University of Ghana http://ugspace.ug.edu.gh A MORPHO-SYNTACTIC STUDY OF THE BULI VERBAL GROUP BY DARIUS ADJONG (10599671) THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON, IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE AWARD OF M.PHIL LINGUISTICS DEGREE JULY, 2018 University of Ghana http://ugspace.ug.edu.gh i DECLARATION I do hereby declare that, except for references to sources of information which have been duly acknowledged, this thesis is a result of my own original research, and that, to the best of my knowledge and belief, it has not been presented either in whole or in part for the award of any other degree or diploma elsewhere. ......................................................... ......................................... DARIUS ADJONG Date (Candidate) ......................................................... ......................................... DR. GEORGE AKANLIG-PARE Date (Supervisor) ....................................................... ......................................... DR. JAMES A. N. SAANCHI Date (Supervisor) University of Ghana http://ugspace.ug.edu.gh ii DEDICATION In memory of my cousin, Clement Atiim Apoochaab† And to my grandmother, Anyetekoniak AwongaaNtiim University of Ghana http://ugspace.ug.edu.gh iii ACKNOWLEDGEMENTS I wish to pay my obeisance to the LORD who has always been my shiled and the keeper of my soul. Working while pursuing a second degree can be very tedious, but He was my constant helper. Oft, when hopelessness set in, my Lord reminded me through The Brooklyn Tabernacle Choir’s ‘My Help’ that He neither slumbers nor sleeps in my trying moments. It is He who has been my sustenance and torchbearer throughout a financially-draining and the seemingly daunting task of a 2-year M.Phil study. If there was anyone who had the slimmest hope of climbing the academic ladder this far, it was certainly me. Lord, for being the torchbearer of the hopeless one, I say “May Your Majestic Name Be exalted In All The Earth And Above.” I am indebted to my supervisors and fathers, Dr. George Akanlig-Pare and Dr. James N. Saanchi, for the time they dedicated to reading this work to give me constructive feedbacks. In spite of their busy schedules, they gave me the needed assistance at all times. It takes empathetic supervisors like them to go beyond their supervisory roles to provide their students shelter. While Dr. Akanlig-Pare secured accommodation for me during my studies in the university, Dr. James N. Saanchi provided a full office space to write this thesis. Yes, that’s how caring and loving my supervisors are! They were not just my supervisors; they were my fathers. May God bless them bountifully. To the two lecturers who taught me Syntax, Prof. Emmanuel K. Osam and Prof. Kofi Saah, I say “Meda Mo Ase”. Being a functionalist, the former introduced Functional Grammar to me, and I have come to love it. The latter also provided me guidance on how to appreciate other theories of syntax, including Chomsky’s Government and University of Ghana http://ugspace.ug.edu.gh iv Binding Theory. I consider it a great privilege to have sat under the enviable tutelage of these two Professors. I would also like to express my profound gratitude to Dr. Isaac Nuokyaa-Ire Mwinlaaru (Department of English, University of Cape Coast) for his useful input in the work. I’m fortunate to be benefitting from the ideas of such a generous scholar. Our conversations over the phone and via e-mail provided me some useful information for this study. Thank you, Doc. I also wish to express my sincere gratitude to other faculty and non-teaching staff of the Linguistics Department. I implore the Lord’s blessings upon all the lecturers who taught me in my first year, namely Prof. Kofi Agyekum, Prof. Nana Aba Appiah Amfo, Dr. Evershed Amuzu, Dr. Fusheini Hudu, Dr. Yvonne Agbetsoamedo, Dr. Mercy Bobuafor , Dr. Seth Ofori, and Dr. Grace Diabah. They introduced me to several linguistics concepts which have consolidated my knowledge of Linguistics. My contact with Dr. Reginald Akuoku Duah, Dr. Clement Appah, and Dr. Paul Agbedor has influenced this work, although I didn’t get the opportunity to be taught by them. My thanks also go to the secretaries of the Department of Linguistics (Mrs. Tagoe and Mrs. Ocran), as well as the technician (Mr. Edmond), and the librarians (Mr. Kisseh and Mr. Amoah). I benefitted from their services a lot. I fondly remember my family and friends. Friends like Miss Ruth Karehina Bodua- Mango are rare to come by. Her empathetic and generous nature made my stay on campus comfortable and memorable. I’m grateful to my cousins, Francis Atiim Apoochaab and Thomas Apoochaab, and my friend, Saraphina Abila Abagi, for their eagerness to provide me Buli data whenever I called on them. To Clement Atiim Apoochaab†, who encouraged me to pursue further education and pledged to support University of Ghana http://ugspace.ug.edu.gh v me financially, I say “May The Lord Forever Keep You In His Bosom.” Atiim, it was your pledge of financial support that motivated me to think of acquiring a second degree. Death denied me your financial sponsorship, but your good counsel has proven invaluable to me in this life’s journey. Finally, I thank my course mates, Joseph, Daniel, Yvette and Nuuratu for the priceless comradeship they provided me. I have learnt a lot from each of them. University of Ghana http://ugspace.ug.edu.gh vi ABSTRACT This thesis explores the Verbal Group (VG) of Buli, a Gur language spoken in northern Ghana, with focus on three main aspects: the morphology of the Buli verb, the grammatical functions of Buli verbal particles, and the distribution of such verbal particles in Serial Verb Constructions (SVCs). Having its underpinnings in Functional Grammar (and skewed towards Functional Typological Syntax), a theory that views language as a set of tools for communication, its data is drawn mainly from naturally occurring spoken texts (spontaneous speech) which was recorded from a range of communicative contexts in Buli-speaking communities. The oral data, which was tape- recorded from informal communication contexts and controlled formal interactions (where necessary), was then transcribed for the analysis and description. These spoken texts were supplemented by elicited data, as well as data based on my native speaker intuition. The analysis is purely descriptive, with data drawn from other better-studied Gur languages (e.g. Dagaare, GurunE, and Dagbani) in order to set the discussion in a broader context. The findings show that although affixation is generally minimal in Buli, a dichotomy can be drawn between dynamic and stative verbs on morphological grounds; while the suffix {-i} is attached to the roots of most dynamic verbs, the suffix {-a} is attached to the roots of most stative verbs. Also, the study reveals some Buli verbal particles and discusses the grammatical functions that each of them performs. The findings further indicate that, verbal particles generally manifest once (before the first verb) within SVCs, and have scope over any other verb that follows. Also, a mixture of some functional categories, specifically the progressive and the perfective aspects, is permissible in Bul, and it is possible for the two morphological units of the University of Ghana http://ugspace.ug.edu.gh vii progressive marker bóràà (i.e. existential bóró and the aspectual marker à) to be split between two serialized verbs in an SVC. University of Ghana http://ugspace.ug.edu.gh viii TABLE OF CONTENTS DECLARATION i DEDICATION ii ACKNOWLEDGEMENTS iii ABSTRACT vi TABLE OF CONTENTS viii LIST OF TABLES xiii LIST OF FIGURES xiv LIST OF ABBREVIATIONS xv CHAPTER ONE 1 GENERAL INTRODUCTION 1 1.0 Introduction 1 1.1 Buli Language 3 1.1.1 Ethnography of Buli 3 1.1.2 Linguistic Features 5 1.1.2.1 Phonology 5 1.1.2.1.1 Consonant System 5 1.1.2.1.2 Vowel System 6 1.1.2.2 Morphology and Syntax 7 1.1.2.3 Tonal Features 9 University of Ghana http://ugspace.ug.edu.gh ix 1.2 Previous Studies and Justification for this Study 9 1.3 Problem Statement 10 1.4 Objectives of the Study 12 1.5 Research Questions 12 1.6 Significance of the Study 12 1.7 Organisation of Chapters 13 1.8 Chapter Summary 13 CHAPTER TWO 15 LITERATURE REVIEW, THEORETICAL FRAMEWORK AND METHODOLOGY 15 2.0 Introduction 15 2.1 The Verb 15 2.1.1 The Verbal System 19 2.1.2 Verbal Particles 23 2.1.3 Serial Verb Constructions (SVCs) 43 2.2 Theoretical Framework 54 2.3 Methodology: Data Type and Data Source 57 2.4 Chapter Summary 59 University of Ghana http://ugspace.ug.edu.gh x CHAPTER THREE 60 VERB MORPHOLOGY AND VERBAL PARTICLES 60 3.0 Introduction 60 3.1 The Buli Verb 60 3.1.1 Verb Types 60 3.1.1.1 Stative verbs 61 3.1.1.2 Dynamic verbs 61 3.1.2 Verb Morphology 66 3.1.3 The Verbal Group 71 3.1.3.1 The Buli Verbal Group 73 3.1.3.2 Tense, Time, and Aspectual Particles in Buli 73 3.1.3.2.1 The Habitual Marker ‘à’ 74 3.1.3.2.2 The Progressive Marker ‘bóràà’ 81 3.1.3.2.3 The Future Tense Marker ‘lì/àlì’ 85 3.1.3.2.4 The Past Time Marker ‘pòòm’ 88 3.1.3.2.5 The Present Time Marker ‘nyìèm’ 90 3.1.3.2.6 The Distant Past Marker ‘jàm’ 92 3.1.3.3 Negation Markers in Buli – ‘kàn’ and ‘àn’ 93 3.1.3.4 Other Verbal Particles 98 3.1.3.4.1 The Assertive Marker ‘yá’ 98 3.1.3.4.2 The Emphatic Marker ‘lá’ 102 3.1.3.4.3 The Affirmative Marker ‘kámā’ 103 3.1.3.4.4 The Purposive Markers ‘àlē sòà / nyīN’ 105 University of Ghana http://ugspace.ug.edu.gh xi 3.1.3.4.5 The Conditional Marker ‘dàn’ 108 3.1.3.4.6 The Repetitive Marker ‘pīlīm’ 110 3.1.3.4.7 The Positive Modal ‘zāā’ 111 3.1.3.4.8 The Negative Modal ‘kāā’ 113 3.1.3.4.9 The Modals ‘fE#’ and ‘zīkā /māgsīkā’ 114 3.1.3.4.10 The Hortative Marker ‘tín’ 116 3.1.3.4.11 The Optative Marker ‘tè (àtè)’ 117 3.1.3.4.12 The Eventuality Marker ‘yāā’ 118 3.2 Chapter Summary 119 CHAPTER FOUR 120 VERBAL PARTICLES IN SERIAL VERB CONSTRUCTIONS 120 4.0 Introduction 120 4.1 Serial Verb Constructions (SVCs) 120 4.1.1 Defining Serial Verb Constructions (SVCs) 123 4.1.2 Marking Tense, Aspect, and Polarity (TAP) in SVCs 128 4.1.2.1 Tense, Time and Aspectual Markers in Buli SVCs 131 4.1.2.1.1 The Future Tense Marker ‘lì’ in SVCs 131 4.1.2.1.2 The Distant Past Marker ‘jàm’ in SVCs 132 4.1.2.1.3 The Progressive Marker ‘bóràà’ in SVCs 133 4.1.2.1.4 The Imperative SVCs 137 University of Ghana http://ugspace.ug.edu.gh xii 4.1.2.2 Negation Marking in Buli SVCs 138 4.1.2.2.1 The Negative Perfective Marker ‘àn’ in SVCs 140 4.1.2.2.2 The Negative Future Marker ‘kàn’ in SVCs 145 4.1.2.2.3 The Negative Imperative SVC 147 4.2 Chapter Summary 150 CHAPTER FIVE 152 SUMMARY, FINDINGS AND RECOMMENDATIONS 152 5.0 Introduction 152 5.1 Summary and Findings 152 5.2 Recommendations 156 5.3 Chapter Summary 157 REFERENCES 158 University of Ghana http://ugspace.ug.edu.gh xiii LIST OF TABLES Table 1.1 The Buli Consonant System 6 Table 1.2 The Buli Pronominal System 8 Table 5.1 The Buli Verbal Particles and their Grammatical Functions 154 University of Ghana http://ugspace.ug.edu.gh xiv LIST OF FIGURES Figure 1.1 Genetic Classification of Buli 4 Figure 1.2 The Buli Vowel System 6 Figure 2.1 The Panare Verb System 15 Figure 3.1 Timeline for the use of ‘póóm’ 89 Figure 3.2 Timeline for the use of ‘póóm bóràà’ 90 University of Ghana http://ugspace.ug.edu.gh xv LIST OF ABBREVIATIONS 1SG First Person Singular 2SG Second Person Singular 3SG Third Person Singular 1PL First Person Plural 2PL Second Person Plural 3PL Third Person Plural ADV Adverbial AFF Affirmative Marker ASP Aspect ASS Assertive Marker COMP Complementizer COND Conditional Marker DEF Definite Marker DEM Demonstrative EMPH Emphatic Marker EVT Eventuality FOC Focus Marker FUT Future HAB Habitual HOR Hortative IMP Imperative MOD Modal NEG Negation/Negative Marker University of Ghana http://ugspace.ug.edu.gh xvi OPT Optative POS Positive PRT Particle (undefined) PERF Perfective PL Plural PROG Progressive Marker PST Past PURP Purposive Marker REP Repetitive Marker SG Singular SVC Serial Verb Construction SVO Subject Verb Object TM Time Marker University of Ghana http://ugspace.ug.edu.gh 1 CHAPTER ONE GENERAL INTRODUCTION 1.0 Introduction Of the different clausal elements in grammar, it is the verb that has often been described as indispensable. The fact that imperative sentences are made of a single verb corroborates this assertion. Thus, the verb can be described as the heart of the clause, since where a clause is made up of a single element, that element must necessarily be a verb. In some languages, grammatical functions like tense, aspect, mood and polarity are marked on the verb via affixation. Thus, verbal affixes are attached to the verb to express functional categories which make up the entire verbal systems of such languages. Verbal affixes are also productive linguistic units, because some of them serve as derivational morphemes that change the word class and the meaning of the verb. For instance, it is not difficult to tell that, apart from marking past tense in English by attaching the inflectional suffix {-ed} to the regular verbs ‘pray’ and ‘climb’ to yield ‘prayed’ and ‘climbed’ respectively, we can derive the nouns ‘prayer’ and ‘climber’ when we attach the noun forming derivational suffix {-er} to the verbs. In tone languages, grammatical functions are often borne by verbal particles, and sometimes by tone which may interact with verbal particles. Like Gur languages such as GurunE, Dagaare, Dagbani, and Sisaali, Buli possesses verbal particles that are University of Ghana http://ugspace.ug.edu.gh 2 assigned essential grammatical and discourse functions, including the marking of tense, aspect, modality, and polarity. Consider each of the Buli constructions presented in (1). 1) a. Nípōwá à nyū dáám kámā. woman.DEF HAB drink alcohol AFF ‘The woman drinks alcohol.’ b. Nípōwá bóràà nyù dáám. woman.DEF PROG drink alcohol ‘The woman is drinking alcohol.’ c. Nípōwá lì nyu# dáámú. woman.DEF FUT drink alcohol.DEF ‘The woman will drink the alcohol.’ d. Nípōwá ka$n nyu# dáámú. woman.DEF NEG.FUT drink alcohol.DEF ‘The woman will not drink the alcohol.’ In each of the examples in (1), the verb nyū ‘to drink’ co-occurs with verbal particles to mark either tense, aspect, modality or polarity. The particle à in (1a) is a habitual marker, which indicates the action of the subject occurs for an extended period of time (i.e. the act of drinking is a habit of the woman). As (1b) shows, bóràà1 is a progressive marker. Also, as shown in (1c) and (1d), the particles lì and ka$n mark futurity; while lì indicates that the action denoted by the verb will take place after the speech time, 1 bóràà, the progressive marker, is derived from the fusion of the existential bóró and the aspectual marker à. University of Ghana http://ugspace.ug.edu.gh 3 kàn is a future negation marker which indicates that an action will not take place after the speech time. Buli has a number of these verbal particles whose syntactic distributions vary, depending on the clausal elements with which they occur. In this descriptive study, we first explore the morphological make-up of Buli verb, and then discuss the various grammatical and discourse functions of verbal particles that co-occur with the Buli verb In order to show how the verbal particles relate with a series of verbs that occur in a single clause, the work also explores the distribution of Buli verbal particles in Serial Verb Constructions (SVCs). 1.1 Buli Language 1.1.1 Ethnography of Buli Buli, a Gur language of the Oti-Volta family, is spoken in the Upper East region of northern Ghana. Native speakers of Buli are the autochthonous inhabitants of towns such as Sandema, Siniensi, Wiaga, Kadema, Fumbisi, Chuchuliga, and Kanjag. Unlike GurunE, Dagaare, and Dagbani which have been sanctioned as mediums of instruction, Buli is not yet used as a medium of instruction. With the exception of radio stations like Bulsa Radio, Ura FM, and Nabina FM that broadcast mainly in Buli2, it is rare to hear it spoken on radio or television stations in Ghana. Three dialects, namely the Northern dialect (spoken mainly in Chuchuliga), the Central dialect (spoken in Wiaga, Sandema, Siniensi etc.), and the Southern dialect (spoken in villages such as Fumbisi, Gbedema, 2 All the three radio stations (Bulsa Radio, Ura FM, and Nabina FM) are found in Buli- speaking communities. University of Ghana http://ugspace.ug.edu.gh 4 and Yiwaasi) have been identified in the literature on Buli. This study focuses on the central dialect, which has received some attention by earlier scholars, with focus on the verb and verbal particles which nhave received minimal attention. The classification of Buli into a language family has seen some inconsistencies. For instance, Greenberg (1963) presented Buli as Kanjaga3 and classified it together with eight other languages (Awuna, Kasena, Nunuma, Lyele, Tamprusi, Degha, Siti, and Sisaali) under the Grunsi subfamily of Gur languages. There are people (including some native speakers in Southern Ghana) who still refer to Buli as Kanjaga. Considering how deeply entrenched the error has become, it is unlikely to be corrected any time soon. Sandema is the district capital and paramountcy of the Bulsa traditional area. Akanlig-Pare and Kenstowicz (2003) peg the number of Bulsas in the Upper East region at 100,000, and Lewis, Simons & Fennig (2015) estimate that the number of speakers in northern Ghana alone is 15, 000. It is expected that the number of Buli speakers exceeds these figures by far, since some Bulsas have abandoned the rural life of subsistence farming in their hometowns and migrated to southern Ghana to pursue other professions. Intermarriage is one of the effects of this migration, so it is not uncommon to find Bulsas who have permanent residences in southern Ghana where they spend their whole life. Native speakers who reside in Bulsa traditional areas are subsistence farmers who cultivate rice, millet, beans, and groundnut. They also rear sheep, goats, donkeys, cattle, and guinea fowls. Figure 1.1 illustrates the linguistic affinity of Buli within the Gur language family: 3 Kanjaga is coined from Kanjag, a Buli speaking area in Upper East. University of Ghana http://ugspace.ug.edu.gh 5 Figure 1.1: Genetic Classification of Buli 1.1.2 Linguistic Features This section presents a brief discussion of some aspects of the phonology, morphology, and syntax of Buli. 1.1.2.1 Phonology 1.1.2.1.1 Consonant System Phonologically, Buli has a total number of twenty-three (23) consonant sounds. These consonants are presented in Table 1.1 below. University of Ghana http://ugspace.ug.edu.gh 6 Table 1.1: Buli Consonant System Bilabial Labio- Alveolar Alveo- Palatal Velar Labio dental Palatal Velar Stops p b t d k g k p gb Fricatives f v s z Affricates t˛ d¸ Nasals m n ¯ N Nm Flap r Lateral l Glide j w 1.1.2.1.2 Vowel System Buli has nine (9) short vowels and three (3) long vowels. Each of the vowels can be nasalized. Figure 1.2 below shows the vowels in Buli. Front Back high i i: u u: I V mid-high e o mid-low E ç low a a: Figure 1.2: Buli Vowel System University of Ghana http://ugspace.ug.edu.gh 7 1.1.2.2 Morphology and Syntax As it is for most Gur languages, Buli has a canonical SVO clausal pattern. Thus, an unmarked clause will have its verb preceded by its subject and followed by its object. It is also possible to topicalize an object or a subject in order derive different clausal patterns. Consider example (2) below. 2) a. Akambe kò bó:súkú. Akambe kill.PERF python.DEF ‘Akambe killed the python.’ b. Bó:súkú, Akambe kò. python.DEF Akambe kill.PERF ‘It is the python that Akambe killed.’ Sentence (2a) is an unmarked sentence, since each of its clausal elements occupies their traditional syntactic positions – the verb kò ‘kill is preceded by the subject Akambe and followed by the object bó:súkú ‘the python’ so that the canonical SVO pattern is derived. This is not the case for (2b) where the object bó:súkú ‘the python’ is topicalized such that it precedes the subject Akambe to yield the clausal pattern OSV. Nouns that occur at the subject and object positions may occur alone or with post- modifiers like adjectives, numerals, or determiners. Sulemana (2012) shows that there is no fixed order for multiple adjectives that post-modify the Buli noun. Such multiple adjectives, however, typically precede other modifiers within a complex noun phrase. Locative elements are body part terms such as zúk ‘head’ and NááN ‘back’ which have undergone grammaticalization to mean ‘on’ and ‘behind’ respectively. This is in line with Heine and Kuteva’s (2007:13) observation that concrete body part terms in most University of Ghana http://ugspace.ug.edu.gh 8 African languages undergo grammaticalization such that they come to be ‘reinterpreted as locative adpositions in certain specific contexts.’ In the case of Buli, and like many other Gur languages such body part terms have not undergone desemanticization. Thus, both the lexical and the grammaticalized meanings of such body part terms co-exist, making the terms polysemous. In Buli, case marking is not done via morphological markings on the verb. Instead, word-order determines the case of arguments. With regard to grammatical relations, the language has a nominative-accusative system; the nominal which occurs at the subject position is the nominative case and the nominal which occurs at the object position is the accusative case. Buli has different pronouns whose choice is dependent on the animacy and person of the referent, as presented in Table 1.2. Table 1.2: The Buli Pronominal System PERSON SUBJECT OBJECT 1 n,$ mí mí SG 2 fí fí 3 wà wà 1 tàmá, tì tàmá, tì PL 2 nàmá, nì nàmá, nì 3 bà bà Inanimate SG dì/kà/kù/bù dì/kà/kù/bù PL Nà/sì/tì Nà/sì/tì University of Ghana http://ugspace.ug.edu.gh 9 1.1.2.3 Tonal Features Buli has three (3) identifiable level tones: a high tone represented by the acute accent ( @), a mid-tone represented by the Macron (¯), and a low tone represented by the grave accent ( ›). Tone is contrastive in Buli. Consider the distinct meanings created by tonal contrasts in the words provided in (3). High Mid Low (3) a. síúk ‘path’ sīuk# ‘navel’ sìùk ‘fish’ b. ká ‘FOC’ ka # ‘to swear’ kà ‘it’ c. kç@k ‘mahogany’ kç#k ‘feather’ kç$k ‘ghost’ d. túrí ‘bean’ tur# ī ‘ear’ tùrì ‘abruptly’ e. bíík ‘child’ bīīk ‘to grow fat’ bììk ‘language’ f. baN# ‘to forget’ bàN ‘bangle/lizard’ 1.2 Previous Studies and Justification for this Study Buli has received a few documentation over the past few decades. It is noteworthy that such works have focused on its phonology, with Akanlig-Pare (1988, 1994, 1996, 1997, 2002, 2005) exploring in depth its sound system, syllable structure, phonological processes, and especially its tonal system. The same author has co-edited a book, Studies in Buli Grammar, which is a collection of works by different authors on selected aspects of Buli grammar (including interrogative clauses, serial verb constructions and University of Ghana http://ugspace.ug.edu.gh 10 embedded clauses). Although not comprehensive, the work serves as a significant foundational material for later explorations of the language. Other scholars who have explored aspects of Buli syntax are Hiraiwa (2003) and Schwarz (2016); while the study of the former investigated relative clauses, that of the latter investigated thetic structures in Buli grammar. Both studies, like previous studies, adopted the formalist approach. Unlike the Masters and Doctoral dissertations of Akanlig-Pare (1994; 2005) which focused on Buli Phonology, those of Schwarz (1995; 2005) focused on its noun class system, morpho-syntax and tonology. Schwarz’s (1995; 2005) dissertations, however, are in German. Also, Schwarz’s few articles written in English are on Buli phonology and grammar, with those on grammar focusing on isolated aspects such as preverbal negative markers. The work of Sulemana (2012) gives us some knowledge about the patterning of elements of the Buli noun phrase, but it also employs the generative approach (specifically, Chomsky’s X-bar theory). Previous studies neither focused on the Buli verb nor explored the language from a functional perspective; hence, the need for this study. 1.3 Problem Statement Unlike Gur languages like Dagaare, Dagbani and Kasem which have been sanctioned as mediums of instruction, Buli is not yet used as a medium of instruction. Consequently, its literate population is suspected to be less than 5%. In comparison, Lewis’ (2009) estimate of Dagaare4 speakers’ literate population is between 5% - 10%. With the exception of radio stations like Bulsa Radio, Ura FM and Nabina FM that 4 Dagaare, like Buli, belongs to the Gur language of the Oti-Volta family, and it is one of the eleven languages sanctioned as media of instruction in schools. University of Ghana http://ugspace.ug.edu.gh 11 broadcast mainly in Buli5, it is rare to hear it spoken on radio or television stations. Listeners of all these radio stations live in Buli speaking communities like Sandema, Wiaga, Fumbisi and Siniensi. We saw in in Section 1.2 that, although Buli has garnered some interest among linguists, it is its Phonology that has been explored most, leaving aspects like morphology, syntax and semantics understudied. It is also evident from my discussion that the few studies on the language’s syntax have been limited to the formalist approach. This makes it necessary for the exploration of the language from a functional approach, in order to draw the language into functional typological studies and give it the rigour of mainstream linguistics analyses. The previous studies provide valuable knowledge on the structure of Buli; however, they rely mainly on constructed data rather than copious naturally occurring texts. Therefore, they provide us little knowledge about Buli as a tool for realizing meaning, and they do not show how the language has developed and evolved through time. The functionalist approach values the contribution of diachronic data to the explanation of synchronic phenomena. Since language is in constant change, there is the need for the adoption of the functionalist approach in our description of languages, including understudied languages like Buli. It is these limitations that have provided the impetus for the present study. Drawing on diachronic evidence (where available) to account for the synchronic data, the study will describe the Buli verb and verbal particles. University of Ghana http://ugspace.ug.edu.gh 12 1.4 Objectives of the Study This study aims at: 1. describing the morphological make-up of the Buli verb. 2. identifying Buli verbal particles, and describing their grammatical functions. 3. describing the distribution of Buli verbal particles in SVCs. 1.5 Research Questions The study is guided by the following research questions: 1. What is the morphological make-up of the Buli verb? 2. a. Which verbal particles co-occur with the Buli verb? b. What grammatical functions do Buli verbal particles perform in the clause 3. How are Buli verbal particles distributed in SVCs? 1.6 Significance of the Study The present study is significant in three (3) main ways. First, since functional typology is yet to attract much attention in West Africa (Caffarel, Martin and Matthiessen, 2004), the study serves as a major contribution towards functional typological studies in the West African context. This provides the impetus for the exploration of Buli in relation to better studied languages (especially, those within the Gur family of languages in Ghana), thereby contributing to existing knowledge on the similarities and differences between languages in the word. Also, since the description of the verb of any language leads to a discussion of other clausal elements, this study provides a foundation for writing a grammar for Buli, and serves as a reference point for the production of literacy materials that will ensure effective pedagogy in Buli. University of Ghana http://ugspace.ug.edu.gh 13 In relation to the above, the present study will enhance the development of Ghanaian languages, by providing a model for exploring other endangered indigenous Ghanaian languages alike. Such explorations will be useful for linguistics studies like semantics, discourse analysis, translation, as well as second language pedagogy. 1.7 Chapterisation of Thesis The rest of the thesis is divided as follows. Chapter 2 reviews related literature on the verbal system of other languages, especially Niger-Congo languages (within which Buli falls). Grammatical functions of verbal affixes and verbal particles in such languages, as well as their distribution in SVCs are discussed in some detail. The theory that underpins this study, Functional Grammar (FG), is also discussed briefly with focus on some tenets on which the theory is founded. The methodology of the study, comprising my data source and type, is also presented in the second chapter. In Chapter 3, I discuss the morphology of the Buli verb, showing affixes that occur with the Buli verb. In the same chapter, I identify Buli verbal particles and discuss their grammatical functions, and further show the distribution of some of these verbal particles in SVCs in the fourth chapter. In Chapter 5, I present a summarize my findings, make recommendations for future studies, and conclude the study. 1.8 Chapter Summary This introductory chapter has provided an essential background to the study. It has shown the linguistic and genetic affinity of Buli, and provided some ethnographic information about its speakers, commenting on their historical origin, population size, geographic location, and occupation. The chapter has also discussed some linguistic University of Ghana http://ugspace.ug.edu.gh 14 features of Buli briefly. Other relevant pieces of information that provide the foundation for the study, including previous studies and justification for the current study, the research problem, objectives of the study and research questions, as well as the relevance of this study have also been presented. University of Ghana http://ugspace.ug.edu.gh 15 CHAPTER TWO LITERATURE REVIEW, THEORETICAL FRAMEWORK AND METHODOLOGY 2.0 Introduction This chapter discusses some related literature and the theory that underpins this study. Works reviewed here include those on Buli, as well as works on other closely related Gur languages like Dagaare, GurunE, and Dagbani. A few works are drawn from non- Gur languages, inasmuch as they are related to the current study and provide some basis for drawing certain conclusions. The chapter also highlights the methodological procedure for the study, specifically the type and source of my data. 2.1 The Verb Verbs, as a grammatical category, express concepts that have the least temporal stability (Givón 2001:52; Payne 1997:47). Givón (2001) uses a time-stability scale to compare verbs with nouns, noting that while nouns occupy the most time-stable end of the time- scale, verbs occupy the opposite end of the scale. His description of verbs as temporally unstable clausal elements is based on his observation that verbs present a concept as changing (rapidly or subtly). Worth noting is how he distinguishes the verbs, using prototypes to show that verbs are graded such that all verbs do not necessarily possess identical features or qualities. This implies that whether a lexeme is a verb or not is not dependent on its ability to exhibit all properties that define the class of verbs. For instance, verbs such as beat, kill, and slap are prototypical verbs, since they code rapid actions that are capable of causing a change in the state of a noun entity (patient). Such University of Ghana http://ugspace.ug.edu.gh 16 verbs have arguments that are assigned the semantic role of agents who volitionally initiate the action depicted by the verb. Other verbs (less prototypical ones) may code ‘events of longer duration’ or ‘longer enduring states’. Verbs that code events of longer duration include read, work etc. and those that code events of enduring states include love, sleep etc. (cf. Givón 2001:52). The choice of a particular verb type, be it stative or dynamic, will determine the elements that can co-occur with the verb. Thus, as we will see in the next chapters, the presence of certain Buli verbal particles in a clause is determined by the type of verb used. Payne (1997) also asserts that the extent to which a linguistic form meets the morphosyntactic pattern of prototypical verbs determines whether it is a verb or not. A linguistic form that follows the morphosyntactic pattern closely has the propensity of being a verb, while a linguistic form that does not follow the morphosyntactic pattern is unlikely to be a verb. Payne (1997:47) puts such morphosyntactic properties of verbs into two groups – distributional or configurational properties, and structural properties. While the former is concerned with the functions that verbs perform in larger constructions such as verb phrases, clauses and texts, the latter deals with the internal make-up or composition of the verb itself. The latter, structural properties, are more relevant to this study, since they reveal certain unique features of the verb that make it possible for it to mark grammatical functions like tense, aspect, modality, and agreement. University of Ghana http://ugspace.ug.edu.gh 17 Payne (1997) further observes that in polysynthetic languages verbs are usually the most complex lexical categories as compared to other categories. He attributes this to the fact that, in such languages grammatical operations such as verb agreement/concord markers, semantic role markers, valence decreasing devices, evidentials, and tense/aspect/mode markers are borne by the verbs. He uses the verb system of Panare, a Cariban language spoken in southern Venezuela, to support his claim. According to him, the verb system of Panare can be complex, since it is inflected severally to express grammatical functions. This is illustrated in Figure 2.1 below. Figure 2.1: Panare Verb System (Payne 1997:61) As Figure 2.1 shows, the Panare verb is complex, since most grammatical functions are incorporated into the verb morphologically via affixation. Yet, Payne (1997) notes that the structure of the verb can be more complex than what has been presented here. It should be pointed out that, even in isolated or analytical languages which are said to have a simple verb due to the lack of multiple affixes, it is not uncommon to have a few grammatical functions (especially tense, aspect, and negation) expressed by verbal affixes and other functions borne by verbal particles. Saah (2003) has shown that affixes University of Ghana http://ugspace.ug.edu.gh 18 that are attached to Akan6 verbs include personal pronouns, tense, aspectual and negation markers as illustrated in example (1). (1) a. ç- re- n- di akutu. 3SG-PROG-NEG-eat orange ‘S/he will not eat oranges.’ b. ç- bE- di akutu. 3SG-FUT-eat orange ‘S/he will not eat oranges.’ c. Kofi fa-a ahina no Kofi take-PST pot DEF ‘Kofi took the pot.’ (Saah 2003:181) It can be inferred from the examples that verbal affixation is a productive means of marking tense, aspect, polarity and other such functional categories. In (1a) the prefix {re-} is a progressive marker, and the prefix {n-} functions as a negative marker. The negation marker is a homorganic nasal affix, because it undergoes a place of articulation change to be in consonance with the place of articulation of the consonant that precedes the verb to which it is attached (Dolphyne1987; Saah 2003). The prefix {-bE} in (1b) is a future marker, and {ç-} in both (1a) and (1b) is a pronoun prefix. 6 Akan is a Kwa language of the Niger-Congo family. It is spoken in most parts of Ghana. University of Ghana http://ugspace.ug.edu.gh 19 In Akan past tense7 marking is not restricted to the addition of a particular affix to the verb. Saah (2003) reports that past tense is marked by lengthening the final vowel of the verb, if the verb is followed by an object or a complement. This implies that the suffix {a-} in (1c) is not the only past tense marker in Akan; it is possible for other vowel suffixes to mark past tense, where the verb is followed by an object or a complement. Also, in sentences where the object or complement is not syntactically realized such that the verb terminates the sentence, a low tone suffix {-ì/-è} is attached to the verb, taking into consideration the ATR features of the vowel in the stem verb (Dolphyne 1987; Osam 1994). 2.1.1 The Verbal System Discussing the verbal system of Gur languages and that of other Niger-Congo languages is a strenuous task, because the status or functions of some verbal particles remain unclear (cf. Welmers 1973). Welmers (1973) observes that, although verbal particles often mark functional categories like tense, aspect, modality and polarity, such particles do not always perform these grammatical functions. He adds that there are some verbal particles which clearly fall outside such functional domains but have been misconstrued and consequently assigned wrong grammatical labels. Consequently, Welmers (1973) adopts the term “verbal construction”, claiming that such a term saves him from the 7 What has been treated as the past tense in the works of Dolphyne (1987) and Saah (2003) is considered as Completive aspect by Osam (1994). University of Ghana http://ugspace.ug.edu.gh 20 inconsistencies and complications that may characterize his description of the verb systems of Niger-Congo languages. [the term ‘verbal constructions’ avoids] the confusions and complications inherent in distinguishing categories such as tense, aspect, and mode. Some constructions, to be sure, may have specific reference to time, such as past; others may have specific reference to mode, such as conditional. But the forms or constructions of Niger-Congo languages do not fall into neat sets with different types of morphological structure. For each construction, of course, the semantic reference must be defined, and in many cases familiar labels are adequate to suggest the function of given constructions. In many other cases, however, there has been widespread misuse of labels, and apparently widespread misunderstanding of usage. (Welmers 1973:344) The fact that some researchers often hold divergent views about the labels of verbal particles in Niger-Congo languages affirms this assertion. For example, the Dagbani post-verbal particle la, which occurs frequently in Oti-Volta languages, has been given different labels and interpretations in the literature. Olawsky (1999) identifies it as an emphatic marker, but glosses it as a focus marker and an aspectual marker which has both habitual and aspectual readings. Although Fusheini (2006) and Issah (2013) identify it as a focus marker, they hold different views about the type of focus marker it is; while the former claims that the particle marks contrastive focus on full NPs, adjuncts, emphatic pronouns and wh-phrases, the latter asserts that it marks presentational focus on NP complements, adjuncts or an entire VP. A similar particle has been treated as a focus marker in both Dagaare and GurunE by Bodomo (1997) and Atintono (2005) respectively. The use of different theories is one of the factors that University of Ghana http://ugspace.ug.edu.gh 21 sometimes account for the different labels assigned to functional categories by different researchers, but this is not always the case. The works of Fusheini (2006) and Issah (2013) are relevant to this study, because la exists in Buli also. As we will see in the third chapter, the particle functions as an emphatic marker. Dixon (2010) notes that even though identical linguistic labels are used to describe different grammatical categories, such labels do not necessarily perform the same functions or carry the same meaning. Assigning grammatical categories of different languages identical labels, according to him, makes the field of Linguistics a science. Consequently, he points out that the features of any two languages vary significantly; hence, using the features of a language to describe another can be misleading. The empirical data for the present study will help us identify the functions of the Buli verbal particles and label them as such. In several Niger-Congo languages, researchers have given functional categories different labels. Harley’s (2008) observation that grammatical distinctions that were previously analysed as tense are being reanalysed as aspect or modality supports this claim. The fact that similar verbal particles in Niger-Congo languages have been given different labels is not surprising. This is because, it is often difficult to ascribe unique functions to certain morphemes or markers in Niger-Congo languages. It is also the case that some particles perform multiple functions, and different particles may perform similar or identical grammatical functions. In still other cases, it is tone that is assigned grammatical and discourse functions. Literature on Akan show that the verb system of the language has been analysed differently by different scholars. While earlier studies (c.f. Welmers 1973; Stewart 1971; Boadi 1974; Dolphyne 1987) suggest that Akan possesses a tense-aspect system, Osam (1994) describes Akan as primarily an aspectual University of Ghana http://ugspace.ug.edu.gh 22 language that is in the process of developing tense markers. His assertion seems to support Manfredi’s (1991) observation that Kwa languages are tenseless. It should, however, be emphasized that the use of different theories may account for the application of different grammatical labels to identical functional categories. Welmers (1973) observes that conditionality is expressed by construction markers in Swahili and Kpelle8. He further proposes the formula Pronoun – Construction Marker – Verb Base – Affix with verb base (P-C-V-A) for the verbal systems of most Niger- Congo languages, noting that: the verbal systems of Niger-Congo languages are best described in terms of a unidimensional list of ‘verbal constructions’ rather than in terms of a bi-dimensional or a multi-dimensional grid with intersecting categories such as tense, aspect, and mode. (Welmers 1973:343) This unidimensional character of the verbal system of Niger-Congo languages, he notes, is so obvious in the Bantu languages that researchers who use the term “tense” to describe certain morphological forms in their discussion of time (present, past and future) are ‘compelled’ to use the same label to describe other similar morphological forms to generate weird terms like “subjunctive tense” and “conditional tense”. The issue of assigning different labels to the same verbal particles persists in Gur languages. This, as I have mentioned, is sometimes attributable to the different theoretical positions of scholars. Besides, most Gur languages possess a large number of verbal particles, and this makes the grammatical and discourse functions of some of 8 Swahili and Kpelle are Bantu languages within the Niger-Congo group of languages. University of Ghana http://ugspace.ug.edu.gh 23 them unclear in some cases. For instance, Bendor-Samuel (1971) identifies as many as 30 verbal particles for Dagbani. The functions of the verbal particles are predominantly grammatical; hence, their absence may distort the intended meaning of the speaker. In the next sections, I show the grammatical functions of such verbal particles by reviewing literature drawn largely from Gur languages and minimally from other less related languages. Specifically, their roles as tense, aspect, mood and polarity markers have been given much attention. The syntactic behaviour of these particles in SVCs have also been discussed in some detail. 2.1.2 Verbal Particles Literature on Niger-Congo languages, including the few works reviewed in this study, suggest that languages within the Niger-Congo family are mostly aspectual (cf. Bhat 1999; Welmers 1973; Dakubu 1989; Bodomo 1997; Saanchi 2006). Such studies seem to support Welmers’ (1973) assertion that “the linear concept of Time, with a Past, Present and Future, is foreign to African thinking, in which the dominant factor is a virtual absence of the Future.” This is debatable, since there are studies on African languages that show that some African languages possess tense markers. For example, Akanlig-Pare (2005) and Schwarz (2005) have identified and discussed the preverbal particle lì as a future tense marker in Buli. Works on Kwa languages, including Saah (2003), also give us some reason to believe that African languages are not strictly University of Ghana http://ugspace.ug.edu.gh 24 aspectual; such languages show distinctions between different tenses with either affixes or verbal particles. The findings of Bhat’s (1999) typological studies on languages indicate that Niger- Congo languages are aspect-prominent. The fact that Olawsky (1999), Atintono (2005) and Saanchi (2006) respectively treat Dagbani, GurunE, and Dagaare as aspectual languages seem to confirm the view that Gur languages and by extension Niger-Congo languages are predominantly aspectual. In her work, Schwarz (2005) discusses two main aspects (perfective and imperfective) for Buli, and shows that the language marks future tense. Her description contradicts the works of Kröger (1992) and Akanlig-Pare (2005) which propose three tenses (present, past, and future) for Buli. It should, however, be noted that Akanlig-Pare (2005) indicates that the present tense and the past tense are realized with the perfective and the imperfective aspects. This gives us the reason to believe that Buli is predominantly aspectual. Apart from the future tense which is clearly marked by àlì or its variant lì, other verbal particles are aspectual markers. The labels which Kröger (1992), Schwarz (2005) and Akanlig-Pare (2005) assign to Buli verbal particles provide us a starting point for discussing the functions of these verbal particles. University of Ghana http://ugspace.ug.edu.gh 25 Bodomo (1997) notes the dichotomy between the perfective and the imperfective aspects of Mabia verbs. He observes that while Mabia verbs that mark the perfective aspect describe an event or action as ended or completed, those that mark the imperfective aspect describe the action as uncompleted or yet to be completed. Bodomo (1997) further identifies two aspects of the Dagaare verb – the perfective and the imperfective, adding that the perfective can be divided into two types on morphological grounds. The first type of the perfective, which he describes as the perfective aspectual form, does not have any morphological markings; it has the citation form or the dictionary entry form. He notes that the other type of the perfective, the perfective intransitive aspectual form, which occurs when the verb is not followed by any object, takes suffixes that interact with pre-verbal particles to express temporal, aspectual, and polarity features. Bodomo (1997) does not divide the imperfective into types like the perfective aspectual form. However, he notes that the imperfective aspectual form, like the first type of the perfective aspectual form, takes suffixes, using the examples in (2) to illustrate his claim: (2) a. teEne dictionary form b. teEne perfective aspectual form c. teEn-EE perfective intransitive aspectual form d. teEn-nE imperfective aspectual form (Bodomo 1997:81) As the examples reveal, the perfective aspectual form teEne is identical to the dictionary form which takes no suffix, but the perfective intransitive aspectual and the imperfective aspectual form take the suffixes {-EE}and {-nE} respectively. While {-EE} shows that the action coded by the verb is bounded (completed), {-nE} indicates that University of Ghana http://ugspace.ug.edu.gh 26 the action is unbounded (progressive). Buli verbs do not take suffixes; the perfective is marked with a low tone on the verb, and the progressive is marked by bóràà. In his discussion of the Dagaare Verb Phrase (VP), Saanchi (2006) shows that it is possible for the Dagaare main verb to co-occur with optional pre-verbal and post-verbal particles. These pre-verbal particles, he indicates, have the grammatical function of marking tense, aspect, and polarity. Following Dakubu (1989) and Bodomo (1997), Saanchi (2006) notes that a single verbal particle may perform more than one grammatical function. He supports his claim by indicating that the pre-verbal particle kU$N is a future negation marker. The function of this particle is akin to the future negative marker kàn in Buli which indicates that an action is not going to take place in the future. The multiple functions that some particles perform are cross linguistically attested. A typical case in point is the future tense marker lì which can indicate a wish, an intention, a prediction or a hope, aside marking futurity. Saanchi (2006) identifies as many as fifteen (15) pre-verbal particles {ba, da, daN, dIE, kUN, là, maN, mUç, na, naN, sIrIN, ta, tçç, tI, yaaN} and only one post-verbal particle lá in Dagaare, and shows the different discourse function(s) that each of them performs. Contrary to Bodomo’s (1997) classification of time adverbials za)a) ‘yesterday’ and daari ‘three days ago’ as pre-verbal particles, Saanchi (2006) asserts that they are not pre-verbal particles. Considering the fact that the two time adverbials, like most adverbials, are syntactically mobile, he proposes that they are not pre-verbal particles. He further notes that the particles should be treated as adverbials, recalling that Dong (1981) does not treat them as verbal particles. In the linguistics literature, University of Ghana http://ugspace.ug.edu.gh 27 adverbs are not considered as verbal particles, because adverbs are optional elements which can occupy different positions within the clause. Thus, their position in relation to the verb is less rigid. The fact that verbal particles are often grouped into two types (pre-verbal and post-verbal) according to their position in relation to the main verb indicates that verbal particles are different from adverbs. Also, the work of Saanchi (2006) reveals that Dagaare marks tense. He observes that the pre-verbal particles da, maN, and ba mark past tense, habitual aspect and negation respectively. He further shows that while là is a pre-verbal repetitive marker that indicates a situation that has occurred before, its high tone counterpart lá is the only post-verbal marker in Dagaare and it functions as an affirmative marker. This is exemplified in example (3). (3) a. a bie da ba maN wa kyE. DEF child PAST NEG HAB come here ‘The child did not use to come here.’ b. a pIIrI da maN là çç lá a kyI. DEF sheep PAST HAB REP chew AFF DEF millet ‘The sheep used to re-eat the millet.’ (Saanchi 2006:57) As the examples in (3) reveal, pre-verbal particles are assigned different grammatical functions in Dagaare. In (3a) past the tense marker da precedes the negative marker ba which also precedes the aspectual marker maN. Saanchi (2006) observes that the order of occurrence is invariable, since any attempt to re-order them yields ungrammatical sentences. Apart from the three pre-verbal particles in (3b), there is also a post-verbal particle là which serves as an affirmative marker. University of Ghana http://ugspace.ug.edu.gh 28 In his discussion of Gurunԑ grammar, Atintono (2004; 2013) observes that aspect is viewed as more important than tense and modality. He refers to particles that co-occur with the verb as verbal modifiers, and identifies nine groups for them. He, like Bodomo (1997), treats time adverbs as verbal particles. Among the particles which he identifies are the time adverbs daam ‘yesterday’, daarԑ ‘two days ago’, and dayita ‘three years ago’. A careful study of the items that he describes as verbal modifiers will reveal that adverbs fall outside what he describes as verbal modifiers. The fact that his examples do not include other types of adverbs (e.g. place and manner) supports this claim. Besides, Atintono (2004) attempts to arrange his verbal modifiers sequentially to show their syntactic distribution in relation to the verb. As we are aware, time adverbs, like most other types of adverbs, are optional elements that are mobile within the sentence. Comrie’s (1976) observation that a number of West African languages have no tense markers supports the claim that Niger-Congo languages are predominantly aspectual. Noting that some West African languages lack tense markers, he explains that the Imperfective aspect relates closely with the present tense, and the perfective aspect relates with the past tense. Following this assertion, Olawsky (1999), who identifies two forms of the Dagbani verb (the perfective and the imperfective), indicates that knowledge of the relationship between the perfective and the past on one hand and the relationship between the imperfective and the present on the other is required for one to understand Dagbani verbs. He explains that where there are no adverbials or other markers to indicate time, the imperfective form of the Dagbani verb is interpreted as the present and the perfective is also understood as the past, noting that a time marker may be added to the imperfective form to give it a past reading. The perfective, according to him, is the unmarked form and it describes a completed action. Consider the following examples: University of Ghana http://ugspace.ug.edu.gh 29 (4) a. o nyu kom. he drink water ‘He drank water.’ b. n di nyuli. I eat yam ‘I ate yam.’ c. piEƒu Nç nyuri kom pam. sheep DEM drink-IPF water much ‘This sheep drinks a lot of water.’ d. jaNa diri kçdu. monkey eat-IPF banana ‘The monkey eats bananas.’ [sic] (Olawsky 1999:32) As examples (4a) and (4b) reveal, the perfective (the unmarked form) takes no particles; the two verbs nyu ‘drink’ and di ‘eat’ are in their bare forms. It can also be inferred from the English glosses that a past reading in implied. The imperfective forms of these are seen in sentences (4c) and (4d) where each of them takes the suffix {-ri}; thus, nyu ‘drink’ becomes nyuri, and di ‘eat’ becomes diri to mark the imperfective. Olawsky (1999) indicates that the habitual carries a progressive reading, when the particle la is inserted between the verb and the object. Thus, as exemplified in (5) below, the examples in (4c) and (4d) above will have a continuous meaning when the post verbal la is present. (5) a. piEƒu Nç nyurila kom pam. sheep DEM drink-IPF-FOC water much ‘This sheep is drinking a lot of water.’ b. jaNa dirila kçdu. monkey eat-IPF-FOC banana ‘The monkey is eating bananas.’ [sic] (Olawsky 1999:32) University of Ghana http://ugspace.ug.edu.gh 30 Olawsky’s (1999) work reveals that the particle la occurs in both the imperfective and the perfective, but he notes that this particle (also found in Dagaare, GurunE, and Buli) is not related to tense and aspect. Although he glosses it as a focus marker and suggests that it is an emphatic marker, he proposes that further studies need to be conducted to ascertain its ‘disjunctive occurrence’. Other verbal particles that Olawsky (1999) identifies are ya, da, and mi. Like la, each of these three particles occurs as a suffix of the verb. Commenting on the use of ya, Olawsky (1999) indicates that it occurs when the object of the perfective verb is absent. His assertion suggests that the occurrence of ya and other verbal particles is not independent of the verb’s transitivity. The assertive marker yá in Buli is akin to ya in Dagbani both syntactically and semantically; in both languages the particle occurs mostly in the absence of the object of a perfective transitive verb, and its occurrence adds emphasis to the action or event coded by the verb. However, while in Dagbani the particle cannot co-occur with an object, in Buli it is permissible for the object to co-occur with the object (in which case ya occurs after the object). The inability of ya to co-occur with an object in Dagbani is probably due to its occurrence as a verbal suffix. We are aware that a transitive verb will typically have its object argument following it immediately, and it is impossible to move a verbal suffix (bound morpheme) to follow an object. Two facts from the two languages give credence to my assertion: 1) in Buli yá occurs as a free morpheme and can co-occur with the object, but it cannot precede the object 2) in Dagbani it is permissible for an adverbial complement (but not an object) to follow yá. I use Olawsky’s (1999) examples (see 6a and 6b) and the Buli examples provided in (6c) and (6d) to explain my claim. University of Ghana http://ugspace.ug.edu.gh 31 (6) a. O nyu-ya He drink-PF/EMPH ‘He has drunk.’ (Olawsky 1999:32) b.*O nyu-ya kom He drink-PF/EMPH water (Olawsky 1999:33) c. Wà nyù yá. He drink.PERF ASS ‘He has drunk.’ d. Wà nyù nyīām yá. He drink water ASS ‘He has drunk water.’ In examples (6a) and (6c) the objects of the verb nyù ‘drink’ are omitted. This makes the presence of yá in the two sentences obligatory. It can also be inferred from (6b) and (6d) that, whereas it is permissible for yá to co-occur with an object in Buli, the co- occurrence of ya with an object yields an ungrammatical sentence in Dagbani. Commenting on the verbal particles ni and yEn, Olawsky (1999) notes that the two particles are semantically similar and are often used interchangeably. He, however, points out that, although yEn has often been treated as a future tense marker, it is only ni that is the “real” future tense marker. Justifying his position, he explains that yEn can better be construed as a “prospective aspect marker”, since it describes an action that is about to happen in the past or future. Besides, it occurs in indirect speech to refer to the past and co-occurs with time-depth markers that carry past time reference. He adds that, unlike yEn which may be preceded by the negation marker bi, ni which he University of Ghana http://ugspace.ug.edu.gh 32 considers as the true future tense marker has its own negation marker ku. This is to say that ku occurs as a suppletive future negation marker of ni. In her discussion of the pre-verbal negative markers in Buli, Schwarz (1999) notes that the verbal system of the language exhibits two major features with regard to negation. These, she identifies as a) double negation and b) the presence of a special negative marker that is restrained to the indicative. Her article gives some insight into the system of negation in Buli. However, what seems unclear is her position on double sentential negation, where she shows that a negative sentence differs from its simple affirmative at two different places. Consider the examples she uses to illustrate her assertion. Affirmative Negative (7) a. Ká la#m Da#a# la#𝑚"𝒂#/ cop / meat neg cop meat+’neg2’ It is meat It isn’t meat. b. Wà kùrì(ya#) WàN kùríyà/ Cl1 / pound cl1+neg1/pound+’neg2’ She has pounded She hasn’t pounded. c. Wàà ku#r(i#) Wà kàN ku#r𝒊 #/ Cl1+à/pound cl1 /neg1/pound+’neg2’ She pounds She doesn’t pound (Schwarz 1999:91) Using the examples in (7), Schwarz (1999) asserts that Buli exhibits double sentential negation (see bolded texts in example 7). My data reveals a mismatch between what she presents as the affirmative forms and their negative forms. The forms she presents as affirmative forms are examples of assertions that speakers make about events that they have enough evidence to support. An addressee who disagrees to such assertions University of Ghana http://ugspace.ug.edu.gh 33 replies in a similar fashion without the part labelled neg2. This implies that, contrary to the negative forms she presents, the constructions Da #a # la#m, Wàn kùrí yá, and Wà kàn ku#rī are the forms that actually occur in the speech of native speakers. There is, however, evidence to prove that the negative forms she presents do occur in speech, but in different contexts. For instance, a speaker poses the question Da#a # la𝒎# 𝒖ā:? ‘Is it not meat?’ when he/she is uncertain about the referent and seeks clarification from the addressee. Here, the addressee may respond Da#a# la𝒎# 𝒖ā ‘It’s not meat’ to indicate that the referent is not what the speaker suspects. In Buli verbal extensions may be determined by the preverbal negative marker in the sentence and the presence or absence of a verb complement. For instance, the negative marker àn in Buli is used in the perfective and it must necessarily co-occur with the assertive marker yá, when the verb has no complement. Where a verb complement is present, the addition of the assertive marker is not required (unless in interrogative constructions). How Schwarz (1999) presents and glosses the nominal and verbal extensions as neg2 makes them appear to be negative markers on their own or part of a discontinuous negative marker (what she glosses as neg1), even though she notes that they are not part of neg1. It should be observed that Schwarz (2005) does not make this assertion in her dissertation which is more current. What she mentions in her dissertation is the general absence of nominal extension in the affirmative and the tendency for the glottal stop to terminate a negated sentence. University of Ghana http://ugspace.ug.edu.gh 34 One other thing that makes Schwarz’s (1999) presentation quite unclear is how she merges the negative marker àn (which she calls the predicate initial marker) with subject pronouns such as wà ‘he/she’ and bà ‘they’ to derive the forms wàn and bàn respectively, with the nasal undergoing homorganic nasal assimilation in consonance with the sound that immediately follows it (see example 7b). Such representation can be misleading, because it treats n (instead of àn) as the negative marker. It is common to have a ‘contamination’ (cf. Kröger 1992) of subject pronouns and other elements that follow them, where such elements are preceded by a vowel. This occurs in rapid speech, so in order to achieve much clarity, it is better to isolate each verbal particle, by presenting (7b) as Wà àn-kùríyà. The fact that the final vowels of the pronouns are weakened when they occur at a clause-final or an object position gives us the reason to believe that it is the vowel of the pronoun that is elided when the negative marker àn follows it. Schwarz (1999) further identifies five negative preverbal markers (kàá, kàn¤, kán$, kàn and àn) for Buli and shows the different grammatical functions each of them performs. She observes that while kàá is used when the sentence is in the imperative and has an imperfective reading, kàn@ is used when the sentence is imperative and has a perfective reading. Consequently, she refers to the former as Imperative I and the latter as University of Ghana http://ugspace.ug.edu.gh 35 Imperative II. In the same vein, she asserts that kàn is used when the sentence is in the indicative and has an imperfective reading, but àn is used in indicative sentences that have perfective reading. kàn and àn, are classified as Indicative I - IPF and Indicative II - PERF respectively by Schwarz (1999). Here also, we must indicate that her classification and discussion of these negation particles is inconsistent with what she presents in 2005 where she identifies only kàn and àn as negation markers. It appears that the particle kàá which Schwarz (1999) identifies as a negative preverbal marker that marks the imperfective aspect in the imperative is what she presents as káá in 2005. Schwarz (2005) observes that káá is a fusion of a truncated form of kán, and the imperfective marker á. There is enough reason to believe that the form káá is a combination of ká (derived from kán) and the imperfective marker á. káá can best be described as a contaminated form of kán + á. Such contaminated forms, as we have already mentioned, are not rare in Buli. Schwarz (2005) shows that kàn marks future negation and has a perfective reading in the indicative. However, the reason for which she identifies kán $ as a future negation marker and kàn as an indicative marker that has a futuristic reading in her earlier work is unknown. The fact that she turns to identify both particles as future markers makes University of Ghana http://ugspace.ug.edu.gh 36 the inconsistencies in her identification or discussion of the preverbal negative markers evident. The particle kan appears to occur as a future negation marker in other Gur languages. For example, Atintono (2004) identifies kán as a future negation marker in GurunE. (8) a. Ba kán bóbé na’am la zina. 3PL NEG.FUT tie chieftaincy DEF today ‘They will not install the chief today.’ b. A $ kán iNE. S/he NEG.FUT do ‘S/he will not do it.’ (Atintono 2004:122) Worth mentioning is the fact that the negative forms of some Buli verbs are suppletives and cannot be used in combination with negation markers. It was shown in example (7a) that da#a# ‘not be’ is a negative suppletive of the copula verb ká ‘to be’. Other pairs of suppletives include ta # ‘have’/ ka# ‘lack’, bo#ro# ‘exist’/ka#ro # ‘be non-existent’, and se#b ‘know’/ze# ‘not know’. The works of Akanlig-Pare (1988, 1994, 1996, 1997, 1999, 2002, 2005), as we have indicated, focus on the phonology of Buli. These works are underpinned by different phonological theories, including the phoneme theory, the syllable theory, and University of Ghana http://ugspace.ug.edu.gh 37 autosegmental phonology which shows the indispensability of tone in any effective discussion of a language. Akanlig-Pare (2005) points out that Buli verbs are not conjugated to mark tense, aspect and polarity (TAP), and adds that such grammatical relations are expressed either via tone alternations alone or through the interaction between tone and the preverbal particles àlì(lì), à, kàn, and àn. Although his work does not discuss the verbal system of the language in much detail, it a) provides some insights into the verbs’ morphology, b) highlights some of the pre-verbal particles and c) illustrates how tone interacts with the pre-verbal particles to mark TAP on the verb. When we study Akanlig-Pare’s (2005) work, it does not take us much time to realize that he proposes a tense-aspect distinction of three tenses (present, past and future) and two main aspects (perfective and imperfective) for Buli. His comment on the language’s TAP inflections supports this observation: Though we identify three tenses in Buli, that is, the present, past and the future, it is not clear whether the present and past in particular, can be separately identified from the TWO main aspects in Buli, namely the perfective and the imperfective. (Akanlig-Pare 2005:59) University of Ghana http://ugspace.ug.edu.gh 38 He further reiterates his claim that tone is a sine qua non for any discussion of Buli TAP, and underscores the fact that it is difficult to ‘tease apart’ tense and aspect inflections, because they are intertwined with tone. It is truly the case that phonological analysis is often inseparable from morpho-syntactic analysis. What is unclear, however, is how he treats tense as part of aspect. Tense is usually separated from aspect, so that a language is said to possess either tense or aspect or both, with different verbal categories and/or tone used to draw a dichotomy between the two aspects of the language. The sixth chapter of Akanlig-Pare’s (2005) work, which is dedicated to Buli tono-syntax, presents a detailed discussion of the indispensable role tone plays in understanding Buli grammar. Adopting a non-linear approach to phonological analyses, he uses the Autosegemental theoretical framework, as well as Lexical Phonology, to explore the interface between tone and Buli grammar. His work is able to account for how tone creates a grammatical contrast between the perfective which is marked solely by tonal interactions and the imperfective which requires a combination of pre-verbal particles and tonal alternations. The work further reveals that verbal particles interact with subject pronouns and object pronouns, and how all these elements form a phonological domain on which tone operates. The following examples are taken from his work: University of Ghana http://ugspace.ug.edu.gh 39 (9) a. m $ à lE.# 1SG IPF insult ‘I insult.’ b. m $ à lE& fV$ 1SG IPF insult 2SG.OBJ ‘I insult you.’ c. n$ lE&. 1SG insult ‘I have insulted.’ d. n$ lE & fV.$ 1SG insult 2SG.OBJ ‘I have insulted you.’ (Akanlig-Pare 2005:257) As the English glosses in (9a) and (9b) reveal, the imperfective is marked by the preverbal particle à. In (9a) the verb is assigned a default mid-tone, but in (9b) a rising tone is realized. Akanlig-Pare (2005) shows that the change in tone of (9b) is as result of the object pronoun that follows the verb. He explains that there is a bond between the verb and the object pronoun which he calls Internal Object, in whose presence there is activation of an agreement HL tone. H causes the mid-tone of the verb to become a gliding tone and L is assigned to the pronoun. The perfective, as exemplified in (9c) and (9d) does not take pre-verbal particles; there is a complex tone alternation University of Ghana http://ugspace.ug.edu.gh 40 to indicate perfective and this is determined by the subject pronoun as well as the absence or presence of an internal object pronoun. In his study, Akanlig-Pare (2005) identifies preverbal particles that mark the habitual and the present progressive, which he subsumes under the imperfective aspect. The habitual takes only the pre-verbal particle à (see 9a and 9b), but the present progressive requires a combination of the temporal adverbial bóró9 and the aspectual marker à to create the chain bóró+à. Kröger (1992), Akanlig-Pare (2005) and Schwarz (2005) mention the contamination of this chain and its accompanying tonal alternation to yield the form bórà:. Akanlig-Pare (2005) further shows that, like the habitual, tonal alternations in the present progressive is not independent of the subject and the object pronouns as well as the verbal particles. Consider the examples below: 9 bóró has grammaticalized into a locative verb and an existential. Where it functions as an existential, I gloss it as EXIST in this work. University of Ghana http://ugspace.ug.edu.gh 41 (10) a. m $ bóró à lE.# 1SG ADV ASP insult ‘I am insulting.’ b. Wà bóró à lE@ fV$. 3SG ADV ASP insult 2SG ‘He is insulting you.’ c. Wà bóró à lE # bí:k. 3SG ADV ASP insult child ‘He is insulting a child.’ (Akanlig-Pare 2005:255) Drawing on the explanation provided for the tonal alternations in the habitual, we are able to account for the default mid-tone of the verb in (10a) and its high tone in (10b). In (10a) there is no internal object that is bound to the verb so the aspectual tone is assigned to the pre-verbal particles such that the verb bears a default M tone, but in (10b) the presence of the internal object after the verb creates a bond between the two elements so that HL tone is activated to generate a high tone for the verb and a low tone for the internal object. The object bí:k ‘child’ in (10c) is an external object and not bound by the verb; hence, its tone is not altered and a default mid-tone is assigned to the verb. University of Ghana http://ugspace.ug.edu.gh 42 Like Schwarz (2005), Akanlig-Pare (2005) identifies the morpheme àlì(lì) as a future tense marker. However, unlike the former who considers nì10 as a variant of the future marker, the latter does not make mention of any variant form. What he points out is the contrast between the imperfective and the future – that the tone of the subject does not affect the tonal realization in both the future and the imperfective. While AGR is activated in the imperfective for an internal object to take a low tone, it remains inactive in the future so that the internal object takes a mid-tone just like the default mid-tone of the verb. This explains why the object pronouns wa # ‘him’ in (11) bear a mid-tone. (11) a. n $ lì lE # wa# 1SG FUT insult 3SG ‘I will insult him.’ b. Wà lì lE# wa# 3SG FUT insult 2SG ‘He will insult him.’ Cross-linguistic studies have revealed that the future marker sometimes performs modality functions, such that a future construction may not always make reference to the future. It sometimes expresses intention, hope, and other such grammatical 10 nì expresses modality. It shows willingness or desire to perform an action. University of Ghana http://ugspace.ug.edu.gh 43 functions related to modality. This accounts for Bybee, Perkins and Pagliuca’s (1994) observation that: future is less a temporal category and more a category resembling agent-oriented and epistemic modality, with important temporal implications.” It will be shown in the next chapter that the future tense marker àlì/lì in Buli is not restricted to expressing futurity; it expresses modality. (Bybee, Perkins and Pagliuca 1994:280) 2.1.3 Serial Verb Constructions (SVCs) Studies on Niger-Congo languages (Oyelaran 1982; Bodomo 1997; Lee 2003; Osam 2004) provide evidence that verbs are capable of being serialized or concatenated. Languages that exhibit this characteristic have been generally referred to as serializing languages. The examples below provide evidence of SVCs in Yoruba, Dagaare, Buli, and Akan respectively. University of Ghana http://ugspace.ug.edu.gh 44 (12) a. Wón ránti pé àlejò ni owó. they remember say guest FOC money ‘They remember that money/wealth is transient.’ (Oyelaran 1982:109) b. Ayuo da de la a bie zegle. Ayuo PST take-perf fact def child seat-perf ‘Ayuo seated the child.’ (Bodomo 1997:108) c. Wà lì pa# gbáNká tE$ wa.# she FUT take book.DEF give him. ‘She will give him the book.’ (Lee 2003:98) d. Aba frE$-E$ Esi sómà-à no Aba call-COMPL Esi send-COMPL 3SG OBJ ‘Aba called Esi and sent her.’ (Osam 2004:17) Inferring from the examples in (12), we realize that each sentence has two verbs which are not coordinated. This makes them monoclausal sentences with serialized verbs that share common subjects. In (12a) the two verbs, ránti ‘remember’ and pé ‘say’, share the pronominal subject wón ‘they’ The same can be said of (12b) where de ‘take’ and zegle ‘seat’, share the subject Ayuo. In (12c) the verbs pà ‘take’ and tE$ ‘give’ share University of Ghana http://ugspace.ug.edu.gh 45 the pronominal subject wà ‘she’ and in (12d) the verbs frE$E$ ‘called’ and sómàà ‘sent’ share the subject Aba. Verb reduplication is different from Serial Verb Constructions (SVCs) in that the latter have different verbs to code different events. Verb reduplication, on the other hand, involves the repetition of the same verb either fully11 or partially. It should be observed that it is possible for a reduplicated verb to co-occur with a verb of a different semantic domain in a serial verb construction. The examples in (13) will be used to explain the two phenomena, verb reduplication and Serial Verb Constructions (SVCs) in Buli. (13) a. Akúm lì ge#b ge#b lāmmú. Akum FUT cut cut meat.DEF ‘Akum will cut the meat into several pieces.’ b. Bílèòkú yìtì zà:nì yá. toddler.DEF rise.PERF stand.PERF ASS ‘The toddler has stood up.’ c. YèsìNká chèN chèN bè yá. madman.DEF go.PERF go.PERF be.lost.PERF ASS ‘The madman got lost.’ (after roaming aimlessly) d. Akúm lì ge#b ge#b lāmmú dìk. Akum FUT cut cut meat.DEF cook ‘Akum will cut the meat into several pieces and cook it.’ 11 Fully reduplicated verbs in Buli often yield meanings that are different from the meanings expressed by the individual verbs. University of Ghana http://ugspace.ug.edu.gh 46 While (13a) is an example of verb reduplication, (13b) is a serial verb construction. In (13a) the verb ge#b ‘to cut’ is repeated, and in (13b) two different verbs, yìtì ‘rise’ and zà:nì ‘stand’ share the same subject bílèòkú ‘the toddler’. Repeating a verb has semantic implications which are related to the manner in which the action is performed or should be performed. For instance, chèN ‘walk’ is repeated in (13c) to indicate that the agent yèsìNká ‘the madman’ roamed aimlessly. In the same way, ge#b ‘cut’ is repeated in (13d) to show that the meat was cut into several pieces. As shown in (13c) and (13d), reduplicated verbs may occur with other verbs to form SVCs. A topical issue that arises in the discussion of SVCs is the identification of the head of the multiple verbs. This has been necessitated by the fact that, where grammatical categories are to be marked on a single verb, it is almost always marked on the head verb. Lee (2003) suggests that, considering the head-initial parameter of Buli, the first verb (V1) is the head in SVCs; hence, it is V1 that bears grammatical categories that mark tense, aspect and polarity. This is evident in the tonal variation of the verbs pa# ‘take’ and tE$ in (12c). The former verb takes the default mid-tone as a result of the preceding future tense marker lì. The low tone of tE$ is attributable to the absence of the preceding future tense marker lì. Thus, since tE$ is not the head verb, the low tone of the future tense marker which prevents any interaction between the tone of the subject pronoun (wà) and that of the verb (pa#) is absent. The low tone of tE$ is an University of Ghana http://ugspace.ug.edu.gh 47 indication that there is tonal interaction between the tone of the verb and an adjacent tone to modify its default mid-tone. Although grammatical functions are often marked on the head verb of SCVs, there are languages that permit the marking of such functions on all the verbs with an SVC. For instance, Osam (2004) reports that in Akan, verbal inflections that express grammatical functions such as tense/aspect and negation are generally marked on all the verbs within an SVC. Consider (14) below: (14) a. Araba tç@-ç$ mpEtsea má-à abofra no. Araba buy-COMPL ring give-COMPL child DEF ‘Araba bought a ring for the child.’ b. Araba à-n $-tç@ mpEtsea à-m $-má abofra no. Araba COMPL-NEG-buy ring COMPL-NEG-give child DEF ‘Araba did not buy a ring for the child.’ (Osam 2004:40) Osam (2004) further mentions the uniformity in the marking of tense/aspect and negation in Akan SVCs. He notes that where the first verb marks a particular aspect, say the completive aspect, the other verb(s) within that SVC will also mark the same aspect. In the same vein, in a negative SVC, negation is marked morphologically on all the verbs. Examples (14a) and (14b) give credence to his claim. In (14a) each of the verbs is marked morphologically to indicate the completive aspect, and in (14b) University of Ghana http://ugspace.ug.edu.gh 48 both verbs (tç@ ‘buy’ and má ‘give) are marked twice to express negation and completive aspect. Osam (2004), however, points out that some sub-dialects of Fante (a dialect of Akan) possess a unique feature in this regard, since they do not always exhibit this uniformity in marking negation. (15) a. Mò-bç-@ tç@ bi à-mà wò. 1SG SUBJ-FUT-buy some CONS-give 2SG OBJ ‘I will buy some you.’ b. Mò-ró-n$-tç@ bi m-$ má wò. 1SG SUBJ-PROG-NEG-buy some NEG-give 2SG OBJ ‘I will not buy some for you.’ c. Mò-n-$ kç-@ tç@ bi à-mà wò. 1SG SUBJ-NEG-FUT-buy some CONS-give 2SG OBJ ‘I will not buy some for you.’ (Osam 2004:40) Osam (2004) explains that, although negation is typically marked morphologically on each of verbs (as exemplified in 15b), there are instances where this uniformity in not seen. Using (15c) as an example, he indicates that, where the first verb is a form of n- kE, the second verb does not take a negation marker. The implication here is that the scope of negation extends from the first verb to all other verbs that follow the first. University of Ghana http://ugspace.ug.edu.gh 49 Some scholars (cf. Nylander 1997; Delplanque 1998; Dimmendaal 2001) have made attempts to justify the need for SVCs in some languages. They argue that West African languages do not have three-place predicates, and this necessitates the presence of a second verb to host a third argument. It is noteworthy that this claim is debatable, because there are Gur languages (including Buli) that possess three-place predicates. tE# ‘give’, and tōm ‘send’ are two of such Buli verbs. Some Kwa serializing languages have also been shown to possess verbs that can host three arguments. At least, Ameka (2013) provides attested evidence to confirm that Ewe, Akan and Likpe possess trivalent verbs. He provides examples of such verbs as ná ‘give’, fíá ‘teach/show’ bíá ‘ask’ for Ewe; ma ‘give away’ kyɛ ‘give’, brɛ ‘bring’ mane ‘send’ fɛm ‘lend’ for Akan; and tə ́‘give’ or té ‘show, teach, sell’ yifó ‘do’ for Likpe. For each of these languages, a three-place predicate construction is provided to show that three-place predicates exist in kwa languages. University of Ghana http://ugspace.ug.edu.gh 50 (16) a. Kofí fíá agble-a ɖeví-á-wó (Ewe) Kofí show farm-DEF child-DEF-PL ‘Kofi showed the farm (to) the children’ (Ameka 2013:9) b. Fə tə́ si-kpi (Likpe) 2SG give 1SG CM-fear ‘You frighten me.’ (Ameka 2013:18) c. Kofí ma-a Áma siká (Akan) Kofí give-PAST NAME money ‘Kofi gave Ama money.’ (Ameka 2013:23) We saw in example (13) that while different verbs occur in SVCs, there are evidence of verb reduplication in Buli. Verb reduplication may result in not only a change in the word class of the verb, but also a change in the verb’s meaning. Hsiao’s (2003) work on reduplication in Buli shows that a verb can be repeated in a sentence to imply that the process or activity indicated by the verb occurred severally. Thus, verb reduplication is only possible when the speaker is uncertain or has no knowledge about the number times that the event specified by the verb occurred. Consider the following examples. University of Ghana http://ugspace.ug.edu.gh 51 (17) a. n $ dìg dìg dZ´n$ta tE $ n-$ dek& . 3SG cook cook soup give myself ‘I cooked soup for myself many times.’ b. n $ dìg dZ´n$ta tE$ n$-dek& . 3SG cook soup give myself ‘I cooked soup for myself.’ (Hsiao 2003:87) It might have been noticed that the serial verb construction in (17a) is different from those provided earlier – unlike those provided earlier, this has the first verb (dìg ‘cook’) of its two verbs reduplicated. This creates the semantic implication that the number of times the speaker cooked for himself/herself is unknown to him or her. If the verb is not reduplicated as shown in (17b), the implication is that the event occurred once. Thus (17b) carries the implication that the speaker cooked soup for himself on a single occasion. It is possible to reduplicate any of the verbs or all the verbs more than once within an SVC, depending on the speaker’s intended meaning. Thus in (17b), for instance, the speaker may reduplicate the verb tE$ ‘give’ to imply that he/she was a benefactive of the soup on several occasions. If each of the two verbs is reduplicated, the construction will then imply that the speaker neither has no knowledge of the number of times the event (cooking) occurred nor the number of times he/she benefitted from the soup. Although Hsiao (2003) does not discuss the marking of grammatical categories such as tense, aspect and polarity on the verb in reduplicated verbs, he provides a few examples on the negation of reduplicated verbs. He mentions only that the negation marker cannot University of Ghana http://ugspace.ug.edu.gh 52 be reduplicated. The examples which he provides show that grammatical categories are marked on the first verb (the head of the multiple verbs). Thus, in each of the examples in (18), the negation marker manifests once and precedes only the first verb. (18) màn díg dìg dZ´nt$ a à tE$ n-$ de&k I (non-emphatic)-not cooked cooked soup give myself ‘I did not cook myself soup many times.’ b. n$ kàn dīgá: dīgī dZ´nt$ a à tE$ n$-de&k I not cook cook soup for myself ‘I do not cook myself soup very often.’ (Hsiao, 2003:93) It will be recalled from the discussion that the transitive nature of a verb determines the occurrence of certain verbal particles. For instance, we saw that in Buli the assertive marker yá obligatorily follows a perfective transitive verb when an object is absent. Reference was also made to Olawsky’s (1999) observation that the Dagbani perfective transitive requires the emphatic marker ya in the absence of an object. Linguistics literature (e.g. Ameka 2013, and Osam 2016) show that verbs can be grouped into types based on their valency and transitivity. By valency, we are referring to the ‘power of a verb’ to take a certain number of arguments/participants. My use of the two terms, arguments and participants, imply that valency is not only a syntactic concept; it is also a semantic concept (cf. Osam 2016:117). The notion of transitivity is evoked in the discussion of valency, since the transitive nature of a verb determines the number of arguments/participants that the verb requires. University of Ghana http://ugspace.ug.edu.gh 53 Discussing the transitive classes of Akan verbs, Osam (2016) reveals that there are: a) verbs that are strictly intransitive b) verbs that are strictly transitive c) verbs that are used ditransitively and d) verbs that are used both transitively and intransitively. Intransitive verbs are monovalent, since they take a single participant or argument which typically occurs as a subject; transitive verbs are bivalent, since they require two arguments or participants (one being the subject and the other being the object); ditransitive verbs are trivalent, since they require three arguments or participants (one being the subject and two being objects). These ditransitive/trivalent verbs are what Ameka (2013) refers to as three-place predicates. Crystal (2008) notes that there are verbs that require no arguments or participants. He describes such verbs as avalent, and identifies rain as an example of such verbs. My data, however, does not show the existence of avalent verbs in Buli. Most Buli verbs, as we have seen, are ambitransitive12. dE# ‘to eat’ is a typical ambitransitive verb. There are verbs that do not require object complement(s) but clausal complements. Such verbs are limited in number in most languages. Norris (2003), in her discussion of embedded clauses in Buli, identifies three types of constructions that serve as clausal complements, thereby categorising the verbs into three groups; verbs that require finite clausal complements, verbs that require non-finite clausal complements, and verbs that can take either finite complements or non-finite clausal complements. She identifies po#li# ‘to think’, ma#:ri# ‘to help’, and ɲe # ‘seem’ as verbs that fall into the first category. Also, she identifies àya#:li# ‘to want’, zE#ri # ‘to object’, and ki#si # ‘to hate’ as those that 12 Ambitransitive verbs can be used transitively or intransitively. University of Ghana http://ugspace.ug.edu.gh 54 take non-finite complements, and the verbs sa#:lim# ‘to beg’ and we#:ni# ‘to say’ as those that can take either a finite or a non-finite verb. Although Norris (2003) does not discuss in any detail the syntactic behaviour of verbal particles in such complex constructions, she mentions briefly that the nature of negation in some of her verb types. She shows that the negation markers kàn and àn are employed in different construction types, depending on the speaker’s intended meaning. For instance, she indicates that kàn can occur in either the matrix clause or the clausal complement of àya#:li# ‘want’ constructions, depending on the speaker’s intended meaning. Also, she observes that both kàn and àn are capable of occurring in the matrix clause. 2.2 Theoretical Framework This study has its underpinnings in Functional Grammar (FG) general, and it is skewed towards Functional Typological Syntax in particular. Proponents of FG approach their task of studying language differently, depending on the extent to which each of them prioritizes ‘sociocultural or cognitive’ factors that shape the structure of language. This explains the existence of different theories of Functional Grammar, including Emergent Grammar, Systemic Functional Grammar, Role and Reference Grammar, Relational Grammar, as well as Functional Typological Syntax to which this work is closely related. In spite of their different approaches to the study of language, all functionalists are unified by one core belief – the belief that language is a tool for communication. This explains Dik’s (1981:2) observation that ‘the main preoccupation of any linguistic University of Ghana http://ugspace.ug.edu.gh 55 enquiry should be to account for the communicative function of language, since language is a tool for communication and not just a device for the expression of our thoughts.’ It is based on their belief that language is a tool for communication that functionalists (e.g. Dik 1981; Nichols 1984; Halliday 1994; DeLancey 2001; Givón 1984, 2001; Van Valin and Lapolla 1997; Tomasello 1998; Dryer 1999; Butler 2008) hold the view that language can be better understood based on the functions it performs in natural communication situation or context. Thus, being a tool for communication, language can be understood and consequently described accurately when we consider the functions it performs. A couple of assumptions are used here to explain the position of functionalists further. In the words of Nichols (1984:97) Functional Grammar: … analyzes grammatical structure, as do formal and structural grammar, but it also analyzes the entire communicative situation: the purposes of the speech event, its participants, its discourse context. Functionalists maintain that the communicative situation motivates, constrains, explains, or otherwise determines grammatical structure, and that a structural or formal approach is not merely limited to an artificially restricted data base, but it is inadequate even as a structural account. We can infer from Nichols’ (1984) assertion that functionalists generally believe that syntactic structures are not independent of the communicative situation; it is the communicative situation (comprising the interlocutors, the communicative goals, and other socio-cultural factors) that determines the forms that language structures take. University of Ghana http://ugspace.ug.edu.gh 56 DeLancey (2001:4) makes a similar claim by noting that the structure of our ‘experience and our cultural models’ influence our language significantly, since we use language as a ‘set of tools for communicating our experience.’ Apart from studying the functional motivations for the structures seen in languages, Functional Typological Syntax (FTS) examines cross-linguistic variations and the extent to which patterns of grammatical features are distributed across the languages of the world. Based on this, functionalists within FTS are able to make cross-linguistic generalizations about different languages. This study is limited in scope in this regard, since no attempt is made at making cross-linguistic generalizations about languages. Nonetheless, it is motivated by the typological functionalists’ belief that grammatical features are to some extent cut across different languages. Thus, the current study draws data from better-studied languages, especially Gur languages, to show how Buli relates with such languages in terms of grammatical features. A major merit of the functional approach to an exploratory study of this nature stems from the fact that it relies on empirical data. Grammatical features of dominant languages have often been applied to less dominant languages, because documentation on such dominant languages, especially English and Latin abound (cf. Welmers 1973). Functional Grammar gives the researcher the opportunity to describe a language as it is employed in the natural communication setting, and consequently reduce the propensity of the researcher to make wrong assumptions and generalizations about the grammar a language. Languages are distinguishable by their nature and based on the grammatical features that they possess. University of Ghana http://ugspace.ug.edu.gh 57 Also, Functional Typological Syntax helps the researcher to identify and discuss recurrent patterns of grammar in different languages. Thus, we adopt FTS of Functional Grammar to describe the grammatical features of Buli as they perform grammatical functions, and show how such features are recurrent in other related languages. Another merit of Functional Grammar is that it does not treat various language components (phonology, morphology, syntax, semantics, and pragmatics) as independent. Therefore, it will help the researcher to show the interconnections between the different components of Buli. Specifically, it will help show how the verb and verbal particles interact with themselves and with other clausal elements at the different levels of the language. 2.3 Methodology: Data Type and Data Source Being an exploratory study with its underpinnings in Functional Grammar, the study relied mainly on empirical data collected from fieldwork conducted in three Buli speech communities, namely Sandema, Fumbisi and Chuchiliga.13 This was aimed at revealing dialectal differences (where necessary). In these three communities, naturally occurring texts, specifically radio panel discussions, radio announcements, every day interactions between speakers and the researcher’s interactions with other native speakers were collected through tape recordings and transcribed for the analysis. Such naturally occurring texts are very reliable for studies of this nature, since speakers’ choice of linguistic items is not independent of the communication situation. I, therefore, relied 13 Dialects spoken in Sandema, Fumbisi, and Chuchiliga are the Central, Northern and Southern dialects respectively. University of Ghana http://ugspace.ug.edu.gh 58 largely on naturally occurring texts, with the aim of identifying how the verb and verb particles are employed in different communication situations, in order to provide an accurate description of the Buli verb and verbal particles. Since the examples used for the analysis comprise a chunk of naturally occurring texts drawn from a wide range of communicative contexts, I do not attempt to give them unique labels; this, nevertheless, isolates them from data collected from other sources which have been duly cited. Relying entirely on naturally occurring texts poses its own challenges, one being the propensity to collect inaccurate data when utterances of speakers are unclear. This leads to the transcription of inaccurate data and consequently errors in the analyses. Besides, native speakers’ speech is not always devoid of grammatical errors, and researchers are sometimes unable to detect them. In order to overcome this challenge, I elicited some data to complement the naturally occurring texts. This gave me the opportunity to seek for clarity from the speakers, if their utterances lacked clarity. The elicitation was guided by two data sheets of phrases and sentences. These are the first and second volumes of Dakubu’s West African language data sheets published in 1976 and 1980 respectively. Each of the data sheets contains a tall list of phrases and sentences drawn from different languages and translated into English. Information from three language consultants (one from each of the three selected Buli- speaking communities), as well as my native speaker intuition, influenced the analysis of the data. Although constructed data were not drawn from written texts, some Buli texts, including Schott’s (1993) Bulsa Sunsuelima: folktales of the Bulsa in Northern Ghana, Akanlig-Pare and Wangara’s (2007) Biamagsika Nyuenta, Kröger’s 1992 Buli-English Dictionary, and a translated version of the New Testament Bible served as reference University of Ghana http://ugspace.ug.edu.gh 59 materials. Particularly, they provided me essential information about Buli orthography and lexicography, thereby ensuring accuracy in the presentation of my data. The analysis was purely descriptive, with focus on the morphological make-up of the verb, the grammatical functions of the verbal particles, as well as the syntactic behaviour of the particles in SVCs. 2.4 Chapter Summary The chapter has set the study in context by presenting a review of some literature on the verb and verbal particles of other languages. Data have been drawn mostly from other Gur languages and minimally from less related languages to discuss the grammatical functions that verbal affixes and verbal particles perform, as well as the syntactic distribution of verbal particles in SVCs. To show the relation between Buli and such other languages, literature on Buli have also been reviewed in the same regard. The discussion also included the basic assumptions of Functional Grammar (and Functional Typological Syntax) and the justification for which the theory has been adopted for this study. Also, the methodological procedures for the study, including the data source, data type, and data analysis have been presented in some detail. University of Ghana http://ugspace.ug.edu.gh 60 CHAPTER THREE VERB MORPHOLOGY AND VERBAL PARTICLES 3.0 Introduction This chapter discusses the morphological make-up of Buli verb and the particles that co-occur with the verb. It, thus, describes the internal composition of the Buli verb and discusses the grammatical functions that verbal particles perform. 3.1 The Buli Verb The clause has generally been defined as a construction that has at least a subject and a verb; it may have objects and adjuncts as complements. In imperative constructions, other clausal elements are unexpressed so that only verb is present. The verb can therefore be described as the most indispensable14 element in a clause. Consider example (1) below: (1) a. Bííká nyù ká dáám jīnlá. child.DEF drink.PERF FOC wine today ‘The child has drunk wine today.’ b. Bííká nyù ká dáám. child.DEF drink.PERF FOC wine ‘The child has drunk wine.’ c. Bííká à ku#m. child.DEF HAB cry. ‘The child cries.’ d. Ka#li#! Sit! 14 Verbless clauses have been reported in some Oti-Volta languages (cf. Saanchi 2006) University of Ghana http://ugspace.ug.edu.gh 61 The examples provide evidence that the verb is the most obligatory element of the clause. In (1a) four clausal elements are present – a subject NP bííká ‘the child’, an object NP dáám ‘alcohol’, a transitive verb nyù ‘drink’, and an adverbial jīnlá ‘today’. In (1b) the adverbial is omitted, leaving the subject, object and the verb. The verb ku#m ‘cry’ in (1c) is an intransitive verb; it possesses only a subject (bííká ‘the child’) leaving out an object and an optional adjunct. Example (1d) is an imperative sentence composed of only the verb ka#li# ‘sit’ as the only element present. It should be added that, although a subject is not physically realized in (1d), an implied subject (the second person singular pronoun fí ‘you’) is present. This is because speakers often do not address our addressees, when we are giving them commands. 3.1.1 Verb Types 3.1.1.1 Stative verbs Crystal (2008) draws a distinction between stative verbs and dynamic verbs using the two criteria – the syntactic and the semantic criteria. He points out that, syntactically, stative verbs do not occur in the progressive aspect or in the imperative, and gives examples of such verbs as have, concern, know, and can. Consider the use of some stative verbs in example (2) below: (2) a. The farmers have cattle. *The farmers are having cattle b. The women know the thief. *The women are knowing the thief. c. *Have! d. *Know! University of Ghana http://ugspace.ug.edu.gh 62 As examples (2a) and (2b) reveal, to use the stative verbs ‘have’ and ‘know’ progressively yields ungrammatical sentences. Also, as shown in (2c) and (2d), it is ungrammatical to use such verbs in the imperative. The ungrammaticality of the examples indicated by the asterisks in (2) can best be understood semantically – stative verbs denote state of affairs but not actions. He adds that verbs that express relational processes (as in 2a) and those that express cognitive processes (as in 2b) are typical examples of stative verbs, but issues the caveat that isolating stative verbs from dynamic verbs can be complicated, since certain verbs are classified as both stative and dynamic. In Buli and in most other Gur languages, just as in English, stative verbs rarely occur in the progressive and in imperatives. Consider the use of the stative verbs sēb ‘to know’, tā ‘to have’, and bāg ‘to be able’ in example (3) below. (3) a. AkúsúN tà yám. AkúsúN have sense ‘AkúsúN is wise.’ b. * AkúsúN bóràà tā yám. AkúsúN PROG have sense c. BàN bàg sírí. lizard be.able bee ‘A lizard is stronger than a bee.’ d. * BàN bóràà bāg sírí. lizard PROG be.able bee University of Ghana http://ugspace.ug.edu.gh 63 e. Tì sèb nícháánōā. 1PL know visitor.DEF ‘We know the visitor.’ f. * Tì bóràà sēb nícháánōā 1PL PROG know visitor.DEF While examples (3a), (3c) and (3e) are grammatical, (3b), (3d), and (3f) are ungrammatical. The presence of the progressive marker bóràà in each of the asterisked sentences makes them ungrammatical, since the verbs sēb ‘to know’, tā ‘to have’, and bāg ‘to be able’ are stative so cannot be marked for progressive. It is equally ungrammatical to use these verbs in the imperative. This explains the ungrammaticality of forms such as *Tā! ‘Have!’, *Bāg! ‘Be able!, and *Sēb! ‘Know!’. Worth noting is the fact that some dynamic verbs are morphologically similar to stative verbs, and may therefore be misconstrued as stative verbs. The semantics of such verbs gives us a clue that they came into existence through metaphorical extensions; hence, their near polysemy with the stative verbs. A typical example is the semantic relation between sēb ‘to know’ and sēbā ‘to be careful/cautious’ which is often used as a warning or a caution remark to an addressee. While the former is stative, the latter is dynamic and can be used in the imperative. zE# ‘not know’, the negative suppletive form of sēb ‘to know’, is another stative verb. The verb kā ‘to be’ and its negative suppletive dāā ‘not be’ are linking verbs considered as a sub-group of stative verbs that denote a state of being. Example (4) illustrates the use of the two copulas kā ‘to be’ and dāā ‘not be’. University of Ghana http://ugspace.ug.edu.gh 64 (4) a. Nídōāwá kā yèsìN. man.DEF to.be lunatic ‘The man is a lunatic.’ b. Nídōāwá dāā yèsìN. man.DEF not.be lunatic ‘The man isn’t a lunatic.’ In (4) verb kā ‘to be’ and its negative suppletive dāā ‘not be’ function as linking verbs or copulas such that the quality denoting noun yèsìN ‘lunatic’ becomes the subject complement of nídōāwá ‘the man’. 3.1.1.2 Dynamic verbs These verbs are the opposite of stative verbs both syntactically and semantically. Syntactically, dynamic verbs are capable of occurring with the progressive marker bóràà and in the imperative. Unlike stative verbs which denote states of affairs and do not express actions, dynamic verbs form a wider group that encode activity, process, and bodily sensation (Crystal 2008). The subject of a stative verb is usually an agent who initiates the action or activity; hence, the agent has control over the action and can cause a change in it. The English examples in (5) below illustrate this: (5) a. The farmer cut the tree. b. The farmer is cutting the tree. c. Sit down! d. Keep quiet! University of Ghana http://ugspace.ug.edu.gh 65 The verbs ‘cut’, ‘sit’, and ‘keep’ in (5a), (5c) and (5d) are dynamic verbs. Since cut is a dynamic verb, it can be used in the progressive as illustrated in (5b). The use of the verbs ‘sit’ and ‘keep’ in the imperative as exemplified in (5c) and (5d) affirms that they are also dynamic verbs. In Buli, dynamic verbs include Nmārīsī ‘to write’, chE# ‘to fell’, Nç#b ‘to chew’, tūsī ‘to push’, and Nārī ‘to fetch’. Example (6) shows that these verbs, and others alike, can be used progressively and in the imperative. (6) a. Dç$glìèwá bóràà tūsī lāmmú. cat.DEF PROG push meat.DEF ‘The cat pushes the meat.’ b. Kpārōāwá bóràà chE # tī:mmú. farmer.DEF PROG fell tree.DEF ‘The farmer fells the tree.’ c. Tūsī! Push.IMP ‘Push!’ d. ChE # tī:mmú! Fell.IMP tree.DEF ‘Fell the tree.’ Examples (6a) and (6b) illustrate the use of tūsī ‘to push’ and chE# ‘to fell’ respectively in the progressive. In each of the two sentences, the action initiated by the subject (dç $glìèwá ‘the cat’ in 6a and kpārōāwá ‘the farmer’ in 6b) is a continuous one that can be affected by the agent who has control over it. Their occurrence in the imperative University of Ghana http://ugspace.ug.edu.gh 66 (shown in 6c and 6d) indicates that, being actions that can be affected by an agent, they can be given as a command to an addressee. 3.1.2 Verb Morphology Morphological markings on Buli verbs to indicate functional categories (e.g. tense, aspect, modality and polarity) are generally minimal. Nonetheless, based on their forms, a dichotomy can be drawn between the large group of dynamic verbs and the relatively small group stative verbs. A study of the forms of most disyllabic dynamic verbs will reveal that, their roots have the structure CVC, but they surface with the suffix {-i}. Thus, morphologically, final vowels of disyllabic dynamic verbs are predictable. I illustrate my observation further, using examples. Verb Root Suffix Verb Gloss (7) a. tūs -i tūsī to push b. sāt -i sātī to slip c. sūr -i sūrī to wash d. dāl -i dālī to pull e. mān -i mānī to cook f. lōb -i lōbī to lay (egg) g. sāg -i sāgī to teach Trisyllabic verbs display a similar morphological make-up; they have the suffix {-i}. Here also, it appears that only the initial vowel is unpredictable. Apart from this unpredictable vowel in the first syllable, /i/ typically occurs as the nucleus element of University of Ghana http://ugspace.ug.edu.gh 67 the succeeding syllables. This implies that any vowel may manifest in the first syllable of a polysyllabic work, but /i/ manifests in subsequent ones so that syllables like ri, mi, bi, and li occur as non-intial syllables. This makes /i/ behave like an epenthetic vowel. The examples in (8) provide evidence to my claim that {-i} occurs as a suffix of polysyllabic dynamic verbs, and that with the exception of the nucleus of the first syllable, the nucleus of any other syllable in a polysyllabic word is predictable. Verb Root Suffix Verb Gloss (8) a. Nmārīs -i Nmārīsī to write b. kpāmīs -i kpāmīsī to fold c. zāmīs -i zāmīsī to learn/study d. tūlīs -i tūlīsī to reply e. gōmīs -i gōmīsī to get ready f. fçb# īl -i fç#bīlī to blame g. kābīl -i kābīlī to turn h. sālīg -i sālīgī to polish Dynamic verbs permit the addition of the nominalizer {-ka} to derive nominals. Thus, forms such as tūsīkā ‘act of pushing’, sāgīkā ‘act of teaching’, Nmārīsīkā ‘act of writing’, sālīgīkā ‘act of polishing’ are nominals that exist in usage. Unlike dynamic verbs which display the suffix {-i}, the few verbs which fall within stative verbs take the suffix {-a}. Such stative verbs are monovalent and assign University of Ghana http://ugspace.ug.edu.gh 68 qualities to the subject which occurs as their only argument. Example (9) illustrates that stative verbs possess the suffix {-a}. Verb Root Suffix Verb Gloss (9) a. pāgr -a pāgrā ‘be hard’ b. tūīl -a tūīlā ‘be hot’ c. nāl -a nālā ‘be nice’ d. bāās -a bāāsā ‘be soft’ e. gēl -a gēlā ‘be short’ f. mās -a māsā ‘be sweet’ g. vç# -a vçā# ‘be alive’ h. dç# -a dç#ā ‘to lie down’ i. tç# -a tçā# ‘be bitter’ j. mīīs -a mīīsā ‘be bitter’ k. jēt -a jētā ‘be thin’ l. mç#n -a mç#nā ‘be fair’ m. sç#bl -a sç#blā ‘be dark/black’ n. wç#N -a wçN# ā ‘be tall’ o. yçg# s -a yç#gsā ‘be cold’ Interestingly, there are a few stative verbs that can be used dynamically. There is a close resemblance between such stative verbs and the forms they take when they are used as dynamic verbs, because the stative forms are almost always deducible from the dynamic forms and vice versa. Pairs such as dç#ā/dūāgī ‘be lying down/to lie down’, mç#nā/mūnī ‘be ripe/to ripe’, yç#gsā/yç#gī ‘be cold/to cool’, kālā/kālī ‘be seated/to sit’, tūīlā/tōlīN ‘be hot/to heat’ corroborate my assertion. The examples in (10) illustrate the distinction between stative verbs and their dynamic counterparts. University of Ghana http://ugspace.ug.edu.gh 69 (10) a. Aluetemi kālā. Aluetemi be.seated ‘Aluetemi is sitting.’ b. Aluetemi lì kālī. Aluetemi FUT to sit Aluetemi will sit.’ c. Nyíámmú yçg# sā kámā. water.DEF be.cold AFF ‘The water is cold.’ d. Nyíámmú yç#gī kámā. water.DEF to cool AFF ‘The water has cooled.’ e. JèntìNá tūīlā kámā. soup.DEF be.hot AFF ‘The soup is hot.’ f. Mí tōlīN jèntìNá kámā. 1SG to heat soup.DEF AFF ‘I have heated the soup.’ Being stative verbs, none of the verbs here permits the nominal suffix {-ka} to be attached to it. Instead, these stative verbs take a different nominalizing suffix {-im}, so that nominals like pāgrīm ‘hardness’, tūīlīm ‘hotness’, nālīm ‘beauty’, bāāsīm ‘softness’, and gēlīm ‘shortness’ are derived from the verbs pāgrā ‘be hard’, tūīlā ‘be hot, nālā ‘be beautiful’, bāāsā ‘be soft’ and gēlā ‘be short’ respectively. A study of the forms of the verbs and their nominals will reveal that the final vowel of the verb is elided before the nominalizer {-im} is suffixed. This supports my observation that {- a} is a suffix. University of Ghana http://ugspace.ug.edu.gh 70 My data reveal that there are a few stative verbs which exhibit distinct morphological forms. The existential verb bó(ró)15 ‘to exist/be there’ and its negative suppletive ká(ró)16 ‘non-existent/be absent’ are typical examples of such verbs. The suffix {-ro} becomes an obligatory element when a verb complement is absent. (11) a. Anampansa bóró. Anampansa be.there ‘Anampansa is around.’ (lit.: Anampansa is there). b.*Anampansa bóró dòk. Anampansa be.there room c. Anampansa bō dòk. Anampansa be.there room ‘Anampansa is in the room.’ d. *Anampansa bō Anampansa be.there As (11a) shows, the suffix {-ro} is required when the verb does not have a complement. However, it is ungrammatical for the suffix to occur when a complement is present (see 11b). While (11c) is grammatical, (11d) is not. This is because, with the absence of {-ro}, the verb bō ‘be there’ requires a complement. The negative form ká(ró) exhibits a similar morphological behaviour, except that the verb has a metaphorical extension which makes it possible for kā ‘not be there’ to occur without a complement. 15 The reduced form of bó(ró) has a mid-tone (i.e bō) 16 The reduced form of ká(ró) has a mid-tone (i.e. kā) University of Ghana http://ugspace.ug.edu.gh 71 (12) a. Awulie káró. Awulie be.not.there ‘Awulie is not around.’ (lit.: Awulie is not there) b.*Awulie káró dòk. Awulie be.not.there room c. Awulie kā dòk. Awulie be.not.there room ‘Awulie is not in the room.’ d. Awulie kā. Awulie be.not.there ‘Awulie is dead.’ (lit.: Awulie is not there) From (12a) to (12c), we can infer that káró behaves like bóró. First, like bóró, káró cannot occur with a complement as illustrated in (12b). Also, káró must be reduced to kā when a complement is present, as shown in (12c). (12d) provides evidence to my observation that kā occurs without a complement, but it gives the construction a metaphorically extended meaning. When used without a complement, kā typically makes reference to death. 3.1.3 The Verbal Group The Verbal Group has been viewed as a grammatical unit that has a lexical verb as its head. Syntactically, it is possible to have only the lexical verb functioning as a ‘group’. This happens when the the lexical verb occupies a syntactic slot within a clause. With our knowledge that the phrase falls between the morpheme (word) and the clause in the ranking of syntactic elements, we can say that it is possible to have a shift in the rank of a verb upwards to become a phrase when it occupies a syntactic slot. The examples in (13) are used to explain. University of Ghana http://ugspace.ug.edu.gh 72 (13) a. KpóNkú lì lōb jéín. guinea fowl.DEF FUT lay egg ‘The guinea fowl will lay an egg.’ b. KpóNkú kàn lōb jéín. guinea fowl NEG.FUT lay egg ‘The guinea fowl will not lay an egg.’ c. KpóNkú lòb jéín. guinea fowl.DEF lay.PERF egg ‘The guinea fowl has laid an egg.’ The verb phrases in (13a) and (13b) consist of two elements – in (13a) the future tense marker lì and the verb lōb ‘lay’, and in (13b) the VP is made up of the future negation particle kàn and the verb lōb ‘lay’. These two verb phrases in (13) and (13b) come between theirs subject kpóNkú ‘the guinea fowl’ and their objects jéín ‘egg’. In (13c), only the verb lòb ‘lay’, occurs as the VP. Examples (13a) and (13b) provide evidence that verbal particles co-occur with a main verb to modify it. Some linguists have discussed the sequential arrangement and syntactic positioning of verbal particles. For example, adopting Baker’s (1995) semantic model of verb classification, Atintono (2004) identifies nine pre-verbal positions and two main post- verbal positions for GurunE verbal modifiers. It must be pointed from the outset that no deliberate attempt is made to provide a sequence for the occurrence of Buli verbal particles in this work. As we have already indicated, it is rare to find multiple Buli verbal particles occurring in a fixed sequence within the a single clause. Thus, in the sections that follow, we isolate each of the verbal particles that occurs with the Buli verb and discuss the grammatical functions it perfoms. Also, worth mentioning is the fact that grammatical functions are not performed by only verbal particles; such University of Ghana http://ugspace.ug.edu.gh 73 functions are sometimes borne by tone which may interact with verbal particles. Since this study focuses on the verbal particles, we isolate the verbal particles and discuss their grammatical functions, without any conscious attempt to discuss the functions tone performs or how tone interacts with verbal particles to mark functional categories like tense, aspect, modality and polarity. 3.1.3.1 The Buli Verbal Group The Buli Verbal Group is made up of a main verb and verbal particles that co-occur with the main verb to perform grammatical and discourse functions. It should be emphasized that the Buli Verbal Group is discontinuous, since other clausal elements intervene between the main verb and the verbal particles to break what could have formed a sequence. 3.1.3.2 Tense, Time, and Aspectual Particles in Buli Tense is defined as the grammaticalized expression of location in time (Comrie 1985:9). In line with this, Velupillai (2016) describes tense as a timeline which serves as a means of locating an event at a point in time. From these definitions, we may talk about the present, the past and the future, depending on the partitioning of the timeline – the present coincides with the speech time, the past is posterior to the speech time, and the future is anterior to the speech time. Aspect, on the other hand, is viewed as the “…different ways of viewing the internal temporal constituency of a situation.” (Comrie 1976:3). Thus, aspect indicates the temporal structure of an event with regard to the way in which the event occurs in time. Here, the event is either viewed as on-going or completed, beginning, continuing or ending, or iterative (Bhat 1999:43). University of Ghana http://ugspace.ug.edu.gh 74 Generally speaking, tense has been shown to be less prominent in Gur languages. Saanchi (2006) describes aspect as a fundamental category in Dagaare, since all Dagaare verb phrases are marked for it. Along the same lines, Atintono (2013) notes that, in discussions pertaining to the grammar of Gurunԑ, aspect is viewed as more important than tense and modality. These observations give credence to the findings of typological studies like Bhat (1999) which show that Gur languages are largely aspect- prominent, other than tense-prominent or mood-prominent. Nonetheless, there is the need for further studies into the verbal system of individual Gur languages to reveal their specific features with regard to the expression of tense or aspect. Buli is predominantly aspectual – with the exception of the future tense marker lì which clearly marks future, most other particles mark aspect. In the following subsections, we discuss the functional categories of some Buli verbal particles that mark aspect, tense, time, or aspect. 3.1.3.2.1 The Habitual Marker ‘à’ This preverbal particle is used in the indicative to present a habitual reading of the verb to show that an action lasts for an extended period of time. It expresses the idea that the action occurs frequently as the habit of the subject. It is also possible for a universal worldview17 to be expressed with the habitual marker. Consider the use of the habitual marker à in example (14). 17 A universal worldview is expressed with a generic noun which has no determiners. University of Ghana http://ugspace.ug.edu.gh 75 (14) a. Pòlìká à nyū dáám. police.DEF HAB drink alcohol ‘The policeman drinks alcohol.’ b. Dùìnsá à nyū zíím. mosquito.PL HAB drink blood ‘Mosquitoes suck blood.’ The presence of the habitual marker à in (14) indicates that the action of denoted by the verb occurs regularly. Thus, the construction implies that the policeman is used to drinking alcohol, an act which has develop into a habit. Like example (14a), (14b) has a habitual reading; however, the use of the generic noun dùìnsá ‘mosquitoes’ gives it a universal sense. An easy way of showing that the habitual marker indicates that an action occurs as a habit, is to introduce a time adverb as exemplified in (15) below: (15) a. Mì kòwá à dā záá dāmíé:ná. 1SG father.DEF HAB buy millet ADV ‘My father buys millet every day’ b. Pòlìká dāāNá dE@ à nyū dáám. police.DEF ADV HAB drink alcohol. ‘The policeman drinks alcohol these days.’ The introduction of the adverbs dāmíé:ná ‘every day’ and dāāNá dE @ ‘these days’ in (15a) and (15b) respectively gives them a habitual meaning; thus, they indicate that the actions denoted by the verbs dā ‘buy’ and nyū ‘drink’ occur on a regular basis. University of Ghana http://ugspace.ug.edu.gh 76 Vowel lengthening is not fundamental in Buli as it is in Dagaare. What looks like vowel lengthening in Buli is actually a reduction in form of the focus marker ká to the vowel á which is in turn assimilated totally by the vowel of the monosyllabic verb which precedes it. I draw on the Dagaare example in (16a) and the Buli examples (16b and 16c) to support my claim. (16) a. A bie kUç-rEE a zie DEF child weed-IMPERF DEF place ‘The child is weeding the place.’ (Saanchi 2006:53) b. Kpārōāwá à kpā nàNgà:Nká. farmer.DEF HAB weed.FOC backyard.DEF ‘The farmer weeds the backyard.’ c. Bííká à dE#E # túé. child.DEF HAB eat.FOC beans ‘The child eats beans.’ As the English gloss in (16a) shows, the morpheme -rEE in kUçrEE ‘weed’ is an imperfective marker. Saanchi (2006) indicates that the vowel of the imperfective marker {E} is -ATR because the vowels {U, ç} of the verb have the feature -ATR. Thus, where the vowel(s) of the verb is +ATR, the vowel(s) of the imperfective marker is/are also +ATR. In Buli, the focus marker ká necessarily follows transitive verbs which have open syllable verbs in order to enable the verb host another argument. This is to say that the full forms of the verbs kpāā ‘weed’ and dE#E# ‘eat’ in (16b) and (16c) are kpā + ká and University of Ghana http://ugspace.ug.edu.gh 77 dE# + ká respectively. It is as a result of the reduction in form of ká to á in speech which accounts for the constructions in (16b) and (16c). Thus, the substantive forms of (16b) and (16c) are (17a) and (17b) respectively. (17) a. Kpārōāwá à kpā ká nàNgà:Nká. farmer.DEF HAB weed FOC backyard.DEF ‘The farmer weeds the backyard.’ b. Bííká à dE # ká túé. child.DEF HAB eat FOC beans ‘The child eats beans.’ The fact that disyllabic and polysyllabic verbs cannot be lengthened supports my claim that vowel lengthening is not fundamental in Buli. The ungrammatically of (18b) and (18d) confirms this observation. (18b) is ill-formed, because Nārī ‘fetch’ is a disyllabic verb, making it impermissible for the focus marker ká to be reduced to á. A similar situation pertains to (18d) where the verb Nmārīsī ‘write’ is polysyllabic. Other disyllabic/polysyllabic verbs, including dārī ‘pull’, sūrī ‘wash’, bīlī ‘clean’ and bā:lī ‘sew’ do not permit a reduction in the form of the focus marker ká. (18) a. Awén à Nārī ká nyíám. Awén HAB fetch FOC water ‘Awén fetches water.’ b. *Awén à Nārīī nyíám. Awén HAB fetch water University of Ghana http://ugspace.ug.edu.gh 78 c. Félíká à Nmārīsī ká gbáN European.DEF HAB write FOC book ‘The European writes a book.’ d. *Félíká à Nmārīsīī gbáN European.DEF HAB write book During speech, anticipatory assimilation occurs when a pronominal subject is preceded by the habitual marker. This causes sequences like mí + à, wà + à, fí + à, nè + à, tè + à, bà + à and sì + à to be rendered as máà, wàà, fáà, nàà, tàà, bàà, and sàà respectively.18 Consider the examples in (19). (19) a. Máà kūlī 1SG.HAB go.home ‘I go home.’ b. Mí à kūlī. 1SG HAB go.home ‘I go home.’ c. Tàà pūūsī wén 1PL.HAB pray God ‘We pray.’ d. Tì à pūūsī wén 1SG HAB pray God ‘We pray.’ 18 Kröger uses the term “contamination of the pronoun” to describe such forms them, but Schwarz (2005) treats them as full pronouns and refers to them as imperfective subjective pronouns. University of Ghana http://ugspace.ug.edu.gh 79 e. Wàà dāā zóm. 3SG.HAB buy.FOC flour ‘He buys flour.’ f. Wà à dāā zōm. 3SG HAB buy.FOC flour ‘He buys flour.’ While (19a), (19c) and (19f) occur in speech, it is rare to hear (19b), (19d) and (19f) in native speakers’ speech. During speech, the pronominal subjects (mí ‘I’, tì ‘we’, and wà ‘he/she’) merge with the habitual marker a$ so that the ‘contaminated’ forms (máà, tàà, and wàà) are derived. This explains why Schwarz (2005) describes them as imperfective subjective pronouns, although they carry in them an aspectual marker (a property which pronouns do not possess). It should be added that these forms cannot be substituted with nouns. Stative verbs do not take the habitual marker, because they denote a state which cannot be changed volitionally by an external agent. Therefore, property denoting verbs such as nālā ‘be nice’, māsā ‘be sweet’, tç#ā ‘be bitter’ sōblā ‘be dark’, mōnā ‘be fair’ etc. cannot co-occur with à. This accounts for the ungrammaticality of examples (20b), (20d) and (20f). (20) a. Gàrùgèlìkú nàlá. smock.DEF be.nice ‘The smock is nice.’ b. *Gàrùgèlìkú à nālā. smock.DEF HAB be.nice University of Ghana http://ugspace.ug.edu.gh 80 c. Tì:mmú tç$á. medicine.DEF be bitter ‘The medicine is bitter.’ d. *Tì:mmú à tç#ā. medicine.DEF HAB be.bitter e. CháámíNá dòblá. sheabutter.DEF be.heavy ‘The sheabutter is heavy.’ f. *CháámíNá à dōblā. sheabutter.DEF HAB be.heavy Another group of Buli stative verbs are verbs of perception/cognition and verbs of relation. The former type of verbs includes sE#b ‘to know’ and its negative suppletive zE# ‘not know’, Nmā ‘to blame’, and the latter group comprises tā ‘to have’, and sʋ#ā ‘to own’. We illustrate the use of these verbs in (21). (21) a. Mí sE $b nààwá. 1SG know chief.DEF ‘I know the chief.’ b. *Mí à sE#b nààwá. 1SG HAB know chief.DEF c. Kpārōāwá tà níígá. farmer.DEF to have cattle ‘The farmer has cattle.’ d.* Kpārōāwá à tā níígá. farmer.DEF HAB to have cattle University of Ghana http://ugspace.ug.edu.gh 81 As example (21b) reveals, it is ungrammatical for the stative verb of cognition sE#b ‘to know’ to take the habitual marker. In the same vein, as exemplified in (21d), the verb of relation tā ‘to have’ cannot take the habitual marker. 3.1.3.2.2 The Progressive Marker ‘bóràà’ The progressive marker, as its name implies, expresses the idea that an action has started, and it is still in progress or on-going at the time of speech. In this case, the action runs simultaneously with the speech. It should be mentioned that bóràà is derived from the fusion of the existential bóró and the aspectual marker à19. It will be shown it in the next chapter that, in some communicative contexts, the two particles are shared between two verbs of a serial verb construction to mark the progressive aspect. Thus, (22b) carries the implication that an action runs concurrently with the speech time or is still on-going at the speech time. (22) a. Nááwén bóró. God be.there ‘God exists.’ (lit.: God is there) b. Nááwén bóràà māārī núrúbíík. God PROG help mankind ‘God is helping mankind.’ (lit.: God is there helping mankind) During speech, the aspectual marker ‘à’ is fused into bòró, so that bóráá is heard. The fusion of the existential marker bóró and the aspectual marker à, with its accompanying sound and tonal modifications, facilitates easy articulation of the two separate forms. It 19 In SVCs that require the sharing of bóró and à between two serialized verbs, I gloss à as ASP (aspectual). University of Ghana http://ugspace.ug.edu.gh 82 is difficult to tease apart the two forms, since they are always rendered as the single unit bóràà in speech. Wherever necessary, I will provide different glosses for the two particles (bóró and à) that form the progressive marker. I use the examples in (23) to further support my obvervation that bóràà is a progressive marker. (23) a. Mí kòwá bóràà dā záá. 1SG father.DEF PROG buy millet ‘My father is buying millet.’ (lit.: My father is there buying millet.) b. Pólíká bóràà nyū dáám. police.DEF PROG drink alcohol ‘The policeman is drinking alcohol.’ (lit.: The policeman is there drinking alcohol) Each of the examples in (23) suggests that an action has begun and is still on- going; in (23a) the buying of the millet has ended, and in (23b) the drinking of the alcohol is still on-going. It should be indicated that the introduction of bóró sometimes depends on the type of verb. Action verbs which depict movement (e.g. chēN ‘go’ and jō ‘enter’) cannot take the particle bóró. In this case, the expression becomes ambiguous, because it is the context that determines whether the expression has a progressive reading or a habitual reading. The ungrammaticality of (24b) and (24d) is attributable to the co-occurrence of the movement verbs chēN ‘go’ and jō ‘enter’ with bóràà. University of Ghana http://ugspace.ug.edu.gh 83 (24) a. Atie à chēN tālīm. Atie ASP go farm ‘Atie goes to the farm/Atie is going to the farm.’ b. *Atie bóràà chēN tālīm. Atie PROG go farm. c. Níchāāníwá à jō dòk. visitor.DEF ASP enter room ‘The visitor enters the room/The visitor is entering the room.’ d. *Níchāāníwá bóràà jō dòk. visitor.DEF PROG enter room The high tone counterpart of à (i.e. á) is used in the imperative mood. Unlike à which does not occur at a sentence-initial position, á may occur at a sentence-initial position. This is seen when a command is made directly to a singular addressee, in which case the speaker needs not address the addressee by name or by the second person singular pronoun fí ‘you’. Consider example (25) below. (25) a. (Fí) á pōtī sùNkpáámú! 2SG IMP deshell groundnut.DEF ‘(you) Be deshelling the groundnut!’ b. (Fí) á sūrī gà-tìáká! 2SG IMP wash cloth.DEF ‘(you) Be washing the clothes!’ The second person pronoun fí ‘you, as indicated by the brackets, is implied in each of the examples – the addressee receives a direction or a command directly from the speaker. In such statements, the speaker often requires that the addressee undertakes an University of Ghana http://ugspace.ug.edu.gh 84 action whose time of completion is undetermined. Also, a second action (of the speaker or a third party) is sometimes intended or expected by the speaker to run concurrently with the first action. Hence, as illustrated in (26), a second part of the sentence can be added as illustrated. (26) a. á pōtī sùNkpáámú àtè n$ jām. IMP deshell groundnut.DEF COMP 1SG come ‘Be deshelling the groundnut till I come.’ b. á sūrī gà-tìáká àtè wà sō bísíNá. IMP wash cloth.DEF COMP 3SG bathe child.PL.DEF ‘Be washing the clothes, while he bathes the children.’ (Be washing the clothes so that he bathes the children.) In each of the examples in (26), the end or time of completion of the actions encoded by the verbs pōtī ‘deshell’ and sūrī ‘wash’ is not specified. It is also possible to have a subject before the particle á in some cases. This happens when there is more than one addressee or when the action indicated by the verb in the utterance is to be performed by a third party. Consider the examples in (27). (27) a. Nídōāmá á nāg gīNgānìNá. man.PL.DEF IMP beat drum.PL.DEF ‘The men should be beating the drum!’ b. Nípōmá á sīā yííní. woman.PL.DEF IMP sing song.DEF ‘The women should be singing the song!’ University of Ghana http://ugspace.ug.edu.gh 85 c. Bà á nāg gīNgānìNá. 3PL IMP beat drum.PL ‘They should be drumming!’ d. Nì á sīā yííní. 2PL IMP sing song ‘(You should) Be singing the song!’ In (27a) and (27b), the speaker instructs the addressee that an action must be performed by a third party. Thus, in (27a) nídōāmá ‘the men’ should beat the drum, and in (27b) nípōmá ‘the women’ should sing the song. The referents (subjects) bà ‘they’ and nì ‘you’ are required in (27c) and (27d) respectively, because they are plural. As expected, when a pronominal subject precedes the imperative marker á, there is a fusion of the pronominal subject and the imperative marker into a single form. Therefore, in speech, the sequence bà + á, and nì + á in (27c) and (27d) respectively are rendered as bàá and nàá. 3.1.3.2.3 The Future Tense Marker ‘lì/àlì’ The pre-verbal particle àlì and its variant lì, mark futurity. Speakers occasionally use this future tense marker interchangeably with nì20, but nì typically carries a prospective reading. To draw a clear distinction between the use of àlì/lì and nì can be quite challenging, since future tense markers may express modality like hope or a wish. This is to say that, because the future marker àlì/lì takes up functions of modality performed by nì, the latter may be misconstrued as marking futurity. 20 Schwarz (2005) renders the future markers lì/alì as le/ale, and presents nì which she describes as their variant, as ne. University of Ghana http://ugspace.ug.edu.gh 86 The examples in (28) show that the use of the future tense marker àlì (lì) does not always yield a future time reading; apart from expressing futurity, it has other semantic implications related to modality. (28) a. Tì lì chèN yābā. intention 1PL FUT go market ‘We will/shall go to the market.’ b. Mí lì dE# kòòdūkú21. request 1SG FUT go banana.DEF ‘I will eat the banana.’ c. NyìN-yògsā àlì ja#m. hope good health FUT come ‘There shall be good health.’ d. Nmòrùkú lì nī. prediction rain.DEF FUT rain ‘It will rain.’ (28a) indicates that the speaker has the intention of going to the market, and (28b) makes a request by expressing the desire of the speaker to eat some of the addressee’s bananas. A good scenario for (28c) is when an addressee is in ill-health and the speaker wishes the addressee recovers. In that case, (28c) expresses hope, but not futurity. (28d) may be uttered to make a prediction of a downpour in the face of a cloudy weather. It is also necessary to point out that, although it is possible for all the sentences in (28) to have a future reading, each of the constructions has a basic meaning that is not related 21 kòòdūk ‘banana’ is a loan word taken from Akan. It is a phonologically modified form of the Akan word kwedu ‘banana’. University of Ghana http://ugspace.ug.edu.gh 87 to futurity. Native speakers usually add temporal adverbials such as jīnlE @ ‘today’, chūm ‘tomorrow’, vç#nV#N ‘in the next two days’, etc. when they want to evoke the futuristic sense. Time adverbials have been added to the examples in (28) and presented in (29) to indicate futurity. (29) a. Tì lì chèN yābā vç#nV#N. 1PL FUT go market next two days. ‘We will go to the market in the next two days.’ b. Mì chūm lì dE# kòòdūkú. 1SG tomorrow FUT go banana.DEF ‘I will eat the banana tomorrow.’ c. NyìN-yògsā àlì jām dādíídáí. good health FUT come one day ‘There will be good health one day.’(lit.: Good health will come one day) d. Nmòrùkú lì nī jùnōáí. rain.DEF FUT rain evening ‘It will rain in the evening.’ As we hinted earlier, nì basically expresses modality rather than marking futurity. In (30a) the speaker expresses the desire to eat some of the addressee’s rice, and in (30b) the speaker expresses the intention of going to the farm. University of Ghana http://ugspace.ug.edu.gh 88 (30) a. Mí nì dE # mùmìNá. 1SG MOD go rice.DEF ‘I will eat the rice (desire).’ b. Tì nì chèN yābā. 1PL MOD go market ‘We will go to the market (intention).’ 3.1.3.2.4 The Past Time Marker ‘pòòm’ This particle pòòm is used to mark past time. A speaker uses it to indicate a state or an event that existed not too long ago. It may occur alone or with the progressive marker bóràà. A dichotomy can be drawn between two: pòòm typically occurs with stative verbs to indicate that a state or an event which existed not too long ago no longer exists. Consider the use of pòòm in (31) below. (31) a. Adem pòòm tāā bàN. Adem PST own.FOC bangle ‘Adem owned a bangle (some time ago). = He no longer owns it. b. Baaba pòòm bōō yérí. Baaba PST be.there.FOC home ‘Baaba was at home.’ = Baaba is no longer there. In (31a) the speaker implies that the referent (Adem) had a bangle some time ago. A similar situation pertains to (31b) which expresses the idea that the referent (Baaba) was at home not too long ago. I use the timeline in Fig. 3.1 below to illustrate the use of póóm. University of Ghana http://ugspace.ug.edu.gh 89 change in state or change in possession Past Time Speech Time Future Time Figure 3.1: Timeline for the use of ‘póóm’ ‘pòòm’ co-occurs with the progressive marker bóràà when the verb is dynamic to express the idea that an action was still being performed by a referent at the time. It behaves like the past progressive in English, and it indicates that the speaker witnessed the occurrence of the event. It typically answers the question [Subject]+ pòòm+ bóràà nyE# ká bòà? “What were you doing?”. This implies that the action has been completed. Here, the speaker/interrogator has a particular time in mind and seeks to find out what transpired during that period of time. I use example (32) to explain this further. (32) a. Ajíndém pòòm bóràà kpāā mā. Ajindem PST PROG weed.FOC AFF ‘Ajíndém was weeding.’ b. Agúúk pòòm bóràà gīsī ká píísìNá. Aguuk PST PROG search FOC sheep.PL.DEF ‘Agúúk was searching for the sheep.’ The constructions in (32) show that an action was being undertaken at a particular during in the past. A speaker or an interrogator who seeks to know what transpired during that period will pose the question Agúúk pòòm bóràà nyE# ká bòà? “What was Agúúk doing?”. Examples (32a) and (32b), respectively, will be the responses to this question, if Ajíndém was weeding or searching for the sheep at the time being referred to. Since the interrogator has a particular time in mind, he/she may specify that time. University of Ghana http://ugspace.ug.edu.gh 90 Thus, an interrogator may ask, Mí nì wū fù lá, fí pòòm bóràà nyE # ká bòà? “What were you doing when I called you?”. (33a) will be the response if the addressee was praying when the interrogator called, and (33b) will be the response if the addressees were washing clothes when the interrogator called. (33) a. Mí póóm22 bóràà pūūsī ká wén. 1SG PST PROG pray FOC God ‘I was praying.’ b. Tàmá 23 póóm bóràà sūrī ká kç$lìmá. 2PL PST PROG wash FOC cloth.PL ‘We were washing clothes.’ Fig. 3.2 below illustrates the use of póóm bóràà . Action was on-going completed, but it has been completed. Past Time Speech Time Future Time Figure 3.2: Timeline for the use of ‘póóm bóràà’ 3.1.3.2.5 The Present Time Marker ‘nyìèm’ The particle nyìèm indicates that an activity or an action occurs at a specific period of time as a habit of the speaker or the referent. Example (34a) suggests that the thief would have smoked weed at specified period of time. A scenario will make this clearer. When a noisy thief is very quiet on a particular occasion and one wants to find out the 22 The low tone of pòòm becomes high, when the pronominal subject bears a high tone. 23 The pronominal subject tàmá is different from tì, in that while the former makes reference to a portion of a entire group of people, the latter refers to everyone within the entire group (cf. Sulemana, 2012) University of Ghana http://ugspace.ug.edu.gh 91 cause of his quietness, (34) may be given as a response to indicate that the thief would have smoked whenever he disturbs. Thus, nyìèm indicates that although an action is not being undertaken at the speech time, it happens at a point in time under the influence of something, and it will happen if it is triggered by again. (34), therefore, suggests that the thief will make noise, once he smokes. ‘nyìèm’ may also co-occur with the progressive marker, bóràà, to indicate that an activity would have been being performed at a certain time. (34b) and (34c) express the idea that the referent would have been undertaking an action at a certain time; thus, in (34b) the referent zúéwá ‘the thief’ would have been smoking at a specified time, and in (34c) the referent bísíNá ‘the children’ would have been deshelling groundnut at a specified time. Consider the constructions in (34). (34) a. Zúéwá nyìèm nyūū wúúk. thief.DEF PRES drink.FOC weed ‘The thief would have smoked weed.’ b. Zúéwá nyìèm bóràà nyūū wúúk. thief.DEF PRES PROG drink.FOC weed ‘The thief would have been smoking weed.’ c. BísíNá nyìèm bóràà pōtī ká sùNkpáám. child.PL.DEF PRES PROG deshell FOC groundnut ‘The children would have been deshelling groundnut.’ A general observation can be made here for such constructions, that the focus marker ká obligatorily follows closed monosyllabic verbs and disyllabic/polysyllabic verbs, but open monosyllabic verbs usually cause the focus marker to be reduced to the vowel University of Ghana http://ugspace.ug.edu.gh 92 á which in turn assimilates the final vowel of the preceding verb totally. The presence of the focus marker makes it possible for the final argument to be hosted. (35) a. Aniak nyìèm bóràà līmsī ká wà nç#Nkú. Aniak PST PROG wait FOC 3SG boyfriend.DEF ‘Aniak has been waiting for her boyfriend.’ b. Nípōōbíní nyìèm bóràà sē ká chūmbēná. lady.DEF PST PROG roast FOC corn.DEF ‘The lady has been roasting corn.’ c. Nípōōbíní nyìèm bóràà sēē chūmbēná. lady.DEF PST PROG roast.FOC corn.DEF ‘The lady has been roasting corn.’ The verb līmsī ‘wait’ in (35a) below is disyllabic and requires the focus marker to follow it. Similarly, as illustrated in (35b), the focus marker is required after the open monosyllabic verb sē. As we have already seen in some of the previous examples, the focus marker is usually fused into an open monosyllabic verbs. Thus, in (35c) ká is reduced to the vowel /-a/ which is in turn assimilated totally by the final vowel of the verb sē, such that the final vowel of the verb appears to be lengthened. 3.1.3.2.6 The Distant Past Marker ‘jàm’ The particle jàm is a grammaticalised form that has evolved from the verb jām ‘come’ to function as a distant past time marker. It is used to indicate that an event which once took place has ended, or a state which once existed has changed over time. Often, it is the case that the speaker is uncertain about the exact time the event took place, and uses jàm to indicate that the event happened in the distant past or a very long time ago. Due to this, the particle is prevalent in the opening of folktales. jàm is also used to recount University of Ghana http://ugspace.ug.edu.gh 93 histories, including those whose evidence are difficult to find. The examples in (36) below illustrate the function of jàm. (36) a. Mì kòwá jàm tā ká níígà. 1SG father.DEF PST have FOC cattle ‘My father once had cattle.’ b. A$kūsūN jàm mç#nā kámā. A$kūsūN PST be.fair AFF ‘A$kūsūN used to be fair.’ c. Bà jàm bīāk Yéézú ká Betilihem. 2PL PST give birth Jesus FOC Bethlehem ‘Jesus was born in Bethlehem…’ Both (36a) and (36b) express the idea that a state has ceased to be as it was previously. Thus, in (36a) the speaker implies that his father no longer owes cattle, and in (36b) the speaker tells us that AkusuN’s once fair complexion is no more. The use of jàm in (36c) is an indication that the time of Jesus’ birth is took place several years ago. 3.1.3.3 Negation Markers in Buli – ‘kàn’ and ‘àn’ Two negative markers, kàn and àn, occur in Buli to negate the truth value of either specific elements within the sentence or the entire sentence. The choice of any of these particles is strictly determined by the tense/aspect in operation (cf. Akanlig-Pare 2005). Whereas kàn is typically a future negation marker, àn is a negative perfective marker. Consider the use of kàn in (37b) and the use of àn in (37d). University of Ghana http://ugspace.ug.edu.gh 94 (37) a. Tì lì chēng kúmmú 1PL FUT go funeral.DEF ‘We will attend the funeral.’ b. Tì kàn chēng kúmmú. 1PL NEG.FUT go funeral.DEF ‘We will not attend the funeral.’ c. Apók nyà wáámú yá Apók see.PERF snake.DEF ASS ‘Apók has seen the snake.’ d. Apók àn nyà wáámú. Apók NEG see.PERF snake.DEF ‘Apók hasn’t see the snake.’ Whereas (37b) is the the negative form of (37a), (37d) is the negative form of and (37c). kàn is used in (37b) to indicate that an event will not occur posterior to the speech time. Thus, the subject tì ‘we’ referred to by the speaker will not attend the funeral which is yet to take place. In (37d) the negation marker àn occurs with the perfective verb nyà ‘to see’ to indicate that the referent, Apók, hasn’t seen the snake. A study of the examples in (37) will reveal that the two polarity particles exhibit different syntactic distributions. kàn occurs as the negative form of the future marker lì, whose syntactic position it occupies. We may put this in simple terms by saying that In negative future sentences, the negative future marker kàn replaces the affirmative future marker lì. Therefore, the negative form of the affirmative construction presented in (38a) is (38b) but not (38c). University of Ghana http://ugspace.ug.edu.gh 95 (38) a. Mí chúm lì kō būūkú 1SG tomorrow FUT kill goat.DEF ‘I will kill the goat tomorrow.’ b. Mí chúm kàn kō būūkú. 1SG tomorrow NEG.FUT kill goat.DEF ‘I will not kill the goat tomorrow.’ c. *Mí chúm lì kàn kō būūkú. 1SG tomorrow FUT NEG.FUT kill goat.DEF The particle kàn may have a habitual reading in certain contexts (see example 39b). To avoid such ambiguity, speakers often use time adverbials together with the future negation marker (see example 38b above). (39) a. Abila à Nçb# bīāk. Abila HAB chew dog ‘Abila eats dog.’ b. Abila kàn Nçb# bīāk. Abila NEG chew dog ‘Abila doesn’t eat dog/Abila will not eat dog.’ The particle àn also occurs in the negative progressive aside the negative perfective. As expected, àn must necessarily co-occur with bóràà to express the idea that an action is not on-going at the time the speech was being made. In the negative progressive àn precedes bóràà so that the sequence àn + bóràà + V24. is yielded. I use (40) to explain further. 24 In speech, it is rare for native speakers to utter each linguistic unit distinctively. University of Ghana http://ugspace.ug.edu.gh 96 (40) a. BísíNá bóràà so#. child.PL PROG bathe ‘The children are bathing.’ b. BísíNá àn bóràà so#. child.PL NEG PROG bathe ‘The children aren’t bathing.’ c. Bà bóràà sēē bóósúk. 3PL PROG roast.FOC python ‘They are roasting a python.’ d. Bà àn bóràà se# bóósúk. 3PL NEG PROG roast python ‘They are not roasting python.’ The examples in (40b) and (40d) affirm my observation that the particle àn is used in the negative progressive. It is also noteworthy that unlike in future negative constructions where the future negation marker kàn replaces the affirmative future marker lì, àn does not replace the progressive marker. Instead, it is placed before the progressive marker to derive àn + bóràà + à. During rapid speech, it is quite difficult to perceive or isolate each of the elements as a distinct linguistic unit. The sequence àn + bóràà in (40b), for instance, will be rendered as àm + bóráá. Where the subject is a pronoun as in (40d), the fusion becomes more severe, since the subject is affected. Thus, bà àn bóràà in (40d) will be rendered as bàm + bóráá. University of Ghana http://ugspace.ug.edu.gh 97 Data available also show that the high tone counterpart of kàn (i.e. kán) is used in the imperative to command an addressee to put a stop to an act. We use the examples in (41) to explain this. (41) a. Nāg bííká! beat.IMP child.DEF ‘Beat the child!’ b. Kán nàg bííká! NEG.IMP beat child.DEF ‘Don’t beat the child!’ c. Kō kpónkú! kill.IMP guinea fowl. DEF ‘Kill the guinea fowl!’ d. Kán kò kpónkú! NEG.IMP kill guinea fowl. DEF ‘Don’t kill the guinea fowl!’ Both (41a) and (41c) are commands that require that the addressee performs an action. In (41a) the speaker instructs the addressee to beat the child, and in (41c) the speaker instructs the addressee to kill the guinea fowl. (41b) and (41d), the negative forms of (41a) and (41c) respectively, are commands that require that the addressee does not perform the actions. As the English glosses show, the particle kán, is a negative imperative marker that is used to instruct an addressee to cease the performance of an action. University of Ghana http://ugspace.ug.edu.gh 98 3.1.3.4 Other Verbal Particles In the preceding discussion, I focused on verbal particles that express tense, aspect, polarity and time. The expression of mood was also given some attention, since some of the particles discussed carry in them certain aspects of mood. Here, I discuss verbal particles that perform other grammatical and discourse functions. Verbal particles that mark assertiveness, conditionality, purpose, modality etc. are discussed under different subsections. 3.1.3.4.1 The Assertive Marker ‘yá’ A speaker uses the assertive marker ‘yá’ to indicate forcefully that an action took place, when he/she is confident about its occurrence. It usually occurs immediately after a transitive verb when the valency of the verb is reduced so that the object or complement is not syntactically realized. Its presence also makes it impermissible for the focus marker ká to follow the verb. During speech, the assertive marker yá is often reduced to the clitic ‘i’. This clitic is capable of influencing the vowel of the verb phonologically so that /i/ becomes /I/ when the vowel of the verb is -ATR. This is illustrated in (42). (42) a. TúéNá bè yá / TúéNá bèí. Bean.PL.DEF cook.PERF ASS ‘The beans is cooked.’ b. Fíókú kpì yá / Fíókú kpìí. baboon.DEF die.PERF ASS ‘The baboon is dead.’ c. Ab$ ìlí sàtI$ yá / Ab$ ìlí sàtI$I @ Abili slip.PERF ASS ‘Abili has slipped.’ University of Ghana http://ugspace.ug.edu.gh 99 Example (42) illustrates the use of the assertive marker yá, and how it can occur as a clitic. Reading from the clitic forms, we realize that bè + yá, kpì + yá, and sàtI$ + yá become bèí, kpìí and sàtI $I @ respectively. (42c) confirms my observation that the clitic ‘i’ is realized in speech as /I/ to agree with the -ATR feature of the verb’s (i.e. sàtI$) final vowel. Although the assertive marker usually co-occurs with intransitive verbs, it also co- occurs with transitive verbs when the object of the transitive verb is unspecified or unexpressed. The object of a transitive verb may be unexpressed, when the object is one that is implied by the verb and automatically understood by the addressee. Such verbs have been described in the linguistics literature (cf. Nwachukwu 1985; Essegbey 2010; Korsah 2011; Anyanwu 2012) as Inherent Complement Verbs (ICV)25. Consider the examples provided in (43) below. (43) a. Kpīāká lòb yá. fowl.DEF lay.PERF ASS ‘The fowl has laid an egg.’ b. Bīlēōkú sàm yá. toddler.DET urinate.PERF ASS ‘The toddler has urinated.’ c. Nūīmú zàànì yá. bird.DEF defecate.PERF ASS ‘The bird has defecated.’ 25 Nwachukwu (1985) defines an ICV as a verb whose ‘citation form is obligatorily followed by a meaning-specifying complement.’ University of Ghana http://ugspace.ug.edu.gh 100 The verbs lòb ‘lay (egg)’, sàm ‘urinate’, and zàànì ‘defecate’ in (43) are ICVs, because their object complements are implied and need not to be expressed by the speaker; the addressee knows the object, since it is specified by the verb. Thus, the constructions in (44) are unlikely to be heard in native speakers’ speech, because the verbs lòb ‘lay’, sàm ‘urinate’ and zàànì ‘defecate’ do not require the speaker to mention the objects jéín ‘egg’, sìnsām ‘urine’, and bíntá ‘faeces’ before the utterances can be understood. (44) a. ?Kpīāká lòb jéín. fowl.DEF lay.PERF egg ‘The fowl has laid an egg.’ b. ?Bílìòkú sàm sìnsām. baby.DEF urinate.PERF urine ‘The baby has urinated.’ c. ?Nūīmú zàànì bíntá. bird.DEF defecate.PERF faeces ‘The bird has defecated.’ The objects of such verbs are expressed when the object is different from what the addressee has in mind. In (45), for instance, the object must be expressed, since the verb sàm ‘urinate’ does not specify the object zíím ‘blood’ but sìnsām ‘urine’ (45) Bílìòkú sàm zíím. child.DEF urinate.PERF blood ‘The child has urinated blood.’ Buli possesses transitive verbs that permit anti-causative constructions. Verbs of such anti-causative constructions require that the assertive marker yá follows them. The examples in (46) illustrate the use of yá in causative and anti-causative constructions. University of Ghana http://ugspace.ug.edu.gh 101 (46) a. Akànkó mç$bì kpálábíká. Akanko break.PERF earthenware bowl.DEF ‘Akanko has broken the earthenware bowl.’ b. Kpálábíká mç$bì yá. earthenware bowl.DEF break.PERF ASS ‘The earthenware is broken.’ c. *Kpálábíká mç$bì earthenware bowl.DEF break.PERF The meaning expressed in (46a) is different from that which is expressed by (46b). While there is a causing agent Akànkó in (46a), there is no causing agent in (46b). Thus, in (46a) the change in the state of patient, kpálábíká ‘the pot’, is caused by an agent (Akànkó), but in (46b) the causing agent is unexpressed. The ungrammaticality of (46c) is caused by the absence of the assertive marker yá. This supports my observation that anti-causative constructions require that the assertive marker yá follows them. It is not possible for the assertive marker to occur after the object in negative sentences. Therefore, while (47b) is grammatically acceptable as the negative form of (47a), (47c) is considered ungrammatical. (47) a. Akànkó mç$bì kpálábíká yá. Akanko break.PERF earthenware bowl.DEF ASS ‘Akanko has broken the earthenware.’ b. Akànkó àn mç$bì kpálábíká. Akànkó NEG break.PERF earthenware bowl.DEF ‘Akànkó has not broken the earthenware.’ c. * Akànkó àn mç$bì kpálábíká yá. Akànkó NEG.PERF break.PERF earthenware.DEF ASS University of Ghana http://ugspace.ug.edu.gh 102 3.1.3.4.2 The Emphatic Marker ‘lá’ Syntactically, lá occurs as a pre-verbal particle. I use example (48) to illustrate the function of lá and its occurrence as a pre-verbal particle. (48) a. BúdùktìNá pòm chāāb. billy-goat.PL.DEF butt.PERF each other ‘The billy-goats butted each other.’ b. BúdùktìNá lá pòm chāāb. billy-goat.PL.DEF EMPH butt.PERF each other ‘The billy-goats butted each other (EMPH).’ c. Chīnní mç$bì yá. calabash.DEF break.PERF ASS ‘The calabash is broken.’ d. Chīnní lá mçb$ ì. calabash.DEF EMPH break.PERF ‘The calabash is broken (EMPH).’ In (48b) the presence of lá makes the speaker’s assertion stronger than in (48a) where the particle is absent. Similarly, a claim stronger than (48c) is made in (48d) with the emphatic marker lá. Thus, (48b) indicates strongly that the billy-goats butted each other, and (48d) indicates strongly that the calabash broke. University of Ghana http://ugspace.ug.edu.gh 103 3.1.3.4.3 The Affirmative Marker ‘kámā’ This post-verbal particle functions as an affirmative marker which typically occurs in a response to an interrogation to indicate that an action has occurred. Thus, its presence in an utterance affirms that an action, which an interrogator seeks to find out whether occurred or not, took place. Example (49c), therefore, will typically occur as a response to the question Fì yàlì yá? “Are you married?” (49) a. Wà kùlì yá. 3SG go.home.PERF ASS ‘He has gone home.’ b. Wà àn kùlì yá. 3SG NEG go.home.PERF ASS ‘He hasn’t gone home.’ c. Wà yàlì kámā. 3SG marry.PERF AFF ‘She is married (indeed).’ d. *Wà àn kùlì kámā. 3SG NEG marry.PERF AFF The particle behaves like the assertive marker yá discussed in Section 3.1.3.4.1. Like the assertive marker, kámā can occur after a verb which has no complement and after object complements. A dichotomy can, however, be drawn between the two particles, based on their functions and syntactic distribution. While yá may occur in both affirmative and negative constructions, kámā (being an affirmative marker) occurs solely in affirmative constructions. As I have illustrated in (49c), the occurrence of yá and a negation marker yields a grammatical sentence, but an ungrammatical sentence University of Ghana http://ugspace.ug.edu.gh 104 is yielded when kámā co-occurs with a negation marker. Thus, whereas (49b) is the negative form of (49a), (49d) cannot be the negative form of (49c). There is a close relation between the affirmative marker kámā and the focus marker ká, attributable to their inability to co-occur in the same sentence (as shown in 50b). Where the verb is open and monosyllabic and the focus marker is reduced to a vowel, the initial syllable of kámā is elided so that mā serves as the affirmative marker as shown in (50d). This happens when the object of a transitive verb is suppressed. (50) a. Bà dE $ ká sāāb. 3PL eat.PERF FOC tuo-zaafi26 ‘They have eaten tuo-zaafi.’ b.*Bà dE $ ká sāāb kámā. 3PL eat.PERF FOC tuo-zaafi AFF c. Bà dE$ kámā. 3PL eat.PERF AFF ‘They have eaten (indeed).’ d. Bà dE$-E$ mā. 3PL eat.PERF-FOC AFF ‘They have eaten (indeed).’ Example (51b) below affirms my observation that the affirmative marker kámā cannot be rendered as mā when the object is specified. 26 tuo-zaafi is a popular dish prepared from millet or maize flour. University of Ghana http://ugspace.ug.edu.gh 105 (51) a. Bà dE$ sāāmú kámā. 3PL eat.PERF tuo-zaafi.DEF AFF ‘They have eaten the tuo-zaafi (indeed).’ b. *Bà dE$ sāāmú mā. 3PL eat.PERF tuo-zaafi.DEF AFF Closed monosyllabic verbs behave like polysyllabic verbs, with regard to their relation with the affirmative marker, because closed syllables and polysyllabic verbs do not permit kámā to be rendered as mā. Consider the examples provided in (52). (52) a. Wà à dōm kámā. 3SG HAB bite AFF ‘He bites (indeed).’ b.* Wà à dōm mā. 3SG HAB bite AFF c. Wà à chēng kámā. She HAB go AFF ‘He goes (indeed).’ d. * Wà à chēng mā. 3SG HAB go AFF 3.1.3.4.4 The Purposive Markers ‘àlē sòà / nyīN’ The two particles, àlē sòà and nyīN, are used interchangeably and they often co-occur with the complementizer tè in speech.27 They perform the function of indicating the reason or purpose for which an action was performed or an event took place. When 27 The particle tè is sometimes omitted in speech. University of Ghana http://ugspace.ug.edu.gh 106 used in a complex clause, they indicate that the action or event expressed in the second clause (dependent clause) occurred as a sequel to the one expressed in the first clause (independent clause). Thus, they project the action expressed by the independent clause as the cause of the action expressed in the dependent clause. Hence, the action expressed in the dependent clause cannot take place without the action expressed in the independent clause. I use (53) to illustrate this. (53) a. Achūm ālè jām lá, àlē sòà tè Achūm TM come.PERF EMPH PURP COMP bísíNá sò. child.PL.DEF bathe.PERF ‘The children bathed because of Achum’s arrival.’ b. Wáámú àlē sòà tè kpésíNá à kā:ī. snake.DEF PURP COMP fowl.PL ASP squawk ‘The fowls are squawking because of the snake.’ We can infer from these examples that the particle àlē sòà indicates purpose/reason. Its occurrence between the verbs ja#m ‘come’ and so # ‘bathe’ in (53a) indicates that the children bathed because of Achūm’s arrival. Thus, if Achūm had not arrived, the children wouldn’t have bathed. In (53b) àlē sòà indicates that the action of the fowls (i.e. squawking) is caused by the subject referent wáámú ‘the snake’. We have indicated that interchanging àlē sòà with nyīN does not result in any meaning change. Therefore, as the English glosses show, the constructions we saw in (53a) and (53b) are equivalent to those in (54a) and (54b) respectively. University of Ghana http://ugspace.ug.edu.gh 107 (54) a. Achūm ālè jām lá, nyīN tè Achūm TM come.PERF EMPH PURP COMP bísíNá sò. child.PL.DEF bathe.PERF ‘The children bathed because of Achum’s arrival.’ b. Wáámú nyīN tè kpésíNá à kā:ī. snake.DEF PURP COMP fowl.PL.DEF ASP squawk ‘The fowls are squawking because of the snake.’ Interestingly, my data suggest that there are speakers who combine àlē sòà / nyīN in a single statement. In this situation, nyīN necessarily precedes àlē sòà so that the form nyīN àlē sòà is derived. This accounts for the sameness in meaning of the three statements presented in (55). (55) a. NóótíNá àlē sòà tè bííká à kçs# ī. cold.DEF PURP COMP child.DEF ASP cough ‘It is because of the cold that the child is coughing.’ b. NóótíNá nyīN tè bííká à kç#sī. cold.DEF PURP COMP child.DEF ASP cough ‘It is because of the cold that the child is coughing.’ c. NóótíNá nyīN àlē sòà tè bííká à kçs# ī. cold.DEF PURP PURP COMP child.DEF ASP cough ‘It is because of the cold that the child is coughing.’ University of Ghana http://ugspace.ug.edu.gh 108 3.1.3.4.5 The Conditional Marker ‘dàn’ This preverbal particle is a conditional marker, and therefore occurs in the protasis28 of a complex sentence. Its presence indicates that the occurrence of an activity specified in the apodosis29 is a consequence of the activity specified in the protasis in which it occurs. Thus, the occurrence of an activity coded by the verb in the apodosis is dependent on the one coded by the verb in the protasis. Example (56) illustrates function of dàn. (56) a. Fì dàn nyà nídōā māN, yālī 30 2SG COND get man good marry ‘If you get a good man, marry him.’ b. Bà dàn yìg zúéwá, bà nāg wà kò. 3PL COND catch thief.DEF 3PL beat 3SG kill ‘If they catch the thief, they should lynch him.’ In each of the two sentences, the particle dàn occurs in the protasis31 to indicate that the action coded in the apodosis occurs as a consequence of the one expressed in the protasis. Thus, in (56a) the marriage of the addressee is dependent on her ability to find a good man, and in (56b) the lynching will occur as a consequence of the addressee’s ability to find the thief. 28 the dependent clause 29the independent clause 30 yālī is a full clause. It shares the same subject (fì) with the verb nyà of the dependent clause. 31 My data seem to suggest that syntactically the protasis always precedes the apodosis. University of Ghana http://ugspace.ug.edu.gh 109 As we are aware, it is also usually possible for such conditional markers to indicate time. Consider the examples presented in (60) below. (57) a. Tì dàn pāā kúmmú yénní, tì lì kāārī. 1PL COND reach funeral house.DEF 1PL FUT boohoo ‘When we get to the bereaved’s house, we will boohoo.’ b. Nípōk dàn tā pūūk, wà kàn Nçb# jéín. woman COND have belly 3SG NEG chew egg ‘When a woman is pregnant, she doesn’t eat egg.’ As the use of ‘when’ in the English glosses suggests, each of the clauses indicates the time that an event or activity will take place. In (57a), the boohooing will take place when the referents arrive at the bereaved’s house, and (57b) expresses the idea that a woman does not eat egg during pregnancy. The particle dàn also co-occurs with the chain bóró + à to form dàn + bóró + à in the subordinate clause when the speaker intends to make a request or give a command (in the independent clause) whose occurrence is only possible if the action expressed in the subordinate clause is on-going. Thus, (58a) indicates that Atie will be called on condition that he is fighting, and (58b) indicates that Adoli will not be beaten unless he is drinking alcohol. (58) a. Atie dàn bóró-à kpālīN, wū wà tE$ mì. Atie COND EXIST-ASP fight, call.IMP 3SG give 1SG ‘If Atie is fighting, call him for him.’ b. Adoli dàn bóró-à nyū dáám, nāg wà. Adoli COND EXIST-ASP drink alcohol beat.IMP 3SG ‘If Adoli is drinking alcohol, beat him.’ University of Ghana http://ugspace.ug.edu.gh 110 The conditional clause is negated with kàn, irrespective of the type of verb. This is because the conditional marker dàn co-occurs with only kàn which usually intervenes between the conditional marker and the verb. Example (59) illustrates this. (59) a. Kòòdùkú dàn kàn bāāsā, kán dā! banana.DEF COND NEG be.soft NEG.IMP buy ‘If the banana isn’t soft, don’t buy it!’ b. Gmórúkú dàn kàn nī yá, n $ kàn bōrī zááNá. rain.DEF COND NEG rain ASS, 1SG NEG.FUT sow millet.DEF ‘If it doesn’t rain, I will not sow the millet.’ Each of the examples shows that dàn co-occurs with the negation marker kàn. (59a) expresses the idea that the banana should not be bought unless it is soft, and (59b) expresses the idea that the millet cannot be sowed unless it rains. 3.1.3.4.6 The Repetitive Marker ‘pīlīm’ The repetitive marker pīlīm can be taken to mean ‘again’ in English, since a speaker uses it to describe the recurrence/repetition of an activity. It can co-occur with different tense/aspect and negation markers, whose presence influences its syntactic distribution. Its usage is explained with the sentences in example (60). (60) a. Amusa lì pīlīm jām. Amusa FUT REP come ‘Amusa will come again.’ b. Nícháánōā pìlìm dE$-E$ mā. visitor.DEF REP eat.PERF-FOC AFF ‘The visitor has eaten again.’ c. Nì pīlīm gōg. 2PL REP dance.IMP ‘(You) dance again!’ University of Ghana http://ugspace.ug.edu.gh 111 The use of pīlīm and the negation marker kàn in the indicative requires the presence of particle à. This is not the same for àn which does not require the particle á, but instead requires the assertive marker yá when there is no object complement. Consider the following examples. (61) a. Amusa kàn pīlīm á jām. Amusa NEG.FUT REP PRT come ‘Amusa will not come again.’ b. *Amusa kàn pīlīm jām. Amusa NEG.FUT REP come c. Amusa àn pīlīm jàm yá. Amusa NEG REP come.PERF ASS ‘Amusa didn’t come again.’ 3.1.3.4.7 The Positive Modal ‘zāā’ The particle zāā is a modal marker which is used in the affirmative to express certainty about the occurrence of an event or the presence of a state expressed in an embedded clause. Etymologically, zāā is composed of the verb zā ‘stand’ and the grammaticalised verb āyín ‘say’ which is rendered as the clitic ā onto zā. As we are aware, apart from undergoing grammaticalisation, the verb ‘say’ in most languages is capable of hosting complement clauses. I use (62) to explain my claim that āyín ‘say’ has undergone grammaticalisation to become a complementizer in Buli. In (62a) āyín occurs as a full verb, but in (62b) it occurs as a complementizer. University of Ghana http://ugspace.ug.edu.gh 112 (62) a. Wà āyín wà tā-ā yám. 3SG say 3SG have-FOC sense ‘He says he is wise.’ b. Wà à po#li # (āyín) wà tā-ā yám. 3SG HAB think COMP 3SG have-FOC sense ‘He thinks he is wise.’ Since āyín is a full verb in (62a), it cannot be omitted. In (62b), however, it can be omitted because it occurs as a grammaticalized particle which functions as a complementizer. Having established that the verb āyín can function as a complementizer, we show with examples that it fuses with the verb zā ‘stand’ to yield the positive modal marker zāā. Example (63) illustrates the use of zāā. (63) a. Wà zāā wà dá ká lógrī. 3SG MOD.POS 3SG buy FOC car ‘He shall (certainly) buy a car.’ b. Wà zāā wà nyū-ū dáám. 3SG MOD.POS 3SG drink-FOC alcohol ‘He shall (certainly) drink alcohol.’ c. Kpārōāwá zāā wà nūm ká záá. farmer.DEF MOD.POS 3SG grind FOC millet ‘The farmer shall (certainly) grind millet.’ The use of “shall” in the English glosses is an indication of a strong assertion. It should be observed that the full form of zāā (i.e. zā + āyín) rarely occurs in speech. This is to say that the constructions in (64) and other such constructions rarely occur in the speech of native speakers. University of Ghana http://ugspace.ug.edu.gh 113 (64) a. Wà zā āyín wà dá ká lógrī. 3SG stand COMP 3SG buy FOC car ‘He shall (certainly) buy a car.’ b. Wà zā āyín wà nyú ká dáám. 3SG stand COMP 3SG drink FOC alcohol ‘He shall (certainly) drink alcohol.’ c. Kpārwá zā āyín wà nūm ká záá. farmer.DEF stand COMP 3SG grind FOC millet ‘The farmer shall (certainly) grind millet.’ 3.1.3.4.8 The Negative Modal Marker ‘kāā’ This modal marker is antonymous with zāā; while zāā expresses a strong assertion that an action will certainly take place, kāā expresses strong doubt over the occurrence of an action. The internal make-up of zāā (presented in Section 3.1.3.4.7) gives us some reason to believe that kāā is composed of the negation marker kàn ‘stand’ and the grammaticalised verb āyín ‘say’. It should also be observed that the modal kāā is different from káá32 which is derived from the fusion of the negative marker kán and the modal á (i.e. kán + á). I illustrate the use of kāā in (65). (65) a. Avūūk kāā wà dE# nààm. Avūūk MOD.NEG 3SG eat chief ‘Avūūk isn’t going to become a chief.’ b. Akūm kāā wà chēN tálīm. Akūm MOD.NEG 3SG go farm ‘Akūm isn’t going to go to the farm.’ 32 káá is used in the imperative to admonish an addressee to stop performing an action.. University of Ghana http://ugspace.ug.edu.gh 114 A speaker uses the constructions in (65) and other such constructions to expresses doubt about the occurrence of an event. In (65a) the speaker is doubtful that the referent (Avūūk) is going to become a chief, and in (65b) the speaker is doubtful that the referent (Akūm) will go to the farm. It should be indicated here also that, like the full form of zāā (i.e. zā + āyín), the full form of kāā (i.e. kàn + āyín) rarely occurs in speech. It is, thus, rare to hear the examples (65a) and (65b) uttered as (66a) and (66b) respectively. (66) a. Avūūk kàn āyín wà dE# nààm. Avūūk NEG COMP 3SG eat chief ‘Avūūk isn’t going to become a chief.’ b. Akūm kàn āyín wà chēN tálīm. Akūm NEG COMP 3SG go farm ‘Akūm isn’t going to go to the farm.’ 3.1.3.4.9 The Modals ‘fE#’ and ‘zīkā /māgsīkā’ These particles are modal markers, since they are used to express the idea that an action needs to be performed or should take place. Semantically, a subtle distinction can be drawn between them. While fE# is used to draw an addressee’s attention to the fact that the occurrence of an event is a necessity and must therefore happen, zīkā expresses the idea that the occurrence of an event is relevant. Thus, although both carry the implication that it is required that an event takes place, fE# makes a strong assertion that the occurrence of an event is a sine qua non. zīkā/māgsīkā33 occurs as an advice which indicates that it is important that an event/activity takes place. Although the advice is 33 zīkā and māgsīkā are derived from the verbs zīk and māgsī. The two verbs can roughly be translated as ‘okay/sufficient’ in English. University of Ghana http://ugspace.ug.edu.gh 115 given to show that an event/activity is essential for a desired result to be achieved, the performance of the activity is not viewed as obligatory. Another feature common to the two particles is that both of them host complement clauses. Also, both particles can only take the expletive kV $ ‘it’ which must be followed by the complementizer à34. The use of the two particles are explained further with example (67). (67) a. Kù à fE# àyīn fì nyú tììmmú. 3SG COMP MOD COMP 2SG drink medicine.DEF ‘It is a must that you drink the medicine.’ b. Kù à zīkā fì nyú tììmmú. 3SG COMP MOD 2SG drink medicine.DEF ‘It is necessary that you drink the medicine.’ In (67a) the speaker implies that the medicine must be drunk by the patient as a necessary condition for him/her to regain good health. Here, the speaker asserts that drinking the medicine is a pre-requisite for the patient’s recovery and refusal to drink might result in persistent poor health. Example (67b), on the other hand, is an advice to the patient about how relevant it is for him/her to take the medicine. zīkā may be substituted with māgsīkā in (67b), since both of them give the sentence the idea of “it’s necessary that…” The forms zīkā and māgsīkā are deducible from the verbs zīk and māgsī respectively. Both of them carry the meaning ‘It’s okay/fine’ They usually occur in constructions like KV$ māgsī ‘It’s okay’ and KV$ zīk ‘It’s okay’ as responses to a benevolent gesture 34 The particle is functionally distinct from the habitual marker. University of Ghana http://ugspace.ug.edu.gh 116 3.1.3.4.10 The Hortative Marker ‘tín’ ‘tín’ is used to exult or urge an addressee to perform an action which the speaker perceives to be against the addressee’s will. Thus, it is used to encourage an addressee to perform an activity which the addressee is reluctant to perform. The use of ‘tín’ is exemplified in (68). (68) a. Tín kàlì. HOR sit ‘Sit!’ b. Tín bàsì kūmsīNá. HOR stop cry.PL.DEF ‘Stop crying! (lit.: Stop the cry!)’ Both (68a) and (68b) are uttered by a speaker to request that an addressee performs an activity which the speaker perceives to be against the will of the addressee. The presence of ‘tín’ gives the construction the implication that the addressee should perform the activity even if he/she has no intention or desire to perform it. Thus, (68a) encourages the reluctant addressee to sit down, and (68b) entreats the addressee who is unwilling to stop crying to stop. A scenario can be created for each of the statements to make the use of ‘tín’ clearer. (68a) may be uttered to a lender (an addressee) who demands payment from a borrower at all cost and is refusing to sit down to be engaged in a peaceful dialogue. In such a situation (68a) can be uttered to the lender to ask him to relax and sit down. (68b) may also be uttered to a child who is crying endlessly over his missing money to stop crying in spite of the loss. University of Ghana http://ugspace.ug.edu.gh 117 3.1.3.4.11 The Optative Marker ‘tè (àtè)’ The particle tè has evolved as a grammaticalised form of the verb tE# ‘give’ which is interpreted as ‘let’ or ‘make’ depending on the context. A distinctive feature of this grammaticalised particle is its ability to co-occur with tE# ‘let’ from which it evolved. Statements in which both forms occur are those that either seek permission from an addressee or make a direct request or wish. Consider the use of tè in (69). (69) a. Nì tE# tè tì chēng. 2PL let OPT 1PL go ‘Let us go.’ b. TE # (tè) nààwá kāābī bç$glVk$ ú. let OPT chief.DEF sacrifice altar.DEF ‘Let the chief perform the sacrifice.’ However, it should be pointed out that it is also possible to make a request or express a wish without the optative mood marker tè. Example (70) illustrates this observation. (70) a. TE# tì chēng. let 1PL go ‘Let us go.’ b. Tì chēng. 1PL go ‘Let us go.’ Where the subject pronoun is the first person plural pronoun tì as presented in (70), care must be taken in order that the pronoun is not misconstrued as the optative marker. University of Ghana http://ugspace.ug.edu.gh 118 3.1.3.4.12 The Eventuality Marker ‘yāā’ This particle may be glossed roughly as ‘eventually’ in English, but it is not an adverb. It is used to indicate that an action/activity long expected to have taken place eventually happened. The particle also emphasizes a particular length of time when reporting past events, so it is often employed in narratives to specify a time being referred to or spoken about. Consider the examples in (71). (71) a. Mí pówā yāā kā kìkìrúk. 1SG wife EVT to.be witch ‘My wife has (eventually) become a beast.’ b. NāNchāāmá yāā kā dʋ #a#ba#. rival.PL EVT to.be friend.PL ‘The rivals have (eventually) become friends.’ c. Kpālūngkú yāā lì pāā kpīāká yénní lá, … hawk.DEF EVT TM reach hen.DEF house.DEF FOC, … ‘When the hawk (eventually) arrived at the hen’s house, …’ Examples (71a) and (71b) report a change in the states of two affairs. In the former, the speaker reports that his wife (who was probably very modest) has changed completely, by comparing his wife to a beast metaphorically. In (71b) the speaker reports that two rivals have finally become friends. The use of yāā in both examples indicates that a long existing situation or state has changed over time. Example (71c) is an excerpt of a narration in which the particle yāā has been used to show that some events had unfolded prior to the hawk’s arrival at the hen’s house. University of Ghana http://ugspace.ug.edu.gh 119 3.2 Chapter Summary I have shown in this chapter that although verbal affixation is generally minimal in Buli, a dichotomy can be drawn between stative and dynamic verbs morphologically. The chapter has also revealed that grammatical functions such as tense, aspect, modality, conditionality and polarity. Some of these verbal particles have been identified, and their functions have been discussed. University of Ghana http://ugspace.ug.edu.gh 120 CHAPTER FOUR VERBAL PARTICLES IN SERIAL VERB CONSTRUCTIONS 4.0 Introduction The works of Bodomo (1997), Ameka (2006; 2013), Osam (2004) and Lee (2003) provide evidence that SVCs are present in Dagaare, Ewe, Akan and Buli respectively, as well as other Niger-Congo languages. We also saw in Section 2.1 that Lee’s (2003) work on Buli SVCs presents a brief discussion of the general properties of Buli SVCs, including their headship, and transitivity restriction. This chapter mainly aims at revealing how grammatical categories like tense, aspect, and polarity operate within SVCs. In order to provide a good basis for the discussion and to show how the feature SVCs has evolved over the years, I discuss some issues that have characterized the use of the concept in linguistics literature. 4.1 Serial Verb Constructions (SVCs) SVCs have received much attention in the linguistics literature, after Stewart (1963) applied it to Akan. An SVC may be defined as a clause that has two or more verbs that encode a single event. In SVCs, verbs are concatenated in a clause without any overt marking of coordination or other linking expressions. This explains Foley and Olson’s (1985:18) observation that while serialized verbs may require the same actor for both predicates, each verb in the series may have arguments not shared by other verbs. University of Ghana http://ugspace.ug.edu.gh 121 Contrary to claims that serializing languages lack three-place predicates, Buli possesses three-place predicates such as sā ‘smear’, sāk ‘teach’, dāk ‘show’, tE# ‘give’, and tōm ‘send’. Consider the following examples: (1) a. Bà tE$ tí nyíám 3PL give.PERF 1PL water.DEF ‘They have given us water.’ b. *Bà tE$ tí 3PL give.PERF 1PL c. Wà sà bìsìNá kpáám. 3SG smear.PERF child.PL.DEF pomade ‘She has smeared the children pomade.’ d. *Wà sà bìsìNá 3SG smear.PERF child.PL.DEF As example (1) illustrates, the verbs tE$ ‘give’ and sà ‘smear’ are three-place predicates, which is why tE$ ‘give’ hosts the arguments bà ‘they’, tí ‘us’ and nyíám ‘water’ in (1a) and sà ‘smear’ hosts the arguments wà ‘she’, bìsìNá ‘the children’ and kpáám ‘pomade’ in (1c). The ungrammaticality of (1b) and (1d) shows that each of the two verbs necessarily requires a third argument. Of course, earlier studies have also provided evidence that three-place predicates exist in Ewe, Likpe (Sɛkpɛlé) and Akan35 35 Each of these three Kwa languages is a serializing language. University of Ghana http://ugspace.ug.edu.gh 122 (cf. Ameka 2013). The phenomenon exists in most other Gur languages, including Dagaare and Dagbani. The verbs jù ‘burn’ and gèb ‘cut’ are not three-place predicates, since, as illustrated with (2b) and (2d), they do not obligatorily require a third argument. The forms àlè bòlùmmú ‘with the fire’ and àlè ge#bīk ‘with a knife’ in (2b) and (2d) respectively are adjuncts which do not bear theta roles. (2) a. Bà jù zu#e#wá àlè bòlùmmú. 3PL burn.PERF thief.DEF with fire.DEF ‘They have burned the thief with the fire.’ b. Bà jù zu#ew# á. 3PL burn.PERF thief.DEF ‘They have burned the thief c. Mì gèb bòrùbòrùkú àlè ge#bīk. 2PL cut.PERF bread.DEF with knife ‘I have cut the bread with a knife.’ d. Mì gèb bòrùbòrùkú 2PL cut.PERF bread.DEF ‘I have cut the bread.’ University of Ghana http://ugspace.ug.edu.gh 123 4.1.1 Defining Serial Verb Constructions (SVCs) Although SVCs are often viewed as a cross-linguistic concept, it seems impractical to provide a universal definition for it. This is because languages, no matter how closely- related they are, usually possess varying morpho-syntactic categories and constructions. A couple of definitions by different scholars, however, will suffice here. According to Aikhenvald (2006), a serial verb construction (SVC) is a sequence of verbs which act together as a single predicate, without any overt marker of coordination, subordination, or syntactic dependency of any other sort. Serial verb constructions describe what is conceptualized as a single event. They are monoclausal; their intonational properties are the same as those of a monoverbal clause, and they have just one tense, aspect, and polarity value. SVCs may also share core and other arguments. Each component of an SVC must be able to occur on its own. Within an SVC, the individual verbs may have same, or different, transitivity values. (Aikhenvald 2006:1) Buli SVCs exhibit some of the properties identified by Aikhenvald (2006). I use the examples below to illustrate the presence of some of these properties. (3) a. Apolinya dà Nàndììntà dE $ yá. Apolinya buy.PERF food eat.PERF ASS ‘Apolinya has bought food and eaten it.’ b. Baaba lì pūūsī wén tE$ tì. Baaba FUT pray God give 1PL ‘Baaba will pray for us.’ University of Ghana http://ugspace.ug.edu.gh 124 Each of the constructions presented in (3a) and (3c) has two verbs: while (3a) has the verbs dà ‘buy’ and dE$ ‘eat’, and (3c) has the verbs pūūsī ‘pray’ and tE$ ‘give’. These verbs, as we see from the examples, form a single predicate which is not joined by coordinators or subordinators. Like verbs in other Gur languages, most Buli verbs are ambitransitive36. This makes it likely for verbs within Buli SVCs to either possess the same or different transitivity values. While the verb dà ‘buy’ in (3a) is a prototypical transitive verb, dE$37 ‘eat’ is a prototypical intransitive verb. Also, we see in (3b) that marking futurity in Buli SVCs requires the future tense marker lì to manifest once, so that the scope of the future marker extends to cover all other verbs within the construction. All these are in conformity with the properties that Aikhenvald (2006) attributes to SVCs. Nonetheless, we do not claim that Buli SVCs exhibit all the properties he provides. Against the backdrop of the preceding discussion, I assert that the definable properties often proposed for SVCs are not applicable to all serializing languages in equal measure; there are prototypes in SVCs that make it possible for constructions that do not possess identical properties in all regard to be considered as SVCs. There is 36 ambitransitive verbs can be used both transitively and intransitively. 37 dE$ ‘eat’ may be used transitively, when a speaker wants to specify the entity that is eaten. University of Ghana http://ugspace.ug.edu.gh 125 evidence to prove that all the features that were initially ascribed to Akan SVCs, do not apply entirely. Osam’s (2004) revelation that some sub-dialects within the Fante dialect of Akan do not always exhibit uniformity in the marking of tense on serialized verbs is a case in point. Using causative constructions, Osam (2004) also proposes that it is possible for serialized verbs to possess different subjects. Here, he treats the NP following V1 (the causative verb) as the verb’s object and as the subject of the V2.38 It is as a result of the disparities that characterize discussions on SVCs that Osam (2004:32) notes that ‘attempting to define the concept in a uniform manner is just like attempting to define the impossible.’ With the hope of finding a precise description for the concept, Haspelmath (2016) reviews some earlier works on SVCs. In the work, he suggests that earlier scholars (e.g. Foley & Olson 1985, Aikhenvald 2006) did not approach the concept SVCs from the right angle. He explains that earlier scholars’ attempts to treat SVCs as a universal concept by providing ‘a fit-all languages’ definition for the concept and prescribing universal properties that SVCs in all serializing languages must possess cross- linguistically is a misleading and gainless task. He adds that it is such a tedious, yet wasteful task, that has accounted for the broad definitions ascribed to the concept SVC whose scope is broadened whenever it is extended to a new language. 38 Osam (2004) treats causative constructions that exhibit the patterns of SVCs under what he terms as Object-Sharing (Switch Subject) University of Ghana http://ugspace.ug.edu.gh 126 Haspelmath (2016) further draws an analogy between Vulpes Vulpes (red fox) and SVCs, indicating that the former are natural kinds which possess identical features globally and can be described using the same universal features, but the latter is not an “innate category of universal grammar that can manifest itself in any language, and should manifest itself in (more or less) the same way in each language.” (pg. 293) Consequently, he proposes that SVCs must not be viewed as a universal category as earlier linguists have done, since different languages possess different grammatical categories and constructions. This, according to him, will not only save syntacticians from the task of combing through the grammar of languages with the hope of finding universal properties that are applicable to all serializing languages, but also grant them the laxity of professing language-specific definitions to allow for interesting comparisons between SVCs of different languages. Haspelmath (2016) proposes the following definition for SVCs and provides ten (10) generalizations which he claims hold for serializing languages. I adopt his definition and generalizations here not only because his work is more current, but also because he summarizes the issues that often come up in the discussion of SVCs: A serial verb construction is a monoclausal construction consisting of multiple independent verbs with no element linking them and with no predicate argument relation between the verbs. (Haspelmath 2016:296) University of Ghana http://ugspace.ug.edu.gh 127 The 10-point generations which he claims are applicable to SVCs are paraphrased as follows: 1. the verbs in all SVCs possess identical tense value. 2. the verbs in all SVCs possess the same mood value. 3. the verbs in SVCs have the same temporal or event-locational modifiers. 4. all SVCs have a single intonation contour, just like single-verb clauses 5. the order of occurrence of verbs within SVCs that express a cause- relationship is fixed and tense-iconic – cause verb precedes effect verb. 6. where only a single person, tense, mood or negation marker occurs within an SVC, the marker occurs in a peripheral position, either before the first verb or after the final verb. 7. all the verbs within an SVC share at least one argument. 8. all serializing languages exhibit at least the same-subject feature, aside other possible constructions. 9. the second verb within different-subject SVCs is always intransitive. 10. it is impossible for an SVC to have two different agents. It can be inferred from these generalizations that they make room for constructions that hitherto fell outside SVCs (if earlier descriptions of the concept are used as a University of Ghana http://ugspace.ug.edu.gh 128 benchmark), to be considered as SVCs. For instance, the example in (3b) falls within the characteristic of SVCs presented in Point six (6) of Haspelmath’s (2016) generalizations. Thus, in (3c) the future tense marker lì manifests once (i.e. before the first verb pūūsī ‘pray’) and has scope over the second verb tE$ ‘give’. In this chapter, I describe Buli SVCs, focusing mainly on how verbal particles, especifically tense, aspect, and polarity markers are distributed within such SVCs. My discussion here will be comparative in nature, as data will be drawn from other better studied languages in order to show how Buli relates with such languages with regard to SVCs. 4.1.2 Marking Tense, Aspect, and Polarity (TAP) in SVCs Cross-linguistic studies on SVCs have sought to find out how grammatical categories are marked on serialized verbs. Some of such studies (e.g. Bodomo 1997; Osam 2004; Campbell 2017) have shown that morphological marking of tense and aspect in serial verbs is characterized by uniformity, since the same tense or aspect marker is marked overtly on each of the verbs. For instance, Bodomo (1997) asserts that all the verbs in SVCs of Mabia languages are marked for the same TAMP, presenting a general constraint for all Mabia SVCs: For any construction C to qualify as an SVC, all the different verbs in C must be in the scope of one TAMP node. (Bodomo 1997:111) University of Ghana http://ugspace.ug.edu.gh 129 This assertion implies that a single grammatical category operates within a serial verb construction, such that an identical tense, aspectual, modality or polarity marker either has scope over all the other verbs within the SVC or is attached to each of the verbs within the SVC. The examples in (4) are used to illustrate this: (4) a. Ba zo-ro gE-rE di-re la bondiri. they run-imp go-imp eat-imp fact food ‘They run there and eat food (repeatedly) .’ (Bodomo 1997:112) b. Ama á-tç$ m$pàbòá á-kyE$ Fiifi. Ama PERF-buy shoes PERF-give Fiifi ‘Ama has bought shoes for Fiifi.’ (Osam 2004:38) [ c. Gbèé-!E @ hu)@ é-ku $) é-sE$E$ é-bà. dog-DEF too 3SG.PERF-break 3SG-back 3SG.PERF-come ‘The dog too has returned.’ (Campbell 2017:396) We can infer from the examples that, with regard to tense/aspect marking, Dagaare, Akan and Ga exhibit some uniformity, since identical tense/aspectual categories are marked on each of the verbs within the SVC. In (4a) each of the three verbs zo ‘run’, gE ‘go’ and di ‘eat’ are marked for imperfective by the aspectual suffix {-re} whose vowel undergoes tongue root modification to be in harmony with the vowel of the root word. In (4b) each of the two verbs tç$ ‘buy’ and kyE$ ‘give’ are marked for University of Ghana http://ugspace.ug.edu.gh 130 perfective with the marker á on the two verbs tç $ ‘buy’ and kyE$ ‘give’. Similarly, in (4c), the uniformity in the marking of the perfectective aspect is shown by tone. In spite of the seemingly prevalent uniformity that SVCs in several languages exhibit, its application is neither rigid nor applicable to all languages. Osam’s (2004) observation that Akan permits a blend of the perfective and the progressive aspects corroborates my claim. It is shown in example (5) that a blend of different functional categories is permissible in Akan. As the English glosses show, in both (5a) and (5b) there is a mixture of the perfective and the progressive aspects. This is indicated with the different aspectual markers on the verbs. (5) a. Abam é-yì nám @ nó ró-wé. Abam PERF-take meat DEF PROG-chew ‘Abam has taken the fish and is eating it.’ b. Esi á-hòr $ ènt$ ùhó nó rè- hátá. Esi PERF-wash towel DEF PROG-hang ‘Esi has washed the towel and is hanging.’ (Osam 2004:39) The discussion in the next sections will reveal that this mixture of different aspects is possible in Buli. University of Ghana http://ugspace.ug.edu.gh 131 4.1.2.1 Tense, Time and Aspectual Markers in Buli SVCs 4.1.2.1.1 The Future Tense Marker ‘àlì/lì’ in SVCs In affirmative future SVCs, the future tense marker lì occurs once (before the first verb), and it has scope over all the verbs within the clause. I use the examples presented in (6) to explain this: (6) a. Bà lì pā bàsànī dà. 3PL FUT take bitch.DEF sell ‘They will sell the bitch.’ b. Kàrìmō:rīká lì yītī zà:nì kààrì. muslim.DEF FUT rise stand scream ‘The muslim will rise and scream.’ We can infer from each of the examples in (7) that the future tense affirmative marker lì must occur once within the serial verb construction. Also, we see that the tense marker must precede only the first verb. Each of the constructions presented in (7) is ungrammatical, because the future tense marker lì precedes each of the verbs within the construction. Thus, in (7a) lì precedes each of the two verbs (pā ‘take’ and dà ‘buy’), and in (7b) it precedes each of the three verbs (yītī ‘rise’, zà:nì ‘stand’, and kà:rì ‘shout’) University of Ghana http://ugspace.ug.edu.gh 132 (7) a. *Bà lì pā bàsànī lì dà. 3PL FUT take bitch.DEF FUT sell b.* Kàrìmō:rīká lì yītī lì zà:nì lì kààrì. muslim.DEF FUT rise FUT stand FUT scream It is equally ungrammatical for the future tense marker to occur before any other verb, and not the first verb, when it manifests itself once within the SVC. This accounts for the ungrammaticality of (8b) and (8c) where the future tense marker does not precede the first verb but precedes the second and the third verbs respectively. (8) a. Adjong lì dE# tàlì tE# bìsìNá. Adjong FUT eat leave give child.PL.DEF ‘Adjong will give the children some leftovers.’ b. *Adjong dE# lì tàlì tE# bìsìNá. Adjong eat FUT leave give child.PL.DEF c. *Adjong dE# tàlì lì tE# bìsìNá. Adjong eat leave FUT give child.PL.DEF 4.1.2.1.2 The Distant Past Marker ‘jàm’ in SVCs Like the future tense marker, the distant past particle jàm occurs once within an SVC. The particle, which indicates that a situation that once existed has changed over time, occurs immediately before the first verb and not any other verb. However, its occurrence before the first verb implies that the event coded by the concatenated verbs no longer exists. I use the examples in (9) to explain this. University of Ghana http://ugspace.ug.edu.gh 133 (9) a. Wà jàm tá wà pōwá á nāg kámā. 3SG PST have 3SG wife.DEF PRT beat AFF ‘He used to beat his wife.’ b. Bà jàm yìg mè nàg kámā. 3PL PST hold 3SG beat AFF ‘They once beat me.’ (lit.: They once caught me and beat me.) In each of the two examples, the distant past marker jàm occurs once, and it precedes the first verb; in (9a) it precedes tá ‘have’, and in (9b) it precedes yíg ‘hold’. As illustrated in (10a) below, ungrammatical constructions are yielded when jàm manifests more than once or occurs before each of the verbs. It is also ungrammatical for jàm to occur before the second verb and not the first (as shown in 10b). (10) a. *Wà jàm tā wà pōwá jàm nāg kámā. 3SG PST have 3SG wife.DEF PST beat AFF b. *Bà yìg mí jàm nàg kámā. 3PL hold 1SG PST beat AFF 4.1.2.1.3 The Progressive Marker ‘bóràà’ in SVCs The data provide evidence of mixed tense/aspect between the perfective aspect (marked with a low tone on the first verb) and the progressive aspect. This finding makes Bodomo’s (1997) generalization that serialized verbs in Mabia languages are marked for the same TAMP doubtful. Consider the examples presented in (11) below: University of Ghana http://ugspace.ug.edu.gh 134 (11) a. Mí pà lāmmú bóràà Nç#b. 1SG take.PERF meat.DEF PROG chew ‘I have taken the meat, and I am chewing it.’ b. Abili dà bā:sīkū:lī bóràà dām. Abili buy.PERF bicycle PROG ride ‘Abili has bought a bicycle and is riding it.’ From (11), we see that Buli permits a mixture of the perfective and the progressive aspects within an SVC. In Buli, it is uncommon to have uniformity in the marking of TAMP on serialized verbs. This non-uniformity is evident in the imperative (discussed in Section 4.1.2.1.4) where the verbs within the SVC bear different tones.39 An important observation must be made about the progressive aspect in SVCs. It is possible to present each of the events coded by the verbs as ongoing. Here, the progressive marker bóràà is split such that the existential bóró and the apectual precede the first verb and the second verb respectively. I use the examples in (12) to illustrate this. (12) a. Atubalie bóró dìg jèntà à tE# wà chōlā. Atubalie EXIST cook soup ASP give 3SG husband ‘Atubalie is cooking soup for her husband.’ b. Avuuk bóró dāg tāngbāní à sāk núrmá. Avuuk EXIST show shrine.DEF ASP teach person.PL.DEF ‘Avuuk is showing the people the shrine.’ 39 In the imperative SVC, the first verb takes mid tone, but the second verb takes a low tone (Akanlig-Pare 2005). University of Ghana http://ugspace.ug.edu.gh 135 These examples support my claim that the progressive marker bóràà (bóró + à) can be split between the two verbs when the event coded by the SVC is ongoing. In (12a) bóró precedes V1 (dìg ‘cook’) and à precedes V2 (tE )# ‘give’).40 This is also seen in (12b) where the first verb dàg ‘show’ is preceded by bóró and the second verb sāk ‘teach’ is preceded by à. In SVCs that contain more than two verbs, the chain bóró+à is split between either the first and the final verbs or the second and the final verbs, depending on the intended meaning. This is exemplified in example (13) below: (13) a. Bííká bóró yìg gbāNká chīērī à kāāsī. child.DEF EXIST grip book.DEF tear ASP spoil ‘The child is holding the book, and is tearing it into waste. b. Bííká yìg gbāNká bóró chīērī à kāāsī. child.DEF grip.PERF book.DEF EXIST tear ASP spoil ‘The child has gripped the book, and is tearing it into waste.’ The examples in (13) show that the syntactic position of bóró within an SVC has semantic implications on the sentence. Since bóró precedes the first verb in (13a), the verb has a progressive reading with the implication that the speaker caught sight of the child gripping the book and tearing it into waste, but in (13b) the process of gripping 40 V1 means first verb, V2 means second etc. University of Ghana http://ugspace.ug.edu.gh 136 the book was completed (as the first verb yìg ‘grip’ is marked for the perfective) before the speaker saw the book being torn. Notwithstanding this mixture of tense/aspect, there can be uniformity in the marking of the perfective aspect within an SVC. This is seen in the perfective where each of the verbs bears a low tone. Consider the examples provided in (14). (14) a. Wà pà lāmmú Nç$b yá. 3SG take.PERF meat.DEF chew.PERF ASS ‘He has chewed the meat.’ (lit.: He has taken the meat and chewed it.) b. Wà tùsì mí lòànsì yá. 3SG push.PERF 1SG drop.PERF ASS ‘She has pushed me down.’ Example (14) supports my observation that it is possible for each of the verbs in an SVC to mark the same functional category. Thus, in (14a) the verbs pà ‘take’ and Nç$b ‘chew’ are marked for the perfective with low tones. Similarly, the two verbs in (14b), tùsì ‘push’ and lwànsì ‘drop’ are marked for the perfective with low tones. University of Ghana http://ugspace.ug.edu.gh 137 4.1.2.1.4 The Imperative SVCs In the imperative form of an SVC, the first verb takes a mid-tone and the second takes a low tone (cf. Akanlig-Pare 2005)41. Consider the examples presented in (15): (15) a. pā tE$ take give (to give) b. yītī zà:nì rise stand (stand up) c. tūsī lwànsì push drop (push down) d. chālī jàm run come (rush in) e. ¯E # sàgì do teach (demonstrate) (Akanlig-Pare 2005:287) The examples show imperative SVCs which possess two verbs following each other. More than two verbs can occur in such SVCs, and it is possible for TAMP particles and other such functional categories to occur within imperative SVCs. The imperative marker ‘á’, for instance, may precede the second verb when the speaker wants to command the addressee to perform an activity. I use the examples in (16) to explain this further. 41 Akanlig-Pare’s (2005) discussion of SVCs is restricted to constructions that contain two verbs. University of Ghana http://ugspace.ug.edu.gh 138 (16) a. Yītī á kūlī! rise IMP go.home ‘Rise and head towards home!’ b.Nmālī sāāmú á dE#! tear tuo-zaafi.DEF IMP eat ‘Be eating the tuo-zaafi. (lit.: Be tearing the food and eating it.) It can be inferred from (16) that the imperative marker á can occur before the second verb in an imperative SVC. In (16a) á precedes the second verb kūlī ‘go home’ and in (16b) it precedes the second verb dE # ‘eat’. 4.1.2.2 Negation Marking in Buli SVCs In Section 2.1.3, I hinted that, generally, negation in Akan SVCs is marked morphologically through the affixation of the homorganic nasal prefix n– to each of the verbs. Osam (2004) provides an exception to this norm, showing that although Akan SVCs exhibit uniformity in negation marking, some sub-dialects of Fante42 may require the use of an n-kE format in the future negative, such that the negation marker is attached to only the first verb so that the scope of negation covers all the verbs in the SVC. He uses the following example to support his claim. (17) a. Mò-n $-kç$-tç @ bí à43-mà wò. 1SG SUBJ-NEG-FUT-buy some CONS-give 2SG OBJ ‘I will not buy some for you.’ 42 Fante is one of the two broad categories of Akan (the other being Twi). Its sub-dialects include Gomoa, Ekumfi, Nkusukum, and Breman (see Osam 2004) 43 Osam describes the particle à as a consecutive (CONS) which occurs mostly in affirmative constructions. University of Ghana http://ugspace.ug.edu.gh 139 (Osam 2004:20) In (17), the negation prefix {n-} is attached to only the first tç @ ‘buy’, but not to the second verb mà ‘give’. Thus, once marked on the first verb, the negative marker has scope over all other verbs that follow. In Buli negative SVCs, the negation marker precedes only the first verb immediately, and this marker has scope over all other verbs within the SVC. Consider the distribution of the perfective negative marker àn in (18). (18) a. Aboka àn yìtì zà:nì yá. Aboka NEG rise.PERF stand.PERF ASS ‘Aboka hasn’t stood up.’ b. *Aboka àn yìtí àn zà:ní yá. Aboka NEG rise.PERF NEG stand.PERF ASS c. *Aboka yìtì àn zà:ní yá. Aboka rise.PERF NEG stand.PERF ASS In (18), we see that negation is marked once, and it occurs on only the first verb which it must necessarily precede. This accounts for the ungrammaticality of (18b) and (18c). This pattern of negation is not restricted to the negative perfective; it is also seen in the future negative. Thus, in Buli future negative sentences, the negative marker kàn precedes the first verb and has scope over other verbs in the series. (19) confirms my observation: University of Ghana http://ugspace.ug.edu.gh 140 (19) a. Aboka kàn yītī á zā:nī. Aboka NEG.FUT rise PRT stand. ‘Aboka will not stand up.’ b. *Aboka kàn yītī kàn zā:nī. Aboka NEG.FUT rise NEG.FUT stand c. *Aboka yìtì kàn zā:nī. Aboka rise.PERF NEG.FUT stand Also, as illustrated in (19a), the future negative marker kàn must manifest once (and must precede only the first verb) in negative constructions. It is ungrammatical for the future negative marker to manifest more than once or precede any other verb aside the first. This accounts for the ungrammaticality of (19b) where kàn manifests before each of the two verbs, as well as (19c) where kàn manifests before the second verb zà:ní ‘stand’ instead of the first verb yìtí ‘rise’. 4.1.2.2.1 The Negative Perfective Marker ‘àn’ in SVCs We have seen in examples (18) and (19) that the negative perfective marker àn must occur before the first verb, and not the second verb. This explains why (20b) below is considered as the negative form of (20a). When more than two verbs are present in the SVC, the negation pattern does not change; the negation marker must manifest once, and it must precede the first verb. (20d), the negative form of (20c), illustrates this. University of Ghana http://ugspace.ug.edu.gh 141 (20) a. Nícháánīmá bòòrì dáámú tE$ chāāb. visitor.DEF.PL pour.PERF wine.DEF give.PERF each other ‘The visitors have shared the wine with each other.’ b. Nícháánīmá àn bòòrì dáámú tE$ chāāb. visitor.PL.DEF NEG pour.PERF wine.DEF give.PERF each other ‘The visitors haven’t shared the wine with each other.’ c. YībīsīNá kò kpīāká làb tE$ chāāb. twin.PL.DEF kill.PERF fowl.DEF share.PERF give.PERF each other ‘The twins have killed the fowl and shared it.’ d. YībīsīNá àn kò kpīāká làb tE $ chāāb. Twin.PL.DEF NEG kill.PERF fowl.DEF share.PERF give.PERF each other ‘The twins haven’t killed the fowl and shared it.’ These examples affirm that the negation marker must always precede the first verb. In (20b), the negation marker àn precedes bòòrì ‘pour’, but not tE $ ‘give’. (20c) has three verbs, namely kò ‘kill’, làb ‘share’, and tE$ ‘give’. Its negative form presented in (20d) possesses only a single negation marker àn, which occurs immediately before the first verb. It is also possible for more than three verbs to occur in an SVC, but this does not affect the pattern of negation. The SVC in (21a) has as many as five verbs (yìgì ‘catch’, kò ‘kill’, sè ‘roast’, làb ‘share’ and tE$ ‘give’), yet its negative form presented in (21b) has a single negative marker which occurs immediately before the first verb. University of Ghana http://ugspace.ug.edu.gh 142 (21) a. Tì yìgì būūkú kò sè làb 1PL catch.PERF goat.DEF kill.PERF roast.PERF share.PERF tE$ chāāb. give.PERF each other ‘We caught the goat, killed it, roasted it, and shared it amongst each other.’ b. Tì àn yìgì būūkú kò sè 1PL NEG catch.PERF goat.DEF kill.PERF roast.PERF làb tE$ chāāb. share.PERF give.PERF each other. ‘We didn’t catch the goat, kill it, roast it, and share it among each other.’ We saw in Section 4.1.2.1.3 that a mixture of the perfective and the progressive aspects is possible in Buli. The negative forms of such constructions usually require the placement of the perfective negation marker àn and the progressive marker bóràà before the first verb and the second verb respectively. As (22) below illustrates, the negative form of (22a) is (22b) which has àn preceding the first verb (pòtì ‘peel’) and the progressive marker bóràà preceding the second verb (dík ‘cook’). Where the verbs are more than two, the negation pattern does not change; àn precedes the first verb and bóràà precedes the final verb. This explains why in (22c), where a third verb is introduced, the negation marker àn precedes the first verb pòtì ‘peel’ and bóró+à precedes the final verb dík ‘cook’. University of Ghana http://ugspace.ug.edu.gh 143 (22) a. Ataale pòtì bànchìbìká44 bóràà dík. Ataale peel.PERF cassava.DEF PROG cook. ‘Ataale has peeled the cassava, and is cooking it.’ b. Ataale àn pòtì bànchìbìká bóràà dík. Ataale NEG peel.PERF cassava.DEF PROG cook ‘Ataale has not peeled the cassava, and is not cooking it.’ c. Atale àn pòtì bànchìbìká sùrì bóràà dík. Atale NEG peel.PERF cassava.DEF wash.PERF PROG cook. ‘Atale hasn’t peeled the cassava, hasn’t washed it, and isn’t cooking it.’ Where an SVC possesses three verbs, it is possible for the progressive chain bóró+à to be split between the second and the third verbs. It, however, must be pointed out that such syntactic distribution is not without semantic nuances. Consider the change in meaning of (22c) presented in (23) with bóró+à split between the second verb sùrì ‘wash’ and the third verb dík ‘cook’. (23) Atale àn pòtì bànchìbìká bóró sùrì à dík. Atale NEG peel.PERF cassava.DEF EXIST wash ASP cook. ‘Atale hasn’t peeled the cassava, and isn’t washing it for cooking.’ In (23), the splitting bóró+à between the second and the third verbs results in each of them being construed as a progressive or continuous action. On the other hand, when this splitting does not take place and bóràà precedes the third verb (as exemplified in 44 bànchìbìk ‘cassava’ is borrowed from the Akan word bànkyè which means ‘cassava’. University of Ghana http://ugspace.ug.edu.gh 144 22c above), the second verb carries a perfective reading while the third verb carries the progressive reading. Constructions like (23) usually occur in speech when the actions expressed by the second and third verbs are those that run concurrently. Example (24) is used to explain this assertion further. (24) a. Adem dà kìnkàlì bóró pìèsì à Nçb@ . Adem buy.PERF sugarcane EXIST peel ASP chew ‘Adem has bought sugarcane, and is peeling and eating it.’ b. Adem àn dà kìnkàlì bóró pìèsì à Nç@b. Adem NEG buy.PERF sugarcane EXIST peel ASP chew ‘Adem isn’t chewing sugarcane.’ (lit.: ‘Adem hasn’t bought sugarcane and isn’t peeling and eating it.’) From example (24a) we can infer that the acts of peeling and eating the sugarcane run concurrently. Thus, the referent (Adem) is eating the sugarcane while peeling it. The example also confirms that it is possible for bóró+à to be split between the second and the third verbs: bóró precedes pìèsì ‘peel’, and à precedes Nç@b ‘chew’ in both examples. Apart from showing different aspectual combinations of the perfective and the progressive, (24b) shows that negation manifests itself once within such SVCs. Thus, the negative marker àn occurs immediately before the first verb dà ‘buy’ and has scope over the verbs pìèsì ‘peel’ and Nç@b ‘chew’. University of Ghana http://ugspace.ug.edu.gh 145 4.1.2.2.2 The Negative Future Marker ‘kàn’ in SVCs The future negation marker kàn exhibits a similar syntactic behaviour in SVCs as the perfective negation marker àn; the future negation marker kàn manifests itself once (before the first verb), and does not recur before any other verb within the SVC. A unique feature of the negative future SVC is that the future negation marker kàn usually co-occurs with the particle á to create the discontinuous marker kàn… á (where the verbs are two). The position of each of the two particles is invariable; while kàn precedes the first verb, á precedes the final verb. The fact that (25b) is the negative form of (25a) corroborate my assertion. (25) a. Bà lì pā bàsànī dà. 3PL FUT take bitch.DEF sell ‘They will sell the bitch. b. Bà kàn pā bàsànī á dà. 3PL NEG.FUT take bitch.DEF PRT sell ‘They will not sell the bitch.’ c. *Bà kàn pā bàsànī dà. 3PL NEG.FUT take bitch.DEF sell University of Ghana http://ugspace.ug.edu.gh 146 (25b) is grammatical, because it has both the future negation marker kàn and á (the former preceding V1 immediately, and the latter preceding V2 immediately). The ungrammaticality of (25c) is caused by the absence of the particle á. Where the verbs in the negative future SVC are three or more, kàn and á are split between the first verb and the final verb. Here, kàn precedes the first verb immediately, and á precedes the final verb immediately. Consider the examples presented in (26). (26) a. Kàrìmō:rīká lì yītī zà:nì kààrì. muslim.DEF FUT rise stand scream ‘The muslim will rise and scream.’ b. Kàrìmō:rīká kàn yītī zà:nì á kààrì. muslim.DEF NEG.FUT rise stand PRT scream ‘The muslim will not rise and scream.’ c. Adjong lì dE# tàlì tE$ bìsìNá. Adjong FUT eat leave give child.PL.DEF ‘Adjong will give the children some leftover.’ d. Adjong kàn dE# tàlì á tE$ bìsìNá. Adjong NEG.FUT eat leave PRT give child.PL.DEF ‘Adjong will not give the children any leftover.’ As examples (26b) and (26d) reveal, marking negative future in SVCs that possess three or more verbs requires the occurrence of the future negation marker kàn and the particle á before the first and the final verbs respectively. In (26b) kàn precedes the University of Ghana http://ugspace.ug.edu.gh 147 first verb yītī ‘rise’ immediately, and á precedes the third verb kààrì ‘scream’. A similar situation pertains to (26d) where kàn precedes the first verb dE# ‘eat’ immediately and á precedes the final verb tE$ ‘give’ immediately. The co-occurrence of kàn and á in a negative future SVC also pertains to SVCs that do not have any object. Here, both verbs are intransitive. Consider example (27) below. (27) a. Súómú lì chèng bè. rabbit.DEF FUT go lost ‘The rabbit will get lost.’ b. Súómú kàn chèng á bè. rabbit.DEF NEG.FUT go PRT lost ‘The rabbit will not get lost.’ c. *Súómú kàn chèng bè. rabbit.DEF NEG.FUT go lost Drawing on the explanation I provided to (25), it can be deduced that the ungrammaticality of (27c) is as a result of the absence of the particle á. 4.1.2.2.3 The Negative Imperative SVC In Section 4.1.2.2.2 we saw that, in the indicative, the two particles kàn and á occur in future negative SVCs. It was also shown that while the future negation marker kàn precedes the first verb immediately, á precedes the second verb (where the clause University of Ghana http://ugspace.ug.edu.gh 148 possesses two serialized verbs) or the final verb (where more than two serialized verbs are present). In the negative imperative, kán does not co-occur with á. Like other preverbal particles, it manifests once and precedes only the first verb. As expected, kán has scope over any other verbs that follows the first. Consider the imperative construction provided in (28a) and its negative form in (28b). (28) a. DE# tàlì tE$ bìsìNá! eat leave give child.PL.DEF ‘Give the children some leftover!’ b. Kán dE$ tàlì tE$ bìsìNá! NEG.IMP eat leave give child.PL.DEF ‘Don’t give the children any leftover!’ c. *Kán dE $ tàlì kán tE$ bìsìNá! NEG.IMP eat leave NEG give child.PL.DEF d. *DE# tàlì kán tE $ bìsìNá! eat leave NEG give child.PL.DEF From (29b), we see that the negative imperative particle kán occurs before only the first verb of the construction and has scope over those that follow it. Thus, kán precedes the first verb dE $ ‘eat’. It is ungrammatical for it to manifest twice or occur before each of the verbs (as illustrated in 28c) or precede only the second verb (as shown in 28d). University of Ghana http://ugspace.ug.edu.gh 149 In the negative imperative, the particle á never co-occurs with kán which precedes V1 at sentence initial position. I exemplify this, using (29) and (30) below. (29) a. Kán tā bííká kùlì. NEG.IMP take child.DEF go.home ‘Don’t take the child home!’ b. Kán tE# bííká chèng bè. NEG.IMP let child.DEF go lost ‘Don’t let the child get lost!’ c. Kán chèN bè. NEG.IMP go lost ‘Don’t get lost!’ Each of the constructions in (29) is grammatical, since kán does not co-occur with á. If the particle á is introduced in each of the constructions, they become ungrammatical. (30) illustrates my observation. (30) a. *Kán tà bííká á kùlì. NEG.IMP take child.DEF IMP go.home b. *Kán tE$ bííká á chèng bè. NEG.IMP let child.DEF IMP go lost c. *Kán chèN á bè. NEG.IMP go IMP lost University of Ghana http://ugspace.ug.edu.gh 150 The particle á can co-occur with the negative imperative marker kán. This happens when the speaker is commanding an addressee to halt an action that is being performed by the latter. Also, the combination results in tonal modification which causes the low tone of the progressive marker to become high. Consider the examples provided in (31). (31) a. Kán yìk mííká á dālī. NEG.IMP hold child.DEF PRT pull ‘Don’t be pulling the rope.’ b. Kán pà fíNká á nāg píísíNá. NEG.IMP take cane.DEF PRT hit sheep.PL ‘Don’t be caning the sheep.’ c. Kán tà gbāNká chīērī á kāāsī. NEG.IMP hold book.DEF tear PRT spoil ‘Don’t be tearing the book into waste.’ In each of the examples, kán occupies a sentence-initial position (immediately before V1), and á precedes the second verb when there are two verbs within the clause. Where there are more than two verbs (as exemplified in 31c), á precedes the final verb. 4.2 Chapter Summary The discussion in this chapter focused on the distribution of verbal particles that mark in SVCs. It has been shown that such verbal particles typically manifest once in SVCs. They precede the first and have scope over any other verb that follows the first verb. University of Ghana http://ugspace.ug.edu.gh 151 Nonetheless, depending on a speaker’s intended meaning, the existential marker bóró and the aspectual marker à, which form the progressive marker bóràà, may be shared between serialized verbs. University of Ghana http://ugspace.ug.edu.gh 152 CHAPTER FIVE SUMMARY, FINDINGS AND CONCLUSION 5.0 Introduction In Chapters 3 and 4, I discussed the morphology of the Buli verb, the grammatical functions of verbal particles, and the distribution of such verbal particles in SVCs. This chapter basically presents a synopsis of the entire study, focusing on the findings of the study. Based on my findings, I advance some recommendations for future studies before I conclude the study. 5.1 Summary and Findings In the introductory chapter, I presented a general overview of the study. I discussed some linguistic features and the genetic affinity of Buli, and provided some ethnographic information about its native speakers, commenting on their historical origin, population size, geographic location, and occupation. The chapter also presented some information the linguistics features of Buli, including its phonology, morphology, syntax and tonal properties. Other relevant information that laid the foundation for the current study, such as the research problem, objectives, previous studies and justification for the current study, as well as the relevance of this study, were also presented. The second chapter focused on three key areas: the literature review, the theoretical framework, and the methodology. Here, I reviewed literature on the verb in general, showing how functional categories like tense, aspect, modality and polarity are marked on the verb with affixes and verbal particles. To set the study in a broader context, I did University of Ghana http://ugspace.ug.edu.gh 153 not restrict the review to studies on the verb of only Gur languages; works on the verb of a few other languages to which Buli is less related (e.g. Panare, Akan, etc.) were reviewed. Most of the literature, however, were drawn from Gur languages like Dagaare, GurunE, and Dagbani which are closely related to Buli. Works on Buli, especially those that have discussed some aspects of Buli grammar were given attention. The chapter also explained some basic assumptions that underlie Functional Grammar, the theoretical framework that underpins this study. Here, my justification for the use of the theory, that functionalists view language as a tool of communication that can be effectively described in its communication context, was highlighted. Since the study is skewed towards Functional Typological Syntax (FTS), the major position of proponents of this theory, that languages should be studied cross-linguistically in order to reveal how grammatical features are spread across them was also discussed briefly. Finally, the chapter discussed the methodological procedures in the current study, showing the data source and type. Chapter 3 served as the starting point for the discussion of the data. The discussion revealed that affixation is minimal in Buli, and that grammatical functions are generally performed by verbal particles. It was also shown that a dichotomy can be drawn between dynamic and stative verbs on morphological grounds. With exemplification, I revealed that while the roots of dynamic verbs display the suffix {-i}, the roots of stative verbs display the suffix {-a}. A list of some of the verbal particles discussed and their grammatical functions has been provided in Table 5.1 below. University of Ghana http://ugspace.ug.edu.gh 154 Table 5.1: The Buli Verbal Particles and their Grammatical Functions S/N PARTICLE GRAMMATICAL FUNCTION 1 à Habitual Marker 2 bóràà Progressive Marker 3 lì/àlì Future Tense Marker 4 pòòm Past Time Marker 5 nyìèm Present Time Marker 6 jàm Distant Past Marker 7 kàn Future Negative Marker 8 àn Perfective Negative Marker 9 yá Assertive Marker 10 lá Emphatic Marker 11 kámā Affirmative Marker 12 àlē sòà / nyīN Purposive Marker 13 dàn Conditional Marker 14 pīlīm Repetitive Marker 15 zāā Positive Modal Marker 16 kāā Negative Modal Marker 17 fE # Modal Marker 18 zīkā /māgsīkā Modal Marker 19 tín Hortative Marker 20 ‘tè (àtè) Optative Marker 21 yāā Eventuality Marker University of Ghana http://ugspace.ug.edu.gh 155 In the fourth chapter, I discussed the distribution of some of these verbal particles in Serial Verb Constructions (SVCs). Attention was given to the distribution of Tense, Aspect, and Polarity particles in SVCs, as this has been a topical issue in the linguistics literature on SVCs. Five general observations can be made about the findings in the fourth chapter. 1. Verbal particles generally manifest once within an SVC. For instance, the negation markers kàn and àn occur once within the serial verb construction. This is applicable to the future tense marker lì/àlì. 2. Within an SVC, verbal particles necessarily precede the first verb and have scope over any other verb that follows the first verb. 3. A mixture of different aspects in Buli SVCs is permissible. The discussion revealed that at least the perfective and the progressive aspects may be marked in a single SVC. 4. The fourth chapter has also shown that the presence of the negative future marker kàn before the first verb necessitates the presence of the particle á before the second verb. This is not the case for the imperative which has only the negative imperative marker kán occurring before the first verb. 5. Again, the fourth chapter revealed that the progressive chain bóró+à may be split between the verbs in an SVC, depending on the speaker’s intended meaning. University of Ghana http://ugspace.ug.edu.gh 156 5.2 Recommendations This work serves as a starting pointing for further investigations into the verbal system of Buli. Based on the findings and the scope of this study, I provide some recommendations for future studies. 1. I do not claim that the list of verbal particles provided here is exhaustive, although about 30 verbal particles (including those that were not given unique sub-headings have been identified and discussed in this study. Therefore, there is the need for future studies to consider identifying and discussing the functions of other Buli verbal particles (if there are). Besides, the functions of some of the verbal particles remain unclear, so further studies on these verbal particles will help achieve much clarity. 2. Although studies on SVCs have mainly focused on discussing the syntactic and semantic types of SVCs in different languages, this study did not investigate the types of Buli SVCs. Hence, future studies on Buli SVCs should explore the types of SVCs in Buli, in order to provide some basis for typological studies on SVCs involving Buli. 3. The verb is the most complex element of the clause and any attempt to discuss it leads to the discussion of the entire clause. This is because the meaning and syntactic behaviour determines the presence of certain arguments and the interpretations of such arguments (cf. Levin 1993). There is the need for future studies to investigate other features of the Buli verb (especially its types), since some of such features determine the presence of certain verb particles. University of Ghana http://ugspace.ug.edu.gh 157 4. In Buli, it is rare to find a large number of verbal particles in a single clause. Nonetheless, future studies may investigate the sequential arrangements of the few verbal particles that may co-occur in a clause, to show how they relate to the verb and to each other within the clause. 5.3 Chapter Summary This chapter has presented a chapter-by-chapter summary of each chapter, in order to reveal the subject matter of the study. It has also highlighted the findings of the study and proposed some recommendations for future studies. University of Ghana http://ugspace.ug.edu.gh 158 REFERENCES Aikhenvald, Alexandra. 2006. Serial verb constructions in typological perspective. In Aikhenvald, Alexandra and Dixon, R.M.W. (eds.), Serial verb constructions: A cross-linguistic typology, 1-68. Oxford: Oxford University Press. Akanlig-Pare, George. 1988. Towards a phonological study of tone in the Wiani dialect of Buli. Department of Linguistics, University of Ghana. Akanlig-Pare, George. 1994. Aspects of Buli phonology, Department of Linguistics, University of Ghana: M.Phil Thesis. Akanlig-Pare, George. 1996. The Tonal structure of phonological nouns in Buli. Legon: Department of Linguistics, University of Ghana. Akanlig-Pare, George. 1997. Tonal structure of Buli phonological nouns. Gur Papers / Cahiers Voltaïques 2: 63-67. Akanlig-Pare, George. 1999. Nominal derivation in Buli. Gur Papers / Cahiers Voltaïques 4: 99-109. Akanlig-Pare, George. 2002. Vowel harmony in Buli. In New Directions in Ghanaian Linguistics, eds. Felix K. Ameka and Emmanuel Kweku Osam, 265-278. Legon: Department of Linguistics, University of Ghana. Akanlig-Pare, G. and Kenstowicz, M. 2003. Tone in Buli. In Studies in Buli Grammar, eds. Michael Kenstowicz and George Akanlig-Pare, 1- 34. (Working Papers on Endangered and Less Familiar Languages Volume 4). Cambridge: MIT. Akanlig-Pare, George. 2005. Buli Tonology: a non-linear approach. PhD Thesis, University of Ghana. University of Ghana http://ugspace.ug.edu.gh 159 Akanlig-Pare George. 2007. The Tono-syntax of Pronominal Objects in Buli. In Dakubu, M. E. K., E. K. Osam, G. Akanlig-Pare & Kofi K. Saah (eds.) Studies in the Languages of the Volta Basin 4.1:87-102, 2007 Akanlig-Pare, George and Wangara, Peter. 2007. Biamagsika Nyuenta. Legon- Trondheim Linguistics Projects, Department of Linguistics, University of Ghana. Akua, Campbell. 2017. A Grammar of Gã. PhD Thesis. Rice University. Ameka, Felix. 2006. Ewe serial verb constructions in their grammatical context. In Aikhenvald, Alexandra and Dixon, R.M.W. (eds.), Serial verb constructions: A cross-linguistic typology, 124-143. Oxford: Oxford University Press. Ameka, Felix K. 2013. Three-place predicates in West African serializing languages. Studies in African Linguistics, 42(1), 1-32. Anyanwu, Ogbanna. 2012. The syntax and semantics of inherent complement verbs in Igbo. Theory and Practice in Language Studies, 2(8), 1561. Atintono, Samuel A. 2004. A morpho-syntactic study of the Gurenɛ verb. M.Phil Thesis, University of Ghana. Atintono, Samuel A. 2005. Serial verb constructions in Gurene: an overview. Legon Journal of the Humanities, 16, 55-71. Atintono, Samuel A. 2013. The semantics and grammar of positional verbs in GurunE: a typological perspective. PhD Thesis. University of Manchester. Baker, Carl L. 1995. English syntax. 2nd Edition. Cambridge, Massachussetts: MIT Press. Bendor-Samuel, John T. 1971. Niger-Congo, Gur. In Thomas A. Sebeok (eds.), Current trends in linguistics: Linguistics in Sub-saharan Africa, vol. 7, 141- University of Ghana http://ugspace.ug.edu.gh 160 178. The Hague: Mouton.
 Bhat, Shankara D. N. 1999. The prominence of tense, aspect and mood. Amsterdam: John Benjamins. Boadi, Lawrence A. 1974. Focus-marking in Akan. Linguistics, 12(140), 5-58. Bodomo, Adams. 1993. Complex predicates and event structure: An integrated analysis of serial verb constructions in Mabia languages of West Africa. Trondheim: University of Trondheim. Bodomo, Adams. 1997. The structure of Dagaare. Stanford, CA: CSLI Publications. Bodomo, Adams. 2000. Dagaare. Munchen, Germany: Lincom Europa. Bodomo, Adams. 2004. The syntax of nominalized complex verbal inflection in Dagaare. Studia Linguistica 58.1 1-22 Butler, Christopher. S. 2008. Cognitive adequacy in structural-functional theories of language. Language Sciences, 30(1), 1-30. Bybee, J. L., Perkins, R and Pagliuca, W. 1994. The evolution of grammar: tense, aspect and modality in the languages of the world. Chicago: University of Chicago. Comrie, Bernard. 1976. Aspect. Cambridge: CUP Comrie, Bernard. 1985. Tense. Cambridge: CUP Crystal, David. 2008. A dictionary of linguistics and phonetics (The Language Library). John Wiley & Sons Incorporated. Dakubu, Mary E. Kropp. 1989. The Dagaare verbal phrase. MS, Department of Linguistics, University of Ghana, Accra.
 Dakubu, Mary E. Kropp. 2005. Collected language notes on Dagaare grammar. University of Ghana http://ugspace.ug.edu.gh 161 Accra: Institute of African Studies.
 DeLancey, Scott. 2001. On functionalism. University of Oregon, Linguistics Department. MS. http://pages.uoregon.edu/delancey/sb/fs.html Delplanque, Alain. 1998. Focalisation et identification en Dagara. Gur Papers/Cahiers Voltaiques 3. 19-26. Dik, Simon C. 1981. Functional grammar. USA: Foris Publications.
 Dimmendaal, Gerrit J. 2001. Areal diffusion and genetic inheritance: an African perspective. In Robert M. W. Dixon & Alexandra Y. Aikhenvald (eds.), Areal diffusion and genetic inheritance: problems in comparative linguistics, 358- 392. Oxford: Oxford University Press. Dixon, R. M. W. 2010. Basic linguistic theory, Vol. 1: Methodology. Oxford: OUP. Dolphyne, Florence A. 1987. On negating the consecutive verb in Akan. Journal of West African Languages, 19(1), 70-80. Dolphyne, Florence. A. 1988. The Akan (Twi-Fante) language: Its sound systems and tonal structure. Ghana Universities Press. Dong, David. A.1981. The verb phrase in Dagaare. Accra: Language Centre, University of Ghana.
 Dryer, Matthew. 1999. Are grammatical relations universal? In Bybee, Joan, Haiman, John and Thompson, Sandra (eds.) Essays on language function and language type, 116-143. Amsterdam: John Benjamins. Essegbey, James. 2010. Inherent complement verbs and the basic double object construction in Gbe. In Topics in Kwa syntax, pp.177-193, Springer, Dordrecht. Foley,W., and Olson, M. 1985. Clausehood and verb serialization. In J. Nicholas, and A. Woodbury (eds.), Grammar inside and outside and outside the clause. pp.17 - 60. Cambridge: Cambridge University Press University of Ghana http://ugspace.ug.edu.gh 162 Foley, William A. 2010. Events and serial verb constructions. Complex predicates: Cross-linguistic perspectives on event structure, ed. by Mengistu Amberber, Brett J. Baker & Mark Harvey, pp. 79–109. Cambridge: Cambridge University Press. Fusheini, A. H. 2006. Focus marking in Dagbani. Ms UBC-Canada.
 Givón, Talmy. 1984. Syntax: A functional-typological approach. Amsterdam: John Benjamins. Givón, Talmy. 2001. Syntax: A functional-typological introduction. Vol.1. Amsterdam: John Benjamins. Greenberg, Joseph H. 1963. The languages of Africa (Vol. 25). Indiana University. Halliday, Michael A. 1994. An introduction to functional grammar (Vol. 2). London: Edward Arnold. Harley, Matthew. 2008. Tense, aspect and mood in Tuwuli. In Felix Ameka & Mary Esther Kropp Dakubu (eds.), Aspect and modality in Kwa languages, 291-330. Amsterdam: John Benjamins. Haspelmath, M. (2016). The serial verb construction: Comparative concept and cross- linguistic generalizations. Language and Linguistics, 17(3), 291-319. Heine, B., & Kuteva, T. 2007. The genesis of grammar: A reconstruction (Vol. 9). Oxford University Press. Hiraiwa, K. 2003. Relativization in Buli. Studies in Buli Grammar. In Kenstowicz Micheal & Akanlig- Pare, George (eds.), Studies in Buli Gammar: Working Papers on Endangered and Less familiar Languages 4 (pp. 45-84). Massachusetts: MiTWPL. University of Ghana http://ugspace.ug.edu.gh 163 Hsiao, Franny. 2003. Reduplication in Buli. In Kenstowicz Micheal & Akanlig- Pare, George (eds.), Studies in Buli Gammar: Working Papers on Endangered and Less familiar Languages 4 (pp. 85-96). Massachusetts: MiTWPL. Hopper, Paul. 1991. On some principles of grammaticization. In Traugott, Elizabeth C. & Heine, Bernd (eds.), Approaches to grammaticalization, pp. 17-35. Amsterdam: John Benjamins. Issah, Samuel. A. 2013. The function of the post verbal la in Dagbani. Studies in African Linguistics, 42(2), 153-174 Korsah, Sampson. 2011. Transitivity and inherent complementation in Ga. MPhil thesis, University of Ghana. Kröger, Franz. 1992. Buli-English dictionary: with an introductory grammar and an index English-Buli. Lit Verlag. Lee Youngjoo. 2003. Serial verb constructions in Buli. In Studies in Buli Grammar, eds. Michael Kenstowicz and George Akanlig-Pare, 1-34. (Working Papers on Endangered and Less Familiar Languages Volume 4). Cambridge: MIT. Levin, Beth. 1993. English verb classes and alternations: A preliminary investigation. University of Chicago Press. Lewis, M. Paul (ed.). 2009. Ethnologue: Languages of the world (16 th edn.). Dallas, Tex.: SIL International. Online version: http://www.ethnologue.com/ (4 February, 2018) Lewis, M.P., Simons, G. F., & Fennig, C. D. (eds.). 2015. Ethnologue: Languages of the World (18th ed.). Dallas, Texas: SIL International. Online version: http://www.ethnologue.com. Manfredi, V. (1991). Ágbọ̀ and Ẹ́hụgbò: Ìgbo linguistic consciousness, its origins and limits. PhD. Dissertation, Harvard University
 University of Ghana http://ugspace.ug.edu.gh 164 
 Mélançon, Lucien and Prost, André. 1972. Dictionnaire Buli. Département de Linguistique Générale et de Langue Négro-Africaines de la Faculté des Lettres et Sciences Humaines de l'Université Dakar, No. 20. Nichols, Johanna. 1984. Functional theories of grammar. Annual Review of Anthropology, 13, 97-117. Norris Rebecca L. 2003. Embedded clauses in Buli. In Kenstowicz Micheal & Akanlig- Pare, George (eds.), Studies in Buli Gammar: Working Papers on Endangered and Less familiar Languages 4 (pp. 131-146). Massachusetts: MiTWPL. Nwachukwu, P. A. 1985. Inherent complement verbs in Igbo. Journal of the Linguistic Association of Nigeria, 3, 61-74. Nylander, Dudley K. 1997. Some myths about serial verbs. South African Journal of African Languages 17(3). 85-88. Olawsky, Knut J. 1999. Aspects of Dagbani grammar with special emphasis on phonology and morphology. Munich: LINCOM. Osam, Kweku E. 1994. Aspects of Akan grammar: a functional perspective. PhD Thesis, University of Oregon, Eugene. Osam, E. Kweku. 2004. The Trondheim lectures—An introduction to the structure of Akan: Its verbal and multiverbal systems. Legon: Department of Linguistics. Osam, Kweku E. 2016. Valency changing processes in Akan. In Doris L. Payne, Sara Pacchiarotti & Mokaya Bosire (eds.), Diversity in African languages, 117– 139. Berlin: Language Science Press. Owu-Ewie, Charles. 2006. The language policy of education in Ghana: A critical look at the English-only language policy of education. In Selected proceedings of the 35th annual conference on African linguistics (pp. 76-85). Somerville, MA: Cascadilla Proceedings Project. University of Ghana http://ugspace.ug.edu.gh 165 Oyelaran, Olasope. 1982. On the scope of the serial verb construction in Yoruba. Studies in African Linguistics, 13(2), 109-146. Payne, Thomas. 1997. Describing morphosyntax: A guide for field linguists. Cambridge: Cambridge University Press. Saah, Kofi K. 2003. The nature and functions of Akan verbal affixes. In
M.E.Kropp Dakubu and E.K. Osam, eds. Studies in the languages of the Volta Basin I: Proceedings of the annual colloquium of the Legon-Trondheim Linguistics Project, 4-6 December 2002., pp.180-196. Saanchi, James A. 2006. A grammar of spatial constructions in Dagaare. PhD Thesis, University of Ghana. Schott, Rudiger. 1993. Bulsa sunsuelima. Folktales of the Bulsa in northern Ghana. LIT Schwarz, Anne. 1995. Untersuchungen zum nominalen klassensystem des Buli (Nordghana). Magisterarbeit, Universität Hamburg. Schwarz, Anne. 1999. Preverbal negative markers in Buli. Gur Papers / Cahiers Voltaïques 4: 91-98. Schwarz, Anne. 2000. Perfective verb tone in Buli. Gur Papers/Cahiers Voltaïques 5:149-155. Schwarz, Anne. 2003. Low tone spreading in Buli. Gur Paper /Cahiers Voltaïques 6: 121- 130. Schwarz, Anne. 2005. Aspekte der morphosyntax und tonologie im Buli (mit schwerpunkt auf dem Buli von Wiaga). PhD Dissertaion. Humboldt-Univ, Berlin. Schwarz, A. 2016. All-in-one and one-for-all: Thetic structures in Buli grammar and discourse. Diversity in African languages, 81-100 Stewart, John. M. 1963. Some restrictions on objects in Twi. Journal of African Languages, 2(2), 145-149. University of Ghana http://ugspace.ug.edu.gh 166 Stewart, John M. 1971. Niger-Congo, Kwa. In: Sebeok TA, Berry J, Greenberg J (eds) Linguistics in Sub-Saharan Africa. Current trends in linguistics, pp. 179- 212. Mouton: The Hague. Sulemana, Abdul-Razak. 2012. The structure of the determiner phrase in Buli. M.Phil Thesis. University of Ghana. Tomasello, Michael. 1998. Introduction: A cognitive-functional perspective on language structure. In Tomasello, Michael (ed.), The new psychology of language (pp. vii-xxiii). London: Lawrence Erlbaum. Van Valin Jr, R. D., & LaPolla, R. J. (1997). Syntax: Structure, meaning, and function. Cambridge University Press. Velupillai Viveka. 2016. Partitioning the timeline: A cross-linguistic survey of tense’. Studies in Language 40(1), pp. 93-136. Welmers, W. E. 1946. A descriptive grammar of Fanti. Language, 22(3), 3-78. Welmers, W. E. 1973. African language structures. Berkley, Los Angeles & London: University of California Press.