secwepemctsin while playing with me, so that I could learn 83 it. Only myself, the mother, and Julienne were present at these elicitation sessions. Due to my lack of fluency in the language, I only used limited Secwepemctsin words with Julienne, and communicated mostly in English with Julienne in and out of the sessions. Julienne's mother's use of Secwepemctsin with her was consistent. The sessions started in September 1997 were mostly videotaped by either myself or the mother, and written notes were added after the sessions had finished. When a session was not videotaped, written notes were taken. Depending on the situation and Julienne's mood, each session ranged from 30 minutes to 60 minutes. Early on in the collection process, it was noticed that Julienne did not respond well to one type of elicitation process. This was expected of a child, and many different ways of elicitation were used throughout the data-collection period. They included imitation, in which Julienne repeated after her mother; a 'Simon says' game in Secwepemctsin, in which Julienne, her mother, and myself took turns issuing commands; observation of Julienne at play with her mother, and spontaneous conversations and games between the child and the mother. In each visit to Julienne's home, I spent approximately 7 to 10 hours in the presence of Julienne in addition to the actual elicitation session. It is estimated based on this experience at her home that Julienne uses English to spontaneously communicate with her older siblings all the time, while she uses English to spontaneously communicate with her parents and younger siblings about 70 to 80% of the time. It is therefore assumed that Julienne is much more comfortable and much more fluent in English than in Secwepemctsin. Based on Julienne's performance in English and Secwepemctsin at the end of the research period, there is little doubt that Julienne is now most fluent in English. 5.1.3 Data Transcription and Coding 84 A total of 305 intelligible utterances (see (4) above), including English, Secwepemctsin, and mixed utterances, were transcribed from the elicitations and entered into a database with the following fields: utterance number utterance language (Shuswap, English, or mixed) complete or incomplete number of morphemes per utterance yes/no or non-yes/no imitation or non-imitation pronominal reference (yes or no) non-independent pronouns used (enter each pronoun) independent pronouns used (enter each independent pronoun; if none, enter no) • date collected additional notes A brief explanation of each of the fields is given below. Utterance Each utterance is recorded in English or Secwepemctsin orthography. An utterance is determined by a continuous string produced by Julienne. If she pauses, and does not continue, the output string is considered broken, and the end of the utterance is reached. However, interruption of her speech which resumes immediately do not constitute the end of an utterance. For instance, (5a, b) each shows an utterance. 5 a Yes you can. It means I see him, hmm, which one? Was it a girl or a boy? (7;2;28)48 b I feed, me-, mete.t-0-en re n-sqexe (7;2;28) I feed, fee-, feed(redup)-3sg.obj-lsg.subj det lsg.poss-dog /feed, fee-, I feed my dog Each of Julienne's utterance is given with her age at time of utterance. If an utterance is referred to without being given m the main text, the utterance number is cited, and it can be looked up in the appendix, where all of Julienne's utterances are given. 85 Language Utterances that are completely in Secwepemctsin or English are coded as so. Utterances are "mixed" when using morphemes from both languages. Words or expressions that cannot be easily categorized as English or Secwepemctsin are also coded as "mixed"; these include expressions like "hey", "huh" and proper names. However, if a name or expression that is unidentifiable as a particular language appears in a longer utterance that is completely Secwepemctsin or English, then the unidentifiable word is ignored, and the rest of the utterance is classified accordingly. For example, utterance (6) is classified as "mixed", while (7) is "Secwepemctsin ". 6 Lizzie (6;3) 7 mama, kec-t-se.ts.m-e t k cereal (6;3) mom, give-tr-lsg.subj.redup-imp obi irr cereal Mom, give me some cereal Completeness Utterances that are holistic in nature are generally complete; for instance, if the child says "yes" in response to a yes/no question, "yes" is considered a complete utterance. Incomplete utterances consist of false starts and completions of previous utterances. False starts are utterances that do not finish (8). 8 ntsetswe7, uh, me-, me- (7;2;28) lsg.indpr, uh, fee-, fee-I, uh, fee-, fee- (Julienne is trying to remember the word for 'I feed', meteten) Completion of previous utterances include words or syllables that complete the child's own previous utterances or the mother's utterances. A typical completion case is given in (9). 86 Q49 J. setsinem-s Sandra (7;2;28) sing-3sg.subj Sandra Sandra sings. M: setsinem-sing-Sing-? J: -ce Sandra (7;2;28) -2sg.imp Sandra You (sing)! Sandra! Number of Morphemes per Utterance The same set of rules to calculate the number of morphemes per utterance is used for English, Secwepemctsin, and mixed utterances. The rules are as follows50: Since the notion "word" is not easily definable in an agglutinative language like Secwepemctsin, and Secwepemctsin has too many lexical items that are not free morphemes (lexical suffixes, for example), only morphemes will be counted. Any item with an identifiable independent meaning is counted as one morpheme. This includes clitics, affixes, and bound or free lexical items. Morphemes in exclamations are counted as one morpheme. "Huh?" may be used to mean "What did you say?" Morphemes in full and partial imitation utterances are counted. Morphemes in incomplete utterances are counted. • Morphemes have to be complete morphemes in order to be counted; i.e., an incomplete morpheme, though recognizable, is counted as 0; for example, (10) is recognizable as speqwelt 'cw "book", but it is counted as 0 morphemes because it is incomplete. 10 -qwel'tcw (7;2;28) -book book (speqweTtcw = 'book') Yes/No or Non- Yes/No Any utterance containing English or Secwepemctsin "yes" or "no" is classified as Yes/No. The conversation participants' utterances will be coded either " M " for Julienne's mother, or "S" for myself. Productivity of these morphemes is ignored. 87 Imitation or Non-imitation If the utterance repeats a previous utterance by her mother, or by the researcher, it is considered an imitation. These also include partial imitations with expansions on the utterances (11). 11 M: n-tsetswe7 ell-1 sg.indpr and-I, and? J: n-tsetswe7 ell tsetse (7;2;28) lsg.indpr and younger, sibling, of. same, sex I and younger sister Pronominal Reference An utterance is coded "yes" if it has any kind of morpheme encoding pronominal reference, "no" if it has no morphemes encoding pronominal reference. Bound Pronouns Used Each pronoun used is entered. English pronouns include "I, you, he, she, it, we, they, my, your, etc." and Secwepemctsin pronouns are entered as gloss (e.g.: Isg.ind for 'first person singular indicative') because they are all clitics or suffixes and may change their form depending on the environment. 88 Independent Pronouns Used Independent pronouns are entered in this field if they occur in the utterance. If there are no independent pronouns in the utterance, "no" is entered. Additional Notes Relevant additional notes are entered here, such as context, gestures, etc. 5.2 General Analysis 5.2.1 Results From a total of 305 utterances by Julienne, 60.98% (186 utterances) were in Secwepemctsin, 29.83% (91 utterances) were in English, and 9.18% (28 utterances) were mixed. Within each language category, there are significantly more complete utterances than incomplete utterances. (12) shows detailed numbers of utterances in each language. 12 Break-down of Julienne's Sample by Language Secwepemctsin English Mixed Total 186 91 28 305 60.98% 29.83% 9.18% 100% 13 Break-down of Julienne's Sample by Completeness Complete Incomplete Total 230 75 305 75.41% 24.59% 100% 89 Although the large percentage of complete Secwepemctsin utterances is encouraging, the Secwepemctsin incomplete utterances comprise 86.67% of all incomplete utterances. This shows that Julienne finds it more difficult to start and complete utterances in Secwepemctsin than in English or mixed language situations. In comparison, there are only 5 cases (6.67%) of incomplete utterances in English, and 5 such cases in mixed language utterances. 14 Julienne's incomplete utterances Secwepemctsin Incomplete: 86.67% (65) False starts: 31 Completions: 34 English Incomplete: 6.67% (5) False starts: 2 Completions: 3 Mixed Incomplete: 6.67% (5) False starts: 2 Completions: 3 Virtually all of Julienne's imitations are in Secwepemctsin (27 in Secwepemctsin, 0 in English, 2 in mixed languages). The lack of imitation in all 91 English utterances shows her comfort with English, since she can produce spontaneous utterances on her own. Moreover, assuming that imitation is a learning strategy for the language-acquiring child, the asymmetry between Secwepemctsin and English or mixed-language imitation indicates Julienne's attempt to improve her Secwepemctsin skills. The MLU table in (15) shows that Julienne's acquisition of English is far ahead of her acquisition in Secwepemctsin in terms of mean length of utterance (MLU). 90 15 MLU in Secwepemctsin, English, and mixed languages Secwepemctsin English Mixed languages Length of Number of utterance utterances (morphemes) Length of utterance (morphemes) Number of utterances Length of utterance (morphemes) Number of utterances 0 2 0 0 0 0 1 69 1 12 1 8 2 27 2 6 2 3 3 28 3 12 3 4 4 15 4 10 4 2 5 16 • 5 . 8 5 1 6 11 6 9 " 6 2 7 8 7. 9 7 0 8 2 ' 8 3 8 1 9 5 9 5 9 0 10 3 10 5 10 3 11 0 11 2 11 1 12 0 12 2 12 1 13 0 13 2 13 1 14 0 14 1 14 0 15 and more 0 15 and more 5 15 and more 1 Total number of morphemes = 560 =3.01 186 Total number of morphemes = 552 =6.07 91 Total number of morphemes Total number of utterances = 140 =5 28 Total number of utterances Total number of utterances The longest Secwepemctsin utterance in the database is recorded at 10 morphemes per utterance, whereas the longest English utterance is 25 morphemes long, and the longest mixed language utterance is 15 morphemes long. The mean length of utterances (MLUs) for Secwepemctsin, English, and mixed languages are 3.01, 6.07, and 5 respectively. There is a big gap between the MLU for Secwepemctsin and the MLU for English, confirming that Julienne's performance in English is far more advanced than that of Secwepemctsin. Furthermore, the chart in (16) shows that it is in Secwepemctsin where there is a peak at one morpheme and a sharp decline in number of occurrences as the length of utterance increases. This is an indication that it is more difficult for Julienne to have longer utterances in Secwepemctsin because she is less fluent in this language. 91 16 Chart of Julienne's MLU JI's MLU 80 | Number of morphemes There is something to be said about the one-morpheme utterances in Secwepemctsin. Of all 69 occurrences, 56.52% (39/69) are either me7e 'yes' or tala 'no', used in response to Julienne's mother and the researcher's elicitations51. Julienne's passive role in the use of Secwepemctsin is thus manifested. Other single morpheme utterances are evenly distributed between numerals, single nominals, predicate stems without pronominal markings (which are typical of Julienne's predicates), and free use of single bound morphemes, used primarily as completions to previous utterances. 51Note that the elicitations are not designed to deliberately solicit yes/no answers. Julienne's yes/no answers are used correctly, although the contexts in which they are used generally need further elaborations to the yes/no words. 92 Julienne uses yes/no answers very frequently. There are 43 utterances with yes/no word occurrences, 39 of which are single-morpheme utterances; in other words, only 9.3% (4/43) utterances have elaborations other than the yes/no word. There are 12 yes/no utterances in English, and in this case, Julienne elaborates on 66.6% (8/12) of them. In mixed language utterances, there are 3 yes/no occurrences, and all of them have elaborations on the yes/no response. Therefore, I interpret the absence of elaborations on the Secwepemctsin yes/no utterances as another sign that Julienne is not as comfortable in this language as she is in English. 5.2.2 Secwepemctsin Vocabulary and Phrase Structure There are a total of 91 different morphemes recorded in Julienne's Secwepemctsin utterances, one third of which belong to the nominal category. Note that the classification of adjectivals is approximate, as it is often difficult to establish whether this category is present in Salish languages (Demirdache and Matthewson 1994, Davis, Lai, & Matthewson 1997). 17 Secwepemctsin morphemes found in Julienne's speech Type of morpheme Number of types Nominal (including emphatic pron. stems) 30 Verbal ('action' words) 22 Functional (bound pronouns, determiners, etc) 18 Adjectives (including numerals and deictics) 14 Other (including yes/no and interjections) 7 Total 91 Since Secwepemctsin is characterized as having a wide range of bound grammatical morphemes, a brief look at such morphemes in Julienne's language sample would illustrate the extent to which she has acquired Secwepemctsin. The following are the functional types found in Julienne's vocabulary. 93 18 Functional morphemes in Julienne's sample: 18 -ce imperative marker -en lsg.subj -es/us 3sg.conj k irrealis marker -k 2sg.ind -ken lsg.ind -kt lpl.ind kucw exclusive marker n- lsg.poss re determiner s- nominalizer s(i) 2sg.obj -s 3sg.ind -t- transitivizer te oblique determiner w7ec progressive/aspectual marker 7- 2sg.poss (redup) first person singular reduplication A total of 18 different functional morphemes are found, 11 of which are pronominal morphemes. Compare these to the 35 different pronominal morphemes shown in appendix C, and it becomes clear that Julienne is actively using less than half of the bound pronominal morphemes available in the language. If the entire range of functional morphemes in Secwepemctsin (including causative marker, passive marker, particles, and determiners) are taken into account, the percentage that Julienne actually uses is significantly lowered. An additional point to be considered is that Julienne's database contains imitations as well as spontaneous utterances, and to a certain extent some of her expressions are idiomatically memorized, due to the lack of positive evidence of productivity; as a result, not all of the functional morphemes recorded here are productive, lowering even more the number of actively used functional morphemes. Due to space and time limitations, I will not explore further the productivity of Julienne's morphological vocabulary in this thesis. Instead, I will discuss general characteristics of 94 Julienne's grammar and compare them to the characteristics of the adult grammar as presented in Chapter 1. 5.2.3 General Aspects of Syntax: Adult versus Julienne Radical-Head Marking Language In adult Secwepemctsin, pronominal marking on the predicate is obligatory, since it is a radical head-marking language, but in Julienne's Secwepemctsin, predicates are optionally marked by pronominals. In the two utterances in (19), want is uninflected in one case and inflected in another. This is clearly a violation of the adult grammar, which always requires pronominal clitics or affixes to license pro and to allow single predicate utterances. 19 a n-tsetswe7 qwenmin te dance (7;4;6) lsg.indpr want obi dance I want to dance b ta7 k s-qwenmim-0-en ntsetswe7 te q'imeke7 (7;2;28) negirr nom-want-3sg.obj-lsg.subj lsg.indpr obi pen I don't want the pen Adult Secwepemctsin allows a clause to be composed of only a single predicate, and apparently so does Julienne's Secwepemctsin grammar (20). However, there are few examples for this sentence type. 20 oh, xwist-si-n, ki7ce (7;2;28) oh, like-2sg.obj-lsg.subj mother Oh, I love you, Mom! 95 Optional DPs and Their Order In (21a), the predicate appears to be complete in and of itself, with pronominal subject and object correctly inflected on it. Although this predicate is used with the word for mother, it is quite clear from the pause (signalled by a comma) that mother is not functioning grammatically as an argument of the clause. (21b), on the other hand,, shows that Julienne does employ overt DPs optionally. 21a oh, xwist-si-n, ki7ce (7;2;28) oh, like-2sg.obj-lsg.subj mother Oh, I love you, Mom! b n-tsetswe7 xwe.xwis-t-si-n 7-enwi7 (7;2;28) lsg.indpr like(redup)-tr-2sg.obj-lsg.subj 2sg.indpr I love you. Word orders available in adult Secwepemctsin are SVO, VOS, and VSO, while Julienne's samples show SVO (22) and VSO (23) orders. 22 n-tsetswe7 illen te s-t7illen (7;2;28) lsg.indpr eat obi food I eat food 23 ta7 k s-qwenmi.m.-0-en ntsetswe7 teq'imeke7 (7;2;28) net irr nom-want(redup)-3sg.obj-lsg.subj lsg.indpr obi pen I don't want the pen (Not T don't want my pen') Determiners are obligatory on straight and oblique arguments in adult Secwepemctsin, but optional in Julienne's grammar. Julienne uses the direct determiner re regularly; she has never used the absent determiner /, and has oyergeneralized the oblique determiner te. Compare the inconsistent use of the determiner for the argument of touch in the following utterances. 24 tsut-0 te Simon, tkenstek-e re sq'wext (6;10;5) say-3sg.ind obi Simon, touch-imp det leg Simon says, touch the leg 96 25 tsut- 0 te Simon, tkenste-t-0-en say-3sg.ind obi Simon, touch-tr-3sg.obj-lsg.subj Simon says, touch the belly button te qw'u (6;10;5) obi bellybutton Summary Given the above criteria, Julienne's Secwepemctsin grammar is decidedly different from that of the adult's. It appears that she is using an agglutinative language (Secwepemctsin) as if it were an analytic language (English), evidenced by her optional argument marking on predicates and proliferated use of overt DPs. These observations indicate that her Secwepemctsin grammar is deeply influenced by Secwepemctsin-external forces; a possible analysis arises that she is simply mapping Secwepemctsin lexical items onto an English syntactic template. However, this analysis would have to be rejected, based on the comparison given in the following table. 26 Adult Secwepemctsin Grammar versus Julienne's grammar versus Relexified English52 Adult grammar Julienne's grammar Relexified English Word order SVO, VOS, VSO SVO, VSO SVO Overt DP Optional - rare Optional - frequent Obligatory overt DP Arguments Generally pro Extensive use of overt DP Obligatory DPs Pronominal marking on predicate Obligatory Optional Virtually absent Use of determiners Obligatory on arguments Optional Obligatory Use of independent pronouns Seldom used; for contrastive purposes only Extensive use Argument pronouns obligatory; contrastiveness used in restrictive contexts only Julienne's Secwepemctsin grammar is neither adult Secwepemctsin, nor relexified English. It is different from English and Secwepemctsin in all of the categories presented in table (26). However, in several cases, Julienne's Secwepemctsin grammar appears to be in between 2Given that less than a quarter (24.2%, 45/186) of Julienne's Secwepemctsin utterances are over 5 morphemes long (recall that independent lexical items in this language are rarely singly morphemic), and many of these are incomplete utterances, the usable data drawn from her sample for a syntactic analysis is extremely limited. 97 English and Secwepemctsin. For instance, while Secwepemctsin overt DPs are always rare, and English overt DPs are always obligatory, Julienne's Secwepemctsin overt DPs are optional but very frequent. While pronominal marking on predicates is obligatory in Secwepemctsin and virtually absent in English, it is optional in Julienne's Secwepemctsin predicates. In section 5.3,1 will single out the special case of independent pronouns, and examine how Julienne uses them in comparison to how adults use them. The following diagram supplements the table in (26) in illustrating that Julienne's grammar characterizes both English and Secwepemctsin, but also has innovative properties that are neither English nor Secwepemctsin. 27 Overlapping of grammars Adult Julienne's Secwepemctsin English Secwepemctsin (27) illustrates that Julienne's Secwepemctsin grammar contains properties of English, Secwepemctsin, and also properties that are non-English and non-Secwepemctsin. It is therefore inappropriate to conclude that Julienne's grammar is only ifluenced by the languages that are used in her environment. The Secwepemctsin-external and English-external forces that affect her Secwepemctsin performance could possibly come from Universal Grammar. It would therefore be helpful to see what monolingual children acquiring languages similar to Secwepemctsin do, which would help explain how much of Julienne's grammar is coming from non-Secwepemctsin, non-English sources. More discussion on this subject will follow the study of Julienne's Secwepemctsin independent pronouns. 98 5.3 A Special Case: The Comparison of Pronominal Systems in Adult Secwepemctsin and Julienne's Secwepemctsin Julienne's use of independent pronouns is extremely frequent. Nearly half (46.6%, 48/103) of , all her utterances with pronominal reference employ independent pronouns. While adult speakers only use such pronouns as arguments in contrastive contexts, Julienne does not adhere to this restriction. This section examines the occurrences of independent pronouns in Julienne's language sample, and compares them to use by adults. 5.3.1 Overregularization of Independent Pronouns: Data Extensive Occurrence Of the 7 independent pronouns available in Secwepemctsin, (see table in (28)) Julienne has only used the three singular forms. Their number of occurrences are recorded in (29). 28 Independent pronominal paradigm (adopted from Kuipers 1974) 1 sg. 2sg. 3 sg. n-tsetswe7 7-enwi7 newi7-s 1 pi. incl. 1 pi. excl. 2 pi. 3 pi. wll-enwi7-kt wll-enwi7-s-kucw wll-enwi7-mp wll-enwi7-s Julienne's use of independent t pronouns Pronoun Number of occurrences n-tsetswe7 7' 37 7-enwi7 'you' 10 newi7-s 'he/she/it' 1 According to Julienne's mother, she and Julienne's father, like most fluent speakers, do not use independent pronouns often, and when they do, the first and second persons singular are most frequently used. This would explain why only one of Julienne's independent pronoun 99 occurrences is neither first nor second singular. The single use of the third person singular independent pronoun is interesting. It is used deictically in the following context: 30 newi7-s nts'qexe (7;2;28) 3 sg.indpr horse It's a horse (pointing to a picture of a unicorn) Julienne does not elaborate on this in later utterances, and for this reason I assume that this is a complete utterance, and not a DP containing the independent pronoun with its overt referent (i.e., the horse). It is interesting that Julienne would employ the third person singular independent pronoun in this case, as she does use Secwepemctsin deictics appropriately in other contexts: 31 yiri7 ntsetswe7 te q'imeke7 (7;2;28) deic lsg.indpr obi pen That's my pen As far as the plural independent pronouns are concerned, it is probably the case that Julienne does not know the plural independent pronouns, since her parents rarely, if ever, use them. There are a total of 103 utterances with Secwepemctsin pronominal reference. 45 utterances containing 48 uses of independent pronouns are found. There are 3 utterances in which independent pronouns are used more than once. One of them is the spontaneous form in (32), the other an incomplete utterance (utterance 233), and the third, a repetition (utterance 237). The spontaneous form in (32) is well-formed in adult speech given the appropriate context. 32 n-tsetswe7 xwe.xwist-si-n,. , 7-enwi7 (6;11;16) lsg.indpr like(redup)-2sg.obj-lsg.subj 2sg.indpr I love you 100 Independent Pronouns versus Bound Clitics/Affixes In the 17 full sentences containing independent pronouns, 35.3% (6/17) are used in accordance with the adult grammatical use; namely, the independent pronouns co-occur with the correct pronominal clitics and/or affixes on the predicate (33). (34) gives the predicates with the correct morphology, with zero morphology, and with inappropriate morphology. 33 Full sentences containing independent pronouns: 17 utt.005 Mom, wiwkten re 7enwi7 te kelc. utt.009 Mama, kectsetsme te ntsetswe7 te stemstete7stem utt.010 Mama, ntsetswe7 meq'-ken utt.011 Mama, kectsetsme ntsetswe7 tek orange juice! utt.012 ki7ce, k7ep re ntsetswe7 utt. 041 est'il ntsetswe7 utt.130 ntsetswe7 ta7a qwenmin yi7ene utt. 131 ntsetswe7 ta7a qwenmin yi7ene utt.187 ntsetswe7 xwexwistsin 7enwi7 utt. 194 ntsetswe7 illen te st7illen utt.204 ta7 k s qwenmimen ntsetswe7 te q'imeke7 utt.228 ntsetswe7 ell re Lizzie lecus utt.231 ntsetswe7 ell re Sandra lecus-ken utt.241 newi7-s nts'qexe utt.251 yiri7 ntsetswe7 te q'imeke7 utt.298 ntsetswe7 qwenmin te dance utt.301 ntsetswe7 qwenmin te dance 34 Grammatical, ungrammatical, and lack of pronominal morphology on predicates in Used with correct morphology (6) Used with zero morphology (9) Used with inappropriate morphology (2) wiwkten T see it' kectsetsme '(you) give me' xwexwistsin T like/love you' qwenmimen 'I want' meq'ken 'I'm full' n-q'imeke7 'my pen' k7ep 'sick' est'il 'stop' illen 'eat' qwenmin 'want' (4) lecus 'pretty' yiri7 'that (one)' lecusken 'I'm pretty' newi7-s ntse'qexe 'It's a horse' Compare the above with non-imitation full clauses that contain pronominal reference, but without independent pronouns. 101 35 Full non-imitation sentences with pronominal reference: 19 utt.001 Mama, kectsetsme t k cereal utt.007 kectsetsme te stemstete7stem utt.008 qwenmimen tek pop utt.015 *I pul'te re cmeye utt.016 *I wi7 secwen utt.030 *tkensteke... ntsetswe7 utt. 031 tkensteten n-qweten utt.032 tkensteke re sq'wext utt.036 tsut re Simon, *tkensteten nqw'u utt.037 tkensteten nqw'u utt.048 OK, tsut re Simon, te... [etc] utt. 137 *ts'umqsente re kenkeknem (for T kiss the black bear') utt. 141 Oh, xwistsin, ki7ce utt. 156 *illenten ('it is eating') utt. 176 *w7ec seysemus (' seysus') (repeated in utt. 178, 180)53 utt.208 *setsinem-s, Sandra (imperative setsinem-ce) utt.211 setsinem-ken,-ken utt.239 ?I feed, me, me, meteten, re nsqexe utt.302 *kectsetsme re ts'umqsentsme 36 Grammatical and ungrammatical morphology on predicates in non-imitation full sentences Used with correct morphology (10) Used with zero morphology (0) Used with inappropriate morphology (9) tsut 'says' -kectsetsme '(you) give me' (3) qwenmimen Twant' xwistsin T like/love you' setsinemken T sing' tkensteke 'touch it' tkensteten 'I touch it' (2) none pul'te 'kill' wi7 secwen 'finish bathing' ts'um'qsente 'kiss it!' ts'um'qsentsme 'kiss me' illenten 'eat' w7ec seysemus 'playing' setsinems 'sings' tkensteke 'touch it' tkensteten 'I touch it' It is a good sign that there are no forms used with zero morphology in the table above. This shows that when a clause is without independent pronouns, Julienne always uses pronominal ^Notice that the ungrammaticality of this utterance is caused by Julienne's productive analysis of Secwepemctsin morphology. The morphological breakdown of the utterance is given in (i) below. It contains a passive marker and a conjunctive marker, which are usually what accompanies a clause having a progressive particle w7ec. The utterance is only ungrammatical because the attested form here, seysus, is an irregular verb, (i) w7ec seyse-m-us prog play-pas-3sg.conj It's/They're playing 102 morphology to mark the arguments. This shows that Julienne knows arguments must be present in a clause, whether these are realized as overt lexical items, such as DPs, or pronominal morphemes, which are the clitics and affixes. What is interesting is that she appears to be treating DPs and pronominal morphology as if they were of equal status, since both can satisfy the argument requirement. The predicates used with the correct pronominal morphology are generally words with high frequency of usage for children, such as 'give', want', 'say', 'see', suggesting the co-occurrence of pronominal markings on many of these predicates could be due to idiomatic usages of the predicates, and not due to knowledge of adult grammar.. However, many of the predicates with inappropriate or zero pronominal morphology also appear to be words that should normally be used in a child's speech, or in conversation with a child, and it is not clear why these forms are used incorrectly by Julienne. However, most of the inappropriate forms are attested in the language, but are used in the wrong context by Julienne. For example, Julienne overgeneralizes forms such as ts'um'qsentsme 'kiss me (imperative)' andpixl'ste 'kill it (imperative)', but the forms are attested in the language. Furthermore, from translated utterances (such as utterances 20-29 in appendix E) , it is not difficult to see that Julienne does indeed understand many of the pronominal markings on predicates. These are thus encouraging signs that Julienne has a fairly thorough understanding of Secwepemctsin pronominals, and may simply be blocked from actively producing them54. ^It is widely known in the language acquisition field that a normal language-learner's comprehension is better than her performance. Due to the special circumstances surrounding Julienne, her language comprehension-production gap may be wider than normal (Ingram, p.c). This will have to be validated by systematic comprehension/production tests and analyses. Optimally the results should be compared and contrasted with comprehension/production studies of regular monolingual acquisition and bilingual acquisition. 103 Let us return to Julienne's over-use of independent pronouns. Often (52.9%, 9/17), independent pronouns are used by Julienne as a regular argument without pronominal marking on the predicate, exactly the kind of construction that English has. (37) provides the details. 37 Usage of pronominal morphology on the predicate Pronominal morphology on predicate With independent pronoun in clause Without independent pronoun in clause correct 35.3% (6/17) 52.6% (10/19) inappropriate 11.1% (2/17) 47.4% (9/19) zero morphology 52.9% (9/17) 00.0% (0/19) When an independent pronoun is available, Julienne tends not to use bound pronominal morphology on the predicate. When an independent pronoun is not available, she always uses bound pronominal morphology. Some kind of pronominal marking is necessary in all clauses, regardless of whether it is bound or free. Therefore, Julienne's independent pronouns and bound clitics/affixes appear to be of equal status functionally - both can satisfy a clause's need for arguments. These observations may give clues about the categorial difference/similarity between Julienne's independent pronouns and bound clitics/affixes. In the 45 utterances containing independent pronouns, a striking 91% (41/45) are used in non-imitation cases, yielding a high number of spontaneous independent pronoun use. This finding is significant. If Julienne is most comfortable using independent pronouns rather than pronominal bound morphemes, the implication is that she is more comfortable with an analytic grammar system. There is little doubt that this is a result of direct influence from English, the dominant language of her peers and her community. Indeed, a telling example of English influence in her Secwepemctsin acquisition is manifested in (38). 104 38 ntsetswe7 ta7a qwenmin yi7ene (6;11;16) lsg.indpr neg want deic I don't want this This is almost a direct translation from English, with a one-to-one correspondence between Secwepemctsin lexical item and English lexical item. Compare (38) to the grammatical adult form in (39). 39 ta7a k s qwenmim-0-en yi7ehe neg irrnom want-3sg.obj-lsg.subj deic I don't want this Literally 'it is not the case want-it-I this' 5.3.2 Adult Independent Pronouns versus Julienne's Independent Pronouns A brief comparison of adult Secwepemctsin independent pronouns and Julienne's Secwepemctsin independent pronouns is presented below. The characteristics of independent pronouns in adult Secwepemctsin are summarized in (40). 40 Characteristics of Adult Secwepemctsin independent pronouns a) They can be in argument position or in predicate position b) They are used rarely, and only in emphatic contexts c) There can only be one independent pronoun per clause when both arguments are third person d) In clauses with only third person arguments, the independent pronoun is subject-oriented e) Independent pronouns always refer to humans f) Independent pronouns can be possessors in a possessive phrase Does Julienne adhere to these restrictions? Julienne does use independent pronouns as arguments of a clause, but she does not use them predicatively. There are two possibilities to explain why there are no predicative uses of independent pronouns. First, independent pronouns cannot be predicates in Julienne's grammar. Second, Julienne cannot use independent pronouns predicatively because she has not yet learned the cleft construction. The 105 first possibility can be rejected based on the fact that both English and Secwepemctsin can have clefted pronouns, and if Julienne's Secwepemctsin performance is affected by English, there is no reason for her not to cleft pronouns. That she has not learned the cleft construction is highly probable, as syntactic constructions involving gaps are generally considered more advanced (see O'Grady 1997 and references therein), and Julienne does not use cleft constructions in her sample. Julienne's independent pronoun usage is rather broad, as opposed to the adults' use in restrictive contrastive contexts only. Independent pronouns found in Julienne's language samples are mostly used in non-intensifying contexts, and the sheer frequency of their occurrence (46.6% of all her utterances with pronominal reference) suggests she is not bound by the restriction found in the adult grammar. The lack of contrastiveness in her independent pronouns may be related to the fact that her independent pronouns and bound pronouns are of equal syntactic status. In other words, independent pronouns may simply be the unbound counterparts of the pronominal clitics and affixes, and the bound and independent pronouns are all of the same category. It is unclear what can trigger this reanalysis. It has already been mentioned that Julienne has used independent pronouns for both arguments in a transitive clause (see (32) above). It is possible to get two non-third person independent pronouns in a transitive clause, but only with appropriate discourse contexts. For Julienne, no context is necessary for any kind of independent pronoun use, not even for having two independent pronouns per clause. Julienne does not appear to follow the subject-orientation restriction of independent pronouns either, but this is due to the lack of sufficient data, since virtually all of her independent 106 pronouns are first and second person singular. Insufficient third person data is also the reason it is not possible to tell whether she adheres to the "human effect"; namely, that independent pronouns always refer to humans. As shown in (41), the data shows that with regards to independent pronouns, Julienne's grammar is again neither completely Secwepemctsin nor completely English. More importantly, the distribution of pronominal morphology is optional, which is both un-Secwepemctsin-like, and un-English-like. 41 Adult Independent pronouns Julienne's independent pronouns versus English Adult Secwepemctsin independent pronouns Julienne's Secwepemctsin independent pronouns English Context Contrastive (argument) or focus (predicate) only; never out of the blue Not contrastive; can be used out of the blue No restriction on overt unstressed pronouns; stressed pronouns are used contrastively Co-occurrence withD Not complements of D; adjoined to DP Optionally take D Never take D, but are themselves intransitive Ds Argument aosition Yes Yes Yes Predicate xjsition Yes No Yes Subject •orientation In clauses with third person arguments Insufficient data No restriction Pronominal morphology an predicate Obligatorily co-occur with pronominal clitics and affixes Optional Third person singular obligatorily present; rest of paradigm absent 5.3.3 Syntactic Analysis of Julienne's Independent Pronouns In sections 3.5 to 3.7, I discussed in detail a syntactic analysis of adult Secwepemctsin independent pronouns. The following section compares Julienne's utterances in relation to them. The observations about Julienne's Secwepemctsin independent pronouns are as follows. In terms of internal syntax, they take D optionally; this suggests that they behave like nominals. 107 In terms of external syntax, they always occur in argument position, but never in predicate position; independent pronouns can either co-occur with, or replace, pronominal clitics or affixes on the predicate. In terms of semantics, the independent pronouns do not show contrastiveness. The Familiarity Hierarchy The usage of independent pronouns in adult Secwepemctsin is discourse-oriented. Julienne's usage is not. There is no evidence that Julienne uses independent pronouns in the contrastive way that adults use them. Her independent pronouns appear to be regular pronouns. Since Julienne uses virtually no third person independent pronouns, no data is available to show whether there are asymmetrical behaviours regarding the subject position and third person independent pronouns, and hence no evidence is available to show whether Julienne is following the Familiarity Hierarchy in terms of nominal usage in clauses. However, following from the observation that Julienne's independent pronouns are not contrastive, there is no reason to expect her independent pronouns to show subject orientation, since the IPR occurs when one argument is more familiar than another. There also is insufficient data to show whether Julienne's grammar is constrained by the ONI Law or whether it shows third person subject clefting asymmetry. Category of Julienne's Pronouns Julienne's independent pronouns are argument DPs, but there is no positive evidence that they are adjoined to argument DPs like adult Secwepemctsin independent pronouns. Lack of third 108 person independent pronouns co-occurring with overt DPs indirectly suggests the lack of adjunction structure. Julienne's independent pronouns, again virtually all first and second person singular, occur optionally with determiners. However, they occur without determiners more often than with. (42) is an example of a first person singular preceded by the oblique determiner, occurring as the indirect object of a clause (ungrammatical in adult grammar), and (43, 44) are examples of the direct determiner preceding an independent pronoun. 42 mama, ke-c-t-se.ts.m-e tentsetswe7 te k orange juice (6;3;0) mama, give-2sg.obj-tr-lsg.subj(redup)imp obllsg.indpr obi irr orange juice Mama, give me some orange juice! 43 ki7ce, k7ep re ntsetswe7 (6;3;0) mother sick det lsg.indpr Mother, I'm sick 44 temtu-m-0-en re 7enwi7 ell Sandra (6;10;25) dream-tr-3sg.obj-lsg.subj det 2sg.indpr and Sandra I dreamed of you and Sandra Since Julienne does use determiners for ordinary nominal arguments in her utterances, and she also uses determiners on independent pronouns, we might assume that independent pronouns are DP projections projected from overt Ds in Julienne's grammar. However, it is interesting to see why she sometimes omits determiners for independent pronouns that are arguments. Some possible categories that independent pronouns might assume, as suggested in 3.4.7, are N, D°, and DP. The adult Secwepemctsin independent pronouns are claimed to be not Ns, D°s, nor DPs. They are analyzed as their own maximal projection XP, which adjoins to DPs in argument positions to provide contrastiveness. 109 The fact that the determiner plus independent pronoun string does exist in Julienne's sample suggests that Julienne at least has available the construction in (45). 45 [detD [indpr]]DP : - v Why is the determiner sometimes absent, then? In many cases where her independent pronouns lack the determiner, it is clear that such pronouns are arguments. 46 ntsetswe7 xwe.xwi.s-t-si-n 7enwi7 (6;11;16) lsg.indpr like.redup-tr-2sg.obj-lsg.subj 2sg.indpr I love you From the perspective of determiners, it appears that Julienne is treating independent pronouns as lexical nominals. In the adult grammar, I have analyzed independent pronouns as heading their own maximal projections which can adjoin to lexical DPs or pro. Julienne has no third person independent pronouns with nominal referents, so there is no evidence of an independent pronoun being adjoined to a lexical DP. However, in Julienne's use of first and second person independent pronouns (which in the adult grammar are attached to pro), it is apparent that she treats them as arguments. She has sometimes attached a determiner to these pronouns, suggesting she has reanalyzed independent pronouns as lexical nominal projections that are complements of D°, as in (47). 47 DP (det) indpr Other possible structures for Julienne's independent pronouns are in (48) and (49). 110 48 DP (det) indpr 49 DP D NP (det) indpr 0 The structures in (48) and (49) can account for the fact that there are no predicative independent pronouns Julienne's sample, the lack of predicative independent pronouns would not follow from (47), as the NP should be able to be predicative. That Julienne's independent pronouns do not co-occur with overt nominals follows from (47), in which the NP position is already filled. It also possibly follows from (48), in which there is no internal structure to DP. However, it does not follow from (49), since there is no explanation why the NP in (49) is never filled. Since I have claimed that Julienne does not use independent pronouns as predicates because she has not acquired cleft constructions, the structures in (47) and (48) are equally possible as structures for Julienne's independent pronouns. I speculate that Julienne is confused about the status of independent pronouns in Secwepemctsin. On the one hand, she feels that such pronouns are arguments, and therefore should receive DP status by being marked with a determiner. On the other hand, she knows from English that while pronouns can be arguments, they don't ever take determiners. I will assume that Julienne's independent pronouns can have the structures in (47) and (48). Notice that this suggests independent pronouns in Julienne's grammar are R-expressions. If her independent pronouns have the structure in (47), taking either overt D or null D, then once she acquires cleft constructions, she will use independent pronouns in predicate position. If on the 111 other hand her independent pronouns have the structure in (48), with unanalysable internal structure, then her independent pronouns would not be able to occupy predicate position. Suppose that my speculation above is correct. Would Julienne's use of independent pronouns as argument DPs (with a determiner) affect how she treats bound pronouns on the predicates? Data suggests the answer to be no. In (50, 51), where independent pronouns are D heads (due to the lack of determiner), predicates are optionally marked with bound pronouns. 50 Mama, ntsetswe7 meq'-ken (6;3;0) Mama, lsg.indpr full-1 sg.ind Mama, I'm full 51 ntsetswe7 ta7a qwenmin yi7ene (6;11;16) lsg.indpr neg want-0 deic I don't want this When independent pronouns are full DPs, the predicates are also optionally marked with bound pronouns (52, 53). 52 re 7enwi7 tsxwente sk'epqen ell re sq'wext (7;4;6) det 2sg.indpr pickup-imp head and det leg You pick me up by the head and legs (one arm under the neck and one arm under the legs) . 53 ki7ce, k7ep re n-tsetwe7 (6;3;0) mother, sick det lsg.indpr Mother, I'm sick If her use and lack of bound pronominal marking on the predicate is not due to her use or lack of determiner on the independent pronouns, then what affects it? This question is of interest because, while in the adult grammar, pronominal bound morphemes (licensing pro) are obligatory and lexical pronouns and R-expressions are optional, in Julienne's grammar, the bound pronominal morphemes and the overt arguments appear to be equal in status; that is, the 112 presence of either bound pronominal morphemes or overt arguments satisfies the requirement for predicates to have a way of identifying arguments, as discussed earlier. It is plausible that the influence of English is so overwhelming that for Julienne, Secwepemctsin independent pronouns are the equivalent of English unstressed pronouns, while Secwepemctsin bound pronouns are the equivalent of verb inflection. There is evidence in the literature of English acquisition that children at a very early age drop overt subjects and sometimes even overt objects; this is why Julienne sometimes does not have independent pronouns - she is applying pro-drop. It is also the case that English speaking children have uninflected verbal stems at certain stages in their acquisition. This explains why Julienne does not always have bound pronouns in Secwepemctsin. The problem with this kind of reasoning, however, is that Julienne is already past the age for displaying pro-drop inconsistencies. Monolingual English children tend not to display pro-drop past the age of 2 or 3, and Julienne's fluency of English is adequate for children her age. The English nominative personal pronouns I, you(sg), he, she, it, we, youfpl), they have all been used in Julienne's sample correctly. The accusative forms me, you(sg), him, her, it, us and the possessive forms my, your(sg), his, have also been used appropriately. A relevant question to ask at this point, also, is this: if Julienne uses independent pronouns in Secwepemctsin as the equivalents of English pronouns, what would she employ to show contrastiveness of these pronouns? A likely prediction is that she would use stress, as English pronouns are stressed to show contrastiveness55. Finding out whether Julienne does What would be an interesting topic for further exploration is this: what is Julienne missing in the input so that she does not acquire the contrastiveness of independent pronouns? The presence of English cannot be a factor, as many generations before her have successfully acquired independent pronouns in a predominantly English-speaking environment. 113 contrastively stress Secwepemctsin independent pronouns is a step towards reaching an explanation for her grammatical system. 5.4 Remaining Questions The one question that has constantly been asked during the examination of Julienne's Secwepemctsin grammar is this: what is the source for Julienne's grammar? That is, what makes her grammar the way it is? Julienne's Secwepemctsin grammar clearly is not only affected by the Secwepemctsin adult input she receives. If it were, her grammar would be completely the same as that of fluent Secwepemctsin speakers. Julienne's Secwepemctsin grammar is also not completely affected by English. If it were, she sould be speaking a relexified version of English, in which English words are substituted by Secwepemctsin words, but the grammar is English. Julienne's grammar may be a hybrid of adult Secwepemctsin and adult English. However, there are properties of her Secwepemctsin grammar that are distinctively non-Secwepemctsin and non-English. For example, she uses determiners optionally on arguments, but in both Secwepemctsin and English, determiners are obligatory on nominals. Another example is the equal status of her Secwepemctsin DP arguments and pronominal clitics/affixes. Julienne's clauses always need some kind of argument marking; this follows from both Secwepemctsin nd English. The argument marking can be bound clitics/affixes (following from Secwepemctsin), overt DPs (following from English), or both (following from Secwepemctsin). Her choice of which kind of argument marking system to use appears to be arbitrary, but in neither Secwepemctsin nor English can the argument marking system in clauses be arbitrary. For the 114 reasons stated here, Julienne's Secwepemctsin grammar cannot be simply a hybrid grammar made from Secwepemctsin and English alone. ' ! . The next possibility is that Julienne's Secwepemctsin grammar is a hybrid of Secwepemctsin and English held together by Universal Grammar. Intuitively, this is the most plausible possibility, as it can account for the Secwepemctsin- and English-external force creating the innovations in her Secwepemctsin samples, which are neither Secwepemctsin nor English. If this is the case, then Universal Grammar should account for the optional determiners on arguments and the arbitrary use of argument markings in clauses. The use of determiners can be parametrized easily. Julienne's determiner parameter may have been initially set to [-determiner], and her optional use is triggered by positive data in Secwepemctsin and English. Since this case study looks only at one stage in her acquisition process, it is likely that there are still parameters that have not been set yet - the determiner parameter being one of them. However, English-speaking children are generally believed to start using detenniners at around age 2 or so (O'Grady 1997, Ingram 1989), and Julienne, who uses English determiners appropriately in her language sample, is already seven years old. With regards to the arbitrary use of argument marking in clauses, Julienne's system is clearly a marked one cross-linguistically. There are few (if any) languages that can, within themselves, choose arbitrarily to have either only bound pronominals on predicates or only free DP arguments without any bound pronominal morphology. The role that Universal Grammar plays in monolingual and bilingual acquisition is therefore of great interest to the present topic. Bound by time and space limits, I will not explore the 115 possibilities of Universal Grammar's effects on Julienne. This case study will undoubtedly benefit from such a discussion in the future. 5.5 Chapter Conclusion It may at first appear that Julienne does not produce many "real" Secwepemctsin sentences, in the sense that her utterances are not very adult-like. Even though she is still a child, it is not typical for a seven-year old to be talking with mostly ungrammatical sentences by adults' standards; that is, a typical seven-year old acquiring a language under normal circumstances may speak in shorter and simpler sentences than adults, with occasional grammatical errors, but the majority of her speech should be well-formed. With Julienne, her Secwepemctsin performance is not advanced, but she is obviously not learning the language under the most normal circumstances, having only her parents as providers of Secwepemctsin input. Many syntactic characteristics of Secwepemctsin still surface in Julienne's grammar. Although in performance she is inconsistent with many grammatical constraints, in comprehension she shows very thorough understanding of Secwepemctsin. Her production data greatly underrates her knowledge of Secwepemctsin. It is therefore in this light that her language sample should be viewed. Since Julienne's language samples in the database are collected when she is instructed specifically to use Secwepemctsin, and only 60.98% of the total utterances are in Secwepemctsin, it can be deduced that she uses much less Secwepemctsin in other contexts, when she is surrounded by non-Secwepemctsin speaking children and adults. The investigations and results reported here illustrate that Julienne performs much better in English 116 than in Secwepemctsin in all respects. It also appears that English's analytic nature plays a role in influencing Julienne's overregularization of independent pronouns. The generalizations that can be drawn about independent pronouns from Julienne's language sample are the following. Syntactically, the bound pronominal morphemes and overt arguments are of equal status, since Julienne does not allow clauses with no argument marking whatsoever, but if either bound pronominal morphology or overt arguments, or both, are present, then the clause is felicitous. The independent pronouns are R-expressions, or DPs, which will always be able to occupy argument positions. Depending on the internal structure of these independent pronouns, they may or may not occupy the predicate position once Julienne's Secwepemctsin performance advances. In terms of semantic generalizations, Julienne's Secwepemctsin independent pronouns are not contrastive, unlike adult independent pronouns. Why she has not acquired the contrastiveness of independent pronouns is a missing piece in the acquisition puzzle. The observation that syntatically the independent pronouns and bound clitics/affixes are of equal status may be indicative of her reanalysis of independent pronouns, which appear to be free counterparts of the bound pronominals, with no difference in function between them. Why this reanalysis has occurred is not clear. 117 C H A P T E R 6. C O N C L U S I O N S A N D I M P L I C A T I O N S 6.0 Introduction In this chapter, I summarize what has been achieved through the study of the syntax and acquisition of Secwepemctsin independent pronouns. The findings about the grammar of independent pronouns are contributive to Salish Linguistics as well as to the existing studies of pronouns. The case study reported in Chapter 4 is an example that, with persistence, much passive knowledge of a little-used language can be acquired. The sections under 6.2 discuss the importance of providing positive contexts for language use, and of using an endangered language creatively so that it may continue to survive. 6.1 Summary In this thesis, I have studied the syntax of Secwepemctsin independent pronouns, and applied the analysis I reached to explain the interesting usage of independent pronouns by a young learner of Secwepemctsin. It is only when an analysis of the adult grammar has emerged that the examination of acquisition data can be viewed in a clear way. My analysis of the syntax and semantics of independent pronouns in adult Secwepemctsin together with the predominantly English-speaking environment in which she has been raised throws light on the innovations in Julienne's system, and partially predicts why Julienne's independent pronouns behave so differently from the adult's. 6.2 Looking Forward and Beyond 118 6.2.1 Interesting Differences in Comparative Salish Linguistics The finding that independent pronouns in Secwepemctsin are DP-adjoined, and have no binding properties of their own, is of great importance in a comparative Salish context, since it has been shown that in Halq'emeylem, independent pronouns behave like R-expressions. In contrast to Secwepemctsin, Halq'emeylem independent pronouns are lexical elements that take determiners and are DPs when they are in argument position, while they drop determiners when they are in predicate position. In the studies of independent pronouns of both languages, syntactic behaviour follows from syntactic category and position, thus showing that in both cases the internal structure of DP determines its external behaviour. 6.2.2 To Keep the Language Alive Since Secwepemctsin is a radical head marking language, if the pronominal markings on the predicate are not mastered by the language learner, the ultimate consequence is that complete fluency of the language cannot be achieved. Julienne, the child studied in this thesis, is one of very few children to be in the fortunate position of having active Secwepemctsin-speaking parents who have instilled in her very good passive knowledge of the language, even though in Secwepemctsin production she does not perform as well as in comprehension. It thus appears that in order for a child to successfully acquire a language, maximum input in the home helps a learner to achieve a level of knowledge that is good, but is not quite enough for active usage so far. 119 Lack of a social context in which to use a minority language, and the overwhelming influence of a dominant language are the two main impediments to introducing and retaining a minority language. For this reason, it is extremely important that learners of such a language be provided with as intensive a learning environment as possible, maximizing a learner's exposure to the language. With young learners, it is optimally the case that exposure to the target language be given before the emergence of words by an infant, as perception of language starts as soon as a baby is born, if not earlier. It is important to provide as much target language input as possible, first in the home, and then, as the child expands her social circle, in the playground, daycare, and school. This, of course, would require that the elder speakers of the language be involved in the project of bringing back the language, since many parents of children today do not speak the minority language. The point, therefore, is that the language would need to be present in the home as well as in the larger social context to facilitate acquisition. As has been illustrated with the case study on Julienne, merely parental input is not enough. I am not aware of any formal study of results of aboriginal language immersion programs. However, there is a Secwepemctsin immersion program at the Chief Atham School in Chase, B.C, that has been reputed to be a very successful program56. This is something to be applauded, but there is no doubt that the learning must take place outside the school as well, most possibly with parental participation so that children would continue to use the language as much as possible outside the classroom. The presence of a supportive community with the common goal of re-introducing the minority language to a new generation is also extremely necessary in order to provide contexts for language use in all aspects of a child's daily life, such as school, home, community T. V., radio, neighbourhood road or store signs, etc.. ,6Much research is needed in finding out the results of such programs in aboriginal language education. 120 Another barrier to overcome in bringing children up with a minority language is that children, at some point, become unwilling to learn, or to continue using, the target language. Often, children of multi-lingual backgrounds start rejecting a language if they feel that using the language would make them different (in a negative way) from their friends, and since this is undesirable, they disassociate themselves from the minority language. This generally starts when children become involved with the environment outside the home, and can last into the adolescent years, sometimes even into adulthood. Therefore, it is extremely necessary also to have the presence of a supportive community with the common goal of opposing the stigma associated with minority language usage by providing positive contexts for language use. Perhaps then, the stimulation and the motivation of a language learner would be greatly improved. In this, context, I would also like to mention that although Julienne's acquisition of Secwepemctsin as a first language is unusual, it is not the only recent case in Salish. A speaker of another Salish language also provided only Salish language input to her children since their birth57. It is known that the children always understood the language, but never uttered a single utterance in that language. These children are in their teens now, and their mother is still persistent in speaking to them in the target language. From a recent conversation with one of the teens mentioned here, I was told that she now has many friends of different ethnic and linguistic backgrounds, and she envies her friends for being able to talk to their grandparents in their native tongue. I have also recently heard that these teens are now starting to speak the target language. Because they have been receiving input since an early age, they already have a great deal of knowledge of the language, and can use it with relative ease. I am almost a certain that this change in their attitude comes from their seeing that it is no longer socially "Permission from this family has not been obtained, and therefore no detailed information can be given at this point. 121 negative for young people to connect with the culture, and therefore the language, of their elders. It is not "uncool" anymore. It is therefore of utmost importance that people are educated to be proud of their cultural and linguistic heritage. This case thus provides evidence that although the prospects of teaching a minority language to a child in an adverse linguistic environment are extremely difficult, the results are promising. The ultimate goal in reviving a language is for the new speakers to achieve a level of fluency where these speakers can use the language productively and creatively. By productively and creatively, I mean that the speakers have enough knowledge about the language to be able to coin new words and create new expressions - to name objects and express notions that never before had existed in the language. This is in fact a crucial part of language survival. If speakers of a language cannot create a way to name new objects and express new notions to keep up with the changes in the society surrounding that language, then the language will be abandoned because many old objects will have disappeared, and contexts for many old expressions will have been lost. There is a danger that in such a situation, the dominant language will be used exclusively to name the new objects, and to express the new notions. In summary, it is necessary for a language learner to have maximum exposure to the target language, to have motivation to learn the language, and to have relevant context for the language. These factors are necessary to ensure that a language survives and gets passed on to generations to come. 122 6.2.3 Possible Future Changes It is of course best that minority languages be kept in their pure form by being passed on from generation to generation. However, language change is unavoidable, no matter how large the number of speakers is. Thinking in pessimistic terms, if minority languages manage to be maintained, but characteristics of dominant languages continue to permeate into minority languages, what will emerge is that the direction in which the minority language changes will be increasingly towards the dominant language, resulting in a pidgin or creole language. The smaller the number of fluent speakers there are in the minority language, the faster these changes will take place. Thus, even with the best of efforts, it is possible that new speakers of many North American aboriginal languages will make their languages appear more like the dominant language of their area by assuming the dominant language's characteristics, such as sound systems, word order or, as in the case of Julienne, the morphological type. 6.3 Final Summary and Conclusion A detailed description of the syntax of Secwepemctsin independent pronouns is given in this thesis. I have presented an analysis in which independent pronouns in argument position are adjuncts to DP. The DP may either be a pro or a full lexical DP. I have shown that binding effects follow from this analysis, and I have contrasted the behaviour of Secwepemctsin independent pronouns with those of Halq'emeylem. Secondly, I have shown a strong syntactic subject-object asymmetry with respect to third person independent pronouns, and I have analyzed this asymmetry in terms of the mapping of familiarity hierarchies onto syntactic positions. 123 In Chapter 4, I turn to the semantics of independent pronouns. Using the alternative based view of semantics of Rooth (1992) and Buring (1995). I analyze Secwepemctsin independent pronouns as contrastive focus and contrastive topic depending on their syntactic position. I show that the semantic behaviour of Secwepemctsin independent pronouns may follow from a contrast-based analysis of topic and focus. In Chapter 5, I give a general overview of the syntax of a child who has received linguistic input in Secwepemctsin and English all her life, but, based on her use of independent pronouns, I show that she continues to have difficulty with the agglutinative aspect of the Secwepemc language. I suggest that not only is this due to the massive influence of the more analytic dominant language, English, parametric possibilities of universal grammar must play a role in influencing the child's grammar, which exhibits non-Secwepemctsin and non-English grammatical characteristics. In section 5.3 ,1 illustrate that the new speaker of Secwepemctsin examined in this case study uses independent pronouns differently from fluent adult speakers, and that the differences follow from her reanalysis of independent pronoun to liken the bound pronominal clitics/affixes of the language. Although Julienne does not use independent pronouns like adult speakers do, her usage of independent pronouns in Secwepemctsin, as well as her overall Secwepemctsin syntax, still exhibit many Secwepemc language characteristics. The efforts of this child and her family in maintaining this language and culture is thus worthwhile. For the majority of new Secwepemctsin speakers of this generation, the kind of exposure to the language that Julienne has is unavailable. It is therefore extremely important that increased linguistic resources be 124 provided for Secwepemc children if they are to learn the language from their parents' or grandparents' generation, and in turn teach their own children the language. While many interesting syntactic phenomena and bilingual acquisition phenomena surfaced through the study of contrastive pronouns and the case study on Julienne, further work should be devoted in both areas for an empirically complete picture. 125 REFERENCES Abney, Stephen 1987. The English Noun Phrase in Its Sentential Aspect. Ph.D. Dissertation. MIT. Aissen, Judith 1992. Topic and Focus in Mayan. Language 68:43-80. Allen, Shanley 1994. Acquisition of Some Mechanisms of Transitivity Alternation in Arctic Quebec Inuktitut. Ph.D. Dissertation, McGuill. Anagnostopoulou, Elena, and Anastasia Giannakidou 1995. Clitics and Prominence, or Why Specificity is not Enough. Papers from the 31st Regional Meeting of the Chicago Linguistics Society. Volume 2: The Parasession on Clitics: 1-12. Baker, Mark 1991. On Some Subject/Object Asymmetries in Mohawk. Natural Language and Linguistic Theory 9:537-576. Blake, Susan J. 1997. Another Look at Passives in Sliammon (Salish). Proceedings of the 32nd International Conference on Salish and Neighbouring Languages. Peninsula College. Port Angeles, Washington. Buring, Daniel 1995. The 59th Street Bridge Accent - On the Meaning of Topic and Focus. Ph.D. Dissertatin, Universitat Tubingen. Cardinaletti, Ana and Michal Starke 1996. Deficient Pronouns: A View from Germanic: A Study in the Unified Description of Germanic and Romance. Chapter in Comparative Germanic Syntax, Volume II, eds. Thrainsson, Hoskuldur, Epstein, Samuel David, and Peter, Steve. Dordrecht: Kluwer. Chomsky, Noam 1995. The Minimalist Program. Cambridge, Massachusetts: MIT Press. Davis, Henry 1994. Tali-ho! Papers for the 29th International Conference on Salish and Neighbouring Languages: 117-144. Pablo, Montana. Davis, Henry1996. Agreement Parameters in Salish. Proceedings of Canadian Linguistic Association, 1996, Brock University. Davis, Henry 1997a. Remarks on Nouns and Nominalizations Salish. Paper presented at the University of Washington. Davis, Henry 1997b. Turning the Pronominal Argument Hypothesis on its Head. Invited talk, Workshop on Structure and Constituency in the Languages of the Americas. University of Manitoba. Davis, Henry, Dwight Gardiner, and Lisa Matthewson 1993. A Comparative Look at wh-Questions in Northern Interior Salish. Papers for the 28th International Conference on Salish and Neighbouring Languages, 79-95. University of 126 Washington. Seattle, Washington. Davis, Henry, I. S. Lai, and Lisa Matthewson 1997. Cedar Roots and Singing Detectives: Attributive Constructions in English and Salish. Proceedings of the 32nd International Conference on Salish and Neighbouring Languages, 291-316. Peninsula College. Port Angeles, Washington. Demirdache, Hamida 1998. Condition C. Atomism and Binding, eds. Bennis, H, P. Pica, and J. Rooryck. Dordrecht: Foris Publications. Demirdache, Hamida, Dwight Gardiner, Peter Jacobs, and Lisa Matthewson 1994. The Case for D-Quantification in Salish: 'All' in St'at'imcets, Squamish, and Secwepemctsin. Papers for the 29th International Conference on Salish and Neighbouring Languages, 145-203. Demirdache, Hamida, and Lisa Matthewson 1995. On the Universality of Syntactic Categories. Procedings of NELS 25: 79-93. Diesing, Molly 1992. Bare Plural Subjects and the Derivation of Logical Representations. Linguistic Inquiry 23:353-380. Dixon, May, and A. H. Kuipers 1974. A Shuswap Course. Leiden. Dixon, R. M. W. 1979. Ergativity. Language 55:59-138. Dixon, R. M. W. 1994. Ergativity. Cambridge, New York: Cambridge University Press. Dorian, Nancy 1981. Language Death: The Life Cycle of a Scottish Gaelic Dialect. Philadelphia: University of Pennsylvania Press. Erteshik-Shir, Nomi 1993. The Dynamics of Focus Structure. Ms., Ben Gurion University of the Negev. Galloway, Brent 1993. A Grammar of Chilliwack Halkomelem. University of California Publications in Linguistics 96. Berkeley: University of California Press. Gardiner, Dwight 1993. Structural Asymmetries and Preverbal Positions in Shuswap. PhD. Dissertation, Simon Fraser University. Gardiner, Dwight 1996. Determiner Phrases in Secwepemctsin (Shuswap). Paper for the 31st International Conference on Salish and Neighbouring Languages, 175-183. U.B.C. Vancouver. Gardiner, Dwight 1998 Topic and Focus in Shuswap (Secwepemctsin). In Salish Language and Liguistics: Theoretical and Descriptive Perspectives, eds. Czaykowska-Higgins, E. and Dale Kinkade. New York. Mouton de Gruyter. 127 Gardiner, Dwight, Lisa Matthewson, and Henry Davis 1993. A Preliminary Report on Word Order in Northern Interior Salish. Papers for the 28th International Conference on Salish and Neighbouring Languages. University of Washington. Seattle, Washington. Gerdts, Donna B. 1988. Object and Absolutive in Halkomelem Salish. New York: Garland. Gerdts, Donna B. 1989. Object Agreement in the Halkomelem Salish Passive: A Morphological Explanation. In General and Amerindian Ethnolinguistics: In Remembrance of Stanley Newman (Contributions to the Sociology of Language), eds. M. R. Key and H. Hoenigswald. Berlin: Mouton de Gruyter, 185-200. Giusti, Giulianna 1994. Enclitic Articles and Double Definiteness: A Comparative Analysis of Nominal Structure in Romace and Germanic. The Linguistic Review 11:241-255. Hyams, Nina, and Kenneth Wexler 1993. On the Grammatical Basis of Null Subjects in Child Language. Linguistic Inquiry 24(3):421-459. Hukari, Thomas E. 1976. Transitivity in Halkomelem. Paper presented at the 11th International Conference on Salish and Neighbouring Languages. Ignace, Marianne 1995. Shuswap Language Framework Accessment. Report. Secwepemc Cultural Education Society and School District 75. Ingram, David 1989. First Language Acquisition Method, Description, and Explanation. New York: Cambridge University Press. Jelinek, Eloise, and Richard Demers 1983: The Agent Hierarchy and Voice in Some Coast Salish Languages. International Journal of American Linguistics 49-2:167-185. Jelinek, Eloise, and Richard Demers 1994. Predicates and Pronominal Arguments in Straits Salish. Language 70:697-736. Kamp, Hans, and Uwe Reyle 1993. From Discourse to Logic. Dordrecht: Kluwer. Kayne, Richard 1995. The Anti-symmetry of Syntax. Cambridge, Massachusetts: MIT Press. Kinkade, M. Dale 1983. Salish Evidence Against the Universality of'Noun' and 'Verb'. Lingua 60:25-40. 1 Kinkade, M. Dale 1991. The Decline of Native Languages in Canada. In Endangered Languages, eds. R. H. Robins and E. M. Uhlenbeck. Berg Publishers. Krifka, Manfred 1998. Scope Inversion under the Rise-Fall Contour in German. Linguistic Inquiry 29(1): 75-112. Kroeber, Paul D. 1991. Comparative Syntax of Subordination in Salish. Ph.D. Dissertation, University of Chicago. 128 Kuipers, Aert 1990. A Report on Shuswap. Paris: Peeters/Selaf. Kuipers, Aert 1974. The Shuswap Language: Grammar, Text, Dictionary. The Hague: Mouton. Lai, I. S. 1997. Secwepemctsin Complex Predicates. Proceedings of the Northwest Linguistics Conference, 209-220. University of British Columbia. Vancouver. Lai, I. S. 1998. Secwepemctsin Independent Pronouns: Evidence for Subject - Object Asymmetry. Proceedings of the 33rd International Conference on Salish and Neighbouring Languages, 307-321. University of Washington. Seattle, Washington. Lai, I. S. and Marianne Ignace 1998. A Preliminary Study of Secwepmec Language Acquisition by a Young Chid. University of Washington. Seattle, Washington. Longobardi, Giuseppe 1994. Reference and Proper Names: A Theory of N-Movement in Syntax and Logical Form. Linguistic Inquiry 25(4): 609- 665. Marcus, Gary, Steven Pinker, Michael Ullman, Michelle Hollander, John Rosen, and Fei Xu. 1992. Overregularization in Language Acquisition. Monographs of the Society for Research in Child Development, 1992, 57(4, Serial No. 228). Matthewson, Lisa 1996. Determiner Systems and Quantificational Strategies: Evidence from Salish. Ph.D. Dissertation. University of British Columbia. Vancouver. Matthewson, Lisa, Henry Davis, and Dwight Gardiner 1993. Coreference in Northern Interior Salish. Papers for the 28th International Conference on Salish and Neighbouring Languages. University of Washington. Seattle, Washington. Matthewson, Lisa and Henry Davis 1995. The Structure of DP in St'at'imcets (Lillooet Salish). Papers for the 30th International Conference on Salish and Neighbouring Languages, 55-68. University of Victoria. Victoria. Montler, Timothy 1993. Relative Clauses and Other Attributive Constructions in Saanich. In American Indian Linguistics and Ethnography in Honour of Laurence C. Thompson, eds. Mattina, Anthony, and Timothy Montler. University of Montana Occasional Papers in Linguistics No. 4. Missoula, Montana. Newman, Stanley 1977. The Salish Independent Pronoun System. InternationalJournal of American Linguistics 43:302-314. Neeleman, Ad, and Fred Weerman 1997. LI and L2 Word Order Acquisition. Language Acquisition, 6(2), 125-170. O'Grady, William 1997. Syntactic Developement. Chicago: The University of Chicago Press. Postal, Paul M. 1969. On So-called "Pronouns" in English. In Modern studies in English, 129 eds. Reibel David, and Sanford Schane. Englewood Cliffs, N.J.: Prentice-Hall. Pye, Cliff 1992. Language Loss among the Chilcotin. InternationalJournal of the Sociology of Language 93: 75-86. Ritter, Elizabeth 1995. On the Syntactic Category of Pronouns and Agreement. Natural Language and Linguistic Theory 13:405-443. Roberts, Taylor 1994. Subject and Topic in St'at'imcets (Lillooet Salish). M A . Thesis, University of British Columbia. Royal Commission on Aboriginal People. 1996. Report of the Royal Commission on Aboriginal People, Volume I. Looking Forward, Looking Back. Ottawa. Rooth, Mats 1992. A Theory of Focus Interpretation. Natural Language Semantics 1(1) :75-116. Speas, Margaret 1990. Phrase Structure in Natural Language. Dordrecht: Kluwer. Thompson, Laurence C. 1979. Salishan and the Northwest. In The Languages of Native America: Historical and Comparative Assessment, eds. L. Campbell and M. Mithun. Austin: University of Texas Press. 692-765. Wang, Qi, Diane Lillo-Martin, Catherine Best, and Andrea Levitt 1992. Null Subject Versus Null Object: Some Evidence from the Acquisition of Chinese and English. Language Acquisition 2-3: 221-254. Williams, Edwin 1997. Blocking and Anaphora. Linguistic Inquiry 28(4):577-628. Wiltschko, Martina 1995. Presuppositions in German Dislocation Construction. Folia Linguistica 29(3-4):265-295. Wiltschko, Martina 1998. The Syntax of Independent Pronouns in Halq'emeylem. Paper presented at the Workshop on Structure and Constituency in the Languages of the Americas. March, 1998. Wiltschko, Martina 1998. On the Internal and External Syntax of Indipendent Pronouns in Halq'emeylem. Papers from the 33rd International Conference on Salish and Nighbouring Languages, 428-447. University of Washington. Seattle, Washington. 130 A P P E N D I X A ABBREVIATIONS aut=autonomous caus=causative conj=conjunctive det=determiner perf=perfect dir=direction emph=emphatic excl=exclusive fc=fiill control hab=habitual imp=imperative ind=indicative indpr=independent pronoun inter=interrogative intr=intransitive irr=irrealis neg=negative nom=nominative obl=oblique obj=object pas=passive pl=plural poss=possessive ptc=particle redup=reduplication sg=singular subj=subject top=topic tr=transitive unsp=unspecified. A P P E N D I X B PHONETIC INVENTORY OF SECWEPEMCTSIN (adapted from Kuipers 1974 and Gardiner 1993). The orthography used in this thesis is followed by the IPA symbol in brackets. Consonants: Labial Alveolar Lateral Palatal Velar Uvular Glottalized Stops plain p(p) t(f) k(k) kwCk") q(q) qw(qw) 7(?) glottalized P' (P') k'(k') kw'(kw') q'(q') qwV) Affricates plain ts(tj) glottalized ts'(ts) m Fricatives s(s) 11(1) c(x) cw(xw) x(x) xw(xw) h(h) Resonants plain m(m) n(n) 1(1) y(y) r(Y) w(w) g(?) gw(r) glottalized m'(m') n' (n') l'(l') y'(y') r'(Y') w'(w') Vowels: i u e 0 a 131 APPENDIX C PRONOMINAL PARADIGMS Intransitive Clitic Paradigm (Kuipers 1974) 1 sg. 2sg. 3 sg. 1 pi. incl. 1 pi. excl. 2 pi. 3 pi. Indicative -ken -k -0 -kt -kucw -kp -0 Conjunctive -wen -(w)cw -(w)s -(w)t -kucw -(w)p -(w)s Possessive n- -7 -s -kt -kucw -mp -s Transitive Subject Suffix Paradigm (Kuipers 1974) 1 sg. 2sg. 3 sg. 1 pi. incl. 1 pi. incl. 2 pi. : 5 pi. -(e)n -(e)c -(e)s -(e)t/-(e)m -kucw -(e)p (e)s Transitive Object Suffix Paradigm (Kuipers 1974) 1 sg. 2sg. 3 sg. 1 pi. incl. 1 pi. excl. 2 pi. 3 pi. -sem-/-sm- -S1-/-S- 0 -el-/-l- -kucw -ulm-/-lm- 0 Independent pronominal paradigm (adopted from Kuipers 1974)' 1 sg. 2sg. 3 sg. n-tse.ts-we7 1 sg.poss-tse(redup)-deic 7-enwi7 2sg.poss-emph newi7-s emph-3sg.poss 1 pi. incl. 1 pi. excl. 2 pi. 3 pi. wll-enwi7-kt wll-enwi7-s-kucw pl-emph-3 sg. poss-excl wll-enwi7-mp pl-emph-2pl.poss wll-enwi7-s pl-emph-3 sg.poss APPENDIX D Actual-determinate Hypothetical-indeterminate Present Absent Irrealis Direct re 1 k Oblique te/t' tk/t'k 'The independent pronouns are all composed of a stem (henceforth "emphatic stem") and the possessive clitic. The emphatic stem for the first singular form is different from the rest of the paradigm. According to Newman (1977), the first person singular stem in Secwepemctsin is derived from the proto-Salish first person singular emphatic stem *?mtfa, while the stem for the rest of the Shuswap paradigm is derived from the proto-Salish second person singular stem *nawi. 132 APPENDIX E JULIENNE'S L A N G U A G E S A M P L E S=Secwepemctsin; E=English; M=Mixed languages # Language Utterance Age 1 S mama, kectsetsme tk cereal 6;3;0 2 E I got us 3 plates 6;3;0 3 E If you got more than one, you gotta say "syelts" 6;3;0 4 S re 7nwi7 tsxwente sk'epqen ell re sq'wext [pick me up by the head and legs] 7;4;6 5 S Mom, wiwkten re 7nwi7 te kelc 7;0;6 6 E You mean you're gonna grab me? 6;3;0 7 S kectsetsme te stemstete7stem [give me a drink] 6;3;0 8 S qwerimimen tek pop 6;3;0 9 s Mama, kectsetsme te ntsetswe7 te stemstete7stem 6;3;0 10 s Mama, ntsetswe7 meqLken 6;3;0 11 s Mama, kectsetsme ntsetswe7 tek orange juice! 6;3;0 12 s ki7ce, k7ep re ntsetswe7 6;3;0 13 E Outside? 6;3;0 14 E We're gonna make it? 6;3;0 15 M I pul'te e cmeye 6;3;0 16 M I wi7 secwem 6;3;0 17 E Wanna play? 6;3;0 18 M wanna seyse? 6;3;0 19 S ta7a, met'em-ken 6;3;0 20 E Do you want money 7;0;6 21 E Daddy wants money 7;0;6 22 E I want money 7;0;6 23 E We will go to the store 7;0;6 24 E I wil go to the store 7;0;6 25 E I love you 7;0;6 26 E I love my Mommy 7;0;6 27 E she puts away 7;0;6 28 E we will run 7;0;6 29 E I will jump 7;0;6 30 S tkensteke... ntsetswe7 6;10;15 31 S tkensteten nqewten 6;10;15 32 s tkensteke re sq'ewxt 6;10;15 33 s tsut te Simon, tkensteke re sq'exwt < ' 6;10;15 34 s tkensteten nsq'exwt 6;10;15 35 E mama, can we play a different game? 6;10;15 36 s tsut te Simon, tkensteten te qw'u! [laughts] 6;10;15 37 s tkensteten nq'wu? 6;10;15 38 M mama! [dances and sings] 6;10;15 39 M huh? 6;10;15 40 S est'il... 6;10;15 133 41 S est'il ntsetswe7 6;10;15 42 s . ce! 6;10;15 43 s ntsetswe? 6;10;15 44 E Can can can we do the can can [sings] 6;10;15 45 E I'm doing my buttons... 6;10;15 46 S qwetseq? [die] 6;10;15 47 E But you know most of the words, mama! 6;10;15 48 M OK, tsut re Simon, te... How do you say sleep again? 6;10;15 49 S eti7ce 6;10;15 50 S 10 hours! 6;10;15 51 E Good night, I'm tired. Good night Mama, Good night Sandra. 6;10;15 52 S. ta7a ta7a 6;10;15 53 S ta7a 6;10;15 54 s ta7a... ntsetswe7 6;10;15 55 s te ntsetswe7 6;10;15 56 s -ken. Can I? 6;10;15 57 s ta7a 6;10;25 58 s temtumen re 7nwi7 ell Sandra 6;10;25 59 s wecqinmects te 7qewten 6;10;25 60 s tsiqw ell yect re 7qewten 6;10;25 61 s me7e 6;10;25 62 s ta7a 6;10;25 63 s me7e 6;10;25 64 s ta7a 6;10;25 65 s me7e 6;10;25 66 s me7e 6;10;25 67 s me7e 6;10;25 68 s ta7a 6;10;25 69 s ta7a 6;10;25 70 s ta7a 6;10;25 71 s ta7a 6;10;25 72 s ta7a 6;10;25 73 s ta7a 6;10;25 74 s ntsetswe7 6;10;25 75 s me 6;10;25 76 s ta7a 6;10;25 77 s ta7a 6;10;25 78 s ntsetswe7 6;10;25 79 s swetikre... 6;10;25 80 s me7e 6;10;25 81 s ntsetswe7 6;10;25 82 . E Mama, can we have a- 6;10;25 83 s ta7a 6;10;25 84 E you 6;10;25 85 s 7nwi7 6;10;25 86 s 7nwi7 6;10;25 87 s ta7a 6;10;25 88 E m-hm, but not, but this is boring. 6;10;25 134 89 E Can we do something else? 6;10;25 90 E It's up to you. 6;10;25 91 E and Carrie 6;10;25 92 E Veronika. I know, let her come over all the way from the other country to Canada. 6;10;25 93 E u-uh! What airplane would you take to get to Canada? 6;10;25 94 E Katie, from grade one 6;10;25 95 S mus 6;10;25 96 M me7e, cause Katie. 6;10;25 97 S me7e 6;10;25 98 M mama! 6;10;25 99 S me7e, ell equipment 6;10;25 100 M re stemins re Katlin re Carrie in grade two 6;10;25 101 M ell Natlie in grade three... ell re Samantha Gufferson in grade sesele 6;10;25 102 M Samantha play re appointment? 6;10;25 103 En appointment 6;10;25 104 M today? uhm.. seysus 6;10;25 105 S seyse - seysus, seyses, seysus. 6;10;25 106 M nschool-nschool-nschool 6;10;25 107 E My fun time! [reads from book] 6;10;25 108 E mama, why do I... board games? 6;10;25 109 S -sus? 6;10;25 110 S 7nwi7? 6;10;25 111 s ta7a 6;10;25 112 s ta7a 6;10;25 113 E uh, nobody 6;10;25 114 s re 7tsetse 6;10;25 115 M Meghan 6;10;25 116 M Jo Thomas 6;10;25 117 M tsetse? Lizzie. 6;10;25 118 M Meghan 6;10;25 119 E Ron. How did you know? 6;10;25 120 S Jessica ell Meghan 6;10;25 121 E I don't know 6;10;25 122 E Meghan. Why can't we do things that- 6;10;25 123 S me7e 6;10;25 124 M Lizzie [very unwillingly] 6;10;25 125 S weyt-k ki7ce 6;11;16 126 S me7 eti(t)ce 6;11;16 127 s me7e 6;11;16 128 s me7e 6;11;16 129 s sesele 6;11;16 130 s ntsetswe7 ta7a qwenmin yi7ene 6;11;16 131 s ntsetswe7 ta7a qwenmin yi7ene 6;11;16 132 s ta7a k s qwenmimen 6;11;16 133 s yi7ene sesele 6;11;16 134 s qweqwi7ts [seqw7its =rabbit] 6;11;16 135 s keknem [kenkeknem = blackbear] 6;11;16 135 136 S nek'u 7 sesele kelles mus tsilkst teqmekst 6;11;16 137 S ts'um'qsente re kenkeknem rnmamrnmarrurima... 6;11;16 138 s me7e, ntsetswe7 6;11;16 139 s 4cen 6;11;16 140 s ta7a 6;11;16 141 s oh, xwistsin, ki7ce [hugs and kisses mom good night] 6;11;16 142 s me7e 6;11;16 143 s ntsetswe7... yi7ene 6;11;16 144 s me7e 6;11;16 145 s kelles! nek'u7 sesele kelles, one two three 6;11;16 146 s nek'u7 te? 6;11;16 147 s sesele! 6;11;16 148 s kelles! 6;11;16 149 s nek'u7 sesele kelles mus tsilkst 6;11;16 150 s teqmekst! 6;11;16 151 s nk'u7 sesele kelles mus tsilkst - teqmekst nek, nek7ups, nek7ups! 6;11;16 152 s tem7emtsuta! [!!! means "temllemkuk7e"] 6;11;16 153 s teqmekst 6;11;16 154 s seqw'iyts! 6;11;16 155 E they're... nibbling 6;11;16 156 s illenten 6;11;16 157 s te kelc 6;11;16 158 s te awyiit 6;11;16 159 s te? 6;11;16 160 s illenes 6;11;16 161 s w7ec 6;11;16 162 s illenes te - 6;11;16 163 s illente te - 6;11;16 164 s w7ec te book te - 6;11;16 165 s w7ec re illen te - 6;11;16 166 s illenes te wohu 6;11;16 167 s ta7a 6;11;16 168 s ta7a 6;11;16 169 s illen 6;11;16 170 s. w7ec illenes te, te, mayan ell... te pumpkins 6;11;16 171 s w7ec te illenes te - 6;11;16 172 s w7ec re illen 6;11;16 173 s w7ec re (shows arms swimming) 6; 11; 16 174 s re sexwumes 6;11;16 175 s pesi 6;11;16 176 S " w7ec seysemus 6;11;16 177 s seysus 6;11;16 178 s w7ec seysemus 6;11;16 179 s seysus, seysus! 6;11;16 180 s w7ec seysemus 6;11;16 181 s seysus 6;11;16 182 s ell re qwas- re qwose- i . . . sey... illens te qmut ell te eqwe [koso?] 6;11;16 183 s me7e 6;11;16 136 184 S te me7... illen... 6;11;16 185 S te illenes te sulensum 6;11;16 186 S te sqexe re illenes re seysus.,.mus [seysus/seysemus] [the horse eats and plays?]. 6;11;16 187 S ntsetswe7 xwexwistsin 7enwi7 6;11;16 188 E Spice your life up! He puts on his jackets 7;2;28 189 M no, weytk 7;2;28 190 S mama? 7;2;28 191 E Yeap. The spice girl picture 7;2;28 192 S uh, ntsetswe7 Mens te... 7;2;28 193 S illen 7;2;28 194 s ntsetswe7 illen te st7illen 7;2;28 195 E yeap 7;2;28 196 M mama? 7;2;28 197 E um, the second one. My turn, OK. The picture of Spice Girls 7;2;28 198 S re Spice girl te pi7ture 7;2;28 199 M megcen te... picture, I like that one, sometimes it shows, sometimes it doesn't 7;2;28 200' S nts'sqexe te picture 7;2;28 201 M you say, ntsetswe7 7;2;28 202 S ta7 k s qwenmin 7;2;28 203 S -mens. ntsetswe7 te q'imeke7 7;2;28 204 s ta7 k s qwemimen ntsetswe7 te q'imeke7 7;2;28 205 s -ken 7;2;28 206 s setsinem-k? 7;2;28 207 s setsinem 7;2;28 208 s setsinems, Sandra 7;2;28 209 s -ce, Sandra 7;2;28 210 E One-horse-open-sleigh! You came close to the camera! 7;2;28 211 s setsinem-ken, -ken 7;2;28 212 E Can we do something else? 7;2;28 213 E uh, no! How about I say it in English and you say it in Shuswap? 7;2;28 214 E Here, Sandra, here! You hold it! 7;2;28 215 E OK, she picks this, it, up [ungrammatical on purpose for fun] 7;2;28 216 E No, Sandra, I'll hold it now 7;2;28 217 E OK, Now make a funny face. 7;2;28 218 E OK 7;2;28 219 E mm, yes you can! 7;2;28 220 E Yes you can. It means I see him., mm, which 7;2;28 one? Was it a girl or a boy? 221 E Oh, It means, it see, I seen you. 7;2;28 222 M ntsetswe7... Hey! I have to say it in Shuswap? 7;2;28 223 M OK. Now you tape me. ntsetswe7 ell 7nwi7 ell re s-tsetse... 7;2;28 224 S ntsetswe7 ell tsetse 7;2;28 225 S lecus 7;2;28 226 M yeah, lecusa? 7;2;28 227 S lecuskt 7;2;28 228 S ntsetswe7 ell re Lizzie leciis- 7:2:28 137 229 S -ken 7;2;28 230 M -kucw. I said that before! 7;2;28 231 S ntsetswe7 ell re Sandra lecus-ken 7;2;28 232 S -ken 7;2;28 233 S ntsetswe7 el re lizzie lecus... 7;2;28 234 S -kucw 7;2;28 235 E Let's switch [camera] now 7;2;28 236 E I put it [should be I feed it] 7;2;28 237 M I feed the horse? OK. ntsetswe7, say ntsetswe7 7;2;28 238 S ntsetswe7, uh, me- me- 7;2;28 239 M I feed, me, me-teten, re nsqexe 7;2;28 240 S me7e 7;2;28 242 S sesele 7;2;28 241 S nwi7-s nts'qexe [points ahead to her poster of unicorn, points to another pic of horse] nts'qexe, nts'qexe 7;2;28 243 E Mommy, Can you make one of those? 7;2;28 244 S me7e, re cake 7;2;28 245 S ntsetswe7- 7;2;28 246 S yiri7ntsetswe7te... 7;2;28 247 S n-cake 7;2;28 248 S n-stsq'ey 7;2;28 249 E Look! That's the symbol of the God! 7;2;28 250 S Nicola ell te Hercules. 7;2;28 251 S yiri7 ntsetswe7 te q'imeke7 7;2;28 252 E How do you say that's my spice Girls poster? 7;2;28 253 S Jo-Thomas ta7a n-tsetse 7;2;28 254 S nuxwenxw 7;2;28 255 S ta7a 7;2;28 256 S me7e! 7;2;28 257 E Spice up your life! 7;2;28 258 E No. That's what Frosh spice she says, blah blah, spice world, uhm, spice, yeah, I think so she says blah blah blah; spice world, or spice girl. 7;2;28 259 E Yeah I know, I look like the Backstreet boys and Spice Girl 7;2;28 260 E Like both. 7;2;28 261 E Can we use the camera and play a different game? 7;2;28 262 E Sandra, why don't you do something 7;2;28 263 E No 7;2;28 264 E OK, last one 7;2;28 265 S 7nwi7-mommy- 7;2;28 266 E Oh, Hi, Jo-Thomas! 7;2;28 267 E OK. Hi, Sandra! 7;2;28 268 S ta7an- 7;2;28 269 S 7nwi7, ta7a n-qwenmins te - 7;2;28 270 E How do you say book, mama? 7;2;28 271 S spe- 7;2;28 272 S qweltcw 7;2;28 273 S me7e 7;2;28 138 274 E You may go to recess now. 7;2;28 275 E This is the first time I'm on colour TV! 7;2;28 276 S Me7 texin te JoThomas 7;2;28 277 E where? 7;2;28 278 E Where's stop? 7;2;28 279 E We won't do any hard ones. Now get a blank piece of paper 7;2;28 280 E You brought 10? 7;2;28 281 E I found it. But it's a whole sentence! 7;2;28 282 E heart 7;2;28 283 S p ii smen 7;2;28 284 E Oh, rub 7;2;28 285 E Just one word! 7;2;28 286 E Get, the, controller, pen? Mommy, no more sentences! 7;2;28 287 E No more sentences OK? Sandra? 7;2;28 288 E mmm, let's see let's see let's see 7;2;28 289 S Oh, metsin! 7;2;28 290 E Give! 7;2;28 291 E horses! 7;2;28 292 E But you said no more sentences! 7;2;28 293 E Daddy- 7;2;28 294 E feeding. 7;2;28 295 S ntsetswe7 7;4;6 296 E I forget how to say want 7;4;6 297 S qwenmin- 7;4;6 298 S ntsetswe7 qwenmin te dance 7;4;6 299 E No 7;4;6 300 E No 7;4;6 301 S ntsetswe7 qwenmin te dance 6;10;15 302 S kectsetsme re ts'u m'qsentsme 7;4;6 303 E I don't know what you're gonna play. It's up to you [understands kucw=sandra and mom, but not Julienne] 6;10;25 304 S kectsetsme te ste7 [give me a drink] 7;0;6 305 S piitucw 7;4;6 139 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items