UNIVERSITY OF GHANA AN AUTOSEGMENTAL ANALYSIS OF PHONOLOGICAL PROCESSES IN DAGARA. BY KUUBEZELLE NERIUS (10160616) THIS THESIS IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL FULFILLMENT OF THE REQUIREMENT FOR THE AWARD OF M.PHIL LINGUISTICS DEGREE. JULY 2013. University of Ghana http://ugspace.ug.edu.gh i DECLARATION I, KUUBEZELLE NERIUS, do declare that, with the exception of secondary data, drawn from the works of Dr. S. K. Bemile, and other referemces which have been duly acknowledged, the analysis in this thesis is my original work under the supervision of Dr. George Akanlig-Pare and Dr. James A. N. Saanchi and it has been submitted neither in part nor in whole previously for the award of a degree anywhere else. ……………………... Date ……………………. KUUBEZELLE NERIUS (10160616) SUPERVISORS: ……………………… Date ……………………. Dr. George Akanlig-Pare ……………………….. Date ………………….. Dr. James A. N. Saanchi University of Ghana http://ugspace.ug.edu.gh ii DEDICATION This thesis is dedicated to: My son Nuo-ire Arnold Janssen and my wife Bole Yvonne who I abandoned for two years to pursue this M.Phil. Course. And Monsieur et Madame Lotegeh a , who are currently at Ouagadougou, for sowing the Mustard Seed of education in me. University of Ghana http://ugspace.ug.edu.gh iii ACKNOWLEDGEMENTS GOD has been so good to me in all my endeavours and in all circumstances; in my academics, when I am travelling, and even when I am sleeping. Almighty Father, I say thank You!!! In the academic arena, my inexpressible appreciations primarily go to my friendly and fatherly supervisors; Dr. George Akanlig-Pare and Dr. James A. N. Saanchi, for their scholarly guidance and comments that helped fine tuned this Thesis. Thank you for your friendship and the scholarly guidance. I wish to also acknowledge with profound gratitude, the academic inputs, generously made, by all the lecturers in the department and those outside the department, during my two years academic experiences in Legon. Dr. P. K. Agbedor and Dr. James A. N.Saanchi, my Referees, thank you for the favourable recommendations to Graduate School to facilitate my admission into the M.Phil. Programme. May the Almighty God continue to guide and bless you abundantly. Prof. Kofi Agyekum, HOD of the year, accept my award of Heart Felt Gratitude, for all the administrative privileges that I enjoyed from your good office. I have always asked the Good Father to grant you all the Graces you need to continue to be kind to all who encounter you. Prof. Kofi K. Saah, thank you for the concern you always express for the health and well-being of my little boy, Nuo-ire Arnold Janssen. Your daily enquiries: “Nerius ow is your son” as been so inspiring. You will remain in my memory as a Sweet Dad and Grandpa to me and my son. University of Ghana http://ugspace.ug.edu.gh iv I also thank all the non-teaching Staff of the Department, especially Mr. Frimpong in t e Library. r. Frimpong, w enever t e term “controversial” slips into your mind, then I am saying thank you for your patience and tolerance. To all my colleagues, Graduate Students with whom I had fun and cordial relationship, thank you for accommodating me for the two years of togetherness. I would want to thank in a special manner; Mark Gandaa Dundaa, the first colleague whom I encountered for the first time during our first lecture, for the closeness we shared since then; Abdul-Razak, whom I frequently pestered to edit my Term Papers, even though he was saddled with his own Thesis at that time. Thank you, Razak, for your editorial works. I wish to acknowledge the contributions of the Management of University for Development Studies (UDS) for the study leave granted me and the contributions of every member of my family, friends and benefactors, to the successful completion of the course. The most notable are Mr. Zumah Nuobier Joseph and Mr. Okyere-Dadzie Anthony of French Language Centre-UDS, Wa Campus, Dr. Kambou Moses at UCC Ghanaian Languages Department, Dr. and Mrs. S. K. Bemile, Mr. and Mrs. Kuupol, Mr. Cyril Bangdome, Mr. Peter Damian Derzu, Mr. George Sagraza, Yvonne Bole (my Dear Wife), and Monsieur et Madame a , at Ouagadougou. onsieur et a ame a , ue irias- e vous ce moment? Je ne peux pas vous remercier pour m avoir sauv tre berger. Vous m avez envoy l cole uan mes parents voulaient ue e sois berger. e voici au our hui au University of Ghana http://ugspace.ug.edu.gh v niveau e la a trise, grace vous. Je vous remercie beaucoup pour toutes vos souveillances et bienveillances. ue Dieu ous nnisses !!! University of Ghana http://ugspace.ug.edu.gh vi ABSTRACT The stored information about the grammar in the mental lexicon of the interlocutors does not only include the morphological, syntactic and semantic components of grammar, but also its phonological component. Phonology is a component of the grammar of language which involves the production, perception as well as the interpretation of sounds. This thesis discusses the phonology and phonological processes that are specific to Dagara, a dialect of Dagaare in the Upper West Region, within Autosegmental phonology framework, espoused in Goldsmith in 1976. Assimilatory processes including Vowel Harmony, Consonant Nasalization, Homorganic Nasal Assimilation, Glide Formation, Labialization and Rhotacism, as well as some syllable structure processes are highlighted. Data for the analyses are drawn from both primary and secondary sources and my native speaker s intuition. T e primary data are gathered through elicitation using Summer Institute of Linguistics, West African Area Wordlist 1, as well as recordings of recitals using an audio sound recorder. Secondary data are drawn from existing literature on the dialect. University of Ghana http://ugspace.ug.edu.gh vii TABLE OF CONTENTS Declaration i Dedication ii Acknowledgement iii Abstract vi Table of Contents vii List of Tables xv List of Figures xviii List of Abbreviations and Sybols xxiii CHAPTER ONE: BACKGROUND STUDY 1.0 Introduction 1 1.1 The Dagaare language and its people 1 1.2 Dialect variation in Dagaare 3 1.3 Problem Statement 6 1.4 Scope of the Research 7 1.5 Aims and Objectives of the Study 7 1.6 Significance of the Study 8 1.7 Data collection and Methodology 8 1.8 Theoretical framework 10 1.8.1 Generative Phonology 10 1.8.2 Autosegmental Phonology 11 University of Ghana http://ugspace.ug.edu.gh viii 1.8.2.1 Basic tenets of Autosegmental Phonology 14 1.8.2.1.1 Skeletal tier 14 1.8.2.1.2 Universal Association Convention 15 1.8.2.1.3 Linkage condition 16 1.8.2.1.4 Obligatory Contour Principles 17 1.8.2.1.5 Well Formedness Condition 18 1.8.3 Motivation for Autosegmental Phonology 19 1.8.4 Tiers used in the thesis 23 1.9 Literature Review 25 1.10 Organization of Work 29 1.11 Chapter Summary 30 CHAPTER TWO: DAGARA SOUND SYSTEM 2.0. Introduction 31 2 .1 Vowels 31 2.1.1 Oral Vowels 31 2.1.2 Nasal Vowels 35 2.1.3 Tongue Root Vowels 40 2.1.4 Diphthongs in Dagara 41 2.2 Consonants 43 University of Ghana http://ugspace.ug.edu.gh ix 2.3 Phonetic Feature Description of Dagara Sounds 47 2.3.1 Major Class Features 48 2.3.1.1 [ +/- Consonantal] Sounds 48 2.3.1.2 [ +/- Syllabic ] Sounds 48 2.3.1.3 [ +/- Sonorant ] Sounds 49 2.3.2 Cavity Features 49 2.3.2.1 Primary stricture 50 2.3.2.1.1 [ Coronal ] Sounds 50 2.3.2.1.2 [+/- Anterior ] Sounds 50 2.3.2.1.3. [ Labial ]; [+/- Round ] Sounds 50 2.3.2.2 Tongue Body Features 51 2.3.2.2.1. [+/- High] Sounds 51 2.3.2.2.2. [+/- Low] Sounds 52 2.3.2.2.3 [+/- Back]; [ Dorsal ] Sounds 52 2.3.2.2.4 [+/- ATR] Sounds 52 2.3.3 MANNER FEATURES 53 2.3.3.1 [+/- Continuant] Sounds 53 2.3.3.2 [+/- Lateral] Sounds 53 University of Ghana http://ugspace.ug.edu.gh x 2.3.3.3 [+/- Nasal] Sounds 54 2.3.3.4 [+/- Strident] Sounds 54 2.3.3.5 [+/- Del. rel] 54 2.3.4. Laryngeal feature 55 2.3.4.1 [+/- Voiced] Sounds 55 2.4 The Syllable 55 2.4.1 Syllable Types in Dagara 59 2.4.1.1 The V/V: syllable type 59 2.4.1.2 The VC syllable type 60 22.4.1.2.1 The VVC syllable shape 61 2.4.1.3 The CV syllable Type 62 2.4.1.3.1 The CV/CV: syllable shape 62 2.4.1.3.2 The CVV syllable shape 63 2.4.1.4 The CVC syllable type 64 2.4.1.4.1 The CVC syllable shape 64 2.4.1.4.2 The CVVC/CV:C syllable shape 65 2.4.1.5 The C syllable shape 66 2.4.2 Dagara Word Structure 67 University of Ghana http://ugspace.ug.edu.gh xi 2.4.2.1 Simple Stem Words 67 2.4.2.1.1 Mono-syllabic words 67 2.4.2.1.2 Disyllabic words 68 2.4.2.1.3 Trisyllabic words 68 2.4.2.1.4 Quadrisyllabic words 69 2.4.2.2 Complex Stem Words 70 2.5.0 Chapter Summary 71 CHAPTER THREE: PHONOLOGICAL PROCESSES IN DAGARA 3.0 Introduction 72 3 .1 Assimilation Processes 72 3.1.1 Vowel Harmony 73 3.1.1.1 Verb Stem and Imperfective Suffix Harmony 76 3.1.1.2 Verb Stem and Nominal Suffix Harmony 81 3.1.1.3 Noun Stem and Nominal Suffix Harmony 84 3.1.1.4 [+ATR] Verb and Negation Particles Harmony 87 3.1.1.5 [-ATR] Stems and Diminutive Suffix Harmony 90 3.1.1.6 [-ATR] Stems and [+ATR] Stem Harmony 92 3.1.2 Consonant Nasalization 94 University of Ghana http://ugspace.ug.edu.gh xii 3.1.3 Homorganic Nasal Assimilation 96 3.1.4 Consonant Labialization 99 3.1.5 Glide Formation 101 3.1.6 Rhotacism 103 3.2.0 Syllable Structure Processes 106 3.2.1 Elision 107 3.2.1.1 Vowel Elision 107 3.2.1.2. Consonant Elision 110 3.2.2 Compensatory Lengthening 113 3.2.3 Resyllabification 116 3.2.4 Vowel Epenthesis 118 3.3.0 Chapter Summary 120 CHAPTER FOUR: TONE AND TONAL PROCESSES IN DAGARA 4.0 Introduction 122 4 .1 Pitch and Tone 122 4.1.2 Tone Bearing Unit 124 4.3 Dagara Tone Systems 126 4.3.1 Level Tones 127 4.3.1.1 High Tone 127 University of Ghana http://ugspace.ug.edu.gh xiii 4.3.1.2 Mid Tone 128 4.3.1.3 Low Tone 128 4.3.2 Contour Tones 129 4.3.2.1 Falling Contour Tone 129 4.3.2.2 Rising Contour Tone 130 4.4 Tone Combination in Dagara 131 4.5 Tone Stability 131 4.6 Tone Terracing 135 4.6.1 Declination 136 4.6.2 Downstep 136 4. 7.0 Function of Tone 139 4.7.1 Lexical Tone 139 4.7.1.1 Minimal Pairs 139 4.7.1.2 Minimal Tripplets 139 4.7.1.3 Minimal Quadruplets 140 4.7.2.0 Grammatical Tone 140 4.7.2.1 Nominative/Possessive and Accusative/ Dative Tone 141 4.7.2.2 Aspect Tone 142 4.7.2.3 Mood (Subjunctive II) Tone 143 University of Ghana http://ugspace.ug.edu.gh xiv 4.7.2.4 Negation Tone 144 4.7.2.5 Locative Expression Tone 145 4.7.2.6 Nominal Suffix Tone 146 4.8 Tonal Processes 147 4.8.1 Tone Assimilation 147 4.8.2 Contour Tone Simplification 149 4.8.2.1 Tone Absorption 149 4.8.2.2 Tone Deletion 152 4.8.4 Tone Dissimilation 153 4.9 Chapter Summary 156 CHAPTER FIVE: SUMMARY AND CONCLUSION 5.0 Introduction 158 5.1 Summary 158 5.2 Conclusion 161 APPENDICES 162 REFERENCES 178 University of Ghana http://ugspace.ug.edu.gh xv LIST OF TABLES TABLE PAGE Table 1. Dialect Variation at Segmental Level 4 Table 2. Dialect Variation in Tone 4 Table 3. Oral Vowel in word examples 33 Table 4. Long Oral vowels in word examples 34 Table 5. Short and Long Oral Vowels in contrast 34 Table 6. Short nasal vowels in words 37 Table7. Long Nasal Vowels 37 Table 8. Short and Long Nasal Vowels in Contrast 38 Table 9. Oral and Nasal vowels in contrast 39 Table 10. Tongue Root Harmony 41 Table 11. Diphthongs in Dagara 42 Table 12 Consonant Chart 44 Table 13. Consonant Phonemes Distribution in Words 45 Table 14. The V/V: Syllable Shape 59 Table 15. The VC/V:C syllable shape 60 Table 16. The VVC syllable shape 61 Table 17. The CV/CV: syllable shape 62 Table 18. The CVV Syllable Shape 63 University of Ghana http://ugspace.ug.edu.gh xvi Table. 19 The CVC syllable shape 64 Table 20. The CVVC/CV:C syllable shape 65 Table 21. The C Syllable Shape 66 Table 22. Mono-syllabic Simple Words 67 Table 23. Disyllabic Simple Words 68 Table 24. Trisyylabic Simple Words 69 Table 25. Quadrisyllabic Simple Words 69 Table 26. Complex Stem Words 70 Table 27. Verb Stem and Imperfective/Progressive Suffix 76 Table 28. Verb and Nomial Suffix 81 Table 29. Noun Stem and Nominal Suffix 86 Table. 30. [-ATR] Noun Stems and Diminutive Suffix Harmony 90 Table 31. [-ATR] Stems and [+ATR] Dominant Stem Harmony 92 Table 32. Consonant Nasalization 94 Table 33. Homorganic Nasal Assimilation 96 Table 34 Consonant Labialization 99 Table 35. Glide Formation 101 Table 36. Rhotacism 104 Table 37. Final Vowel Elision in CV/ CVV words 108 Table 38. Vowel elision in CVVC words 108 University of Ghana http://ugspace.ug.edu.gh xvii Table 39 Consonant Elision in compounding process 110 Table 40.Consonant and Vowel Elision in fast Speech 112 Table. 41. Compensatory Lengthening 114 Table 42. Resyllabification in Dagara 117 Table 43. Vowel Epenthesis in Loan Word 119 Table 44. Tonal Combination 131 Table 45. Tone Stability 132 Table 46. Pronouns 141 Table 47. Non-Low Tone Absorption 151 Table 48. Contour Tone Simplification by deletion 152 Table 49. Low –Low Tone dissimilating to become Rising-Low 154 Table 50. Low Tones dissimilating to Low-Falling tone pattern 155 University of Ghana http://ugspace.ug.edu.gh xviii LIST OF FIGURES FIGURE PAGE Fig.1.The Genetic-Linguistic Relationships of Dagaare 5 Fig.2. Linear Representation of Segmental and Suprasegmental Features 11 Fig.3. Autosegmental Representation of Segmental and Suprasegmental Features. 13 Fig.4. Skeletal Tier 14 Fig.5. Universal Association Convention 16 Fig. 6. Linkage Condition 17 Fig. 7. Obligatory Contour Principle 18 Fig.8. Well Formedness Condition 19 Fig.9. Contour Tone and Double Articulated Segment Representation 20 Fig. 10. Assimilatory Derivation Process 22 Fig. 11. Syllable Tier 24 Fig.12. Wrongful Arrangement of Tiers 28 Fig.13. Short Oral Vowel Chart 32 Fig. 14 Short Nasal Vowel Chart 36 Fig. 15. Onset and Rhyme Model Structure 57 Fig. 16. Moraic Structural Model 58 Fig. 17. Structural Representation of V/V : Syllable Shape 60 Fig. 18. Structural Representation of VC/V:C Syllable Shape 61 University of Ghana http://ugspace.ug.edu.gh xix Fig 19. Structural Representation of VVC Syllable Shape 62 Fig.20. Structural Representation of CV/CV: Syllable Shape 63 Figure 21. Structural Representation of CVV Syllable Shape 64 Fig. 22. Structural Represention of CVC Sylllable Shape 65 Fig. 23. Structural Representation of CVVC Syllable Shape 66 Fig 24. Structural Representation of C Syllable Shape 66 Fig. 25. Verb and Imperctive Suffix Harmony Process 78 Fig. 26. Verb Stem and Nominal Suffix Harmony 83 Fig.27. Noun Stem and Nominal Suffix Harmony 86 Fig. 28. Harmony across Word Boundary 89 Fig.29. [-ATR] Noun Stems and Diminutive Suffix Harmony 91 Fig.30. [-ATR] Stems and [+ATR] Dominant Stem Harmony 92 Fig.31. Autosegmental representation of Consonant nasalization 95 Fig.32. Autosegmental representation of Homorganic Nasal Assimilation 97 Fig.33. Autosegmental Representation of Consonant Labialization 100 Fig 34. Glide formation in Autosegmental Representation 102 Fig.35. Rhotacism Autosegmental Representation 105 Fig.36. Vowel Elision in Autosegmental Representation 108 Fig.37. Consonant Elision in Autosegmental representation 111 University of Ghana http://ugspace.ug.edu.gh xx Fig.38. Compensatory Lengthening in Autosegmental Representation 114 Fig. 39. Resyllabification in Autosegmental Representation 117 Fig.40. Vowel Epnthesis in Autosegmental Representation 120 Fig.41. High Tone 127 Fig.42. Mid tone 128 Fig.43. Low Tone 129 Fig.44. Falling Contour Tone 130 Fig.45. Rising Contour Tone 130 Fig. 46. Autosegmental Representation of Tone Stability 133 Fig.47. Declination illustrated in sequence of Dagara High tones 136 Fig.48. Declination illustrated in sequences of Hausa High tones 137 Fig 49. Downstep in the sequence of Dagara Low-High tones 138 Fig.50. Minimal contrastive Pairs 139 Fig. 51. Minimal contrastive Triplets 139 Fig.52. Minimal Contrastive Quadruplets 140 Fig.53. Use of Tone to Distinguish Pronouns 142 Fig.54. Mid Tone for Aspect Marking 142 Fig.55. Mid Tone for Subjunctive Expression 144 University of Ghana http://ugspace.ug.edu.gh xxi Fig.56.Tone for Negative Expressions 145 Fig.57. Tone for Locative Expressions 146 Fig.58. Tone Assimilation 148 Fig.59. Autosegmental Representation of Non-low tone absorption 151 Fig.60. Contour Tone Simplification by Deletion 153 Fig.61. Representation of Low Tones Dissimilating to Rising-Low Pattern 155 Fig.62. Representation of Low Tones dissimilating to Low- Falling pattern 156 University of Ghana http://ugspace.ug.edu.gh xxii LISTS OF ABBREVIATIONS AND SYMBOLS 1PL First Person Plural 2PL Second Person Plural 3SG Third Person Singular Ant Anterior -ATR Unadvanced Tongue Root +ATR Advanced Tongue Root C Consonant/ Coda Cons Consonantal Cont Continuant CV Consonant Vowel del rel Delayed Release H High tone HL Falling tone HUM Human Hz. Hertz Ibid Ibidem i.e that is L Low tone Lab Labial LH Rising tone M Mid tone M. A. Manner of Articulation MIT Massachussetts Institute of Technology N Nucleus NEG Negation University of Ghana http://ugspace.ug.edu.gh xxiii O Onset OCP Obligatory Contour Principle P. A. Place of Articulation PL Plural PROG Progressive SILWL Summer Institute of Linguistics African Area Word List Son Sonorant SPE Sound Pattern of English Sylla Syllabic TBU Tone Bearing Unit UAC Universal Association Convention UCC University of Cape Coast UEW University of Education, Winneba UG University of Ghana V Short Vowel V: Long Vowel WFC Well Formedness Condition SYMBOLS Low tone High tone Falling tone Rising tone ó Syllable / / Phonological Brackets [ ] Phonetic Brackets ( ) Brackets University of Ghana http://ugspace.ug.edu.gh 1 CHAPTER ONE BACKGROUND STUDIES 1.0 Introduction This introductory chapter starts with a brief statement on the Language and its speakers with a sketch of dialectal variation in the language. The Problem Statement, the Scope of the research, the Aims and Objectives of the research, the Significance of the research as well as data collection and the Methodology used are discussed. It also contains an explanation of the Theoretical Framework and the motivation for using the framework in the phonological and tonal representations. A Literature review and the Organization of the thesis conclude the chapter. 1.1 The Dagaare Language and its people Bendor-Samuel (1971:144) classifies Dagaare as belonging to the northern branch of the Western Oti-Volta group of the Gur branch of the Niger-Congo language family. Dagaare shares very close affinity with other languages of the Gur branch of the Niger-Congo language family like Dagbanli, Moore, Gurunɛ, Mampruli, Kusaal and Buli. Bodomo (2000: 3) refers to these West Oti-Volta languages as Mabia. Geographically, Nanbigne (2008:32- 33) situates Dagaaba within latitudes 90 N and 110 N and longitude 20 W and 30 W covering a major part of Upper West Region an some towns in t e Nort ern egion an urt er stretc es beyon t e international bor ers into urkina Faso an te Ivoire. University of Ghana http://ugspace.ug.edu.gh 2 Dagaare is not however confined to the Upper West Region and part of the Northern Region. Occupational mobility has caused some native speakers to leave their traditonal land to other regions in Southern Ghana for their livelihood. This explains why there are several important Dagaare speaking communities in Accra, Kumasi, Obuasi, Sunyani, Kintampo, Bolgatanga Tamale and other significant towns and villages throughout Ghana. There are no accurate and up-to-date census figures in recent times to account for the number of native speakers of Dagaare. However, it is estimated to be the fourth largest indigenous language in Ghana after Akan, Ewe and Dagbani (Bodomo 1997). There have been controversies over the years with regard to the right nomenclature to designate the language and the people. Some scholars in Ghana use the term Dagaare exclusively to refer to the language and the term Dagaaba (Dagao, singular) to refer to its speakers (cf. Bodomo 1997, 2000, Saanchi 2006, Nanbigne 2008 etc.). On the other hand, other scholars in Ghana and across the borders of Burkina Faso and te d Ivoire use the term Dagara as a generic term to refer to the language, the speakers and the land (cf Bemile 1983, 1985 Delplanque 1983, om 2004 etc). Despite these controversies, the speakers both in Ghana and beyond recognize the language as one t at c anges gra ually in a “ ialect continuum” as one moves from north to south and from east to west (Dakubu 2005: 1). In this thesis, the term Dagaare is used to refer exclusively to the language. University of Ghana http://ugspace.ug.edu.gh 3 According to Bodomo (1997), there are four major dialects that make up the language, though other scholars, in Ghana and African francophone linguists in Burkina Faso, would include Wule spoken around Lawra-Babile and across the Black Volta, and others as dialects of Dagaare. They include Dagara (or Northern Dagaare), spoken in and around Lawra, Nandom, Hamile and their surroundings as well as in Burkina Faso in towns such as Dano, Dissin, Gaoua, and Diebougou; Central Dagaare spoken in and around Jirapa, Nadwoli, and Daffiema; Southern Dagaare (Waali) spoken in Kaleo and Wa and; finally Birifor which is mainly spoken in Tuna, Kalba and surrounding areas in the Northern Reigion and in towns across the Black Volta as well as in te Ivoire. There is however no single standard form recognized by all as there are variations, though some are mutually intelligible. 1.2 Dialect variation in Dagaare In view of the fact that there is no agreed single standard form of the language, it is important to pay attention to variation when describing the language. The dialects differ considerably at phonological, lexical and grammatical levels, as noted by Dakubu (2005), but since the focus of the thesis is on phonology, it is prudent to limit ourselves only to the phonological variation. There are some phonological features in Dagaare which are not common to all the four major dialects. Bemile (1983), Bodomo (1997), Dakubu (2005) and Saanchi (2006), all agree that the Birifor and Dagara dialects have some features that other dialects do not have. They all identify the bilabial implosive /ɓ/, the voiceless velar fricative /x/, and the alveolar lateral ingressive / 'l/ as uncommon phonological features in Central dialect and Southern dialect. University of Ghana http://ugspace.ug.edu.gh 4 On the other hand, there is a phonological feature common to the Southern and Birifor dialects but which is not in the Central and Dagara dialects. For instance, in the Southern and Birifor dialects, the voiced alveolar fricative /z/ does not exist. It has been replaced by the voiced alveo-palatal affricate /ʤ/. The Central and Dagara dialects however still have both /z/ and /ʤ/ in their sound system. The table 1 below displays some examples of dialect variation at the segmental level. Table 1. Dialect Variation at Segmental Level Dagara Central Dagaare Southern Dagaare Birifor Gloss ɓ f l f l ɓ to slap 'l ny g ny g 'l to light x n w l w l xanf sweat z z ʤ ʤ head Besides the variation at the segmental level, there is also variation in the use of suprasegmentals, especially tone in the four dialects. Some morphemes have different tones in the different dialects. Dakubu (2005) remarks that the Central dialect has a falling tone in some words where other dialects do not. Table 2 displays some examples of variation in tone drawn from Dakubu (2005 :6) University of Ghana http://ugspace.ug.edu.gh 5 Table 2 Dialect Variation in Tone. Dagara Central Dagaare Southern Dagaare Birifor Gloss z ʤ ʤ ea b b b badaba friend z ɛ z ɛ ʤ ɛ ʤ ɛ millet University of Ghana http://ugspace.ug.edu.gh 6 Figure 1. The Genetic-Linguistic Relationships of Dagaare (adopted from Dakubu 2005:4) Niger-Congo Mande Volta-Congo Adamawa Gur Kulango Lobi Central etc. Grusi Oti-Volta Kurumfe South-West East (Delo) Western Eastern North-West South-West West East Buli ( Konni) Senufo North (Kasem, Lyele) Gurma, Bimoba, Konkomba Dagbanli, Mampruli, Kusaal, Hanga Dagaare, Safaliba Moore Gurunɛ Sisalla Chakale, Vagla, Deg Tampulma Kwa (Akan, Ga, Ewe) Benue-Congo ( Yoruba, Ibo (Igbo), Bantu languages) University of Ghana http://ugspace.ug.edu.gh 7 1.3 Problem Statement Since the Language was reduced to writing in 1929 by Catholic Missionaries, the Central dialect among the four major dialects has been the focus of linguistic research and analyses by both missionary linguists and academic linguists (Saanchi 2006:4-5). The major dialects exhibit obvious linguistic variation which could be at lexical, phonological and grammatical levels, yet only the Central dialect has received prominent attention in literature and sociolinguistic functions in Ghana to the disadvantage of the others. It is this dialect that represents the other dialects at the National level and is being taught at the basic schools as well as at the tertiary institutions such as the Training Colleges and the Universities including University of Education, Winneba, (UEW), University of Cape Coast (UCC), and the University of Ghana (UG). For a holistic development of the language in Ghana, however, there is need for linguistic research and analyses in terms of phonetics, phonology, morphology, syntax and semantics in the other dialects. Though Dagara phonology is somehow being developed, much of the work is undertaken by francophone African linguists in Burkina Faso. Bodomo (1997:5) attests to t is as e a mits t at, “most o t e linguistic analyses o t is group of Dagaare [Dagara] have been undertaken by French and francophone A rican linguists in urkina Faso”. University of Ghana http://ugspace.ug.edu.gh 8 Linguistic works in Dagara in Ghana are but a few (Bemile 1983, 1984, 1985, 1990, 2010 etc.). It is therefore certainly motivating to carry out phonological analyses in this dialect in Ghana. Besides it has been observed that the spoken form of Dagara varies from the written form. Obviously some kind of phonological processes occur in the language which need investigation. 1.4 Scope of the Research The thesis is limited to Dagara phonology and the phonological processes that are specific to it. It discusses the Dagara sound system, i.e. vowel and consonant phonemes, its syllable structure in relation to other dialects especially the Central dialect. It also discusses the major findings, i.e. the phonological processes specific to Dagara. Since Dagara is a tone language like other African languages, tone is not left out of the discussion. The phonological representation of the analyses is in the autosegmental framework. The research is carried out in selected villages under Nandom and Hamile. The villages covered include Nandom-Kogle, Nandom-Gengenkpe, Hamile- Kandemegaŋ and Kyebogo. These villages are some of the typical Dagara speaking communities in Ghana and are selected on the score of accessibility. 1.5. Aims and Objectives of the Study The primary aim of this Thesis is to contribute to the existing literature on Dagara phonology. The specific objectives to meet include the following: I. To discuss the Dagara sound system. II. To give an account of the phonological processes in Dagara University of Ghana http://ugspace.ug.edu.gh 9 III. To discuss Dagara Tone and Tonal patterns as well as the tonal processes that occurs in Dagara. IV. To formalize the description of the phonological processes and tonal processes within the Autosegmental framework. 1.6 Significance of the Study The importance of the findings will be significant in several ways. On the descriptive perspective, this thesis may be the first attempt to elaborately deal with the phonological processes in Dagara in English. On the theoretical perspective, this is the first attempt to discuss phonological processes and tone representation in the language within the autosegmental framework. It will also help to address the deficiencies in the description of the phonological process and ultimately add to the literature on the dialect in particular and the language in general. 1.7 Data Collection and Methodology A blend of primary data, secondary data and my native speaker intuition were relied on in this work. Most of the secondary data are drawn from Bemile (1983, 1984 & 1985). For primary data, I spent about three weeks (in the month of October 2012) in the field eliciting data, using Summer Institute of Linguistics West African Area Wordlist 1 (SILWL 1) in Boone Douglas (1989), through interviews with four native speaker informants. The elicitations were recorded using an audio recorder. Recitations of prayers, with a blind aunt who is a mono-lingual Dagara and non-literate as lead person, were recorded at Holy Family Parish, Hamile. University of Ghana http://ugspace.ug.edu.gh 10 A total of about six hundred and thirty (630) words are elicited in the selected Dagara speaking communities. At Nandom-Kogle, the informant, a retired educationist, Mr. Sylvanus Baaro produced one hundred and fifty (150) Dagara words of the SILWL1. Mr. Peter Damian Derzu at Kyebogo, another retired educationist, who prior to his retirement was teaching Dagaare at Tumu College of Education, also produced one hundred and fifty (150) Dagara words of the SILWL1. At Hamile-Kan emegaŋ, r. yril ang ome, a retire educationist who is currrently teaching Dagaare at Holy Family Senior Secondary School at Hamile, also produced two hundred (200) Dagara words of the SILWL1. Finally, Mr. George Sagraza at Gengenkpe, a native speaker and a young teacher, produced one hundred and thirty (130) Dagara words of the SILWL1. I then transcribed the recorded data using my native speaker intuition and also in consultation with the language informants where there was need for clarification and verification. Relevant data are then extracted and used for this analysis. It must, however, be mentioned that it was not easy carrying out the data elicitation in t e iel . It was i icult getting t e language elpers au iences. There were a number of disappointments in attempts to meet with them for some reasons. The period of the fieldwork coincided with the peak of the harvest season which, of course, was a very busy period for most of them. Besides, funeral celebrations and bad weather were also contributory factors to the difficulties faced during the data elicitation. University of Ghana http://ugspace.ug.edu.gh 11 1.8 Theoretical Framework. Though the mode of phonological representation is within Goldsmith s (1976) Autosegmental P onology, a review o omsky an Halle s linear generative approach to phonological representation is inevitable. This section, therefore, discusses the limitations of Chomsky and Halle (1968) linear approach to phonological representation. It then explains how autosegmental phonology operates and the motivation for its use in this thesis. 1.8.1 Generative Phonology The Generative phonology theory espoused in Chomsky and Halle (1968): Sound Pattern of English (SPE) was a dependable theory for phonetic and phonological analyses before the emergence of autosegmental phonology. It characterized segments as unstructured, homogeneous feature matrices, “w ere every segment has a specification for each of the two dozen distinctive features orderly bound together as a unit” (Odden 2005: 314). Phonological representations of segments in generative phonology, therefore, consist of a string of segments arranged in a linear form like the serial arrangement of the orthography. For instance, for the Dagara word ŋ ‘know , eac segment is broken own into t eir istinctive eatures an arranged linearly with each representing an aspect of the pronunciation of the entire word. Figure 2 below shows the linear generative phonology representation of the word ŋ. University of Ghana http://ugspace.ug.edu.gh 12 Fig. 2 . Linear representation of segmental and suprasegmental features b a ŋ +cons +sylla +cons +voiced -ATR +nasal +labial Low tone +back In this representation, the Low tone is lumped up with other features and considered as part and parcel of the segment /a/. Viewing segments in this way implies that in the event of segment deletion, every distinctive feature goes with the segment in question. However sometimes, tone and other suprasegmental phonemes which are neither vowels nor consonants resist the effect of the deletion of a segment. Consequently, generative phonology could not succeed in explaining such facts about sounds. It, indeed, fails woefully in accounting for suprasegmental features of sounds such as Tone in phonological analysis (Odden 2005:314), thus the emergence of new phonological theories. 1.8.2 Autosegmental Phonology. Autosegmental phonology was developed within the tradition of classical Generative phonological theory of Chomsky and Halle (1968), following the works of Williams (1971) and Leben (1973) on tone systems in West African languages such as Margi, Igbo and Mende. But the principal and remarkable innovations of autosegmental phonology are exemplified in Goldsmith (1976) in his dissertation to Massachusetts Institute of Technology (MIT). University of Ghana http://ugspace.ug.edu.gh 13 Autosegmental phonology is not a departure from the principles of Generative p onology co i ie in omsky an Halle s Sound Pattern of English (SPE) in 1968. It is a Non-linear version of phonological analyses of Generative phonology while SPE is a linear version of phonological analyses. Indeed, Goldsmith (1990:1) confesses that autosegmental phonology is a “ irect continuation o t e tra itional works in generative p onology co i ie in omsky an Halle s Sound Pattern of English in 1968”. The only difference between the SPE and autosegmental t eory is “t e evelopment o a multi-linear phonological analysis in which different features may be placed on separate tiers, an in w ic t e various tiers are organi e by ‘association lines ” in autosegmental P onology (Goldsmith, 1979:202). It can therefore, best be viewed as an attempt to supply a more adequate understanding of the phonetic side of linguistic representation (Goldsmith, 1976:16). Goldsmith (1976:16) further explains that autosegmental phonology is a “theory of how various components of the articulatory apparatus, i.e. the tongue, t e lips, t e larynx an t e velum are coor inate ” in the process of sound production. It, therefore, implies that phonological features lead their own independent lives and not grouped together in unordered bundles (segments) as it is made to believe in generative phonology. In this vein therefore, autosegmental phonology places segments and suprasegmentals (especially tone and other prosodic features) on different tiers parallel to each other with the suprasegmentals being linked to the segments by University of Ghana http://ugspace.ug.edu.gh 14 association lines. Every segment on each tier is specified for a set of features specific and unique to that tier and segments on each tier are associated with segments on the other tiers by association lines. Association lines in the derived or phonetic representation indicate a relationship of simultaneity, while at deeper levels of representation they specify a more abstract relationship among the segments on the separate tiers (http://hum.uchicago.edu/~jagoldsm/papers/autosegmentalunderspecificationwnos ke.pdf). The autosegmental phonolgical representation of the Dagara word ŋ will therefore take the form as in Fig. 3 below: Fig.3. Autosegmental Representation of Segmental and Supresegmental Features. Tone tier L Skeletal tier x x x Segment tier b a ŋ In the figure 3 above, the segments are arranged on multiple separate and independent tiers and each relates to the other tiers by the association lines. The Low tone and the segment /a/ relate to each other via the skeletal tier signifying the simultaneous co-ordination of the segment and the suprasegmental at the point of articulation. Thus, it is indeed an improvement over linear generative phonology since it allows for representation of sequence of segments and suprasegmentals simultaneously. University of Ghana http://ugspace.ug.edu.gh 15 1.8.2.1 Basic Tenets of Autosegemental Phonology The application of autosegmental phonology is based on some fundamental principles and conditions. These principles and conditions, though originally meant for tone analysis, apply to all other phonological representations. They include the following principles and conditions: the skeletal tier, linkage condition, Universal Principle of association, Obligatory Contour Principle and Well-formedness Condition. 1.8.2.1.1 Skeletal Tier The skeletal tier plays a crucial role in the organization of the entire phonological structure. It is the mediating point on which the elements on the different tiers anchor. A segment which is not linked to a position in the skeletal tier is not phonologically realized. The skeletal tier holds units represented by consonants and vowels; hence they are sometimes represented as C-slots and V-slots or X-slots. The skeletal tier can relate with other tiers on one-to-one, one-to-many or many-to-one basis. Fig. 4 below illustrates the position or relation of the skeletal tier to other tiers in autosegmental representation of the words ‘to lock and s ‘hail in Dagara. Fig. 4. Skeletal Tier a. Tone tier L Skeletal tier X X Segmental tier k p a University of Ghana http://ugspace.ug.edu.gh 16 b. Tone tier H Skeletal tier X X X X X X X X Segmental tier s a b i k u b e In 4a) above, the Low tone on the tonal tier relates to the /a/ on the segmental tier via the V on the skeletal tier in one-to-one fashion, while one C on the skeletal tier relates to the segment / kp/ in a one-to-two fashion, indicating that the segment is a double articulated one. But in 4b) the Vs on the skeletal tier relate to only one High tone on the tonal tier which illustrates the many-to-one fashion of the relationship, whereas the skeletal tier relates to the segmental tier on one-to-one basis. 1.8.2.1.2 Universal Association Convention (UAC) The UAC embodies three principles which indicate the relationship between tones and Tone Bearing Units (TBUs). These principles guide the assigments of tone to TBUs as outlined by Durand (1990:249) below: A. Mapping: Associate vowels with tones in a one-to-one fashion left to right until we run out of tones or vowels. B. Dumping: If after applying (mapping) some tones are still free (that is unassociated), link them to the last vowel to the right. C. Spreading: If after applying (mapping) some vowels are still free, link them to the last tone on the right. The following data on Margi (a Chadic language in Northern Nigeria) adapted from Oyebade (1998:129) best illustrate the three principles in Fig. 5. University of Ghana http://ugspace.ug.edu.gh 17 Fig.5. Universal Association Convention Tone tier H L HL H L Segmental tier sa dla bdlu sa-na dla-na a. Mapping: Tone tier H L HL H L Segmental tier sa dla bdlu sa-na dla-na b. Dumping : Tone tier H L HL H L Segmental tier sa dla bdlu sa-na dla-na c. Spreading : Tone tier H L HL H L Segmental tier sa dla bdlu sa-na dla-na 1.8.2.1.3 Linkage Condition Though segments are represented on autonomous tiers, they are not phonologically realized if they are not linked to other tiers. This condition thus demands each tonal feature and tone bearing segment on the separate parallel tiers be linked to each other via the skeletal tier by association lines before it can be phonologically realized (Goldsmith, 1990:53). Association lines represent concurrent articulations so that when tone is linked to the segment, then it means that both tone and segment are articulated simultaneously. This condition actually caters for the phenomena of floating tone and tone stability. A syllable structure process, during fast speech in Latɛ, which results in high vowel deletion, leaves its tone floating but the linkage condition allows the tone to dock on the consonant next to it, thus making the consonant a syllabic consonant as illustrated in Akrofi Ansah (2002:8), demonstrates the import of the linking condition. University of Ghana http://ugspace.ug.edu.gh 18 Fig. 6. Linkage Condition Tone tier L H L H L H Skeletal tier X X X X X X XX X XX X Segmental tier f u r i f u r i f r i Tonal Tier L H L H Skeletal Tier X X X X X X X Segmental Tier f r i f r i In figure 6 above, the segment /u/ is linked to the skeletal tier in the underlying form but it is delinked during fast speech and for that matter it is not realized phonetically in the pronunciation of the word. The /u/ therefore remains silent while its Low tone docks onto the following sound /r/ hence the resultant is the syllabic consonant formtion in the Language. The word therefore suffaces as [ r ] rat er t an r ] during fast speech. 1.8.2.1.4 Obligatory Contour Principle (OCP) This principle spells out categorically that concatenation of two or more adjacent, identical tones are fused into a single tone before t ey are “mappe onto” t eir correspon ing vowels. W at t is means is t at, a wor wit more than one Low tone vowel such as ‘hunter’ have to be mapped onto one Low tone (that is, one-to-many) instead of a one-to-one mapping of the two Low tones. The principle also applies perfectly to other segments on the other tiers. The representation in Figure 7a is in consonance with the OCP whilst 7b is a violation of the principle as shown below: University of Ghana http://ugspace.ug.edu.gh 19 Fig. 7. Obligatory Contour Principle a ) Tone teir H L b) * H L L Skeletal tier X X X X X XX X X X X X XX Segmental tier n a b a g l ɛ n a b a g l ɛ From the illustration above in Fig. 7a the OCP permits the two Low tones to be merged into just a single Low tone (one-to-may representation). 1.8.2.1.5 Well Formedness Condition (WFC) Well Formedness Conditions are universal principles which govern the multi- tier structure of autosegmental representation. These conditions actually determine the way tone and vowels are associated on the parallel tiers and allow the addition or deletion of association lines at any point throughout a phonological derivation. These are stupilated in Goldsmith (1979:207) as follows: 1. Each vowel must be associated with (at least) one toneme. 2. Each toneme must be associted with (at least) one vowel 3. Association lines may not cross. By the Well Formedness Condition, all the tiers remain independent throughout the derivation with association lines linking them. At no point should the association lines cross in the course of the derivation. The representation in figure 8a) is a violation of the Well-Formedness Condition because the Low tone has crossed the High tone to map onto the last ton bearing unit via the skeletal tier, eventhough the two are not adjacent. University of Ghana http://ugspace.ug.edu.gh 20 The correct configuration which conforms to the Well-Formedness condition is 8b. Fig.8 Well Formedness Condition a). Tone tier * L H b) L H L Skeletal tier X X X X X X X X X X X X X X Segmental tier D a g a r e D a g a r e 1.8.3 Motivation for Autosegmental Phonology The motivation for the use of the autosegmental framework in this thesis is based on its ability and flexibilty in representing complex segments such as, “ ontour tone”, Tone tability, Floating tones, ouble articulate segments an germinates among others and also in providing explanation for their occurrences which were otherwise problematic to deal with in linear generative phonology. Autosegmental phonology provides an excellent representation for Contour tone and double articulated segments via association lines in one-to- many and many-to-one mapping principle. ontour tone vowel is a vowel “w ose sur ace tone is rising or falling, a situation that can often be the result of a concatenation o Low an Hig tone” (Goldsmith, 1979:205). Since Dagara is a tonal language and also has double articulated segments, autosegmental phonology is the best option for the phonological representation in this analysis. University of Ghana http://ugspace.ug.edu.gh 21 Thus, the segments / / and /kp/ are represented respectively as in figure 9 below: Fig9. Contour Tone and Double Articulated Segment Representation a) H L b) C a k p In 9a) although the segment /a/ is made up of a sequence of articulations and behaves phonologically as a sequence of features, it also behaves phonologically as a single segment. Whilst in 9b) although the segment /kp/ consists of a double articulation and behaves phonologically as a single feature specification, it also behaves as two segments (Sagey, 1986:22). These facts about phonological representation cannot be captured within the linear generative phonology. “Tone tability” is anot er elusive phonological phenomenon which linear generative phonology could not account for in phonological analysis and which autosegmental phonology has successfully accounted for. According to Gol smit (1979:205) tone “stability re ers to t e resistance o t e tonal features of a vowel to deletion, even when the vowel that bore the tonal eatures is elete or esyllabi ie ”. For instance, autosegmental phonology places tonal features and vowels on separate and autonomous tiers such that the deletion of one does not affect the other and with the association conventions and well-formedness conditions in place the tone or the vowel is re-associated to the adjacent tier (http:// hum.uchicago.edu~jagoldsm/papers/autosegmentalunderspecificationwnoske.p df). University of Ghana http://ugspace.ug.edu.gh 22 “Floating tones” are also a e uately an better an le in autosegmental phonology than in linear generative phonology. A Floating tone is a tone whose existence is autonomous to major segments and yet has a status in the grammar of a language. Autosegmental representation has a way of dealing with the presence and behaviour of such floating elements, which consisted of autosegments existing purely on the tonal tier. (http://hum.uchicago.edu~jagoldsm/papers/autosegmentalunderspecificationwn oske.pdf.) Finally, autosegmental phonology makes it possible to account for phonological derivation in a more self-explanatory way with respect to phonological processes, especially in the step by step representation of the assimilatory processes. Word compounding in Dagara sometimes results in consonant nasalization, an assimilatory process which linear phonology has not been able to explain satisfactorily the derivational processes involved. For example, / z m+ / [z mm l ] ‘Fish + ‘small (pl) ‘small fishes Figure11 shows graphically, the step by step derivation of the assimilatory process that accounts for consonant nasalization of the above example in Dagara in word formation process. University of Ghana http://ugspace.ug.edu.gh 23 Fig. 10. Assimilatory Derivation Process Nasal Tier -N +N -N Skeletal Tier X X X + X X X X Segment Tier z m b l Nasal Tier -N +N -N Skeletal Tier X X X X X X X Segment Tier z m b l Nasal Tier -N +N -N Skeletal Tier X X X X X X X Segment Tier z m b l Nasal Tier -N +N -N Skeletal Tier X X X X X X X Segmental Tier z m l In the derivation above, the straight lines ( ) are the association lines that link segments and features on the different tiers via the skeletal tier. The broken lines ( ) show a progressive spreading of the Nasal feature onto the oral stop /b/. In the next stage we observe that, the oral segment /b/ is now fully associated to the Nasal feature and is delinked from the minus Nasal (-N) feature. The two parallel lines ( ) across the association lines signify the delinking of an association line from a segment either from a feature or the University of Ghana http://ugspace.ug.edu.gh 24 skeletal tier. For instance, the segment /b/ is delinked from the –N feature on the Nasal tier as well as from the skeletal tier. The arrow symbol ( ) shows the progression of the derivational process. 1.8.4 Tiers used in this Thesis Tiers are the multi-levels at which the different phonological features are arranged in Non-linear generative phonology. They usually consist of sequences of features which differ from one tier to the other. There are various kinds of tiers that can be specified for various phonological analyses. In this Thesis, the following tiers are applicable to the phonological representation: i. Sekeletal / CV Tier It is the mediating point on which the elements on the different tiers anchor. The skeletal/CV tier holds units represented by consonants and vowels; hence they are sometimes represented as C-slots and V-slots or X-slots. The skeletal/CV tier can relate with other tiers on one-to-one or one-to-many basis (cf section 1.7.2.2). ii. Segmental Tier This is the level at which phonetic representation relevant to the description are arranged. In other words, the underlying phonetic segments occupy this tier (cf. Fig. 10 above). iii. Place / Manner of Articulation Tier (P. A. /M. A. Tier) This tier specifies the Place or Manner of articulation features of the segment that are significant to the phonological analysis. This applies in University of Ghana http://ugspace.ug.edu.gh 25 the formalization of phonological processes such as Homorganic Nasal Assimilation, Labialization and Glide formation. iv. Tonal Tier The tonal tier specifies whether the underlying tone on the segment is high (H), mid (M), low, (L) or a contour tone (HL, LH). This is applicable in the discussion and representation of tone in the languages (cf. 8b above) v. Tongue Root Tier This tier spells out the tongue root position features of the phonological segments whether it is an Advanced Tongue Root feature (+ATR) or Unadvanced Tongue Root feature (-ATR). This tier is very relevant in the discussion of Vowel Harmony as a phonological process in Dagara. vi. Syllable Tier This tier indicates the units of sounds that make up a word. It is applicable in the discussion of the syllable type in Dagara and syllable structure processes such as elision, linking, and segment deletion. The CV-tier model of the syllable structure is used in this Thesis. Figure 11 below shows the structre syllable of the word gli ‘to unt . Fig. 11 Syllable Tier Syllable Tier ó ó CV Tier C V C C V SegmentalTier b g l University of Ghana http://ugspace.ug.edu.gh 26 1.9 Literature Review The amount of linguistic work on the phonological processes in any of the Dagaare dialects is not comprehensive. Early missionary and academic linguists works on Dagaare p onology were mainly on t e escription o t e segmental and suprasegmental phonemes with the focus on only the Central dialect. Kennedy (1966), a field work report on Dagaare phonology for instance, is mainly a description of segmental and suprasegmental phonemes. Bodomo (1997) is also mainly a description of the segmental and suprasegmental phonemes of the language. Bodomo (2000) however mentions in passing some phonolgical processes such as vowel harmony, vowel assimilation, and labialization, but with evidence limited to the Central dialect of Dagaare. Saanchi (1997) also discusses vowel harmony pointing out three types including ATR harmony, Rounding harmony, and Cross Height Vowel Harmony drawing data from the Central dialect to support his claim. Delplanque (1976) is a work on the phonological description of the Dagara language as spoken at Dissine (Burkina Faso). He attempts to satisfy both descriptive and interpretative requirements by developing both static conception of the language (as a system of features and units), and a dynamic conception (as a system of rules). He shows the relevance of the three kinds of unit or tactical level: the phoneme (the minimal distinctive unit), the syllable (the minimal prosodic unit), and the phonological word (the minimal lexical constituent) (Delplanque, ibid: 7). In deed, it is a detail work on the Dagara language with regards to the distinctive features classification of the segments as well as the prosodic features; the syllable and tone. But he has not discussed University of Ghana http://ugspace.ug.edu.gh 27 the phonological processes in Dagara and besides, his work is in French language which is only limited to francophone readers to the disavantage of the majority of the readers of anglophone background. Bemile (1983 and 1984) also follows a similar trend as he discusses Dagara segmental and suprasegmental phonemes. He posits that there are forty-five (45) segmental phonemes in Dagara. He justifies this claim and comes out with a list of six hundred and thirty (630) vowel phoneme oppositions and three hundred and fifty-one (351) consonant phoneme oppositions with examples drawn from Dagara dialect. Bemile (1985) discusses some assimilatory processes including; Vowel Harmony, Consonant Nasalization, Homorganic Nasal assimilation and syllable structure processes such as final Consonant and Vowel Elision as well as some Tonal processes. The analysis is however within the linear generative phonology framework and in German language which is not accessible and comprehensible to the many readers in general. The current thesis extends the discussion of the phonological processes to include Consonant Labialization, Rhotacism, Glide formation, and Resyllabification in Dagara. The formalization of the phonological processes in this thesis is within non-linear generative phonology theoretical framework, and in a language which is more accessible and comprehensible to the academia in Ghana, in particular, and other Anglophone countries in general. At theoretical level, there are some works on phonological processes and tone in cognate languages within the autosegmental framework which serve as good guide at the theoretical level. Akanlig-Pare (2005) discusses tone and some phonological processes such as re-syllabification, compensatory lengthening, University of Ghana http://ugspace.ug.edu.gh 28 vowel epenthesis and Vowel Harmony in Buli within the autosegmental framework. Agoswin (2010) also discusses a number of phonological processes in the Kusaal language including Nasalization, Homorganic Nasal assimilation, labialization, palatalization, vowel truncation, consonant deletion, aspiration and glottalization. He also discusses the tonology of the language. Though the phonological processes are discussed within linear generative phonology framework, autosegmental phonology is used for the tonal representation. Both Buli and Kusaal belong to the Mabia language group with Dagaare. Dagara may exhibit some kind of similarity at the level of the phonological processes and tonology with them, but it is anticipatory that the triggering phonological conditions will not be the same. There are other works on the analysis of phonological processes within the autosegmental framework in some Ghanaian languages which have been keenly studied for a review. These works include Bota (2002), Akrofi Ansah (2002), Asante (2009) and Frimpong (2009) among others. There are some fundamental errors in some of these works in terms of the use of the tenets of the theoretical framework in the phonological representations. Though the theoretical framework stresses that the skeletal tier is the mediating point on which the elements of the different tiers anchor, it is observed that in the works such as Akrofi Ansah (2002) and Frimpong (2009), the position of the skeletal tier is seriously inverted. University of Ghana http://ugspace.ug.edu.gh 29 Figure 12 highlights some examples of erroneous placement of the skeletal tier in the parallel arrangements of the various tiers in the works of Akrofi Ansah (2002) and Frimpong (2009). Fig. 12 . Wrongful Arrangement of Tiers a) Skeletal tier C V C V C V C V Segmental tier m ʥ i m ʥ i [ATR] tier [-ATR] [+ATR] [-ATR] [+ATR] Skeletal tier C V C V C V C V Segmental tier m ʥ i m i ʥ i [ATR] tier [-ATR] [+ATR] [+ATR] (Akrofi Ansah, 2002:48-49. Fig. 23). b) +ant cor Place of Articulation tier ŋ g o Segmental tier C C V Sekeletal tier [+nas] [-nas] Nasal tier (Frimpong, 2009: 22). In the arrangements of the tiers in Akrofi Ansah (2002), the skeletal tier is placed above the other tiers, whereas in Frimpong (2009), the skeletal tier is University of Ghana http://ugspace.ug.edu.gh 30 placed below the other tiers. In both cases, the skeletal tier is not assuming that mediating role on which the features and the segments anchor as the theory stipulates. It is rather the segmental tier that is mediating between the feature tier and the skeletal tier. This is a fundamental error which this thesis will avoid or correct. 1.10 Organization of the Work The Thesis is composed of five chapters. Chapter One consists of the introduction, Dagara and its speakers, a sketch on dialectal variation in Dagaare, the Problem Statement, Objectives and Significance of the study, Data collection and Methodology, Theoretical framework, Literature Review and Organization of the work. Chapter Two discusses the Dagara vowel inventory and vowel distribution, consonant inventory and consonant distribution, the phonetic description of the sounds, the syllable type and structure of word in Dagara. Chapter Three contains the phonological processes in Dagara. It discusses assimilatory processes including Vowel Harmony, Consonant Nasalization, Homorganic Nasal Assimilation, Rhotacism, Glide Formation, Labialization and Syllable Structure Processes such as Vowel Elision, Consonant Elision, Liaison (Linking), and Vowel Epenthesis. Chapter Four dicsusses Tone and the function of Tone in Dagara. Chapter Five is the summary of the thesis and concluding remarks. University of Ghana http://ugspace.ug.edu.gh 31 1.11 Chapter Summary In this chapter, we have given a brief statement on Dagara and its speakers with a sketch of dialectal variation in the language. We have also outlined the problem statement; the scope of the research, the aims and objectives of the research, the significance of the research as well as data collection and the methodology involved in the data collection is highlighted. We have also given an explanation of Autosegmental phonology with emphasis on how it evolved and the basic tenets on which it is applied. We also justified the motivation for using the framework in the phonological and tonal representations. A literature review and the organization of the thesis conclude the chapter. University of Ghana http://ugspace.ug.edu.gh 32 CHAPTER TWO DAGARA SOUND SYSTEM 2.0 Introduction The knowledge of the sounds of a language is very crucial for any phonological analysis. As Ladefoged (2003:1) rig tly points out “wit out t e knowle ge o t e soun s, one cannot escribe t e p onology o a language”. This Chapter therefore, deals with the inventory of Dagara vowel and consonant phonemes and their distributions. It also outlines the Distinctive Features (DFs) of the sounds since phonological processes are phonetically motivated based on articulatory and auditory systems. It then discusses the syllable and syllable types in Dagara as well as their combination in words. 2 .1 Vowels Vowels are produced without any obstructions of the airstream from the larynx to the oral cavity during their production. There are eighteen (18) vowels in Dagara (cf Bemile 1983, 1985). These vowels may broadly be categorized as oral and nasal vowels in terms of their places and manners of articulations. They may further be sub-categorized with regards to Tongue root position; Advanced Tongue Root / Unadvanced Tongue Root (+/-ATR). 2.1.1 Oral Vowels There are nine (9) oral vowels /i, , e, , u, , o, ɔ, a / in Dagara, unlike in the Central dialect of Dagaare which is attested to have ten (10) oral vovels (Saanchi 1997). Oral vowels are produced with the velum raised as high as the pharyngeal wall and thereby blocking the airstream in the lungs from passing University of Ghana http://ugspace.ug.edu.gh 33 through the nasal cavity in the course of their production. The airstream thus, escapes through the oral cavity. The chart below shows the places and manners of articulation of the oral vowels. Fig.13 Short Oral Vowel Chart From the chart above the oral vowels may be phonetically described as below: /i/ Close front unrounded advanced / / Close front unrounded unadvanced /e/ Close-mid front unrounded advanced /ɛ/ Open-mid front unrounded unadvanced /a/ Open central unrounded unadvanced /ɔ/ Open-mid back rounded unadvanced /o/ Close-mid back rounded advanced / / Close back rounded unadvanced u ɛ o ɔ Close-mid Close-mid Open-mid Open-mid Open Open Front Close Back Close e a i Central University of Ghana http://ugspace.ug.edu.gh 34 /u/ Close back rounded advanced Table 3. shows word examples of the short oral vowels. Table 3. Oral Vowel in word examples. Vowel Example Gloss. /i/ .r types / species ky millet / / t we, our, us d to eat, to burn /e/ d to take tome work / / ny to shit m r to stick on /u/ d to scrub, kul to marry/go out / / she/he/ it, him/her/, his/her b r milk /o/ d to climb ɓ to take off bark of a tree /ɔ/ yɔ to roam about bɔ to search for somethng /a/ kp to nail, to lock to stick something University of Ghana http://ugspace.ug.edu.gh 35 All the nine (9) oral vowels have their long counterparts. A long vowel is a short vowel whose sound duration is lengthened. Table 4 shows the examples of long oral vowels in words. Table 4 Long Oral vowels in words Vowel Examples Gloss /i:/ p :r to find something by chance v :r to go round / to wander about / :/ p :r to sweep m :l to mix / to clean /e:/ y : names p : stomachs / :/ y : penises n : mouths /u:/ g : to fail to/ not be ble to : alligator / :/ z :r tail b :r goats /ɔ:/ dɔ : man yɔ : brother /o:/ p :le girl f :li to whistle /a:/ p :r vagina d :r sticks Vowel length is contrastive in Dagara consequently, the short oral vowels contrast in meaning with their long vowel counterparts as the table below shows: Table 5. Short and Long Oral Vowels in contrast University of Ghana http://ugspace.ug.edu.gh 36 Vowel Example Gloss /a/ b friend /a:/ b : to grow /ɔ/ dɔ to weed /ɔ:/ dɔ : man /i/ ky ru pouring /i:/ ky :ru taboo / / b to ripen/cooked / :/ b : or /e/ gb to be blunt /e:/ gb : boats / / ny to shit / :/ ny : noses / / g r to pluck out / :/ g :r thorns /u/ g take care /u:/ g : fail to 2.1.2 Nasal Vowels All the nine (9) oral vowels have their nasal counterparts phonetically in Dagara.Similary, in the Central dialect all the oral vowels, except /e/, have their nasal counterparts (cf. Bodomo 1997, Saanchi 1997). This phenomenon distinguishes Dagaare from other Gur languages such as Kusaal and Buli, in which vowel nasalization occurs as a result of an oral vowel being in contiguous environment with a nasal consonant (cf. Agoswin 2010 and Akanlig-Pare 1994, 2005 respectively). In the production of the Nasal vowels, the velum is lowered such that the airstream from the lungs escapes through both the oral and the nasal cavities. University of Ghana http://ugspace.ug.edu.gh 37 The nine (9) nasal vowels, / , , , , ɔ /, have the same vowel qualities as their oral counterparts except for their spontaneous nasality. They are represented on chart in figure 14 below. Fig. 14 Short Nasal Vowel Chart Table 6. Short nasal vowels in words. Vowel Sound Examples Gloss / p to hide behind v shame / shyness / / I / my p in the least way / / gb forehead t t alone / / t to infiltrate s to roast / / v v wasp k : death / / v to expand ɔ Central Front Close Back Close Close-mid Close-mid Open-mid Open-mid Open Open University of Ghana http://ugspace.ug.edu.gh 38 k will not /ɔ / vɔ to pull out dɔ to grow lean / hatred / / gb to pile up kp to break open an egg / / p to compress something kp kp very hot Like their oral counterparts, all the nine (9) nasal vowsels in Dagara also have their long vowel equivalents. Since vowel lenght is contrastive in Dagara, the short nasal vowels also constrast in meaning with their long nasal vowel counterparts. The tables 7 and 8 show examples of long nasal vowels and short nasal vowels in contrastive distribution with their long nasal vowel counterparts in minimal pairs or near minimal pairs respectively in the language. Table 7. Long Nasal Vowels. Vowel Sound Examlpes Gloss / :/ p : an arrow / a key d : quiet / :/ b : broth / juice t : poison / medicine / :/ gb : on the forehead p : in the belly / :/ kpɛ : big / elder brother gb : on feet/ foot / :/ k : death University of Ghana http://ugspace.ug.edu.gh 39 v : fire / :/ b : n what is it? k : probably/ perhaps / /ɔ :/ tɔ : exactly gyɔ : very tall / :/ b : mournfully / :/ t : to erect s : father Table 8. Short andLong Nasal Vowels in contrast Vowels Example Gloss / / z to scar away / : / z : yesterday / / p to intercept / :/ p : arrow/ key / / z sit down / :/ z : blood / gb forehead / :/ gb : on the horehead / / b to seal / :/ b : slowly (movement) / / ny drink :/ ny : smell / / k will not / :/ k : probably / / p to shave / :/ b : mournfully /ɔ / tɔ to abandon /ɔ :/ tɔ : exactly University of Ghana http://ugspace.ug.edu.gh 40 These nasal vowels similarly contrast with the their oral vowel counterparts as well in Dagara. The table below shows some examples of minimal pairs or near minimal pairs of oral and nasal vowels in contrasts. Table 9. Oral and Nasal vowels in contrast. Vowels Example Gloss /a/ kp kp to lock repeatedly / / kp kp very hot /ɔ/ dɔ to weed /ɔ / dɔ hatred /i/ p to swell / / p to intercept / / do / / I /e/ gb to be blunt / / gb forehead / / s to sow / / s to roast / / p ɔ stomach / / p ɔ to rot /u/ d to scrub / / ny drink /o/ k to dry / / k to cry 2.1.3 Tongue Root Vowels All the eighteen (18) vowel phonemes, may further be sub-categorized into two main sets based on the position of the tongue root during their production, ie. University of Ghana http://ugspace.ug.edu.gh 41 Advanced Tongue Root position (+ATR) or Retracted/Unadvanced Tongue Root position (-ATR). The advance tongue root (+ATR) vowels are produced by pushing the tongue root forward, thus enlarging the pharyngeal cavity often raising the body of the tongue as well (Durand 1990:45). These vowels include / i, e, u, o, , , , /. The Retracted/Unadvanced Tongue Root vowels on the other hand, are produced with the tongue root pushed backward toward the pharynx thereby narrowing the space between the pharynx and the tongue root. The set of (- ATR) vowels include / , , , ɔ, a, , , , ɔ /. emile (1994) urt er uses t e terms “superior vowels” an “in erior vowels” to refer to the advanced tongue root (+ATR) vowels and the retracted tongue root (-ATR) vowels respectively. This dichotomy of vowels in Dagara is strictly adhered to in writing and in speech. There is a co-occurrence constriction on vowels of the two sets in the same lexical, mono-syllabic or simple multi-syllabic, non-compound word. Compounding process however sometimes voilates the harmonic system, since the it sometimes involves stems of different vowel qualities. In other words, there is a strict vowel harmony in a simple stem Dagara word, as it demands that only (+ATR) (“superior”) vowels co-exist with (+ATR) (“superior”) vowels in a simple Dagara word and (–ATR) (“in erior”) vowels co-exist with (–ATR) (“in erior”) vowels. The vowels /a, /, however, by virtue of the tongue root postion during their production, are considered as neutral and may co-occur with (+ATR) (superior) vowel(s) in some words but they are most commonly associated with the (– University of Ghana http://ugspace.ug.edu.gh 42 ATR) (inferior) set of vowels. The table below shows examples of the strict co- occurrence constrictions of the two sets of vowels in mono-syllabic or simple multi-syllabic non-compound words. Table 10. Tongue Root Harmony [ +ATR ] Gloss [ -ATR ] Gloss t r load t ɔ r mortar k r funeral k ɔ r gourd p farm p ɔ within gb b r toe t ɛ ŋk bl beard gbo gb ky l jackal n ɔ ɓɛ gl kɛ r sandals (slippers) 2.1.4 Diphthongs in Dagara This section outlines the possible ways in which vowels may combine in a simple Dagara word. We have already discussed with examples in section 2.1.1 and 2.1.2 that both oral and nasal vowels length is a consequent of two identical vowels in combination. We shall therefore discuss diphthongs, another type of vowel combination in Dagara. Westermann & Ward (1966) cited in Bemile (1985:46) defines a diphthong as follows: A diphthong is a gliding sound in which the tongue starts in the position of one vowel and immediately leaves it to glide towards another vowel position by the most direct route, without any diminution and subsequent reinforcement of the breath force. A diphthong strikes the ear as one syllable… (Westermann &War 1966: 43-44). University of Ghana http://ugspace.ug.edu.gh 43 In other words, a diphthong is a combination of two different vowels in a word that surfaces as a single syllable. In Dagara, when different vowels occur in a word, they must strictly adhere to the ATR harmony. The table below shows some possible different vowel combinations in Dagara words. It must however be mentoined that the list given here is not at all exhaustive. Bemile (1984) contains an elaborate list of diphthongs in contrast with practical illustrations in Dagara. Table 11. Diphthongs in Dagara Diphtong Example Gloss / ie / bie child / io / b g tomorrow / o / d ŋ a room / / t m chin / / d today / ɔ / s ɔ ŋ rainy season / ui / z iŋ because of.... / iu / d to chase / ue / k hoes / uo / y r name / ou / to regurgitate / / n dry season / / k to grow lean / oe / k e funerals / ɔ / n ɔ r mouths University of Ghana http://ugspace.ug.edu.gh 44 / ɔ / wɔ to bruise, sustain burns / ɔ / k ɔ water / eu / leu immediately / ɔ / pɔ oaths / ɔ / ɔ b saturated / / p basket / / k malt, types of malt / / s witch / / sɛ mat / ɔ / kɔ gourds / / v leaves / / m very red 2.2 Consonants Consonants are produced with various articulators and degrees of modulations of the airstream in the oral cavity. Ladefoged (1982) describes the production of consonants as radical obstruction in the mid sagittal region (the midline) of the vocal tract. There are twenty-seven (27) consonant phonemes; twenty-five (25) consonants and two (2) approximants (cf. Bemile 1983, 1985). These consonant phonemes are produced at eight (8) different places of articulations and seven (7) different degrees of airstream modulations in the oral cavity during their articulations. The consonant chart below shows their phonetic values ie. the places and manners of articulations of the phonemes. The horizontal axis shows the eight places of articulations and the vertical axis shows the seven manners of University of Ghana http://ugspace.ug.edu.gh 45 articulations. Where sounds are in pairs, those to the left are voiceless and those to the right are voiced. Singly occurring sounds are naturally voiced except the velar and glottal fricatives Table 12 Consonant Chart Bilabial Labial- Velar Labio- Dental Alveolar Alveo- Palatal Palatal Velar Glottal (Plosive)explosives Implosive p b ɓ kp gb t d k g Affricates ʨ ʥ Fricatives f v s z x h Nasal m ŋm n ɲ ŋ Laterals egressive Ingressive l 'l Trill r Glides/Approximants w j All the sounds in Dagara are represented in the orthography as they appear on the phonetic chart except the alveo-palatal affricates [ʨ] ,[ʥ] and the the University of Ghana http://ugspace.ug.edu.gh 46 palatal nasal [ɲ] which orthographically appear as , and in the written form, represpectly. The glottalized bilabial implosive /ɓ/, the glottalized ingressive lateral / 'l /, and the voiceless velar fricative /x/ are sounds only attested in Dagara and Birifor; the Central and Southern dialects of Dagaare lack them. The trill sound /r/ is prohibited at word-initial position though, it may occur at word-medial or secondary syllable-initial and word-final positions. All other consonants may occur at word-initial or word-medial position. But at word- final position, only seven consonants, inclu ing / b, g, m, n, ŋ, l, r / are exclusively allowed. Table 13 shows examples of the twenty-five consonants and the two glides distribution in words. Table 13. Consonant Phonemes Distribution in Words Sound Word initial Word medial Word final /p/ pɔ g ‘woman’ n b pɛgr ‘fingernail’ ------- /b/ b r ‘animal milk pɛ rb ‘buttocks ɛ b ‘man /ɓ/ ɓ r ‘to split ɓ ɓ ‘parboil ------------ /t/ t g ‘to dig w t ‘to suffer ------------- /d/ d ɛ n ‘to play b nd r ‘food -------------- /k/ kɔ r ‘to cough k gle ‘a stool ----------------- /g/ g g ‘a drum gl ‘earthen pot t ns g ‘gound /kp/ kp gtir ‘baboon n kpɛ ɛ ‘elder --------------- University of Ghana http://ugspace.ug.edu.gh 47 /gb/ gb b r ‘toe nɔ gb mb ‘lips -------------- /ʨ/ ky p r ‘brain daky n ‘wall ------------- /ʥ/ gy l ‘egg wɛ gya ‘cloth ------------- /f/ f ‘to sieze n ɔ ɓɛ gl kɛ r ‘Sandals ------------- /v/ v r ‘breathe v v ‘wasp -------------- /s/ s ‘father p sa ŋ ‘sorrow ------------- /z/ ɛl ‘tongue z z ŋ ‘bat -------------- /x/ x r ‘to yawn x n x n ‘sweat profusely --------- /h/ ha ‘to buy excessively helhel ‘to tear --------- /m/ m t k ɔ ‘tears tɛ mɛ ‘tobacco u m ‘fish ŋm/ ŋm m ‘dove n ŋm n ‘pigeon ------------ /n/ na ŋ ‘scorpion ba ndaa ‘lizard n ŋmin ‘God /ɲ/ nya a kp n ‘ghost nya a ny ‘cat --------------- /ŋ/ ŋ ‘this (one) kp ŋkp ŋnyu gb ‘shoulder a ŋ ‘body /l/ l l ‘bird w le ‘tree branc l ‘scar /'l/ 'lɛ ‘to enjoy ɛ pɛ 'lɛrɛ ‘spi er ------------- /r/ ------------- kɛ r ‘tall ɛ gr ‘ irt /w/ w gr ‘to bark w w ‘confess repeatedly ----------- University of Ghana http://ugspace.ug.edu.gh 48 /j/ y n ‘year yiryel ‘ amily matter ------------ 2.3 Phonetic Description of Dagara Sounds This section discusses the inherent properties of Dagara sounds and groups them into natural classes in consonance with the Distinctive feature theory, based on the non-hierarchical binary feature system, outlined in Chomsky and Halle (1968) and in the works of Katamba (1989), and Durand (1990). An inventory of the sounds of a language alone does not offer enough bases to make general statements about the nature and behaviour of sounds for p onological analysis. T e knowle ge o “t e basic p onological ingre ients, calle Distinctive Features, w ic p onemes are ma e up o ” (Katamba, ibi : 35), is very crucial for phonological analysis of any language and for that matter, the description of Dagara. Distinctive features (DFs) are minimal contrastive units that enter into the composition of sounds. They are in effect, the building blocks of sounds (Akanlig-Pare 1994:18). Oduro Kwarteng (2009:11) following Katamba (1989) also views distinctive features as the acoustic and articulatory features that seek to group sounds into natural classes based on their shared features. But in my view, they are the lenses through which the internal structure of a sound can be defined and distinguished from other sounds, thus making it possible to make general statements about the nature and behaviour of sounds. Some distinctive features are binary in terms of the presence (+) of the feature or absence (-) of the feature in the sound. Whilst others are unary features because they have a single value and specify only sounds that have them. University of Ghana http://ugspace.ug.edu.gh 49 Sounds whose productions involve only the active articulators are tagged with the unary feature. The sounds are classified and discussed under; Major class features, Cavity features, Tongue Body features, Manner features and Laryngeal features. 2.3.1 Major Class Features These features define the major classes of sounds that are relevant in phonological analysis. These features sub-divide speech sounds into vowels, consonants, obstruents, sonorants, glides, and liquids. We will classify the sounds in Dagara under the three major class features; Consonantal / non- consonantal, Syllabic / non-syllabic, and Sonorant / non-sonorant, guided by the definitions given in Durand (1990). 2.3.1.1 [ +/-Consonantal ] Sounds Consonantal sounds are produced with a constriction in the vocal tract at least equal to that found in the fricative consonants; non-consonantal sounds are made without such constriction. Obstruents (plosives, fricatives, and affricates), nasals, and liquids are [+cons]. The [-cons] class includes the vowels, the glides (a category made up of the semi- vowels) and the glottal sounds [ʔ] and [h] (Durand, ibid: 42). However, the consonantal classification include / p, b, ɓ, t, , k, g, kp, gb, ʨ, ʥ, f, v, s, z, x, h, m, n, ɲ, ŋ, ŋm,l, 'l, r, w, j / in Dagara. The [-Cons] sounds include the 18 vowels only. 2.3.1.2 [+/-Syllabic ] Sounds “Syllabic sounds are those which constitute peak of syllables, non-syllabic sounds are those w ic are at t e margins o syllables” (Duran , ibi : 41). T e University of Ghana http://ugspace.ug.edu.gh 50 sounds that constitute peak of syllables in Dagara, ([+syllabic] sounds), include the eighteen vowels; / i, , e, ɛ, u, , o, ɔ, a, , , , ɛ , , , , ɔ / and the liquids / r, l / at coda position. All other sounds including the stops, nasals, and glides are always at the margins of syllables, thus [-sylla]. 2.3.1.3 [ +/-Sonorant ] Sounds Sonorant sounds are produced with a vocal tract configuration sufficiently open for the intra-oral air pressure to be approximately equal to the ambient air pressure. By contrast, obstruents are produced with a constriction sufficient to generate intra-oral pressure much greater than that of the surrounding air (Durand, ibid: 42). The [+son] soun s in Dagara inclu e t e vowels / i, , e, ɛ, u, , o, ɔ, a, , , , ɛ , , , , ɔ /, the nasals / m, n, ɲ, ŋ, ŋm/, t e liquids / l, 'l, r /, and the glides / w, j /. The [-son] sounds include the stops, fricatives, affricates and the glottal stop. 2.3.2 Cavity Features These features are associated with the places of the articulation of the sounds. They specify where, in the vocal tract, the modifications of the airstream take place during the production of particular sounds (Katamba, 1989: 43). The features relevant for the description of Dagara sounds are under two broad areas; Primary stricture which include Coronal/non-coronal, Anterior/non- anterior, Labial/non-labial; /Round/non-round, and Tongue-body features which include High/non-high, Low/non-low, Back/non-back/Dorsal and Advanced/Unadvanced Tongue Root (-/+ATR). Their definitions are based on what Katamba (1989) as given ollowing t ose in omsky an Halle s S PE. University of Ghana http://ugspace.ug.edu.gh 51 2.3.2.1 Primary stricture 2.3.2.1.1 [ Coronal ] Sounds “All sounds produced with the blade of the tongue raised towards the front teeth, the alveolar ridge and the hard palate are designated with this feature. For non-coronal consonants, the blade of the tongue remains in a neutral position” (Katamba, ibi : 44). T is feature is a unary feature as the active articulator, ie. the blade of the tongue, is a reference point for its specification. Dagara sounds with this feature specification, ie. coronal sounds, include / t, d, ʨ, ʥ, s, z, n, ɲ, l, 'l, r, j, i, , u, /. Non-coronal soun s inclu e; / p, b, ɓ, k, g, kp, gb, (ʔ), , v, x, , m, ŋm, ŋ, w / an t e rest o t e vowels. 2.3.2.1.2 [ +/- Anterior ] Sounds The anterior sounds are produced from the alveolar ridge forward. Katamba (1989) distinguishes the production of anterior and non-anterior sounds as follows: In the production of anterior sounds, the main obstruction of the airstream is at a point no farther back in the mouth than the alveolar ridge; for non-anterior sounds the main obstruction is at a place farther back than the alveolar ridge. Labials, dentals, and alveolar are anterior while all other sounds are not (Katamba, ibid: 44). The [+ant] sounds in Dagara thus, include / p, b, ɓ, t, d, m, n, f, v, s, z, l, 'l, r, w /, while all other sounds are [-ant]. 2.3.2.1.3 [ Labial ]; [ +/- Round ] Sounds These features are attributed to sounds whose productions involve a varying degree of the airstream obstruction at the lips or the protrusion of the lips. University of Ghana http://ugspace.ug.edu.gh 52 These features draw consonants and rounded vowels into one category; the Labial feature is a consonant specification, whereas Round feature is a vowel specification. Katamba (1989:44) points out that, there is a considerable degree of overlapping between sounds covered by these two features. Labial sounds include; Bilabial, Labial-velar and Labio-dental consonants as well as rounded vowels. All other sounds are non-labial/ [-round], (Katamba, ibid: 44). The sounds in Dagara that fall under these features are / p, b, ɓ, kp, gb, , v, m, ŋm, w / and Rounded vowels / o, ɔ, u, , , ɔ , , /. All other sounds are neither Labial nor Round. 2.3.2.2 Tongue Body Features The features discussed in this section relate to the positions of the tongue body and the tongue root relative to their neutral positions. The neutral position of the body of the tongue is said to be the position which it assumes in the production of a mid front vowel (Katamba, ibid: 45). The features under consideration are; High/non-high, Low/non-low, Back/non-back and Advanced/non-advanced tongue root. 2.3.2.2.1 [ +/- High ] Sounds “Hig soun s are ma e wit t e tongue [body] raised from neutral position while non-high sounds are made without such raising of the body of the tongue” (Katamba, ibi : 45). T is eature covers bot consonants an vowels including alveo-palatal, palatalized, palatal, velar consonants, glides and high vowels. Dagara sounds in this category include / ʨ, ʥ, ɲ, j, w, k, g, kp, gb, x, ŋ, i, , u, , , , , /. All other sounds are [-high]. University of Ghana http://ugspace.ug.edu.gh 53 2.3.2.2.2 [+/- Low] Sounds “Low sounds are produced with the tongue depressed and lying at a level below that which it occupied when at rest in neutral position; non-low sounds are produced wit out epressing t e level o t e tongue in t is manner” (Katamba, ibid: 45). This feature specification covers open vowels and pharyngeal consonants. All other sounds are non-low. Guided by the description above, sounds that are [+low] are the open vowels / a, / since there are no pharyngeal consonants in Dagara. All other sounds are [-low]. 2.3.2.2.3 [+/- Back ]; [ Dorsal ] sounds The features [Back] and [Dorsal] are virtually the same but whereas the feature [Back] is binary and a vowel specification, [Dorsal] is a unary feature for consonant specification. Sounds produced with the body of the tongue retracted from neutral position are [+back] or [Dorsal]. Sounds produced with the body of the tongue either in neutral position or pushed forward are non-back (Katamba, ibid: 46; SPE: 305). The sounds that are described as [+back] in Dagara include back vowels, / u, , o, ɔ, , , , ɔ /; the low vowel / a/ and the consonants that are specified for the feature [Dorsal] are the velar consonants / k, g, ŋ, x /. The Labial Velar sounds / kp, gb, ŋm, w / also take part of their specification from the feature [Dorsal] and from the feature [Labial]. All other sounds are not specified for the feature [ +Back] or [Dorsal]. University of Ghana http://ugspace.ug.edu.gh 54 2.3.2.2.4 [ +/- Advanced Tongue Root ] Sounds This is exclusively a vowel feature which pertains to the movement of the tongue root during their productions. In the production of Advanced Tongue Root [+ATR] vowels: the tongue root is pushed forward, thus expanding the resonating chamber of the pharynx and possibly pushing the tongue body upward; Unadvanced Tongue Root [-ATR] vowels are produced with the tongue root at neutral position (Katamba, ibid: 47). T is eature ivi es t e vowels o Dagara into two istinct sets. T e set o AT vowels are / i, e, u, o, , , , /. The set of [-ATR] vowels are / , ɛ, , ɔ, a, , ɛ , , ɔ , /. 2.3.3 Manner Features The manner features characterize the degree and kinds of modulation that the airstream undergoes in the oral cavity as it comes into contact with the articulators during the production of sounds. These features include Continuants/non-continuants, Lateral/non-lateral, Nasal/non-nasal, Strident/non-strident and Delayed Released/ Instantaneous Release. 2.3.3.1 [ +/- Continuant ] Sounds Continuant sounds are produced with a continuous airflow in the oral cavity. Durand (1990: 51) explains that continuant sounds are produced with a primary constriction which allows the air to flow through the mid-sagittal region of the vocal tract, while sounds produced with a sustained occlusion are non- continuant. The [+ cont] sounds in Dagara include / f, v, s, z, x, h, l, 'l, r, w, j, / and all the vowels, while the [-cont] sounds are the stops and affricates. University of Ghana http://ugspace.ug.edu.gh 55 2.3.3.2 [ +/-Lateral ] Sounds “A lateral soun is pro uce i t e air low t roug t e centre o mout is blocked and air only escapes over or both sides of the tongue. In non-lateral soun s air lows t roug t e centre o t e mout ” (Katamba, ibi : 50). The [+lateral] sounds in Dagara are / l, 'l, /. All other sounds are [-lateral]. 2.3.3.3 [ +/- Nasal ] Sounds In the production of a nasal sound the velum is lowered to allow air to escape through the nasal cavity. Oral sounds are produced with the velum raised so as to block access to the nasal cavity and to allow air to go out only through the mouth (Katamba, ibid: 50). Nasal sounds in Dagara include the nasal stops / m, n, ɲ, ŋ, ŋm / and nasal vowels / , , , ɛ , , , , ɔ /. 2.3.3.4 [ +/- strident] Sounds Sounds with the feature are acoustically characterized by a more random noise than nonstrident sounds. Only fricatives and affricates can be strident (Katamba, ibid: 50). Thus, strident sounds in Dagara include / f, v, s, z, x, h, ʨ, ʥ/. 2.3.3.5 [+/- Del. rel ] Sounds The feature delayed release is a tag on sounds produced with a complete closure of the vocal tract and a gradual release of the airstream. Chomsky and Halle (1968: 318) explain that during the delayed release, turbulence is generated in the vocal tract so that the release phase of the affricate is acoustically quite similar to the cognate fricative. The instantaneous release is University of Ghana http://ugspace.ug.edu.gh 56 normally accompanied by much less or no turbulence. Sounds with this feature in Dagara are the affricates, / ʨ, ʥ /. The rest of the sounds are instantaneous release. 2.3.4 Laryngeal feature 2.3.4.1 [ +/- voiced ] Sounds “Sounds produced with vibration of the vocal cords are voiced; voiceless sounds are produced with a glottal opening so wide that it will prevent vocal vibration i air low t roug it” (Duran , ibi : 54). oice soun s in Dagara include the vowels, the glides / w, j /, the liquids / l, 'l, r /, the nasals / m, n, ɲ, ŋ, ŋm / and the voiced obstruents / b, ɓ, d, g, ʥ, v, z /. 2.4 The Syllable Much as the value of the syllable cannot be overemphasized in phonological analysis, it is, in t e wor s o Kenstowic an Kissebert (1979), “probably t e most elusive o all p onological an p onetic notions” (Kenstowicz and Kisseberth, ibid: 256). Goldsmit (1990) points out une uivocally t at “t e syllable is a unit of phonological description which has never ceased to be discussed at length in t e p onological literature o t is century” (Goldsmith, ibid: 103). There have been many controversies over the years concerning the definition and nature of the syllable. Various linguists have given various definitions from their individual perspectives, a few of which we will review to enable us attempt establishing the syllable structure in Dagara. Phonetically, Roach 1991 describes the syllable as follows: University of Ghana http://ugspace.ug.edu.gh 57 .......... as consisting of a centre which has little or no obstruction to airflow and which sounds comparatively loud; before and after that centre (that is, at the beginning and end of the syllable), there will be greater obstruction to the airflow and/or less sound (Roach, ibid: 67) Inferring from this definition, the centre of the syllable is nothing less than a vowel, while the peripheries of the syllable are consonants. For instance, the Dagara word n ‘skin as /a/, produced with little or no obstruction of airflow in the centre; whilst /g/and /n/, are produced with relatively greater obstruction of airflow. From the phonological view point, several definitions have been postulated. Laver (1994: 114) gives a phonological definition of the syllable as “a complex unit made up of nuclear and marginal elements.” T e nuclear elements by is estimation are vowels and syllabic consonants such as nasals and liquids and the marginal elements are consonants. In the case of Dagara, the nuclear position is exclusively for vowels since there are no syllabic consonants in the language. The word z l ‘tongue is a Mono-syllabic, which consists of the vowel / / as its nuclear element and /z/ and /l/ as its marginal elements. levins (1995:20) also consi ers t e syllable as being “t e p onological unit which organizes segmental melodies in terms of sonority” wit explanation that the sonority of a sound is its loudness relative to other sounds produced with the same input energy. In Katamba (1989: 153), the syllable is described as “t e unit in terms o w ic p onological systems are organi e ”, w ile Kenstowicz (1994: 250) says that it is “an abstract unit o proso ic organi ation t roug w ic a language expresses muc o its p onology.” Gol smit (1990: 108) owever paints a University of Ghana http://ugspace.ug.edu.gh 58 clearer picture of what the syllable is by categorically stating that the syllable is a phonological constituent composed of zero or more consonants, followed by a vowel, and ending with a shorter string of zero or more consonants. There is however a unanimous consensus that, the internal structure of the syllable is made up of an obligatory nucleus (N), an optional consonantal onset (O), and coda (C), where the onset and the coda are the marginal elements and are filled by non-syllabic sounds, whilst the nucleus is occupied solely syllabic sounds. There are equally different views concerning the structural representation of the internal structure of the syllable in the literature. Some of the competing models of representation of the internal structure of the syllable in phonological analysis are the Onset Rhyme structural model, the Moraic structural model and the CV structural model. The Onset and Rhyme structural model proponents, (Pike and Pike 1947; Halle and Vergnaud 1978; Fudge 1969 etc.), posit that syllable is composed of a binary intermediary branching with; Onset and Ryhme, with the Ryhme further branching into Nucleus and Coda. The structure adapted from Blevins (1995) illustrates this model. Fig. 15 Onset and Rhyme Model Structure. ó Onset Rhyme Nucleus Coda University of Ghana http://ugspace.ug.edu.gh 59 Proponents of the Moraic structural model (Hyman 1985; Hayes 1989 etc.), relate the syllable-internal structure to syllable weight; monomoraic (light) and bimoraic syllable, and posit that initial consonants do not contribute to the syllable weight, thus they are attached directly to the syllable node, whilst vowel and final consonants are moraic simce they contribute to syllable weight. The structure below illustrates this model. Fig. 16. Moraic Structural Model ó µ µ p ɔ g The proponents of the CV structural model (Anderson 1969; Khan 1976; Clements and Keyser 1983 etc.) posit that the syllable is a flat structure without subconstituent but the segments themselves. Katamba (1989:156) explains that the CV-model assumes that, the syllable is assumed to be a Three-Tier structure consisting o a yllable No e ‘ó ; a -Tier, whose C and V elements dominate consonantal and vowel segments and a Segmental Tier consisting a of a bundle of distinctive feature matrices, which represent consonant and vowel segments. The V element of the CV-model represents a syllable Nucleus while the C element represents a syllable Onset or Margin. In this work, the CV-model expounded in Clements and Keyser (1983) is adapted in representing the various syllable structures in Dagara because it is University of Ghana http://ugspace.ug.edu.gh 60 more convenient model for the representation of double sounds than the other models. 2.4.1. Syllable Types in Dagara Dagara belongs basically to the language group of V, VC, CV, CVC and C syllable typology. These basic syllables however are modified into various syllable shapes. 2.4.1.1. The V syllable type The syllable nucleus may stand alone as syllable either as short vowel (V) or a long vowel (V:) in Dagara. Its occurrence is, however, limited to syllable initial position and in few words. Almost every vowel can stand independently as a syllable. Examples of this shape of syllables are given in table 14 below. Table 14. The V/V: Syllable Shape Word Gloss definite article I/my S/he, her/his : bury : yes i.ri spicies The structural representation of the V/V: is shown below in figure17 below: University of Ghana http://ugspace.ug.edu.gh 61 Fig. 17. Structural Representation of V/V: Syllable Shape a) V syllable shape b) V: syllable shape Syllable Tier ó ó CV Tier V V V Segment Tier u : ‘t e u :‘bury Syllable Tier ó ó CV Tier V C V Segmental Tier r .r ‘spicies’ 2.4.1.2 The VC syllable type This syllable type appears in words in the following shapes: VC, V:C, and VVC,. All the vowel may occupy the V position in these syllable shape in Dagara. Examples of words of these syllable shape attested in Dagara are given in the tables below: Table 15. The VC/V:C syllable shape Word Gloss r get up b behaviour r to surprise University of Ghana http://ugspace.ug.edu.gh 62 ɔ :r type of fruit :l horn The representation in the CV-model structure in figuer 18 Fig. 18. Structural Representation of VC/V:C Syllable Shape ii) VC syllable shape ii) V:C syllable shape Syllable Tier ó ó CV Tier V C V V C Segmental Tier r : l r ‘get up :l ‘ orn 2.4.1.2.1 The VVC syllable shape This syllable shape is not attested in other dialects of Dagaare. Vowels that occupy the V position in this syllable shape are different vowels in a permitted sequence. Examples of the VVC syllable shape is attested in Dagara word are in table 16 below: Table 16. The VVC syllable shape Word Gloss ɛ b to pinch ɔ b saturated ag fly a ŋ body o n dry season The structural representation of this syllable shape is shown in figure 19 below. University of Ghana http://ugspace.ug.edu.gh 63 Fig 19. Structural Representation of VVC Syllable Shape Syllable Tier ó ó CV Tier V V C V V C Segmental Tier ɛ b a g ‘to pinc ‘ ly 2.4.1.3 The CV syllable type The CV syllable is a universal phonological notion. Blevins (1995:220) declares that all languages have CV syllables in the same way as Draga (2007) who posits that it figures in all language-specific inventories and has the status as the least marked syllable type. The CV type also appears in words in the following shapes: CV, CV: and CVV. 2.4.1.3.1 The CV/CV: syllable shape A majority of Dagara words are of this shape. All consonants and vowels can fill the positions of the CV slots in these syllable shapes. Examples of words containing these syllable shapes are given below. Table 17. The CV/CV: Syllable Shape Word Gloss Word Gloss head s .b .k .b hail gbe forehead t arrive tɛ .mɛ tobacco kɔ to cultivate University of Ghana http://ugspace.ug.edu.gh 64 b : river t : medicine ku : cradle p : arrow Figure 20. Structural Representation of CV/ CV: Syllable Shape a) Syllable Tier ó ó ó ó CV Tier C V C V C V C V Segmental Tier s b k b .b .k .b ‘ ail b) Syllable Tier ó ó CV Tier C V C V Segmental Tier t ɛ m ɛ t .m ‘tobacco c) Syllable Tier ó CV Tier C V V Segmental Tier p i: : ‘arrow’ 2.4.1.3.2 The CVV syllable shape . The vowels in this sequence that occupy the V-slot in CVV syllable shape are constrained by tongue root vowel harmony. In other words, the vowels in sequence must either belong to the set of [+ ATR] or [-ATR]. Some examples University of Ghana http://ugspace.ug.edu.gh 65 of words composed of this syllable shape include the following in table 18 and the illustration of the structural representation in figure 21. Table. 18 The CVV syllable shape Word Gloss p ɛ basket pie ten kɛ malts t carry s ɔ knife Fig. 21. Structural Represention of CVV Sylllable shape. Syllable Tier ó ó CV Tier C V V C VV Segmental tier p ɛ t p ɛ ‘basket t ‘carry 2.4.1.4 The CVC syllable type This syllable type is the next preferred after the CV type. It manifests in many words in Dagara in the following shapes: CVC, CVVC, and CV:C,. The V-Slot may be occupied by any vowel, but the coda C-Slot is exclusively reserved for word final consonants which comprise of / b, g, m, n, ŋ, l, r /. University of Ghana http://ugspace.ug.edu.gh 66 2.4.1.4.1 The CVC syllable shape This is yet another least marked syllable shape which is attested in many words in Dagara. Examples of words of this syllable shape are in table 19 below. Table 19. The CVC Syllable Shape Word Gloss Word Gloss t ŋ mountain t g pull b l weak pɔ b beat b r breast ny m you g n skin/book The structural representation of this syllable shape is illustrared in figure 22. Fig 22. Structural Representation of CVC Syllable Shape Syllable Tier ó ó CV tier C V C C V C Segnental Tier p ɔ b g n a) pɔ b ‘beat b) g n ‘skin/book 2.4.1.4.2 The CVVC/ CV:C syllable Shape The vowels in sequence in CVVC syllable shape are also constrained by tongue root vowel harmony. Some examples of words with this syllable shape are in table 20 and and the CV-model representation in figure 23 respectively below: Table 20. The CVVC/CV:C syllable shape Word Gloss Word Gloss n ɔ r mouth p :r rock p r back k :r hoe University of Ghana http://ugspace.ug.edu.gh 67 y r name ɓ :.r cold t r load g :r thorns t ɔ r mortar g :l reared Fig. 23 Structural Representation of CVVC Syllable Shape Syllable Tier ó ó CV Tier C V V C C V V C Segmental Tier n ɔ r g : r n ɔ r ‘mout g :r ‘thorns 2.4.1.5 The C syllable type This syllable shape is attested only in few disyllabic simple words or compound words with two final consonants in sequence. The only consonants that are syllabic are the liquids / l, r / and they are restricted to word final or Coda positions. The following data in table 21 show words with syllabic consonants at final position. Table 21. The C syllable shape Word Gloss b. cocodiles ɔ b.r painful/ chewing og.l dented/ oblong g.r separate University of Ghana http://ugspace.ug.edu.gh 68 sa .s g. grasshopper This syllable shape is structurally representd in CV-model as in figure 24 below. Fig.24. Structural Representation of C Syllable Shape Syllable Tier ó ó ó ó CV Tier V C C V C C Segmental Tier b g r ‘crocodiles ‘seperate 2.4.2 Dagara Word Structure A word is composed of one or more syllables; therefore, there is no gainsaying that the structure of the syllable and that of the word are basically related. Hyman (1975) best describes this relationship in the following words: …..the sequential constraints which operate at the beginning of a word should be operative at the syllable, even if this syllable is word-initial. Similarly, the same sequential constraints which operate at the end of a word should be operative at the end of a syllable…. (Hyman, ibi : 189). There is a direct correlation between the syllable shape and word structure in Dagara in consonance wit Hyman s explanation, since the sequential constraints that operate at the beginning and end of the syllable are also in the word. The words in Dagara may be categorized as simple stem words and complex stem words. University of Ghana http://ugspace.ug.edu.gh 69 2.4.2.1 Simple Stem Words Simple stem words in Dagara range mono-syllabic to quadri-syllabic forms which correspond directly to the different shapes of the syllable discussed above. 2.4.2.1.1 Mono-Syllabic Words Mono-syllabic words reflect directly the basic syllable shapes. Examples of mono-syllabic words taking the forms of the various syllable shapes are in table 22. Table 22. Mono-syllabic Simple Words Syllable shape word gloss V to do V: : to bury VC ɔ :r type of a fruit V:C :l horn VVC o n dry season VCC br crocodiles CV p to swell CV: n : cattle CVV b ɔ goat CVC kpeŋ strong CV:C p :r rock CVCC b gr sacrifice University of Ghana http://ugspace.ug.edu.gh 70 2.4.2.1.2 Disyllabic words Words in this category consist of two different syllable shapes in various permutations. Some examples of disyllabic words in different combination of the basic syllable shapes are in table 23. Table 23. Disyllabic Simple Words Syllable shape Word Gloss V.CV .r spicies V:.CV :.r dust VC.VC og.f separation CV.CVV l .b money CV.CV l .le chicken CV.CVC b .b r day CVC.CVC b g.loŋ armpit 2.4.2.1.3 Trisyllabic words Words in this category consist of three syllable shapes in differnt permutations. Some examples of trisyllabic words in different combinations of the basic syllable shapes are in table 24. Table 24. Trisyllabic Simple Words Syllable shape Word Gloss CVC.CV.CV m.b .lu fooliness CVC CV.CV yel. m . ŋ truth CVC. CV.CV m. p . g light (weight) CV.CV.CV kpa . k . r hard (sufface) CV.CVC.CVC b .sɛ n. ɛr cheese CV.CV.CVC gbɛ .n .ɓɛ l foot CV.CV.CV . v . r court yard University of Ghana http://ugspace.ug.edu.gh 71 CV.CV.CV b .b .le child CVC.CV:.CVV k r.n .m ɔ tortis (sea) CVC.C.CVC bag.r.ŋm n collar bone 2.4.2.1.4 Quadri-syllabic words Words in the category consist of four syllable shapes in differnt permutations. Some examples of possible quadri-syllabic words in different combination of the basic syllable shapes are in table 25. Table 25. Quadrisyllabic Simple Words Syllable shape Word Gloss CV.CV.CV.CV k . l .k . l round CV.CV.CV.CVV gbɛ .s .l .p ɔ sole (foot) CV.CVC.CV.CV gbo .gb r.kyi.le jack CV.CV.CV.CV ɛ . pɛ .'lɛ.rɛ spider CVC.CVC.CV.CVVC kpa ŋ.kpa ŋ.nyu .gb l elbow 2.4.2.2 Complex Stem Words They are usally compounds or phrasal words which consist of a concantenation of stems of different syllable shapes or stems and suffixes to designate a concept. The complex stem words are words which usually undergo some kind of morphophonological processes. The combinations of these stems or/and suffixes, usually trigger some assimatory or syllable structure processes which are duely discussed in Chapter Four. Some examples of complex stem words are given in table 26 below. Table 26. Complex Stem Words. Underlying Form Compound Word University of Ghana http://ugspace.ug.edu.gh 72 CVVC CVC.CV CVVC.CVC.CV / t ɛ m / + / k b.l / [ t ɛ ŋ.k b.l ] ‘chin ‘ air ‘ bear VVC CVC CVC.CV VVC.CVC.CVC.CV / a ŋ/ / g ŋ / + / k b.l / [ a ŋ.g ŋ.k b.l ] ‘bo y ‘skin ‘ air ‘bo y air CVV CVC CVC CVC CVC.CVC.CVC.CVC / m ɔ / + / b ŋ / + / y ŋ / + / b / [ m ɔ .b ŋ.y ŋ.b ] ‘grass ‘ know ‘sense ‘see s ‘w eat CV CVC CV CV CV CV.CVC.CV.CV.CV / gb / + /gb r + / tɔ / + / / + /-ru / [ gb . gb r. tɔ . . ru ] ‘pile ‘rus ‘ it ‘ ea -PL ‘ ake oor ‘rus an it ea s CV CV CVCC CV CVV CV CV.CV.CVCC.CV.CVV.CV /b / + / k / + / m gr / +/ b /+ / w ɔ /+ /m / [b . k . m gr. b w ɔ . m ] ‘ rien ‘with ‘rope ‘is ‘sack ‘in ‘ a rien wit as rope in sack (‘ ake rien ) 2.5 Chapter Summary The chapter discussed the sound system of the Dagara dialect with emphasis on the consonant and vowel inventory, the phonetic description of the sounds, the syllable structure and the structure of words in Dagara. It has pointed out that, there are eighteen vowel phonemes; nine short oral vowels and nine short nasal vowels, all the eighteen short vowels have their long counterparts which contrast, and that they are subcategorized into Advanced Tongue Root (+ATR) vowels and Unadvanced Tongue Root (-ATR) vowels. It also pointed out that, there are twenty-seven consonant phonemes; twenty- five consonants and two approximants, the distribution of the consonant University of Ghana http://ugspace.ug.edu.gh 73 phonemes is constrained; whereas all consonants can be at word-initial and/or word-medial, word-final position is exclusively occupied by / b, g m, n, ŋ, l, r /, and that there are no syllabic consonants in Dagara. It further, within the Distinctive Feature theory framework described and classified the Dagara sounds into natural classes, under Major class features, Cavity feature, Manner features and laryngeal feature. It then examined the syllable structure, highlighting that Dagara belongs to languages with V, VC, CV, CVC and C syllable typology, with a various syllable shapes. Finally, it examined the structure of words in Dagara and observed that, the sequential constraints that operate at the beginning and end of the syllable are also in the word and that, the words in Dagara may be classified as simple stem or complex stem words. University of Ghana http://ugspace.ug.edu.gh 74 CHAPTER THREE PHONOLOGICAL PROCESSES IN DAGARA 3.0 Introduction Phonological processes refer to the changes that take place in sounds when segments are juxtaposed. According to Wolfgang (1984:31) “phonological processes serve the communicative function of language by serving their proper functions: pronunceability and perceptibility.” Phonological processes are universal and phonetically motivated based on articulatory and auditory systems. Though these universal processes may apply in all languages, eac linguistic community ‘selects a set o processes, (http:// www.ling.hawaii.edu/faculty/donegan/papers/201Xhistphon.pdf). This chapter, thus, discusses the phonological processes that are specific to Dagara, a dialect of the Dagaare language. It discusses both assimilatory and syllable structure processes and formalizes the various phonological processes within autosegmental representation. 3.1 Assimilatory Processes A phonological process is called assimilation, if as a result of its application two or more segments in form agree in their value for some phonological feature(s) or feature class(es) ( akovi , 2007:335). Katamba (1989: 80) further University of Ghana http://ugspace.ug.edu.gh 75 explains that, assimilation is the modification of a sound in order to make it more similar to some other sound in its neighbourhood, with the aim of making a smoother, more effortless, more economical transition from one sound to another. When a sound is modified to look more like the sound that precedes it, the assimilation is in a progressive/preservative direction. On the other hand, when the sound is modified to look more like the sound that follows it, the assimilation is in a regressive/anticipatory direction. In autosegmental phonology, assimilation is construed as an autosegmental spreading along with a hierarchical feature representation. Adherents of autosegmental phonology, such as Sagey (1986) , advocate that autosegmental spreading is a better representation of the process of assimilation than the linear phonology representation of assimilation, as a changing individual feature in feature matrix, because of the relative simplicity of describing the derivations of the assimilation processes to reflect their relative naturalness (Sagey, ibid: 10). The directions of the autosegmental spreading are described, as left to right spreading and right to left spreading, in tandem with the progressive/preservative and regressive/anticipatory directions, respectively in linear phonology. The assimilatory processes discussed in this thesis include; Vowel Harmony, Consonant Nasalization, Homorganic Nasal Assimilation, Glide Formation, Labialization and Rhotacism, within the autosegmental phonology representation. University of Ghana http://ugspace.ug.edu.gh 76 3.1.1 Vowel Harmony The phenomenon of vowel harmony has been explained by various linguists from the individual perspective. According to Goldsmith (1990:304) a vowel harmony system is one in which the vowels of a language are divided into two (or more) (possibly overlapping) subsets with the condition that all vowels in a given word (or domain, generally) must come from a single subset. Kenstowicz (1994) also explains that vowel harmony is a phonological state in which the vowels in a given domain share or harmonize for a particular feature and emphatically states that vowel harmony as a phonological process “differs from other processes affecting adjacent vowels in that, typically all of the vowels within the language participate in the harmonic constraint” (Kenstowicz, ibid: 347). Clements (1976:57) further says that vowel harmony consists of a co- occurrence restriction upon the vowels that may occur in a word. In other words, all the vowels in a word must be drawn from one or another of two mutually exclusive sets. Dagara is a natural control-ground, following the definitions or explanations given above, for the confirmation of the phenomenon of vowel harmony. All University of Ghana http://ugspace.ug.edu.gh 77 the eighteen vowels are sub-categorized into two distinct sets based on the position of the tongue root during their production (cf. Chapter two section 2.1.3), and there is a strict co-occurrence restrictions of the two sets of vowels in mono-syllabic or simple multi-syllabic, non-compound words. Invariably, therefore, Dagara exhibits a horizontal harmony. A horizontal harmony, as explained by Akanlig-Pare (1994:86), is the type of harmony involving the root of the tongue. The harmonic processes occur regularly, bi- directionally, between stems and suffixes; with the stem vowel quality usually being the triggers while the suffix vowels are the targets of the harmonic process in a progressive or preservative direction, and the suffix vowels being the triggers while stem vowels are the targets of the harmonic process in a regressive or anticipatory direction. In a stem vowel triggered harmonic process, the two distinct sets of vowels have equal ability to determine the harmonic process provided they are situated in the stem. But in a suffix vowel triggered harmonic process, [-ATR] stem vowels are usually the targets whilst [+ATR] vowels are the triggers. There is also evidence of a harmony across word boundary in Dagara that operates in a progressive or preservative direction. From the harmonic patterns discussed above, there is no gain saying that, the domain of the harmony in Dagara is the phonological word. This is characteristic of Gur languages as Hudu (2013:53), following Dakubu (1997), notes that the phonological word in Central Gur languages, to which Dagbani belongs, typically consists of a thematic CV syllable with full range of articulatory contrast and a mono- or bi-syllabic suffix with restricted vowels University of Ghana http://ugspace.ug.edu.gh 78 and consonants. In Dagara, however phonological word may consist of other syllable shapes besides the CV syllable. Under this section we shall discuss the harmonic processes between; verb stem and progressive/ imperfective suffix; verb stem and nominal suffix; noun stem and nominal suffix; verb stem and the split negation morphemes in the indicative and the imperative, all of which operate in the progressive direction. We shall also discuss the harmonic processes between dominant diminutive suffixes and noun stems, as well as a noun stem and a dominant adjectival stem which operate in regressive directions. 3.1.1.1 Verb Stem and Imperfective Suffix Harmony In Dagara, the imperfective suffix particle is basically re/r when the verb stem is an open syllable (CV) type. But where the verb stem ends in a nasal or a nasalized vowel, the alveolar trill /r/ is replaced by an alveolar nasal /n/ and it becomes ne/n and where the verb stem is of a CVC syllable shape, with the coda consonant being a liquid, then the trill is dropped and the mid-vowels /e, / is added to the verb stem to form the imperfective. The choice of the suffix is based on the [ATR] feature value of the stem vowels, in line with Dakubu (1997) observation that the vowels of suffixes in Gur languages ‘ten to be at least partly etermine by t e features of the t ematic syllable vowel (Dakubu, ibid: 83). The examples in the table below show the harmony between verb stem and the progressive or imperfective: Table 27. Verb Stem and Imperfective Suffix [+ATR] Verb + Imperfective Particle / p / + / -re/ [p re ] University of Ghana http://ugspace.ug.edu.gh 79 ‘swell PROG ‘swelling / 'le / + / -ne/ [ 'l n ] ‘tie P OG ‘tying /y / + /-re / [y re ] ‘open PROG ‘opening / ky r/ + /- e / [ ky re] ‘pour PROG ‘pouring [ k ] +/ -ne / [k n ] ‘cry PROG ‘crying /kul / + /-e / [ kule] ‘marry /go home PROG ‘marrying/going home [-ATR] VERB + IMPERFECTIVE PARTICLE / d ɛ / + / nɛ / [d ɛ nɛ ] ‘play PROG ‘playing /t ɛ r / + / ɛ / [ t ɛ rɛ ] ‘think P OG ‘t inking / p l / + / / [ p l ] ‘swear PROG ‘swearing / k / + / r / [ k r ] ‘give PROG ‘giving In an attempt to explain the above harmonic process, Classical generative phonology posits rules to establish an underlying representation such that, the suffixes regularly alternate for the feature value in agreement with the stem, thereby, systematically varying the quality of their vowels in accord with the stem. In such simple and discretional positing of one form as the underlying representation, it is often difficult to interprete and explain the naturalness of the assimilation process and thus, shrouding the phenomenon of vowel harmony in a puzzle in the past. This naturalness of the assimilatory process in University of Ghana http://ugspace.ug.edu.gh 80 vowel harmony is, however, easily captured in autosegmental phonology, without the need for positing an underlying form for the suffix in the derivation. For instance, since the domain or the trigger of the harmony is the stem or the suffix, this naturalness is captured by the representation as autosegment spreading from the stem to the suffix or vice versa from the hierarchy of the feature tier onto the segmental tier via the skeletal tier in the representation. In other words, a [+/-ATR] vowel in a verb stem spreads its feature quality onto a [+/-ATR] vowel in the suffix, causing it to vary the quality of its vowel in accord with the stem, in the case of a progressive or preservatory harmony. While a [+ATR] vowel in a suffix or stem spreads its feature quality onto a [-ATR] vowel in the stem, in the case of a regressive or anticipatory harmony, causing it to vary the quality of its vowel in accord with the suffix or the stem. An illustration of the autosegmental representation of this naturalness of the assimilatory process in vowel harmony is in figure 25. Fig. 25. Verb and Imperfective Suffix Harmony Process [+ATR] stem + [-ATR] suffix Output / y / ‘open + / -rɛ / [ y re ]‘opening University of Ghana http://ugspace.ug.edu.gh 81 Underlying Form Tongue Root Tier [+ATR] [-ATR] Skeletal Tier X X X X X Segmental Tier y r ɛ [+ATR] Spreading Tongue Root Tier [+ATR] [-ATR] Skeletal Tier X X X X X Segmental Tier y r ɛ [-ATR] Delinking: Tongue Root Tier [+ATR] [-ATR] Skeletal Tier X X X X X Segmental Tier y r ɛ Output Form: Tongue Root Tier [+ATR] Skeletal Tier X X X X X Segmental Tier y r e University of Ghana http://ugspace.ug.edu.gh 82 [-ATR] stem + [-ATR] suffix Output /k / ‘give’ + /-re/ [ k rɛ] ‘giving’ Underlying Form: Tongue Root Tier [-ATR] [+ATR] Skeletal Tier X X X X Segmental Tier k r e [-ATR]Spreading: Tongue Root Tier [-ATR] [+ATR] Skeletal Tier X X X X Segmental Tier k r e Delinking of [+ATR]: Tongue Root Tier [-ATR] [+ATR] Skeletal Tier X X X X Segmental Tier k r e Output Form: University of Ghana http://ugspace.ug.edu.gh 83 Tongue Root Tier [-ATR] Skeletal Tier X X X X Segmental Tier k r ɛ 3.1.1.2 Verb Stem and Nominal suffix Harmony Many nouns are derived from other lexical categories in Dagara with derivational morphemes. A category of verbal nouns are derived from verbs with the addition of the suffixes - / u or –f . While the suffixes - /-u are allophonic, varying depending on the vowel quality in the verb stem, the suffix –f is invariable irrespective of vowel quality in the verb stem. There is a restriction on the spreading of [+ATR] stem vowel quality on the suffix –f . The possible explanation of this phenomenon is that, /f/ is an opaque consonant that blocks the spreading of the [+ATR] feature from the stem to the suffix. Segments opacity effects on harmonic process are not uncharacteristic of language with tongue root vowel harmony, especially in the Gur languages. Hudu (2013: 63), in discussing consonantal opacity in Dagbani tongue-root harmony notes that, the continuant coronals [l, ɾ, s] block the spread of [+ATR] from the root vowel to targets such as epenthetic vowels, affix and clitic vowels. The table below shows the harmonic process between verb stems and the nominal suffix Table 28. Verb and Nomial Suffix a) [-ATR] Verbs + [-ATR] Nominal suffix Verb Stem suffix Derived Noun University of Ghana http://ugspace.ug.edu.gh 84 / t r/ + / / [t r ] ‘ send suffix ‘the sending / dɔ m / + / / [ dɔ m ] ‘to squat suffix ‘the squatting / ky r / + / / [ ky r ] ‘to owry suffix ‘ owry b) [+ATR] Verbs + [+ATR] Nominal suffix Verb Suffix Derived Noun / gbe m / + / -u/ [ gbe mu] ‘to tig ten suffix ‘t e tig tening / gb l / + / -u / [ gb lu ] ‘to penetrate suffix ‘the penetration / b n / + / -u / [ b nu ] ‘to put down suffix ‘putting down c) [+/-ATR] Verbs stem and [-ATR] Suffix Stem suffix Derived Noun / gm / + / -f / [ gm f ] ‘crus suffix ‘crus ing / d ɔ / + / -f / [ d ɔ f ] ‘raise suffix ‘raising / t ɛ / + / -f / [ t ɛ f ] ‘support su ix ‘supporting / g / + / -f / [ g ] ‘c ase su ix ‘t e c asing / t ] + / -f / [ t ] ‘carry su ix t e carrying / m / + /-f / [ m ] ‘s ake suffix ‘t e s aking University of Ghana http://ugspace.ug.edu.gh 85 Figure 26 shows the process of vowel harmony in the derivation of the above nouns in autosegmental representation. Fig. 26. Verb Stem and Nomial Suffix Harmony a) [+ATR] Verb Stem Hormonizing [-ATR] Suffix gb l ‘to penetrate + gb lu ‘penetration Underlying Form: Tongue Root Tier [+ATR] [-ATR] Skeletal Tier X X X X Segmental Tier g b l [+ATR] Spreading : Tongue Root Tier [+ATR] [-ATR] Skeletal Tier X X X X Segmental Tier g b l [-ATR] Delinking: Tongue Root Tier [+ATR] [-ATR] Skeletal Tier X X X X Segmental Tier g b l Output Form: University of Ghana http://ugspace.ug.edu.gh 86 Tongue Root tier [+ATR] Skeletal Tier X X X X Segmental Tier g b l u b) [+ATR] Verb Stem Hormonizing [-ATR] Suffix / t r / ‘sen + / u / [ t r ] ‘t e s aking Underlying Form: Tongue Root Tier [-ATR] [+ATR] Skeletal Tier X X X X Segmental Tier t r u [-ATR] Spreading: Tongu Root Tier [-ATR] [+ATR] Skeletal Tier X X X X Segmental Tier t r u [+ATR] Delinking: Tongue Root Tier [-ATR] [+ATR] Skeletal Tier X X X X Segmental Tier t r u Out put Form: University of Ghana http://ugspace.ug.edu.gh 87 Tongue Root Tier [-ATR] Skeletal Tier X X X X Segmental Tier t r 3.1.1.3 Noun Stem and Nominal Suffix Harmony Some nouns can also be derived from other nouns by adding a nominal suffix to them. Abstract nouns in Dagara are derived from concrete nouns by adding a nominal suffix; - /-u or -l /-lu to them. This category of nouns depict character, style, state, size and quality. The [ATR] vowel quality in the noun stem triggers the harmonic process. In other words, a [+/-ATR] vowel in a noun stem spreads its feature quality onto a [+/-ATR] vowel in the suffix to vary in accord with the stem. Table 29 shows examples of nouns derived from other nouns. Table 29. Noun Stem and Nominal Suffix a) [-ATR] Nouns and [-ATR] Suffix Noun Suffix Derived Noun / pɔ g / + / -l / [ pɔ gl ] ‘woman suffix ‘womanhood / s ɔ / + / -l / [ s ɔ l ] ‘witchcraft suffix ‘witchery / s n / + / - / [ s n ] ‘guest suffix ‘hospitality b) [+ATR] Nouns and [+ATR] suffix Noun Suffix Derived Noun / y r / + / -lu / [ y lu ] ‘ ouse suffix ‘patriclan/kins ip / bie / + [ -lu / [ biilu ] ‘c il suffix ‘c il oo University of Ghana http://ugspace.ug.edu.gh 88 / mb l / + / -u / [ mb lu ] ‘ ool suffix ‘ oolis ness The above assimilation is formalized in autosegmental phonology in figure 27. Fig.27 Noun Stem and Nominal Suffix Harmony / y r / ‘ ouse + / -l / [y lu] ‘patriclan/kinship Underlying Form: Tongue Root Tier [+ATR] [-ATR] Skeletal Tier X X X X X Segmental Tier y r l [+ATR] Spreading: Tongue Root Tier [+ATR] [-ATR] Skeletal Tier X X X X X Segmental Tier y r l [-ATR] Delinking: Tongue Root Tier [+ATR] [-ATR] Skeletal Tier X X X X X Segmental Tier y r l Segment Deletion: University of Ghana http://ugspace.ug.edu.gh 89 Tongue Root Tier [+ATR] Skeletal Tier X X X X X SegmentalTier y r l Vowel Lenghtening: Tongue Root Tier [+ATR] Skeletal Tier X X X X X Segmental Tier y l u Output Form: Tongue Root Tier [+ATR] Skeletal Tier X X X X X Segmental Tier y : l u 3.1.1.4 [+ATR] Verb and Negation Particles Harmony Negation in Dagara is expressed in the indicative and imperative by means of two discontinuous morphemes; b ........... / and ta.............. / , for the indicative and imperative structures respectively. The structural arrangements are similar to the structure of a negative clause in French, where the verb is placed between the two negative markers. These morphemes are underlyingly [-ATR], but the second components of the morpheme undergo a [+ATR] harmony when preceded by a verb with [+ATR] vowels, whilst the first components remains invariably [-ATR]. Thus, their allophonic versions are b …….. i and t …… i for indicative and imperative, respectively. It is University of Ghana http://ugspace.ug.edu.gh 90 important to note here that, the harmony process is between the verb and the second negation particles and does not affect the first negation particles; hence the process is across word boundaries in the progressive direction. om (2004:34-35) highlights this phenomenon, acknowledging the fact that it is vowel harmony across word boundaries in the progressive direction, but with inaccurate interpretations or explanations. The following negative clauses below illustrate the phenomenon. Indicative Imperative bɛ bɔ r Ta bɔ r ɛ 3SG NEG loose NEG NEG loose NEG ‘He / She did not get lost ‘Don t be lost bɛ d Ta d 3SG NEG eat NEG NEG eat NEG ‘ / e i not eat ‘ on t eat [+ATR] Vowel and Negation Morphemes Harmony bɛ * bɛ do i 3SG NEG climb NEG ‘ / e i not climb Ta d * Ta d i. NEG take NEG ‘Do not climb it bɛ k l *ɛ bɛ k l e 3SG NEG go home NEG ‘S/he did not go home k l * . Ta k l e. NEG go home NEG ‘Do not go home University of Ghana http://ugspace.ug.edu.gh 91 What is linearly desribed above can be formalized in autosegmental phonology, as in figure 28, showing the autosegmental feature spreading from the [+ATR] verb onto the second conponent of the negation. Fig. 28. Harmony across Word Boundary Undrlying Form: Tongue root Tier [+ATR][-ATR] Skeletal tier X X X X X X Segmental Tier b d [+ATR] Spreading: Tongue Root Tier [+ATR][-ATR] Skeletal Tier X X X X X X Segmental Tier b d [-ATR] Delinking Tongue Root Tier [+ATR] [-ATR] Skeletal Tier X X X X X X Segmental Tier b ɛ d University of Ghana http://ugspace.ug.edu.gh 92 Output Form: Tongue Root Tier [+ATR] Skeletal Tier X X X X X X Segmental Tier b ɛ i 3.1.1.5 [-ATR] Noun Stems and Diminutive Suffix Harmony The diminutive suffixes –le (singular) and li (plural) impose their [+ATR] feature value on stems with [-ATR] vowels in word formation process in Dagara. The harmony process thus, is in the regressive direction. In some of the derivations are preceded by some syllable structure processes which are duly discussed in subsequent sections in this thesis. Examples showing this imposition of [+ATR] features of the diminutive suffix are given in table 30. Table. 30. [-ATR] Noun Stems and Diminutive Suffix Harmony / sɔ r / + / -le / [ s le ] ‘path diminutive ‘small path / s ɔ / + / le / [ s le ] ‘knife diminutive ‘small knife / gbɛ r / + / -le / [ gb le ] ‘leg diminitive ‘sheen/ small leg / t ɔ r / + / -le / [ t le ] ‘mortar iminutive ‘small mortar / lɛ r / + / -le / [ l le ] ‘axe diminutive ‘small axe / y ɔ r / + / -le / [ y le ] ‘penis diminutive ‘small penis University of Ghana http://ugspace.ug.edu.gh 93 / kɔ g / + / -le / [ k gle ] ‘chair diminutive ‘stool / d g / + / -le / [ gle ] ‘pot diminutive ‘small pot Figure 29 shows the autosegmental representation of this regressive assimilatory process. Fig. 29 [-ATR] Noun Stems and Diminutive Suffix Harmony d g ‘pot + -le ‘ eminutive gle ‘small pot Underlying Form: Tongue Root Tier [-ATR] [+ATR] Skeletal Tier X X X X X Segmental Tier d g l e [+ATR] Left Ward Spreading: Tongue Root Tier [-ATR] [+ATR] Skeletal Tier X X X X X Segmental Tier d g l e [-ATR] Delinking: Tongue Root Tier [-ATR] [+ATR] Skeletal Tier X X X X X Segmental Tier d g l e University of Ghana http://ugspace.ug.edu.gh 94 Output Form: Tongue Root Tier [+ATR] Skeletal Tier X X X X X Segmental Tier g l e 3.1.1.6 [-ATR] Stems and [+ATR] Dominant Stem Harmony Vowel harmony, in Dagara, do not usually operate in words which are compounds of two different stems consisting of vowels from the two distinct sets; [+ATR] vowels and [-ATR] vowels. There is nevertheless, a dominant [+ATR] vowel adjectival stem, r ‘see ( ‘see s ), which harmonizes [- ATR] vowel stem in word compounding. Again, some of the derivations here involeve some syllable structure processes which are subsequently discussed. Examples of the exceptional phenomenon are given in table 31. Table 31. [-ATR] Stems and [+ATR] Dominant Stem Harmony / s ɔ / + / b r / [ s b r ] ‘ kni e ‘see ‘knife blade / gbɛ r / + / b / [ gb b ] ‘leg seeds ‘toes / lɛ r / + / b / [ l b ] ‘axe ‘see s ‘axe blades / y ɔ r / + / b r / [y b r ] ‘penis ‘see ‘tip of penis University of Ghana http://ugspace.ug.edu.gh 95 Figure 30 shows an example of the formalization of the above process in autosegmental phonology. Fig. 30. [-ATR] Stems and [+ATR] Dominant Stem Harmony / gb r / + / / [ ] ‘leg ‘see s ‘toes Underlying Form: Tongue Root Tier [-ATR] [+ATR] Skeletal Tier X X X X X X Segmental Tier g b ɛ r b [+ATR] Leftward Spreading and Segment Deletion : Tongue Root Tier [-ATR] [+ATR] Skeletal Tier X X X X X X Segmental Tier g b ɛ r b [-ATR] Delinking : Tongue Root Tier [-ATR] [+ATR] Skeletal Tier X X X X X Segmental Tier g b ɛ b Output Form : Tongue Root Tier [+ATR] Skeletal Tier X X X X X University of Ghana http://ugspace.ug.edu.gh 96 Segmental Tier g b b 3.1.2 Consonant Nasalization Nasalization is a Manner of articulation assimilation process, whereby an oral segment acquires nasality from an adjacent nasal segment. In most Gur languages, Kusaal, Buli and Dagaare, in particular and other African languages in general, the process of segment nasalization is a common. For instance, in Kusaal and Buli, oral vowels only become nasalized when they are found in contiguous position to a nasal consonant (cf. Agoswin 2010 and Akanlig-Pare 2005; 1994 respectively). In Dagara, however, nasalization process strictly involves two voiced oral stops (/b/ and /g/), when they are in contiguous position to a nasal segment. Consonant nasalization is induced by a morphological process, for example, when two stems or a stem and a suffix are put together to form a compound word. When the first stem ends in a nasal and the second stem or the suffix begins with a voiced bilabial oral stop /b/ or voiced velar stop /g/, the nasal feature in the first stem spreads on to assimilate the voiced oral stop. The process is a total assimilation in a progressive direction. The following data adapted from Bemile (1985:101, 104) in table 32 attest to this fact in Dagara Table 32. Consonant Nasalization /z m + -b l- / [ mm l ] ‘ is ‘small - PL ‘small fishes /s + -bɛ/ [s mɛ] ‘guest -PL ‘guests University of Ghana http://ugspace.ug.edu.gh 97 /sɛ n + -bɛ/ [sɛ nmɛ] ‘girl friend - PL ‘girl friends / a ŋ + g n/ [ ŋŋ n] ‘body skin skin (human) Figure 31 shows an illustration of the consonant nasalization assimilation process in the autosegmental representation. Figure 31 Consonant nasalization a) / z m / + /-b l- / [ u mm l ] ‘ is ‘small -PL ‘small fishes (keta boys) Underlying Form : M. A. Tier -Nas + Nas -Nas Skeletal Tier X X X X X X X Segmental Tier z m b l [+Nasal] Spreading Rightward : M. A. Tier -Nas +Nas -Nas Skeletal Tier X X X X X X X Segmental Tier z m b l Oral Segment deletion and Nasal segment germinating : M. A. Tier -Nas +Nas -Nas Skeletal Tier X X X X X X X Segmental Tier z m b l i University of Ghana http://ugspace.ug.edu.gh 98 Output Form : M. A. Tier -Nas +Nas -Nas Skeletal Tier X X X X X X X Segmental Tier z m l 3.1.3 Homorganic nasal Assimilation (HNA) Homorganic nasal assimilation is one of the commonest places of articulation assimilation attested in a variety of languages in the world. Durand (1990; 100) explains that, it is the assimilation process whereby a nasal consonant inherits the place of articulation of a plosive. In Dagara, homorganic nasal assimilation process operates during word formation processes. When two stems are brought together; the preposed stem ending in a nasal consonant and the second beginning with an oral plosive or even another nasal to form a compound word, the nasal consonant in the preposed stem, inherits the place of articulation of the initial consonant of the scond stem, whether oral or another nasal consonant. In Table 33 below are examples of some words in Dagara, showing the homorganic nasal assimilation process during compounding, while figure 32 shows the autosegmental formalization of the process in the regressive direction of the assimilation. Table 33. Homorganic Nasal Assimilation /t ɛ m / + / k bl / /t ɛ mk bl / [t ɛ ŋk bl ] ‘c in ‘ air ‘bear /t nɛ / + /b g/ + / kpɛ ɛ / /t nb gkpɛ ɛ / t mb gkpɛ ɛ ] University of Ghana http://ugspace.ug.edu.gh 99 ‘earth ‘hole ‘big ‘lake /g n/ + / pɛ l / /g npɛ l/ g mpɛ l] ‘skin ‘sheet ‘a sheet of paper /z m / + / daa / /z mdaa/ [z ndaa] ‘fish male ‘male fish / tɛ mɛ / + / n ɔ / /tɛ mn ɔ / [tɛ nn ɔ ] ‘tobacco ‘sweet ‘sweet tobacco / g ŋ / + / mimir/ /g ŋmimir/ [g mmimir] ‘cloth ‘eye ‘attractive cloth Figure 32 shows Autosegmental phonology formalization of the homorganic nasal assimilation process. Fig. 32 Autosegmental representation of Homorganic Nasal Assimilation a) t ɛ m + k bl /t ɛ mk bl / [t ɛ ŋk bl ] ‘c in ‘ air ‘bear Underlying Form : P. A. Tier [Lab] [Doral] Skeletal Tier X X X X X X X X X Segmental Tier t ɛ m k b l Dorsal feature Leftward Spreading : P.A. Tier [Lab] [Dorsal] Skeletal Tier X X X X X X X X X Segmental Tier t ɛ m k b l Labial feature Delinking : University of Ghana http://ugspace.ug.edu.gh 100 P. A. Tier [Lab] [Dorsal] Skeletal Tier X X X X X X X X X Segmental Tier t ɛ m k b l Output Form : P. A. Tier [Dorsal] Skeletal tier X X X X X X X X X Segmental Tier t ɛ ŋ k b l b) z m + daa /z mdaa/ [z ndaa] ‘fish ‘male ‘male fish Undrelying Form : P. A. Tier [Labial] [Coronal] Skeletal Tier X X X X X X Segmental Tier z m d a: Coronal feature Leftward Spreading : P. A. Tier [Labial][Coronal] Skeletal Tier X X X X X X Segmental Tier z m d a: Labial feature Delinking : University of Ghana http://ugspace.ug.edu.gh 101 P. A. Tier [Labial] [Coronal] Skeletal Tier X X X X X X Segmental Tier z m d a: Output Form : P. A. Tier [Coronal] Skeletal Tier X X X X X X Segmental Tier z n d a: 3.1.4 Consonant Labialization Labialization is a process which consonants acquire by being contiguous to rounded vowels. In Dagara there are no underlying labialized consonants, however, consonants that are followed immediately by rounded vowels are produced with lip rounding in the process. Akanlig-Pare (1994) describes this process as a secondary articulation process, since the primary stricture to which the rounding feature is added, is not altered. Labialization is a regressive assimilatory process. Some examples of consonant labialization are in Dagara in table 34. Table 34. Consonant Labialization / k / [k w ] ‘kill / g r / [g w :r] ‘t orns / t / [tw ] ‘ to ollow University of Ghana http://ugspace.ug.edu.gh 102 / / [dw ] ‘to swim / zɔ / [z wɔ ] ‘ to run / sɔ gl / [s wɔ gl ] ‘to i e / n ɔ / [n w ɔ ] ‘sweetness / m ɔ / [m w ɔ ] ‘rippen / bɔ / [b wɔ ] ‘look or / pɔ / [p wɔ ] ‘to swear / li / [fw li] ‘w istle / v li / [vw li] ‘look into Figure 33 illustrates the autosegmental representation of consonant labialization process. Fig.33 Autosegmental Representation of Consonant Labialization a) / bɔ / [b wɔ ] ‘look or Underlying Form Leftward [+Round] Spreading M. A. Tier [-Round][+Round] [-Round][+Round] Skeletal Tier X X X X Segmental Tier b ɔ b ɔ [-Round] feature delinking Output Form M. A. Tier [-Round] [+Round] [+Round] University of Ghana http://ugspace.ug.edu.gh 103 Skeletal Tier X X X X Segmental Tier b ɔ bw ɔ 3.1.5 Glide Formation A glide is a term used in phonetics to refer to a transitional sound as the vocal organs move towards or away from an articulation, (Crystal, 1992: 153). This is a manner of articulation assimilation process whereby a sound takes on the manner of articulation features of an adjacent sound. In CVC words in Dagara, the voiced velar stop / g / or the velar nasal /ŋ / in coda position becomes a Labial velar approximant, /w/. The manner of articulation feature, [+Cont] of the vowel, spreads in a progressive direction on to the [-Cont] consonants to make them surface as a Labial velar approximant /w/ which is a [+Cont]. Table 35 contains some examples of Dagara words in whose production, the final velar stop and nasal glide to a Labial velar approximant. Table 35. Glide Formation / pɔ g/ [ pɔ w ] ‘woman / tɛ g / [ tɛ w ] ‘to exchange / k g / [ k w ] ‘oak tree / ny g / [ ɲ w ] ‘to mix / ny g / [ ɲ w ] ‘to burn / naŋ / [ naw ] ‘scorpion / b ŋ / [ b w ] ‘donkey University of Ghana http://ugspace.ug.edu.gh 104 The autosegmental representation of the glide formation process is shown in figure 34. Fig 34 Glide formation in Autosegmental Representation. a) / k g / [ k w ] ‘oak tree Underlying Form [+Cont] Feature Spreading M. A. Tier [+Cont] [-Cont] [+Cont] [-Cont] Skeletal Tier X X X X X X Segmental Tier k g k g Ouput Form M. A. Tier [+Cont] Skeletal Tier X X X Segmental Tier k w b) / n ŋ / [ n w ] ‘scorpion Underlying Form [+Cont] Fearure Spreading M. A. Tier [+Cont] [-Cont] [+Cont] [-Cont] Skeletal Tier X X X X X X Segmental Tier n ŋ n ŋ Output Form M. A. Tier [+Cont] University of Ghana http://ugspace.ug.edu.gh 105 Skeletal Tier X X X Segmental Tier n w 3.1.6 Rhotacism Rhotacism is a bi-directional Manner of assmilation process identified in Dagara. Gummere (1960:254), defines rhotacism as the change of /s/ between vowels to /r/, but Saanchi (2006) broadly defines the phenomenon as the conversion of other sounds to /r/ and remarks that it is a marked feature of Dagaare, of which Dagara is a dialect, compared to other Oti-Volta Languages (Saanchi, ibid:14). Rhotacism as an assmilation process is an attested fact in other languages, especially non-African languages, as an unmarked process. e no (1987) reports of an intervocalical /d/ rhotacism in Dominican Spanish. He claims, among others, that /d/changes into /r/ not because of their large number of shared features but because of the assimilation of certain features of adjacecnt segments would unequivacally produce a predictable tap, ( e no, ibid: 365). Davies (1982) also sugguests that in Hier.Luwian (an Anatolian language) dentals can change into /r/, with the following evidence in the language: ……….in some texts of the first millennium, we find words written with the ra/ri and ru signs instead of the expected ta/ti and tu signs. The texts in question are inconsistent in their spellings; KARATEPE has both “ [mis]-ia-ti” and “ -mi-ia +ra/i” for ablative of amis-‘my (Davis, ibid: 245-246). In fast speech in Dagara, the voiceless alveolar stop /t/ in intervocalic position changes into /r/ across word boundaries in a defined phonological environment. In an utterance that contains a sequence of mono-syllabic words with CV/CV: University of Ghana http://ugspace.ug.edu.gh 106 syllable shapes, if the initial consonant of the preceding word is the voiceless alveolar stop, /t/ and the vowel is a high front vowel, the original voiceless alveolar stop /t/ surface phonetically as an alveolar trill /r/, especially in fast speech. What happens is that the [+Cont] features of the non-consonantals spread onto the [-Cont] features of the /t/ to result in the trill. The following examples of such utterances below illustrate the phenomenon. Table 36. Rhotacism / p ri n ŋm n k t / [p ri n ŋm n k r ] ‘Pray or us / k t na / k r na] ‘Take care o us to ay / zɔ t n mb l / [zɔ r n mb l ] ‘Have mercy on us / ta t t / [ta t r ] ‘Don t ollow us / zɔ t kpɛ / [zɔ r kpɛ ] ‘run an enter / t / r ] ‘Go up an collect / w t ʨ n / [w r ʨ n] Come an lets go Explaining rhotacism within linear generative phonology exposes yet another limitation of the theory because it lacks the formal tool to capture unequivocally the nature and the direction of the assimilation process. Autosegmental approach has the machinery; the multi-tier parallel arrangement of autosegments, to capture explicitly the nature and direction of the process. Figure 35 illustrates in graphical terms how autosegmental phonology captures University of Ghana http://ugspace.ug.edu.gh 107 the nature of rhotacism. Fig 35 Rhotacism in Autosegmental Representation / / ‘go up an collect ] Underlying Form : M. A. Tier [+Cont] [-Cont] [+Cont] Sekeletal Tier X X X X Segmental Tier d t Bi-directional Spreading of [+Cont] Feature : M. A. Tier [+Cont] [-Cont] [+Cont] Skeletal Tier X X X X Segmental Tier d t Delinking of [-Cont] Feature : M. A. Tier [+Con] [-Con][+Con] Skeleltal Tier X X X X Segmental Tier d t Output Form : M. A. Tier [+Cont] [+Cont] University of Ghana http://ugspace.ug.edu.gh 108 Skeletal Tier X X X X Segmental Tier d r In autosegmental analysis, the derivation can be explained from the view point that, the non-consonantal segments are acting as the triggering context for the assimilation process from the manner of articulation Tier from both directions. The [+Cont] features of the non-consonantal segments spread bi-directionally onto the stop /t/. The stop /t/, thus, assimilates the continuance of the adjacent non-consonantal segments and surfaces as a trill /r/. Though the trill /r/ is also a type of a stop, it is much closer to [+Cont] sounds than a complete stop, since during its production there is still the continuous follow of airstream along the sides of the tongue blade in the vocal tract similar to the production of a [+Cont] Sound. 3.2.0 Syllable Structure Processes Syllable structure processes are those processes that lead to the lost or addition of a sound in the word for morphophonological reasons. For instance, morphological process of compounding, and also some times in fast speech, sometimes results in segments lost. Similarly, the adaption of loan words into a language also results in segments insertations. These processes affect the basic syllable structure of words in a language by altering syllable shape of words. In Dagara, the most common syllable structure processes induced by morphological process include apocope, the ommision of final segment of a word, and syncope, the ommision of segments within a word. The syllable structure processes discussed under this section include Elision, Compensatory lenghtening, Resyllabification and Vowel Epenthesis. University of Ghana http://ugspace.ug.edu.gh 109 3.2.1 Elision Elision, in Dagara, takes the form of both apocope and syncope; that is, the omission of final segment of a word and/or the omission of a segment within a word respectively. This commonly occurs in word compounding process during which a vowel or a consonant is elided when two stems are put together to form a new word. 3.2.1.1 Vowel Elision A vowel can be elided in two instances during word compounding. In the first instance, final vowel of CV or CVV syllable shape of an initial stem is elided before the second stem is added in order to form a compound word. The data in table 37attest to this type of vowel elision. Table 37. Final Vowel Elision in CV /CVV syllable Stem1 Stem2 Compound word / pɛ r / + / p la / [ pɛ rp la ] ‘sheep ‘white ‘white sheep / n r / + / z ɛ / [ n rz ɛ ] ‘cock ‘re ‘re cock / sa mɛ / + / y g / [sa my g ] ‘guest ‘many ‘many guest / bie / + / dɛ b / [ bidɛ b] ‘child ‘male’ ‘son / gbɛ :/ + / ʨe n / [gbɛ ʨe n ] ‘ egs ‘walking ‘ eet / kɛ : / + / k r / [kɛ k r ] ‘malt ‘ol ‘ol malt /ba: / + / p :la / [ bap :la] ‘ og ‘young ‘puppy University of Ghana http://ugspace.ug.edu.gh 110 In the second instance, a back, high vowel within a CVV syllable shape of an initial stem is elided before the second stem is added in order to form a compound word. The data in table 38 attest to this type of vowel elision. Table 38. Vowel Elision in CVV Syllable Stem1 Stem2 Compound word / n ɔ / + / pɔ g / [ nɔ pɔ g] ‘ owl ‘woman ‘ en / s ɔ + / kp r / [ sɔ kp r ] ‘kni e ‘sheave ‘knife sheave / t / + / v :r / [ t v :r ] ‘baobao ‘leaf ‘baobao leaf / g ɔ / + / t ɛ / [ gɔ t ɛ ] ‘thorn ‘tree ‘thorn tree The autosegmental representation of the above syllable structure process is illustrated in figure 36. Fig. 36 Vowel Elision in Autosegmental Representation a) / p r / + / p la / [ p r-p la ] ‘sheep ‘white ‘white sheep Underlying Form : Syllable Tier ó ó ó ó CV Tier C V C V C V C V Segmental Tier p ɛ r p l a Final Vowel Elision : Syllable Tier ó ó ó ó University of Ghana http://ugspace.ug.edu.gh 111 CV Tier C V C V C V C V Segmental Tier p ɛ r p l a Resyllabification process : Syllable Tier ó ó ó ó CV Tier C V C C V C V Segmental Tier p ɛ r p l a Syllable truncation : Syllabe Tier ó ó ó ó CV Tier C V C C V C V Segmental Tier p ɛ r p l a Output Form : Syllable Tier ó ó ó CV Tier C V C C V C V Segmental Tier p ɛ r p l a University of Ghana http://ugspace.ug.edu.gh 112 3.2.1.2 Consonant Elision Westermann and Ward (1966), cited in Bemile (1985:203), note that, in quick or slovenly speech, it is found out that consonants are pronounced with less vigorous articulations and consequently tend to weaken and drop out. This observation is very remarkable in Dagara as consonant elision is a common practice not only in word formation processes, but also in ‘ uick or ‘slovenly speech as noted by Westermann and Ward. In some compounding processes, final consonants in initial stems of CVC/CVCC syllable shapes are elided when the second stem begins with another consonant or a vowel. Where the Onset consonant of the second stem is the same as that of the Coda consonant of the initial stem, the Coda consonant is elided before a compound is formed. The data in table 39 drawn from Bemile (1985) attest to the above syllable structure process in Dagara. Table 39 Consonant Elision in compounding Stem1 Stem2 Compound word / lɔ b / + / bɛ r / [ lɔ bɛ r ] ‘throw ‘leave ‘throw away / sɛ n / + / n : / [sɛ n : ] ‘girl/boy friend ‘hand ‘girl/boy friends hand / zɔ ŋ / + / ŋa / [zɔ :ŋa ] ‘blind person ‘this ‘this blind person / k l / + / le / [ k le ] ‘dawadawa ‘diminutive ‘small dawadawa / lɛ b / + / wa: / [ lɛ wa: ] ‘turn back ‘coming ‘the return University of Ghana http://ugspace.ug.edu.gh 113 / pɔ g / + / a :n / [pɔ a :n] ‘woman ‘barren ‘barren woman / z m / + /s :la/ [z -s :la] ‘fish ‘smooth ‘mudfish / pɛ n / + / ulu / [ pɛ ulu ] ‘rag ‘grey ‘grey rag / ka ŋ / + / gyɛ lɛ / [ka gyɛ lɛ ] ‘guinea owl ‘eggs ‘guinea owl eggs The autosegmental representation of the above syllable structure process is illustrated in fig 37. Fig.37 Consonant Elision in Autosegmental representation / lɔ b / + / b r / [ lɔ b r ] ‘throw ‘ eave ‘throw away Underlying Form : Syllable Tier ó ó CV Tier C V C C V C Segmental Tier l ɔ b b ɛ r Consonant Elision : Syllable Tier ó ó CV Tier C V C C V C Segmental Tier l ɔ b b ɛ r Output Form : University of Ghana http://ugspace.ug.edu.gh 114 Syllable Form ó ó CV Tier C V C V C Segmental Tier l ɔ b ɛ r Besides this morphological process during which ssegments are elided, in ‘ uick an ‘slovenly speec , as note by Westermann an War , some segments in words or an entire syllable in a stretch of an utterance maybe elided causing a restructuring of the syllable structure of adjacent words. The following utterances exemplify segment elision in speech process. Table 40. Consonant and Vowel Elision in fast Speech a) / kul n b .b :r / [ ku.l b .b :r ] CVC CV V CV.CV:C CV.CV V CV.CV:C “go ome wit t e c il ren” b) / saab n z ɛ r na d / [ saa. b z ɛ . r d ] CV:C CV CVVC CV V CV CV:.CV CVV.CV V CV “ T. . wit soup t at I ate” c) / ɓ b na ɓ m / [ɓ .b ɓ m ] CVC CV V CV CV CV.CV V CVC “It is slapping t at e slappe me” d) / f yɛ b n bɛ nyɔ g w n / [ f yɛ .b bɛ nyɔ g w n] V CV CVC CV CV CVC CV CV V CV CV.CV CV CVC CVC “it is your brot er t at t ey caug t an brought” 3.2.2 Compensatory Lengthening. Compensatory lengthening can be defined as the lengthening of a segment triggered by the deletion or shortening of a nearby segment, (Hayes, 1989:260). Compensatory lengthening is an attested process in many languages including University of Ghana http://ugspace.ug.edu.gh 115 the Gur languages, (cf. Akanlig-Pare 1994; 2005 and Agoswin 2010), in which the lengthened segment is usually a short vowel.When a segment within a syllable is deleted, the sound duration of the short vowel is extended to compensate for the loss of duration that is caused by the loss of adjacent element. In Dagara, compensatory lengthening is a highly marked phenomenon as all the short vowels have their long counterparts. Its occurrence is conditioned by compouding process. In the derivation of some nouns from other nouns or other word classes using the nominal suffix,-l , the final segment of the stem to which it is affixed, is usualy deleted which consequently create the condition for the lengthening of the sound duration of the short vowel that precedes the deleted segment. The deleted segment which triggers the extension of the duration of the short vowel is usually a vowel or the alveolar trill /r/ consonant. This phonological process may entail other processes during its derivation. Table 41 shows some examples of the compensatory lengthening in Dagara. Table. 41. Compensatory Lengthening Stem Suffix Compound word / z ɛ / + -l [ z :l ] ‘re suffix ‘reddness / bie / + -l [ bi:lu ] ‘child suffix ‘childhood / y r / + -l [ y :lu] ‘house suffix ‘kinship Figure 38 shows the autosegmental phonology formalization of the compensatory lengthening in Dagara. University of Ghana http://ugspace.ug.edu.gh 116 Fig. 38 Compensatory Lengthening in Autosegmental Representation / bie / ‘child + /-l / [ bi:lu ] ‘childhood Underlying Form : Syllable Tier ó ó Skeletal Tier X X X X X Segmental Tier b i e l Vowel Deletion : Syllable Tier ó ó Skeletal Tier X X X X X Segmental Tier b i e l Vowel Lengthening : Syllable Tier ó ó Skeletal Tier X X X X X Segmental Tier b i l Tongue Root Tier [+ATR] [-ATR] University of Ghana http://ugspace.ug.edu.gh 117 [+ATR] Spreading : Syllable Tier ó ó Skeletal Tier X X X X X Segmental Tier b i: l Tongue Root Tier [+ATR] [-ATR [-ATR] Delinking : Syllable Tier ó ó Skeletal Tier X X X X X Segmental Tier b i: l Tongue Root Tier [+ATR] [-ATR] Output Form : Syllable Tier ó ó Skeletal Tier X X X X X Segmental Tier b i: l u Tongue Root Tier [+ATR] 3.2.3 Resyllabification A type of Resyllabification process whereby the final segment of a word runs onto or is linked to the initial segment of the following word is also attested in University of Ghana http://ugspace.ug.edu.gh 118 Dagara. This phonological phenomenon which is similar to liaison, especially attested in French (Durand 1990), is a common phonological process in the Gur languages and other Ghanaian languages. For instance, resyllabification process attested in Dagara when nouns whose final segment is an alveolar trill, /r/, is followed immediately by the question word, ŋmi ‘ ow many , or a numeral beginning with a vowel at syllable initial. When a noun ending in a trill, /r/, is followed by thequestion word, ŋmi ‘ ow many or a numeral beginning with a vowel at initial syllable, the trill is linked to the first syllable of the question word or the numeral in question. This brings about the restructuring of the underlying syllable structure at the surface level. The examples in table 42 are evidence of this phonological fact in Dagara. Table. 42. Resyllabification in Dagara Noun Numeral/ Q-word Underlying form Resyllabification / p :r/ + / .yi / / p :r. .yi / [ p :. r . y ] ‘s eep ‘two ‘two sheep’ CV:.CV.CV / b :r / + / .ta / /b :r. .ta / [ b :. r . ta] ‘goats ‘t ree ‘t ree goats CV:.CV.CV / :r/ + / .w / / :r. .w / :. r . w ] ‘sticks ‘nine ‘ nine sticks CV:.CV.CVV /bagr/ + /a.n :/ /bagr.a.n :/ [bag. r . n : ] ‘shoulder ‘five ‘five shoulders CVC.CV.CV: /s ɔ br/ + / a.n / / s ɔ br.a.n / [s ɔ b. ra. n ] ‘witc es ‘eig t ‘eig t witc es CVVC.CV.CV /y :br/ + / .n :r/ /y :br. . na :r/ y :b. r . n :r] ‘twins ‘ our ‘ our twins CV:C.CV.CV:C University of Ghana http://ugspace.ug.edu.gh 119 /b gr/ + / .ŋm n/ /b gr. .ŋm n/ b g. r . ŋm n] ‘law ‘how many ‘how many laws CVC.CV.CVC The autosegmental representation of this syllable structure process is illustrated in figure 39 below: Fig. 39 Resyllabification in Autosegmental Representation a) /b :r/ ‘goats + / .t / ‘t ree /b :r. .t / ‘t ree goats [ b :. r . t ] Underlying Form : Syllable Tier ó ó ó CV Tier C V V C V C V Segmental Tier b : r t Syllable Truncation/Spreading : Syllable Tier ó ó ó CV Tier C V V C V C V Segmental Tier b : r t Output Form : Syllable Tier ó ó ó Skeletal Tier C V V C V C V Segmental Tier b : r t a University of Ghana http://ugspace.ug.edu.gh 120 3.2.4 Vowel Epenthesis Vowel epenthesis is a common phonological process in dealing with loan words in many Languages, especially where the syllable structure of borrowed words do not conform to the syllable structure requirement of the borrowing language. According to Uffmann (2001:1), vowel epenthesis is a process in loan word adaptation in order to satisfy the constraints on phonotactics and syllable structure in the borrowing language. Languages with a preference for CV syllable structure often epenthesize vowels in positions where they serve to break up consonant clusters or resyllabify coda consonants. Dagara, like most African languages, borrow a lot of lexical items from other languages. Some of the source languages of Dagara loan words include Isaala, Hausa, Akan, French and English. However, in this thesis the discussion on vowel epenthesis is limited to English loan words in Dagara. The constraints on phonotactics and syllable structure of English and Dagara are not the same. For instance, whereas it is permissible in English to have Onset consonants cluster, Onset consonants cluster is highly constrained in Dagara. Consonant cluster is only permissible at word medial or word final position, where they serve as syllable boundaries, and even then there is still a constraint on the consonant sequencing. Due to this Morpho-phonological constraint that Dagara poses, vowels are epenthesized in English loan words to generate licit output in compliance with the Dagara syllable structure requirements. No conclusive statement can be made with regards to which vowels serve as epenthetic vowels. However, observable pattern in the data University of Ghana http://ugspace.ug.edu.gh 121 shows that the epenthetic vowel is often a duplication of the stem vowel or / i / by default. The data below in table 43, adapted from Bemile (1985), show how vowels are epenthesized in English words to make them comply with the Dagara syllable structure. Table 43. Vowel Epenthesis in Loan Word English Syllabifcation Dagara Syllabification Gloss [brik] CCVC [b .r .k ] CV.CV.CV brick [ gla:s ] CCV:C [g .l .s ] CV.CV.CV glass [tra ze(z)] CCV.CV(C) [t r s ] CV.CV.CV trouser(s) [trʌk] CCVC [ t .r .k ] CV.CV.CV truck [sku:l] CCV:C [s .k :l] CV.CV:C school [spriŋ] CCCVC [s .p .r .ʥ ] CV.CV.CV.CV spring The loan word does not only conform to the syllable structure of Dagara, it is also adapted to conform to Dagara tonology since tone is an integral part of the syllable in Dagara. The illustration of epenthetic process formalized in autosegmental representation is in figure 40 below. Fig.40 Vowel Epnthesis in Autosegmental Representation. English Dagara Gloss [spriŋ [s .p .r .ʥ ] spring University of Ghana http://ugspace.ug.edu.gh 122 Underlying Form Output Form Syllabe Tier ó ó ó ó ó CV Tier C C C V C C V C V C V C V egmental Tier s p r i ŋ s p r ʥ 3.3.0 Chapter Summary This Chapter constitutes the core part of the thesis. It discussed both assimilatory processes and syllable structure processes and formalized them within Autosegmental framework. For the assimilation processes, it identified and discussed Vowel Harmony, Consonant Nasalization, Homorganic Nasal Assimilation, Glide formation, Consonant Labialization, and Rhotacism. It observed that, there is a very strict vowel harmony system in Dagara, which operates bi-directionally, involving stems and suffixes, [-ATR] noun stems and a [+ATR] dominant stem and across word boundaries as well. It also noted that all the assimilatory processes in Dagara involve Place of Articulation and Manner of Articulation features. For syllable structure processes, it identified and discussed Elision which involves vowel and consonant, Compensatory Lengthening, Resyllabification and finally Vowel epenthesis. It again, observed that both assimilatory and syllable structure processes are triggered by morphological processes. From the theoretical perspective, the formalization of the phonological processes within the Autosegmental framework is, “a natural testing-ground for the thesis that autosegmental phonology is not restricted to tonal phenomena, University of Ghana http://ugspace.ug.edu.gh 123 but constitutes a general theory of phonological representation”, (Clements, 1976: 57). University of Ghana http://ugspace.ug.edu.gh 124 CHAPTER FOUR DAGARA TONE AND TONAL PROCESSES 4.0 Introduction Dagara tonology has received detailed description in Bemile (1983, 1984, and 1985) within linear generative phonological framework. This chapter, however, presents a description in a non-linear approach using Autosegmental framework. The analysis in this chapter is based largely on the corpus of data in these earlier works and Autosegmental framework is used in order to overcome the challenges of tonal representations that confronted the linear generative phonological analysis in the earlier works. We shall discuss; the relationship between Pitch and Tone, Tone Bearing Unit (TBU), Dagara Tonal System, Tone Combination in Dagara, Tone Stability, Tone Terracing, the Functions of Tone and Tonal Processes. 4.1. Pitch and Tone The relationship between Pitch and Tone in phonological analysis is of particular interest to Phonologists. Pitch and Tone relationship is such an intricate type that, it is virtually impossible to talk about tone languages without the mention of pitch. Brosnahan and Malmberg (1970:17) cited in Akanlig-Pare (1994:138) define pitc acoustically as t e “perceive impression o re uency or repetition o a wave” w ic implies t at, its easible to establis t e nature o pitc by counting the peaks of air pressure in the record of its waveform. University of Ghana http://ugspace.ug.edu.gh 125 La e oge (1993:186) also explains t at t e pitc o a soun is t at “au itory property that enables a listener to place it on a scale going from low to high wit out consi ering t e acoustic property”. A speec soun t at goes up in frequency also invariably goes up in pitch, thus the pitch of the sound is relative to its fundamental frequency. Frequency is a technical term for acoustic property of pitch and it is measured in Hertz (Hz.). Pitch can contrast across a wide range of frequencies, depending on several factors. Katamba (1989) notes that, pitch relates to the rate of vibration of the vocal cords. The tauter the vocal cords are, the faster they vibrate and the higher the pitch of the perceived sound. The rate of vibration of the vocal cords is also dependent on the thickness of the vocal cords and the amount of air flow through the glottis, (Akanlig-Pare, 1994: 139). Fox (2000), however, points out that pitch is not in itself a phonological feature; it is a phonetic feature with a variety of prosodic functions based on which languages in the world are categorizes phonologically as tone, accent, or intonation languages. Tone, on t e ot er an , “refers to the organization of speech into discrete, linguistically significant categories which serve (as do non-prosodic phonological characteristics) to discriminate among distinct linguistic units” (Clements, 1979:536). Pike (1948) explanation that, a tone language is one that has lexically significant contrastive, but relative pitch on each syllable has been the model definition in the literature. Akanlig-Pare (1994), following Cruttenden (1986), thus describes tone as a feature of the lexicon, in terms of, prescribe pitc or syllable. Yip (2007) also postulates t at a “language wit University of Ghana http://ugspace.ug.edu.gh 126 tone is one in which an indication of pitch enters into the lexical realization of at least some morp emes” (Yip, ibi : 230). Tone therefore, like segmental phonemes, can be used to distinguish lexical items or to convey grammatical distinctions in many tone languages in the world. 4.2. Tone Bearing Unit (TBU) Much as tone is a significant phonological phenomenon, especially in Asian and African languages, there have been divergent linguistic opinions on which phonological unit tone should be associated with. Yip (2002:73) admits that it is not always clear whether to associate tone to segments, syllables, or moras. Scholarly works such as Schachter and Fromkin (1968) and Gandour (1974) on tone in Akan and Thai languages postulate that vowels and consonants segments are Tone Bearing Units in Akan and Thai respectively. Goldsmith (1976) also proposes in autosegmental phonology that tone should be associated with vowels and vowels associated with tones, thus holding the same view as Schachter and Fromkin. However, Bao (1999:134), Fox (2000:217-8), and Yip (2002:67) among others, basing on the phenomena of ‘tone stability an ‘Floating tone , ebunk t e notion t at t e segment is t e tone-bearing unit. They argue that, in the event of phonological processes such as segment deletion, elision or contraction, metathesis and reduplication, tone can remain unaffected. University of Ghana http://ugspace.ug.edu.gh 127 Other works including Trubetzkoy (1967), Hyman (1985), Odden (1988) and Hayes (1989), considering syllable quantity, weight and length, also claim that the mora is the tone-bearing unit in phonological analysis. While works including Clements and Ford (1979), Hyman (1988) and Clements (1984), also hold the view that the syllable should be assigned tone in phonological analysis. For instance, Clements and Ford (1979) cited in Odden (1995:449) argue that: There has been some ambiguity in previous uses of the term tone- bearing unit. It is maintained here that tones are not directly associated with vowels or other segments, but rather with higher-level units (‘tone- bearing units ) suc as t e syllable or syllable-final (rhyme), in which vowels typically function as peaks of prominence (Clements and Ford, ibid:181). With dissenting opinions such as the above concerning which phonological unit is associated with tone, Hyman (1988:47) recommends that the issue of tone-bearing units should not be viewed as cross-linguistic but rather as language specific, since tone languages may differ in the choice of the tone- bearing unit. Akanlig-Pare (2005) consequently suggests a more lucid alternative for the choice of tone-bearing units in tone languages in the following statement: ….. i a language isplays a situation w ere t ere is one to one correlation of a tone to a mora, such that monomoraic light syllable bears only a single and for that matter, a simple tone such as a level, high, mid or low, whilst for a complex tone such as rising or falling contour, there must be a bi-moraic heavy syllable, there is the justification in claiming the mora as the TBU. However, if both light monomoraic and heavy bimoraic syllables can bear the same type of tone, simple and complex in one and the same language, then the TBU in that language must be the syllable (Akanlig-Pare, ibid:153-4). University of Ghana http://ugspace.ug.edu.gh 128 In most Gur languages, the syllable is considered the TBU because both light monomoraic and heavy bimoraic syllables can bear the same type of tone, simple and complex. Akanlig-Pare (2005:154) claims that the tone bearing unit in Buli is the syllable, because both light monomoraic and heavy bimoraic syllables can bear the same type of tone, simple and complex in the same language. Antilla and Bodomo (1996) cited in Yip (2002:141), also claim the syllable to be the tone bearing unit in Dagaare, of which Dagara is a dialect. Bemile (1983, 1984, and 1985) similarly support the claim that the syllable is the tone bearing unit, wit emp asis t at, “tone is an integral part o t e syllable” in Dagara (Bemile 1983:10). 4.3 Dagara Tone Systems A pioneering work on Dagaare phonology in Kennedy (1966) notes that, the language is a register tone type and postulates two levels of tone; High tone and Low tone, and a Downstepped. Bodomo (1997) and Antilla and Bodomo (1996) support this assertion with data limited to the Central dialect of the Dagaare language. Bemile (1983, 1984, and 1985) also note that there are five tonemes (three level tones) High, Mid, and Low, (and two contour tones) Falling contour tone and Rising contour tone. However, a critical analysis of tone in this work reveals that, there are three level tones but no underlying contour tone in Dagara, contrary to previous analysis by Bemile. The Rising and Falling Contuor tones are phonetic realization of a sequence of H, M, or L. For instance, a Falling tone could University of Ghana http://ugspace.ug.edu.gh 129 could be phonetic realization of H-L, H-M, or M-L, while a Rising tone could be as a result of L-H, L-M or M-H. In this analysis these tones are represented as follows: High (H) is marked with an acute accent ( ), the Mid (M) tone is the default ton and it is not marked with any diacritic sign, the Low (L) is marked wit a grave accent ( ), the Falling (F or HL) contour tone is marked with a circonflex accent ( ) an t e ising ( or LH) contour is marke a a ek sign ( ). The signs; (H) for High tone, (M) for Mid tone, (L) for Low tone, (HL) for Falling tones and (LH) for Rising tones, are used subsequently for the Autosegmental representation. 4.3.1 Level Tones Pike (1948:5) explains a level tone as the type in which, within the limits of perception, involves a level of steady pitch height during the production on the syllable. The three level tones; High tone, Mid tone and Low tone are tonemes, as they contrast in words to show lexical and grammatical distinctions in the Dagara dialect. 4.3.1.1 High (H) Tone The High Tone is produced with a relatively high pitch of the voice at the phonetic level. It can occur on all syllable shapes of various words types. The examples below show an autosegmental representation of the High tone. Fig. 41 High Tone a) pag b) y u o c) p i: r H H H University of Ghana http://ugspace.ug.edu.gh 130 p g ‘to meet y ‘open p :r ‘sheep d) t o m e e) s a b i k u b e H H t m ‘to be bitter s b k b ‘hail 4.3.1.2 Mid (M) Tone The Mid tones are produced with a relative pitch of the voice not as high as that of the High tones and not as low as the Low tone at the phonetic level. Its occurrence is on all syllable shapes. Some examples of words with Mid tone are shown in autosegmental representation below. Fig. 42 Mid tone a) k u r b) d u o r i c) ky l M M M kur ‘iron ore douri ‘to pile up ky l ‘jubilation d) k e r k e r e) b a g r f) k a r M M M kerker ‘raw bagr ‘s oul er kar ‘break into pieces 4.3.1.3. Low (L)Tone The Low tone on the other hand is produced with a relatively low pitch at the phonetic level. It may occur on all syllable shapes as well as on the various word types.Some word bearing Low tones are shown in the representation below. University of Ghana http://ugspace.ug.edu.gh 131 Fig. 43 Low Tone a) b a b) z ɛ l c) z :r L L L b “friend” zɛ l ‘tongue z :r ‘tail d) p ɔ l e) bagr f) L L L pɔ l ‘pride b gr ‘sacrifice ‘to do 4.3.2 Contour Tones Pike (1948:5) explains that, a contour tone is the type in which during the pronunciation of the syllable on which it occurs there is a perceptible rise or fall, or some combination of rise and fall, such as rising-falling or falling- rising. This research reveals that, contour tones in Dagara, develop as a sequence of level tones either as a result of the tone of one vowel spreading to a neighbouring vowel or as a result of vowel elision or a sequence of different level tones on a long vowel and they are not tonemes as previously analyzed. The following level tone sequences: Low-High, Low-Mid, and High-Low, may develop into Low-Rising, and High-Falling, contours respectively, similar to languages such as Yoruba and Gwari (Hyman and Schuh 1974). 4.3.2.1 Falling (HL) Contour Tone This is a concatenation of High and Low tones in the same syllable. This tone commonly occurs on syllable shapes such as V:C, and CV:C in monosyllabic or disyllabic words or as derived form resulting from some other phonolgical University of Ghana http://ugspace.ug.edu.gh 132 or morphological processes. During the pronunciation of the syllable on which they occur, there is a perceptible fall and rise or some combination of rising- falling of the pitch (Pike, 1948). Some examples of this type of tone in Dagara are shown in the representation below: Fig. 44 Falling Contour Tone a) ɔ:r b) ɓ a: g r c) ka:ŋ H L H L H L ɔ :r ‘a type o ruit ɓ :gr ‘taking k :ŋ ‘guinea fowl d) gb i: l e) p o: b f) gb i: b H L H L H L gb :l ‘whole p :b ‘act of sharing gb :b ‘art o kneeling 4.3.2.2 Rising (LH) Contour Tone This is a concatenation of Low and High tones in the same syllable. This toneme commonly occurs on syllable shapes such as CV, CVV and CV:C in monosyllabic or disyllable words or as derived form resulting from some other phonological or morphological processes. During the pronunciation of the syllable on which it occurs there is a perceptible rise and fall or some combination of falling-rising of the pitch (Pike, 1948). Examples of this type of tone in Dagara are shown in the representation below: Fig. 45 Rising Contour Tone a) g ɔ b) ba: gr c) pɔ : l L H L H L H g ɔ ‘ thorn b :gr ‘groin pɔ :l ‘a path University of Ghana http://ugspace.ug.edu.gh 133 d) t ɛ:r a e) gb ɔ:b r a f) k aso:g L H M L H L M L H tɛ ra ‘had gbɔ :br ‘pinching kas g ‘ in the wier basket 4.4. Tone Combination in Dagara The five tones may combine to give several different melodies on two or more syllabe words. The table below shows some possible tone melodies with word examples. Table 44. Tonal Combination Tone Melody Word Example Gloss HH p r tamarin LL pɔ lɛ paths MM pile mats HM p le small ram HL z ɛ millet LH b gr obstacle LM b le be rolling ML kas g a wierd basket M LH kas :g in a wierd basket H LH n :g in the kraal University of Ghana http://ugspace.ug.edu.gh 134 HL HL 'l :gr'l :gr very fast LHLH :ŋ :ŋ straight 4.5 Tone Stability Fox (2000) describes tone stability as the situation whereby the lost of a syllable or tone bearing unit through elision, contraction, or reduplication does not affect its associating tone, but it is preserved and maybe re-associated with the next syllable. This is a tonal feature which is difficult to explain and represent in linear generative phonology and thus, provides evidence to confirm of the independence of tone from segmental phonemes to support the superiority of autosegmental phonology over linear generative phonology. In Dagara, compounding processes or fast speech often result in vowel elision, while the tone is maintained. In the event of such syllable structure process, the stranded tone re-associates with the nearest vowel to the left or right. This tonal process is results into derived contour development which does not change the meaning of the word. The data in table 45 shows the re-association of tone in the event of a segmental deletion or elision. Table 45. Tone Sability / g ɔ / + / t ɛ / [ gɔ -t ɛ ] ‘t orn ‘tree ‘t orn tree / g r / + / v la / g r -v la ] ‘ erbal concoction ‘goo ‘goo erbal concoction / g / / w le / [g -w le ] ‘ebony ‘branc ‘ebony branc University of Ghana http://ugspace.ug.edu.gh 135 / kp r / + / faa / [ kp r-faa ] ‘s irt ‘ba ‘ba s irt / w / + / n / [ w n ] ‘come ‘wit ‘come wit An illustration of autosegmental representation of the above data is below: Fig. 46 Autosegmental Representation of Tone Stability. Underlying form : Tone Tier L H L Skeletal Tier X X X X X X Segmental Tier / g ɔ / + / t ɛ / Back, High Vowel Deletion : Tone Tier L H L Skeletal Tier X X X X X X Segmental Tier / g ɔ t ɛ / Low tone floating : Tone Tier L H L Skeletal Tier X X X X X X Segmental Tier / g ɔ t ɛ / University of Ghana http://ugspace.ug.edu.gh 136 Floating tone re-associating to the right Tone Tier L H L Skeletal Tier X X X X X Segmental Tier / g ɔ t ɛ / Output Form : Tone Tier L H L Skeletal Tier X X X X X Segmental Tier [ g ɔ t ɛ ] ‘t orn tree Underlying Form : Tone Tier L H M Skeletal Tier X X X X X X X Segmental Tier / k p a r / + / f a: / ‘s irt ‘ba Vowel Deletion : Tone Tier L H M Skeletal Tier X X X X X X X Segmental Tier / k p a r f a: / University of Ghana http://ugspace.ug.edu.gh 137 High tone floating : Tone Tier L H M Skeletal Tier X XX XX X X Segmental Tier /k pa r f a: / Floating tone re-associating to the left : Tone Tier L H M Skeletal Tier X X X X X X Semental Tier / k pa r f a: / Output Form : Tone Tier L H M Skeletal Tier XX XXX X Segmental Tier [k pa r-f a: ] ‘ba s irt 4.6 Tone Terracing Tone terracing is a property of many tone languages, especially, the sub- Saharan Africa languages. Tone terracing occurs in a language when the pitch of successive basic tones are either raised or lowered over an utterance (Akanlig-Pare, 1994:150). Though Dagara is a discrete-tone-level Language, it exhibits such characteristics of tone terracing over an utterance which can be escribe as “ eclination”, ollowing onnell & La (1990) an ( onnell 2001) and downstep . University of Ghana http://ugspace.ug.edu.gh 138 4.6.1 Declination According to Connell & La (1990:2), eclination is “a gra ual mo i ication (over the course of a phrase or utterance) of the phonetic backdrop against w ic t e p onologically speci ie F0 targets are scale ”. onnell (2001) further explains that, while in tone languages it can occur regardless of tonal combination (i.e. in mixed tone sequences), the existence and degree of declination can most clearly be seen in a phrase consisting of tones all of which have the same phonological value. In Dagara, the intonation pattern over an utterance sometimes shows a kind of falling of the pitch level. When an unbroken sequences of high tone spans over an utterance, there is a slight lowering of pitch from the beginning to the end of the utterance. For instance, in the production of the phrase; n r tɔ ‘bulls heads are very big , t ere is a perceptible eclination o t e ig pitc level from the beginning to the end. This is some what similar to the phenomenon described in Connell and Ladd (1990) and Connell (2001). Figure 47 illustrates the declination in the sequence of Dagara High tones in the phrase n r tɔ ‘bulls ea s are very big . Fig.47 Declination illustrated in sequences of High tones 200 100 H H H H HH H b r tɔ University of Ghana http://ugspace.ug.edu.gh 139 The phenomenon of declination is not only peculiar to Dagara, it is attested in Hausa language in the literature. Lindau (1986) cited in Connell (2001) using data from Hausa illustrates declination in a sequence of Hausa High tones in t e sentence ‘ y g ( uu ii came ome) as in figure 48 below: Fig.48 Declination illustrated in sequences of Hausa High tones 200 100 H H H H H H Muudii yaa zoo gi aa 4.6.2 Downstep Downstep refers to the lowering of a High tone in a given environment (Connell 2001). Stewart (1965) first acknowledged that in a sequence of two High tones, the second High tone is lower in pitch than the first tone. Also, in a sequence of High-Low-High tones, the High tone after the Low tone is lowered in pitch than the first High tone. In Dagara, when Low and High tones are in sequence stretching over an utterance, the High tones following the Low tones drift downward while the Low tones maintain their pitch. For instance, in the production of the sentence; m ɔ g w r 'l ɔ r ɔ ‘a fool caught a horse and University of Ghana http://ugspace.ug.edu.gh 140 tie it at t e river bank an it escape to t e market , the High tone following the Low tone are stepped lower than the underlying High pitch level. This downward trend is illustrated in the diagrams in figure 49 below. Fig 49 Downstep in the sequence of Dagara Low-High tones L H L H L H L H L H L L a) L H L H L H L H L H L L The diagrams above illustrate the sequences of Low and High Pitches in an utterance. Diagram a) illustrates the relative pitch levels in their underlying form, while b) illustrates the out put form in which the High pitch following the Low pitch is lowered to sound lower than its underlying form. 4.7. The Functions of Tone in Dagara Generally, tone languages employ pitch in speech to depict semantic and syntactic distinctions, otherwise, known as Lexical and Grammatical tone. This section thus, discusses the lexical and grammatical functions of tone in Dagara. 4.7.1 Lexical Tone When tone is used to differentiate the meaning of lexical items, as they appear in isolation in a lexicon, then it is playing a lexical function. Meanings of lexical items with the same segmental phonemes depend on the tonal patterns University of Ghana http://ugspace.ug.edu.gh 141 for their distinctions. The lexical items so distinguished by tone patterns may or may not belong to the same lexical or grammatical category. In Dagara, the tonemes can pattern in several different melodies on words for the purpose of distinguishing meanings. Below are examples of some attested contrastive tonal patternings on words to create such semantic distinctions between them. 4.7.1.1 Minimal Pairs Fig.50 Minimal contrastive Pairs L H i r i r r ‘get up/ rise r ‘to c oose/ elect H L pu o r p u o r p r ‘stomac p r ‘back/be in 4.7.1.2 Minimal Triplets Fig. 51 Minimal Contrastive Triplets H M M L M bu l e b u l e b u l e b le ‘gerninating bule ‘cistern b le ‘ki H L M L kp e l e kp e l e kp e l e kp l ‘ break at once kp le ‘termites kp l ‘type of disease University of Ghana http://ugspace.ug.edu.gh 142 4.7.1.3. Minimal Quadruplets Fig. 52 Minimal Contrastive Quadruplets H M L bagr bagr bagr b gr ‘grinding bagr ‘shoulder b gr ‘sacri ice H M L M M L H kogle kogle kogl e kogle k gle‘ small oak tree k gle ‘small stool kogle ‘ rie up k gl ‘emanciating 4.7.2. Grammatical Tone When pitch is employed to indicate tense, aspect and mood, as well as to distinguish content morphemes from functional morphemes which in all respect have the same segmental features, then it is assuming a grammatical tone function in the language. In some tone languages, especially the sub- Saharan African Languages pitch variation in the verb stem or noun stem signals a change in the grammatical category, or different tone patterning over a stretch of an utterance discriminates between tense, aspect and mood. In Dagara, content words (verbs, nouns, adjectives and adverbs), cannot be associated with any particular tone or tonal patterns in a regular manner. It is rather functional morphemes, (free or bound morphemes), which are identified regularly with some tones in given environments. This section discusses the type of tonemes that regularly associate with the morphemes to encode grammatical function in the language. Specifically, we shall focus on tones which distinguish; Nominative/Possessive pronouns from University of Ghana http://ugspace.ug.edu.gh 143 Accusative/Dative pronouns, Aspect, Mood (indicative, imperative and Subjunctive II), Negations, Locative expressions and Nominal Suffix Morphemes. 4.7.2.1 Nominative/ Possessive and Accusative/Dative pronouns In Dagara, all the pronouns take the same segmental form. Except for the first person singular nominative and possessive pronouns and their accusative/dative forms, the only means of distinguishing the nominatives/ possessives from their accusatives, is by means of their tone differences. Whereas the nominative/possessive pronouns bear Low tones, their accusative/dative forms are Mid tone in order to distinguish them respectively. The table below shows the Low and Mid tonal contrast that distinguishes the nominative and possessive pronouns, on one hand, from their accusative and dative forms the other hand. Table 46. Pronouns Nominative/Possessive pronoun Accusative/ Dative pronoun / / ‘I , ‘My / mɛ / ‘me / f / ‘you , ‘your / f / ‘you / / she/he , ‘his/her / / ‘him/her / t / ‘we , ‘our / t / ‘us / ny / ‘you , ‘your / ny / ‘you / bɛ / ‘they , ‘their / bɛ / ‘them / / ‘it , ‘its / a / ‘it University of Ghana http://ugspace.ug.edu.gh 144 In an utterance which has the same forms, for example, the third person singular or plural [+ Human] forms, it is only by the tonal pattern that the distinction can be made as in the sentences below: Fig. 53 Use of Tone to Distinguish Pronouns L M L M L M L / b k bɛ n tome / / kul na / ‘ t ey gave t em work ‘ she/he married er/ im . 4.7.2.2. Aspect Tone Dagara has two aspectual forms; perfective and imperfective or progressive forms. The perfective aspect usually takes the same form as the present and past indicative.The imperfective or progressive, however, is basically marked with a verbal suffix -re/-rɛ or the variants -ne/-nɛ and –e/-ɛ depending on the phonological condition of the verb stem (cf. Section 3.1.1.1 of Chapter Three). The tone of the suffix is always a Mid tone irrespective of the tone in the verb stem. Below are sentences showing the imperfective forms of the verbs. Fig 54 Mid Tone for Aspect Marking L M L Present Progressive: / ir-e na / ‘ I am getting up L M L Future progressive: / na ir-e na / ‘I will be getting up University of Ghana http://ugspace.ug.edu.gh 145 L H M L Present progressive: / f ir-e na a pɔg / ‘you are c oosing t e woman L H M L Future Progressive: / f na ir-e na a pɔg / ‘you will c oose t e woman L M L Present Progressive : / yi-re na / ‘ / e is going out L M L Future Progressive: / na yi- re na / ‘ She/ e will be going out 4.7.2.3. Mood (Subjunctive II) In Dagara, the expression of Mood is in three forms, namely; the indicative, imperative and subjunctive. Though verbal particles maybe used to encode the type of mood, tone also functions on the verb to indicate mood. For instance, the particle that encodes the indicative mood is / na / with a low tone and that of the subjunctive II is / naa / with a Mid tone. But the imperative mood is expressed in the infinitive form of the verb with its underlying tone without a verbal particle. In the subjunctive II Mood, the tone of the verb changes to a Mid tone in conformity with tone of the particle, irrespective of the underlying tone of the verb in the infinitive form. The following constructions illustrate the imperative, the indicative and the subjunctive II moods. University of Ghana http://ugspace.ug.edu.gh 146 Fig.55 Mid Tone for Subjunctive Expression Imperative: L / ir / ‘get up Indicative: Present Progressive: L M L / ir-e na / ‘ I am getting up Subjunctive II: L M L /f naa ir- e na / ‘You s oul ave been getting up Imperative : H / ir / ‘ c oose Indicative: Present Progressive: L H M L / ir-e na a pɔg / ‘ I am c oosing t e woman Subjunctive II: L M L / naa ir-e na a pɔg / ‘ I s oul ave been c oosing t e woman 4.7.2.4. Negation Negation in expressed by the use two split morphemes which are characterized by an underlying Mid tone. The first components of the negation morpheme /bɛ/ and /ta/, for indicative and imperative respectively, are invariable. While the second components /i, , e, ɛ/ are allophonic depending on the [ATR] vowel University of Ghana http://ugspace.ug.edu.gh 147 quality of the verb that preceeds them, as in ‘bɛ…… /i, bɛ.......ɛ/e an ‘ta…… /i, ta..........ɛ/e , for the indicative and the imperative Moods, respectively. Below are some negative constructions in Dagara which demonstrate the consistent underlying tone of the negation morphemes. Fig. 56 Tone for Negation Indicative Imperative L M H M M H M / f bɛ kyer e e / / ta kyen e / ‘You are not going ‘Do not go L M L M M L M / bɛ d ɛ / / ta d ɛ / ‘ I ave not eaten ‘ Do not eat 4.7.2.5. Locative Expressions Location in Dagara is expressed in two ways; by the use of free morphemes, which are mainly body parts expressions and by the use of tone. When a body part expression is used, it is placed postpositionally relative to the noun that is being located. When tone functions grammatically to show locations, depending on the underlying tone of the noun that is being located, a High tone or Mid tone distinguishes the nominative from the locative form. Where the nominative is mono-syllabic with a Low tone, a High tone is added and it becomes a Rising Contour in the locative form. But in disyllabic or trisyllabic nominatives, if the final syllable bears a Low or Mid tone, it changes to a High tone in the locative form. Finally, if the nominative is mono-syllabic with a University of Ghana http://ugspace.ug.edu.gh 148 long vowel of identical High tone, the second tone changes to a Mid tone in the locative form. The following tonal patterns of the examples below illustrate the distinction between nominative from locatives: Fig. 57 Tone for Locative Expressions Nominative Locative L M L H / davoro / / davoro / ‘court yar ‘in t e court yar L L H / ŋ / ŋ ‘room ‘ in t e room L L H /bɔg/ / bɔg/ ‘ ole ‘in t e ole H H M /n / /n / ‘ an ‘in t e an 4.7.2.6. Suffixal Morphemes Generally, suffix morphemes in Dagara are characterized with an underlying Mid tone. The verbal suffix which encodes the imperfective (cf. Section 4.7.2.3) always remains a Mid tone irrespective of the tone pattern of the verb stem to which it is affixed. Similarly, the nominal suffixes; - /-u, -l /-lu, -f /-fu (cf. Section 3.1.1.2 and 3.1.13 chapter 3), which are affixed to other content word (verb, noun, University of Ghana http://ugspace.ug.edu.gh 149 adjectives or adverb) stems to derive new forms of nouns, are also characterized by a Mid tone irrespective of the underlying tone patterns of the stems to which they are affixed. The diminutive suffixes, -le ‘singular or –li ‘plural are also associated with Mid tone. 4.8 Tonal Processes The juxtaposition of tonemes usually triggers some processes in tone languages similar to the phonological processes that result from the juxtaposition of segmental phonemes. Fox (2000) notes that, tonal processes are phonological processes which affect tones, changing or modifying them in particular contexts and thereby obscuring their identity. Tonal processes are universal phenomena in tone languages but language specific rules apply to account for them. This section discusses the tonal processes and the phonological conditions that trigger them in Dagara. The processes identified in Dagara include; Tone Assimilation, Contour Tone Simplification and Tone Dissimilation. 4.8.1. Tone Assimilation Hyman and Schuh (1974) explain that if two syllables differ in tone, the tone of the first syllable naturally enlarges its domain by spreading into the second syllable, a process similar to segmental feature assimilation in the progressive or preservative process. Tone assimilation in Dagara is a synchronic process which takes place always to the right hence it is a progressive or preservative assimilatory process type. University of Ghana http://ugspace.ug.edu.gh 150 In Dagara, this process involves a Mid tone spreading from a preceding syllable onto the following syllable with either a Low or High tone. Specifically, when a syllable with long vowel bearing a Mid tone precedes a syllable with Low or High tone, the Mid tone enlarges its domain to assimilate the following tone. Evidence of this process is sourced from the resultant change of the underlying tone in the verb in the infinitive to Mid tone when the subjunctive II Mood particle /naa/, which is a syllable with a long vowel bearing a Mid tone, precedes it. The examples below illustrate the derivation process of the subjunctive form from the verbs ‘go an d ‘eat in t e in initive orm. Fig. 58 Tone Assimilation Underlying Form. L M H / naa kyen / Mid Tone Spreading L M H / naa kyen / Delinking of High Tone L M H / naa kyen / Output Form L M / naa kyen / ‘ / e s oul ave gone University of Ghana http://ugspace.ug.edu.gh 151 Underlying Form L M L / naa d / Mid Tone Spreading L M L / naa d / Delinking of Low Tone L M L / naa d / Output Form L M / naa d / ‘She/ e s oul ave eaten 4.8.2 Contour Tone Simplification Tone simplification is a synchronic tone rule like other tonal processes. It is the process where a contour tone tends to be levelled or converted into level tones in a given environment. Hyman and Schuh (1974) note that in some tone languages, contour tones are simplified to a non-low tone in the environment of non-low tones. In Dagara, contour tone simplification is achieved by means of tone absorption and tone deletion. 4.8.2.1 Tone Absorption According to Hyman and Schuh (1974), the absorption process “involves a contour tone followed by a tone identical to the end point of the preceding contour. [In such sequences], the second component of the contour, by virtue of its identity with the following tone, is absorbed into the next syllable” University of Ghana http://ugspace.ug.edu.gh 152 (Hyman and Schuh, ibid: 90). Tone absorption is also a synchronic tone rule in Dagara. Tone absorption differs from tone assimilation on the basis that, the tone in absorption does not enlarge its domain from the preceding syllable onto the following syllable, it rather moves from its domain to be absorbed by the following tone in the next syllable. The process of tone absorption involves a movement of tone from the end point of the contour to the contiguous syllable of identical tone. Tone absorption is particularly common in tone languages in which contours can develop as a result of sequences of Low-High tones or Low-Mid tones. In Dagara, the end point of a Rising tone is absorbed when it is followed by a syllable with a non-low tone. Tone absorption usually occurs during compounding process, particularly when a noun is qualified by an adjective. In this compounding process the long vowel with contour tone of the noun stem is shortened and the High tone which remains subsequently gets absorbed by the High or Mid tone of the following adjective. The High point of the Rising contour tone delinks and moves to be absorbe by the High or Mid tone of the following adjective. The resultant tonal pattern of the compound will thus be either L-H or L-M. Evidence of tone absorption is shown in the data in table below: University of Ghana http://ugspace.ug.edu.gh 153 Table 47. Non-Low Tone Absorption Noun Adjective Compound / kp :gl / / b r / [ kp gl b r] ‘belt ‘see ‘ belt stran / :gr / / y g / [ gr y g ] ‘ irt ‘ plenty ‘ plenty irt / pɔ :l / + / w g / [pɔ l w g ] ‘ pat ‘long ‘long pat / b :gr / + / k bl / [ b g k bl ] ‘groin ‘ air ‘pubic air / pɔ :l / + / faa / [ pɔ l faa ] ‘pat ‘ba ‘ba pat The representations show the absorption process in the non-linear order. Fig. 59 Autosegmental Representation of Non-low tone absorption Input Form : L H H /p ɔ: l / /w o g o / Vowel Shortening L H H /p ɔ l / / w o g o/ High Tone Delinking and Movement L H H /pɔ l / / w o g o / University of Ghana http://ugspace.ug.edu.gh 154 Output Form L H [p ɔ l w o g o ] ‘long pat 4.8.2.2 Tone Deletion This process usually occurs in tone languages in which sequences of High-Low or Mid-Low develop in to a Falling contour, such that when the contour is followed by a High or Mid tone in compounding process, the resultant tonal pattern of the compound will sequences of level tones. In Dagara, contour tones simplification by deletion pertains strictly to Falling contour tone. A Falling contour tone disintegrates into non-low (High or Mid) level tones when it is followed by a non-low tone (High or Mid tones) in noun- adjective compounds. In this compounding process the long vowel with contour tone of the noun stem is shortened and the Low point of the Falling tone deletes before the adjective with High or Mid tones follows. The resultant tonal pattern of the compound, thus become either High-High or High-mid. Evidence of tone simplification is shown in the following data below: Table 48. Contour Tone Simplification by Deletion Noun Adjective Compound / kɔ :r / + / kpɛ ɛ / [ kɔ :r kpɛ ɛ ] ‘cheeck ‘big ‘big cheek / k :ŋ / / m m r / [ k ŋ m m r ] ‘guinea fowl ‘living ‘living guinea owl / pɛ :r / + / b r/ [ pɛ r b r ] ‘anus ‘see ‘buttock University of Ghana http://ugspace.ug.edu.gh 155 / k :ŋ / + / k : / [ k ŋ k : ] ‘guinea fowl ‘ ea ‘ ea guinea owl / pɛ r / + / koŋ / [ pɛ r koŋ ] ‘anus ‘lean The autosegmental representations of tone simplification are illustrated below Fig.60 Contour Tone Simplification by Deletion representation Input Form H L H / k :ŋ/ / mimir / Vowel Shortening and Low Tone Delinking H L H / k ŋ/ + / mimir/ ‘guinae owl ‘living Output H [ k ŋ mimir] ‘living guinea fowl 4.8.3 Tone Dissimilation Another tone process common in tone languages, which is often considered as a synchronic, is dissimilation. Hyman and Schuh (ibid: 100) explain that, a syllable that is assigned an underlying tonal representation, but when it is in proximity with a syllable of an identical tone, it changes its underlying tone. In the literature languages like Mandarin and Fe?Fe? are cited for the case of dissimilation (Hyman and Schuh ibid:101). Hyman and Schuh further note that, McCawley (I970) mentions that, the dissimilation rule of Mandarin is L-L becomes R-L. University of Ghana http://ugspace.ug.edu.gh 156 This tonal phenomenon is also observed in Dagara, quite similar to Mandarin, on one hand, when a syllable of an underlying Low tone is in proximity with another syllable of an underlying Low tone, the tonal pattern changes to either Rising-Low in one instance, but differs on another hand when Low-Low pattern results in a Low-Falling pattern in another instance. In the first instance, when a verb with an underlying Low tone precedes a verbal particle with identical tone or a noun with an underlying Low tone is qualified by an adjective of identical tone, the tonal pattern changes from Low- Low to Rising-Low. Evidence of Low-Low tones dissimilating to become Low-Rising tonal pattern is in the data that follows below: Table 49. Low –Low Tone dissimilation to become Rising-Low Verb Verbal Particle Verbal Phrase / zɔ ŋ / / n / [zɔ ŋ n ‘to be blin e affirmative ‘ma e blin / wo ŋ / + / n / wo ŋ n ] ‘to be ma e ea ‘a imative ‘ma e ea / gb / / n / [ gb n ] ‘to be blunt affirmative ‘ma e blunt Noun Adjective Adjectival Phrase / zɔ ŋ / / ka ŋ / [ zɔ ŋ ka ŋ ] ‘blin person ‘certain ‘certain blin person / wo ŋ / ka ŋ / [wo ŋ ka ŋ ] ‘dea person ‘certain ‘certain ea person / gbe / / ka ŋ / [ gbe ka ŋ ] ‘ ore ea ‘certain ‘certain ore ea The autosegmental representation of the above process is illustrated below: University of Ghana http://ugspace.ug.edu.gh 157 Fig. 61. Representation of Low Tones Dissimilating to Rising- Low pattern Underlying Form Output Form L L H L / z ɔ ŋ / + / na/ [ zɔ ŋ na ] ‘to be bin e ‘a irmative ‘ma e blin Input Form Output Form L L H L / zɔ ŋ / / k ŋ / [ zɔ ŋ k ŋ] ‘blin person ‘certain ‘certain blin person The second instance is, when the subject personal pronouns which bear Low tones precede verbs with underlying Low tones, the tone of the verb changes from Low to Falling tone. Evidence of Low-Low tones dissimilating to become high-Falling tonal pattern is in the data that follows below: Table 50. Low Tones dissimilating to Low-Falling tone pattern Subject Pronoun Verb Phrase / / + / zɔ ŋ / [ zɔ ŋ ] 1 G ‘to be blin ‘s oul I be blin ? / f / + / tɔ / [f tɔ ] 2SG ‘to touc ‘s oul you touc ? / / + / d / [ ] 3SG.+HUM ‘to eat ‘ / e s oul eat / / + / ga ŋ / [ ga ŋ ] 3SG/PL.-HU ‘ to cross ‘it/t ey s oul cross / t / / kp / [ t kp ] 1PL.+ HU ‘ ie ‘let us ie University of Ghana http://ugspace.ug.edu.gh 158 / ny / + / v / [ ny v ] 2PL ‘t res ‘you s oul t res / bɛ / / y g / [ bɛ y g ] 3PL ‘take o ‘t ey s oul take o An illustration of autosegmental representation of the above data below: Fig. 62 Low Tones dissimilating to Low-Falling tone pattern Underlying Form Output Form L L H L / / + / zɔ ŋ / [ zɔ ŋ ] I ‘be blin ‘s oul I be blin ? L L H L / / + / d / [ d ] 3SG.+HUM ‘to eat ‘ / e s oul eat 4.9 Chapter Summary This Chapter discussed genreal issues about tone including Pitch and Tone relationship and Tone Bearing Unit in tone languages. It also examined the nature of tone in Dagara, highlighting tonal feature such as Tone Stability and Declination and Downstep and illustrating how they manifest in Dagara. It finally discussed lexical and grammatical Function of Tone as well as some Tonal Processes in Dagara which include Tone Assimlation, Contour Tone Simplification and Tone Dissimilation. The analysis and phonological representations of tone and tonal processes are done within the framework of Autosegmental Phonology. University of Ghana http://ugspace.ug.edu.gh 159 CHAPTER FIVE SUMMARY AND CONCLUSION 5.0 Introduction This Chapter contains Summary of what has been discussed in the various Chapters of the Thesis and Conclusion. 5.1. Summary In chapter One, I have given a brief statement on Dagara and its speakers with a sketch of dialectal variation in the language. I indicated that there are four main dialects: Northern Dagaare (Dagara), Central Dagaare, Southern Dagaare, and Western Dagaare (Birifɔr); which differ considerably at the phonological, lexical and grammatical levels. I then gave a genealogical sketch of Dagaare language family. I also discussed the Problem Statement, the Scope of the research, the Aims and Objectives of the research, the Significance of the research as well as the data and the Methodology involved in the data collection. I then gave an explanation of the Theoretical Framework (Autosegmental phonology) within which the work is situated, with emphasis on how it evolved and its basic tenets as well as the motivation for using the framework in the phonological and tonal representations. I ended the chapter with a review of relevant literature in both the language and the Theoretical Framework and the organization of the thesis. University of Ghana http://ugspace.ug.edu.gh 160 In Chapter Two, I discussed the sound system of the Dagara dialect highlighting on the vowel and consonant inventory, the phonetic description of the sounds, the syllable structure and the structure of words in Dagara. I pointed out that there are eighteen vowel phonemes; nine short oral vowels and nine short nasal vowels, with all the eighteen short vowels having their long contrastive counterparts. And also that all the vovels are subcategorized into two distinct sets: [+ATR] vowels and (-ATR) vowels with strict co- occurrence constrictions. I also indicated that, there are twenty-seven consonant phonemes; twenty-five consonants and two approximants, whose distribution in words are constrained. All consonants maybe at word-initial and/or word-medial, except /r/, which may only be at word medial or in secondary syllable onset position or coda position whilst word-final position is exclusively occupie by / b, g m, n, ŋ, l, r /. I also described and classified the Dagara sounds into natural classes within the Distinctive Feature theory framework, under Major class features, Cavity feature, Manner features and laryngeal feature. I then examined the syllable structure and observed that Dagara belongs to the group of languages with V, VC, CV, CVC and C type with varied syllable shapes. I also highlighted that the sequential constraints that operate at the beginning and end of the syllable are reflect on words whether simple stem or complex stem. Chapter Three which constitutes the core part of the thesis, discussed both assimilatory and syllable structure processes and formalized them within University of Ghana http://ugspace.ug.edu.gh 161 Autosegmental framework. For the assimilatory processes, I identified and discussed Vowel Harmony, Consonant Nasalization, Homorganic Nasal Assimilation, Glide formation, Consonant Labialization, and Rhotacism. I observed that, there is a very strict vowel harmony system in Dagara, which operates bi-directionally, involving stems and suffixes, [-ATR] noun stems and a [+ATR] dominant stem and across word boundaries as well, except for the nominal suffix /-f / blocks Vowel Harmony process. I noted too that all the assimilatory processes in Dagara involve Place of Articulation and Manner of Articulation features. For syllable structure processes, Elision (both vowel and consonant), Compensatory Lengthening, Liaison or Linking and finally Vowel epenthesis are identified and discussed. Chapter Four discussed issues that border on Pitch and Tone distinction, pointing out that, pitch is a phonetic feature whereas tone is a phonological feature like segmental phonemes. I discussed the controversial issue of Tone Bearing Unit in tone languages, the Tone System and Tonal features such as Tone Stability, Declination and Downstep in Dagara. I also looked at the Functions of Tone and finally identified and discussed the phonological conditions which triggered some tonal processes such as Tone Assimilation, Contour tone Simplification, and Tone Dissimilation. The analysis and phonological representations of tone and tonal processes are within the framework of Autosegmental Phonology. University of Ghana http://ugspace.ug.edu.gh 162 5.2. CONCLUSION Though this thesis may seem quite elaborate, it cannot be said to be exhaustive. There may still be some phonological and tonal processes that have not been identified and discussed in this work. This is just the first attempt to discuss the phonological and tonal processes in Dagara, it is therefore meant to trigger academic review and to set the pace for future research work in the dialect, as well as in other dialects of the Dagaare. On the theoretical perspective, the formalization of the phonological and tonal processes wit in t e Autosegmental ramework is, “a natural testing-ground for the thesis that autosegmental phonology is not restricted to tonal phenomena, but constitutes a general theory of phonological representation”, ( lements, 1976: 57). University of Ghana http://ugspace.ug.edu.gh 163 APPENDIX A: SUMMER INSTITUTE OF LINGUISTICS, WEST AFRICA AREA WORD LIST 1. S/N LIST ID LIST GLOSS (ENGLISH) DAGARA WORD 1 10 eye m m r 2 20 ear t br 3 30 nose nyɔ ɔ r 4 40 mouth n ɔ r 5 43 tooth ny m 6 45 tongue ɛ l 7 48 lips nɔ gb mb 8 60 chin t ɛ m 9 70 jaw y glɛr 10 80 beard t ɛ ŋk bl 11 90 cheek kɔ r 12 100 face n ŋe 13 110 forehead gbe 14 120 head 15 150 hair (head) k bl 16 160 hair (body) ŋg ŋkɔ bl 17 170 fur k bl 18 180 feather (s nɔ k bl 19 190 horn l 20 200 neck ny 21 210 throat kɔ kɔ r 23 220 shoulder bagrgm n 24 230 armpit b gl ŋ 25 236.1 the bush m ɔ 26 250 breast b r 27 252 udder b gr 28 260 milk (mother's) b rk ɔ 29 262 milk (animal's) b r 30 265 butter ka a 31 267 cheese b sɛ n ɛr 32 280 chest ny 33 310 belly p r 34 320 navel ny 35 340 side of body l mb gr 36 342 rib ny b r 37 360 back p r University of Ghana http://ugspace.ug.edu.gh 164 38 370 buttocks p b 39 380 tail z r 40 400 leg gbɛ r 41 410 hip s ŋkp gr 42 420 thigh gbɛ kpɛ ɛ 43 440 knee 44 450 foot gbɛ n ɓɛ l 45 460 sole of foot gbɛ s l p ɔ 46 500 arm (forearm) kp ŋkp ŋ 47 505 wing b gr 48 540 elbow kp ŋkp ŋnyu gbi l 49 550 hand n 50 560 palm n s l p ɔ 51 580 finger n b r 52 582 thumb n b aa 53 585 toe gb bir 54 590 fingernail n b pɛ gr 55 591 claw gb b ɛ mɛ 56 600 body a ŋ 57 610 skin (of man) a ŋg n 58 611 hide (animal skin) g n 59 620 wound n t r 60 621 scar l 61 630 bone kɔ br 62 640 meat; flesh nɛ n 63 650 fat / grease ka a 64 651 oil ka a 65 660 vien gy lu 66 680 egg gyɛ l 67 700 blood 68 710 saliva m t ar 69 720 tears m t k ɔ 70 740 urine d r 71 750 sweat x n 72 760 feces / excrement b n 73 761 dung n gb n 74 800 heart s kyir 75 830 liver sarbr 76 860 guts /bowels ny gɛ 77 890 brain kyap or 78 1000 person n r/ n saalɛ 79 1050 chil (child) b bile 80 1080 elder n kpɛ ɛ University of Ghana http://ugspace.ug.edu.gh 165 81 1100 man (vir) ɛ b 82 1101 husband s r 83 1150 son bi ɛ b 84 1151 boy ble 85 1200 woman pɔ g 86 1201 wife pɔ g 87 1250 daugther pɔ gy a 88 1251 girl p le 89 1300 father s 90 1310 mother m 91 1320 brother yɛ b 92 1321 elder brother yɛ kpɛ ɛ 93 1322 younger brother y bile 94 1330 sister y p le 95 1331 elder sister y p kpɛ ɛ 96 1332 younger sister y p bile 97 1340 friend b / b pɛr 98 1350 mother's brother m ɛ b 99 1355 father's sister pur 100 1360 child (offspring bie 101 1365 twins y br 102 1380 stranger s n 103 1385 guest s n 104 1390 enemy ɔ ɔ mɛ 105 1400 chief /king n 106 1410 owner s b 107 1415 slave gba gb a 108 1430 judge sɛr rɛ 109 1450 God n ŋm n 110 1460 spirit /ghost nya a kp n 111 1461 shadow s le 112 1470 name y r 113 1480 voice kɔ kɔ r 114 1485 language kɔ kɔ r 115 1490 story (tale) s lu 116 1500 animal d ŋ 117 1502 wild animal wɛ ŋ 118 1505 domestic animal yird ŋ 119 1508 sacrifice b gr 120 1610 dog baa 121 1620 jackal gbo bg rky le 122 1630 hyena gb gb r 123 1650 cat nya a nyu University of Ghana http://ugspace.ug.edu.gh 166 124 1660 lion gbe ŋ 125 1670 leopard nyu 126 1710 rat ɛ rbaa 127 1720 bat (fruit) a a b ŋ 128 1750 elephant wɔ b 129 1760 hippo ɛ n 130 1770 buffalo wɛ n b 131 1780 monkey ŋm m 132 1790 baboon kp gt r 133 1810 goat b ɔ 134 1820 sheep pɛ r / p r 135 1850 pig baa 136 1890 horse w r 137 1900 cow (cattle) n b /n 138 1910 cow (female) n pɔ g 139 1920 bull n ɛ r 140 1930 ox n ɛ k ɔra 141 2000 bird lile 142 2010 chicken lile 143 2012 hen nɔ pɔ g 144 2015 cock n ra 145 2020 eagle s la 146 2030 vulture sigdum 147 2040 guinea fowl k ŋ 148 2050 dove ŋm m 149 2060 pigeon n ŋm m 150 2110 tortoise (land) k r 151 2111 tortoise (water) k rn m ɔ 152 2120 lizard bandaa 153 2130 crocodile b 154 2150 snake w b 155 2190 worm (earth-) s bɛ rnyɛ 156 2200 fish m 157 2250 crab ga g ra 158 2310 louse ky b 159 2320 fly (house) naz ɔ 160 2330 honey s gr 161 2332 honeybee s gb r 162 2340 grasshopper sa s gr 163 2350 ant (soldier) l 164 2360 termite y gra 165 2370 spider b ɛr / ɛ pɛ ɛ rɛ 166 2380 scorpion na ŋ University of Ghana http://ugspace.ug.edu.gh 167 167 2500 tree t ɛ 168 2510 bark (tree) t pɛ gr 169 2520 leaf v r 170 2530 branch w l 171 2535 stick/ club daa 172 2540 root ny gr 173 2550 thorn g ɔ 174 2570 flower p r 175 2580 fruit wɔ m 176 2590 seed b r 177 2600 grain k b r 178 2610 wheat m ɔ b ŋyamb 179 2620 millet ky 180 2630 barley ? 181 2640 maize kam n 182 2650 rice m n 183 2660 banana (s) sweet kudu 184 2665 plantain bɔ ɔ ɛ 185 2670 palm tree (eg. Date bɛ t ɛ 186 2680 yam ny r 187 2685 manioc /cassava ba ky 188 2690 groundnut /peanut s m r 189 2710 tobacco tɛ mɛ 190 2720 grass m ɔ 191 2725 thatch k mp l 192 2750 forest wɛ gya 193 2760 "the bush" k rp ɔ 194 2770 desert /wilderness gb gb la 195 2790 field (s) w ɛ 196 2800 place 197 2810 country t ŋ 198 2820 tribe /ethnic group kɔ kɔ r 199 2830 village t mbile 200 2840 home /compound y r 201 2850 house /hut y r/ g r 202 2860 roof g r 203 2870 loft /granary b gr 204 2880 wall ky ne 205 2890 door n ɔ r 206 2900 gate p n 207 2910 fence ? 208 2930 path s le 209 2940 road sɔ rkpɛ ɛ University of Ghana http://ugspace.ug.edu.gh 168 210 2970 well (water) b le 211 3000 thing (object) b m 212 3001 thing (affair) y l 123 3100 clothing b s r 124 3110 pagne (wrap-around) ga ŋ 215 3120 buobuo /shirt kp r 216 3130 trousers k rlan 217 3150 shoe (s) k gt ɛ 218 3152 sandals n ɔ grɓɛ l kɛ r 219 3160 bracelet b ŋ 220 3170 necklace kɔ kɔ rlɛ gɛ 221 3180 ring (finger) n t ba 222 3200 rope m gr 223 3210 string /thread m gr 224 3220 mat (sleeping) sɛ ŋ 225 3250 hoop net z mky g 226 3260 net ky g 227 3310 chair kɔ g 228 3320 stool k gle 229 3350 drum (n) g gaa 230 3360 boat /canoe gb r 231 3410 calabash /gourd ŋma n 232 3420 basket p l /p ɛ 233 3430 load t r 234 3450 rubbish / garbage s gɛ 235 3460 hole (in ground) b g 236 3510 mortar (gringing) t ɔ r 237 3520 pestle t l 238 3550 medicine t 239 3555 poison lɔ g 240 3570 salt ny r 241 3580 fuofuo /ugali ny saab 242 3600 pot (cooking) gle 243 3610 water pot (earthern) y r 244 3620 cooking pot (metal) s ɛ 245 3700 iron (metal) k r 246 3710 knife s ɔ 247 3720 bush-knife sɔ kyɛ ra 248 3730 axe lɛ r 249 3740 hoe k r 250 3750 arrow p 251 3760 bow (weapon) t m 252 3770 spear k in University of Ghana http://ugspace.ug.edu.gh 169 253 3775 fish-spear ky mb l 254 3780 shield (n) ? 255 3799 war ɛ ɛ br 256 3810 law b gr 257 3910 charcoal saala 258 3920 fire v 259 3930 firewood daar 260 3940 smoke z r 261 3950 ash(es) t mpɛ l 262 3950 fire place ŋ 263 4010 night (time) t sɔ g 264 4011 darkness l g 265 4050 moon ŋm r 266 4051 month ky g 267 4060 star ŋm b r 268 4070 sun ŋm na /m tɔ g 269 4071 heat of day t l 270 4072 daylight v ɛ l 271 4080 day-time ny n 172 4100 day of 24 hours bib r 273 4120 morning b ɓ ra 274 4180 evening n ɔ ra 275 4210 sky sal ŋ 276 4220 cloud (s) s g mɛ / r 277 4230 fog /mist l 278 4250 wind sɛ sɛ b 279 4260 storm sɛ sɛ kpɛ ɛ 280 4300 water k ɔ 281 4310 rain saa 282 4315 lightening saany gr 283 4316 thunder saat n 284 4320 hail s b k b 285 4340 dew m ɛ l 286 4400 watercourse k ɔ pɔ l 287 4410 river b 288 4420 stream bagbɛ r 289 4450 lake tamb gkpɛ ɛ 290 4460 sea p /man 291 4510 mountain t ŋ 292 4511 hill k nk ne 293 4512 rock p r 294 4520 stone k s ɛ r 295 4522 pebble k sɛ b r University of Ghana http://ugspace.ug.edu.gh 170 296 4550 earth (soil) t nɛ 297 4551 ground t sɔ g 298 4560 sand b re 299 4570 dust r 300 4580 clay y gr 301 4585 mud mɛ mɛ rɛ 302 4600 year y n 303 4610 rainy season s ŋ 304 4620 dry season n 305 5000 how many ŋm n 306 5010 one b n 307 5020 two y 308 5030 three ta 309 5040 four n r 310 5050 five n 311 560 six y ɔb 312 5070 seven yɔ p 313 5080 eight n 314 5090 nine w 315 5100 ten pie 316 5120 twelve pie n y 317 5150 fifteen pie n n 318 5160 twenty l ɛ r 319 5170 thirty l ɛ r n pie 320 5180 hundred k ba 321 5190 thousand t r 322 5210 hot weather t l 323 5220 cold (weather) gr 324 5310 long (thing) w g 325 5311 tall w g 325 5312 deep l 326 5320 short (thing) ŋma a 327 5321 short (not tall) ŋma a 328 5330 big kpɛ ɛ 329 5332 thick (thing) t g 330 5332 fat bɛ r 331 5240 wide yɛ l 332 5350 small ky mpile 333 5352 thin (thing) m l 334 5354 thin (person) gb le/ baale 335 5360 narrow r 336 5380 round (adj.) k l k l 337 5410 heavy t rbr University of Ghana http://ugspace.ug.edu.gh 171 338 5412 difficult kp ŋ 339 5420 light in wieght mp g 340 5422 easy mɔ lɛ 341 5430 hard (surface) kp k r 342 5432 strong kp ŋ 343 5440 soft (surface) b l / l bl b 344 5442 weak b l 345 5450 smooth s l 346 5455 shiny ny grɛ 347 7470 clean /pure r t 348 5475 defiled s ŋ 349 5480 good v la 350 5481 well (adv.) kpɛ mɛ 351 5490 bad faa 352 5510 bitter t 353 5520 sour m r 354 5540 sweet n ɔ 355 5560 TRUE yelmiŋa 356 5565 truth yelmiŋa 357 5580 FALSE r 358 5585 untruth /lie r 359 5610 straight 360 5620 crooked gɔ ŋ 361 5640 right (correct) miŋa 362 5650 right (side) d r 363 5660 left side g b 364 5710 new p l 365 5720 young pɔ l 366 5750 old (worn) ny ŋ / k r 367 5810 all 368 5820 many / much y g 369 5830 crowd (of people) n y g 370 5850 few b laŋ 371 5860 some am nɛ 372 5870 other /different k ŋ 373 5910 red ɛ 374 5930 yellow d ɔ r 375 5940 green v r 376 5950 black sɛ bl 377 5952 dark- colored sɛ bl 378 5953 dark sɔ b 379 5960 blue s n 380 5980 white p l University of Ghana http://ugspace.ug.edu.gh 172 381 5982 bright-colored ny l ny l 382 6000 who? an 383 6010 I 384 6020 you (thou) f 385 6030 (s)he 386 6060 we t 397 6070 you (pl.) ny / ny m 388 6080 they bɛ /bɛ l 389 6100 what? b n 390 6110 this ŋ 391 6120 that l 392 6200 where? ny nɛ 393 6210 here k 394 6220 there b 395 6230 yonder b p r 396 6240 at b n 397 6250 (going) towards l r 398 6260 (coming) from y 399 6280 near p ɛ l 400 6290 far a a 401 6300 before b ŋ 402 6310 behind /after p r 403 6350 in p ɔ 404 6370 above s 405 6380 below p l 406 6400 how? ŋm ŋm n 407 6420 with n 408 6430 also m 409 6435 and an 410 6440 if mɛ 411 6450 when/ b r 412 6460 then / that time s ŋ 413 6470 today n 414 6480 yesterday a a 415 6490 tomorrow b g 416 6500 why? b s 417 6510 because b ns 418 6550 no , not k ŋ 419 6610 alive, to be v rɛ 420 6615 life nyɔ v r 421 6620 dirty /to become ɛ g 422 6630 become dry k 423 6640 full /filled p l University of Ghana http://ugspace.ug.edu.gh 173 424 6650 ripe, be m ɔ 425 6660 rotten p ɔ 426 6670 sharp rɛ 427 6679 blunt, dull gb 428 6690 become wet ɓ / b r 429 6760 sit (down) 430 6761 be seated t ŋ 431 6762 to remain /reside kp ɛ r 332 6770 stand up r 433 6771 to be standing r 434 6772 to stop bɛ r 435 6780 lie down g 436 6781 to be lying g 437 6840 sleep (v) g r 438 6850 dream n 439 6860 fear mb ɛ 440 6870 angar; be angry s r 441 6880 hunger; be hungry kɔ 442 6890 thirst; be thirsty k ɔ ny r 443 6900 shame; be ashamed v 444 6910 illness; be ill b l 445 6920 sorrow; be sad p sa ŋ 446 6930 joy ; be joyful p p ɛ l 447 7010 to bite 448 7012 gnaw ny 449 7020 eat 450 7025 food b n r 451 7030 drink ny 452 7032 to smoke (something) ɔ 453 7050 vomit t 454 7110 to cough; a cough kɔ r 455 7115 sneeze ky r 456 7150 suck mɔ gr 457 7152 suck (breast) ɛn 458 7170 spit ky r 459 7190 blow pɛ lb 460 7192 the wind blows 461 7195 breathe v r 462 7210 whistle l 463 7230 yawn (v) x ri 464 7310 sing y li 465 7315 song y l 466 7320 dance (v) sɛ b University of Ghana http://ugspace.ug.edu.gh 174 467 7325 dance (n) sɛ rb 468 7340 play (v) ɛ n 469 7350 laugh (v) l 470 7360 weep k 471 7370 to bark w gr 472 7372 cry out ky r 473 7375 make noise p gɔ mɛ 474 7400 say y l 475 7410 to talk / speak ɛ r 476 7420 answer s g 477 7430 tell y l 478 7440 proclaim mɔ ɔ l 479 7450 ask (question) sogr 480 7455 ask for ɛ l 481 7460 to command t n 482 7465 to rule k 483 7470 obey s g 484 7475 refuse agr 485 7480 to swear an oath p l 486 7490 to curse (someone) ŋmɛ aŋ 487 7495 insult t 488 7510 see nyɛ 489 7520 look at k 490 7530 show w l 491 7535 teach w l 492 7550 hear w 493 7560 listen to kyɛ l 494 7570 to smell (something) ny ri 495 7575 smell ; stink nyu u 496 7600 know (something /one b ŋ 497 7610 learn a n 498 7614 remember t ɛ r 499 7615 forget y r 500 7620 count (v) sɔ r 501 7622 read sɔ r 502 7625 book g n 503 7630 write ŋma a 504 7650 think t ɛ r 505 7680 to suffer /feel bad w t 506 7700 to love nɔ ŋ 507 7710 to want (something) bɔbr 508 7715 desire bɔbr 509 7720 to want (to do bɔbr University of Ghana http://ugspace.ug.edu.gh 175 something) 510 7730 need (V) bɔbr 511 7750 to seek bɔ 512 7760 find (v) nyɛ 513 7790 carry t 514 8000 take 515 8002 seize 516 8004 catch nyɔ g 517 8005 hold nyɔ g 518 8010 to lift / raise ɛ g 519 8020 give k 520 8022 gift kyɔ gt a 521 8030 pay y 522 8035 money L b r 523 8040 cost r 524 8050 to get /rceive 525 8060 steal 526 8070 to hide (something) sɔ gl 527 8080 lie (tell lies) r 528 8090 deceive bɛ l 529 8100 buy da 530 8105 sell k ɔ r 531 8110 to marry (a wife) kul 532 8130 bear child ɔ g 533 8131 bear twins ɔ g y br 534 8135 be born ɔ g 535 8210 to die kp 536 8220 kill k 537 8250 to live v rɛ 538 8310 to leave, depart bɛr 539 8320 to go (somewhere) ky n 540 8330 come (to) w 541 8335 come from y 542 8340 arrive t 543 8341 return lɛ b 544 8342 go out y 545 8334 enter kpɛ 546 8344 go up 547 8346 follow t 548 8347 bring w n 549 8350 send (someone) tɔ ɔ l / t 550 8410 fly (v) ag 551 8415 jump v r University of Ghana http://ugspace.ug.edu.gh 176 552 8430 swim 553 8431 dive m r 554 8432 float ɔ gl 555 8450 walk ky n 556 8460 run ɔ 557 8465 flow ky r 558 8470 fall l 559 8480 turn l ɛ b 560 8510 scratch a ŋ 561 8512 scratch oneself a ŋ 562 8530 rub ɔ ɔ r 563 8550 wipe m l 564 8650 pour p r / ky r 565 8660 wash pɛ g 566 8662 to bathe (oneself) s 567 8662 wash one's hands pɛ gn r 568 8665 wash (clothes) pɛ g b s ri 569 8667 to wash a pot pɛ g y r 570 8670 sweep p r 571 8675 broom saar 572 8710 open (v) y 573 8720 shut (v) p g 574 8800 break (tr) k 575 8810 split (wood) ɓ r 576 8812 tear (v tr) ky ɛ r 578 8814 divide p 579 8820 cut ŋma a 580 8830 saw ɔ g 581 8840 chop kp re 582 8850 stab (pierce) kyɔ r 583 8880 fight ɛ b 584 8900 hit, strike v 585 9010 beat ŋmɛ / pɔ b/ t ɔ 586 8930 to hurt someone ? 587 8940 help (v) s ŋ 588 8950 heal s n 589 8955 healer s nɛ 590 9010 throw (v) lɔ b 591 9015 throw away lɔ bɛ r 592 9020 push (v) 593 9030 pull (v) v r /t g 594 9050 press 595 9052 squeeze (v) ŋmɔ ɔ r University of Ghana http://ugspace.ug.edu.gh 177 596 9100 tie l 597 9105 hitch gb gl 598 9110 untie kyɛ r 599 9200 build mɛ 600 9210 make m l 601 9220 do 602 9230 work (v) t 603 9235 work t m 604 9240 to create m l 605 9250 forge (v) ? 606 9255 blacksmith sa a 607 9260 begin p l 608 9270 finish b r 609 9280 to sew sɛ 610 9290 weave w b 611 9295 weaver w br 612 9400 dress s 613 9410 undress yaa 614 9450 to braid, to plait p l 615 9500 hunt (v) b gl 616 9505 hunter n b glɛ 617 9510 shoot ŋmɛ 618 9520 gun m l 619 9550 cook (v) d g 620 9700 cultivate kɔ 621 9710 to plow kɔ 622 9720 plant (v) b r 623 9730 to dig (a hole) t g 624 9750 bury u u 625 9800 burn (something) ny g 626 9810 kindle tɔ ŋ 627 9820 to burn / be alight ny g 628 9830 extinguish kp r 629 9900 shiver m g 630 9950 swell p University of Ghana http://ugspace.ug.edu.gh 178 APPENDIX B: PRAYERS DAPAAR POƆ PUORU T s na be apaar p ɔ, a f yuor kpɛ ɛ , a f naaloŋ wa. ka t sagr f n ɔr teŋ u, a sɛg lɛ bɛ na sagra f n ɔr dapaari p ɔ. K t bund r zina a sɛg t , d a t yelbebe suur k t , a sɛg lɛ t na m d rɛ suur kurɛ taar a. ka t taa wa sagr bɛlɛ ɛ, ir t yel faar zaa p ɔ. Amina. H LOR ’S PRAYER Our Father, Who art in Heaven, Hallow be thy name, thy kingdom come, thy will be down on earth as it is in Heaven. Give us this day our daily bread, and forgive us our trespasses as we forgive those who trespass against us. And lead us not into temptation, but deliver us from evil. Amen. Y M RI Yaan Maria, f paali n garas ɛ, Naagmin be f zie, Naagmin maal f na gan pɔgbɛ za, ɛ a f p ɔ bie yezu nyɛrɛ puoru. aria soŋ, puori Naagmin k t , t na yelbebe em, p p n an t k aar, Amen. HAIL MARY Hail Mary, full of Grace, the Lord is with you, Blessed are you amongst women. Blessed is the fruit of thy womb Jesus. Holy Mary, Mother of God, pray for us sinners now and at sthe hour of our death. Amen. University of Ghana http://ugspace.ug.edu.gh 179 REFERENCES Agoswin, A. Musah (2010). Aspects of Kusaal phonology. M.Phil. Thesis, University of Ghana. Akanlig-Pare, George (2005). Buli Tonology: A Non-Linear Approach. Ph.D. Thesis, University of Ghana. Akanlig-Pare, George (1994). Aspects of Buli Phonolgy. M.Phil Thesis, University of Ghana. Akrofi, Ansah Mercy. (2002). Aspects of the Phonology of Latɛ. M.Phil. Thesis, University of Ghana. Anderson, John (1969). Syllabic or Non-Syllabic Phonology. JL5: 136-143 Antilla A. and A. Bodomo (1996). Stress and Tone in Dagaare. Ms. Standford University of TromsØ. To appear in the proceedings of NELS 31. Asante, Roger K. (2009). Topics in Nkonya Phonology. M.Phil. Thesis. University of Ghana. akovi , Eric (2007). Local Assimilation and Constraint Interaction. In Paul de Lacy (ed.) The Cambridge Handbook in Phonology. Cambridge:C.U.P. Bao, Zhiming (1999). The Structure of Tone. New York: Oxford University Press. Bemile, Sebastian K. (1983). g r honeme Contrast. Vol. 1. Africana Saraveinsia Linguistica. Nr.1. Saarbrücken: Universität des Saarlandes. Bemile, Sebastian K. (1984). g r honeme Contrast. Vol. 2. Africana Saraveinsia Linguistica. Nr.1. Saarbrücken: Universität des Saarlandes. Bemile, Sebastian K. (1984). k r): A language Course. Accra. Unpublished. Bemile, Sebastian K. (1985). Grundzüge der Phonologie des g r . Africana Saraveinsia Linguistica. Nr. 9. Saarbrücken: Universität des Saarlandes. University of Ghana http://ugspace.ug.edu.gh 180 Bemile, Sebastian K. (1990). g r rthography. apers in g r tudies 1 (2) pp. 1-31 Bemile, Sebastian K. (2010). g r rover s. Berlin: Dietrich Reimer Verlag. Bendor-Samuel, J. T. (1971). Niger Congo, Gur. In Sebeok, T. (ed.), Current Trends in Linguistics. The Hague: Mouton Blevins, Juliette (1995). The syllable in Phonological Theory. In Goldsmith John A. (ed), The handbook of Phonological Theory. Cambridge: Blackwell Publishers. Bodomo, Adams (1997). The Structure of Dagaare. Stanford Monographs in African Languages. California: CSLI Publications. Bodomo, Adams. (2000). Dagaare. Languages of the World/Materials 165. Muenchen: Lincom Europa. Boone, Douglas (1989). Towards an African Area Standard Word List. Paper presented at the SIL International Language Assessement Conference (ILAC), Horsley Green , England. Bota, Grace (2002). Some Aspects of Bono Phonology: An Autosegmental analysis. M.Phil. Thesis. University of Ghana. Brosnahan, L. F. and B.Malmberg (1970). Introduction to Phonetics. Cambridge: W. Deffer and Sons Ltd. e no, N e a ael A. (1987). Intervocalic /d/ Rhotacism in Domincan Spanish: A Non-Linear Analysis. Hispania Vol.70. No. 2. Pp 363-368. American Association of Teaching of Spanish and Portuguese. Chomsky, N. and Halle, M (1968). Sound Pattern of English. New York: Harper & Row. Clements, G. N. (1984). Principles of tone association in Kikuyu. In: G.N. Clements and J. Goldsmith (eds.). Autosegmental Studies in Bantu Tone, 281-339. Dordreccht: Foris Clements, G.N. (1979). The Description of Terraced-Level Tone Languages. Language, Vol. 55, No. 3, 536-558. Linguistic Society of America. http://www.jstor.org/stable/413317. 18/04/2013 12:25 University of Ghana http://ugspace.ug.edu.gh 181 Clements, G. N. and K. C. Ford (1979). Kikuyu Tone Shift and its Synchronic Consequences. Linguistic Inquiry 10 (2). pp.179-210. Clements, G.N. and Keyser, S. J. (1983). CV Phonology. Cambridge: MIT Pres. Connell, Bruce (2001). Downdrift, Downstep, and Declination.Typology of African Prosodic Systems Workshop. Bielefeld University, Germany. < Coral.spectrum.uni.bielefed.de/TAP/Connelle.pdf> 18/04/20013. Connell, Bruce & D.R. Ladd (1990). Aspects of Pitch realization in Yoruba. Phonology,Vol.7,1, 1-29 Cruttenden, A. (1986). Intonation. C.U.P. Crystal, David (1992). A Dictionary of Linguistics and Phonetics, 3rd Edition. UK: Blackwell Publishers. Dakubu, M. E. Kropp (1997). Oti-Volta Vowel Harmony and Dagbani. Gur Papers 2. 81-88. Dakubu, M. E. Kropp (2005). Collected Language Notes on Dagaare Grammar. Collected Language Notes No. 26. Institute of African Studies, Legon. Davies, Morpurgo Anna (1982). Dental Rhotacism and Verbal Ending in the Luwian Languages. . ....96.Bd.,2. H. Vandenhoeck &Ruprecht (GmbH & Co.KG) Delplanque, Alain (1976). Phonologie Transformationelle du Dagara. Th se de me Cycle. Institute des Langues et Civilization, University des Paris III Draga, Zec (2007). The Syllable. In Paul de Lacy (ed.) The Cambridge Handbook of Phonology. Cambridge: CUP Durand, J. (1990). Generative and Non-linear Phonology. New York : Longman. Fox, Anthony (2000). Prosodic Features and Prosodic Structure: The Phonology of Suprasegmentals. New York: Oxford University Press. Frimpong, Victoria (2009). Aspects of Esahie Phonology. M.Phil. Thesis University of Ghana. University of Ghana http://ugspace.ug.edu.gh 182 Fudge (1969). Syllables. Journal of Linguistics. CUP. Gandour, Jackson (1974). On the Representation of Tone in Siamese. In: Maddieson, Ian (ed.), The tone tome Studies. Studies on tone from the UCLA tone project UCLA Working Papers in Phonectics 27, pp. 18-46. Goldsmith, J. A. (1976b). Autosegmental Phonology. Ph.D. Dissertation, MIT Goldsmith, J. A. (1979). The Aims of Autosegmental Phonology. In: D. Dinnsen (ed.) Current Approaches to Phonological Theory. (pp. 202-222) Bloomington: Indiana University Press Goldsmith, J. A. (1990). Autosegmental & Metrical Phonology. Basil Blackwell. Goldsmith, J. A. and Noske, Manuela (Forthcoming). Autosegmental Phonology and Underspecifiation Theory. In: Auroux, S., Koerrner, K. Niederehe, H.J. and Versteegh, K. (eds.) History of the Language Science. New Kork: Walter de Gruyter. <> August 24, 2012 Gummere, John F. (1960). Classroom applications of Rhotacism and Vowel Weakening. The Classical World Vol. 53 No. 8 http://www.jstor.org/stable/4344365 >February 18, 2013. Halle, Morris and Jean-Roger Vergnaud (1978). Metrical Structure in Phonology. Ms, MIT. Hudu, Fusheini (2013). Dagbani tongue-root harmony: triggers, targets,and blockers. Journal of African Languages and Linguistics. Vol. 34. 1. 81-88 Hayes, Bruce (1989). Compensatory Lengthening in Moraic Phonology. Linguistic Inquiry, Vol.20. No.2, pp 253-306, The MIT Press. August 17, 2012. Hymann, Larry M. (1975). Phonology. Theory and Analysis. New York: Holt, Rinehart and Winston. Hyman, Larry (1985). A Theory of Phonological Weight. Dordrecht: Foris Hyman, Larry M. & Russell G. Schuh (1974). Universals of Tone Rules: Evidence from West Africa. Linguistic Inquiry, Vol. 5, No. 1, pp. 81. MIT Press http://www.jstor.org/stable/4177809. 18/04/2013 11:43 University of Ghana http://ugspace.ug.edu.gh 183 Kahn, Deniel (1976). Syllable-based generalization in English. Doctoral Dissertation, MIT. New York: Garland Press. Katamba, Francis. (1989). An introduction to Phonology. New York: Longma Kennedy, Jack (1966). Collected Field Reports on The Phonology of Dagaari. Collected Language Notes No. 6. Accra: Institute of African Studies, University of Ghana. Kenstowicz, Michael (1994). Phonology in Generative Grammar. U S A: Blackwell Publishers. Kenstowicz, Michael and Kisseberth, Charles (1979). Generative Phonology: Description and Theory. New York: Academic Press Inc. Ladefoged, Peter (1992). A Course in Phonetics. Harcourt Bruce Javanovitch, New York. Ladefoged, Peter (1993). A course in Phonetics (3rd Edition). Orlando: Harcourt Brace and Company. Ladefoged, Peter (2003). Phonetic Data Analysis: An Introduction to Fieldwork and Instrumental Techniques. Oxford: Backwell Laver, J. (1994). Principles of Phonetics. Cambridge: CUP Leben, W. R. (1973). Suprasegmental Phonology. Ph.D. Dissertation MIT. Lindau, M. (1986). Testing a model of intonation in a tone language. Journal of the Acoustical Society of America, Vol 80, 757–764. McCawley, J.D. (1970). Some Tonal Systems That Come Close to Being Pitch A y m B ’t Quite Make It. In: M. Cambell et al (eds.). pp. 526-532. Nanbigne, Edward (2008). Culture, Context and Conflict: A study of the Changing Discourse on Care among the Dagaaba of Northern Ghana. Ph.D. Thesis. University of Ghana. University of Ghana http://ugspace.ug.edu.gh 184 Nathan Geoffery S. and Patricia J. Donegan. Natural Phonologgy and Sound Change.> August 18, 2012 Odden, David (1988). Anti antigermination and the OCP. Linguistic Inquiry 19, pp. 451- 475. Odden, David (1995). Tone : African Languages. In: J. A. Goldsmith (ed.). The Handbook of Phonological Theory. Cambridge: Blackwell Publishers Inc. Odden, David (2005). Introducing Phonology. Cambridge: CUP. << http:// journals.Cambridge.org/article_SO952675707001315>> August 18, 2012. Oduro, Kwarteng Thomas. (2009). Aspects of Anufo Phonology. MPhil. Thesis. University of Ghana. Oyebade, F.O. (1998). Course in Phonology. Lagos: Shebiotimo Publication Pike, Kenneth and E. Pike (1947). Immediate Constituents of Mazatec Syllable. Pike K. L. (1948). Tone Languages. Ann Arbor: The University of Michagan Press. Roach, Peter (1991). English Phonectics and Phonology. A Practical Course. Second edition. Cambrdge: CUP Saanchi, A. J. (1997). The Vowel System of Dagaare. Gur Papers/ Cahier Voltaїques 2, 129-135 Saanchi, A. J. (2006). A Grammar of Spatial Constructions in Dagaare. Ph.D. Thesis. University of Ghana. Sagey, Elizabeth Carol (1986). The representation of Features and Relations in Non-Linear Phonology. Ph.D. Thesis MIT. Schachter, P and Fromkin, A.Victoria. (1968). Phonology of Akan: Akwapim, Asante, and Fante. Working Paper in Phonetics 9. Los Ageles : University of California. om , D. Joachim (2004). Dagara orthography. Journal of Dagaare Studies Vol. 4. Stewart, J. M. (1965). The Typology of the two Tones Systems. Bulletin of Institute of African Studies, University of Ghana. Trubetzkoy, N.S. (1967). Principes de Phonolgie (First Published in 1939). Paris: Klincksieck. University of Ghana http://ugspace.ug.edu.gh 185 Uffmann, Christian (2001). Patterns of Vowel Epenthesis (not only) in Shona Loanwords. In: Caroline Féry, Antony Dubach Green & Ruben van de Vijver (eds.). Proceedings of HILP5, x–x, University of Potsdam. Westermann, Diedrich & Ward, Ida C. (1966). Practical Phonetics for Students of West African Languages. London: Oxford University Press. Williams, Edwin (1971). Underlying Tone in Margi and Igbo. In: Linguistic Inquiry 7(3). 463-484 (1976.) Wolfgang, U. Dressler (1984). Explaining Natural Phonology. Phonology, 1, 29-51. doi:10.1017/S0952675700000282. August 20, 2012. Yip, Moira (2002). Tone. Cambridge: C.U.P. Yip, Moira (2007). Tone. In: Paulde Lacy (ed.). The Cambridge Hand book of Phonology. Cambridge: CUP. University of Ghana http://ugspace.ug.edu.gh