AN INITIAL RECONSTRUCTION OF PROTO-BORO-GARO by DANIEL CODY WOOD A THESIS Presented to the Department of Linguistics and the Graduate School of the University of Oregon in partial fulfillment of the requirements for the degree of Master ofArts December 2008 11 "An Initial Reconstruction ofProto-Boro-Garo," a thesis prepared by Daniel Wood in partial fulfillment of the requirements for the Master ofArts degree in the Department of Linguistics. This thesis has been approved and accepted by: Dr. Scott DeLancey, Chair of the Examining Committee Date Committee in Charge: Accepted by: Dean of the Graduate School Dr. Scott DeLancey, Chair Dr. Spike Gildea © 2008 Daniel Wood 111 IV An Abstract of the Thesis of Daniel Cody Wood in the Department of Linguistics for the degree of to be taken Master of Arts December 2008 Title: AN INITIAL RECONSTRUCTION OF PROTO-BORO-GARO This manuscript has been approved by the advisor and committee named . below and by Richard Linton, Dean of the Graduate School. - .-----.-------------- Dr. Scott DeLancey \ This study attempts to reconstruct Proto-Boro-Garo (PBG), the ancient language from which the modern Boro-Garo (BG) family evolved. BG is a largely under- documented sub-branch of Tibeto-Burman that is spoken primarily in the Brahmaputra valley of northeastern India. While other comparative studies have focused on PBG phonology, this study concentrates on grammatical elements and syntactic structures. An initial reconstruction is attained by examining data from the limited number of descriptive grammars available on BG languages and using the comparative method to determine the oldest forms of grammatical elements. Where elements correspond across languages, they can be reconstructed for the common ancestor. When they do not, we have evidence for independent innovation. This is accounted for, when possible, by language-internal reconstruction. CURRICULUM VITAE NAME OF AUTHOR: Daniel Wood GRADUATE AND UNDERGRADUATE SCHOOLS ATTENDED: University of Oregon Eugene, OR Universitat Autonoma de Barcelona Bellaterra, Catalonia, Spain University of California, Santa Barbara Santa Barbara, CA DEGREES AWARDED: Master of Arts, Linguistics, 2008, University of Oregon Bachelor of Arts, Linguistics, 2003, University of California, Santa Barbara Bachelor of Arts, Spanish, 2003, University of California, Santa Barbara AREAS OF SPECIAL INTEREST: Language description and documentation; Tibeto-Burman languages; historical and comparative linguistics; morphosyntax; semantics; language change PROFESSIONAL EXPERIENCE: Graduate Teaching Fellow, UO Department of Linguistics, September 2006- present v Catalan Language Tutor, UO Yamada Language Center, January 2006-June 2007 Graduate Teaching Fellow, UO Many Nations Longhouse, January-June 2006 GRANTS, AWARDS AND HONORS: VI Foreign Language and Areas Studies (FLAS) Fellowship, South East Asian Studies Summer Institute (SEASSI), University of Wisconsin, Madison, 2008 (declined) Tuition Fellowship, South East Asian Studies Summer Institute (SEASSI), University of Wisconsin, Madison, 2008 (declined) Small Professional Grant, Center for Asian and Pacific Studies (CAPS), University of Oregon, 2007 Foreign Language and Areas Studies (FLAS) Fellowship, South East Asian Studies Summer Institute (SEASSI), University of Wisconsin, Madison, 2007 Tuition Fellowship, South East Asian Studies Summer Institute (SEASSI), University of Wisconsin, Madison, 2007 Small Professional Grant, Center for Asian and Pacific Studies (CAPS), University of Oregon, 2007 South East Asian Studies Grant, University of Oregon, 2007 University of Oregon General Scholarship, University of Oregon, 2005 Phi Betta Kappa Honor's Society Induction, University of California, Santa Barbara, 2003 Robert E. Wilson Memorial Prize for outstanding senior in Department of Spanish and Portuguese, University of California, Santa Barbara, 2003. Foreign Language Studies Scholarship, Portuguese, University of California, Santa Barbara, 2003 PUBLICATIONS: Wood, Daniel. (under review). "An Initial Reconstruction of the Proto-Boro-Garo Noun Phrase". In eds. S. Morey and M. Post, North East Indian linguistics 3: Papers from the third international conference ofthe North East India Linguistic Society. Wood, Daniel. (accepted for publication). Notice for Translation Equivalence: An Essay in Theoretical Linguistics. By M.K.C Uwajeh. eLanguage: http://elanguage.net/blogs/booknotices/ Vll Wood, Daniel. (accepted for publication). Notice for Aspect in Burmese: Meaning and function by Nicoletta Romeo. eLanguage: http://elanguage.net/blogs/booknotices/ Wood, Daniel & Linda Konnerth. (accepted for publication). Report on the 3rd Annual Meeting of the North East Indian Linguistic Society, Guwahati, Assam, India, 20- 22 January 2008. LTBA. V111 ACKNOWLEDGMENTS I wish to express my most sincere appreciation to Professor Scott DeLancey for the amount of time and wisdom he has shared with me throughout this entire process. His valuable insights and willingness to share his personal library certainly helped improve the overall quality of this work. I would also like to thank Professor Spike Gildea for his never- ending friendliness, support and infectious laugh. Were it not for these two people, this study would still only exist as disconnected thoughts in my head. In addition, special thanks are due to my parents and friends for leaving me alone when I needed to write but who never let me fully disappear into my own mind. For Lobsang, the tiger-killer. IX TABLE OF CONTENTS Chapter Page I. I1'JTRODUCTION 1 1.1 Language Family Background 1 1.2 Typology of the Modern Boro-Garo Languages 6 1.3 Data Sources 7 1.4 Aim and Scope 8 II. THE PROTO-BORO-GARO NOUN PHRASE.................................................. 9 2.0 Introduction 9 2.1 Syntactic Description of PBG NP Word Order 9 2.1.1 Demonstratives and the Nouns they Qualify.................................. 10 2.1.2 Possession 11 2.1.3 Plural Marking 12 2.1.4 Adjectives 13 2.1.5 Classifiers and Numerals 14 2.1.6 Case-inflections and Post-positions 16 2.1.7 Conclusion 16 2.2 PBG Grammatical Forms 16 2.2.1 Personal Pronouns and Demonstratives 16 2.2.2 Adjectives: Verbal Derivation 21 2.2.3 Relative Clauses: Verbal Derivation vs. Relative Pronouns 29 2.2.4 Interrogative Pronouns.............................................. 34 2.2.4.1 PBG who and whose 35 2.2.4.2 PBG what 36 2.2.4.3 PBG why 37 2.2.4.4 PBG where 38 2.2.4.5 PBG when 39 2.2.5 Plural Suffixes 40 2.2.5.1 Honorific Plural.................................................................... 42 2.2.5.2 Personal Pronoun Plurals 43 2.2.5.3 Human Plural........................................................................ 44 2.2.5.4 General Plural....................................................................... 46 x Xl Chapter Page 2.2.6 The PBG Classifier System 47 2.2.7 Postpositions 48 2.2.8 Case-endings 50 2.2.8.1 Nominative 51 2.2.8.2 Accusative 52 2.2.8.3 Dative 53 2.2.8.4 Genitive 53 2.2.8.5 Ablative 53 2.2.8.6 Instrumental 55 2.2.8.7 Locative 56 2.3 Conclusion 56 III. THE PROTO-BORO-GARO CLASSIFIER SYSTEM 57 3.0 Introduction 57 3.1 BG Classifiers: A Synchronic Definition 58 3.2 The Proto-Boro-Garo Classifiers: An Attempt at Reconstruction 62 3.2.1 Class Terms. 62 3.2.1.1 Animal Class Terms 63 3.2.1.1.1 Birds 64 3.2.1.1.2 Fish 64 3.2.1.1.3 Quadrupeds 64 3.2.1.1.4 Insects ~........................................................ 65 3.2.1.2 Plant Class Terms 65 3.2.1.2.1 Tress 66 3.2.1.2.2 Bamboo 66 3.2.1.2.3 Fruit 66 3.2.1.2.4 Tubers 66 3.2.1.2.5 Grass 67 3.2.1.2.6 Rice, Paddy............................................................ 67 3.2.1.3 Natural World Class Terms 67 3.2.1.3.1 Land, Soil............................................................... 68 3.2.1.3.2Fire 68 Xll Chapter Page 3.2.1.3.3 Stones 68 3.2.1.4 Objects of Human Manufacture Class Terms 68 3.2.1.5 Body Parts Class Terms 69 3.2.1.5.1 Face, Eyes 70 3.2.1.5.2 Hand, Arm 70 3.2.1.5.3 Leg 70 3.2.1.5.4 Internal Organs 70 3.2.1.5.5 Mouth 71 3.2.2 ]'Jumeral Classifiers 71 3.2.2.1 NCs Consistently Found Across BG 71 3.2.2.1.1 Humans 72 3.2.2.1.2 Animals 73 3.2.2.1.3 Flat Objects 73 3.2.2.1.4 Round Objects 75 3.2.2.1.5 Long Objects 75 3.2.2.1.5.1 Long, Flexible Objects 76 3.2.2.1.5.2 Long, Rigid Objects 76 3.2.2.1.6 Residual Category.................................................. 76 3.2.2.1.7 Abstract Concepts and Inanimate Objects 77 3.3 The Simplification of the Boro-Garo Classifier System 77 3.4 Conclusion 78 IV. PBG VERBAL MORPHOLOGy..................................................................... 80 4.0 Introduction 80 4.1 Neutral Tense 80 4.2 Negative Suffix........................................................................................ 85 4.3 Auxiliaries and Copulas.. 86 4.3.1 Equational Copula.. 86 4.3.2 Existential Copula 89 4.3.3 Other Auxiliaries and Modal Verbs .. 93 4.3.3.1 PBG *tak 'do, make' Auxiliary............................................ 94 4.3.3.2 PBG *man 'to be able to, manage, finish' Auxiliary........... 98 -- - --- ---------------- Xlll Chapter Page 4.3.3.3 Desiderative 100 4.4 Finite Verb Morphology.......................................................................... 101 4.4.1Future 101 4.4.2 Past Tense 106 4.4.2.1 RemotelNarrative Past 106 4.4.2.2 Recent/Immediate Past 108 4.4.3 Continuative 111 4.4.4 Imperfective 116 4.4.5 Optative 119 4.5 Prefixal Verb Morphology....................................................................... 121 4.5.1 Negative Imperative 121 4.5.2 Causative Derivation 122 4.6 Conclusion 128 V. CONCLUSION 129 APPENDIX: LIST OF ABBREVIATIONS 131 BIBLIOGRAPHY 133 XIV LIST OF FIGURES B~ ~~ 1. Map of India (northeast region circled) 2 2. Boro-Garo Language Map 2 3. Boro-Garo Language Family Tree 6 xv LIST OF TABLES Table Page 1. Pronouns across Boro-Garo 17 2. Modern Boro-Garo Demonstratives . 18 3. The PTB Suffix Paradigm 19 4. Transitive Paradigms for Mhai Kham, Rawang and Limbu 19 5. Indo-Aryan Demonstrative Pronouns 20 6. Cognate Forms of 'thick' across BG 25 7. Cognate Forms of 'good' in Modern BG 26 8. Adjective Prefix across BG 29 9. BG Relative Pronoun 33 10. who across BG 35 11. whose across BG 36 12. vvhat across BG 36 13. why across BG 38 14. where across BG 38 15. when across BG 39 16. BG Plurals 42 17. BG Classifiers and Reconstructed PBG Forms 48 18. Case-endings across Boro-Garo 50 19. Animal Class Terms in Modern BG 63 20. Plant Class Terms in Modern BG 65 21. Natural World Class Terms in Modern BG 67 22. Objects of Human Manufacture Class Terms in Modern BG 69 23. Body Parts Class Terms in Modern BG 69 24. Numeral Classifiers in Boro-Garo 72 25. PBG Neutral Tense 81 26. PBG Negative Suffix 85 27. PBG Equational Copula 86 28. PBG Existential Copula 89 29. Future Suffix in BG 102 XVI Table Page 30. RemotelNarrative Past in BG 106 31. Recent/Immediate Past in BG 108 32. Continuative Suffix in BG 111 33. PBG *diJ} 'stand' 114 34. Imperfective in BG 116 35. Optative Suffix in BG 119 36. Negative Imperative Prefix across BG 121 37. Causative Morphology in BG 123 38. BG Causative Prefixes 126 1CHAPTER I INTRODUCTION 1.1 Language Family Background The Brahmaputra valley of northeastern India is one of the most linguistically diverse regions on earth. The area has been the locus of intense language contact for millenia, with successive layers ofIndo-Aryan, Austro-Asiatic, Tibeto-Burman and most recently Tai languages (Grierson 1903; Jacquesson 2005,2006). All ofthe languages in the region evidence significant influence from the others. This is conspicuously true of the Boro-Garo family. Boro-Garo belongs to the greater Boro-Konyak-Jinghpaw sub-branch of Tibeto- Burman and consists of at least eleven languages: Bodo, Garo, Deuri, Rabha, Tiwa, Dimasa, Kokborok, Atong, Wanang, Moran, Koch. These languages can be further divided into various dialect communities which are found in the northeastern Indian states of Assam, Meghalaya and Nagaland, highlighted in Figure 1, as well as across the national border in Nepal and Bangladesh. The distribution of these languages across northeast India is illustrated in Figure 2. 2Figure 1: Map of India (Northeast region circled) Figure 2: Bora-Garo Language Map The classification of Tibeto-Burman sub-groups has been a point of debate for many years, with many distinct analyses proposed (see Hale 1982 for a comparative summarization of different proposed TB taxonomies). This is due in part to the similarity of typological features among many languages in Southeast Asia, the confusion created by the use of different names for the same language, as well as the prolonged cultural contact between distinct linguistic communities which has lead to massive amounts of cross-linguistic influence with respect to phonology, morphology, syntax and semantics (Haudricourt 1966; Matisoff 1976; Hale 1982). Boro-Garo is a close-knit branch whose genetic unity is indisputable, but its place in the larger family is less clear. Numerous 3attempts have been made to classify the primary branches of Tibeto-Burman, and consequentially, where to place Boro-Garo within the family. Those who have focused on BG have also attempted to classify these languages in terms of their relation to one another since as early as the late 1800's. In one of the earliest publications on any BG languages, Dalton (1872) acknowledges the relationship between Rabha, Kachari and Garo, but does not suggest internal branches to the BG family or place the sub-group in a specific branch of TB. Konow (1903) places BG under its own branch ofTB and divides the family into two sub-groups: Bara and Garo. He places Bodo, Mech, Dimasa, Kachari, Hojai, Tipura, Lalung and Moran under the Bara branch, while the Garo node is further divided into four distinct sub-branches; these consist of a) Garo, Abeng, Dacca and Kamrup, b) Atong, Rabha, Ruga, Tintekiya, Cooch Behar and Koch, c) Wanang and d) Chutiya. This contrasts only slightly with Shafer's (1953, 1955, 1966) sub-grouping of the family, who suggests a Baric branch ofTB that consists of the Barish (BG) and Naga sub-families. Within the BG group, rather than dividing BG into two primary sub-branches as Konow (1903) did, Shafer (1955, 1966) divides the BG languages among five primary sub-branches: a) Western, b) North Central, c) South Central, d) Jalpaiguri and e) Eastern. These divisions, however, seem to be based more-so on geographic location of the communities. Shafer's (1955, 1966) groups consist of a) Bodo, Mech, Dimasa, Hojai, Tipura, Lalung and Moran, a b) North CentralGaro, Abeng, Dacca and Kamrup, c) Atong, Rabha, Ruga, Tintekiya, Cooch Behar, Kontsand Koch, d) Wanang and e) Chutiya, respectively. 4The close genetic ties between the BG languages and neighboring Naga languages, as suggested by Shafer (1953), are recognized by Burling (1959) and reiterated by Voegelin & Voegelin (V&V) (1977). While Burling (1959) focused on the internal distribution of languages within BG, resulting in three distinct branches (Boro, Garo and Koch), V&V's analysis attributes the similarities between BG and languages like Jinghpaw and Rawang as evidence of a common linguistic heritage and propose a Boro-Naga-Kachin subbranch of Tibeto-Burman. The Boro (BG) branch of this family is divided into three distinct groups: a) Bodo, consisting of Mech, Dimasa and Kachari, b) Garo and c) Koch, consisting ofKoch, Atong and Wanang. Benedict (1972) divides the BG family into four branches: a) Bodo, consisting ofBodo and Dimasa, b) Garo, which is further divided into two distinct sub-branches, Garo A (Atong, Rabha and Ruga) and Garo B (Garo, Abeng and Awe), c) Deori and d) Konyak languages Moshang, Namsang, Banpara, Tableng, Tamlu and Chang. While Egerod (1974) applies Shafer's (1955) Baric term to the same family, he divides it into three groups: a) Bodo, consisting of Bodo and Dimasa, b) Garo, consisting of Atong, Rabha, Ruga and Koch, and c) languages related to the Baric languages, such as Moshang, Namsang and Banpara. The languages categorized under this third branch reflect the claims of genetic relation made by Shafer (1955, 1966) and Benedict (1972), who place them under a separate Nagish or Konyak node, respectively. Burling (1983) continues with the classification ofBG, drawing from previous comparative analyses which demonstrate the genetic relationship between Garo and Jinghpaw (Burling 1971). Based on lexical similarities resulting from shared innovation, 5he proposes the Sal sub-group, the name of which is based on the reconstructible form of the cognate lexical item used for 'sun' which is unique to these languages. Though his analysis does not alter the internal classification ofthe BG family, Burling (1983) provides greater evidence for the common linguistic ancestry among the BG languages and other Konyak languages such as Nocte, Konyak and Tangsa. The tripartite division ofBG espoused by Burling (1959) was reiterated in Jacquesson (2005, 2006), but after conducting a more thorough comparative analysis ofBG phonology (Joseph & Burling 2006), a different subgrouping of BG was proposed. While the connection of BG to Konyak and Kachin languages remains undeniable, Joseph & Burling (2006: 1-2) divide the family into four sub-branches: a) Garo, consisting of the various mutually intelligible Garo dialects, b) Koch, consisting of Koch, Rabha, Wanang, Atong and Ruga, c) Boro, consisting of Boro, Kokborok and Tiwa and d) Deuri. This division is maintained by Joseph (2007), who also claims some level of shared innovation between Bodo and Garo. For the purposes of this project, I will adopt the division of modern BG into these four sub-groups as proposed by Joseph & Burling (2006). Figure 3 illustrates the internal division of the family. 6Figure 3: Boro-Garo Language Family Tree PBG· Bodo (taro BO(~Va 1\:0(1 1.2 Typology of the Modem Boro-Garo Languages _ tong In Boro-Garo, as in Sino-Tibetan languages in general, the syllable plays a major role in the phonological system which tends to have a more numerous inventory of initial consonants than syllable-final consonants (Burling 1992; Joseph & Burling 2006). While there is some dispute concerning the number of tones (Burton-Page 1955; Burling 1959; Bhattacharya 1977; Joseph & Burling 2006), we find a fairly consistent two-way contrast between high and low tones in BG languages; Garo is the only exception to this phenomenon. Instead, Garo syllables contrasts between those with a final glottal stop and those without. The lack of phonological tone contrasts in Garo has been attributed to the historical development of the glottal stop from the BG high tone (Burling 1959, 1981, 1992; Joseph & Burling 2001, 2006). The modem BG languages are verb-final. Arguments are freely omitted, especially when argument reference is easily retrieved from the discourse context. When included in propositions, however, there is a strong tendency for A arguments to precede o arguments, though this is by no means strict. Argument order may vary according to 7pragmatic factors of the speech environment. The semantic roles of the arguments are nevertheless maintained by post-NP case-marking clitics, which follow a Nominative- Accusative system. The languages of the BG family also employ a system of post- positions and relator nouns which describe the location of a referent with respect to some grounding element. We also find the consistent use ofnumeral classifiers which prefix onto the numeral word class when modifying a noun referent. While some ofthe categories denoted may vary across the family, these classifiers organize referents into semantic categories according to salient physical characteristics. It is difficult to support the notion of a coherent OV verb phrase in the traditional sense due to the non-restrictive nature ofargument order. However, grammatical behavior of the verbal complex is consistent across all branches of the family, with TAM inflections suffixed onto the main clausal verb. There are very few verbal prefixes, which are more commonly derivational in function. These languages also demonstrate a tendency towards clause-chaining and verb serialization, which has created the necessary conditions for grammaticalization and reanalysis in many instances. 1.3 Data Sources As the majority of the BG languages lack descriptive documentation, data sources for this language family are limited. However, there are descriptive grammars available for at least one language from each of the four branches ofmodem BG (Bodo, Garo, Rabha and Deuri), which are the basis for this comparative work. All examples and data 8used in this thesis have been taken from these grammars, and when available, examples from lesser discussed BG languages have been taken to supplement the reconstruction. 1.4 Aim and Scope This thesis attempts an initial reconstruction Proto-Boro-Garo syntax, paying special attention to the form and function of grammatical elements. All phonological reconstructions are based on Joseph & Burling's (2006) reconstruction ofProto-Boro- Garo phonology. Where the functions of grammatical elements vary between languages, I attempt to provide a logical diachronic pathway for the evolution of these items. In Chapter II, I reconstruct the PBG noun phrase, with special focus placed on inconsistent forms and functions ofNP grammatical items. In Chapter III, I provide an in-depth analysis of the classifier systems across BG in an attempt to identify the basic semantic categories present in PBG grammar. I also account for variation of classifier forms between languages, and discuss class-terms. Chapter IV discusses PBG verbal morphology and offers a reconstruction of most of the finite verb affixes in PBG. 9CHAPTER II THE PROTO-BORO-GARO NOUN PHRASE 2.0 Introduction Noun phrases behave as arguments to the main clausal verb. In modern Bodo- Garo, just as in many Tibeto-Burman languages, NPs may be left out of a proposition if the referent is clear from the discourse context. When included, they can range in the complexity of their internal constituents. A noun phrase may be as simple as a single pronoun, or may contain a more complex array of elements that form a syntactically and conceptually bound unit. This chapter attempts to reconstruct internal NP constituent order, as well as many of the grammatical elements typically found in NPs across the modern BG languages. 2.1 Syntactic Description and Reconstruction of PBG NP Word Order In this section I will reconstruct the order of elements in the Proto-Bodo-Garo NP. The reconstruction will be based on a purely formal approach, paying attention only to the word order of clausal elements in very specific construction types. Once the reconstructed form of a construction type is established, I will then move on to another construction, all the time keeping in mind the previously reconstructed forms to see if the 10 data continue to support, or end up contradicting, my hypothesis. With each subsequent reconstruction of a particular construction, I will add these to the PBG forms in their respective syntactic position, making an ever more complex noun phrase structure for PBG. Since the languages that comprise this branch of Tibeto-Burman are said to have developed at a more recent time-depth, we can assume that propositional forms will be very similar and lend themselves to a rather easy reconstruction ofPBG syntax. 2.1.1 Demonstratives and the Nouns they Qualify Demonstratives may function as a pronoun and alone fill the NP slot. When modifying a noun, they consistently precede the noun qualified across the modem BG languages. This suggests a similar syntactic pattern in Proto-Boro-Garo: PBG *DEM N. (1) Garo: a. alJ-a u-ko nik-a-milJ I-NOM that-Acc see-NEUT-PST 'I saw it/that' (Burling 2004: 214) b. i-a thiS-NOM palJ-ko tree-ACC daralJ-ba anyone-IND den?-na-be cut-NEG.IMP ia-de this-but alJ-ni-sa 1S-GEN-only 'Don't anyone cut this tree. It is mine.' (Burling 2004: 217) (2) Boro: be bithni-khow bun this information-Acc speak 'Speak this information' (Bhattacharya 1977: 122) (3) Deuri: Hela iskul-a atay-o-ne that school-TH all-GEN-FOC 'That school is public.' (Jacquesson 2005: 257) (4) Kokborok: i manwy-rak tub-ay thalJ-di this thing-many take-PROG gO-IMP 'Go taking these things.' (Annamalai 1976: 32) (5) Dimasa: ebo ancha hambi that child good 'That child is good' (Dundas 1908: 5) (6) Rabha: e-kai pan-be nem-a this-REL tree-DEF good-NEuT 'This tree is good.' (Joseph 2007: 314) 2.1.2 Possession Across the modem languages, the genitive-inflected possessor consistently precedes the possessed. From this we can assume that the same was true for PBG. (7) Garo: chi1J-ni achak-pi?-saba lp-GEN dog-DIM-also 'our puppy also' (Burling 2004: 167) (8) Boro: tin-nl blztib-khow how 1S-GEN book-ACC send 'Send my book' (Bhattacharya 1977: 149) (9) Deuri: a a-yo now-o dii-i 1S 1S-GEN boat-Loc EX.COP-NEUT 'I am in my boat' (Jacquesson 2005: 231) 11 (10) Kokborok: Ram-ni uskul hakcal-o Ram-GEN school far-Loc 'Ram's school is far away' (Annamalai 1976: 58) (11) Dimasa: na-ni no baraha 2S-GEN house where.is 'Where is your house?' (Dundas 1908: 20) (12) Rabha: bak-gi tatheg bl-nata pig-GEN leg break-psT 'The pig's leg is broken.' (Joseph 2007: 361) Though there are no examples in any of the grammars where a NP has both a demonstrative and a possessor, it is a robust cross-linguistic pattern for pre-N demonstratives to be the first element in a NP. Thus, we may also assume that demonstratives preceded both the possessor and noun: *PBG: DEM poss N 2.1.3 Plural Marking Plural inflection is suffixed directly onto the noun it modifies in the modem languages, and precedes any case-marking that may be relevant to the noun's grammatical role in the sentence. This order reconstructs to*PBG: DEMposs N-PL. (13) Garo: mande-dag-ni ha?-ba-sog-ba bag?-a-mig persOn-PL-GEN fields-also-village-also much-NEUT-PST 'The Mandis' property (fields and villages) were plentiful' (Burling 2004: 180) (14) Boro: ma-ph:Jr-khow nan-gow what-PL-ACC need-psT 'What are needed?' (Bhattacharya 1977: 123) 12 (15) Deuri: jimosaya Deuri mira-mirasi-laju-wa boy-young-PL-TH cagu-ho route-LOC kirobo-ho gO-LOC 13 jelJ-laju-na gusa-i obstacle-PL-Acc remove-NEUT 'The young Deuris remove the obstacles when they go on the road' (Jacquesson 2005: 280) (16) Kokborok: bUl-rmy-chikla-rag-na rahar-di 3s-girl-young-manY-ACC Send-IMP 'Send the young girls' (Annamalai 1976: 54) (17) Dimasa: subulJ-nishi-ni man-PL-GEN 'of the men/men's' (Dundas 1908: 3) (18) Rabha: bukhi-kai kai-talJ-a mai nikhu-fJ hungry-REL person-PL-DAT rice give-IMP 'Give rice to those who are hungry.' (Joseph 2007: 359) 2.1.4 Adjectives Adjectives vary across the languages with respect to their word order. They may occur either before or after the noun they qualify. This may be a pragmatic function of the adjectives, giving contrastive focus on whichever element comes first. Since there is evidence in most of the languages that adjectives can move more freely with respect to the qualified noun in the NP, we might assume that this was also a phenomenon in PBG: *PBG: DEMposs {ADJ} N-PL {ADJ} (19) Garo: a. dal?-gipa big-REL 'big dog' achak dog 14 b. achak dal?-gipa dog big-REL 'big dog' (Burling 2004: 133) (20) Deuri: de papa tall tree 'a tall tree' (Jacquesson 2005: 111) (21) Kokborok: wag-j;Jla-kasam-na pig-male-black-DAT 'to the black male pig' (Annamalai 1976: 92) (22) Dimasa: iicha hiimba child good 'good child' (Dundas 1908: 5) 2.1.5 Classifiers and Numerals Since it is a grammatical restriction in some BG languages that the c1assifier- numeral morphemes follow the qualified noun, we might assume that this was their syntactic patterning in PBG. The freer word order we find in a few of the modem languages could simply evidence that reanalysis has taken place such that these constituents are now less constrained in the NP, somewhat mimicking the word order patterns of adjectives: *PBG: DEMposs {ADJ} N-PL {ADJ} CL-NUM (23) Garo: a. me?chik ak-bri woman cL-4 '4 women' (Burling 2004: 250) 15 b. kan-sa kodam-ko-in alJ-na cL-l pillOW-ACC-FRO IS-DAT 'Bring me a pillow.' (Burling 2004: 185) ra?ba-bo bring-IMP (24) Boro: m;Jsow-ha athin th;Jn-br;Jy don coW-LaC leg cL-4 Ex.Cop 'A cow has 4 legs' (Bhattacharya 1977: 142) (25) Deuri: a. ii-yo gu-ja gumo dii-i 1S-DATcL-l hat EX.COP-NEUT 'I have one hat' (Jacquesson 2005: 262) b. motosaikl gu-ja ko-ri motorcycle cL-l come-PRF 'A motorcycle came' (Jacquesson 2005: 262) (26) Kokborok: than-sa CL-l cempay basket b;J-thay-g;JnalJ nag-ga fruit-pass need-NEuT kag-cha CL-l japa basket nalJ-ga need-NEuT 'A basket full of fruits and ajapa (basket) are needed' (Annamalai 1976: 130) (27) Dimasa: boto} thai gini bottle CL 2 '2 bottles' (Dundas 1908: 7) 16 2.1.6 Case Inflections and Post-positions Case inflection is consistently the final element in the NP, preceded by a postposition - if part of the proposition. This behavior suggests that this too was the syntactic restriction for these elements in PBG. *PBG: DEMposs {ADJ} N-PL {ADJ} CL-NUM PP=CASE 2.1.7 Conclusion Word order within the BG is consistent across the family. This indicates that no divergence has taken place in the modem languages since the split from the proto- language. The rest of this chapter will focus on reconstructing PBG NP grammatical forms. 2.2. PBG Grammatical Forms Many grammatical NP elements throughout BG are cognate and can be attributed to PBG inheritance. There are, however, a number of divergent forms which evidence independent innovation. Both comparative and language-internal reconstruction are used to determine which grammatical morphemes reconstruct to the PBG level, as well as the diachronic sources for deviant forms. 2.2.1 Personal Pronouns and Demonstratives One might expect to find a clear correlation between pronominal forms across BG due to the family's relatively shallow time depth. However, as Thurgood (1985) 17 demonstrates, it is not uncommon for TB languages to innovate new pronominals, even within a short period of time. We find some evidence of this in the modem BG languages, as not all the person distinctions made are consistently found across the family, nor are all the forms cognate. Table 1 demonstrates the closeness in form for the 1S, 2S and IPI pronouns, which easily reconstruct to the original PBG pronominal forms. They also reflect the Proto-Tibeto-Burman system (Matisoff2003; Jacquesson 2005). Following the sound correspondences proposed by Joseph & Burling (2006), they can be reconstructed as *alJ, *nalJ and *cilJ, respectively. Table 1: Pronouns across Boro-Garo Garo Boro Deuri Dimasa Kokborok Rabha PBG PTB Is aJ](a) an a aJ] aJ] al] *ag *a.r; 2s na?(a) n;)n no nUl] mUI] nal] *nag *na.r; 2s n;'}nthan FORM 3s bi(a) bi ba bo b;) u u(a) 0 Ip na?chig jUIIJ jou jig CUll] cig *cig *cig an?chig Ip chig(a) EXCL 2p na?sol] n;'}nS;)f lou nishi n;)f;)k narol] 3p bisol] bis;)f bau bonishi irol] uamal] biph;)f urOI] The third person pronouns slightly diverge within the family. There is a strong correspondence across the family with respect to the fbi-initial forms, which suggests pre- PBG origin. This PBG pronoun reconstructs as *bi (Joseph & Burling 2006). We are thus 18 left with the task of determining whether the divergent u forms in Garo and Rabha also date back to PBG, or if they were more recently innovated. Demonstratives are typically a reliable etymological source for third person pronominals cross-linguistically. The BG demonstrative system, shown in Table 2, does not provide strong evidence for either claim. Table 2: Modern Boro-Garo demonstratives Garo Boro Deuri Dimasa Kokborok Rabha 'this' (h)i-(a) be la abo i i 'that' (h)u-(a) bi ba houbo u u We find a similar divergence in form among demonstrative pronouns as in the third person pronominal system. Cognate demonstratives in other TB languages, such as Tibetan, further support the reconstruction ofPBG *bi. (28) a. khyi-bo dog-DEM 'that dog' b. myi-bo man-DEM 'that man' (Franke 1929: 112) However, there is evidence elsewhere in TB that supports the reconstruction of *u 3s and *i 'this' vs. *u 'that' into PBG as well. Benedict (1983) reconstructs *(h)i 'this' and *(h)aw 'that' in Western Bodie, which are clearly cognate to the u vs. i demonstratives found in the modern BG system. This is especially true for the Garo allomorphs, hia and hua, respectively. Other evidence supporting a *u 3s reconstruction can be found in the PTB transitive verbal paradigm, outlined in Table 3 19 Table 3: The PTB Suffix Paradigm (from DeLancey 1988) Undergoer 15t 2na 3ra Actor 15t (-nal-I)) -I) 2na -I) (-na) 3ra -I) -na [-u] The -1) 1sand -na 2s suffixes developed from the PTB 1sand 2s pronouns, *1)a and *na(1)) (DeLancey 1988, 1989). Following this pattern, it is a logical step to assume that the 3>3 index has a diachronic connection to a 3s pronominal *u. Table 4 demonstrates how similar cognate forms are also found in the verbal paradigms other modern TB languages. Table 4: Transitive Paradigms for Mhai Kham, Rawang and Limbu (DeLancey 1988) Mhai Kham Rawang Limbu 1>2 I)a- V -n V -I) V -ne 1>3 V -1) V -I)U V -UI) 2>3 n~- V -n e- V -u ke- V -u 3>2 (0)- V -nu e- V ke- V -a 3>1 (0)- V -1)u e- V -I) V -al) 2>1 n~- V -n e- V -I)a ke- V -al) The numerous examples of cognate u forms across TB indicate that the divergent third person u in Garo and Rabha is not an independent innovation, but was rather inherited into PBG from an earlier ancestor, possibly even dating back to PTB. 20 It is not clear from the data what grammatical functions PBG *bi served compared to *u. Only one language demonstrates the continued use of both forms; Garo. Burling (1961: 39) writes that speakers alternate between ua and bia for the distal demonstrative, commonly relying on the ua form for written language, whereas bia is used orally. He also mentions that bia is used in reference to humans in the spoken language, while ua is used in reference to non-human animate and inanimate entities. This difference could reflect a split in PBG 3s and demonstrative pronominal system based on humanness and animacy. As the modern BG languages evolved, the system was simplified in the majority of the communities. In spite of the evidence supporting a reconstructible PBG *u, it is worth noting that the surrounding Indo-Aryan languages have similar demonstrative forms, shown in Table 5. Table 5: Indo-Aryan Demonstrative Pronouns 'this' 'that' Bengali l ea~ehi~ini O~Ul~unl Bh ' ·2oJPun i~hai u~hau~(h)unhi Assamese3 i~iye~eye hi~hiye~heye This cannot be interpreted as evidence that the BG u/i forms are borrowed, given the evidence for their TB provenience. However, it is certainly possible that the existence of similar forms in these neighboring languages has supported their preservation. 1 See Bandyopadhyay (1998) for an introductory grammar to the Bengali language. 2 See Pandey (2003) for a comparative analysis of Bengali and Bhojpuri. 3 See Grierson (1903) for a brief description of the Assamese language. 21 2.2.2 Adjectives: Verbal Derivation In order to behave as noun modifiers within the NP, stative verbs take a derivational prefix. The modem languages vary as to how productive this process is, however. Adjectives in Bodo are formed somewhat productively by prefixing either gV- or mV- onto a stative verb. The verb roots these affixes occur with are constrained, mV- being more restricted than gV-. The vowel in these prefixes is generally copied from the vowel in the verb root. Below are a few examples of adjectives derived from stative verbs in Bodo, including the limited number of examples of the mV- prefix cited by Bhattacharya (1977). (29) Bodo: a. bal] ~ 'to be more' 'more' b. zam ~ 'to be old' gj-zam 'old, ancient' (Basumatary 2005: 63) (30) Bodo: a. ga-khay 'sour' (Bhattacharya 1977: 206) b. ga-ran 'wise, one who knows' (Bhattacharya 1977: 207) (31) Bodo: ma-za1j ma-dam ma-fed ma-slid ma-se1j mu-day me-hukhub me-fem 'good, nice (inherently) 'good scented' 'indistinct, weak' 'weak' 'dry' 'small' 'busy' 'content' (Bhattacharya 1977: 208-9) 22 There are very few examples of the m-initial prefix in the other BG languages. In Rabha, we find mO-10l] 'naked (adj.)' (Joseph 2007: 785) and mi-j81) 'giddy, dizzy (adj, verb)' (Joseph 2007: 787). Other instances of this prefix in the family are rare, but if we explore other languages in the greater Bodo-Jinghpaw-Konyak family, there is evidence ofthis form. The similarity in form and function suggests common ancestry, and indicates that the m- prefix does reconstruct to at least the PBG level. (32) Jinghpaw: a. daw1] m