A descriptive grammar of Koho-Sre: a Mon-Khmer language

Kơho, a Mon-Khmer (Austroasiatic) language, is spoken by an indigenous population of more than 207,000 people located in Lâm Đồng province in the highland region of Vietnam. There are also several thousand additional members of this ethnic group who live in France and the United States (primarily North Carolina). The goal of this dissertation is to describe the Kơho-Sre language in such a manner that it is accessible both to linguists and also to those in the Kơho-speaking community interested in their own language. This grammar-based on a linguistic analysis that is informed by current linguistic theory and best practices in the field-includes phonological, morphological, and syntactic data. A grammatical description of Kơho is needed, in spite of the fact that a literature of the language does exist. This is because (1) adequate documentation is not achieved by the extant literature; (2) materials are dated and do not reflect recent advances in typology and linguistic analysis; (3) many materials are published in Russian and Vietnamese or are not readily available to most researchers; and (4) earlier descriptions are cast in frameworks that are not amenable to contemporary documentary linguistic analysis. This dissertation, based on data collected during fieldwork in Vietnam and North Carolina, supplemented with previously published syntactic and lexicographic materials, provides an overview of the grammatical structure of Sre. Sre is a polysyllabic (usually dissyllabic) language with a synchronic tendency towards reduction of the presyllable (the weaker or minor syllable) and development in the remaining (main or major) syllable of contrastive pitch characteristics associated with vowel length. Vowel length, in turn, is influenced by the main syllable coda. A formerly complex system of nominal classifiers (operating in the pattern: numeral + classifier + noun) has been reduced to three generally used classifiers. Sentence structure is subject + verb + object with a fairly rigid word order with some phrase or clause movement to indicate certain syntactic functions.

A DESCRIPTIVE GRAMMAR OF KƠHO-SRE: A MON-KHMER LANGUAGE by Neil Hayes Olsen A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Linguistics The University of Utah December 2014 Copyright © Neil Hayes Olsen 2014 All Rights Reserved The Univers i ty of Utah Graduate School STATEMENT OF DISSERTATION APPROVAL The dissertation of Neil Hayes Olsen has been approved by the following supervisory committee members: MaryAnn Christison , Chair May 19, 2014 Date Approved Marianna Di Paolo , Member May 21, 2014 Date Approved Timothy Chambless , Member May 11, 2014 Date Approved Lyle Campbell , Member June 17, 2014 Date Approved Mark Alves , Member June 17, 2014 Date Approved and by Edward Rubin , Chair/Dean of the Department/College/School of Linguistics and by David B. Kieda, Dean of The Graduate School. ABSTRACT This dissertation is a descriptive grammar of the Sre dialect of the Kơho language. Kơho, a Mon-Khmer (Austroasiatic) language, is spoken by an indigenous population of more than 207,000 people located in Lâm Đồng province in the highland region of Vietnam. There are also several thousand additional members of this ethnic group who live in France and the United States (primarily North Carolina). The goal of this dissertation is to describe the Kơho-Sre language in such a manner that it is accessible both to linguists and also to those in the Kơho-speaking community interested in their own language. This grammar-based on a linguistic analysis that is informed by current linguistic theory and best practices in the field- includes phonological, morphological, and syntactic data. A grammatical description of Kơho is needed, in spite of the fact that a literature of the language does exist. This is because (1) adequate documentation is not achieved by the extant literature; (2) materials are dated and do not reflect recent advances in typology and linguistic analysis; (3) many materials are published in Russian and Vietnamese or are not readily available to most researchers; and (4) earlier descriptions are cast in frameworks that are not amenable to contemporary documentary linguistic analysis. This dissertation, based on data collected during fieldwork in Vietnam and North Carolina, supplemented with previously published syntactic and lexicographic materials, provides an overview of the grammatical structure of Sre. Sre is a polysyllabic (usually dissyllabic) language with a synchronic tendency towards reduction of the presyllable iv (the weaker or minor syllable) and development in the remaining (main or major) syllable of contrastive pitch characteristics associated with vowel length. Vowel length, in turn, is influenced by the main syllable coda. A formerly complex system of nominal classifiers (operating in the pattern: numeral + classifier + noun) has been reduced to three generally used classifiers. Sentence structure is subject + verb + object with a fairly rigid word order with some phrase or clause movement to indicate certain syntactic functions. To Gloria, my guardian angel Thank you for your patience, support, and love. TABLE OF CONTENTS ABSTRACT ....................................................................................................................... iii LIST OF TABLES ........................................................................................................... viii ABBREVIATIONS AND SYMBOLS .............................................................................. ix ACKNOWLEDGMENTS ................................................................................................. xii CHAPTERS 1 INTRODUCTION ......................................................................................................... 1 1.1 Introduction to the study ............................................................................. 1 1.2 The Kơho people ......................................................................................... 6 1.3 The Kơho language ................................................................................... 10 1.4 Scientific significance of this dissertation ................................................. 18 2 PHONETICS AND PHONOLOGY ........................................................................... 23 2.1 Introduction ............................................................................................... 23 2.2 Phonological description ........................................................................... 23 2.3 Syllable structure ....................................................................................... 30 2.4 Suprasegmentals ........................................................................................ 33 2.5 History of orthographies ............................................................................ 34 3 MORPHOLOGY ......................................................................................................... 41 3.1 Introduction ............................................................................................... 41 3.2 Derivational morphology .......................................................................... 41 3.3 Clitics ........................................................................................................ 50 3.4 Compounding ............................................................................................ 53 4 SYNTACTIC CATEGORIES: WORD CLASSES .................................................... 54 4.1 Introduction ............................................................................................... 54 4.2 Lexical categories: content word classes .................................................. 55 4.3 Functional categories: grammatical word classes ..................................... 74 vii 5 SYNTACTIC STRUCTURES: GRAMMATICAL COMPONENTS ........................91 5.1 Introduction ................................................................................................91 5.2 Sre word order ............................................................................................91 5.3 Simple sentences ........................................................................................93 5.4 Complex sentences .....................................................................................98 5.5 Nominal constituents ...............................................................................102 5.6 Negation ...................................................................................................104 APPENDICES A TEXT: TRADITIONAL VILLAGE WORK .............................................................106 B SWADESH 285-ITEM WORD LIST FOR SOUTHEAST ASIA: KƠHO-SRE DIALECT .............................................................................................108 REFERENCES ................................................................................................................122 LIST OF TABLES 2.1 Consonant phonemes .............................................................................................38 2.2 Vowel phonemes ....................................................................................................38 2.3 Vowel minimal pairs with glosses .........................................................................39 2.4 Examples of permitted Sre main syllable patterns .................................................40 2.5 Sre pitch length by subdialect ................................................................................40 4.1 Example of common nouns ...................................................................................85 4.2 Personal pronouns ..................................................................................................85 4.3 Numerals ................................................................................................................86 4.4 Inventory of Kơho classifiers cited in this section .................................................87 4.5 Coverbs and their glosses .......................................................................................88 4.6 Deictics of location and direction ..........................................................................88 4.7 Temporal units with dəәʔ ‘one of a number of recurring or multiplied instances, or repeated acts; a recurrent event' ........................................................................89 4.8 Temporal units with tuʔ ‘time, occasion' ..............................................................89 4.9 Formation of interrogative pronominals ................................................................90 5.1 Content question interrogatives ...........................................................................105 5.2 Negative elements with ʔαːʔ ‘not' .......................................................................105 ABBREVIATIONS AND SYMBOLS ~ alternates with; alternate form [x] phonetic form /x/ phonemic form * ungrammatical form < derives from > changes into <…> orthographic representation of an original source à becomes # word boundary § section (of a chapter) 1 first person 2 second person 3 third person ADJ adjective BEN benefactive C consonant CAUS causative CLF classifier CMA Christian and Missionary Alliance COM comitative x CONJ conjunction DAT dative DIS distal EXCL exclusive EXIST existence F, f feminine GEN genitive Hz Hertz IMP imperative INCL inclusive INS instrumental IPA International Phonetic Alphabet lit. literally LOC locative M, m masculine MACV Military Assistance Command-Vietnam msec millisecond N noun NEG negative NMLZ nominalizer/nominalization O object PASS passive PL plural xi POSS possessive PREP preposition PROX proximal/proximate PURP purposive PW phonological word Q question marker RECP reciproal RED reduplication REFL reflexive S subject σ Greek letter sigma = syllable SIL Summer Institute of Linguistics (now SIL International) s.o; s.t. someone; something SVO subject + verb + object V, v verb; vowel VOT voice onset time WALS World Atlas of Language Structure WH WH-type question word ACKNOWLEDGMENTS Grammars do not write themselves. Many people who have been involved in influencing, encouraging, nurturing, and mentoring me over the last five decades must be acknowledged for their aid and support during the writing of this grammar. First, I must acknowledge the people who introduced me to the scholarly study of language and culture: †Robert M. Ariss and Paul L. Kirk. In Vietnam, I am deeply indebted to †David and Dorothy Thomas of the Summer Institute of Linguistics, then in Nha Trang, who focused my research and connected me with a support group of fellow researchers. I wish to thank members of the Kơho-speaking community that so graciously gave their time and knowledge of their language under very difficult circumstances during the American War in Vietnam: Uok Cil, K'Wa, K'Bris, and K'Smal. When I pursued graduate studies at the University of Hawai‘i at Mānoa, I benefited from classes taught by Robert ‘Bob' Hsu, †Philip Jenner, †Nguyễn Đăng Liêm, Steve O'Harrow, and Laurence C. Thompson. †Nguyễn Đình Hòa, Steve Sherman, Paul Sidwell, and †William Smalley and shared valuable expertise from their respective fields of endeavor. In North Carolina, my friends and language consultants from the Kơho community there have again patiently and willingly taught me the nuances and of their language. I must thank Broi Toploi, K'Ben Kon Sa, and Uok Cil, all of whom survived the war and now are flourishing with their families in the United States. Most recently, at the University of Utah, I would like to acknowledge the members of my graduate committee: Maryann Christison (chair), Lyle Campbell, xiii Marianna Di Paolo, Tim Chambless, and Mark Alves. These scholars, colleagues, and friends have been wonderfully patient and so supportive of my efforts to document the Kơho language. I owe them my gratitude for believing in my work and me. Lastly and most importantly, I must acknowledge the love and support given to me by my wife, Gloria, while I was pursuing my passion. Without her patience and understanding, I would not have accomplished this task. CHAPTER 1 INTRODUCTION 1.1 Introduction to the study The aim of this dissertation is to provide a descriptive grammar of the Sre dialect of the Kơho-speaking peoples in a theory-neutral format that will be accessible to linguists and other academics. Hopefully, this grammar can serve as a resource for pedagogical purposes and revitalization efforts, should the community desire them. The Kơho,1 who call themselves kon cau, number approximately 200,000 people and inhabit most of Lâm Đông province in highland Vietnam; several thousand now live overseas in France and the U.S.A. They are one of approximately fifty ethnic minority groups living in Vietnam; these people are also known as montagnards (an older French term) or highlanders. Kơho, along with the Chrau, Mnong, and Stieng languages, comprise the South Bahnaric branch of the Mon-Khmer (Austroasiatic) language family. This chapter presents the theoretical framework of the study (§1.1.1), a brief description of how fieldwork was conducted (§1.1.2), and an introduction to the Kơho people (§1.2). A brief discussion of the language classification (§1.3.1), typology 1The ethnonym Kơho [kəәә'hɔ] is derived from a Cham word that refers collectively to a group of several peoples speaking mutually intelligible dialects in the southern part of the highlands of Vietnam (Olsen 1968, 1976). Although, Mà is linguistically a Kơho dialect, it is considered by the Mà community and ethnologists to be a separate ethnic group. The ISO 639-3 code for the Kơho language is kpm; the code for Mà is cma. 2 (§1.3.2), and dialects (§1.3.3) follows. A discussion of the scientific significance of this dissertation (§1.4) concludes the chapter. 1.1.1 Theoretical framework and terminology The purpose of this study is to present a detailed linguistic description of the phonology, morphology, and syntax of the Sre dialect of the Kơho language. In writing the grammar, I used Comrie and Smith's 1971 Lingua questionnaire to organize my data and materials. Besides various vocabulary lists, basic bilingual phrase books, and folkloric works in or about the Kơho, there are only a few lexicographic and syntactic materials of note on this language. In addition to my own data collected in fieldwork with the language, I have also consulted Evans and Bowen (1962) and Manley (1972) with supplemental reference to two important dictionaries (Dournes 1950, updated and revised in Bochet and Dournes 1953; and Drouin and K'Năi 1962). Best practices used in documentary linguistics guided the preparation of this grammar. 1.1.2 Fieldwork and data collection During the course of ethnographic field research among the Kơho-speaking peoples in the Central Highlands region of Vietnam, I had the opportunity to record textual material from two of the groups in the area-the Nồp and the Sre. The resultant corpus comprises speech representing three speakers of Sre, and one of Nồp; from this resource, the idiolect of one of the Sre-speakers was selected as the main focus of this analysis. This particular idiolect was selected because it had the most documentation (see §1.3.3). 3 The following is an inventory of my field notes from 1967-1968: • 1967-8. Unpublished field notes. Cited as Olsen Field Notes (OFN). [red plastic notebook, 3 x 5 in. pages, ‘Central, No. 1270. Made in U.S.A.'] Photocopy made. • 1967-8. Notes on Kơho grammar and dictionary. [red notebook, 8 x 12.5 cm; Cây Tre-Nhụt-Ký printed on cover] • 1968. Progress Report on Research and Field Work Among the Kơho, a Montagnard Tribe of South Viet-Nam. Djiring (Di-Linh), 15 Feb 68, 5 p., typescript. • 1968. Report No. 2. Blao (Bao Loc), 21 March 68, 2 p., typescript. • 1968. Reference data. Blao (Bao-Loc), 6 May 68, 2 p., typescript. All the field recordings were copied, enhanced, and archived at the University of Hawai‘i Language Learning Center in 1974 (Tape # LIF 109.1). In addition, in 2008, selected records were digitized in a split-head (consumer stereo) format at radio station KUER, University of Utah. The following is an inventory of digital and acoustic phonetic (spectrograms) materials generated at the University of Hawai‘i at Mānoa between 1975-1978.2 1. Inventory of spectrograms from the 1967-1968 field recordings (351 spectrograms), 1974. 2. Concordance for Manley's Outline of Sre Structure, 26 p., (222 types/1407 tokens), including frequency count in descending order. Reverse concordance, 23 p., 1 July 1976. 2These materials are in my personal field notes collection. 4 3. Sre (Evans and Bowen and Manley) concordance, 157 p. (1,062 types/8,787 tokens), 7 May 1977. 4. Nop concordance, 14 p., (162 types/773 tokens); reverse concordance, 14 p., 30 June 1976. 5. Sre-English dictionary, 272 p., (12,160 lines, 1,654 main entries, 1,128 subentries); English-Sre finder list, 119 p., (4,695 lines, 1,978 main entries, 19 May 1978. (This is an ongoing project.) An inventory of my papers presented at conferences and other relevant publications is listed in the References section. 1.1.3 Prior research on Kơho The Sre and other Kơho peoples are described in a moderately comprehensive ethnographic literature. This material deals primarily with their folkways and socio-cultural organization; treatment of the language has not been extensive.3 The following is an inventory of the relevant materials available on the language (those involving grammar, dictionary, and analyzed texts). There are several published works involving the grammar: one of the earliest, a 1959 publication in Vietnamese, is Học tiếng Thượng: Kơho [Learn a highland language: Kơho] issued by the Nha Công-tác Xã-hội miền Thượng [Office of Highland Social Work]; the second is Kơho Language Course by Helen Evans and Peggy Bowen (1962); the third is my brief grammar sketch (15 p.) for use by military advisers (Olsen 1968); the fourth is a syntactic study in a case grammar framework (Manley 1972); the fifth is a 3An annotated bibliography of both ethnographic and linguistic references concerning the Kơho-speaking peoples is in preparation by the author (Olsen n.d. a). 5 treatment of affixation by Nguyễn văn Hoan (1973); and finally, there is Lý Toàn Thắng, et al. (1985) Ngữ pháp tiếng Kơho [Kơho grammar]. Kơho lexicography is adequately represented in four major works of note. The first is Jacques Dournes' Dictionnaire srê (köho)-français, published in 1950 using the 1949 orthography (subsequently revised: Bochet and Dournes 1953, a quadrilingual lexicon edition); the second is the more recent and larger (multivolume, 1512-page) work by Father Sylvère Drouin and Professor K'Năi of the Centre Montagnard de Rédemptoristes, the Dictionnaire français--montagnard, which appeared in 1962;4 the third is the trilingual Kơho Vocabulary by Ha Bul Sohao of the Summer Institute of Linguistics (1976).5 The fourth, and most recent, is a Vietnamese-Kơho dictionary by Hoàng Văn Hành, et al. Từ Điển Việt-Kơho, issued in 1983.6 Phonology is the least documented area in Kơho. The classic work is Smalley's important Srê phonemes and syllables (1955). Although Manley devotes a chapter to phonology in his grammar (1972:10-39), it is not a main focus of his research. Despite his emphasis on syntax, Manley does offer several new insights; he further delineates the various Srê dialects. Duong Tan Le's thesis (2003) explores phonological comparisons in Kơho and Mà. There is a recent phonological study from Vietnam, Ngữ âm tiếng Kơho [Kơho phonology] by Tạ Văn Thông (2004). A comprehensive bibliography of published materials on Kơho may be found in the references section following the appendices. 4This work documents the Ryông Tô dialect. 5This work documents the Cil dialect. 6Additionally, I have compiled a Kơho-English lexicon (with over 2,000 entries to date). 6 A grammatical description of Kơho is needed, in spite of the fact that a number of works are listed in this section. This is because (1) adequate documentation is not achieved by the extant literature; (2) materials are dated and do not reflect recent advances in typology and linguistic analysis; (3) many materials are published in Russian and Vietnamese or are not readily available to researchers; and (4) earlier descriptions are cast in frameworks that are not amenable to contemporary documentary linguistic analysis. 1.2 The Kơho people Kơho [kəәˈhɔ] is a Cham ethnonym applied collectively to a group of several peoples speaking mutually intelligible dialects in the Western Highlands (Tây Nguyên) of Vietnam.7 The Kơho autonym kon cau [kɔn caw] is usually translated as ‘sons of men,' but literally means ‘offspring of humans.' Although 207,517 ethnic Kơho were enumerated in Vietnam in 2009 (Vietnam Central Population and Housing 2010), with about 2,000 in North Carolina (U. S. Census 2000), there are no comparable figures for the current number of speakers. In Vietnam, Kơho speakers inhabit most of Lâm Đồng province, and portions of Bình Thuận, Ninh Thuận, Đòng Nai, and Đắc Lắc provinces. Major population centers in the Kơho-speaking area are Đàlạt, Bảo Lộc, and Di Linh. Minority ethnic groups in 7‘Koho est un mot cham qui servait à désigner l'ensemble des Montagnards et qui se trouve pratique pour désigner un dialecte commun à un nombre relativement important de tribus (de parler Môn-Khmer) et compris sans exception dans toute la province du Haut- Donnai' (Bochet 1951). In Moussay's Dictionnaire Căm-Vietnamien-Français (1971:125), the entry reads: ‘KAHO /kahauw/: người Koho; la peuplade des Koho.' The Kơho are an officially designated ethnic group in Vietnam. Other Vietnamese spellings of the ethnonym are Cơho, Cờ-ho, Co-ho, and Kohor. 7 Vietnam, including the Kơho people, have been known by several other designations, including French montagnard ‘mountain dweller,' and Vietnamese người thượng ‘highland compatriot.' See Figure 1.1. Since the 1930s, missionaries, government agencies, and educators using several different alphabets have produced scripture, primers, grammars, and dictionaries. During the 1960s and 1970s, a series of pedagogical materials in Kơho and a dozen other minority languages was produced by SIL (Summer Institute of Linguistics) under contract to the former Saigon government-the Highlander Education Project. Writing primers, science, and health books were used in many classrooms where Kơho was the language of instruction in the primary grades; in the higher grades, Vietnamese was phased in.8 During the American War in Vietnam, most of the heavier fighting was located outside of the Kơho-speaking area, but that war had an effect on these people in several significant ways. Because most of their homeland was designated as a ‘free fire zone' in the late 1960s,9 many Kơho underwent abrupt social and cultural upheaval when entire village communities were forcibly relocated into strategic hamlets strung along National Route 20 and clustered around district and provincial centers between Bảo Lộc and Dàlạt 8Steve Sherman, of the RADIX Foundation, Houston, TX, was able to obtain electronic copies of SIL's Vietnamese minority education materials. They were previously available only on microfiche from SIL. These materials have been shared with the North Carolina community. 9‘Free fire zones' were areas designated (often arbitrarily) by the South Vietnamese or American forces as being de facto ‘insecure' (i.e., enemy territory) and thus subject to indiscriminate and unlimited artillery fire, aerial bombing, and chemical defoliation. 8 (Olsen 1970 and Volk 1979). Many innocent people who remained in their native villages were decimated along with their crops and livestock as a result of these zones. Many highlanders served as soldiers in the Republic of Vietnam's Regional and Popular Forces, which were province- and district-level home guard units, respectively. Others were recruited into Civilian Irregular Defense Groups (CIDG), and other ‘special operations' which were administered by the U.S. Special Forces (the Green Berets). At the conclusion of the Second Indochinese War, what remained of the highlander population tried to pick up the pieces of their lives and rebuild. After reunification in 1976, the Hanoi government initiated a massive program where lowlander Kinh (people of Vietnamese ethnicity) were relocated to many of the highland provinces, making highlanders a minority in their homeland (Hardy 2003). There was severe deforestation as highlanders and Kinh alike destroyed the double- and triple-canopy forests to plant coffee, which, at that time, was enjoying a worldwide economic boom. Unfortunately, Vietnamese coffee flooded the market and prices plummeted, leaving many people economically impoverished.10 At the same time, the government and military were conducting mop-up operations against remnants of the former Saigon administration. Highlanders who had sided with the Saigon and American authorities were imprisoned in reeducation camps, many for decades. Some died from miserable conditions or were executed. The military rounded up guerilla groups comprised of former South Vietnamese military personnel, a highlander autonomy movement called 10See Gerard Greenfield, Vietnam and the world coffee crisis, Urban Renaissance Institute website, 1 March 2002. http://www.urbanrenaissance.org/urbanren/ index.cfm?DSP=content&ContentID 057. Accessed 9 August 2009. 9 FULRO,11 and other antigovernment elements. A modicum of peace finally came to the highlands in 1992, when FULRO holdouts in Cambodia surrendered to a United Nations team. Since then, the government has aggressively persecuted and prosecuted evangelical Protestant home churches, which have many highlander adherents (Human Rights Watch 2006). Because of this situation, thousands of highlanders, including Kơho people, have fled their homeland. Since 1986, several thousand highlanders have been resettled in the United States, mostly in North Carolina. After 1975, and especially following unification in 1976, use of minority languages was discouraged and a Vietnamization campaign emphasized integration of minority peoples into the majority national society. Vietnamese became the language of instruction at all levels of education. All pre-1975 materials were banned. During fieldwork in Vietnam in 1999, I observed that minority languages were not being nourished in the ‘old country.' Realistically, unless Vietnamese government policies change, the overseas Kơho are ‘on their own' as far as language and cultural maintenance emanating from the homeland are concerned. Despite the fact that there is a population of over 200,000 in Vietnam alone, Kơho (like Navajo12) has the potential to become an endangered language. A challenge to overseas Kơho will be maintaining a ‘critical mass' of speakers and transmitting their language to future generations. As mentioned, most Kơho refugees from Vietnam, Cambodia, and United Nations camps in Thailand were able to relocate to North Carolina. See Figure 1.2. This is because Ft. Bragg, the home of 11FULRO is the acronym for Front Uni de Lutte des Races Opprimées (United Front of Struggle for the Oppressed Races). 12Navajo is considered at risk because fewer young people are learning the language despite the demographic size of the group-more than 100,000 (Mary Ann Willie, p.c.). 10 the Special Forces, is located near Fayetteville, North Carolina. Many active duty and retired Green Berets sponsored individuals and families so they could immigrate to the U.S. and reestablish themselves in this country. The Kơho community in North Carolina is flourishing and adapting to American ways while maintaining their culture and language. The families of my language consultants all speak Kơho on a daily basis and even their grandchildren are using the language. This bodes well for Kơho language maintenance in North Carolina. 1.3 The Kơho language Among the Kơho-speaking peoples, Sre has emerged as the prestige dialect because they inhabited areas that were developed during the French colonial period. An important highway, National Route 20, that connects Hồ Chí Minh city (Sàigòn) and Đàlạt, was built through the middle of their territory. Because of this accessibility, the Sre were one of the first Kơho peoples to be converted to Christianity by French and, later, American missionaries. 1.3.1 Language classification Kơho, along with Chrau, Mnong, and Stieng comprise the South Bahnaric branch,13 which in turn is a subdivision on a par with the North Bahnaric and West 13Other synonyms include the terms Mnong-Ma and Chau-Ma (or Che-Ma), and appear primarily in French scholarly works (Dournes 1974:161). French ethnographers and linguists often include the Bölöö (Bơlờ) [bǝ.lǝː], a Stieng people, along with Mnong under their Mnong-Ma rubric. The situation is further complicated by the fact that although one may refer to the Kơho language as such of the Kơho-speaking peoples, there was, historically, no group of people specifically referred to, or that considered themselves as Kơho. After reunification, this changed as the ethnic Vietnamese (Kinh) refer to these people as Kơho 11 Bahnaric branches of the Bahnaric group, a major division of the Mon-Khmer language family (Thomas 1966:194-197). The Mon-Khmer languages, the Munda languages in India, and the Aslian languages of the Malay Peninsula comprise the three components of the Austroasiatic family. The basic vocabulary among the four South Bahnaric languages all range around a 60-percent cognate rate, so that no further subdivision is suggested. The South Bahnaric languages are physically separated from North and West Bahnaric by a geographic intrusion of Austronesian-speaking Chamic peoples; it is thought that the split of Bahnaric into northern, western, and southern groups is historically connected with the appearance of people speaking Chamic languages (not only Cham, but Eđê (Rhadé), Jơrai (Jarai), Rơglai, and others) into the Western Highlands region (Thurgood 1999).14 Recently, sufficient materials have become available on at least one speech community of each of the four South Bahnaric languages so that several scholars have reconstructed proto-South Bahnaric.15 The Sre have much in common with other highland minority peoples in Vietnam, both culturally and ethnically, and despite some linguistic differences, consider themselves along with all the highland groups to be a unified socio-cultural entity. Although the Sre speak a Mon-Khmer (Austroasiatic) language, they are culturally more (also spelled K'Ho and Cơ ho), which is now an official term; many highlanders adopted it when speaking about themselves to outsiders. 14The Austronesian intrusion of Chamic languages into the Western Highlands is detailed in Gregerson, et al. (1976) and Thurgood (1999), while the South Bahnaric areal linguistic setting is discussed in Thomas (1971:18-21). 15Proto-South Bahnaric has been reconstructed by Efimov (1988, 1990), and Sidwell (2000). 12 similar to Austronesian-speaking Chamic peoples because of the intensive political dominance by Champa in the highland region until the eighteenth century. Thus, while the Sre speak a dialect that is mutually intelligible to their western neighbors, the Mà, they differ from them both in social structure and subsistence patterns (Olsen 1976).16 Many Sre are multilingual, possessing a degree of fluency in either French or Vietnamese, depending on their age, education, and military service. Some also speak neighboring languages, such as Mnong, Rơglai, Chru, or Eđê (Rhadé), and more recently English. Traditionally, the Sre17 engaged primarily in wet-rice (paddy) agriculture and lived in the plains and valleys of the Di Linh Plateau (Cao Nguyên Di Linh), which is located in central Lâm Đông province, and comprises the southern-most portion of the Western Highlands (Tây Nguyên) region in central Vietnam.18 The social and cultural 16The Cham influence on Sre social structure and linguistic behavior is very pervasive. The Sre reckon descent matrilineally, as do most Austronesian-speaking peoples in Vietnam, unlike many Mon-Khmer-speaking groups. 17Surprisingly, the ethnolinguistic literature refers to these people by the same name that they use for themselves: (cau) sre ‘irrigated paddy (people).' Alternate citations for the Sre include: Sơre, Cau S're, Srê, and Xrê. For brief ethnographic sketches of the Sre, see Queguiner (1943), and Le Bar, et al. (1964:156-157). 18After unification in 1976, Lâm Ðông and Tuyên Đuc provinces were merged into one administrative unit, retaining the former's designation. During the former Republic of Vietnam administration, the provincial capital of Lâm Ðông was Bảo Lộc (sometimes transliterated as Bơlao, from the indigenous Blao). The province capital is now Dàlạt. The Western Highlands (Tây Nguyên) are also known as the Central Highlands (Cao Nguyên). 13 center of Sre territory is Djiring (Di Linh),19 the principal market town in the district. Population estimates for the Sre account for approximately 30,000 people in Vietnam. Kơho speakers began arriving in North Carolina in 1986; they settled primarily in Charlotte, Greensboro, and Raleigh. Another group joined them in 1992. Both state and local government and refugee services have provided excellent support to the relocated people. Many of the Kơho have moved into various professions, such as realtors, lawyers, restaurateurs, and religious leaders. Most of the children and grandchildren of the people I met speak Kơho and use it at home and in many domains. The younger people talk and text in the language. This bodes well for the maintenance of the Kơho language in North Carolina, which includes the Sre, Cil, and Lat dialects. This study will primarily focus on the Sre dialect as spoken in North Carolina. 1.3.2 Typology As mentioned above, Kơho, along with the Chrau, Mnong, and Stieng languages comprise the South Bahnaric branch of the Mon-Khmer (Austroasiatic) language family. Both Khmer (Cambodian) and Vietnamese are distantly related to the Bahnaric languages, and thus Kơho. Alves, in his Pacoh grammar (2006), a Katuic language, lists typological characteristics that are found in Mon-Khmer (Austroasiatic) languages, noting that some of these also apply to other languages and language families in the Southeast Asia Linguistic Area (1-3).20 These include: 19Di Linh (sometimes written Gi Rinh) is the Vietnamese transliteration of Djiring (which is the French rendering of the Sre place name Ñjring [ɲ̩ ˈɟriŋ], the district seat until 1975. 14 • Sentence structure: topic-comment, a basic subject-verb-object (SVO) order • Noun phrase structure: classifier language, noun-modifier word order, lexically indicated plurality (not affixes) • Modal aspects: time (or tense), aspect, and the interrogative are indicated by adverbs and sentence particles (not conjugated verbs or affixes) • Word-formations: presyllables (e.g., deriving causative verbs) and infixes (e.g., nouns derived from verbs), partial/alternating reduplication, no suffixes • Phonological word shape: sesquisyllabic word structure in which presyllables are unstressed and somewhat reduced in form, consonant clusters on main syllables • Phonology: vocalic/register differences on vowels, four-way place of distinction of consonants (labial, dental, palatal, and velar) These genetic features and areal traits are discussed in detail in the phonology, morphology, and syntax chapters that follow. Hawkins (1983:284,338) classifies Sre, the most documented dialect of Kơho, as a subject-verb-object word order language with adpositions and numeral-noun, noun-adjective (alternatively adjective-noun is a nonbasic order), noun-genitive, and noun-relative clause characteristics. Sre is language Type 9 in his Extended Sample typological classification, expanding on Greenberg's language universals works (Greenberg 1966 [1963], 1978).21 20See also Goddard (2005:33) for an inventory of distinctive phonological and morphological characteristics in the Mon-Khmer languages. 21Other Type 9 languages include Cambodian, Vietnamese, Modern Israeli Hebrew, Indonesian, most Romance languages, and almost all Bantu languages. 15 Syntactically, Kơho is similar in many respects to Vietnamese and other Southeast Asian languages. Sentential word order is generally subject-verb-object (SVO). Kơho is a polysyllabic (usually dissyllabic) language with a synchronic tendency towards reduction of the presyllable (the weaker or minor syllable) and development in the remaining (major) syllable of contrastive pitch characteristics associated with vowel length. Vowel length, in turn, is influenced by the main syllable coda. Since the mid-twentieth century, a formerly complex system of nominal classifiers (operating in the pattern: numeral + classifier + noun) has been reduced to three generally used classifiers, which function in the environment previously filled by the more specialized classifier terms (Manley 1972:120-121). In addition, a diachronically intricate affixation pattern that previously served syntactic functions seems to be currently nonproductive, although stylistic or metaphoric extensions of certain forms (usually verbs) may be generated by the use of polysyllabic compounds composed of an even number of syllables. Verbal aspect may be optionally marked by atelic or telic situation markers (Comrie 1976:44- 48). Personal pronouns distinguish three persons, singular and plural (the second person has masculine, feminine, and informal forms; the first person plural has an inclusive/exclusive dichotomy). Demonstratives are categorized by physical or temporal distance and visibility with reference to the speaker and listener. Interrogative sentences are marked by appropriate final particles; prefixing an assimilating nasal to a nominal substantive forms many of the WH-type interrogatives.22 22WH-type interrogatives indicate the type of questions that begin with wh- in English (e.g., who, what, why, etc.). For a discussion of other syntactic material in Kơho, see Olsen 1968, Manley 1972, Nguyễn văn Hoan 1973, and Lý Toàn Thắng, et al. 1985. 16 Approximately one-fifth of the basic lexicon in Kơho is cognate with neighboring Austronesian languages (especially S. Rơglai).23 Vietnamese, as the national language, exercises a considerable influence on Sre, and the other minority languages in Vietnam. In North Carolina, the influence of Vietnamese is not as significant, but Vietnamese is often used as a lingua franca in intergroup communication, especially between speakers of other minority languages. The influence of English on Kơho in North Carolina is beyond the scope of this study. The Kơho language has several intriguing features of interest to language typology and linguistic theory. These features include a system of diachronically diverse noun classifiers, intricate patterns of reduplication, and a set of chameleon lexemes that appear in form as adpositions, but function as relator nouns. 1.3.3 Dialects24 There are at least twelve Kơho dialects: Chil (Cil, Til), Kalop (Tulop), Kơyon (Kodu, Co-Don), Làc (Làt, Lach), Mà, Nồp (Nop, Xre Nop, Noup), Pru, Ryông Tô (Riồng, Rion), Sop, Sre, Talà (To La), and Tring (Trinh). Linguistically, I consider Mà to be a Kơho dialect, since Mà dialects are mutually intelligible with other Kơho dialects. Mà subdialects include: Cau Mà (Chau Mà), Mà Huang, Ngan, Preng, Cop (Xop), Krung, and (Cau) Tô. This list of dialects and subdialects is not comprehensive or definitive. The 23About 20 percent of the Sre lexicon examined (in this and other material) contains words cognate with Chamic languages, especially in the domains of kinship, politics, and agriculture. 24I am using the term dialect as defined by Di Paolo and Spears (2014:11). 17 ethnographic and linguistic literature presents considerable variation on the designation or existence of the various dialects and subdialects.25 As previously mentioned, Sre is considered the prestige dialect among most Kơho speakers. Sre is one of the main languages of commerce used in the area and is employed in radio and television broadcasts in Kơho-speaking areas (Dương 2003:31). In his fieldwork, Manley noted the existence of three distinct subdialects within the Sre dialect which he terms subdialects A, B, and C. Both subdialects A and B are spoken in Di Linh, apparently in contiguous territories within the town.26 Speakers of subdialect A have inhabited the area for a long time. Because of this, its speakers are wealthier and better educated, so socially subdialect A is considered the prestige dialect (see §1.3 above). Subdialect B is spoken by a group that migrated from north and west of Di Linh into the area in the mid-1900s. Although speakers of subdialect B have intermingled with subdialect A speakers, subdialect B speakers still maintain the characteristics of their dialect. These two subdialects are mutually intelligible and exhibit only minor phonetic and lexical differences. Subdialect B, although not the prestige subdialect, is the basis for materials produced by the Christian and Missionary Alliance, including the translation of the New Testament. Pedagogical materials issued by the Highlander Education Project used an orthography based on subdialect B and that subdialect was taught in local schools. Smalley documented subdialect C, spoken to the south of Di Linh, in his 25 I wish to thank William Labov and Gillian Sankoff for an enlightening discussion concerning languages, dialects, and subdialects. 26Both Manley and Nguyễn văn Hoan based the majority of their research on the speech of Broi Toploi (K'Broi), who is from Di Linh. Subdialect A is his first language. He is also the prinicipal resource for this grammar. I want to acknowledge K'Broi's assistance in delineating the geography of these subdialects. 18 important article ‘Sre Phonemes and Syllables' (1955). Manley notes that subdialect C ‘is apparently different still from [sub-]Dialects A and B-more different, in fact, than A is from B' (1972:11-12, 15, 18-19). This grammar is based on subdialect A, but topics of interest in subdialects B and C will be noted when relevant. Most Kơho speakers in North Carolina are fluent in at least two or three other dialects, in addition to English, Rhade, Jarai, and sometimes Spanish. A few older people are familiar with or speak French. My language consultants told me that although people will converse in different dialects with each other, when they write something, they use Sre, the prestige dialect.27 1.4 Scientific significance of this dissertation This dissertation addresses two areas of scientific significance. The first is that there is no contemporary, theory-neutral syntactic description of any South Bahnaric language.28 The Kơho-Sre literature is introduced above. Of the four South Bahnaric languages, Thomas's Chrau Grammar (1971) is the most accessible syntactic description despite being cast in SIL's tagmemic framework. Both Mnong and Stieng have been adequately documented, but the materials are in scattered articles and diverse journals. The materials have not been collected into a single source with syntactic description in a contemporary theory-neutral format for either language. As mentioned, most publications 27A church service I attended in North Carolina was conducted primarily in Vietnamese with some English interspersed because the congregation consisted of speakers of seven different languages. 28This assertion could be extrapolated to most Mon-Khmer languages, with the possible exception of Khmer (Cambodian) and Vietnamese. 19 from SIL linguists, especially in Vietnam, are cast in tagmemics, which is opaque for theoretical or typological analysis. The second area of scientific significance is not limited to Mainland Southeast Asia languages, but nonetheless is a major problem in linguistic description-agreement on terminology. The terminology conundrum is especially vexing in Mon-Khmer linguistics because the phonological systems in many of the languages are complex and require sophisticated theoretical apparatus to describe the phenomena observed adequately. Both Alves (1997) and Schiller (1999) comment on this situation, noting that problem areas in typical ‘Western' analyses of Southeast Asian languages include ‘inadequate linguistic theories that continue to analyze these languages incorrectly despite the reality of Southeast Asian languages and other languages of the world' (Alves:1). In syntax, there are serious terminological problems in defining concepts such as pronouns, classifiers, and adpositions. Both of these areas of significance (and concern) are addressed in this dissertation. As in any academic endeavor, a descriptive grammar is a hypothesis-or better said, a compilation of interacting hypotheses, which can be challenged, rejected, or refined if researchers (which can include the author of the grammar himself) find new data which suggest different analyses, or if it is found that other possible analyses offer a more enlightening account of the data on hand. The later can happen if modifications in linguistic theory provide more insightful ways of looking at particular phenomena in the language. Documentary linguistics contributes to the science of linguistics and should be held to the same standards and ethics as any discipline. Finally, as Kơho presents linguistic traits that are of significant general 20 typological interest, their importance is emphasized and the phenomena are described in detail in this dissertation, thereby contributing to typology generally. 21 Figure 1.1. Map of Vietnam showing the location of the Kơho people. 22 Figure 1.2. North Carolina showing cities (underlined) where Kơho people live. CHAPTER 2 PHONETICS AND PHONOLOGY 2.1 Introduction The purpose of this chapter is to provide an overview of Sre phonetics and phonology. The first section is a phonological description of Sre (§2.2), followed by a discussion of syllable structure (§2.3). Next, the section on suprasegmentals covers the phenomenon of pitch length (§2.4), and finally a brief history or summary of orthographies (§2.5). Because there are five orthographies extant for the Kơho language and to avoid confusion (and consternation), all data in this grammar are transcribed using the International Phonetic Alphabet (IPA). 2.2 Phonological description Data from Kơho-Sre language descriptions (Manley 1972; Nguyển văn Hoan 1973; Smalley 1955; and Tạ Văn Thông 2004) and my field notes (Olsen field notes and 1968) are used to examine selected phonological phenomena. This section is divided into the phonological inventory of Sre (§2.2.1) and selected phonological processes (§2.2.2). 2.2.1 Phonological inventory Sre phonology patterns similarly with the sound systems seen in other Mon- Khmer languages. Section 2.2.1.1 details the consonant system and §2.2.1.2 discusses Sre 24 vowels. 2.2.1.1 Sre consonants Sre has consonant segments at five places of articulation with voiceless unaspirated and aspirated stops, voiced stops, implosives, fricatives, unaspirated and aspirated nasals, trills, laterals, and glides (Table 2.1). Sre has a series of unaspirated, aspirated, and voiced stops at labial, alveolar, palatal, and velar points of articulation, plus a glottal stop. Two implosives are found at labial and alveolar points of articulation. There are two voiceless alveolar and glottal fricatives. Unaspirated nasals occur at labial, alveolar, palatal, and velar points of articulation. Aspirated nasals occur at labial, alveolar, and palatal positions. According to Manley's Table 3a (1972:14, 26), no aspirated velar forms have been found in the data. There is an unaspirated and aspirated alveolar trill and an unaspirated and aspirated lateral. Finally, there are two glides: a high, back, rounded labial and a high, front, unrounded palatal. Before the palatal finals /c/ and /ɲ/, there is an audible palatal offglide after the vowel [Vj], so that /pwac/ ‘flesh' is pronounced as [pwaʲc] and /ʔaɲ/ ‘I,' 1st person singular' as [ʔaʲɲ]. The voiced alveolar trill [r] often reduces to a flap [ɾ] when it occurs as the second segment in a consonant cluster (e.g., [sɾe] ‘paddy field'). There have been several interpretations of whether aspirated nasals and liquids (/l/ and /r/) should be considered as units segment or as a segment plus /h/ (Smalley 1955:222). Smalley considers /ph, th, ch, kh/ as two segments (CC) rather than as single unit aspirated stops [ph, etc.] because of parallels in /mh, nh, ɲh, lh, rh/: /mhar/ ‘quickly,' /nhap/ ‘covered,' /ɲhat/ ‘drop (n.),' /lhaʔ / ‘asleep,' and /rhjaŋ/ ‘one hundred.' I prefer a 25 different interpretation because of the implications of a morphophonological ‘test,' one of infix insertion, employed by Richard Watson in his description of Pacoh29 phonemes (Watson 1964:141, §1.2.1.2). He posits that if clusters can be divided by morphological infixes as aspirates can, then the clusters should be interpreted as two segments (i.e., CC). For example, the nominalizing infix {-an-} in Pacoh may be inserted between the first and second segments of a verb root (with allophones /-al-/ and /-ar-/) (see 1a-c). (1) a. c<al>leang (stem: cleang) Pacoh <NMLZ> door bar ‘a door bar' ‘to bar door' b. c<ar>rong (stem: crong) <NMLZ> s.t. surrounded ‘something surrounded' ‘to surround' c. k<an>(h)iar (stem: khiar) <NMLZ> broom ‘a yard broom' ‘to sweep yard' Given the data in (1a-c), Watson therefore interprets aspirated stops (in Pacoh) as clusters of stops /p, t, k/ plus /h/ because they parallel the pattern of stops /p, t, k/ plus liquids /r, l/. He notes that aspirates, like clusters, often occur in only main-syllable-initial position (C-), whereas unit consonants can also occur in presyllables-initial and word-final position. Alves, in his Pacoh grammar (2006), observes that such insertion appears only in fossilized remnants and is no longer active in the language (p. 21). He considers the single segment interpretation to be a reflection of the typological tendency in Southeast Asia toward onset cluster reduction. 29Pacoh belongs to the Katuic branch of Mon-Khmer; it is spoken primarily in Thừa Thiên province in central Vietnam. 26 Applying Watson's test to the Sre data, it appears that no infix can be inserted between a word-initial consonant and its aspirated component as in (2a-d). (2) a. p*<infix>han (stem: phan [pʰan]) ‘things' b. m*<infix>ham (stem: mham [mʰam]) ‘blood' c. l*<infix>haʔ (stem: lhaʔ [lʰaʔ]) ‘asleep' d. r*<infix>hjaŋ (stem: rhjaŋ [rʰjaŋ]) ‘one hundred' This observation supports the interpretation of Sre aspirated consonants as being considered single segments (although they are represented in most orthographies as digraphs: <ph th ch kh>, etc.). Manley, in his grammar, concurs with this interpretation, noting that [t]he elements in question function like other single consonants in the language, not like clusters. (For example, voiced stops and implosives can all precede r and l, but never precede h …). (1972:36fn1) The aspiration in nasals and liquids will sometimes have a very faint centralized vowel just prior to aspiration (e.g., rᵊhjaŋ ‘hundred'). Additionally, voiceless stops may or may not have a slight aspirated release in morphemes uttered in citation form; this variation occurs in several of my recorded word lists. Examination of spectrograms generated from my field data confirm acoustically that the aspiration of voiceless stops is evidenced by a voice onset time (VOT) that is, on average, two to three times longer than the VOT for unaspirated stops. VOT is the 27 interval between the plosive burst and the onset of vocal fold vibrations (Ladefoged 2003:96). The distributional phonology of the consonant segments is detailed in the syllable structure section (§2.3). 2.2.1.2 Sre vowels Sre vowels are phonemically categorized by four primary features: height, backness, length, and tentatively, by tongue root position triggered by allophonic overlap. The front vowels are unrounded, while back vowels are rounded. The central vowels vary widely as to lip rounding and height, but mostly occur as allophones of /əә/, especially as presyllables vowels. The high, central vowels [ɨ, ɯ] occur very rarely in the data. The vowels are delineated in Table 2.2. Table 2.3 lists minimal pairs, if they occur in the data. Several vowels almost always occur long [eː, oː, αː], except in some personal names (e.g., [doh] ‘female personal name'). Additionally, there appear to be several processes operating within the vowel system that are difficult to discern and describe. A brief attempt to document these follows. Manley (1972:15-16, 18) discusses allophonic overlap in the high and mid-high front vowels, where the high front vowel /i/ has [i] and [ɪ] as allophones, while the mid-high front vowel /e/ allophones are [i], [ɪ], and [e̝]. In a minimal pair, where /i/ and /e/ contrast, the allophones can be indistinguishable as far as tongue height alone is concerned. For example, in the words /ntiːŋ/ ‘bone' and /nteːŋ/ ‘where', both vowels are long, high, tense, and front [ɪː]. He notes that speakers will invoke a slight tongue-root 28 advancement [ɪ̘ː] to distinguish between the two overlapping allophones.30 Manley notes that in dialect B, allophones of /o/ overlap with allophones of /u/, so an advanced tongue-root (ATR) [u̘] is triggered to disambiguate that overlap. It appears that some of the dialects of Sre are gradually evolving a nascent system to disambiguate overlapping allophones using advanced tongue root. This process has been documented in other Mon-Khmer languages as well. Manley also observed that the vowels /e, o, α/ form a subset that almost always occur long and are never articulated with advanced tongue root (1972:16-17). Exploring the implications of both of the use of advanced tongue root to disambiguate allophonic overlap and the subset of vowels /e, o, α/ are subjects for future research (see Olsen 2008). 2.2.2 Phonological processes These selected phonological processes illustrate the types of rules employed to preserve the fundamental syllable template in Sre. Most of these are degemination or deletion rules. Geminate consonant clusters are not permitted; these clusters reduce to a single consonant. Nasals and liquids are not permitted in consonant clusters. If there are two adjacent nasals, then the first is deleted (3a). If two coronals (dental or alveolar) occur in succession, then the second cannot be a lateral (3b). Elsewhere, the nasal coda of /təәn-/ 30I was able to confirm in subdialect A the advanced tongue-root acoustically by examining my spectrograms of that sound; it is indeed a very subtle phenomenon even when analyzed by instrumental means. The third formant of the vowel remains at the same frequency throughout the utterance while the second formant bends down at the beginning of the vowel sound. The same pattern has been found in African ATR vowel spectrograms. 29 undergoes partial contact regressive assimilation with the following stop (3c-d). (3) a. /təәn + muːʔ/ > [təә.muːʔ] CAUS + go.down CAUS-go.down ‘go down' ‘cause/make s.o. or s.t. go down' b. /təәn + lik/ > [təә.lik] CAUS + come.out CAUS-come.out ‘come out' ‘cause/make s.o. or s.t. to come out' c. /təәn + cʰəәt/ [təәɲ. cʰəәt] CAUS + die CAUS-die ‘die' ‘cause s.o. to die; kill' d. /təәn + kah/ [təәŋ.kah] CAUS + remember CAUS-remember ‘remember' ‘cause s.o. to remember' Phonemically, a palatal glide may not follow another palatal consonant (4). (4) a. /*cj- *chj- *ɟj- *ɲj- *ɲhj-/ b. [+palatal]* +palatal +glide ! " # $ % & Syllable-final glottal stop following a long (marked) vowel /v̀ʔ/# ([Vːʔ]) is deleted when followed by another syllable (5a-d). However, it is not deleted following a short (unmarked) vowel (5e). The examples are from Smalley (1955:219). (5) a. ʔ > Ø / CV: ___# #C b. /dàʔ mɛ/ > [daː mɛ] ‘river' (lit. ‘water mother') c. /bòʔ daʔ/ > [boː daʔ] ‘that head' d. /dàʔ daʔ/ > [daː daʔ] ‘that water' 30 e. /luʔ daʔ/ > [luʔ daʔ] ‘that rock' There is one postlexical rule that frequently occurs because it involves the first person singular pronoun /ʔaɲ/ ‘I.' After the pronoun /ʔaɲ/, the genitive postposition /de/ is realized as /ɟe/ (6a-b). Alveolar /d/ progressively assimilates to a palatal /ɟ/ following the palatal nasal in /ʔaɲ/. (6) a. de > ɟe / ʔaɲ ___ b. sraʔ ʔaɲ ɟe book 1 GEN ‘my book' c. sraʔ kʰaj de book 3 GEN ‘his/her book' This palatal assimilation serves to distinguish and disambiguate the postposition /de/ from the homophone directional preposition /de/, since both forms may appear in the same sentence. The other personal pronouns (/mi, aj1, kʰaj/) are not affected by this rule (as in 6c). This is apparently a Proto-South Bahnaric external sandhi rule because it is also found in Chrau, Mnong, and Stieng. 2.3 Syllable structure Prosodically, syllabic organization is depicted in Figure 2.1. The Greek letter σ (lower case sigma) represents ‘syllable.' As in many Mon-Khmer languages, Kơho has two syllable types: a presyllable 31 and a main syllable.31 The prototypical full phonological word (PW) takes a sesquisyllabic (‘syllable and a half') form (Matisoff 1973:86). Thus, a word in Sre may be defined as consisting of a main syllable, optionally preceded by a presyllable (7). (7) PW = (presyllable) + Main syllable Metrically, such phonological words constitute an iambic construction of the type ( ̆ . --). 2.3.1 Presyllables The fundamental shape of a presyllable is illustrated in (8), (8) c1v1(c2)32 where the onset (c1) may be any unaspirated, unimploded obstruent, the nucleus (v1) is a central vowel /əә/, and the coda (c2) is a liquid or /n/. There are two other possible presyllables forms: /ʔa/, and a syllabic nasal that assimilates to the point of articulation of the first segment of the following main syllable (e.g., [m̩ ˈpaŋ] ‘foot,' see #93 in App. B). Presyllables tend to weaken or disappear in many environments. This presyllable weakening is apparently part of a diachronic process operating as an areal tendency towards monosyllabicity in the mainland Southeast Asia region that crosses language family boundaries (Thomas 1971:18-21). Vietnamese, a Mon-Khmer monosyllabic tone language, is an extreme example of this process. 31In Austroasiatic descriptive literature, a presyllable is also termed a weak or minor syllable; a main syllable may be termed a strong or major syllable (Thomas 1992). 32The use of lower case letters for presyllable symbols follows Wallace 1969. Main syllable symbols are indicated by upper case letters. 32 2.3.2 Main syllables The fundamental shape of a main syllable is detailed in (9), (9) C1(C2[C3])V2(C4[C5]) where the onset consists of C1(C2[C3]). C1 may be any consonant, C2 may be a glide or liquid (with constraints associated with C1). C3 can only be a glide (with constraints associated with C1 and C2). The nucleus (V2) may be any vowel with attendant pitch length. The coda is comprised of (C4[C5]). C4 can only be a glide, and C5 can only be /Ɂ/ (with C4 being either /j/ or /w/) or /h/ (with C4 being /j/). The permitted main syllable patterns are exemplified in Table 2.4. Although the majority of Sre words are monosyllabic (with an optional presyllable), there are also polysyllabic items in the lexicon. Disyllabic examples are illustrated in (10-12). (10) CV.CVC a. cəә.nɑːŋ ‘general term for household items (with a plane surface)' b. səә.nəәm ‘medicine' (11) CVC.CVC a. rəәn.dɛh ‘vehicle' b. səәr.daŋ ‘sugar' (12) CVC.CCVC səәn.djaŋ ‘steep side of a valley' (13) CV.CVC.CVC gəә.dan.ɲɔʔ (gəәjʔ) ‘uncontrolled laughter' 33 2.4 Suprasegmentals This section on suprasegmentals includes a brief observation of Sre speech and a summary of pitch length in the three dialects documented in this study. Concerning Sre speech, Smalley observed that Sre speech is characterized by a marked tenseness and preciseness of articulation. Phrasal groupings may be identified by final stress, and are bounded by space. A sharp syllable-timing, plus the allophones of tone-length, make syllable division fairly easy medially on the phonetic level (1955:218-219). All Sre vowels in subdialect A are phonemically normal in length (with level pitch, short, and unmarked) or long (with falling pitch and marked). Phonetically, Sre maps three vowel lengths onto two phonemic pitch length units:33 short/unmarked and long/marked. Sre subdialects are distinguished at one level by the differences in how each dialect handles pitch accent. The prosodic term MORA34 is used to indicate the relative duration of a segment-vowels in the case. Pitch length in the three Sre dialects is summarized in Table 2.5. Long vowels decrease in duration when they are not at the end of a word, particularly if unstressed. Sometimes they are as short as short vowels, but it is always possible to hear the difference because the pitch on long vowels is different from that of short vowels. 33Pitch length is also referred to as pitch-accent or tone-length in other works (Smalley 1955:218). 34A MORA is a term used to describe the length of segments. It is a relative measurement of metrical time equal to the duration of a short vowel or half a long vowel; it is used as a unit of pitch placement in a syllable. A long syllable or geminate consonants comprise two morae (or moras) (Carr 2008:103; Crystal 2001:222; Matthews 2007:251). 34 The following data are based on my spectrographic measurements of a speaker of subdialect A (recorded in 1968 from K'Bris, of Kao Kwil, a village east of Di Linh). Vowels, for the most part, tend to retain their pitch length characteristics, though attenuated, in polysyllabic compounds. The range of pitch rise and the height attained is conditioned by the position of the syllable within a word. In general, rising pitch tends to attenuate the further a syllable is from final position in a word. To illustrate, the pitch rise in the vowel [úː] of [dúːl] ‘one' in /ɟəәt duːl/ [ɟīt dúːl] ‘eleven' is +800 Hertz (Hz). 35 That is, the fundamental frequency increases by 800 Hz over the duration of the vowel. The pitch rise as the penult vowel in /duːl rhjaŋ/ [dú̞ˑɾhjāˑŋ] ‘one hundred' is +100 Hz. Level pitch is not affected by syllable position. While the pitch environment within a particular syllable conditions vowel length, relative vowel duration is also determined by the position of the syllable within a word. Generally speaking, duration decreases the further a syllable is from the final position in a word. For example, the vowel [i] in [n̩ˈtiː(ŋ)] ‘bone', in isolation, is 340 milliseconds (msec) in duration; as the (main syllable) vowel in [(n̩)tiˑŋ ɾəәˈpas] ‘rib', 170 msec, and in [nˈtiŋ gəәˑr ŋˈkɔˑj] ‘backbone' 130 msec. 2.5 History of orthographies In the preparation of a dictionary and a reference grammar for the Kơho language, a decision on which orthography to use is crucial. A romanized orthography based on the Vietnamese national alphabet (quốc ngữ) was developed in 1935 for the Sre dialect by 35HERTZ (Hz) is the current International System designation for cycles per second (cps). The change in the 10th harmonic of the vowel is used here to indicate the change in pitch. The 800 Hz measurement needs to be reanalyzed. It was measured manually on a spectrogram of a poor quality recording. 35 French colonial administrators and missionaries. That orthography, while festooned with diacritics (like Vietnamese), was the most consistently systematic alphabet (i.e., nearly phonemic) utilized to date. Both missionary and government documents were published in that alphabet. In 1949, a new orthography commission met in Đàlạt to devise an acceptable replacement (Martini 1952). Even among a newly literate people, attachment to a written tradition, however brief, attendant with a particular orthography, lingers on. Recent proposals have gained acceptance only with difficulty in some circles.36 Subsequent orthographies were proposed in 1953 and in the 1960s. The main controversy over the orthography arose between different factions of the Protestant (Tin Lành) churches. One group retained <î> to represent the phoneme /e/ in its publications (a hold-over from the 1936 alphabet); the other used <ê>, which was consistent with the orthography developed by the Christian and Missionary Alliance, in conjunction with the Summer Institute of Linguistics. The latter orthography was employed by the former Republic of Vietnam (Saigon) government for use in educational materials published by the Highlander Education Project. The 1967 New Testament and 1993 Psalms were published in that orthography. The complete Bible was published in 2010 in the CMA/SIL orthography. In North Carolina, the orthography employed depends on which church one is affiliated with. In Vietnam, after 1976, all previous (i.e., south Vietnamese) pedagogical materials were discarded. In 1983, the Vietnamese government introduced a 36I am indebted to the late William A. Smalley for documentation and a personal interview (Honolulu, 1975) to discuss this and related subjects. (In connection with this, see Smalley 1954 and also Manley 1972 [pp. 38-9].) To complicate the situation further, Evans and Bowen in their Kơho Language Course reverse two symbols, using <o> for /𝑎/ and <ọ> for /ɔ/, whereas in the standard orthography <o> represents /ɔ/ (1962:6a, 53- 54), while /𝑎/ (if it occurs in that subdialect) is written as <ọ> or <o>. 36 quốc ngữ-based orthography, which has essentially been rejected by Kơho living in North Carolina. 37 σ Onset Rhyme Nucleus Coda Figure 2.1 Syllable structure tree diagram 38 Table 2.1 Consonant phonemes ______________________________________________________________________ Labial Alveolar Palatal Velar Glottal ______________________________________________________________________ Voiceless unaspirated stops p t c k ʔ Voiceless aspirated stops pʰ tʰ cʰ kʰ Voiced stops b d ɟ g Implosives ɓ ɗ Fricatives s h Nasals m n ɲ ŋ Aspirated nasals mʰ nʰ ɲʰ (*ŋʰ) Trill r Aspirated trill rʰ Lateral l Aspirated lateral lʰ Glides w j ______________________________________________________________________ Table 2.2 Vowel phonemes ____________________________________________________ Front Central Back ____________________________________________________ High i ɨ ~ ɯ u Mid-high e əә o Mid ɛ ɔ Low a α37 ____________________________________________________ 37The low back vowel [α] ‘alpha' occurs in Sre subdialects A and B, but not C. 39 Table 2.3 Vowel minimal pairs with glosses _____________________________________________________ Sre vowel Example English gloss _____________________________________________________ i ʔis ‘alone' iː ʔːis ‘hang up clothes (to dry) e re (wəәl) ‘return (home)' eː reːp ‘(be) near, close; beside' ɛ rɛ ‘swim' (v.) ɛː bɛːp ‘father' a ʔaɲ ‘I (1st person singular)' aː ʔaːɲ ‘carry on back with shoulder straps' ɯ (~ ɨ) lɯɲ ‘gums' ɯː (occurs primarily in personal names; e.g., [mɯːh] ‘male personal name') əә bəәs ‘snake' u ɟun ‘carry' uː ɟuːn ‘a kind of deer' o po ‘suck, feed at breast' oː poːn ‘hide' (v.) ɔ dɔ ‘this (here)' ɔː dɔː ‘wear s.t. on hand or arm' αː ʔ αːʔ ‘no, not (negation marker)' 40 Table 2.4 Examples of permitted Sre main syllable patterns ____________________________________________________ Sequence Form Gloss ____________________________________________________ CV sa ‘eat' CVC gir ‘catch fish with a basket' CVCC ləәwɁ ‘fold' CCV sre ‘irrigated rice paddy' CCVC blah ‘split' CCVCC glaːjʔ ‘satisfy, expiate' CCCV <not attested> --- CCCVC krjaŋ ‘a kind of hard wood' *CCCVCC <not attested>38 --- ____________________________________________________ Table 2.5 Sre pitch length by subdialect 38No examples of the maximal main syllable template have been found in the Kơho materials available. CHAPTER 3 MORPHOLOGY 3.1 Introduction Morphology is the study of the grammatical properties of words and how they relate to one another in a language. Because morphology studies word relationships, it determines what procedures a given language needs to create new words. As such, morphology interacts with all the other components of the grammar-syntax, phonology, and semantics. Morphology is traditionally subdivided into roots or stems and inflectional and derivational affixes, with word formation closely related. There is very little evidence of inflectional morphology (i.e., paradigms) in Sre, Kơho dialects, or other Mon-Khmer languages because of the diachronic trend towards monosyllabicity in many of these languages. Vietnamese is the prime example of a Mon- Khmer language with very little inflection; it is considered to be almost totally analytical and isolating. The second basic process is word formation; it has two components: derivational morphology (§3.2) and compounding (§3.3). 3.2 Derivational morphology Inflectional morphology involves such phenomena as noun inflection and verb conjugation. The Latin language is an example of the extensive use of inflectional 42 morphology. However, addition of inflectional affixes does not change the basic meaning of the root word. Derivational morphology, on the other hand, may involve affixes that change the grammatical category of a root, modify the transitivity (valence) of a verb root, or create a diminutive or distributive form of the root. 3.2.1 Affixation Sre affixation presents several challenges. It is not always obvious whether a particular affix is productive synchronically or is a fossilized or lexicalized form. Some affixes have homophonous forms that perform differently from the source of the affix. It is difficult to discern just which syntactic categories (word classes) are involved because the boundaries are often fluid. This is especially true in distinguishing between affixes and clitics. The affixes and clitics listed in this section are not exhaustive by any means; this is definitely an area that warrants deeper analysis. 3.2.1.1 Prefixes Sre prefixes are generally attached to verbs, but a few do act on other syntactic categories (e.g., nouns). It is often difficult to distinguish between a prefix and a presyllable that is part of the root because of phonological similarity (i.e., homophones) and the volatile or unstable nature of any segments that precede a main syllable. 3.2.1.1.1 Causative prefix təәn-. One of the more productive prefixes in Kơho-Sre is təәn-. This prefix converts intransitive verbs to causative verbs. The derived verb has the meaning of ‘causing someone or something to do something unintentionally,' or ‘to make happen to someone or something else.' Allomorphs are created by phonological rules 43 triggered by the initial consonant of the verb (or noun) root. 39 The rules delineated in the phonological chapter (§2.2.2) pertain to affixes as well. The pattern in (14) serves as a template to illustrate the sentential changes that occur after the prefix is added to the root. Examples (15-18) show regular derivational morphology. (14) a. kʰaj git gəә 3 know it ‘S/he knows it.' b. ʔaɲ təәŋ-git gəә kʰaj ʔin 1 CAUS-know it 3 DAT ‘I cause (or make) him to know it.' (15) cʰəәt təәɲ.cʰəәt ‘to die' ‘to make to die; kill' (16) duh təәn.duh ‘to be hot' ‘to make hot' (17) re: təәn.re: ‘to leave; go home' ‘to make someone go home' (18) sɔh təәn.sɔh ‘to wear, to dress' ‘to dress someone else' The nasal cluster avoidance rule (3a) operates in (19-22). (19) mu:ʔ təә.mu:ʔ (*təәn-mu:ʔ) 40 39There is an allomorph pəәn- that sometimes varies freely with təәn-; no fixed pattern or environment has been determined to explain its occurence. Chrau has a cognate causative prefix ta- (Thomas 1971:70-1). 44 ‘to go down' ‘to lower; to make go down' (20) mut təә.mut (*təәn-mut) ‘to go in, enter' ‘to make go in' (21) ndaw təәn.daw (*təәn-ndaw) ‘to wear on the head' ‘to put on someone's head' (22) ŋac təә.ŋac (*təәn-ŋac) ‘to be well (goodbye)' ‘to wish one well; to farewell someone' The nasal-lateral cluster rule (3b) operates on (23), but does not affect (24-25) for reasons that are as yet unclear. (23) lik təә.lik (*təәn-lik) ‘to go out' ‘to make to go out' (24) leː təәn.leː ‘to melt, dissolve' ‘to make dissolve' (25) ljah təәn.ljah ‘to be short' ‘to make short' The nasal assimilation rule (3c-d) operates on (26-27). (26) gəәs təәŋ.gəәs ‘to have' ‘to beget; to have more' (27) kah təәŋ.kah ‘to remember' ‘to remind; to make to remember' 40Forms in parentheses with an asterisk (*) indicate output that has undergone a phonological rule that alters the affix configuration pattern. 45 Additionally, Nguyễn posits a related form təәr- that generally denotes a certain movement associated with a causative meaning (28) (1973:26; §4.3.1.4). (28) a. ʔaɲ rəәp aj3 təәrwəәl ŋaj-hiŋ. 1 ATELIC take CAUS.return tomorrow ‘I will cause it to be returned tomorrow.' b. Hiw dɔ təәrlah-təәrləәm ɟəәh house this collapse-RED completely ‘This house collapsed completely.' c. təәrlah-təәrliŋ to collapse, demolish-RED ‘to collapse completely 3.2.1.1.2 Causative prefix bəә-. The causative prefix bəә- is added to stative verbs to indicate ‘become the quality of' (29-30). This is similar to the English suffix -en which performs the same operation ‘to make, render (of a given character or quality)': hard > hard-en. (29) soŋ bəә.soŋ ‘straight' ‘to straighten (to cause to be straight)' (30) saːr bəә.saːr ‘hard' ‘to harden (to cause to be hard)' When prefixed to a transitive verb root, it has a causative meaning (31-32). (31) cah bəә.cah ‘to break' ‘to cause to separate, sort out' (32) kap bəә.kap 46 ‘to bite' ‘to cause to bite (press)' 3.2.1.1.3 Transitivizing prefix pəәn-. The prefix pəәn- (and an allomorph təәn-) is added to stative verbs: ‘[to] cause (someone) to have the quality indicated by the verb' (33-34) (Manley 1972:44-45), (33) riŋ pəәn.riŋ ~ təәn.riŋ ‘to.be.equal' ‘to make (s.o.) to be equal' (34) hαːp pəәn.hαːp ‘to.be.envious' ‘to make (s.o.) envious' but when the stative verb begins with a glottal, the alveolar feature of the coda /n/ in the prefixed pəәn- is progressively assimilated into the initial glottal segment of the main syllable (35-37) (Nguyễn 1973:27fn1). (35) ʔum ‘to bathe' pəәn.ɗum ‘to bathe (s.o.)' (36) ʔaːŋ ‘to be bright' pəәn.ɗaːŋ ‘to illuminate (s.t.)' (37) ʔja:ŋ ‘to be comfortable, at ease' pəәn.ɗja:ŋ ‘to put (s.o.) at ease' 3.2.1.1.4 Passivizing prefix gəә-.41 The prefix gəә- is added to monosyllabic verb roots. In Sre, when the predicate is an animate agent, that element is deleted (38b). When the predicate is an inanimate agent, that predicate is retained and preceded by a preposition (the instrumental preposition məә in this case) (39b). (38) a. Kʰaj paːʔ mpoːŋ. 3 open door ‘S/he opens the door.' 41The prefix gəә- has several homophones. In addition, the third person singular pronoun gəә1/2 also has several homophones, the analysis of which is beyond the scope of this study. See §4.2.1.2. 47 b. Mpoːŋ gəә-paːʔ Ø. door PASS-open ‘The door (was) opened.' (39) a. Caːl paːʔ mpoːŋ. wind open door ‘The wind opened the door.' b. Mpoːŋ gəә-paːʔ məә caːl. door PASS-open INS wind ‘The door was opened by the wind.' 3.2.1.1.5 Unresolved prefixes or clitics. There are several allomorphs of the prefix bɛ-: one is a possible similitive proclitic (or perhaps a prefix) added to determiners with the meaning ‘to be like' (40); the other is a possible interrogative proclitic which occurs with a few verb roots making them questions (41). (40) a. dɔ bɛ-dɔ ‘this' ‘like-this' b. ɲcʰi dɔ me cih bɛ-dɔ ‘what's this?' ‘you write like-this' (41) a. ləәh bɛ-ləәh ‘to make, do' ‘what to make or do?' b. kʰaj ləәh hiw bɛ-ləәh kʰaj ‘he builds the house' ‘what to do with him?' 48 3.2.1.2 Infixes Infixes are almost always inserted after the initial consonant of the base (onset C1). They generate cognate nominal elements.42 3.2.1.2.1 Nominalizing infix -əәn-. The infix -əәn- is inserted into transitive verbs. The derived noun has the meaning of ‘that which is the goal of the action indicated by the verb.' This infix has several allomorphs: -əәmp- and -əәrn-. The resulting cognate forms often have instrumental (42), result (43), or locational (44) meanings. After infixation, the derived form resyllabifies. The presyllable ʔa- is ignored by this infix; it operates on the initial onset of the main syllable (46). Exx. (42), (46), and (47) undergo phonological rule (3b). (42) plɛh pəә.nɛh (*p-əәn-lɛh) ‘to make way; withdraw ‘an obstruction to divert water' (43) pat p-əәn-at > pəә.nat ‘to knead, squeeze' ‘s.t. kneaded (clay, dough, etc.)' (44) sɛ s-əәn-ɛ > səә.nɛ ‘to turn, detour' ‘place where detour begins or ends' (46) ʔasuh s-əәn-uh > səә.nuh ‘to blow on a fire' ‘bellows' (47) blɔ bəә.nɔ (*b-əәn-lɔ) ‘to wear in the ear' ‘earring' (48) klɔ kəә.nɔ (*k-əәn-lɔ) ‘to hear' ‘to hear' 42English equivalents are: He slept a deep sleep. and He dreamed/dreamt a dream. 49 (49) par pəәn.ɗar (*p-əәn-ar)43 ‘to fly' ‘wing' (50) gar g-əәn-ar > gəә.nar ‘seed n.' ‘classifier for seeds, kernels' (51) tap t-əәn-ap > təә.nap ‘egg n.' ‘classifier for eggs' 3.2.1.2.2 Nominalizing infix -əәmp-. The infix -əәmp-is inserted into monosyllabic transitive verbs. The derived noun has the meaning of ‘that which is used in the action indicated by the verb.' In the available data, this infix only occurs in verb roots with initial /s-/.44 Additionally, no forms had high vowels [i, ɨ, u] in the root. (52) seːt s-əәmp-eːt > səәm.peːt ‘to plug up' ‘a plug' (53) saːc s-əәmp-aːc > səәm.paːc ‘to fish by draining' ‘the place one stands to fish by draining' (54) sɑːn s-əәmp-ɑːn > səәm.pɑːn ‘to wedge in' ‘a wedge' (55) sɔːl s-əәmp-ɔːl > səәm.pɔːl ‘to illuminate' ‘a torch' (56) soːc s-əәmp-oːc > səәm.poːc 43 See §3.2.1.1.3 above. 44The -əәmp- infix also occurs with initial /s/ in the Kơho dialect Ryong Tô, and other South Bahnaric languages: Chrau /set > s!̆pet/, and Stieng /seːt > sʌpeːt/. All meaning ‘to plug' and ‘a plug,' respectively. Marianna Di Paolo suggests that -əәmp- could be an instance of the coda /n/ in the nominalizing infix /-əәn-/ assimilating to the /-p-/ infix as seen in Ryong To, Chrau, and Stieng. 50 ‘to sting (of a bee)' ‘the sting (of a bee)' (57) sroːm s-əәmp-roːm > səәm.proːm ‘to sheathe (a sword)' ‘a scabbard' (58) sraŋ s-əәmp-raŋ > səәm.praŋ ‘to sting (of a fish)' ‘the stinger (of a fish)' 3.2.1.2.3 Nominalizing infix -əәrn-. The derived noun has the meaning ‘that which is used in carrying out the action indicated by the verb (59-60).' (59) kal k-əәrn-al > kəәr.nal ‘to bolt, bar' ‘a wooden door bolt' (60) ndəәp d-əәrn-əәp > dəәr.nəәp ‘to cover up' ‘to cover up, hide' In his Sre grammar, Manley discusses the two major morphological processes he observed: affixation and reduplication, both of which he notes have almost completely died out as active processes in the language. He adds that since his data were limited, generalizations were difficult to formulate (1972:40-41). Pace Manley states that affixes are added exclusively to verbs. Nguyễn adds that the infix -əәrn- occurs only in roots with back vowels, while the other infixes are not constrained by the root verb vowel (1973:35). Neither of these observations is entirely accurate. The accumulation of more data since the 1970s indicates that Sre affixation is much more complex than previously though. This section only touches on the more prominent and recurrent forms. 3.3 Clitics Zwicky (1977) provides the classic definition of a (simple) CLITIC as ‘a phonological weakening and attachment of a morphologically free form to another 51 phonological expression.' He terms special clitics ‘as forms which are not found in the syntax where the expected corresponding nonclitic form would occur and that these are attached to another expression in the morphosyntax.' The base word that the clitic binds to is called the HOST (Payne 1997:22). Clitics preceding the host are termed PROCLITICS; clitics following the host are ENCLITICS (61). (61) Proclitic + host + enclitic 3.3.1 Reciprocal proclitic tam= The reciprocal proclitic tam= indicates that the plural actors involved in the verbal action do something to each other. The derived verb form requires two subjects or plural actors (62b-63b). (62) a. kʰaj ləәh caw 3 hit someone ‘S/he hits someone.' b. caw dɔ məә caw nɛ tam=ləәh person this and person that RECP=fight ‘This man and that man fight each other.' (63) a. kʰaj ʔəәm bal-məә bau ʔur taːm ɓɔn kʰaj de 3 live with spouse woman in village 3 GEN ‘He stays with his wife in her village.' b. kaːɲ-gəәbəәh bal-məә tam=bau love with RECP=spouse ‘Love each other (and) get married.' 52 3.3.2 Possessive/reflexive enclitic =tam The enclitic =tam denotes personal possession and is added to the end of the noun. (64) dɔ la: sraʔ=tam ʔaɲ ɟe45 this is book=POSS 1 GEN ‘This is my own book.' (65) dɔ la: sraʔ=tam kʰaj de this is book=POSS 3 GEN ‘This is his own book.' The reflexive enclitic is formed by adding =tam to saʔ ‘body.' (66) saʔ=tam body=REFL ‘myself, yourself, him/herself' (67) saʔ=tam ʔaɲ kɔɲ lɔt draː body=REFL 1 want go market ‘I myself want to go to market.' (68) Məә-ya saʔ=tam kʰaj la caw ləәh suəәn. but body=REFL 3 is person work garden ‘But he's a farmer himself. 45This is an instance of rule (6) in Chapter 2. 53 3.4 Compounding Compounding is a generic term for a linguistic unit composed of two or more roots, each of which could function independently in other circumstances. The dominant semantic property is that the meaning of a compound is either more specific, creates a generalized form from overlapping semantic fields, or is entirely different from the combined meanings of the words that comprise the compound (69-72). (69) mɛː.bɛːp < mɛːʔ + bɛːp ‘parents' ‘mother' ‘father' (70) ʔoːj.ʔaw < ʔoːj + ʔaw ‘clothes' ‘blanket, skirt' ‘shirt' (71) ɟəәlu.məәŋan < ɟəәluʔ + məәŋan ‘dishes' ‘large bowl' ‘small bowl' (72) muh.mat < muh + mat ‘face' ‘nose' ‘eye CHAPTER 4 SYNTACTIC CATEGORIES: WORD CLASSES 4.1 Introduction The purpose of this chapter is to provide an overview of syntactic categories in Sre. Syntactic categories (word classes) are the traditional ‘parts of speech' and more modern approaches also include phrasal categories, (e.g., noun phrase, verb phrase, adpositional phrase, etc.) The category of a word (or lexical item or phrasal category) is typically determined by its distribution, that is, its place in a sentence, its morphology (then affixes it may take), and by its function in a sentence. Categories are not usually based on semantic criteria. Each language will have its own distributional criteria (Crystal 2001:366; Carnie 2007:37). The grammatical functions of words in languages of the Southeast Asian linguistic area are quite malleable, presenting the linguist with a categorization problem in terms of syntactic and semantic categories. Daley (1998:12fn7) discusses the problem in her study of Vietnamese classifiers: Some grammarians create their own terminology in ‘despair of imposing ready-made "Standard Average European" category-labels on all form-classes and construction types' (quoting Matisoff 1991:445). While it is difficult to distinguish and name many grammatical categories found in Vietnamese, in most cases I use conventional terms which most nearly fit the function of the word in question. Kơho-Sre syntax shares many features with Vietnamese syntax as the languages are distantly related within the Austroasiatic phylum. 55 Baker (2003:3) notes that ‘a serious consequence of the underdevelopment of this aspect of syntactic theory [differences among the lexical categories] is that it leaves us ill equipped to do typology.' Further he states that The literature contains many claims that one language has a different stock of lexical categories from another. In many cases, these claims have caused controversy within the descriptive traditions of the language families in question. … Nor do we make interesting predictions about what the consequences of having a different set of basic categories would be for the grammar of a language as a whole. Additionally, Baker states that lexical categories involve the traditional components of linguistics including derivational morphology, inflectional morphology, syntax, and semantics. In addition, Most languages-probably all-turn out to have the same three-way distinction between nouns, verbs, and adjectives falling out along reasonably familiar line, once various confounding factors (such as the presence of functional categories) are properly controlled for (Baker 2003:21). This chapter will explore how Sre word classes can best be delineated by the use of putative lexical categories (§4.2), followed by a discussion of functional categories (§4.3). Granted, there are fuzzy boundaries between categories, plus the lexical/functional dichotomy itself may or may not provide the absolute elegant exposition that a good descriptive grammar strives for. 4.2 Lexical categories: content word classes Lexical categories provide the content of a sentence. The categories include nouns, verbs, adjectives, and adverbs. Generally, lexical categories consist of an open class of items, that is, new lexemes and neologisms can be added to the lexicon. There is no limit to the inventory of a particular category. This section includes nominal elements (§4.2.1), verbal elements (§4.2.2), and adjectives and adverbs (§4.2.3). 56 4.2.1 Nominal elements Nominal elements discussed in this section include common nouns (§4.2.1.1), pronouns (§4.2.1.2), names and terms of address (§4.2.1.3), measure and quantity words (§4.2.1.4), and classifiers (§4.2.1.5). 4.2.1.1 Common nouns Nouns are the one obligatory constituent of a noun phrase. Examples of common nouns are listed in Table 4.1. 4.2.1.2 Pronouns Table 4.2 lists Sre personal pronouns. First person plural has an exclusive/inclusive dichotomy; second person singular and plural distinguish masculine, feminine, and affinal forms. 46 Third person pronouns have animate categories divided into [±human] and a nonhuman [±animate] homophonous gh2 ‘it' that is neuter and often functions as an anaphor. Dournes (1950:58) notes that ɗi (masculine/feminine) replaces me and ʔaj1 as a term of respect in addressing in-laws (affines). Evans and Bowen (1962:50) comment that the second person singular familiar affinal ʔi is a generally accepted term used among persons of either sex. More often than not, the second person pronouns are replaced in discourse by the relevant kin term that indicates a degree of respect from the speaker. 46Aj1 has several homophones: aj2 ‘give to (with dative marker ʔin)'; aj3 ‘take'; and aj4 ‘as for.' 57 Another practice is teknonymy, where the parents' name derives from their children (usually the oldest) (e.g., <Mè Sem> [mɛː sɛm], ‘the mother of Sem') (see 73). (73) Njam saʔ, bap sraŋ. ʔi pa tus tɑː m Daːlaːc səәl Good body father Srang. 2 new arrive to Dalat Q ‘Hello, father of Srang. Have you just come to Đà Lạt?' First person singular as a subject pronoun is illustrated in (74); dative pronouns examples are (75-77); and possessive usages are shown in (78-80). (74) ʔaɲ kɑːɲ sraʔ həәʔ 1 want book that ‘I want that book.' (75) ʔaj2 ʔaɲ ʔin sraʔ həәʔ. give 1 DAT book that ‘Give (to) me that book.' (76) dan cʰaŋ baːr-pɛ təәnap bɔl-hɛ ʔin please fry two-three egg PL.1INCL DAT ‘Please fry some eggs for us.' (77) ʔaj2 kʰaj ʔin huːc daːʔ give 3 DAT drink water ‘Give (to) him/her water to drink.' Possession is indicated by adding de (ɟe for 1st person) after the pronoun. 47 (78) a. sraʔ ʔaɲ ɟe b. sraʔ kʰaj de c. sraʔ kɔn de 47See Rule (6) in §2.2.2. This appears to be a Proto-South Bahnaric phonological rule (external sandhi), which also occurs in the other South Bahnaric languages: Chrau, Mnong, and Stieng. 58 book 1 GEN book 3 GEN book child GEN ‘my book' ‘his/her book' ‘the child's book' (79) ʔaj2 ʔaɲ ʔin sraʔ ʔaɲ ɟe give 1 DAT book 1 GEN ‘Give me my book.' (80) kʰaj la bəәyo bɔl-hɛ de. 3 is friend PL.1INCL GEN ‘S/he is our friend.' Anaphoric pronominals include the third person pronoun gəә2 ‘it', which has the features [-human, -animate] and is homophonous with gəә1. It may also refer anaphorically to sentential antecedents, as seen in (81.b and 82, 83). (81) a. Nteːŋ ɗah draːʔ{i} where side market{i} ‘Where is the market?' (Manley 1972:118) b. Gəә2{i} ɗah ma guŋ. Or just c. ɗah ma. it{i} side right road ‘It is on the right side of the road.' (82) ʔaj2 ʔaɲ ʔin pʰɛ{i} həәʔ, ʔaɲ kɑːɲ was gəә2{i} give 1 DAT rice that 1 want measure it ‘Bring me that rice, I want to measure it.' (E&B078/019) (83) dilah ʔaɲ bɛp gəә2{i} ʔaɲ rəәgəәj huːc{i} if 1 sip it 1 able drink ‘If I can sip it, I can drink it.' (E&B044/025) 59 4.2.1.3 Names and terms of address Sre personal names, of either sex, are usually prefixed with <K'-> kəә-, as in <K'Sem> kəә.sɛm. Some names are preceded by <Ha> ha-, as in <Ha Sol> ha.sɔl. The names themselves are usually monosyllabic. Certain naming taboos and a preference for a unique name sometimes violate basic phonological rules. In subdialect B, the (interrogative) vocative form hah (VOC) is used to address family members, using the appropriate kin term (84). (84) a. mbəәh me tus, hah bɛːp from.where 2M come VOC father ‘Where are you coming from, father?' b. ɲcʰi ʔaj1 ləәh, hah mɛː what 2F do VOC mother ‘What are you doing, mother?' 4.2.1.4 Measure and quantity words Measure and quantity words include numerals and classifiers. Although they are technically adjectivals, they are discussed under nominals. 4.2.1.4.1 Numerals. Sre employs a decimal numeral system. Numerals are delineated in Table 4.3. Numerals preceding a noun are interpreted as cardinals (85); numerals following a noun are interpreted as ordinals (86). (85) praw kəәnhaj six month ‘six months' (86) kəәnhaj praw 60 month six ‘the sixth month, June 4.2.1.4.2 Classifiers.48 Many Asian and American Indian languages make extensive use of classifiers (e.g., Vietnamese and Navajo). Many Southeast Asian languages do also. Allan (1977) distinguishes four types of classifier languages: 1) numeral classifier languages (ex. Thai), 2) concordial classifier languages (ex. many Bantu and Australian languages), 3) predicate classifier languages (ex. Navajo), and 4) intra-locative classifier languages (only 3 known: Toba, Eskimo, and Dyirbal) (286-8). Focusing on the first type, numeral classifier languages, he notes that only four sequences are permissible for the paradigm combination of quantifier (QNTF), classifier (CLF), and noun (N) (288)(ex. 111). (111) Allan's four permissible sequences for numeral classifier languages: ⇒ QNTF + CLF + N: Amerindian languages, Bengali, Chinese, Semitic languages, Vietnamese N + QNTF + CLF: Burmese, Japanese, Thai CLF + QNTF + N: Kiriwina (Oceanic) N + CLF + QNTF: Louisiade archipelago (Oceanic) Kơho numeral classifiers fall into Allan's first category: QNTF+ CLF + N (see following for examples). Aikhenvald (2003:98) notes that numeral classifiers do not have to appear on any constituent outside the numeral noun phrase; so, there is no agreement in numeral 48This section is a revision of Olsen 2009d. 61 classifier between the noun and another constituent. She notes that numeral classifiers have other, contingent properties (87). (87) a. The choice of a numeral classifier is predominantly semantic b. Numeral classifier systems differ in the extent to which they are grammaticalized. Numeral classifiers can be an open lexical class. (Emphasis mine.) c. In some numeral classifier languages, not every noun can be associated with a numeral classifier. Some nouns take no classifier at all; other nouns may have alternative choices of classifier, depending on which property of the noun is in focus. There are approximately forty classifiers in Sre.49 All Sre concrete nouns can be counted and must be preceded by a classifier when being enumerated, with certain exceptions. Sre classifiers may be divided into two groups-those that are termed concrete and those that termed abstract (Manley 1972:119). The concrete classifiers fall into three major subcategories (88). (88) a. naɁ for humans b. naj for roundish, solid objects (such as fruit, rocks, footballs, grains of rice or wheat, etc.) c. nəәm for nonhuman animate creatures and all other inanimate objects not classified by naj 49In Chrau, a sister language south of Kơho, there are about forty regular classifiers, five standard classifiers, and fifteen temporary measure classifiers (Thomas 1971:130-135). 62 The three basic classifiers listed above are replacing the more specialized classifier functions. Examples of these classifier constructions are exemplified in (89). (89) a. baːr naɁ caw Ɂuːr two CLF person woman ‘two women' b. pɛ naj luɁ three CLF rock ‘three rocks' c. praw nəәm sraɁ six CLF book ‘six books' In addition to these three classifiers, there are other, more specialized classifiers that are not used that much and appear to be dying out. 50 They include measure classifiers borrowed from French (often via Vietnamese)(90). (90) a. lit liter < Vietnamese lít < French litre b. kiɁ kilogram < Vietnamese kí < French kilogramme c. tʰəәk meter < Vietnamese thước51 There are also traditional indigenous measure classifiers (91). (91) a. boːŋ one length (standing with arm upraised) 50An explanation for this is that many of the objects that required these more restrictive categories are not found in the speakers' increasingly urbanized environment, both in Vietnam and overseas. For a relevant discussion, see Nettle and Romaine 2000:62-66. 51More recently, the term mét (< Vietnamese < French) has come into use. 63 b. laːs one length (distance, when arms outstretched and fingers also extended, from fingertips of one to fingertip of the other) c. nɗɑːm one span (distance from thumb to middle finger of spread hand) d. tal one cubit (length from elbow to fingertips of the same arm) Examples of contemporary classifier constructions include (92a-c). (92) a. baːr lit təәrnɑːm two CLF rice-wine ‘two liters of wine' b. pwan kiɁ pwac four CLF meat ‘four kilograms of meat' c. ɟəәt tʰəәk baːj ten CLF cloth ‘ten meters of cloth' All abstract nouns must be preceded by the classifier ɟəәnaw when counted. ɟəәnaw classifies words, languages, songs, poetry, and legends, etc. (93).52 (93) baːr ɟəәnaw pəәnɗik two CLF poem ‘two poems' Evans and Bowen (1962:14a), in their discussion of Kơho classifiers, demonstrate the anaphoric function of classifiers in discourse (94). 52Manley (1972:123) included bəәta in his list of classifiers, but it patterns more as a nominalizer. 64 (94) a. bol-kʰaj bləәj pɛ nəәm Ɂjar{i}, Ɂaɲ kɔɲ bləәj do 3PL buy three CLF chicken 1 want buy one nəәm{i}. CLF ‘They are buying three chickens, I want to buy one.' b. ɟəәt naj krwac{i} nɗɑː ɗɔŋ tajh. Ɂaɲ kɔɲ ɟəәt ten CLF orange how.much coin Q. 1 want ten naj{i} CLF ‘How much are ten oranges? I want ten.' c. teːŋ dɔ gəәs pram naɁ caw{i}, pɛ naɁ{i} nɛh re place close have five CLF people, three CLF ATELIC return raw. already ‘There are five people here; three have gone home already.' b. do paŋ sraɁ one CLF paper (paŋ = CLF for flat, sheet-like things: cloth, mats, paper) ‘a sheet of paper' (96) a. do naɁ caw one CLF person (naɁ = CLF for counting human beings) b. do mpol/ɲcəәm caw one group(s) people (mpol or ɲcəәm = CLF for a group of people) c. do sɛŋ caw 65 one CLF people (sɛŋ = CLF for objects occurring in a sequence or linear order) Kơho numeral classifiers are semantically determined by the specificity of the noun categorized. The classifiers are not rigid in their application and can apply in several related or overlapping contexts with respect to the noun. Table 4.4 lists all the classifiers cited in this section, illustrating their scope with examples. 4.2.2 Verbal elements Verbal elements discussed in this section include intransitive verbs (§4.2.2.1), transitive verbs (§4.2.2.2), and ditransitive verbs (§4.2.2.3) in Sre. Additionally, stative verbs (§4.2.2.4), copular verbs (§4.2.2.5), coverbs (§4.2.2.6), and aspect auxiliaries (§4.2.2.7) are also discussed. 4.2.2.1 Intransitive verbs Intransitive verbs have a predicate that takes only one argument (i.e., a valency of 1) (97-98). (97) kʰaj loːt. 3 go ‘S/he goes.' (98) ʔaɲ bic. 1 sleep ‘I sleep.' 66 4.2.2.2 Transitive verbs Transitive verbs have predicates that take two obligatory arguments (i.e., a valency of 2) (99-100). (99) dan seːn seŋ dɔ. please read sentence this ‘Please read this sentence.' (100) ʔaɲ nɛh sao hɑːm raw. 1 ATELIC eat full already ‘I have eaten to my fill.' (= ‘I'm satiated.') 4.2.2.3 Ditransitive verbs Ditransitive verbs have predicates that take three arguments (i.e., a valency of 3) (101-102). (101) kʰaj ʔaj2 ʔaɲ ʔin do məәŋan dɑːʔ. 3 give 1 DAT one bowl water ‘S/he gave me a bowl of water.' (102) kʰaj bəәtɔ ʔaɲ ʔin ɗəәs caw 3 teach 1 DAT speak people ‘S/he teaches me to speak Kơho.' 4.2.2.4 Stative verbs There are two types of stative verbs: quantifiers and true statives, which can include some verbs, adjectives, and adverbs. An example of a quantifier is ʔwaʔ ‘to be much, many' (103); an example of a true stative (derived from an adjective) is mwat ‘to 67 be sad' (104). Adjectives are described in §4.2.3.1. (103) gɛh ʔwaʔ caw ʔuːr tam draːʔ. existential many people women LOC market ‘There are many women at the market.' (104) ʔoːŋ kra nɛ mwat suːm. man old that to.be.sad always ‘That old man is always sad.' • Comparative and superlative degree of stative verbs. The comparative degree of statives is rəәlaw ‘to exceed, be greater than' + məә which is a comitative marker meaning ‘than' in this construction (105). (105) cʰi dɔ kraʔ rəәlaw-məә cʰi nɛ. tree this to-be-hard COMPARATIVE tree that ‘This tree is harder than that tree.' The superlative degree of statives is rəәlaw + ɟəәh ‘to be finished, complete' (106). (106) ʔuːr nɛ haːŋ rəәlaw-ɟəәh. woman that to-be-pretty SUPERLATIVE ‘That woman is the prettiest.' 4.2.2.5 Copula verbs There are five copula verbs in Sre: la(h) ‘to be', ɟeːŋ ‘to be/become', lah ɟeːŋ ‘to be', gəәs ‘to become/turn into', and gəәləәh ‘to become.' The usage and meaning of these verbs varies depending on the dialect. Sentences (107-109) illustrate some of these subdialectal differences. (107) a. kʰaj lah (ɟeːŋ) caw ywan. 68 3 to be person Vietnam ‘S/he is (a) Vietnamese.' (Subdialect A) b. kʰaj la bəәyo ʔaɲ ɟe 3 to be friend 1 GEN ‘S/he is my friend.' (Subdialect B) In subdialect B, lah means ‘to speak, say'; ɟeːŋ does not appear in the data observed. (108) a. kʰaj gəәs kwaŋ 3 become official ‘S/he became an official.' (Subdialect A) In subdialect A, gəәs serves a mutative function. b. kʰaj gəәs jəәt lɔʔ sre. 3 have ten CLF paddy.field ‘S/he has ten (wet rice) fields.' (Subdialect B) In subdialect B, gəәs is an existential verb; when it appears initially in a sentence, it often introduces a content question. (109) a. kʰaj gəәləәh kɔːp/chaːʔ-haːp 3 become sick/happy ‘S/he got sick/became happy.' (Subdialect A) b. ɲcʰi gəәləәh kʰaj ɲim what make-happen 3 cry ‘What made her/him cry?' (Subdialect B) In subdialect B, gəәləәh has the general meaning of ‘to make something happen to 69 someone.' It occurs mostly in interrogative sentences. 4.2.2.6 Coverbs Coverbs modify the sense of the main verb without changing its basic meaning; they precede the main verb. The principal coverbs are listed in Table 4.5. The first three are true modals. In subdialect A, up to four coverbs may precede the main verb (see ex. 120, below), but Manley (1972:203-204) notes that intelligibility breaks down if there are more than four and many combinations of multiple coverbs produce ungrammatical sentences. The last two coverbs do not occur in the data examined for subdialect B. Examples of coverb usage are illustrated in (110-119). (110) tuʔ ləәj di ləәm time any be.right only ‘Any time will be all right.' (Subdialect A) (111) gəәs ʔwaʔ caw kəәnɔm cih sraʔ ɓaː di ɓaː existential many people child write paper some be.right some ʔɑːʔ not ‘Many children write, some right, some not.' (Subdialect B) (112) ʔaɲ pal ʔəәm tam ɓɔːn 1 must stay in village ‘I have to/must stay in the village.' (Subdialect A) (113) ʔoːŋ pal ɗəәs lɔh-laŋ man must speak clearly ‘You must speak clearly.' (Subdialect B) 70 (114) kɔːn nɛ rəәgəәj rɛ kid that can swim ‘That kid can swim.' (Subdialect A) (115) kun saʔ gɛn-soŋ rəәgəәj mut tɑːm hiw ʔaɲ, … stoop body consequently can enter in house 1 ‘Stoop over so you can come into my house, …' (Subdialect B) (116) kʰaj loːc ləәh hiw pa mɛʔ-baːp kʰaj ʔin 3 finish make house new mother-father 3 GEN ‘S/he is finishing building the new house for her/his parents.' (Subdialect A) (117) loːc bɛhəәʔ kʰaj re hiw finish that 3 return house ‘After that, s/he went home.' (Subdialect B) In subdialect B, loːc only occurs with bɛhəәʔ with the meaning ‘after that.' It generally appears initially in a subordinate clause. (118) kʰaj mɑːŋ ləәh brwaʔ dɔ 3 used.to do work this ‘S/he is used to doing this work.' (119) caw nɛ təәlɔːŋ bəәsram daʔː təәj man that try-to study language French ‘That man is trying to study French.' Manley cites a ‘monster' sentence that contains four coverbs, which borders on unintelligibility; it was acceptable to his language consultant, though not something he would say in daily conversation (120). 71 (120) kʰaj di təәlɔːŋ mɑːŋ rəәgəәj ɗal-sraʔ 3 be.right try.to be.used.to be.able.to read ‘It is all right for him to try to get used to being able to read.' (M206/185) 4.2.2.7 Aspect auxiliaries Sre has two optional aspect auxiliaries that precede the main verb: nɛh and rəәp. Both of these auxiliaries function not so much as tense markers as indicators of whether an action has been completed or not. Nɛh indicates that the action has been completed (TELIC), while rəәp signals that an action has not been completed and is either ongoing or will be completed in the future (ATELIC). Since these aspect auxiliaries are optional, Sre speakers will often add a temporal phrase to the sentence to narrow down the context or time frame of the utterance (121-123).53 (121) kʰaj lɔːt tam Daːlaːc. 3 go to Dalat ‘S/he goes to Đà Lạt (daily/every month/next year).' or ‘S/he went to Đà Lạt (last year/yesterday).' or ‘S/he will go to Đà Lạt (tomorrow/next month).' (122) kʰaj nɛh lɔːt tam Daːlaːc ŋaj ɔr 3 TELIC go to Dalat day previous ‘S/he went to Đà Lạt yesterday.' (123) nam daʔ kʰaj rəәp lɔːt tam Daːlaːc. 53An observation by Emaneau (1951:63) concerning Vietnamese verbal syntax parallels Sre, ‘ ... verbs do not carry the categories of tense and mode. These, to some extent, are carried by the sentence construction, but to an even greater extent they are left to the extra-grammatical context, linguistic or non-linguistic.' 72 year next 3 ATELIC go to Dalat ‘Next year, I will go to Đà Lạt.' 4.2.3 Adjectives and adverbs Carnie (2007:44) points out that there is a lot of overlap between the distributions of adjectives and adverbs. The major difference between them is syntactic: adjectives appear inside noun phrases and adverbs occur elsewhere. He notes that this predictable complementary distribution, analogous to the phonological attribution, could tentatively place adjectives and adverbs in the same lexical category. This distribution appears to occur in Sre also. Adjectives are discussed in §4.2.3.1 and adverbs in §4.2.3.2. 4.2.3.1 Adjectives Adjectives modify nouns. Adjectives are divided into two groups: derived and nonderived. Derived adjectives are lexical copies of stative verbs. They fall into two major subclasses: those that quantify or measure something related to a noun (124a), and those that define a quality associated with a noun (124b).54 (124) a. bɔl ʔuːr tam draːʔ ʔwaʔ ŋan PL woman LOC market (to.be).many very ‘The women at market are very numerous.' (M208/191) b. ʔaruh nɛ haːŋ ŋan young.girl that.DIS to.be.beautiful very ‘That young girl is very beautiful.' (M208/192) 54Both adjectives (and adverbs) can be intensified by ŋan ‘very' which follows the word it modifies. 73 Nonderived adjectives are true adjectives and consist of numerals and specifiers. Numerals are discussed in §4.2.1.4 above. Specifier is Manley's term for nonnumeric quantifiers, such as təәl ‘each' (which is kɑːp in dialect B), and ʔalaʔ ‘most of the' (which functions as a plural marker for noun phrases in subdialect B). Specifiers precede nouns in a noun phrase (1972:144-145). 4.2.3.2 Adverbs Adverbs in Sre are an interesting category, appearing in reduplication, sentence-initial position, sentence-final position, manner, and derived from verbal elements. Adverbs do not occur in noun phrases, but may appear elsewhere. Manley (1972:216- 223) categorizes them into sentence initial adverbs, preverbal adverbs, manner adverbs, nonmanner adverbs, sentence final adverbs, nominal adverbs, and nonnominal adverbs. There are several adverbial auxiliaries that serve a modal function: mɑː functioning as a punctual ‘immediately, right away' (124a); and mɑːn indicating that the result of the sentence is uncertain (124a). (124) a. ʔaɲ soːr kʰaj gɛn ɲɟɯːt kʰaj ntaw mɑː. 1 tell 3 then stand- 3 -up immediately ‘I told him and he stood up right away.' (E&B140/2BA) b. baɲ lɔːt ŋaːj ʔir rᵊŋɔːt bɔl-hɛ bɔːl-glar NEG.IMP go far too afraid PL.1INCL tired mɑːn. uncertainty ‘Don't go too far away, for fear that/lest we'll be tired.' (E&B127/013) Manley (1972:214) notes that 74 Once the entire complex of constraints on copulas, coverbs, stative verbs, true verbs, and verb-plus-verb-complement combinations is taken into account, it is possible to conceive of such monster verbal clauses as that contained in the following sentence [(209) in original]: (125) ʔaɲ mɑːŋ rəәgəәj təәlɔːŋ loːc pih gɔh pʰan 1 be.used.to be.able.to try.to finish wash clean clothes nɛ tam dul djəә. that in one hour ‘I am used to being able to trying to finish washing those clothes clean in one hour.' While it is very unlikely that any native speaker of Sre would utter such a sentence in casual discourse, it is perfectly within his competence to do so, and in fact, this particular sentence was accepted without hesitation by Manley's language consultant. 4.3 Functional categories: grammatical word classes Functional categories provide the grammatical information required in a sentence. Generally, functional categories consist of a closed class of items, that is, there is a finite limit to the lexicon. No new lexemes or neologisms may be added to the inventory of a particular category. This section includes determiners (§4.3.1), adpositions (§4.3.2), relator nouns (§4.3.3), conjunctions (§4.3.4), interrogatives (§4.3.5), and imperatives (§4.3.6). 4.3.1 Determiners Determiners include location and directional words (§4.3.1.1), and temporal units (§4.3.1.2). 75 4.3.1.1 Locational and directional words Sre locational and directional (deictic) words are listed in Table 4.6. 4.3.1.2 Temporal units There are two important temporal words in Sre: dəәʔ ‘one of a number of recurring or multiplied instances, or repeated acts' (Table 4.7) and tuʔ ‘time, event' (Table 4.8). Baker (2003:109) notes that one can count events by involving a dummy noun (e.g., dəәʔ ‘occasions' or tuʔ ‘times') that provides the criterion of identity that makes enumeration possible. There are analogous Vietnamese constructions to many of the Sre temporal constructions (Nguyễn Đăng Liêm 1971:21): mỗi X một Y Template: ‘each X one Y' mỗi ngày một nhiều ‘more and more numerous each day' (lit. ‘each day one many') mỗi tháng một lần ‘once a month' (lit. ‘each month one time') Contemporary units of time are borrowed from Vietnamese (126). (126) nɛh baːr ɟəә ɟəәt pʰuk ATELIC two hour ten minute ‘It is (already) two ten (2:10).' Pʰuk ‘minute' and ɟəә ‘hour' are Vietnamese loans: from <phút> and <giờ>, respectively. 4.3.2 Adpositions Adpositions can be considered to be analytic case markers as opposed to synthetic case markers like the suffixes found in Turkish or Latin (Blake 2001:9). Adpositions in 76 Sre are grouped into three categories: abstract, concrete nonspatial, and spatial case markers. There is a question whether what Manley terms (spatial) case markers actually function as prepositions (1962:69-72). This is an area for further study. 4.3.2.1 Abstract cases (grammatical cases) The abstact (grammatical) cases include: the genitive postposition de (§4.3.2.1.1) and the dative and benefactive postposition ʔin (§4.3.2.1.2). 4.3.2.1.1 Genitive postposition de. The genitive postposition de (ɟe after ʔaɲ)55 indicates possession; it occurs after kin terms and personal pronouns (127) repeating (78). (127) a. sraʔ ʔaɲ ɟe b. sraʔ kʰaj de c. sraʔ kɔn de book 1 GEN book 3 GEN book child GEN ‘my book' ‘his/her book' ‘the child's book' 4.3.2.1.2 Dative and benefactive postposition ʔin. The dative postposition marks an animate direct object (128-129). The dative sense (DAT) often is ambiguous in that it includes or overlaps a benefactive (BEN) interpretation in many of these sentences (130). There is an allomorph dative preposition te that can also appear in the same environment, but it occurs with some sensory verb constructions often in conjunction with causative verbs (131). This is an area for further research. (128) kʰaj ʔaj2 təәrnɑːm rəәpu ʔin 3 give rice-wine buffalo DAT ‘S/he gave rice wine to the water buffalo.' (129) koɲ ʔaɲ sen-gar ʔəәsɛh caw pɔɑː ɓɔːn ʔin 55This is an instance of rule (6) (see §2.2.2). 77 uncle 1 take-care-of horse people chief village DAT ‘My uncle takes care of the village chief's horse.' The dative case marker can also indicate a benefactive (BEN) sense (153). (130) bi klaw ʔaɲ taɲ do nəәm səәʔ caw ʔur kʰaj ʔin older-sibling male 1 weave one CLF basket CLF female 3 BEN ‘My (older) brother is weaving a basket for his wife.' (131) a. kʰaj cəәm luʔ te sɔ 3 throw rock DAT dog ‘He throws rock at the dog.' b. ʔaɲ təәn-ŋɔːt gəә2 te kʰaj 1 CAUS-fear 3 DAT 3 ‘I cause him to fear it.' 4.3.2.2 Concrete nonspatial case (semantic): comitative and instrumental preposition (bal) məә The comitative preposition (bal) məә indicates an accompaniment relationship (‘together with x'). When only the comitative marker appears in a sentence, the first element bal is optional (132). (132) ʔaɲ lɔːt həә Daːlaːc (bal)-məә gəәp ʔaɲ 1 go to Dalat (COM)-COM friend 1 ‘I go to Đà Lạt with my friend.' (M68/19-20) 78 However, if both comitative and instrumental markers occur in the same sentence, then the instrumental phrase is indicated by məә alone and the comitative phrase is obligatorily indicated by bal məә (133). (133) ʔaɲ lɔːt həә Daːlaːc məә rəәndɛh bal-məә gəәp ʔaɲ 1 go to Dalat INSTR vehicle (COM)-COM friend 1 ‘I go to Đà Lạt with my friend.' (M68/21) 4.3.2.3 Spatial cases (local) As mentioned, there is disagreement about whether these spatial cases are indeed case markers (as per Manley) or are prepositions with a rather wide semantic scope. Spatial (local) cases include: the locative preposition tam (A), tɑːm (B) (§4.3.2.3.1), the goal preposition tus (§4.3.2.3.2), the source preposition bəәh (§4.3.2.3.3), and the direction preposition te (A), de (B) (§4.3.2.3.4). 4.3.2.3.1 Locative preposition tam (Subdialect A), tɑːm (Subdialect B). The locative preposition tam/tɑːm indicates a general location in time and/or space. There appears to be no difference in the sense of this preposition in either subdialect A or B (134). (134) a. kʰaj ʔəәm tam/tɑːm ɓɔːn 3 be-located LOC village ‘He lives in the village.' b. ʔaɲ rɛ tam/tɑːm pwan kəәso/ɟəә 1 swim LOC four kilometer/hour 79 ‘I swam for four kilometers/hours.' 4.3.2.3.2 Goal preposition tus. The goal preposition tus has different interpretations depending on the subdialect. In dialect A, tus indicates the goal or destination where someone or something is going (135). The English preposition ‘until' best translates the sense. In subdialect B, the semantic scope of the word is extended to indicate recent arrival in addition to referring to future events in time and space. In subdialect B, tus often co-occurs with bəәh where it appears to function as a verb (136). (135) a. kʰaj bəәsram tus drim 3 study until morning ‘He studied until morning.' b. kʰaj lɔːt tus Daːlaːc 3 go until/to Dalat ‘He went to Đà Lạt.' (136) a. mbəәh me tus where-Q 2M from ‘Where are you (m.) from?' b. ʔaɲ tus bəәh daː-dəәŋ 1 come from village-name ‘I come from Dà Dơng.' 4.3.2.3.3 Source preposition bəәh. The source preposition bəәh indicates the source or place where someone or something is from (137-138). (137) kʰaj loːt bəәh Daːlaːc. 3 go from Dalat 80 ‘He came from Đà Lạt.' (138) caw ləәh mpal bəәh tɑːm krjaŋ people make mortars from CLF kriang (tɑːm = CLF for trees, plant stalks) ‘People make mortars from kriang wood.' 4.3.2.3.4 Direction preposition te (Dialect A), de (Dialect B). The direction preposition te/de indicates a more specific location than tam/tɑːm, the locative preposition. There appears to be no difference in the sense of this preposition in either subdialect A or B (139). (139) a. kʰaj cəәm luʔ te/de sɔ 3 throw rock at dog ‘He throws rock at the dog.' b. ʔaɲ ʔəәm te/de bəәr mpoːŋ 1 be.located at mouth door ‘I am in the doorway.' 4.3.3 Relator nouns In many languages, mostly located in Southeast Asia, there is a set of locative nouns that act as relator nouns, that is in form they are nouns but express relational concepts usually translated in English by prepositions.56 Relator nouns are a specialized subclass of nouns that behave like adpositions in relating a predicate to a noun phrase (Blake 2001:204). Thompson (1987:200) described Vietnamese relator nouns as follows: 56Relator nouns are also termed relative location nouns; in Mesoamerican linguistics they are called ‘relational nouns.' 81 Many of the members of this small class appear superficially to resemble the prepositions of languages like English or French. This has to do with the fact that they all express position (in space or time) or some vaguer dimension in relation to something else. Although many expressions containing these words are best translated by prepositional phrases in English, the understanding of their basic nominal meanings will help remove some of the puzzling aspects of their uses. Since relator nouns generally indicate relative spatial location, analogous to English prepositions (e.g., on the table, under the house, etc.) and share some characteristics of both nouns and prepositions, there is an ongoing discussion among linguists about whether to consider these words as adpositions or complex prepositions (Blake 2001:185fn15). Manley references Thompson's (1987 [1965]:200-1) definition of relator nouns, and notes that Sre relator nouns are more restricted syntactically than those of Vietnamese. Relator nouns in Vietnamese need not be preceded by a locative particle, may be followed by a determiner (a demonstrative), and may function as the surface subject of a sentence (Manley 1972:127-8). In this section, I examine the pattern of relator nouns in Sre. These data are taken from my fieldwork research materials: a concordance (Olsen n.d. c) I compiled from sentences found in Evans and Bowen 1962 and Manley 1972 and my unpublished Kơho- English dictionary (Olsen n.d. b). I found that Sre relator nouns consist of a closed class with these four examples probably representing the totality of the class. The four relator nouns are: ɗaŋ, ‘above, up(per)'; ɗəәm, ‘under(neath),' ‘low(er), below'; dəәlam, ‘in(side), interiorness'; and bəәdiːh, ‘out(side), exteriorness, away' (140) to (143). (140) a. həә ɗaŋ cəәnɑːŋ to(wards) at above table 82 ‘on top of the table' (M127/p21) b. Ɂaɲ ɗuːh plaj bəәh ɗaŋ cʰi 1 drop fruit from above tree ‘I dropped the fruit from the tree.' (E&B119/094) (141) a. həә ɗəәm cəәnɑːŋ to(wards) at below table ‘underneath the table' (M127/p22) b. seː ɗaŋ nɛh hɔn rau, məә-ja seː ɗəәm ʔɑːʔ-heːt tooth upper ATELIC grow already but tooth below not.yet ‘(His/her) upper teeth have come through already, but not the lower ones.' (E&B144/008) (142) a. tam dəәlam ɓɔːn in/at interiorness village ‘inside the village' (M127/p23) b. tam dəәlam hiw ɡəәs dra məә rɑːc in/at interiorness house exist pillars and beams ‘Inside (of) the house, there are pillars and beams. (E&B142/006) (143) a. tam bəәdiːh ləәɡar jwan in/at exteriorness place Vietnam ‘outside Vietnam' (M127/p24) b. baɲ lɔːt bəәdiːh NEG.IMP go outside ‘Don't go outside.' (E&B125/WS1) 83 The examples for the relator nouns pattern similarly: locative marker + relator noun + noun. However, realtor nouns can serve other syntactic functions when used in patterns other than the ones delineated here. Something parallel also occurs in English, where prepositions, such as inside and outside, may function as nouns, adjectives, or adverbs. 4.3.4 Conjunctions Conjunctions link equal phrases or clauses. There are three important conjunctions in Sre: additive məә ‘and' (homophonous with the comitative/instrumental məә), alternative hala ‘or', and contrastive or contrary məә-ja ‘but' (144-146) are illustrative sentences. (144) Ɂaɲ gəәs do naʔ ʔɔh-ʔur məә baːr naʔ lɔʔ-kɔn-koɲ 1 have one CLF yg.sister and two CLF older.F.cousin ‘I have one younger sister and two older female cousins.' (145) kʰaj rkɛʔ hala lat 3 one-way-of-plowing or another-way-of-plowing ‘He plows one way or another.' (146) di məә-ja gəә ʔɑːʔ njam bɛ cʰi dɔ yes but it.(PL) NEG good like thing this ‘Yes, but it is/they are not as good as this one (thing).' (E033/059) 4.3.5 Interrogatives Interrogative pronominals are formed by prefixing a homorganic nasal {N} to a nominal base form. These are listed in Table 4.9. 84 4.3.6 Imperative markers There are four imperative particles: polite inclusive (Ɂih or laʔ), polite exclusive (jəә), and nonpolite exclusive (tɛʔ). See §5.2.3 for illustrative sentences involving these markers. See §5.2.2 for illustrative sentences and a syntactic discussion of this set of interrogatives. 85 Table 4.1 Examples of common nouns Sre form English gloss caw ‘man, person' luʔ ‘stone, rock' kɔːj ‘rice' pwac ‘meat' jwas ‘shoulder ax' Kəә.brɔj ‘personal name (K'Broi)' Daː.laːc. ‘place name (Dalat = Đà Lạt)' Table 4.2 Personal pronouns Singular Plural57 1st person exclusive ʔaɲ bɔl hi (ʔaɲ = rare) inclusive bɔl hɛ 2nd person masculine me bɔl me feminine ʔaj1 bɔl ʔaj1 affinal ɗi ~ ʔi bɔl ɗi ~ ʔi 3rd person [+human] kʰaj bɔl kʰaj [-human, ±animate] gəә1/2 bɔl gəә1/2 57Manley posits hi and hɛ as dual pronouns. He notes that ‘[T]he first person is the only place the category Dual turns up, and also, the only place the inclusive/exclusive distinction is made, thus suggesting that hi and he may be surviving relics from an older, richer pronominal system (1972:116).' Although, no dual pronouns occur in other South Bahnaric languages, they are found in Katuic languages. 86 Table 4.3 Numerals. ________________________________________________________ duːl one baːr two pɛ three pwan four pram five praw six pɔh seven pʰaːm eight sin nine ɟəәt ten ɟəәt baːr twelve baːr ɟəәt twenty pɛ ɟəәt thirty pwan ɟəәt pram forty-five rᵊhjaŋ hundred rᵊbo thousand pɔh rᵊhjaŋ sin ɟəәt praw seven hundred ninety-six: 796 ______________________________________________________ 87 Table 4.4 Inventory of Kơho classifiers cited in this section ___________________________________________________________ Classifier Scope: Example nouns ___________________________________________________________ blah layers of s.t. (clothing, paper): ʔaw ‘shirt, blouse'; sraɁ ‘a ream of paper'; ʔoj ‘blanket' ɟəәnau abstract nouns for words, languages, songs, poetry, and legends: caw ‘Kơho language' kiɁ58 measure for weight: kilogram: <object to be weighed> lit measure for liquids (liter): <object to be measured> mpol group(s) of people (clan): caw‘clan (or people)' naɁ counting human beings: bar naɁ caw ‘two people' naɁ small, round objects (fruit, rocks, balls, grains of rice): kroac59 ‘orange (fruit)' nəәm nonhuman animate creatures and all other inanimate objects not classified by naj: glɛ ‘type of bamboo', sɔ ‘dog', səәgəәn ‘hat' ɲcəәm a group of people: caw 'a group' paŋ flat, sheet-like things (cloth, mats, paper): sraʔ ‘paper', mpan ‘board, plank' sɛŋ objects occurring in a sequence or linear order: caw ‘people lined up, hiw ‘row of houses' səәɁ measure of volume based on capacity (~20 kg. paddy): <quantity to be measured> ______________________________________________________ Data source: Evans and Bowen 1962:14a, 90-92; Manley 1972:119-123; Olsen field notes. 58This is the current term (< French); an older term is kəәr. 59Kroac can also be used with plaj, the classifier for fruit, etc. 88 Table 4.5 Coverbs and their glosses ________________________________________________________ Coverb [modality] Gloss ________________________________________________________ di [permission] ‘be permitted to; be (all) right, correct' pal [necessity] ‘have to; must; be worthy of' rəәgəәj [ability] ‘be able to; can' loːc [completion] ‘finish' mɑːŋ [accustomed] ‘be accustomed to; be used to' təәlɔːŋ [attempt] ‘try to' __________________________________