Mwinlaaru - PHD Thesis - Final

A SYSTEMIC FUNCTIONAL DESCRIPTION OF THE GRAMMAR OF
DAGAARE
ISAAC NUOKYAA-IRE MWINLAARU
Ph.D
The Hong Kong Polytechnic University
2017
The Hong Kong Polytechnic University
Department of English
A Systemic Functional Description of the Grammar of Dagaare
Isaac Nuokyaa-Ire Mwinlaaru
A thesis submitted in partial fulfilment of the requirements for the degree of Doctor of
Philosophy
September 2016
b
CERTIFICATE OF ORIGINALITY
I hereby declare that this thesis is my own work and that, to the best of my
knowledge and belief, it reproduces no material previously published or
written, nor material that has been accepted for the award of any other degree
or diploma, except where due acknowledgement has been made in the text.
_______________________________________________(Signed)
_______________________________________________ (Name of student)
ii
DEDICATION
In memory of the Late Rev. Fr. Linus Zan Mwinlaaro
And to my ancestors: Mwinlaaru and Kognepuor

sons of Saabyang
son of
Vob
iii
ABSTRACT
Since the 1960‘s, the systematic description of languages has gained

unprecedented attention among linguists. While scholars have documented
robust evidence for universal tendencies in the organisation, functions, and
evolutionary pathways of language, there is still need to account for many
languages that have not yet been described and develop comprehensive
accounts for languages that are partially described. Such comprehensive
descriptions are not only important for advancing linguistic science, but are
also crucial for various applications in the communities where the languages
are spoken. Against this background, the present study provides a system-
based account of the lexicogrammar of Dagaare (specifically, the Lobr
dialect), a Gur language spoken in West Africa. The general objective is to
examine lexicogrammatical systems that realise interpersonal, textual and
ideational meanings at clause rank. The account is based on discourse data,
largely consisting of spoken texts produced by Lobr speakers in Ghana and
Burkina Faso. The analysis and descriptive interpretation of the language are
guided by the general theory of language formulated in systemic functional
linguistics, typological generalisations and transfer comparison.
The findings of the study are organised into four chapters. The first
chapter outlines the basic phonological and grammatical organisation of
Dagaare. It identifies the units of phonology (viz. tone group, phonological
word, syllable and phoneme) and grammar (viz. clause, group, word, and
morpheme) and describes their forms and functions. It also discusses the
various classes of grammatical units.
The second chapter discusses the interpersonal grammar of Dagaare,
focusing on MOOD, the grammar of speech acts, and its interaction with
POLARITY and systems of modal assessment, namely MODALITY (i.e.
desirability & probability) and NEGOTIATION, realised by particles placed
prosodically at clause initial or final position to enact speakers‘ attitude
towards propositions and proposals. The chapter first examines the
interpersonal structure of the clause, identifying two immediate constituents,
the Mood base and the Residue. Three structural functions, comprising
iv
Subject, Predicator and Negotiator (realised by clause initial and clause final
particles), have been identified as forming the Mood base and as the most
salient elements in enacting the clause as a unit of exchange. The chapter then
moves on to discuss different mood types, comprising declarative,
interrogative and imperative, and their sub-types. Next, the chapter examines
the interaction between Subject person and the imperative clause, mood in
elliptical and minor clauses and then the phenomenon of mood metaphor or
indirectness between the mood types and the speech functions they realise.
These are then followed by a discussion of POLARITY, MODALITY and
NEGOTIATION.
The third analysis chapter describes the textual systems of THEME and
INFORMATION. Theme is defined as the element that is given initial

prominence in the clause and is developed by the remainder of the clause, the
Rheme. Different realisations have been identified for three types of Theme:
textual, interpersonal and topical Themes. Notably, topical Themes are often
(but not always) signalled by marking them off the rest of the clause as the
Absolute Theme of the clause (‗left-dislocated constructions‘). On the other
hand, the focus structure of Dagaare is minimally realised by the New
element. New is identified semantically as that element that is singled out in
the information unit as newsworthy. Three focus types are identified, namely
end focus, contrastive focus and broad focus. End focus is the default choice
for positive declarative and is realised by the focus particle nι. Contrastive
focus is realised by thematic equatives, emphatic pronouns and exclusive
markers. Broad focus is identified as zero-realisation and it is the default
choice for imperative and negative clauses. The chapter finally discusses the
relationship between the clause and information unit in Dagaare.
The final chapter examines the system of TRANSITIVITY, comprising
the sub-systems of PROCESS TYPE and AGENCY. Six process types have been
identified, comprising three principal types – material, mental and relational
clauses; and three minor types – behavioural, verbal and existential clauses.
Detailed grammatical evidence is given for the identification of each process
type and their subtypes. Grammatical characteristics are also provided for the
different participant roles across the process types.
v
The chapter continues to identify principles that generalise across the
different process types. A notable generalisation is based on one phenomenon
in Dagaare, the tendency for speakers to leave the Complement (or ‗Object‘)
unrealised in the clause. The general factor motivating this omission is a
textual kind, namely, when the potential Complement is regarded as given
information it can be left unrealised. However, other factors such as
humanness in relation to the noun group realising the Complement and
abstractness in relation to the clause as a whole override this principle.
Further, based on this single variable, the various process types cluster into
two broad semantic types: concrete and abstract clauses. Regarding the system
of AGENCY, clauses divide into middle or effective, depending on whether or
not they embody the feature of agency. Middle clauses represent the process
as being self-engendered while effective clauses represent it as being caused
by an external participant, the Agent.
The study contributes to systemic functional theory and the general
meta-theory of language, functional language typology and African linguistics.
The description is also appliable for the purposes of language education,
translation, orthography, discourse studies and other practical settings where
the Dagaare language is in focus.
vi
RELATED OUTPUT
Book Chapter (refereed)

 Mwinlaaru, Isaac N. in press. Grammaticalising attitude: Clause
juncture particles and NEGOTIATION in Dagaare. In Lise Fontaine &
Akila S. Baklouti (eds.) Perspectives from systemic functional
linguistics. London: Routledge.
 Mwinlaaru, Isaac N., Christian M. I. M. Matthiessen & Ernest

Akerejola. forthc. A system based typology of MOOD in Niger-Congo
languages. In Augustine Agwuele & Adams Bodomo (eds.). The
handbook of African linguistics. London: Routledge.
Journal Articles (refereed)
 Mwinlaaru, Isaac N. & Foong Ha Yap. 2017. A tale of two

demonstratives in Dagaare: Reflections on directionality principles in
grammaticalisation. Language Sciences, 64. 130-151.
DOI: 10.1016/j.langsci.2017.07.005. [SSCI & A&HCI]
 Mwinlaaru, Isaac N. & Winfred Wenhui Xuan. 2016. A survey of
studies in systemic functional language description and typology.
Functional Linguistics, 3(8). 1-41.
Conference presentations (refereed)
 Matthiessen, Christian M.I.M & Isaac N. Mwinlaaru. 2017. Sayerhood

across languages. 27th European Systemic Functional Linguistics
Conference [ESFLC], Universidad de Salamanca, Spain, June 29 –
July 1.
 Matthiessen, Christian M. I. M., Abhishek Kumar Kashyap,

Mwinlaaru, Isaac N. & Espindola, Elaine. 2015. Systemic and
metafunctional variation in the construal of motion across languages.
Functional Language Typology Conference, Shanghai Jia Tong
University, China, December 12 – 16.
vii
 Mwinlaaru, Isaac N. 2014. Profiling interpersonal particles in Dagaare:
The case of the Dagara-Lobr dialect. 25th European Systemic
Functional Linguistics Conference, Universite Paris Diderot – Paris 7,
UFR EILA, Paris, July 10 – 12.
 Mwinlaaru, Isaac N. 2014. Towards an interpersonal grammar of

Dagaare: Notes on mood and polarity. A paper presented at the Faculty
of Arts Lecture Series, University of Cape Coast, Ghana. August 21.
 Mwinlaaru, Isaac N. 2015. Projection as a fractal motif in Dagaare. In

Christian M.I.M. Matthiessen & Kazuhiro Teruya (Orgs.), ―Projection
as a fractal motif‖. A colloquium in 1st Tunisian Systemic Functional
Linguistics Conference & Workshop, University of Sfax, Hammamet,
Tunisia, March 26 – 28.
 Mwinlaaru, Isaac N. 2015. The system of NEGOTIATION in the

lexicogrammar of Dagaare. 1st Tunisian Systemic Functional
Linguistics Conference & Workshop, University of Sfax, Hammamet,
Tunisia, March 26 – 28.
 Mwinlaaru, Isaac N. 2015. A systemic functional description of

information structure in Dagaare. 14th National Chinese Conference on
Functional Linguistics and International Symposium on Systemic
Functional Linguistics for Launching „The Halliday-Hasan Linguistics
Fund‟ & Celebrating Professor Halliday‟s 90th Birthday, Beijing
Normal University, April 23 – 26.
 Mwinlaaru, Isaac N. 2015. Grammaticalisation in Dagaare: From

demonstratives, copula verbs, relational pronouns to the focus particle.
42nd International Systemic Functional Congress, RWTH Aachen
University, Germany, July 27 – 31.
 Mwinlaaru, Isaac N. 2015. Polarity and modality in the lexicogrammar

of Dagaare. Functional Language Typology Conference, Shanghai Jia
Tong University, China, December 12 – 16.
 Mwinlaaru, Isaac N. 2016. The structure and functions of the Dagaare

verbal group. In Christian M. I. M. Matthiessen & Mira Kim (Orgs.), A
viii
typology of verbal units in the overall system of languages. A
colloquium presented at the 43rd International Systemic Functional
Linguistics Congress [ISFLC], Universitas Pendidikan Indonesia
(UPI), Indonesia, July 19 – 23.
 Mwinlaaru, Isaac N. 2017. A prosodic-systemic account of ASPECT in

Dagaare. 27th European Systemic Functional Linguistics Conference
[ESFLC], Universidad de Salamanca, Spain, June 29 – July 1.
 Mwinlaaru, Isaac N. 2017. POLARITY and MODALITY in the grammar

and discourse of Dagaare. West African Linguistics Conference
[WALC], University of Education, Winneba, Ghana, July 31 – August
5.
 Mwinlaaru, Isaac N. 2017. Demonstratives and the division of semiotic

labour: A study of the Lobr dialect of Dagaare. West African
Linguistics Conference [WALC], [Workshop on Mabia Languages],
University of Education, Winneba, Ghana, July 31 – August 5.
ix
ACKNOWLEDGEMENTS
This thesis would not have been possible without the enormous assistance of
many people, whom I would like to thank for their great help. First and
foremost is my Chief Supervisor, Professor Christian M.I.M. Matthiessen, to
whom I am forever indebted for his extraordinary guidance and mentorship.
When I was contacting him to be my PhD supervisor, I told myself that if he
declined, I was going to change my thesis topic. I thank him for
enthusiastically supervising this thesis and for his encouragement, patience
and close collaboration throughout the study. He meticulously read my drafts
and gave me very valuable comments. Professor Matthiessen went beyond the
confines of PhD supervision to offer me true comradeship and familial support
in ways that shaped my overall academic and social outlook. Without him, this
study would not have been possible. I am very humbled by his guidance. I also
thank my Co-supervisors, Professor Foong Ha Yap and Dr. Marvin Kit Man
Lam, for the many insights they gave me. Professor Yap introduced me to
functional language typology and grammaticalisation and I extended the
frontiers of my knowledge of language by benefitting from her experience in a
wide range of languages. She has been a close research collaborator
throughout my studies.
I also thank members of the Board of Examiners (BOE) for this thesis,
including the BOE Chair, Dr. Gail Forey, and my two external examiners,
Professor Bernd Heine of the University of Cologne (Germany) and Professor
Felix Banda of Western Cape University (South Africa). They were generous
to take time out of their busy schedules to read my thesis. I am very grateful to
them for the insightful comments, their constructive criticism and the
inspiration they gave me both during and after my PhD examination. I am
lucky to have worked with such a distinguished scholarly team. Any errors in
the thesis are certainly mine.
Many thanks also to Professors Michael Halliday and Ruqaiya Hasan†
for discussions on this study at its conceptualisation stage. Professor Halliday
has inspired me enormously through his writings, without which this study
x
would not have been possible in its present form. My talks with him and
Ruqaiya over dinner contributed to the methodology of this study.
I am equally grateful to Professor Stephen Evans†, Professor Martin
Warren, Professor David Qian, Dr. Xu Xun-feng and Dr. Li Lan, all of the
Department of English of The Hong Kong Polytechnic University. They
contributed to my research experience through their teaching and/or
administrative involvement in my PhD studies. Dr. Xu Xun-feng was the
examiner for the confirmation of my PhD registration.
I would also like to thank fellow members of the Poly-systemic
Research Group, notably Dr. Kazuhiro Teruya, The Hong Kong Polytechnic
University; Professor Jorge Hita Arús, Universidad Complutense de Madrid,
Spain; Dr. Mohammed Ali Bardi, Al-Maarefa Colleges for Science and
Technology, Saudi Arabia; Dr. Patamma Patpong, Madihol University,
Thailand; Professor Abhisheik Kumar Kashyap, Sun Yat-sen University,
China; Dr. Wu Changzhou, Macquarie University, Australia; and Dr Elaine
Espindola, Pontificia Universidad Católica de Valparaíso , Chile. I am indebted
to them for fruitful discussions on many aspects of my analysis of Dagaare at
conferences, symposia and workshops. I should also thank Professor Jim
Martin, University of Sydney, Australia, for his continuous interest in my
work and for his encouragement.
Many thanks to Professor Adams Bodomo, University of Vienna,
Austria, and members of our Dagaaba Facebook community, Dagaare Yong,
for inspiration. Many of them served as my language consultants, notably
Modeste Soméy, Cosmas Taaba, Caroline Dakurah and Henry Dakurah. I
should also thank the staff of Von FM and Radio FREED of Nandom for
recorded data, especially Benjamin and Adaar Linus and my nephew, Stephen
Mwinlaaro, who helped me throughout the data collection process and also
served as one of my language consultants.
I am grateful also to the Hong Kong Research Grant Council for
generously funding my PhD studies through the Hong Kong PhD Fellowship
Scheme (Award no. PF12-10332), and the University of Cape Coast, Ghana,
for granting me study leave to undertake the study. I thank my mentors and
superiors, Professor Joseph B. A. Afful, Professor Dora F. Edu-Buandoh and
Professor Domwin Dabire Kuupole, for their generous assistance in providing
xi
me the study leave even at a time when there was a lot of pressure on human
resource in the university. Many thanks also to Dr. Moussa Traore and Mr.
Samuel Kwesi Nkansah for supporting my study leave application.
I am highly indebted too to my many friends and colleagues who
engaged me both academically and socially and made my PhD research
experience less painful: Ridvan Aydin, Wang Bo, Ma Yuanyi (Helen), He
Quiping (Ares), Lin Ling (Kathy), Yu Yating (Tiffany), Sonya Chik, Esterina
Nervino, Winfred Wenhui Xuan, Zhang Peijia (Kaela), Nancy Guo, Elva Lin,
Mark Nartey, Raymond Adongo, Elizabeth Agyeiwaa, Ibrahim Mohammed,
Aaron Kofi B. Yankholmes, Wong Ming Chiu (Steven), Locky Law, Carmen
Ng, Kathleen MacDonald, Andy Fung, Eric Lok Ming Cheung, Daniel
Recktenwald, Zheng Yaofei (Jennifer). Many thanks to these most valuable
fellow travellers and those I could not mention for want of space. Their doors
were always open to me and they denied me the loneliness and misery of PhD
life!
I will also acknowledge a few contemporary linguists whose work
informed my thinking about language. These include Michael Halliday,
Christian Matthiessen, Bernd Heine, Paul Hopper, Robert M. W. Dixon, Joan
Bybee, Martin Haspelmath and Tom Güldemann. Many thanks also to the
many writers I have consulted, especially authors of previous systemic
grammars and the many collections on African languages. Thanks again to
Professor Adams Bodomo and also Professor Mary E. K. Dakubu†, Joachim
Somé and other scholars working on the different varieties of Dagaare/Dagara.
Their studies served as a foundation for my own description and made my
work less tedious.
Thanks also to my family for their endurance, patience and prayers
throughout my study. May God richly repay them for their sacrifice.
Glory be to the Almighty God, the Alpha and the Omega, the Great
Ingenious Spider, Dweller of the Crossroads, Most Dependable and Most
Merciful, from whom all good things come!
xii
TABLE OF CONTENTS
..................................................................................................................................Page
CERTIFICATE OF ORIGINALITY ............................................................................... ii
DEDICATION ............................................................................................................... iii
ABSTRACT .................................................................................................................. iv
RELATED OUTPUT .................................................................................................... vii
ACKNOWLEDGEMENTS ............................................................................................ x
LIST OF FIGURES .................................................................................................... xvii
LIST OF TABLES ........................................................................................................ xx
ABBREVIATIONS AND SYMBOLS ........................................................................ xxii
CHAPTER ONE ............................................................................................................. 1
INTRODUCTION .......................................................................................................... 1
1.1 Introduction........................................................................................................ 1
1.2 Geolinguistic Landscape of Dagaare .................................................................. 1
1.2.1 Typological Context of Dagaare .................................................................. 3
1.2.2 Varieties of Dagaare: A Dialect Continuum ................................................. 5
1.2.3 Dagaare in Language Contact ...................................................................... 6
1.3 The Dagara Society ............................................................................................ 7
1.4 Developments in Dagaare Studies .................................................................... 10
1.5 The Present Study ............................................................................................ 13
1.5.1 Problem Statement ..................................................................................... 13
1.5.2 Aim of the Study........................................................................................ 15
1.5.3 Significance of the Study ........................................................................... 15
1.6 Outline of the Thesis ........................................................................................ 17
CHAPTER TWO .......................................................................................................... 18
THEORY AND APPROACH ....................................................................................... 18

2.1 Introduction...................................................................................................... 18
2.2 Linguistic Theory and Language Description ................................................... 18
2.3 Language within a Typology of Systems .......................................................... 22
2.4 Systemic Organisation of Language ................................................................. 25
xiii
2.4.1 Instantiation ............................................................................................... 25
2.4.2 Stratification .............................................................................................. 27
2.4.3 Metafunction ............................................................................................. 31
2.4.4 Rank .......................................................................................................... 33
2.4.5 Axis ........................................................................................................... 35
2.4.6 Semogenesis .............................................................................................. 40
2.5 Methodological Procedures in the Present Study .............................................. 42
2.5.1 Data Source ............................................................................................... 43
2.5.2 Analystical Procedure ................................................................................ 43
2.5.3 Limitations of the Study ............................................................................. 48
2.6 Conclusion ....................................................................................................... 49
CHAPTER THREE....................................................................................................... 50
AN OVERVIEW OF THE DAGAARE LINGUISTIC SYSTEM.................................. 50

3.1 Introduction...................................................................................................... 50
3.2 Phonology ........................................................................................................ 50
3.2.1 Phoneme .................................................................................................... 51
3.2.2 Syllable...................................................................................................... 55
3.2.3 Phonological Word .................................................................................... 57
3.2.4 Tone Group ............................................................................................... 60
3.3 Orthography ..................................................................................................... 62
3.4 Lexicogrammar ................................................................................................ 64
3.4.1 Morphemes ................................................................................................ 65
3.4.2 Word Classes ............................................................................................. 68
3.4.2.1 Nouns .................................................................................................... 68
3.4.2.2 Verbs ..................................................................................................... 74
3.4.2.3 Adverbs ................................................................................................. 82
3.4.2.4 Pronouns ................................................................................................ 92
3.4.2.4 Particles ................................................................................................. 94
3.4.3 The Dagaare Clause ................................................................................... 98
3.5 Conclusion ..................................................................................................... 106
CHAPTER FOUR ....................................................................................................... 108
MOOD AND MODAL ASSESSMENT ...................................................................... 108
xiv
4.1 Introduction.................................................................................................... 108
4.2 Choices in Dialogue and the System of SPEECH FUNCTION .............................. 108
4.3 Interpersonal Structure of the Clause .............................................................. 113
4.3.1 The Mood Base........................................................................................ 115
4.3.2 Meaning of Elements in the Mood Base ................................................... 123
4.3.2.1 Negotiator ............................................................................................ 123
4.3.2.2 Predicator............................................................................................. 125
4.3.2.3 Subject ................................................................................................. 127
4.3.3 Function of the Mood Base ...................................................................... 131
4.3.4 Other Elements in the Interpersonal Structure of the Clause ..................... 132
4.3.4.1 Complement ........................................................................................ 132
4.3.4.2 Adjunct ................................................................................................ 134
4.3.4.3 Vocative .............................................................................................. 136
4.4 MOOD as a System .......................................................................................... 138
4.4.1 Indicative ................................................................................................. 139
4.4.1.1 Declarative Clauses .............................................................................. 139
4.4.1.2 Interrogative Clauses............................................................................ 148
4.4.1.2.1 Polar interrogative clauses ............................................................... 148
4.4.1.2.2 Elemental interrogative ................................................................... 153
4.4.1.2.3 Placement of the Q-element in the clause ........................................ 160
4.4.2 Imperative ............................................................................................... 161
4.4.3 The Imperative Clause and MOOD PERSON .................................................... 164
4.4.4 A Note on MOOD and Grammatical Metaphor .......................................... 165
4.4.5 Mood and Elliptical and Minor Clauses ................................................... 166
4.5 Polarity and Modal Assessment ...................................................................... 170
4.5.1 POLARITY ................................................................................................. 170
4.5.2 MODALITY ................................................................................................ 172
4.5.2.1. Probability ........................................................................................... 173

4.5.2.2 Desirability .......................................................................................... 177
4.5.3 NEGOTIATION ........................................................................................... 179
4.5.3.1 Negotiation in the indicative mood: epistemic and affective stance ........... 181
4.5.3.2 Negotiation in the imperative mood: modulating proposals .................. 194
4.5.3.3 Negotiation concord ............................................................................. 200
4.6 Conclusion ..................................................................................................... 203
xv
CHAPTER FIVE ........................................................................................................ 204
THEME AND INFORMATION ................................................................................. 204

5.1 Introduction.................................................................................................... 204
5.2 The Concept of the Clause as Message ........................................................... 204
5.3 Theme and Rheme ......................................................................................... 205
5.3.1 Topical Theme ......................................................................................... 208
5.3.1.1 Absolute Theme ................................................................................... 209
5.3.1.2 Non-absolute Theme ............................................................................ 214
5.3.2 Textual and Interpersonal Themes............................................................ 217
5.4 Focus of information ...................................................................................... 222
5.4.1 End Focus and the Focus Particle ............................................................. 228
5.4.2 Contrastive Focus .................................................................................... 233
5.4.2.1 Thematic Equatives (or ‗Cleft-constructions‘) ...................................... 234
5.4.2.2 Emphatic Pronouns .............................................................................. 237
5.4.2.3 Exclusive Particles ................................................................................. 241
5.4.3 Further on Broad Focus............................................................................ 247
5.4.4 Information Unit and Clause .................................................................... 251
5.5 Conclusion ..................................................................................................... 254
CHAPTER SIX........................................................................................................... 256
TRANSITIVITY: REPRESENTATION OF EXPERIENCE ....................................... 256

6.1 Introduction.................................................................................................... 256
6.2 The Clause as a Representation of Experience ................................................ 256
6.3 Material Clauses: Processes of Doing-&-happening ....................................... 259
6.3.1 Transitive and Intransitive Material Clauses ............................................. 261
6.3.2 Extending Material Clauses in Delicacy: Additional Participants ............. 270
6.3.3 Concrete and Abstract Material Clauses ................................................... 281
6.4 Mental Clauses: Processes of Sensing ............................................................ 283
6.4.1 Direction of Sensing: Emanating and Impinging ...................................... 288
6.4.2 Projection in Mental Clauses .................................................................... 291
6.4.3 Principal Types of Mental Clauses ........................................................... 293
6.5 Relational Clauses: Processes of Being & Possessing ..................................... 297
6.5.1 Extending the types of relational clauses in delicacy ................................ 299
6.5.2 Intensive Clauses ..................................................................................... 300
xvi
6.5.2.1 Intensive Identification .......................................................................... 300
6.5.2.2 Intensive Attribution .............................................................................. 304
6.5.2.3 Projection in intensive clauses .............................................................. 308
6.5.3 Possessive Clauses ................................................................................... 310
6.5.4 Circumstantial Clauses ............................................................................. 312
6.6 Other Process Types ....................................................................................... 316
6.6.1 Verbal Clauses ......................................................................................... 317
6.6.2 Behavioural Clauses ................................................................................ 324
6.6.3 Existential Clauses ................................................................................... 332
6.7 Generalising across Process Types: The Transitive Model versus the
Ergative Model ................................................................................................ 337
6.7.1 The Transitive Model............................................................................... 337
6.7.2 The Ergative Model ................................................................................. 347
6.7.2.1 Middle Clauses .................................................................................... 352
6.7.2.2 Effective Clauses ................................................................................. 354
6.8 Conclusion ..................................................................................................... 357
CHAPTER SEVEN..................................................................................................... 359
SYNOPSIS: DAGAARE AND BEYOND .................................................................. 359

7.1 Introduction.................................................................................................... 359
7.2 Summary of Aims and Procedures.................................................................. 359
7.3 Summary of Description................................................................................. 360
7.4. Implications of the Study................................................................................ 365
7.4.1 Theoretical Contributions......................................................................... 365
7.4.2 Contributions to Research ........................................................................ 366
7.4.3 Practical Applications .............................................................................. 368
7.5 Final Remarks ................................................................................................ 369
APPENDIX ................................................................................................................ 372
REFERENCES ........................................................................................................... 389
xvii
LIST OF FIGURES
Page
Figure 1.1 Locating the Dagara Speech Community on the Map of

West Africa 2
Figure 1.2 Locating Dagaare in the Niger-Congo Language Phylum 4
Figure 2.1 Analysis, Description, Comparison and Theory in
Relation to one Another 19
Figure 2.2 The Hierarchy of Stratification 28
Figure 2.3 Function Structure for Transitivity and Mood 34
Figure 2.4 The System of MOOD of the English Clause 37
Figure 2.5 A Box Diagram of Syntagm and Structural Configurations 39
Figure 3.1 The Phonological Rank Scale of Dagaare 51
Figure 3.2 Vowel Chart for Dagaare Short Vowels 53
Figure 3.3 Illustration of the Syllable Structure 55
Figure 3.4 The System of TONE in the Syllable 56
Figure 3.5 The System of TONE in the Tone Group 60
Figure 3.6 The Dagaare Rank Scale 64
Figure 3.7 Illustration of the Structural Analysis of a Major Clause 98
Figure 3.8 Dagaare Clause Types Presented as a System Network 100
Figure 4.1 The Semantic System of SPEECH FUNCTION 112
Figure 4.2 The Modal Structure of the Dagaare Clause 118
Figure 4.3 Noun Group and Nominal Clause in Subject Position 120
Figure 4.4 Compound Verbal Group as Predicator 121
Figure 4.5 Realisations of Complement 132
Figure 4.6 Example of Two Complements in One Clause 133
Figure 4.7 Complements as Marked Theme 133
Figure 4.8 Types of Adjuncts by Metafunction 134
Figure 4.9 A System Network for Negotiation in Vocatives 137
Figure 4.10 A System Network of MOOD 141
Figure 4.11 Q-element Conflating with Clausal Functions 155
Figure 4.12 A System Network of the IMPERATIVE TYPE 162
Figure 4.13 A System Network of MODALITY in Dagaare 174
xviii
Figure 4.14 Polarity and Probability Represented as a Cline 175
Figure 5.1 A System Network of THEME SELECTION for Topical
Theme 208
Figure 5.2 A system Network of INFORMATION FOCUS in
Dagaare 223
Figure 5.3 Illustration of Information Structure and Focus Types in
Dagaare 227
Figure 5.4 Information Unit in Relation to Clause on the Rank Scale 254
Figure 6.1 Illustration of the Structure of Clauses of Cliency 275
Figure 6.2 Illustration of the Different Meanings of the Agentive
Particle 279
Figure 6.3 Illustration of Causative Mental Clauses 290
Figure 6.4 ‗Thinking‘ as a Mental and as a Behavioural Process 329
Figure 6.5 ‗Wanting‘ as a Mental and as a Behavioural Process 330
Figure 6.6 The Structure of Existential Clauses 336
Figure 6.7 The Structure of a Material Clause 336
Figure 6.8 The System of AGENCY in Dagaare 352
Figure 6.9 Benefactive versus Non-benefactive Effective Clauses 356
xix
LIST OF TABLES
Page
Table 2.1 Text Archive for the Study 44
Table 3.1 The Dagaare Alphabet Systems 63
Table 3.2 Orthographic Representation of Double Articulatory
Sounds 64
Table 3.3 Illustration of Different Realisations of the Imperfective
Suffix 67
Table 3.4 A List of Dagaare Catenative Verbs 81
Table 3.5 Dagaare Adverbial Particles 90
Table 3.6 Pronominal System of Dagaare 93
Table 3.7 List of Dagaare Particles across Metafunctions 95
Table 4.1 Analysis of Speech Roles and Functions in Text 1 111
Table 4.2 Grammatical Categories that Occur within the Scope
of the Predicator 122
Table 4.3 Realisations of the Negotiator in Non-affirmative Clauses 144
Table 4.4 Realisations of Different Polar Interrogative Types 149
Table 4.5 Question Words Used in Elemental Interrogative Clauses 154
Table 4.6 Daily Greetings and Responses in Dagaare 168
Table 4.7 Realisation of Polarity in the Indicative Clause 171
Table 4.8 The Realisation of Probability by Modal Particles 175
Table 4.9 The Realisation of Desirability by Modal Particles 178
Table 4.10 Negotiation Markers and their General Meanings across
Moods 181
Table 5.1 Thematic Analysis of a Text 206
Table 5.2 Illustration of Emphatic Pronouns in Different Clause
Positions 241
Table 5.3 Dagaare Exclusive Particles 243
Table 6.1 Verbs that Typically Realise the Process in
Impinging Mental Clauses 289
Table 6.2 Phenomenalisation of Sensing versus Subjective Sensing 291
Table 6.3 Contrast between Cognitive and Desiderative Clauses
xx
on Aspect Selection 294
Table 6.4 Characteristics of the Sub-types of Mental Clauses 296
Table 6.5 Illustrating Principal Types of Relational Clauses 299
Table 6.6 Options in Representing Quality in Intensive Clauses 306
Table 6.7 Examples of Verbs Realising the Process in Relational
Clauses 315
Table 6.8 Examples of Verbs Used in Different Types of
Behavioural Clauses 327
Table 6.9 Examples of Verbs Associated with Existential Clauses 336
Table 6.10 Transitivity Components in Relation to Realisation
of the Complement 339
Table 6.11 Generalising across Participant Roles in Process Types 349
Table 6.12 Examples of Verbs that Typically Realise Processes
in Middle and Effective Clauses 355
Table 7.1 Function-rank Matrix 371
xxi
ABBREVIATIONS AND SYMBOLS
// - tone unit boundary IDEO – ideophone

|| - internal clause boundary IND – indicative
||| - initial or final clause position INS – insistent
= clitic INST – instrument
[[ … ]] – downranked clause marker INT – interrogative marker
ACC – accusative IM – immediate
ADH – adhortative IMP – imperative
ADM – admonitive INTJ – interjection
ADP – adposition IPFV – imperfective
ADV – adverbial particle JUNC – juncture
ADVLZ – adverbializer M – mild
AFFR – affirmative MOD – modal
CAUS – causative N – non
CE – counter-expectation NEGO – Negotiator
CEN – central NEG – negative
COM – comitative NMLZ – nominaliser
COND – conditional OP – opinative
CONJ – conjunction PFV – perfective
CONT – continuative PL – plural
COP – copula POS – positive
DEF – definite PUR – purposive
DEM – demonstrative PROJ – projection marker
DIST – distal PROX – proximal
DUP – duplication PRT – particle (undefined)
EX – exclusive particle PST – past
EXCL – exclamation particle REL – relativiser
EXH – exhortative REM – remote
EXIST – existential REQ – requestive
EM – empathy RNG - range
EMP – emphatic S – strong
EVT – eventive particle SG – singular
xxii
FOC – focus VOC – vocative
FUT – future 1 – first person
HAB – habitual 2 – second person
HM – human 3 – third person
HST – hesitative
IDENT – identifying
xxiii
CHAPTER ONE
INTRODUCTION
1.1 Introduction
The aim of this study is to describe the grammar of Dagaare, specifically the
Lobr dialect, as a meaning-making resource. It addresses the question: how
does the Dagaare language work? The approach adopted is functional and
discourse-based. This chapter introduces and contextualises the study. It first
gives a geolinguistic overview of the Dagaare language, comprising its areal
scope, typological environment, dialects and its contact with other languages.
The nature and socio-political set-up of the Dagara society will then be
described. This is followed by a discussion on developments in Dagaare
studies to situate the objective of the present study in the extant literature.
Subsequent sections then specify the research problem, the aim of the study
and its value. The chapter ends with an outline of the thesis.
1.2 Geolinguistic Landscape of Dagaare
Dagaare (aka Dagara) is the language of the Dagara (or Dagaaba) people and,
genetically, it is a member of the Gur language family of the Niger-Congo
phylum. It is spoken in three West African countries, namely Ghana, Burkina
Faso, and Cote d‘Ivoire, with an estimated population between 1,500,000 and
2,000,000 native speakers (cf. Lewis, Simons & Fennig 2016; Bodomo 2000).
It is the dominant language of north-western Ghana and south-western
Burkina Faso. It is, however, a minority language in the north-eastern corner
of Cote d‘Ivoire. The geographical coverage of the native land of the Dagara
stretches from Bole in north-western Ghana through principal towns like Wa,
Jirapa, Nadowli, Daffiama, Nandom, Lawra and Hamale into Burkina Faso, as
far as Oronkua in the north, Diebougou in the north-west and Malba, west of
the Black Volta (Dakubu 2005). In Ghana, there are also Dagaare-speaking
communities in almost all regions, especially in the Brong Ahafo Region, the
Ashanti Region and the capital city of Accra. In Burkina Faso, the principal
speech communities are Batie in the Nounbiel Province, Gaoua in the Poni
1
Province, Diebougou in the Bougouriba Province, and Dano in the Ioba
Province. Outside these provinces, Dagaare is popular in Bobo Dioulasso, the
economic capital and Ouagadougou, the political capital (Modeste Somey,
p.c.). Figure 1.1 locates the speech community of Dagaare on the map of West
Africa.
The language has written registers since the 1950s. However, literacy
is largely restricted to liturgical and evangelical materials as well as materials
developed for mother tongue education. In Ghana, the Central Dagaare dialect
is preferred for literacy and education; while in Burkina Faso, the preferred
dialect is Lobr, each of which is the dominant dialect in their respective
countries.
Figure 1.1: Locating the Dagara speech community on the map of West
Africa
2
Dagaare is one of the nine languages approved by the government of Ghana to
be taught as subjects in schools and to be used as the medium of instruction in
the lower primary school. (English is the medium of instruction from Grade
four onwards). It is also the predominant language of local radio stations and
the Church in the Upper West Region (Ghana) and in south-western Burkina
Faso, especially in the Ioba, Nounbiel, and Bougouriba Provinces and, to a
limited extent, in the Poni province. Dagaare is taught in Wa College of
Education (Ghana) and is offered for degree courses in two public universities
in Ghana, namely the University of Ghana and the University of Education,
Winneba. Surprisingly, however, only 5%–10% native speakers are estimated
to be literate in the language (Lewis, Simons & Fennig 2016).
1.2.1 Typological Context of Dagaare
Figure 1.2 locates Dagaare in the Niger-Congo phylum. 1 Genetically, it

belongs to the Oti-Volta group of Gur languages (Swadesh et. al. 1966;
Bendor-Samuel 1971; Lewis, Simons & Fennig 2016). The typological
grouping presented in Figure 1.2 is based on phonological similarity and the
technique of lexicostatistics (cf. Swadesh et al. 1966; Williamson & Blench
2000; Dakubu 2005). Lexicostatistics is the main criterion and it examines the
extent to which identified sets of languages and dialects are similar or different
in their basic vocabulary. This approach was developed by Morris Swadesh in
the 1950‘s and applied to the Gur languages in his Glottochronology of the
Gur Languages (Swadesh et al. 1966). The most common lexical root
identified among these languages is gur, from which the languages derive their
genetic name.
Among the Gur languages, Dagaare is classified more specifically as a
northern Mabia language (cf. Bodomo (1993, 1997) on the term ‗Mabia‘). The
similarity between the Mabia languages in terms of vocabulary is comparable
to that among Romance languages such as French, Spanish, Portuguese, and
Italian. Although Dakubu (2005) questions their status as individual languages
as opposed to dialects, there is a higher mutual intelligibility among the
1
The name Niger-Congo was introduced by Joseph Greenberg and he classified Dagaare
among the Gur (aka Voltaic among French scholars) family of Niger-Congo (cf. Greenberg
1966).
3
attested dialects of Dagaare (see Figure 1.1) than between them and the other
Mabia languages. For instance, while speakers of the different dialects
identified in Figure 1.1 can sustain communication among themselves, with
each speaker using their own dialect, this is not possible between speakers of
any of the dialect and those of the other Mabia languages.
The relative degree of similarity between Dagaare and other Mabia
languages forms a language continuum or a cline of mutual intelligibility, as
the grouping in Figure 1.2 suggests. The closest to Dagaare is Safaliba, a
language with a small population of about 5, 000 speakers in the Northern
Region of Ghana (Dakubu 2005; Lewis, Simons & Fennig 2016). It is
specifically spoken in Mandare, Tankpe and Buanfo (all near Bole). The next
are Farefare, spoken in north-eastern Ghana, and Moore, the major indigenous
language of Burkina Faso and the largest Mabia language.
Figure 1.2: Locating Dagaare in the Niger-Congo Phylum (Adapted from

Dakubu 2005:4)
4
Dagaare also shares many lexicogrammatical similarities with Dagbani,
Mampruli, Hanga and Kusaal (cf. Swadesh et al. 1966). Buli shares few
vocabulary features with Dagaare and the other Mabia languages. In terms of
areal proximity, Dagaare is bordered by Sisalla (Grusi) to the north and east,
Lobi languages (i.e. Dyan, Gan, Doro, Doghosie, Lobire) to the west, Chakale
(Grusi) to the south-east, and Safaliba (Mabia) and Gonja (Guan) to the south.
Most of the languages around its border are those other than its immediate
sisters (Bodomo 1997; Dakubu 2005). It is separated from Moore to the north
by Sisalla (Grusi) and from Farefare to the east by Sisalla, Kasem (Grusi) and
Buli (Mabia). However, these are all Gur languages and share both
lexicogrammatical and phonological similarities with Dagaare.
1.2.2 Varieties of Dagaare: A Dialect Continuum
The name Dagaare designates a variety of dialects with relative mutual

intelligibility among them. As noted earlier, these dialects share more
common vocabulary items among them than they do with other Gur languages
(see also Swadesh et al. 1966). The Dagaare dialect continuum is normally
divided into four dialect areas, namely, northern, central, southern and western
dialect areas. While these labels have often been confused as designations of
dialects, each rather comprises a set of closely related dialects. The term
‗northern Dagaare‘ has also often been used interchangeably with ‗Dagara‘ in
the extant literature (in Ethnologue, it is rendered ‗Dagara, Northern‘). It
should, however, be noted that ‗Dagara‘ is a pronunciation variant of
‗Dagaare‘ by speakers in the northern dialect area and thus designates all
dialects of the language and the speakers as well. The confusion in dialect
classification generally owes to the fact that many Dagaare dialect
communities in Ghana do not use and may not accept designations for their
dialects beyond derivatives from names of towns (e.g. Jirbaale from Jirapa,
Losaale from Lawra and Nandome from Nandom) and often claim to be the
speakers of ‗Dagaare/Dagara proper‘ (see also Dakubu 2005). Speakers in
Burkina Faso, however, use and accept specific designations such as Birifor,
Lobr and Wiile. In this study, I will use local lables that have been used by
native speaker scholars of the specific dialects.
5
Although dialect variation and classification still needs further
investigation, six principal dialects are currently identified, namely Lobr,
Wiile (aka Wule or Ule), Central Dagaare (entered in Ethnologue as‗Dagaare,
Southern‘ and known among speakers of other dialects as ‗Ngmere‘), Waali
(a.k.a. Waala or Waale) and Birifor. Lobr and Wiile constitute the northern
dialect area. Wiile speakers are found around Lawra in Ghana and in Dano,
Guéguéré, Oronkua and Legmoin (or Lagman) in Burkina Faso; while Lobr
speakers extend from towns and villages around Lawra through Nandom and
Hamale (all in Ghana) into Burkina Faso, covering towns such as Dissin,
Mariatang, and Nyebo (Dakubu 2005). Central Dagaare is spoken in Jirapa,
Daffiama, Nadowli and their surrounding towns and villages. Waali is the
principal southern dialect and is spoken in Wa and towns and villages
surrounding it. Birifor consists of two sub-dialects (i.e. Northern and Southern
Birifor) with considerable variation and together they cover the western dialect
area of Dagaare. Birifor is spoken in Ghana, south of Wa, around Sawla and
Tuna, and in west of the Black Volta river in Burkina Faso, particularly
around Malba, Diebougou, Gaoua and Batié. It is also spoken in the north-
eastern corner of Côte d‘Ivoire.
Some classifications consider Waali as a separate language from
Dagaare and five Waali Dialects have been identified, namely Yeri Waali,
Fufula, Dolimi, Bulengee and Cherii (Lewis, Simons & Fennig 2016).
Swadesh et al. (1966) also list the following as separate languages: Dagara,
Dagaare (including Waali), Nura (the Wiile dialect as it is spoken in Lawra,
Ghana) and Birifor; while Ethnologue identifies Dagara, Dagaare, and Waali,
as separate languages. As mentioned earlier, however, the labels ‗Dagaare‘
and ‗Dagara‘ are pronunciation variants of the same language. In this study, I
will continue to use ‗Dagaare‘ for the language and ‗Dagara‘ or ‗Dagaaba‘
when referring to the people, for the sake of clarity.
1.2.3 Dagaare in Language Contact
The evolution of the various Dagaare dialects is partly explained by contact

between Dagaare and other languages. Several Dagaare-speaking communities
are found to have spoken other languages and there were periods during which
6
these communities were bilingual (Dakubu 2005). Dakubu (2005) observes
that the Birifor once spoke Lobi languages while speakers at Kaleo and the
Yeri sub-dialect of Waale once spoke Jula (Niger-Congo; Mande) and Moore
(Oti-Volta: Mabia) respectively. Goody (1954) also notes in the 1950‘s that
many hitherto Chakale (Oti-Volta: Grusi) speaking communities around Wa
had shifted to the Waale dialect.Today, the minority Sisalla speaking
communities around the Dagara speak Dagaare as a second language, which
shows that the language is still spreading.
Besides these shifts, Dagaare speakers are generally multilingual and
do not only code switch but also borrow from languages such as Hausa (Afro-
Asiatic: Chadic), Akan (Niger-Congo: Kwa), Jula or Dioula (Niger-Congo:
Mande), Moore (Niger-Congo: Gur), English and French. The Waale dialect,
whose speakers are predominantly Muslims, has borrowed words from Hausa
(Afro-Asiatic: Chadic) and Arabic (Afro-Asiatic: Semitic), especially those
expressing religious concepts (Dakubu 2005). On the other hand, dialects such
as Central Dagaare, Lobr and Wiile, whose speakers are predominantly
Catholic, have borrowed a few religious terms from Latin (e.g. mιsa, ‗church
service‘; virgo, ‗virgin; sitaana, ‗the devil‘; sakramatɩ, ‗sacrament‘; paapa,
‗pope‘). Akan is a majority language in Ghana, with about 35% ofthe
Ghanaian population as native speakers (Lewis, Simons & Fennig 2016),
while Jula (or Dioula) is the commercial lingua franca of Burkina Faso.
English and French, on the other hand, are the official languages of Ghana and
Burkina Faso respectively. Thus, while in Ghana many Dagaare speakers
speak Akan and English as a second language; in Burkina Faso, many
speakers speak Jula and French as a second language.
1.3 The Dagara Society
The Dagara (or Dagaaba) are part of a larger socio-cultural group of people in
West Africa called Mole-Dagbani (cf. Goody 1954). The largest ethnic group
among the Mole-Dagbani is Mossi (speakers of Moore), who are found in
Ghana, Burkina Faso, Mali and Togo. Just like their languages, the cultures of
individual ethnic groups of the Mole-Dagbani have evolved unique
characteristics, owing to migration in quest for land, escape from slave raids
7
and wars, divisions by colonialism, and contact with other cultures and with
different foreign religions, notably Islam and Christianity.
Like other Mole-Dagbani, the fairly homogenous ethnic group called
Dagaaba (or Dagara), traditionally, have a decentralised socio-political
system, organised around the Tengan-sob (‗custodian of the spirit of the land‘;
literally, ‗owner of the skin of the land‘) or Tindana (‗custodian or owner of
the land‘).2 The basic unit of this socio-political system is the yir (Lobr) or yiri
(Central Dagaare) (i.e. family), headed by the eldest male member. The use of
the word yir, which literally means ‗house‘, to refer to the Dagaaba family is
significant since all members of the patrilineal family often live together in
one large compound house. Critical decisions in the family are taken by the
family head in consultation with other adult male members (i.e. brothers,
cousins and nephews). Social hierarchy in the family and also the whole
community is determined by age and gender, with younger members deferring
to elder members and female members deferring to male members. Deference
among males and females is, however, relative to age; which shows that age
tends to be the key determiner of social power. Every family is also a religious
cult in itself, the object of worship being their family ancestors and may also
include family gods.
Beyond the family is yiilu (Lobr) or yiilong (Central Dagaare) (i.e. the
patriclan). Some patriclans cut across dialects. Each patriclan is bond by a set
of taboos and a totemic animal (e.g. the porcupine, deer or monkey), which is
believed to have brought fortune to their ancestors during an expedition or
escape from slave raids or wars. Members of a patriclan are yir-dem (meaning
people belonging to one house) and descend from a single male ancestor.
Village settlements cluster according to patriclans, with boundaries
demarcating them from other clans. These clusters are normally generations of
extended family members. A patriclan has no single head. Laws, customs and
critical decisions are decided by a council of heads of the various families.
Although members of each patriclan are spread all over the Dagaaba land,
they tend to maintain close ties with one another, especially by attending
funerals of clan members, even across national borders. In addition to the
2
Tengan-sob and Tindana are dialectal variants. Tengan-sob belongs to the Lobr dialect.
8
patriclan is a lineage system called bɛ́lʋ. In contradistinction with clan, lineage
is matrilineal, deriving from a common female ancestor, and it comprises Da,
Dabire, Hien, Kambire, Kambou, Kambouole, Kpoda, Meda, Somda, Some (cf.
J. D. Somé 2004: 43). Although the lineage system is lost in some
communities such as those of Central Dagaare speakers, it is particularly
prominent among Lobr and Wiile speakers. Among speakers in Burkina Faso
and Côte d‘Ivoire, one‘s lineage serves as one‘s surname.Customarily,
members of the same patriclan are forbidden to inter-marry, but people
descending from the same lineage can inter-marry.
Each lineage and patriclan normally has a corresponding lineage or
clan with whom they have a banter (i.e. teasing) relationship. Members of
clans and lineages in a banter relationship have mutual privileges over one
another to intrude into a clan‘s group fellowship, participate in their joy,
satirise their frivolities and limitations and trivialise matters of grave concern
with verbal wit. The Dagara as an ethnic group maintains a similar banter
relationship with the Farefare (another Mole-Dagbani ethnic group). With
regards to patriclans, the mutual roles between banter clans, are very crucial in
maintaining social cohesion as well as the psychological and spiritual health of
a clan. For instance, if members of a patriclan want to make peace with or
negotiate marriage with another clan, they often invite a banterer of the
receiving clan to be the mediator. In case of peace making, members of the
offended clan normally feel very much obliged to accept the intercession of
their banterer and make peace with the offender. A banter clan can also play
the same role in conflicts involving people from the same patriclan or even
family.
It is also believed that the intervention of members of a banter clan in
critical situations has a positive effect on one‘s emotions and psychological
state. Thus, in times of trouble such as the loss of a dear one, members of the
bereaved clan are often consoled by their banterer. It is, indeed, ideal that
grave news is conveyed by a member of a banter clan, which has the
psychological effect of mitigating the shock of the bereaved or the affected
person.
Spiritually, the intercession of a member of a banter clan can cool
down the anger of the spirit of a dead member of their ally clan and wade off
9
evil. Banter clans are also allowed anything a corresponding clan is tabooed
against, such as touching the corpse of someone who has commited suicide
(either a clan member or otherwise) or a dead totem found in one‘s trap in the
farm. The banter relationship between patriclans, in particular, is therefore
partly a way of circumventing an otherwise strict normative order of the
Dagara society, where, in any critical moment, one group is absolved of
deviations from the code in order to maintain a balance in community life.
The cluster of patriclans in a community (i.e. village or town) is
headed by the Tengan-sob (or Tindana), who is both a spiritual and a political
leader. He settles disputes between clans, pacifies the earth, Tengan (which is
deemed sacred), on behalf of offenders, and fine violators of laws and
customs. He takes decisions in consultation with a council of elders, made up
of respectable members of the community, particularly the eldest. The
Tindana is always a descendant of the first settler in a community. By
extension, in villages and towns where the Dagaaba are found outside their
traditional homeland, they normally consider themselves to be a community,
with the first Dagaaba settler as their leader. As Bodomo (1997) has noted, the
policy of indirect rule by the British colonial government in the late 19 th
century has, however, introduced chieftaincy among the Dagaaba and the
society is now organised around paramountcies and chiefdoms, leaving the
Tindana with only his spiritual role as custodian of the sacred land (see Goody
(1972) for a rich discussion of aspects of Dagara culture).
1.4 Developments in Dagaare Studies
As with many languages in West Africa, the earliest explorers of Dagaare

were Christian missionaries and European anthropologists. The first known
record on the language, according to Dakubu (2005), appears in the work of
the German missionary Christaller in 1889. It consists of the numbers one to
ten, noted to have been collected by a German trader called Krause. Christaller
(1889) used this list to classify Dagaare as genetically belonging to the Moore
(now renamed Mabia) branch of the Gur languages. Similar lists were
collected by colonial administrators and possibly other traders. J. D. Somé
(2004), for instance, notes that, in Burkina Faso, colonial officers recorded
10
names of people, settlements, rivers and hills in Dagaare. Rattary (1932), a
British colonial administrator and anthropologist, is, however, accredited with
being the first explorer of the language itself. He compared the dialects spoken
in Wa and Lawra (both in Ghana), based on ―a fairly extensive word list‖
(Dakubu 2005: 1). Rattray‘s (1932) study together with Christaller‘s
classification laid the foundation for later studies on dialect variation and in
comparative linguistics.
Relatively detailed engagement with the language actually began
around the 1950s with the missionary linguists , whose interest was to identify
its basic structure and vocabulary for the development of material for liturgy
and catechesis . B. A. Somé (2003) identifies some of the earliest materials
written in Dagaare as comprising hymns, prayer books and a catechism. These
were produced in 1953 and in the early 1960s by missionary administrators in
Bobo-dioulasso, Burkina Faso. Goody, an anthropologist, also contributed to
literacy development by publishing his transcription of a version of the Bagre
myth, a sacred religious text, in Dagaare (cf. Goody 1972: 316-377).
The linguistic analysis of missionaries mainly consisted of notes on
aspects of lexicogrammar (e.g. Durand, 1953; Wilson 1962a; Girault 1964)
and phonology (Wilson 1962b; Kennedy 1966; Girault 1967). While most of
these notes focused on individual dialects such as Central Dagaare (e.g.
Wilson 1962a; Kennedy 1966), Lobr (e.g. Girault 1964, 1967) and Wiile (P.
A. Somé 1962), others were across dialects (e.g. Wilson 1962b). P. A Somé‘s
(1962) study is significant for being one of the early works by a native
speaker. Studies that directly explored dialect variation around this period
include Callow (1969), which Bodomo (1997) identifies as the first study
focusing solely on dialect variation. Although Swadesh et al.‘s (1966)
Glottochronology of the Gur Languages includes a statistical comparison of
lexical items across Dagaare dialects, its focus is on language classification
rather than variation.
Native priests and church leaders have continued to be very influential
in the literacy development of Dagaare since the 1970s, particularly in the area
of orthography and the production of literature in the language (cf. J. D. Somé
2004). However, since the 1980s, the language has steadily garnered interest
among linguists around the world, culminating in the founding of the Journal
11
of Dagaare Studies by Adams Bodomo in 2001. The language has been
described by both native and non-native scholars from the point of view of
several linguistic approaches. Dagaare research has also been funded by
several universities and international funding institutions such as the Norway
University of Science and Technology, Norwegian Research Council,
Stanford-Berkeley Centre for African Studies, Hong Kong University, The
Hong Kong Polytechnic University, and The Hong Kong Research Grants
Council.
Work on the orthography has been consolidated and extended (e.g.
Bemile 1990; B. A. Somé 2003; J. D. Somé 2004). J. D. Somé (2004), for
instance, gives an interesting discussion of results on the orthography project
in Burkina Faso. In the area of phonology, scholars have explored tone (e.g.
Dakubu 1982; P. K Some 1995; Antilla & Bodomo 2000), phoneme contrasts
(e.g. Bemile 1983) and the vowel system (e.g. Angkaraaba 1997). Nakuma
(1999) examines phonemes and tones together and Bodomo (1997) also
includes a short discussion on tones and phonemes.
Many studies on Dagaare have, however, focused on lexicogrammar.
These studies can be classified into those that employ a structural-descriptive
approach and those that use the generative grammar approach. On the former,
work has been done on the noun (Bodomo 1994), the noun group (e.g.
Angkaraaba 1980; Bodomo 1994), the verbal group (Dong 1981; Dakubu
1989a), clause structure (Dakubu 1989b; Dakubu 2005) and coordination (Ali
2006). C. C. Somé (1984/85) and Bodomo (1997; 2000) also provide short
overviews of aspects of Lobr and Central Dagaare dialects respectively. Other
studies have focused on more specific lexicogrammatical resources such as
information focus (Dakubu & Saanchi 1997; Delplanque 1998), temporal and
modal markers (Delplanque 2000), aspectual forms of the verb (Saanchi
2003a) and spatial and locative constructions (Saanchi 2003b).
Research using the generative approach began in the early 1990s and is
led by Adams Bodomo and his team of researchers. Several aspects of the
grammar have been described and theorised, including verbal group
complexes (or ‗serial verbs constructions‘) (e.g. Bodomo 1993a; Hiraiwa &
Bodomo 2008), nominalisation (Bresnan & Bodomo 1997) and relativisation
(Bodomo & Hiraiwa 2004). Bodomo (1993b) also gives an X-bar account of
12
the language while Sakurai (2014) has recently examined focus marking in the
Central dialect, using the framework of Lexical Functional Grammar.
A few studies have also adopted a multilingual approach in exploring
Dagaare. It has been contrasted with Cantonese in the area of verbal group
complexes (Luke & Bodomo 2000/01) and ideophones (Bodomo 2006).
Bodomo also explores verbal group complexes in Dagaare as part of a
generative typology project on Gur languages (Bodomo 1993) and compares
the temporal system with another Gur language, Dagbani (Bodomo 1996).
Further, Nakuma (1990) compares Dagaare phonological systems with French,
while Dansieh (2008) situates his study on discourse markers within the
context of translation.
In summary, academic engagement with Dagaare can roughly be
classified into three stages: contrastive analysis based on a number of word
lists collected by traders and colonial administrators (1889-1932), literacy
developments and notes collected by missionaries (1950-1980), structuralist
descriptions since the 1980s and, finally, generative descriptions, beginning in
the early 1990s. These classifications only show landmarks in explorations on
the language and each group shade into the other. In contradistinction with the
functional approach adopted in the present study, both the structuralist and
generative studies present the grammar from the viewpoint of formal elements
in the clause, and unlike the present study, they are not based on the
systematic empirical investigation of naturally occurring texts.
1.5 The Present Study
1.5.1 Problem Statement
As mentioned earlier, studies on Dagaare by missionaries were applied in

producing literature on the language and subsequent research has also
provided material for teaching Dagaare to both native and non-native speakers
and for extending work on orthography. The present study generally derives
from and contributes to these previous studies. On the other hand, it provides
an alternative path to structural and formal oriented explorations of the
lexicogrammar of the language. One limitation of the focus on forms is that it
does not provide us with information about the meaning potential behind the
13
grammatical forms described and it is not based on empirical evidence from
naturally occurring texts in context.
Many of the studies reviewed above, for instance, focus on isolated
units within the clause, such as the verb group and noun group, rather than
giving a comprehensive account of the grammar as a meaning potential. In
addition, they are mainly based on elicited and introspectively constructed
data rather than copious naturally occurring texts. The consequence of these
limitations is that the explanation of some of the formal elements is not
complete and explicit since they are not related to their discourse context and
their systemic environment, namely other linguistic features they interact or
contrast with in the language. Both Bodomo (1993) and Dakubu (2005)
acknowledge the need for a pragmatic (or discourse based) approach for
clarifying the functions of aspects of Dagaare lexicogrammar. The present
study is the first step in this direction though it does not claim to provide a
perfect account without limitations. In fact, many of the issues raised here are
not limited to Dagaare but to the description of many other African languages.
As Heine and Nurse (2000: 5) note:
The quality and quantity of the documentation for African

languages ranges from fairly high to nil. We say ‗fairly high‘
because no African language has been documented or analysed
to the extent of the better researched European or Asian
languages. If we define ‗fairly high‘ as having a reasonably
accurate and comprehensive reference grammar available, then
less than a hundred African languages are in this category. For
most, the documentation consists of an inadequate grammar, an
analysis of part of the language, an article or two. For yet
others, all we have is a reliable word list, or less than that.
Although a few more grammars on African languages have certainly been

produced since the beginning of the 21st century (see e.g. Akerejola (2005) and
Atoyebi (2010) on Oko), the current situation essentially reflects this
observation. In this light, the present study should be considered a preliminary
comprehensive account of Dagaare and a contribution to the study of African
14
languages. The broader objective is to contribute to the long-term research
agenda of functional language typology, which explores the similarities and
differences across languages by grounding the description in discourse and
focusing on meaning. Thus, the study adds to the long tradition of research
exemplified by Hopper and Thompson (1980), Comrie (1989), Bybee, Perkins
and Pagliuca (1994), Heine, Claudi and Hünnemeyer (1991), Caffarel, Martin
and Matthiessen (2004), Güldemann (2008), to mention just a few, although
the focus here is on the particular rather than general typology. The next
section gives a more specific statement of the aim of the study.
1.5.2 Aim of the Study

The aim of the study is to explore the lexicogrammar of Dagaare, particularly
the Lobr dialect, froma systemic functional point of view. This means that the
description is oriented towards context, based on discourse and focuses on
meaning (Caffarel et al. 2004). It is ‗systemic‘ because it views the grammar
as a set of choices in wording for realising meaning. On the other hand, it is
‗functional‘ because it is the grammar of usage and examines the
lexicogrammatical systems that realise each of three metafunctions of
language: ideational, interpersonal and textual functions (cf. Halliday 1970;
Halliday & Matthiessen 2014; see also Chapter 2, Section 2.4.3 for details).
Priority is given to the clause since it is the basic unit for realising meaning in
discourse although other units have also been examined as much as it is
possible within the time limitation of the study. In general, the description
goes beyond formal categories of the language to consider the general
meaning potential underlying their use.
1.5.3 Significance of the Study
The study is useful to scholars and students interested in language and society,
language typology and grammaticalisation, systemic functional typology,
West African languages, and Dagaare studies. It is also valuable to
professionals working with Dagaare in various practical contexts, including
curriculum and language education, translation, and orthography.
As indicated earlier, the systemic functional approach adopted by the
15
study examines language in social context. The analysis and discussion of
Dagaare here is thus oriented towards the function of language in its totality as
a resource for enacting roles and relationships, construing experience and
organising text-in-interaction. In other words, it presents the culture of the
Dagara society by focusing on the key resource for organising meaning. The
results will therefore be interesting to scholars in anthropology, sociology,
ethnography and all others with various interests in language and society.
Studies on language typology, within their practical limitations, have
often focused on isolated forms such as focus markers, relative clauses and
clause final or stance particles, or individual systems such as ASPECT and
TENSE, across languages (e.g. Hopper & Thompson 1980; Heine 1983; Comrie
1989). These studies have been valuable in showing discourse meanings

expressed in a wide range of languages and universal paths through which the
different lexicogrammatical forms expressing these meanings evolve (cf.
Heine 2002a). Their results have served as a guide for the description of
Dagaare in the present study. More importantly, however, the present study
provides a complementary approach to these studies by holistically
investigating how the different lexicogrammatical resources work togetheras a
meaning potential in a single language. The study puts grammar into
perspective as meaning evolving in context. The insights revealed by this
holistic approach are that we can observe how the various systemic features
relate to other features in their environment. The results of the study will
therefore be relevant to typologists and researchers working on other
multilingual studies and are interested in West African languages.
In connection with the above, the study contributes to the growing
body of studies in systemic functional typology. Systemic typology has
increasingly garnered more interest since the beginning of the twenty-first
century and languages such as Tagalog, Thai, Japanese, and Vietnamese have
been described (cf. Caffarel, Martin and Matthiessen 2004& references
therein). As noted by Caffarel et al. (2004), research in this area still needs to
be extended and developed through the exploration of many more languages.
Since the present study is the first attempt to examine a Gur language from a
systemic point of view, the findings will contribute to existing knowledge
about the similarities and differences among languages in their realisations of
16
meaning.
One motivation for the study is also to provide a workable description
that will aid current and future professional activities involving Dagaare.
Current ongoing activities include work on orthography, translation,
lexicography, curriculum development, and language education. In addition to
these, the study aims to motivate discourse analysis in critical areas such as
forensics, media, healthcare, service encounters, politics, education, religion
and popular or folk culture. The description is therefore supported with a lot of
text analysis and the analysis of the grammar is made explicit as much as
possible. Many grammatical forms and features that have not been included in
previous descriptions have been examined for the first time and the meanings
of previously identified forms have been clarified and grounded in discourse.
It is hoped that different professionals will test this description through their
applications and point to its limitations for further research.
1.6 Outline of the Thesis
The thesis is organised into seven chapters. This chapter has contextualised the
study within the language situation surrounding Dagaare. It has also provided
the aim and motivations for the study. Chapter Two will discuss the theoretical
and methodological approaches adopted in the study. Chapter Three gives a
compendious profile of the linguistic architecture of Dagaare, its
characterology, as it were. The analysis here includes both the clause and units
below the clause as well as phonological resources. The chapter thus provides
a general context for readers by anticipating discussion in subsequent chapters.
Chapter Four examines the interpersonal grammar of Dagaare, focusing on the
system of MOOD, POLARITY and the modal assessment systems of MODALITY
and NEGOTIATION (i.e. attitudinal stance). Chapter Five discusses the textual
systems of the clause, comprising THEME and INFORMATION. It shows how
speakers organise the flow of discourse into processable chuncks of meaning.
Chapter Six focuses on the experiential system of TRANSITIVITY. It discusses
the process types mapping out different domains of experienceas well as
generalisations across these process types. Chapter seven summarises the
study and discusses implications of the description of Dagaare in this thesis.
17
CHAPTER TWO
THEORY AND APPROACH
2.1 Introduction
It was mentioned in the preceding chapter that this study adopts a functional
approach to language and that the specific functional approach adopted is
systemic functional linguistics (SFL). The present chapter discusses this
approach to provide a theoretical background for the discussion in subsequent
chapters. The chapter first examines how systemic linguistics conceptualise
the relationship between linguistic theory and language description (Section
2.2). It then proceeds to locate language within a typology of general systems
that define our temporal existence (Section 2.3). Following this, the chapter
will map various regions of language that are relevant to this study, based on
systemic functional theoretical dimensions (Section 2.4). The final sections of
the chapter describe specific methods and procedures employed in
implementing these theoretical issues in the present study (Section 2.5.1 &
2.5.2). Limitations of the study are also highlighted (Section 2.5.3).
2.2 Linguistic Theory and Language Description
Systemic functional linguistics is a general linguistic theory (cf. Caffarel,

Martin & Matthiessen 2004; Halliday 2009). It is ‗general‘ in the sense that it
is a general theory of language for the description of particular languages
(Halliday 2009). Description of particular languages here is defined broadly to
cover work on individual languages as well as accounts that are contrastive
and cross-linguistic in nature. The theory is characterized as ‗linguistic‘
because its primary object of study is language. This means that the questions
it investigates are those that have to do with language. In investigating
language, however, the investigator may have to explore other phenomena
since language is embedded in context and necessarily interacts with other
systems (Halliday 2009).
SFL clearly distinguishes between theory and description as two
complementary resources in linguistic science (Matthiessen 2013a, see Figure
18
2.1; also see Teruya & Matthiessen (2015) and Mwinlaaru & Xuan (2016) for
reviews of studies on systemic language description and typology).
Figure 2.1 Analysis, description, comparison and theory in relation to one

another (Matthiessen2013a: 141)
Theory is a designed system as opposed to an evolved system such as

language. It consists of interrelated concepts that are systematically organised
towards achieving potentially explicit goals (Halliday 2009). Generally, the
components or dimensions in a theory are interdependent and mutually
defining. SFL theory is designed as an enabling resource to guide particular
descriptions. This means that the theory does not posit universal linguistic
categories or structures for language, as a way of guarding against the
imposition of the description of one language upon another — in view of the
past tendency to treat features of dominant languages such as Latin or English
as ‗universal‘. For instance, although the theory posits that every language
organises its lexicogrammatical resources into a fixed, identifiable number of
ranks, it does not claim the universality of specific ranks such as clause,
phrase/group, word and morpheme. Every language is considered as a unique
19
manifestation of the semiotic system called language and categorial and
structural labels must emerge from the context of describing the language.
Likewise, the extent to which languages are different and/or similar in terms
of the packaging of lexicogrammatical resources, for instance, must emerge
from the context of typology oriented descriptions.
The relationship between theory and description is a matter of
abstraction (Caffarel, Martin & Matthiessen 2004). Theory is an abstraction
based on language as we have come to understand it from the description of
one language to another. SFL theory emerged and has evolved as an enabling
resource within the context of discourse-based descriptions of language and,
recently, other semiotic systems. It initially came out of Halliday‘s
explorations of various dialects of Cantonese in the 1940s and description of
Mandarin in the 1950s and has been further developed by his descriptions of
English since the 1960s.
These efforts were contributions to an already existing research agenda
led by J. R. Firth within his system-structure theory. This framework forms the
core of SFL theory, and, over the years, it has expanded and evolved in
response to new challenges posed by new contexts of descriptions and
applications (Halliday 2009; Matthiessen 2007). Apart from the core structure
provided by J. R. Firth‘s system-structure theory, it has built on several other
linguistics schools such as the Prague School, Lamb‘s relational network
theory (formerly, stratificational grammar) and the Copenhagen School, in
particular, work by Hjelmslev (e.g. Hjelmslev 1969[1943]). Concepts have
been drawn from these frameworks and redefined in order to deepen our
understanding of language and enable descriptions.
Description itself is an abstraction of, or rather, a generalisation from
the analysis of particular text instances. Language is observable as text,
defined as spoken or written discourse (and by extension other potential
semiotic resources such as gestures and images). Text, the domain of analysis,
therefore, serves as the entry point for investigators into the linguistic system
they want to describe. Description may aim at the linguistic system of the
community as a whole or it may be aimed at some register (i.e. a sub-system)
in the community. In either case, the description needs to be informed by the
analysis of particular discourses and the analyst will have to shuffle between
20
developing general categories and features (i.e. description) and testing them
on text instances (i.e. analysis).
As mentioned earlier, both theory and descriptions are ‗resources‘ in
the science of language; they provide resources for applications in different
contexts. While linguistic theory primarily serves as a resource for the
description of particular languages and other semiotic systems, the
descriptions that accrue become resources that are ‗appliable‘ for solving
problems in practical contexts. Many descriptions of language have been used
for typological generalisations (e.g. Bybee, Perkins and Pagliuca 1994; Heine
& Kuteva 2002a, b; Matthiessen 2004; Teruya et al. 2007; Matthiessen,
Teruya & Wu 2008; Güldemann, Zerbian & Zimmermann 2015) and as
models for the description of other languages through the technique of
―transfer comparison‖ (Caffarel, Martin & Matthiessen 2004: 15). This kind of
application is, however, cyclical since typological generalisations, in turn, are
useful resources for describing individual languages. Apart from this cyclical
use, however, SFL descriptions have been used in the contexts of education
(e.g. Christie & Unsworth 2005; Rose & Martin 2012), translation (e.g. Steiner
2002, 2005; Kunz et al. 2014 and references therein), stylistics (e.g. Lukin &
Webster 2005; Simpson 2014; Mwinlaaru 2014; Webster 2015), and
computational linguistics (e.g. Henrici 1981; Matthiessen & Bateman 1991;
O‘Donnell & Bateman 2005; Bateman & O‘Donnell 2015). They have also
been used for discourse analysis, towards solving practical problems in
contexts such as forensics (e.g. Martin et al. 2013) and healthcare (e.g. Slade
et al. 2008, 2011, 2015; Matthiessen 2013b).
Thus, just as for theory and description, SFL clearly distinguishes
between descriptions and their applications in different contexts. This
distinction is important since descriptions that are localised or oriented
towards particular problems tend to be narrow and constrained in such a way
that they limit our understanding of the language described and the range of
applications of the description. This observation does not condemn context-
specific descriptions; rather, it calls attention to a distinction between what is
described, language, and the phenomenon for which the description is done.
An investigator who does not make this distinction may end up describing the
former as though it were the latter. One area in which this distinction has
21
proved useful in SFL context is computational application and artificial
intelligence. In these areas, descriptive categories have had to be converted
into mathematical formalisations and abstractions in order to make them
machine friendly (cf. Henrici 1981). These abstract formalisations can,
however, not be taken to be the best descriptive representations of language.
As an appliable theory, SFL is grounded in language while being sensitive to
other systems language interacts with. The next section discusses the general
systems language interacts with.
2.3 Language within a Typology of Systems
SFL theory, and to a large extend functional linguistics, in general, locates

language within a typology of systems that together define our cosmic
environment and our human condition. These systems are ordered in terms of
their increasing complexity and their temporal appearance in the time-scale of
cosmic evolution. They comprise physical systems, biological (including
cognitive) systems, social systems and semiotic systems (Hockett 1942; Pustet
2003; Halliday 2005; Matthiessen 2007). Systems of a higher order are also
systems of orders below it. (The discussion on these systems here is mainly
based on Halliday (2005) and Matthiessen (2007: 545-547) and reference may
be made to them for further details). Physical systems are first-order systems,
and they emerged about 15 billion years ago with the so-called big-bang in the
genesis of cosmos. They are widely and orderly distributed throughout the
universe. The mode of organisation of physical systems is composition; each
component is connected to others by some kind of design to create a unified
whole. Physical systems are governed by the laws of physics and their mode
of cosmogenesis (or change) is causation, the principle of cause-and-effect.
Biological systems are second-order systems. This implies that they are
more complex than physical systems and emerged after physical systems in
the course of cosmogenesis. They are essentially physical systems with an
additional property of ‗life‘. Biological systems are estimated to have
appeared on earth roughly about 4.5 billion years ago. Specific characteristics
of biological systems are that they self-replicate through reproduction and
individuate into unique organisms, clustering into different groups of species.
22
They can only exist under life sustaining conditions and this explains why they
must have appeared many years after the emergence of physical systems.
Their mode of change is through adaptation or innovation rather than
causation. However, like physical systems, their mode of organisation is
composition, comprising cells, tissues and organs. One key aspect of the
evolution of biological systems is the increasing complexity of their cognitive
capacity of which humans have evolved as the most cognitively developed
beings.
Social systems are third-order systems and are thus biological systems
plus the property of value or social order. Social systems range from simple
communities like insect colonies to the sophisticated cosmopolitans of the
modern world. As Matthiessen (2007) notes, the first establishment of social
systems cannot be estimated. It could be as far back as the emergence of
biological systems. For the hominid line of evolution, from which the homo
sapiens emerged, however, social organisation may be dated with the
emergence of primates about sixty million years ago (Matthiessen 2007).
Social organisations are structured into social groups of varying sizes and
functions. In these social groups, biological individuals (organisms) acquire
the value of persons and are connected with others through a network of roles
and relationships. The most basic social group in human society is the family
and as the individual person grows, s/he gradually accumulates different
identities through his/her roles in an array of social affiliations. Thus, like
other systems, social systems are compositional and they evolve in time
through adaptation.
Fourth-order systems are semiotic systems; that is, social systems plus
meaning. In addition to constituency (or composition), semiotic systems are
stratified as ‗content‘ and ‗expression‘, the latter realising the former. Among
semiotic systems, language is defined as a higher-order semiotic, a complex
adaptive system of meaning which is constantly evolving. It is distinguished
from other semiotic systems in the sense that it is semogenic; that is, it creates
meaning, both synchronically and diachronically (Halliday 2009; Heine,
Claudi & Hünnemeyer 1991). In addition, language has evolved beyond the
simple bistratal layers of content and expression (as in animal communication)
into a four stratal system, comprising semantics, lexicogrammar,
23
phonology/graphology and phonetics/graphetics (Caffarel, Martin &
Matthiessen 2004). A final unique feature of language is that it has evolved
into a metafunctional system, embodying several meanings that are
complementary and simultaneous in its structural organisation. That is, it can
mean many things at a time. This metafunctional orientation of language
contrasts it with semiotic systems that are microfunctional, such as animal
communication and some forms of non-verbal communication such as the
traffic light, only meaning one thing at a time (Caffarel, Martin & Matthiessen
2004).
As the highest order of systems, language is inclusive of all the lower
systems. In other words, linguistic ―meaning is socially constructed,
biologically activated and exchanged through physical channels‖ (Halliday
2003: 2; Matthiessen 2007: 547). In George Lakoff‘s terms, language is a
system of ―embodied meaning‖ (Lakoff 2007). Socially, language functions in
discourse to enact roles and identities and establish relationship. It also serves
as the repository and transmitter of cultural and shared knowledge. These
discourse meanings are first and foremost realised by the semantics stratum of
language, which thus become an interface resource that transforms what is not
linguistic meaning into linguistic meaning (Matthiessen 2007).
Biologically, language is said to have co-evolved with the brain, which
is its immediate environment (Halliday 2008: 147, 168-169). This is true for
both the development of language in the human species (i.e. phylogenesis) and
its development in the child learning the mother tongue (i.e. ontogenesis).
Language is the product of the brain‘s capacity to ‗mean‘ and linguistic
creativity and innovation reflects the infinite capacity of the brain to mean.
This productive power of the brain, by extension, includes its capacity to
comprehend. Another biological dimension of language is the articulatory and
auditory functions of the relevant organs of speech and perception
respectively. The transmission of language from one speaker to the other is a
physical mechanism involving air and sound waves. All these inter-systemic
aspects of language are included within the domain of linguistic science.
Hockett (1942) highlights this inter-systemic dimension by charactering
language behaviour as biophysical and biosocial. Together, the inter-
relationships characterise language as a complex system and challenges
24
linguistic theory to accommodate this complexity. In the following section, we
will discuss the core dimensions of language that are relevant to the present
study.
2.4 Systemic Organisation of Language

Systemic functional theory has been developed with the aim to mirror the
nature of language as it has been revealed by already existing descriptions.
The theory is composed of a set of interrelated concepts that are called the
‗dimensions‘ or ‗vectors‘ of language. Together, these dimensions construe
language as a semiotic architecture that can be viewed from different angles.
This architecture is, however, fluid and susceptible to change because it
theorises an evolved system rather than a designed system. A viewpoint from
the different angles reveals different aspects of language. The dimensions of
language serve as a road map for investigators interested in language for
different purposes. For the purpose of this study, concepts in these dimensions
that are relevant for the description of lexicogrammar will be brought into
focus. The dimensions discussed are the cline of instantiation, the hierarchy of
stratification, the spectrum of metafunction, the rank scale, axis and
semogenesis (or, simply, genesis). One important concept which permeates the
discussion on these dimensions is that of complementarity. It explains that
many components of language are paired as two aspects of one and the same
phenomenon and need to be viewed as complementary to each other. An
example is the relationship between speaking and writing, each representing
the same phenomenon called ‗language‘ but in different states of existence
(see Halliday 2008 for a detail discussion on complementarities in language).
2.4.1 Instantiation
In Section 2.2, it was mentioned that, in the course of language description,
the investigator has to move between analysing text and establishing linguistic
categories that are general in the linguistic system. Instantiation is the
theoretical dimension that explains this relationship between text and language
as a system. These two perspectives of language, as text and as system, are not
new in linguistic theory. Antecedents are de Saussure‘s (1916 [2011])
25
distinction between langue and parole and Chomsky‘s (1965) notions of
competence and performance. However, these scholars focus the goals of
linguistics on langue or competence. As Halliday (2009) observes, post-
structural linguists reacted to this skewed characterisation of linguistics by
focusing their attention on text rather than system. The dimension of
instantiation is a synthesis of these two perspectives on language. Rather than
defining linguistic science as a focus on one of these two extreme poles of
language, it characterises it as both (see Halliday 2008: Ch. 3).
System is defined as a meaning potential; a large semiotic resource that
offers choices to speakers in a speech fellowship to make meaning. It is
therefore the totality of the potential range of meanings that members of the
speech fellowship can produce. Text, on the other hand, is the tangible or
observable aspect of language; its mode of existence. Language is produced
and consumed as text and this aspect is what is readily available to the
investigator. The relationship between system and text is that of
‗instantiation‘, defined as a semogenic (i.e. meaning creation) process by
which linguistic resources are selected from the system to produce text. Thus,
as system has been defined as a generalised meaning potential, text is an
instance of this potential – in technical terms, text instantiates the system.
One characteristic of language that is revealed by this semogenic
process is that the system is probabilistic. Given a pair of contrasting features
in the system, we can examine quantitatively the probable number of times
one is chosen over the other in actual text instances. In a corpus of about 18
million words, Halliday and James (1993) found that, for the system of
POLARITY in English, the choice between positive and negative clauses is in a
ratio of 09: 01 respectively. In other words, in any particular English text,
there is a probability that the instances of positive clauses it contains will be
90% and instances of negative clauses will be 10%. Any deviation from this
systemic probability could therefore be interpreted as stylistically motivated.
Generally, it has been hypothesised that for any two terms in a system, the
probable ratio of their probability of occurrence will be either 05: 05 or 09: 01
(e.g. Halliday 2008: 41-42).
Further, since meaning is socially constructed, both the system and
instance manifestations of language realise context. This idea follows Firth‘s
26
notion of context of situation, which he borrows from Malinowski (1923).
Halliday has, however, added Malinowski‘s (1923) notion of context of
culture as well. As text realises and, at the same time, construes the context of
situation in which it is produced, the system realises and as well construes the
culture of the speech fellowship. The relationship between language and
culture posited in SFL is a reactivation of Whorf‘s conceptualisation of the
close interaction between language and world-view (see Carroll, 1956).
In SFL, the relationship between language and culture has been
broadened to include other variables. The link between system and text is
construed as a cline, with a series of intervening regions which have roughly
been identified as register and genre/text-type. Register is characterised as a
sub-potential that readjust the probabilities of the system for the functioning of
a particular institution in the speech fellowship. Genre, on the other hand, is a
variety of language that corresponds to particular situation types; it is a regular
text-type that realises specific social activities in a cultural institution (cf.
Mwinlaaru 2017). Correspondingly, context is a phased phenomenon
comprising culture > institution > situation type > situation (Matthiessen
2013a). Given the registerial variability of language, it is always necessary to
include texts from different descriptions in the description of language.
The implication of the cline of instantiation for language description is
that the analyst needs to have a trinocular vision of language (Halliday,
1996). Although the primary focus of description is on the system, the point of
entry into this system is text, and, once the investigation is initiated, the
analysis will have to move between (i) text (i.e. a view from below) and (ii)
system (i.e. a view from above) and (iii) across several registers (i.e. a view
from roundabout) in order to ensure a complete description. The
complementarity between system and text is therefore said to be that of focus.
In our discussion on stratification in the next section, we will again highlight
another complementarity in language, not of focus but of angle, and indicate
another perspective of trinocularity in language description.
2.4.2 Stratification
It has been mentioned in Section 2.3 that language is distinguished from other
27
semiotic systems as a four stratal system. This dimension of language is
represented in SFL theory as a hierarchy of stratification, a concept which
builds on Hjelmslev‘s glossematics (e.g. Hjelmslev 1954) and Lamb‘s (1966)
stratificational grammar (see Halliday 2008: 13-14; 2009; Bache 2010). This
dimension shows that the resources of language are organised into a hierarchy
of strata which are related by realisation (see Figure 2.2). The model in Figure
2.2 presents stratification as comprising language and context. The strata of
language consist of semantics, lexicogrammar, phonology (including
graphology) and phonetics (including graphetics). These strata together are
embedded in context, since language ultimately realises context. For the
purpose of this study, context here is defined as context of culture (see
discussion in Section 2.4.1 above for other phases of context). The
components of context are field, tenor, and mode.
Figure 2.2: The hierarchy of stratification
Field has two complementary dimensions, namely, field as subject matter and
field of activity. At the level of culture, field as subject matter is the repository
of the world-view, ideologies and inter-subjective (or shared) knowledge of
the speech fellowship. Field of activity, on the other hand, comprises socio-
semiotic processes such as reporting, sharing, recreating, doing, etc. by which
field of subject matter is given substance (see Matthiessen 2015 for detail
discussion of socio-semiotic processes). Tenor, the parameter of power,
28
solidarity and formality, encapsulates the social structure of the speech
fellowship – social roles and relationships and the networks between them.
Mode is concerned with the role played by language (and, by extension, other
semiotic or even social systems) in context. It is a three-tier component
comprising medium (e.g. written, spoken or multimodal), channel (e.g. face-
to-face or virtual) and rhetorical manner or styles of expression (e.g. didactic,
persuasive, exhortatory, polemic and performative styles).
Semantics and lexicogrammar form the content plane of language.
Semantics is one of the two outer strata of language (the other being
phonetics) and interacts with context to construe meaning that is otherwise
non-linguistic into linguistic meaning (see also Section 2.3). Specifically,
semantics is the construction of the world-view of the speech fellowship and
the enactment of the various tenor dimensions of interactions into content that
can be communicated. In other words, language is first and foremost meaning.
Semantics or meaning is defined here as a spectrum of metafunctions,
consisting of ideational, interpersonal and textual functions, each realising the
field, tenor and mode parameters of context respectively (metafunctions are
explained in Section 2.4.3).
Lexicogrammar is the realisation of meaning as wording. It is unique
to language and, as Halliday (e.g. Halliday 2008: 44) has noted, it is the power
house of this higher-order semiotic. It is this stratum which gives substance to
meaning by offering an array of choices to speakers for realising linguistic
structure. Since language is a meaning potential, grammar stands in a natural
relationship to semantics. Cross-linguistic studies have shown that all
languages dedicate lexicogrammatical resources for realising ideational,
interpersonal and textual meanings. For many languages, some of these
resources have been identified as respectively the systems of TRANSITIVITY,
MOOD and THEME at the level of clause rank (cf. Caffarel, Martin &
Matthiessen 2004).
Lexicogrammar is conceived of as a complementarity which is
organised as a scale of delicacy, with grammar as the least delicate angle and
lexis as the most delicate. Grammar here comprises what has traditionally
been divided into syntax and morphology. In the complementarity between
grammar and lexis, grammar consists of close systems and most general and
29
predictable meanings in language while lexis comprises specific, open
systems. Both meanings are wording, defined as the individual words and how
they are syntactically organised. As we move from the grammar pole to the
lexis pole, features (defined as lexicogrammatical meanings) increase in
specificity (or, technically, delicacy) until the most specific or most delicate
features are specified. This movement in delicacy is typically represented in
SFL descriptions as system networks. It must, however, be added quickly that
this statement of movement in delicacy is only a theoretical postulate to guide
language description and neither represent cognition nor real time selections in
actual language use. It may be reasonable to assume that speakers will
naturally do these selections simultaneously. This theoretical representation
has turned out to be useful in modelling computational text generation (see
Henrici 1981; Matthiessen & Bateman 1991).
The idea that grammar and lexis form a continuum has been
corroborated by several typological studies where some linguistic resources
such as modality, tense and aspect tend to lie on the boundary between
grammar and lexis (cf. Heine, Claudi, Hünnemeyer 1991; Traugott & Heine
1991; Bybee, Perkins and Pagliuca 1994). Halliday (2008: 66-67, 70) also
identifies English prepositions as lying in the mid region between grammar
and lexis, sharing the syntactic behaviour of verbs as minor processes and
possessing a grammatical character as close system, with a predictable
syntactic position and generalised meanings.
Although lexicogrammar is the power house of language, its physical
existence is in the form of expression. This expression plane of language is
phonology (or graphology) and phonetics (or graphetics). In English, for
instance, various lexicogrammatical features such as mood and information
focus are expressed by prosodic features such as intonation, tone and stress
(see Halliday & Greaves 2008). Phonology thus expresses linguistic meaning
as sounding. Phonetics is the lower outer stratum of language. As semantics
interacts with the ecosocial environment to construe and enact meaning,
phonetics interacts with the bio-physiological environment of language to
express meaning as sounding. Thus, it should also be noted that the realisation
relationship among strata is a kind of metaredundacy (Halliday 1992;
Matthiessen 2007; Martin 2010), a term originally used by Lemke (1984).
30
That is, lexicogrammar realises context via the realisation of semantics;
phonology realises semantics via the realisation of lexicogrammar and
phonetics realises lexicogrammar via the realisation of phonology.
The hierarchy of stratification is relevant for language description. As
with instantiation, it gives the analyst a trinocular perspective on the resources
of language. In describing the grammar of a language, the investigator may
examine grammatical features (i) from above in the semantics stratum to
identify their discourse functions, (ii) from roundabout in the lexicogrammar
to explore how each feature relate to other grammatical features in its systemic
environment and (iii) from below in the phonological stratum to identify the
sounding realisations of grammatical features and forms.
2.4.3 Metafunction
The spectrum of metafunction represents different modes of meaning in
language (Matthiessen, 2007). It is presented in the theory as a spectrum
because it permeates all strata of language. Meaning is the underlying function
of language and is therefore activated at every stratum of its organisation. This
diffusion of meaning across strata relates to the idea of metaredundancy
mentioned in Section 2.4.2 above. Ideational meaning is the referential or
representation function of language. Every language is a resource for
construing the speakers‘ experience of phenomena in the world in several
alternative ways. The ideational metafunction has (evolved) two modes of
construing experience, the experiential and the logical. They are
complementary because they are often employed together in a single language
in construing ideational meaning in different domains. On the other hand, they
are alternative in the sense that different languages may employ either the
experiential mode or the logical mode in construing the same phenomenon. An
example of the former is the experiential system of TRANSIVITY and the logical
system of TAXIS (i.e. parataxis and hypotaxis) in English (see Halliday &
Matthiessen 2014). A common example in SFL context for typological
distribution of experiential and logical meaning is the construal of time across
languages. While English has been shown to have a serial tense system (e.g.
‗present‘, ‗present in present‘, ‗past‘, ‗past in past‘), thereby, favouring a
31
logical mode, (Halliday 2008: 134-139; Halliday & Matthiessen 2014), many
African languages, such as Dagbani (Niger-Congo, Gur), have been noted for
construing a time-depth tense system (e.g. ‗recent past‘, ‗day-before‘, ‗remote
past‘), which is oriented towards the experiential mode of ideation.
Interpersonal meaning, on the other hand, encodes such variables as
social and interactional roles, identities, relationships, attitude and stance in
discourse. Languages have different resources for enacting meanings
engendered by these variables. These include mood, honorification (as in
Japanese and Korean), gender, status, and negotiation. One interpersonal
semantic system that tends to be universal is SPEECH FUCNTION, the resource
for making propositions and proposals and it is cross-linguistically realised by
the system of MOOD in the lexicogrammar (see Caffarel, Martin & Matthiessen
2004; Teruya et al. 2007; Matthiessen, Teruya & Wu 2008).
Textual metafunction is a function of a special kind. It is the semantic
resource that organises the other two modes, ideational and interpersonal, into
something that can be consumed by the listener/reader, that is, text. This kind
of meaning includes cohesion, coherence, and information focus in discourse.
Grammatical systems that have been identified cross-linguistically as textual
include INFORMATION and THEME (Caffarel, Martin & Matthiessen 2004).
What is being demonstrated here is that while the theory presents these three
modes of meaning to be universal properties of language, the grammatical
categories that realise them emerge from the context of description of
particular languages.
The term modes of meaning that has been used to characterise the
metafunctions show that these meanings are realised differently. Ideational
resources are particulate, interpersonal meanings are prosodic while textual
meanings are periodic in their respective realisations. Ideational resources are
particulate because they are generally realised by segments that are cumulative
and compositional in nature. For example, in the transitivity system of the
English clause, the functions Actor + Process + Goal + (Circumstance)
contribute different meanings to realising the clause as a material clause.
Interpersonal meaning, on the other hand, is prosodic in the sense that no
matter where the locus of realisation is placed in the clause, the meaning is
diffused to characterise the whole. In the realisation of negation in the English
32
clause, for instance, irrespective of where the negative particle is located, the
negative meaning scopes over a whole stretch such as the clause. As Halliday
(2008: 63) and Caffarel, Martin and Matthiessen (2004) demonstrate, this
prosody is easily observable when the negative particle is resonated by non-
assertive forms in the clause (e.g. I don’t have any money for any project).
Textual meanings are realised periodically as a movement of waves of
prominence and troughs of non-prominence in syntagmatic structure. Again,
in English, for instance (as in many other languages, including Dagaare), the
initial position of the clause has been identified to be a topically prominent
position while the end is prominent as newsworthy (cf. Chapter 5 on Dagaare).
Thus, as we move from the beginning of the clause towards the end, thematic
prominence decreases while information prominence increases and the reverse
is the case as we move from final to initial position. According to Halliday
(2009), this distribution of modes of meaning is a semiotic strategy by which
language manages the simultaneous realisation of these three meanings in the
clause. As he notes, speakers can put any spin (i.e. prosody) on any particle as
meaning flows in varying degrees of prominence. In SFL metalanguage, these
different modes of realisation of metafunctional meaning have corresponding
verbal terminologies – ideational meaning is construed (i.e. constructed into
meaning), interpersonal meaning is enacted (i.e. reproduced as meaning) and
textual meaning is engendered (i.e. created).
2.4.4 Rank
The resources of language in each stratum (see Section 2.4.2), particularly

lexicogrammar and phonology, are organised into a hierarchy of units called
the rank scale. In the lexicogrammar of English, these units consist of the
clause, group/phrase, word, and morpheme (see e.g. Halliday & Matthiessen
2014, see also Chapter 3 on Dagaare). Generally, typological studies have
shown that languages with a distinct rank of morpheme would have four units
on the rank scale and those without this unit would typically have three ranks
(cf. Caffarel, Martin & Matthiessen 2004). These ranks are hierarchical in the
sense that they are compositional; each higher rank is composed of ranks
below it, and, by reversion, lower ranks constitute a higher rank.
33
This constituency relationship among ranks is different from the inter-
stratal relationship that holds between strata in the hierarchy of stratification
(see Section 2.4.2), which we referred to as metaredundancy. In the English
rank scale, for instance, a clause consists of groups and phrases, a group
consists of words and a word consists of morphemes. English distinguishes
between groups and phrases. While a group is composed from below by
words, a phrase is a ‗fragment‘ from above, considered as a reduced clause
(Halliday & Matthiessen 2014). Analogous to the clause, it has a minor-
process (or minor verb), which is the preposition, and this process has its own
participant, Range, or Complement, typically realised by a nominal group.
Group and phrase are, however, of the same rank in their function in the clause
as illustrated by the box diagram for transitivity and mood below:
The girl sowed the seeds in the garden

Actor Process Goal Place
Subject Finite/ Predicator Complement Adjunct
noun group verbal group noun group prepositional phrase
Figure 2.3: Function structure for transitivity and mood
The clustering of grammatical units in language into ranks reflects divisions of

semiotic labour in the lexicogrammar. Each rank is the domain for the
realisation of different set of grammatical meanings (technically,
lexicogrammatical features or, simply, features). General features that are
realised at the domain of the clause in English include polarity (negative and
positive), mood, transitivity, information focus and theme. Features such as
voice and tense are realised by the English verbal group while nominal groups
realise features such as countability, number, specificity, definiteness and
animacy. Thus the semiotic labour of every language is divided among
different ranks. The rank scale thus guides the investigator in locating the
domains of the different grammatical meanings in a language.
The lexicogrammatical ranks of a language are identified from above
by first establishing the basic unit of the language for realising discourse
meaning. Cross-linguistically, this unit has been identified to be the clause.
34
Lower units are then established based on their functions in this discourse unit,
that is, the division of semiotic labour among units in the clause. However,
units would often perform functions that are not typical of their rank. This is
explained by the notion of rank shift, where a unit is embedded in another
unit either of a lower status in the scale or of the same rank as itself. Typical
examples in English are post-modification in the nominal group (examples (1)
and (2)) and complementation in the prepositional phrase (example (3)):
1. The man [in the red shirt]

2. The man [[who came here this morning]]
3. in [the red shirt]
This phenomenon of downgrading is called embedding. Embedded clauses

are conventionally indicated by double square brackets as in example (2)
while embedded phrases (as in example (1)) or groups (as in example (3) are
indicated by single square brackets. Clauses that are embedded as in example
(2) are non-ranking clauses – they have been downgraded in the scale – while
non-embedded clauses are ranking clauses. Embedding is, however,
distinguished from hypotactic relations, where ranking clauses are dependent
on an independent clause. An example of a hypotactic clause is when I went
home in I saw him when I went home. Both embedded and hypotactic clauses
are, however, bound clauses.
2.4.5 Axis
While the dimension of rank divides the semiotic labour of the realisation of
grammatical features among units related by composition, the dimension of
axis represents the relationship between grammatical features and the units
that realise them as chain and choice. In other words, axis maps the selections
of features to the realisation of these selections as structure. Feature selection
and its structural realisation represent paradigmatic and syntagmatic axes of
language respectively. It originates from Saussure‘s structuralism and became
the dimension that defined Firth‘s system-structure theory. Halliday developed
Firth‘s notion of the relationship between system and structure further into his
scale and category theory (e.g. Halliday 1961). At this stage, however, it was
35
not indicated which of the two axes depended upon the other. In the present
state of the theory, priority is given to paradigmatic organisation over the
syntagmatic axis. This simply means that in the organisation of language,
paradigmatic features come first and structure is then derived from these
features by realisation. In other words, grammatical structure is the product of
lexicogrammatical meaning.
This relationship between meaning and structure is normally
represented by system networks as illustrated in Figure 2.4 for the system of
MOOD in English (see e.g. Halliday & Matthiessen 2014 for detail discussion).
A system network is read from left to right, which means that one views it in
order of increasing delicacy. The English system of MOOD has the major
clause as its point of origin and technically, this rank is said to be the entry
condition for the system network. Every unit of the rank scale, as has been
indicated in Section 2.4.4, is a point of origin for particular systems. A system
label is conventionally written in small upper case letters (e.g. MOOD) to
distinguish it from a feature label (e.g. [mood]) and a structural function label
(e.g. Mood). While ‗system‘ is the name of a set of related features with a
common point of origin, a feature is simply a name for a particular
grammatical meaning that has been identified in language. In our system
network, the labels such as [indicative] and [imperative] are all grammatical
features. Outside the system network, features are typically placed in square
brackets as has been done here (cf. Henrici 1981) or single quotation marks
(cf. Halliday & Matthiessen 2014).
36
Figure 2.4: The system of MOODof the English clause
As the network extends in delicacy, new features are identified and

each feature may serve as an entry condition for other features, and, for that
matter, other systems. This is indicated in the network with a straight arrow.
There are four systems in the network in Figure 2.4. Rather than using the
system network, we can express this chain relationship among features with a
selection expression such as [indicative: declarative: exclamative], where the
colon (:) sign means that the feature that precedes it is pre-selected as a
condition for the following feature. It is this horizontal relationship of features
in a system network that is referred as delicacy.
However, grammatical features are also related vertically. Features that
are vertically related are terms of a system and together they form a feature
paradigm that offers a choice to speakers to construe, enact or create
discourse meanings. Each paradigm represents a point of choice in the system
network (cf. Matthiessen & Halliday 2009; Martin 1987; 2013). Grammatical
features are first and foremost defined from above in the semantic stratum,
37
based on the discourse functions of the lexicogrammatical resources that
realise them. They form the deep grammar of language (see Halliday 1966;
Huddleston 1981; Henrici 1981; Fawcett 1981). In contradistinction with
transformational-generative theories, SFL does not recognise universal deep
structures in the grammar of languages from which surface structures are
derived through transformational rules. It rather shows that there is only one
structure (i.e. surface structure) that is derived from discourse oriented
lexicogrammatical features.
In Figure 2.4, the structural realisations of features are indicated by
slanted arrows between the feature and the structure that realises it (i.e.
realisation statement). Structure in SFL is a structure of functions (also
called elements of structure) rather than of form. Figure 2.4 shows that the
feature [indicative] is realised by a structure comprising the functions of
Subject and Finite as its minimal number of constituents while [imperative]
has the function Predicator as its minimal realisation. Grammatical functions
are written with an initial upper case letter. The realisation statement ‗Subject ·
Finite‘ does not include a statement of how these two elements are sequenced.
As the network extends in delicacy, however, it is clarified that, for declarative
clauses (see example (4)), the Subject element typically precedes Finite
(conventionally, ‗Subject ^ Finite‘), and, for interrogative clauses (example
(5)), the Finite element typically precedes Subject (Finite ^ Subject):
4. I was given a prize at graduation.

5. What were you given at graduation?
The English Finite element is realised by an auxiliary verb indicating tense or

modality. These are normally modal verbs (We should be going) and the
primary auxiliaries be and have (e.g. John is leaving tomorrow) although tense
may conflate with the Predicator element in a verbal group simplex (e.g. He
cooks late).
As a system network extends in delicacy, realisation statements also
become more delicate and specific. In Figure 2.4, the expression ‗#WH‘
extends the realisation statement for interrogative clauses by indicating that,
for elemental interrogatives, a WH-element occurs in clause initial position
38
(indicated by the clause boundary symbol #), In the most delicate situations,
realisation statements normally include class labels or specific words and
morphemes. Such terminal statements are therefore lexified, and, in the
system network, this may be indicated with an equal (=) sign, interpreted as
lexification. Henrici (1981) refers to them as final realisation statements to
distinguish them from the structural realisation statements, which comprises
functional elements.
We may also distinguish between a structure and a syntagm (cf.
Halliday 1966, 2009). As has been mentioned, a structure is a configuration of
functional elements to realise a systemic feature. A syntagm, on the other
hand, is a configuration of class categories and is important for examining
constituency (e.g. nominal group + verbal group + nominal group). SFL
descriptions focus on structure and syntagm is analysed as delicate resources
that realise structural functions. This approach is illustrated with a box
diagram in Figure 2.5.
The student had made no comments

Subject Finite Predicator Complement
Sayer Process: verbal Verbiage
Theme Rheme
nominal group verbal group nominal group
Deictic Thing Finite Event Deictic Thing
determiner noun verb verb determiner noun
Figure 2.5: A box diagram of syntagm and structural configurations
The first three layers show structural configuration of the clause in terms of
mood, transitivity and theme in that respective order. This exemplifies the
simultaneity or complementarity of interpersonal, ideational and textual
meanings discussed in Section 2.4.4. The fourth layer shows the grammatical
classes that realise the structural functions at group rank. Thus,
Subject/Sayer/Theme are realised by a nominal group. Finite + Predicator and
the corresponding transitivity function Process are realised by a verbal group
while Complement and the corresponding transitivity function Verbiage are
also realised by a nominal group. The fifth and six layers analyse the
functional structure of groups and the realisation at word rank respectively.
39
For example, the nominal group The student realising the
Subject/Sayer/Theme consists of the elements Deitic + Thing with the Deitic
realised by a determiner and the Thing realised by a noun. As the analysis
shows, English, as well as other Germanic languages, has a fairly exotic
interpersonal structure of the clause by being able to split the verbal group into
two elements, Finite and Predicator. The Finite combines with the Subject as
the most interactive element in the clause, the Mood. For instance, speakers
pick them up in arguing about the polarity values (e.g. He had; He hadn‟t) and
they are copied in tagged interrogative clauses: The student had made no
comments, had he?
2.4.6 Semogenesis
In Section 2.3, it was stated that language creates meaning, and, in Section
2.4.1, we examined this property as an instantiation process where meanings
are selected from the overall system of language to create text. Another
dimension that illustrates this semogenic property of language is genesis.
Genesis simply means history. It explains three histories in language, namely,
phylogenesis, ontogenesis and logogenesis (e.g. Halliday & Matthiessen
1999). Phylogenesis is the evolution of language in the human species and is
thus primarily the subject matter of historical linguistics, grammaticalisation
and evolutionary linguistics. Ontogenesis is the development of language in
the individual meaner (or person) either as a first language or a second
language. It is the object of study in language learning disciplines.
Logogenesis is the moment by moment unfolding of meaning in text or
discourse and has been the focus of conversation and discourse (including
genre) studies.
These three histories are valuable in explaining grammaticalisation, a
key area in typology studies. Grammaticalisation can be viewed as the
evolution of grammar (i.e. phylogenetically), the sense in which the term was
originally used by Meillet (see Hopper 1991; Hopper & Traugott 2003). It can
also be viewed as a gradual emergence of grammar in a child learning his/her
mother tongue (see e.g. Halliday 1973, Painter 2009). We may also examine it
as a logogenetic process observable synchronically in the unfolding of
40
discourse (cf. Traugott & Heine 1991; Halliday & Matthiessen 2014;
Matthiessen & Kashyap 2014). As an example, we may consider the following
clauses adapted from Halliday and Matthiessen (2014) in which negative
meaning has been lexicalised (example (1)), semi-grammaticalised (example
(2)) and grammaticalised (example (3)):
1. He failed.
2. He never succeeded.
3. He didn‘t succeed.
These different dimensions of grammaticalisation make it better to describe it

rather than define it. The common definition is, however, given as the changes
in language by which lexical items develop into grammatical forms or by
which less grammatical forms become more grammatical (Hopper &Traugott
2003). In other words, grammaticalisation is primarily defined from above in
the semantics stratum – it is meaning that gets grammaticalised by structure or
form (cf. Halliday 2008: Ch. 5; Halliday & Matthiessen 2014). It is therefore a
realisation process by which discourse meaning is assigned a grammatical
feature in the lexicogrammar, and selects a grammatical rank as its domain.
Grammaticalisation, in its diachronic sense, represents an interstratal
tension between semantics and lexicogrammar. As the meaning potential of
the speech fellowship expands, lexicogrammatical resources are stretched to
meet various semiotic demands. Grammaticalisation is a diachronic-
synchronic complementarity. From a synchronic perspective, it is a property of
language that some meanings are grammaticalised and others lexicalised
(Halliday 2008: Ch. 5). The present study focuses on those meanings that are
grammaticalised. There, however, is a fuzzy boundary between the two as
suggested in our discussion of the complementarity between grammar and
lexis (Section 2.4.2). Halliday and Matthiessen (2014) identify three properties
of grammaticalised meaning: (i) it is organised in the language as a closed
system of mutually exclusive terms, (ii) it is associated with a general category
in the language and (iii) it displays proportionality throughout the language.
As a diachronic process, grammaticalisation can as well be
characterised from a trinocular perspective. From above, in the semantics
41
(including pragmatic meaning), it is viewed as a process by which regular
meanings in discourse come to be generalised and specialised by a category in
the lexicogrammar. From a roundabout perspective, on the cline of
lexicogrammar, this means that certain lexical items gradually lose their
specificity and content to become general grammatical forms or less
grammatical forms become more grammatical (Hopper & Traugott 2003). On
the other hand, it may just be a change in word order (cf. Li & Thompson,
1976). From below, in terms of rank (taking the clause as the point of view),
there may be a corresponding reduction of word forms, fusion of items and
other morphological processes (cf. Heine, Claudi & Hünnemeyer 1991;
Bybee, Perkins & Pagliuca 1994; Hopper & Traugott 2003). Below
lexicogrammar, there may be shifts in tones and intonation and phonetic
reduction of lexicogrammatical forms (cf. Heine, Claudi & Hünnemeyer 1991;
Hopper & Traugott 2003). Diachronic and synchronic aspects of
grammaticalisation represent two viewpoints by the analyst on one and the
same phenomenon. It is the meanings and the corresponding forms (and, by
extension, sounds) that get grammaticalised through time that come to be the
focus of synchronic descriptions of grammar.
2.5 Methodological Procedures in the Present Study

The preceding sections discussed the theoretical landscape of the study. This
section will proceed to describe the methods and procedures used in executing
the objectives of the study. The study adopts an ethnographic tradition to the
study of language as its general research design. This approach involves the
study of language in its social context of use. The enthnographic approach to
linguistics has a long history. In modern linguistics, it develops from two
related research traditions. One is American anthropological linguistics,
starting with the work of Franz Boas through Edward Sapir, Benjamin Whorf
and later scholars such as Mary Haas, Ken Pike and Dell Hymes to post-
Greenbergian functionalism. The other is the European functionalism that
developed from the work of Bronislaw Malinowski and extended by J. R.
Firth, M. A. K. Halliday, J. R. Palmer, just to mention a few.The main tenets
of this approach are the emphasis on the study of language in social context,
42
the emphasis on meaning, and the emphasis on observation of language
behaviour and field work. The sections below outline how this ethnographic
approach has been used in collecting and analysing data for the study.
2.5.1 Data Source

As was mentioned in Chapter 1, the study focuses on the Lobr dialect of
Dagaare. The data for the study comprises naturally occurring written and
spoken texts produced by speakers of this dialect in Ghana and Burkina Faso.
The written data consist of religious registers, specifically, Bible translations
and liturgical materials. This registerial limitation of the data is due to the fact
they were the only written material I could access during the field work.
Textsbooks, short stories and also religious materials on the Central Dagaare
(‗Ngmere‘) dialect were available and photocopied as a comparative guide in
the data analysis. On the other hand, the spoken data range across a number of
registers and text-types. These include an unscripted play, folk tales, live
conservations in varying contexts, radio interviews, announcements and panel
discussions, workshop reports and meetings, monologic speeches, and a
religious film. This variable sources and nature of data is very important for
the general and relatively comprehensive description of the language in this
study. The focus of the study is not on the meanings of linguistic resources
instantiated in the individual texts. The aim is rather to move beyond these
instances to identity systems and categories that are general and proportional
in the language. Table 2.1 locates the texts used for the study across registerial
variables of field (i.e. secio-semiotic process) and mode (see Matthiessen
(2015) and Matthiessen & Teruya (2015) on socio-semiotic processes).
2.5.2 Analystical Procedure

The specific analytical procedures adopted in conducting the study can be
summarised into the following six criteria, the first five of which are identified
by Matthiesen (2015) (also see Haspelmath (2009a) on theory and typological
guidance):
43
Table 2.1. Text archive for the study
Socio-semiotic Written Spoken

processes
dialogue monologue dialogue monologue
doing A tribute installation of
for a queen mother
University
VC
enabling school
textbook
recommending concert
advertisement
Exploring arguments in
conversations
Expounding
reporting biblical narratives in farmers‘
narratives; conversations reports at agric
and meetings; workshop;
radio
announcements
recreating short stories unscripted folktales
play; religious
film
sharing conversations; speeches
radio panel
discussions;
radio
interviews;
meetings;
Facebook
posts &
comments
44
1. Analysis of discourse data
2. theoretical guidance
3. typological guidance
4. transfer comparison
5. language consultants
6. dialect comparison
These criteria are time tested criteria employed by many lingusits in the
description of languages even where this is not explicitly stated. Their
combination in the present study leads to an in-depth study of Dagaare.
First, the analysis of the discourse data took several cyclical steps,
starting with the transcription of the spoken texts. Initially, some of the spoken
texts were selected and transcribed in whole for a complete discourse analysis
in order to identify lexicogrammatical features inherent in the language and
the forms that realise them. Others were transcribed in parts for testing
emerging lexicogrammatical features and forms in their analysis. Initial
analysis of the data was exploratory in nature. Both written and transcribed
texts were read through repeatedly to identify grammaticalised meanings and
forms that realise these meanings. At this stage, the analysis was more
haphazard rather than systematic. Almost everything on the semiotic map of
language outlined above was kept in focus: moving across strata, ranks, and
classes, between observing text instances and making systemic generalisations
as well as maintaining a view on all metafunctions (experiential, logical,
textual and interpersonal). In every phase of this process, new grammatical
meanings and/or forms were identified, some forms identified previously
reinterpreted and some other interpretations discarded. As the analysis
progressed and meanings and forms became clearer, particular systems were
identified and analysed systematically. System networks and structural
paradigms were used as key analytical tools. System networks, in particular,
helped in identifying ranks that serve as the domain of particular grammatical
features and the relationship between features, either as contrastive terms of
the same system, as simultaneous systems, or related in delicacy. System
45
networks were further tested on new data and by constructing paradigms to
assess the acceptability of clauses that the network generates. In the process,
the system networks were revised many times, and others were completely
discarded and their features redistributed among other systems. Selected
spoken discourse was also listened to repeatedly to identify phonological and
phonetic realisations of some grammatical features and forms. The final stage
of the analysis was more systematic. First, texts were selected and chunked
into clauses. Second, the chunked clauses were input in an Excel spreadsheet
and analysed closely using the categories of the languages that have been
attested in the previous analysis (see Section 3 in Appendix). This led to a
more rigorous testing of the categories and revisions were made where
necessary. This analysis procedure has been displayed in the Appendix.
This analytical process was guided by theory, typological
generalisations and transfer comparison. As indicated in discussions in the
preceding sections, theoretical guidance was provided by the general
dimensions of language, particularly as they are articulated within systemic
functional linguistics. Typological guidance was provided by three sources.
One is the typoloical generalisations that have been developed based on the
systemic functional theory (e.g. Matthiessen 2004; Teruya et al. 2007;
Matthiessen, Teruya & Wu 2008; Wang & Xu 2013; Teruya & Matthiessen
2015). The second source is the wider typological generalisations that have
been developed by other functional typologists (e.g. Hopper & Thompson
1980; Heine, Claudi & Hünnemeyer 1991; Bybee, Perkins & Pagliuca 1994;
Lambrecht 1994; Heine & Kuteva 2002a, b, 2007;Aikhenvald & Dixon 2003,
2006; Sphopen 2007a, b, c; Dixon 2010a, b, 2012; Dryer & Haspelmath 2013;
Haspelmath 2015). The third consists of typological studies particularly on
African languages (e.g. Welmers 1973; Heine & Reh 1983; Heine & Nurse
2000; Heine 2011; Williamson & Blench 2000; Güldemann et al . 2015). The
typological generalisations in these studies served as a guide in identifying and
interpreting linguistic forms and structures in Dagaare in terms of both
synchronic categories and diachronic pathways of various forms. Work on
grammaticalisation was particularly useful in resolving odd and ambiguous
categories. However, only passing and footnote references are made to
diachronic processes for the sake of coherence in presenting the synchronic
46
information. In addition to typological guidance, the linguistic categories
identified in previous descriptions of other languages were used as a
descriptive framework for Dagaare (e.g. Halliday & Matthiessen 2014; Teruya
et al. 2007; Caffarel, Martin & Matthiessen 2004; Abdel-Hafiz 2015). It must
also be added that previous descriptions of Dagaare were consulted and
incorporated in the description, where necessary (especially, Bodomo (1997)
and J. D. Somé (2004) on the account on phonology in Chapter 3). Bodomo‘s
(1997) account of Central Dagaare was particular used for transfer comparison
with categories in the Lobr dialect decribed in the present study.
Finally, dialect comparison was done through analysis of texts from
other dialects and through the assistance of language consultants, consisting of
native speakers of various dialects of Dagaare. Dialect comparison helped in
identifying the general principles in the language that underlie the forms and
structures identified in the analysis. This knowledge clarified some conflicting
interpretations of categories. A discussion on dialectal variation in Dagaare is,
however, beyond the purpose of the present study. Again, where necessary,
only footnote and passing references will be made to other dialects during the
discussion in subsequent chapters.
The final consideration was the presentation of illustrative examples.
Examples are given in three layers, comprising (1) original example clauses or
texts in Dagaare, (2) morpheme-by-morpheme interlinear glossing and (3) an
English translation. Leipzig Glossing Rules are employed in the morpheme-
by-morpheme interlinear glossing albeit with a few additions and
modifications to reflect some language specific categories of Dagaare
(available at: http://www.eva.mpg.de/lingua/resources/glossing-rules.php).
Chapter 3 gives a comprehensive overview of the language and is partly
intended to introduce the basic categories of the language to readers before the
more extensive discussion in subsequent chapters. Also, a key to abbreviations
in the glosses has been given in the beginning of this thesis and readers can
refer to it where necessary. The English tranlations are presented in idiomatic
English. However, effort is made to make the translations reflect the
grammatical structure of the Dagaare examples as much as possible. In some
instances, this may lead to translations that are not typical of everyday English.
Where it is not possible to give an acceptable English translation that is close
47
to the Dagaare example, a more plausible English translation is provided in
parenthesis.
2.5.3 Limitations of the Study

Language is a complex system and can neither be fully described in a single
study nor by one individual in his life time (Akerejola 2005). In this light, the
present study is limited in several aspects. In terms of stratification, the study
focuses on lexicogrammar. Particularly, it is skewed towards the grammar
pole of the lexis-grammar continuum and explores lexis only in so far as
lexical items illuminate grammatical meaning. The relationship between lexis
and grammar will be particularly salient in the discussion on process types.
However, as has been mentioned, the study highlights the interaction between
grammar and semantics and between grammar and phonology in terms of
realisation.
Regarding rank, the primary focus of the study is mainly limited to
clause systems and lower ranks will be discussed in the context of their
realisation of particular functions at clause rank level. The study however
covers all three metafunctions. It will examine clause rank systems that realize
interpersonal, textual and ideational meanings. In Chapter 4, however, two
verbal group systems, namely POLARITY and MODALITY are discussed in some
detail due to their close interaction with the system of MOOD, a clause rank
3
system. Resources of ranks lower than the clause are also briefly discussed in
Chapter 3. It should also be noted that quantitative statements of systemic
probabilities and register variation are beyond the scope of the study, although
quantitative counts are occasionally given to support interpreations of
systemic contrasts.
As mentioned earlier, however, the study will include a general sketch
of the lexicogrammatical and phonological resources of the language before
proceeding to present a detail exploration of the clause. The analysis will
include ranks below the clause, such as morphemes and word classes. This
linguistic profile of the language is necessary to place the discussion within
3
Note that the characterisation of POLARITY and MODALITY as group rank systems and MOOD
as a clause rank system is relative to dagaare although this pattern is common in the world‘s
languages. In English, POLARITY and MODALITY are clause rank systems, realised through the
Finite element or the mood Adjunct (cf. Halliday & Matthiessen 2014: Ch. 4).
48
the general semiotic environment of the language. It is hoped that the study
will provide impetus for investigation of other semiotic regions of the
languageas has been identified at the end of Chapter 7.
2.6 Conclusion
In summary, this chapter discussed the theoretical framework and methods
used in the study. It first examined the nature of linguistic theory and
description and the relationship between them. It also considered the systemic
environment of language and showed how language interacts with physical,
biological and social systems. Further, it described the architecture of
language as it is theorised in systemic functional linguistics, highlighting the
dimensions of instantiation, stratification, metafunction, rank, axis and
genesis. The chapter ended with methodological issues in the study, including
the nature and source of data, the data analysis process and limitations of the
study. The next chapter will proceed togive a profile of the linguistic
organisation of Dagaare, focusing on phonology, orthography and
lexicogrammar.
49
CHAPTER THREE
AN OVERVIEW OF THE DAGAARE LINGUISTIC SYSTEM
3.1 Introduction
As mentioned in Chapter 1, the purpose of this study is to examine the

grammatical systems of the Lobr Dagaare clause. In order to maximise our
understanding of the clause systems, however, it is useful to describe the
linguistic resources that realise them. This first chapter of the account will
therefore give a general overview of the organisation of the language albeit in
a sketchy manner. The chapter will take a cross-sectional view of the
dimension of stratification, focusing on phonology and orthography in the
expression plane and on lexicogrammar in the content plane. It begins with an
account of the various units of Dagaare phonology, highlighting their form and
functions (Section 3.2). It then discusses pertinent issues in the orthography of
the language (Section 3.3) and examines different grammatical units in terms
of their classes, forms and meanings (Section 3.4).
3.2 Phonology
Dagaare phonology is organised around four units, namely tone group,
phonological word, syllable and phoneme. These are presented in the rank
scale in Figure 3.1. Graphically, phonological word will be distinguished from
grammatical word by putting it in square brackets (cf. Matthiessen 1987).
Each of the units in the rank scale is the domain for the realisation of at least
one phonological phenomenon. They are discussed below, beginning with the
lowest and smallest unit, the phoneme.
50
Figure 3.1. The phonological rank scale of Dagaare
3.2.1 Phoneme
The phoneme is the domain for the realisation of phonemic (or distinctive)
features. The phonemic system of Lobr Dagaare is made up of twenty-nine
(29) consonants and thirty-three (33) vowels (cf. J. D. Somé 2004; also cf.
Bodomo (1997) on the Ngmere dialect).4 The consonants consist of eight stops
(/b, d, g, gb, k, kp, p, t/), two affricates (/tʃ, dʒ/), six fricatives (/f, h, ɦ,ʔ, s, v,
z/), and 12 sonorants (/ɓ, l, 'l, m, n, ny, ŋ, ŋm, r, w, 'w, j/). These are presented
in (1) together with a specification of their features in terms of voicing, place
and manner of articulation.
(1) Dagaare consonants
consonant description example

/b/ voiced bilabial stop báa, ‗dog‘
/ɓ/ voiced bilabial sonorant (implosive) ɓaa, ‗be cool‘
/tʃ/ voiceless aveolar-palatal affricate cɩr, ‗to mark‘
/d/ voiced aveolar stop dáa, ‗push‘
/f/ voiceless labio-dental fricative faa, ‗seize‘
/g/ voiced velar stop gàa, ‗dates‘
/gb/ voiced velar-bilabial stop gbaa, ‗mamba snake‘
/h/ voiced glottal fricative háa, ‗open widely‘
/ɦ/ voiced glottal fricative (implosive) hɛr, ‗be annoyed‘
4
My account on phonology, especially the phoneme, and also orthography is based on J. D.
Somé (2004) and Bodomo (1997).
51
/ʔ/ voiced glottal stop ƴɛ́r, ‗grind‘
/dʒ/ voiced aveolar-palatal affricate jɩr, ‗look‘
/k/ voiceless velar stop kãa, ‗cream, oil‘
/kp/ voiceless velar-bilabial stop kpãa, ‗remind‘
/lɛ/ voiced aveolar sonorant (lateral) lɛ, ‗again, that‘
/'lɛ/ voiced aveolar sonorant (implosive) 'lɛ, ‗enjoy‘
/m/ voiced bilabial sonorant (nasal) mãa, ‗I, me‘
/n/ voiced aveolar sonorant (nasal) náw, ‗to mash‘
/ny/ voiced aveolar-palatal sonorant (nasal) nyaw, ‗dig out‘
/ŋ/ voiced velar sonorant (nasal) ŋa, ‗this‘(also na)
/ŋm/ voiced velar-bilabial sonorant (nasal) ŋma, ‗cut into pieces‘
/p/ voiceless bilabial stop paw, ‗close‘
/r/ voiced palato-aveolar sonorant (roll) par, ‗to jump‘
/s/ voiceless palatal fricative saw, ‗agree‘
/t/ voiceless aveolar stop taw, ‗pull‘
/v/ voiced labio-dental fricative vɔb, ‗uprooting (n)‘
/w/ voiced labio-velar sonorant (glide) wɔb, ‗elephant‘
/'w/ voiced bilabial sonorant (glide) 'wɔblɩ, ‗be deformed‘
/j/ voiced palatal sonorant (glide) yɛ́r ‗to spread‘
/z/ voiced palatal fricative zɛ̀li, ‗to mill‘
As the feature specification shows, a few of the consonants (i.e. /gb/, /kp/,
/ny/, and /ŋm/) have double articulatory features in terms of their place of
articulation. That is, they combine the characteristics of two simple consonants
in their place of articulation.
The thirty-three (33) vowels of Dagaare comprises nine (9) short
vowels and nine (9) long vowels, corresponding to the short vowels, as well as
fifteen (15) diphthongs (or glides). Figure 3.2 present the short vowels in a
traditional vowel chart.
52
Figure 3.2. Vowel chart for Dagaare short vowels (J. D. Somé 2004: 31-33)
As the vowel chart shows, the vowels differ in quality based on the vertical
and horizontal position of the tongue in their production. On the vertical axis,
a vowel could be close or open, or mid-close or mid-open. On the horizontal
axis, it could be front or back, relative to the part of the tongue raised in its
production. Since the long vowels only contrast with the short ones in length,
they will not be discussed separately except to indicate that they form minimal
pairs with the short vowels, thereby supporting their phonemic status.
Examples are dà (‗buy‘) versus dáa (‗push‘) and kùr (‗tortoise‘) versus kùur
(‗hoe‘). Dagaare vowels also contrast on lip posture and tongue root
position.The values for lip posture are rounded versus unrounded and the
values for tongue root position are advanced tongue root, [+ATR], and
retracted tongue root, [-ATR]. The characteristics of the vowels are
summarised in a feature specification in (2) below. Each feature is
independent of the other in the sense that the [ATR] value of the vowel, for
instance, does not depend on whether it is [±close], [±front] or [±round].
53
(2) Feature specification for Dagaare monophthongs
i e o u ɪ ɛ ɔ ʊ a
close + + + + + - - + -
front + + - - + + - - +
round - - + + - - + + -
ATR + + + + - - - - -
example cir de do vu cɩr cɛr cɔr vʋ va
meaning pour take climb crawl to mark untie to pin pierce hit
As shown in (3), the diphthongs contrast based on tongue root position

(advanced versus retracted), the direction of the glide in terms of tongue
height (e.g. opening versus closing) and the horizontal position of the tongue
(e.g. fronting versus backward).
(3) Dagaare Diphthongs

vowel description example
/ie/ opening front advanced tongue root pie, ‗ten‘
/io/ opening backward advanced tongue root bio, tomorrow
/ou/ closing back advanced tongue root ców, ‗nest‘
/ue/ opening fronting advanced tongue root kùe, ‗hoes‘
/uo/ opening back advanced tongue root púo, ‗(a) farm‘
/ɪɛ/ opening front retracted tongue root sɩ́ɛ, ‗waist‘
/ɪa/ opening front retracted tongue root dɩ́a, ‗today‘
/ɪɔ/ opening backward retracted tongue root cɩ́ɔ, ‗squirrel‘
/aɪ/ closing front retracted tongue zʋ́kpáɩ, „proverbs, riddles‘
/aʊ/ closing backward retracted tongue root pàw, ‗close‘
/ɔɪ/ closing fronting retracted tongue root pɔ́ɩ, ‗oath, vow‘
/ɔɛ / mid-open fronting retracted tongue root kɔ́ɛ, ‗bags‘
/ɔʊ/ closing back retracted tongue root bɔ̀w, ‗hole‘
/oe/ mid-close fronting retracted tongue root loe, ‗sides, areas‘
/ʊɔ/ opening back retracted tongue root bʋɔ, ‗goat‘
It can be observed that only the first five diphthongs in (3) above are [+ATR]
while the rest are [-ATR] or retracted tongue root vowels. With regards to the
vertical position of the tongue, eight of the diphthongs, /ie, io, ɪɛ, ɪa, ɪɔ, ue, uo,
ʊɔ/, are opening glides and three are closing glides, /ɔɪ/, aʊ, aɪ/. Two of them
do not glide vertically and are simply mid-open /ɔɛ/ or mid-close /oe/. On the
54
horizontal axis, on the other hand, there are four fronting glides (/ue, ɔɪ, ɔɛ,
oe/) and three backward glides (/io, ɪɔ, aʊ/) while six do not glide horizontally
and are either front vowels (/ie, ɪɛ, ɪa, /aɪ/) or back vowels (/uo, ʊɔ/).
3.2.2 Syllable
The syllable is the domain of both articulatory combinations (i.e. phonotactics)

and prosody, specifically, syllabic tone and nasality (see further below). It is
the rank immediately above the phoneme and it is composed of at least one
phoneme. The structure of the Dagaare syllable comprises an obligatory Peak
element and optional Onset and Coda elements. The maximal structure is
illustrated for the word bʋ̀r (‗sow‘) in Figure 3.3.
Figure 3.3. Illustration of syllable structure
While the Onset and Coda are realised by consonants, the Peak is typically
realised by vowels. However, it can be realised by syllabic consonants in
specific grammatical environments. These consonants include the nasals /n/
and /m/ and the liquids /l/ and /r/. Examples of syllabic /n̩/ and /m̩/ are
respectively the enclitic focus particle =n (4) and the enclitic pronoun =m (5)
in the clauses below:
(4) Sɛb-Sow Ƴɛr-bie (1996)

A Abɛl wa ɩ =n pi-cɩɩnɛ.
DEF Abel EVT COP.PFV FOC shepherd
(5) Casual conversation
Pʋr kʋ̀ =m!
pour.PFV give.PFV 1SG.ACC
‗Pour (some) for me!‘
Syllabic /l̩/ occurs in the context of the elision of the vowel ɩ or i in verbs
55
ending in –li or-lɩ such as maalɩ (‗make‘) in (6) below:
(6) Maal kʋ̀ =m!

make give.PFV 1SG.ACC
‗Make (it) for me!‘
Here, the final vowel of maalɩ (‗make‘) is elided due to rapidity in speech,
resulting in a syllable realised by only the lateral consonant. An example of
syllabic/r̩/ is where -r is used as imperfective aspect suffix (e.g. siw, ‗get
down‘; siwr, ‗getting down‘) (see Section 3.4.2.2 on aspect). In addition to
these syllabic consonants, the bilabial stop /b/ also manifests as a syllabic
consonant when it occurs as an enclitic form of the second person singular:
(7) A bie na yèl a kʋ̀ =b.

DEF child POS.IND.NFUT say. PFV AFFR give.PFV 2SG.ACC
‗The child will tell you.‘
As mentioned earlier, regarding the function of the syllable, it is the

domain of the realisation of syllabic tone and nasality (also see Section 3.2.4
on tone group). Two tones have been postulated for Dagaare, high and low, as
well as a downstep high tone (cf. Kennedy 1966; Bodomo 1997; J. D. Somé
2004). Downstep is identified as a phonological situation where, in a given
[word], the second of two high tones is not as high as the first high tone. The
system of tone at the syllable rank is presented below:
Figure 3.4 The system of TONE in the syllable (cf. Matthiessen 1987: 30)
Tone selection is marked on the syllable Peak and it can be meaning

distinguishing (or emic) as in the minimal pairs below (see J. D. Somé (2004)
56
for details on syllabic tone):
(8) Minimal pairs illustrating syllabic tone contrasts

Low tone High tone
dà, ‗push‘ dá, ‗buy‘
sò, ‗to own‘ só‗to wash‘
wɛ̀r, ‗diverge, wonder off‘ wɛ́r, ‗operate upon, cut open‘
gùr, ‗to fold‘ gúr, ‗sleep‘
zɛ̀lɩ, ‗beg‘ zɛ́lɩ, ‗to sieve‘
pàw, ‗to cover, to close‘ páw, ‗to get‘
sɛ̀w, to calm down in a dance‘ sɛ́w, ‗be equal‘
zʋ̀mɛ, ‗intuition‘ zʋ́mɛ, ‗insults‘
dʋ̀rʋ, ‗right hand‘ dʋ́rʋ, ‗urine‘
Like tone, nasality is indicated on the vowel realising the syllable Peak. It can
also be meaning distinguishing as in the minimal pairs below:
(9) Minimal pairs illustrating constrasts in nasality

+nasal -nasal
nyɛ, ‗see‘ nyɛ, ‗defecate‘
ŋmɛ, ‗resemble‘ ŋmɛ, ‗hit, beat‘
mɛ, ‗like/as‘ mɛ, ‗build‘
sãa, ‗father‘ saa, ‗rain‘
3.2.3 Phonological Word

The [word] is composed of at least one syllable. It is the domain for the
realisation of harmony systems. The most common and consist harmony
system in Dagaare, as in many other Niger-Congo languages, is [ATR] vowel
harmony. The general principle is that vowels that make up a [word] must all
either be [+ATR] or [-ATR]. Vowel harmony is therefore a [word] rank
prosody. We illustrate this phenomenon in (10) below:
57
(10) Tongue root vowel harmony
[-ATR] [+ATR]
cɛrɛ, ‗untying‘ cere, ‗going‘
pʋrɛ, ‗pouring‘ pore, ‗naming‘
sɔwlɩ, ‗to hide‘ suoli, ‗to tell a story‘
There is, however, one opaque vowel, /a/, which is [-ATR] by default but also
co-occurs with [+ATR] vowels. In (11), for instance, it occurs in ayi together
with [i], which is [+ATR], and in kpantole together with the [+ATR] vowels
[o] and [e]:
(11) Co-occurrence of /a/ with [+ATR] vowels
[-ATR] [neutral]
ayʋɔb, ‗six‘ ayi, ‗two‘
kpakpamɛ, ‗shoulders‘ kpantole, ‗mounds‘
In instances such as ayi and kpantole the vowel /a/ neutralises the the vowel
harmony principle. Root morphemes with /a/ as the only vowel, however,
trigger [-ATR] vowel harmony with suffixes. An example is the root verb ta
(‗reach‘) and its imperfective form tarɛ (‗reaching‘), where the vowel in the
root conditions the choice of the [-ATR] vowel /ɛ/ in the suffix (see Section
3.4.2.2 for details).
Vowel harmony is a guiding criterion in distinguishing phonological
word boundaries from grammatical word boundaries in two ways. First, it
shows that one grammatical word can correspond to two phonological words.
Typical instances of this are found in noun compounds, as in (12) below:
(12) One grammatical word versus two phonological words
Pi-cɩɩnɛ, ‗shepherd‘ (‗lit., ‗sheep-care taker‘)

Sɛb-Sow, ‗scripture‘ (lit., ‗Writing-Holy‘)
Ƴɛr-bie, ‗words‘ (lit., ‗speech-seeds‘)
In words such as these, the first morpheme in the compound is normally a

clipped form. For example, pi-, sɛb- and ƴɛr- are clipped versions of the nouns
piir (sheep), sɛbʋ (‗writing‘)and ƴɛrʋ (‗speech‘). Although each compound is
58
treated grammatically as one word, the two components are phonologically
distinct [words]. This is signalled by the fact that each component in the
compound has a different [ATR] value, with the clipped forms maintaining the
vowel quality of the original nouns. In contrast, when the diminutive suffix, –
le, is added to a root noun it changes the [ATR] quality of the stem to [+ATR],
resulting in one phonological word. An example is kʋɔr (‗gourd‘) versus kole
(‗small gourd / bottle, cup‘) (see Section 3.4.1 on morpheme).
On the other hand, two grammatical words can correspond to one
phonological word. An instance of this is where assimilation spans across
grammatical word boundaries. The assimilated words are often single vowel
pronouns, enclitic forms and single word particles. For instance, the third
person pronoun ʋ, which is originally [-ATR], is normally realized as [+ATR]
vowel [u] when it follows a [+ATR] verb in a clause. This means that
phonologically, it is treated as part of the verb that precedes it. A more
interesting example is that an originally [-ATR] verb, nyɛ (‗see‘), is often
realised as [+ATR] when it takes the pronoun ʋ as Complement and
prosodically conditions the pronoun into [+ATR] [u]. I illustrate this
phenonomenon with a phonetic transcription of the clause Nɩbɛ bɛ nyɛ ʋ ɛ
(‗People have not see him/her‘), taken from a spoken advertisement text, in
(13) below and compares this with a reconstruction of an alternative
realisation in (14):
(13) Concert advertisement

[Nɪbɛ bɛ nyẽ u e]
people NEG.IND.NFUT see.PFV 3SG NAFFR
(14) Reconstructed example

[Nɪbɛ bɛ nyɛ̃ ʊ ɛ]
If we compare the original spoken form in (13) to it‘s alternative realisation in

(14), it becomes clear that the last three grammatical words in the clause (nyɛ
ʋ ɛ / nyẽ u e ), including the non-affirmative particle ɛ or e, constitute one
phonological word. Although both nyɛ and the pronoun ʋ, by default, have [-
ATR] values in isolation and in other environments, they bond together as one
[word] in instances such as (13) and (14). When the pronoun is the Subject of
59
the clause as in Ʋ bɛ nyɛ nɩbɛ ɛ (‗He has not seen people‘), this kind of mutual
phonological conditioning among these words in terms of vowel harmony is
not possible. Other harmony systems in Dagaare will be discussed in Section
3.4.2.2 in relation to aspect.
3.2.4 Tone Group
The tone group is the highest unit in the phonological rank scale and is
realised by at least one [word]. In the unmarked case, it corresponds to the
clause, and it is the domain of the realisation of intonation (see Section 3.4.3
on clause). A systematic analysis of discourse reveals five distinct pitch
movements in the tone group although it is likely this number is not
exhaustive. The various attested tones are presented in Figure 3.2 as a system
network. The pitch movements are graphically represented using notational
conventions developed in systemic phonology (cf. Tench 1992; Halliday &
Greaves 2008).
Figure 3.5 The system of TONE in the tone group (cf. Halliday & Greaves
2008)
It must be noted that the term ‗tone‘ in this context refers to pitch movements
extending over a whole tone group (roughly the grammatical domain of a
ranking clause) as opposed to single syllables (cf. Section 3.2.2). The first two
tones, fall and rise, can be illustrated by the prosodic contrast between
declarative and imperative clauses, as shown in (15) and (16):
60
(15) Declarative clause, fall (tone 1)
// 1 Nyɩ wa */na //
2PL come.PFV AFFR
‗You have come.‘

(16) Imperative, rise (tone 2)
// 2 Nyɩ */wa //
2PL come.PFV
‗You come!‘
Examples (15) and (16) contrast between declarative and imperative

respectively (mood types are discussed in Chapter 4). The tone group in (15) is
realised by a fall on the affirmative particle */na, which is the tonic syllable,
indicated by the star symbol. In (16), on the other hand, the tone group is
realised by a rise on */wa (‗come‘). Tone 3 (level-rising) can be exemplified
by bound (or subordinate) clauses such as relative, nominal and adverbial
clauses (cf. Section 3.4.3). An illustration is given in example (17) with an
adverbial clause.
(17) Bound clause, level-rising (tone 3).

// 3 ʋ na */wa a //
3SG come.PFV
ADVLZ JUNC
‗When he came …‘
In clauses such as (17), the tonic prominence is indicated on the last lexical
item in the tone group and continue to rise on the juncture prosody marker a.
Tone 4 (fall-rising) and tone 5 (rise-falling) can be illustrated by
different phonological realisation of the word ʋ̃ʋ (‗yes‘), as in (18) and (19)
below:
(18) fall-rising
*// 4 Ʋ̃ ʋ//
yes?
(19) rise-falling
*// 5 Ʋ̃ ʋ //
yes.
Example (18)is a typical response to a call or a query for the listener to repeat
a proposition while (19) is a confirmation of a proposition addressed to the
61
speaker. It can be noted that these distinct meanings are carried out by only the
pitch movement.
3.3 Orthography
This section proceeds to examine Dagaare orthography, an alternative mode to
phonology in the expression of meaning in the language. Dagaare orthography
has been described as a developing one although it is already advanced in its
development (cf. J. D. Somé 2004). In its earliest form, it was developed by
European missionaries around the 1950‘s and 1960‘s and has been revised in
several stages since the early 1970‘s by native scholars and church fathers.
Today, there are two orthographies for the language due to the arbitrary
division of Africa by colonialists, which has split the Dagara among three
countries (Tables 3.4 and 3.5) One of the current writing systems was
developed in Burkina Faso by the Sous-Commission Nationale du Dagara
mainly within the context of biblical and liturgical translation (see J. D. Somé
2004 for an overview). The other was developed by the Ghana Alphabet
Committee as a general writing system for all indigenous languages of Ghana
(cf. Bodomo 1997). However, since both systems are based on the Latin script,
there are only few differences between them. Literacy materials in the Lobr
dialect are developed in the Burkina Faso orthography and it is this system
that has been used in the present study. Other dialects written in this
orthography are Birifor (both Northern and Southern) and Wiile. The Ngmere
and Waali dialects are written in the orthography of the Ghana Alphabet
Committee.
As shown in Table 3.4, Dagaare orthography, particularly the Burkina
Faso alphabet system is highly phonemic and there is almost a one to one
correspondence between phonemes and their orthographic representation. As
indicated in Section 3.2, there are some consonants in Dagaare with double
articulatory features. These are represented in the orthography with a
combination of letters as shown in Table 3.5. Similarly, long vowels are
indicated in the orthography by doubling the vowel (e.g. pìir, ‗sheep‘; saa,
‗rain‘) while diphthongs are always represented by the two vowels reflecting
their quality (e.g. kʋɔ, ‗water‘).
62
Table 3.1. The Dagaare alphabet systems
letter Burkina Faso Ghana corresponding example
no. [Lobr] phoneme(s) [Lobr]
upper lower upper lower
case case Case Case
1 A a A a /a/
2 B b B b /b/
3 Ɓ ɓ MH mh /ɓ/
4 C c KY ky /tʃ/
5 D d D d /d/
6 E e E e /e/
7 Ɛ ɛ Ɛ ɛ /ɛ/
8 F f F f /f/
9 G g G g /g/
10 H h H h /h/,
11 'H 'h - - /ɦ/
12 I i I i /i/
13 Ɩ ɩ E e /ɪ/
14 J j GY gy /dʒ/
15 K K K k /k/
16 L l L l /l/
17 'L 'l - - /'l/
18 M m M m /m/
19 N n N n /n/
20 Ŋ ŋ NG ng /ŋ/
21 O o O o /o/
22 Ɔ ɔ Ɔ ɔ /ɔ/
23 P p P p /p/
24 R r R r /r/
25 S s S s /s/
26 T t T t /t/
27 U u U u /u/
28 Ʋ ʋ O o /ʊ/
29 V v V v /v/
30 W w W w /w/
31 'W 'w - - /'w/
32 Y y Y y /y/
33 Ƴ ƴ - - /ʔ/
34 Z z Z z /z/
63
Table 3.2. Orthographic representation of double articulatory sounds
letter Letter corresponding example meaning
no. combination phoneme [Lobr]
upper lower
case case
1 GB gb /gb/
2 KP kp /kp/
3 NY ny /ny/
4 ŊM ŋm /ŋm/
Another consideration in the orthography is the representation of lexical tones

and nasality (see Section 3.2). Only those tones and nasality that are meaning
distinguishing are marked (e.g. pìir, ‗sheep‘ versus píir, ‗hill‘ and nyɛ, ‗see‘
versus nyɛ, ‗defecate‘) (see J. D. Somé (2004) for details on the orthography
used in this study).
3.4 Lexicogrammar
Dagaare lexicogrammar, as with the phonology, is organised around four
grammatical units, namely clause, group, word and morpheme. This four-unit
organisation of lexicogrammar is the most common across languages (cf.
Caffarel, Martin & Matthiessen 2004). The clause is the highest unit for
realising grammatical meaning. The other grammatical units function within
the clause and each unit functions within the rank above it. The relationship
between the units is represented in the rank scale in Figure 3.2. Each unit
internally comprises a number of classes based on their forms and functions.
Figure 3.6. The Dagaare rank scale
64
This section discusses each of the grammatical units, beginning with the
morpheme.
3.4.1 Morphemes
Many words in Dagaare consist of only one morpheme. The morphemes can
be classified into free and bound, with the latter consisting of a limited class of
affixes and mainly of function words such as particles, determiners,
conjunctions and pronominal clitics. These function words are grammatical
items, that is, items specified within the grammatical zone of lexicogrammar.
The discussion here will be limited to affixes (see Section 3.4.2.4 on particles).
Both derivational affixes and inflectional affixes are found in Dagaare (see
also Section 3.4.2.1 on nouns and Section 3.4.2.2 on aspectual markers). The
derivational affixes consist of the locative suffix and the diminutive suffix (cf.
J. D. Somé 2004). The locative morpheme is a class changing suffix that is
added to nouns to derive a locative adverb. It is realised variously as -ɩ, -mɩ, i,
and mi, depending on its [ATR] environment. It is illustrated in (20) below:
(20) The locative suffix
noun locative adverb

gbɛr, ‗leg‘ gbɛrɩ, ‗on/by the leg‘
kʋɔ, ‗water‘ kʋɔmɩ, ‗in water‘
pʋɔ, ‗stomach‘ pʋɔmɩ, ‗inside‘
salom, ‗sky‘ salomi, ‘in the sky‘
vuu, ‗fire‘ vuumi, ‗in fire‘
zukur, ‗back‘ zukuri, ‗at the back‘
The locative suffix is realised as -ɩ or -mɩ in words with [-ATR] vowels and as
-i or -mi in words with [+ATR] vowels. The choice of -mɩ or –mi, on the one
hand, and -ɩ or –i, on the other hand, respectively depends on whether the root
morpheme ends with an open syllable or a closed syllable. As an alternative to
the use of the locative morpheme, adpositions can be used to represent
location as pʋɔ in kʋɔ pʋɔ (‗inside water‘) and vuu pʋɔ, and ƴaw in a gbɛr ƴaw
(‗by the leg‘).
65
The diminutive suffix is a class maintaining suffix and it simply
indicates that the derived noun represents a smaller counterpart of the entity
represented by the root morpheme (21). It is grammaticalised from the noun
bile (‗offspring‘) and can be realized as –bile or the phonologically reduced
form –le in its singular form and –bili or -li in the plural form (cf. Somé 2004).
Examples are given below:
(21) Diminutive morpheme

Noun diminutive form
singular plural
pɔl, ‗gentleman‘ pobile, ‗young man‘ pobili
dàa, ‗wood‘ dale, ‗stick‘ dali
dɛb, ‗man‘ deble, ‗boy‘ debli
naab, ‗cow‘ nale, ‗calf‘ nali
nɛn, ‗meat‘ nenle, ‗small meat‘ nenli
pɔw, ‗woman‘ powle, ‗girl‘ powli
sʋɔ, ‗matchet‘ sole, ‗knife‘ soli
waab, ‗snake‘ wale, ‗little snake‘ wali
As mentioned earlier, it is the root word which changes its form to be in

harmony with the [+ATR] quality of the suffix –le. Also, compared to the
locative morpheme, the diminutive suffix is more productive and can occur
with a wide range of nouns. It is used with proper nouns for instance to
distinguish between a younger person and an older one with the same name, as
in Kwasile (small Kwasi) or even Mamale (‗small mum‘).
The inflectional affixes identified in this study comprises plural
suffixes in nouns, namely -bɛ ~ -be, -rɩ ~ -ri and-we, with -bɛ and -rɩ variable
due to vowel harmony (cf. Section 3.4.2.1 on nouns) and verbal suffixes for
marking imperfective aspect. The imperfective morpheme varies considerably
due to vowel and consonant harmony (see Section 3.4.2.2 on verbs). The
realisation forms identified are -rɛ ~ -re ~ -nɛ ~ -ne ~ -r ~ -n.
Plural marking in Dagaare nouns, as in many other Niger-Congo
languages, partially display a noun classs system, where the choice of suffix
depends on the semantic class of the root noun (cf. Welmers 1973: Ch. 6 & 7;
66
Creissels 2000: 241-243, 247-247; Dimmendaal 2000: 189-191; Williamson &
Blench 2000). The suffix -bɛ (or -be, for [+ATR] roots) is derived from the
third person pronoun bɛ (plural, human) and is generally used for human
nouns as in nɩr (‗person‘) versus nɩbɛ (‗people‘), dɛb (‗man‘) versus dɛɛbɛ
(‗men‘) and potuure (‗disciple‘) versus potuube (‗disciples‘). The suffix forms
-rɩ ~ -ri ~ -li, on the other hand, are used to mark the plural of loaned nouns
such as soja (‗soldier‘) versus sojarɩ (‗soldiers‘); manɩja (‗manager‘) verus
manɩjarɩ (‗managers‘) and tebul (‗table‘) versus tebuli (‗tables‘) Finally,
relational body parts nouns take –we as a plural marker (e.g. niwn, ‗face‘
versus niwe, ‗faces‘). It should be noted that these suffixes are not the only
ways in which plurality is marked in Dagaare nouns. Many plural nouns are
irregular and will be illustrated in Section 3.4.2.1.
Table 3.3. Illustration of different realisations of the imperfective suffix

realisation of suffix example
perfective imperfective
-rɛ kʋ́, kill kʋ́rɛ
bʋ̀r, ‗sow‘ bʋ̀rɛ
-re wo, ‗to confess crime‘ wore
per, ‗struggle‘ pere
-nɛ bãw, ‗know‘ bãwnɛ
ƴãw, ‗put, to respect‘ ƴãwnɛ
-ne wõ, ‗hear‘ wõne
bĩn, ‗put‘ bĩne
-r dɔw, ‗procreate‘ dɔwr
lɛb, ‗return‘ lɛbr
-n sɛ, ‗roast‘ sɩɛn
On the other hand, the form of the imperfective suffix, as mentioned earlier, is
conditioned by harmony systems and is discussed in detail in Section 3.4.2.1
on verbs. This section will simply illustrate the various forms (see Table 3.3).
As with plural marking in the nouns, not all verbs in Dagaare indicate their
imperfective forms with the imperfective suffix but rather by vowel
67
substitution. An example is bʋɔlɩ (‗to call‘) versus bʋɔlɛ (‗calling‘), where the
difference in the perfective and imperfective is indicated by the contrast in the
final vowel (see Section 3.4.2.2 for details).
3.4.2 Word Classes

The word class system can be divided into three major classes, namely nouns,
verbs and adverbs, and six minor classes: pronouns, adpositions, determiners,
conjunctions, particles and interjections. The major word classes are open
systems that continue to expand their membership in a more flexible manner
to meet the ever expanding vocabulary needs of the Dagara society. Minor
word classes are closed systems that typically realise grammatical meanings
and are more gradual in their evolution. The discussion in this section will
focus on only five word classes, namely nouns, verbs, adverbs, pronouns and
particles.
3.4.2.1 Nouns
Morphologically, Dagaare nouns divide into count and non-count nouns based
on their realisation of the system of NUMBER (22). Count nouns show
morphological distinctions between singular and plural nouns. As (22) shows,
singular is realised by zero marking while plural is realised by variant
morphological forms. These examples are illustrative and not exhaustive of
the various forms the plural can take.
(22) Illustrating Dagaare NUMBERsystem

non-count count
singular plural
dãa, ‗beer/pito‘ bie, ‗child‘ bibiir
damnʋ, ‗trouble, commotion‘ gbɛ́r,‗leg‘ gbɛɛ
ɓaarʋ, ‗cold, moisture‘ laa, ‗bowl‘ labɛ
ci, ‗guinea corn‘ libir, ‗money‘ libie
kamaan, ‗maize nyawr, ‗intestine‘ nyawɛ
nɔmʋ, ‗love‘ nyʋɔr, ‗nose‘ nyɛɛ
pɔlʋ, ‗beauty‘ ɔw, ‗mouse‘ ɔn
68
saalʋ, ‗okra‘ pɛrʋ, ‗sheep‘ píir
suolu, ‗story/stories‘ san, ‗debt‘ samɛ
vɩɛlʋ, ‗goodness, glory‘ sɔw, ‗rabbit‘ sɔn
ya, ‗wisdom‘ yir, ‗house‘ yíe
ƴawr, ‗cold‘ yùor, „pot‘ yìe
zɩ́ɛ, ‗millet‘ yúor, ‗name‘ yee
As has been discussed in the preceding section, Dagaare count nouns

show evidence of internal semantic classification based on their plural forms, a
phenomenon that is quite common among Niger-Congo languages, especially
those of the Bantu family (cf. Welmers 1973: Ch. 6 & 7; Williamson &
Blench 2000).5 The general principles for the noun class system in Dagaare
are, however, not very apparent and further research is needed to systemically
identify them. Only a few principles have been identified in the present study,
comprising humanness, loaned words, kinship and social status, relational
body parts, discreteness and collectiveness. These are illustrated below:
(23) General human nouns
singular plural
nɩr, ‗person‘ nɩbɛ
dɛb, ‗man‘ dɛɛbɛ
pɔw, ‗woman‘ pɔwbɛ
sãan, ‗visitor‘ sãamɛ ~sãabɛ
(24) Loaned nouns

singular plural
polisi, ‗police‘ polisiri
nɛɛsɩ, ‗nurse‘ nɛɛsɩrɩ
dɔkta, ‗doctor‘ dɔktarɩ
cɛnsɩ, ‗roofing sheet‘ cɛnsɩrɩ
5
This thesis was completed and already being examined when I became aware of Miehe et
al.‘s (2012) edited volume on noun class systems in Gur languages so I could not benefit from
their much detailed account in my analysis here, especially Miehe‘s (2012) chapter on the
Dagara cluster. I will take it into account in my intended book project.
69
(25) Nouns of kinship/social status
singular plural
sãakʋm, ‗grandfather‘ sãakʋm mɩnɛ
makʋm, ‗grandmother‘ makʋm mɩnɛ
sãa, ‗father‘ sãamɩnɛ
nàa, ‗chief‘ na-mɩnɛ
(26) Nouns indicating relational body parts
singular plural
niwn, ‗face‘ niwe
zukur, ‗back‘ zukuwe
lombowr, ‗side‘ lombowe
(27) Nouns representing discrete entities

Singular plural
libir,‗money‘ libie
yir, ‗house‘ yíe
yùor, ‗pot‘ yìe
nyuur, ‗yam‘ nyie
(28) Nouns indicating collectiveness
Singular plural
bie, ‗child‘ bibiir
pɛrʋ, ‗sheep‘ píir
náab, ‗cow‘ nii
waab, ‗snake‘ wiir
As mentioned in our discussion on morphemes, general human nouns

normally form their plural by the suffix -bɛ (23) while loaned words, including
modern occupational terms, have developed a new suffix, -rɩ, -ri or -li,
depending on the nature of the stem and [ATR] quality of the word (24). Thus,
while the plural form of traditional occupations such as kʋɔra (‗farmer‘) and
yɛyɛrɛ (‗trader‘) are respectively kʋɔbɛ and yɛyɛrbɛ, the plural forms of polisi
70
(‗police‘), nɛɛsɩ (‗nurse‘), dɔkta (‗doctor‘) and faara (‗priest‘) are polisiri,
nɛɛsɩrɩ and dɔktarɩ respectively. Also, in addition to the relation of -bɛ as –be
as mentioned in Section 3.4.1, its realisation can also be conditioned by a
neighbouring nasal vowel as in sãamɛ (‗visitors‘). Social relationship terms
such as sãakʋm (‗grandfather‘), makʋm (‗grandmother‘), sãa (‗father‘) and
nàa (‗chief‘) takes the morpheme mɩnɛ. As the variation in the example (25)
show, this morpheme is mostly realised as a postposition. In addition,
relational body parts (i.e. face, back and sides) take the suffix –we in their
plural (26). On the other hand, nouns representing discrete entities normally
undergo vowel mutation, taking the glide –íe in their plural. Examples are libir
(‗money‘; traditionally meaning ‗cowries‘), yir (house), and yùor (‗pot‘) in
example (27). Others include simir (‗peanut‘) versus simie (‗peanuts‘) and
mimir (‗eye‘) versus mimie (‗eyes‘). Finally, nouns whose plural indicates
collectiveness, such as bie and pɛrʋ, also undergo vowel mutation, typically
taking the long vowel /i:/ in their plural (28).
These semantic correlates of nominal suffixes are, however, tendencies
rather than absolute, and therefore do not systematically account for the all
noun class suffixes in Dagaare. As Bendor-Samuel (1971), notes for Gur
languages in general, it is often the case that while the majority of nouns in a
particular noun class correspond to a particular semantic category, there are
always a few members of the class which do not belong to this category. For
instance, while itis expected that wur (‗horse‘) would belong to the
collectiveness class, its plural is wuie, characteristic of discrete nouns, and the
plural form of the kinship term yɛb (‗sibling‘) is yɛbr. Non-human nouns such
as saab (‗dish‘) and laa (‗bowl‘) also take the suffix -bɛ, predominantly
associated with human nouns, in their plural (i.e. sabɛ and labɛ respectively).
Further, Dagaare is very rich in deverbal nouns and every verb with
content meaning can be nominalised (29). The realisation of lexical
nominalisation is irregular and follows the principle of vowel and consonant
harmony. Further research is, however, needed to identify the underlying
principles of these realisations. Some illustrations are given in (29) below:
71
(29) Deverbal nouns
Verb deverbal noun
ɩ, ‗do‘ ɩb (sing.), ‗action‘; ɩ-ɩrɩ (pl.) ‗acts, behaviour‘
be, ‗be‘, existential verb beba, ‗being, existence‘
cen, ‗go‘ cenu
dɩ, ‗eat‘ dɩb
ga, ‗lie down‘ gaa,
ir, ‗get up‘ iru
lɩɛbɩ, ‗turn, become‘ lɩɛbʋ, ‗becoming‘
wa, ‗come‘ waa
Finally, in Dagaare qualities are realised as nouns (but also as verbs)

rather than by a separate class of adjective (see adjectival verbs in 3.4.2.2
below). The variation in the construal of qualities, or what Pustet (2003: 13)
calls ‗property concepts‘, across languages has been discussed extensively by
linguists, at least since Sapir (1921: 117ff). It has been noted that, while in
languages such as those of the Indo-European family they are typically
realised by a class of adjective, in many other languages, they are realised as
either verbs (as in Wolof), nouns or both (as in Chinook) (cf. Dixon 1977;
Croft, 1991: 130-133; Stassen 1997: 200-205; Creissels 2000: 249-250; Pustet
2003: 7-16). In Dagaare qualities are construed both nouns and verbs and they
will be referredto as adjectival nouns and adjectival verbs respectively, not to
show that they display any special morphological characteristics distinct from
nouns and verbs (they do not) but to highlight their semantic uniqueness. I will
first illustrate adjectival nouns here and show their syntactic behaviour.
Adjectival verbs are discussed in Section 3.4.2.2 on verbs. As (30) shows,
adjectival nouns are normally count nouns and thus show singular-plural
distinctions:
(30) Adjectival nouns

Singular plural
pʋ̀la, ‗white‘ pʋ̀lɩ
sɛla ‗black, dark‘ sɛlɛ
ʋlʋ, ‗silver colour‘ ʋ́lɛ
72
zɩ̀ɛ, ‗red‘ zɩɩr
dambol, ‗fool‘ dambole
faa, ‗bad, not nice‘ faar
vla ~ vʋ̀la, ‗good‘ vɩɛlɩ
Although all adjectival nouns are relatively abstract and non-specific, they
display different degrees of abstractness and specificity among themselves.
Nouns that inherently imply a property of an entity such as dambol (‗fool‘)
and also baal (‗sick person‘) are relatively less abstract than those that
construe general qualities such as ‗faa (‗bad, not nice‘)‘ vla (‗good‘) and those
representing colour such as sɛla (‗black‘).
Syntactically, the nominal status of adjectival nouns implies that, like
other nouns, they can occur as Heads of nominal groups and function as
Subjects and Complements in a clause. Examples (31) to (33) illustrate the use
of adjectival nouns in Subject and Complement positions.
(31) Ɩ ɩ =n baal.
1SG COP.PFV FOC sick:person.SG
‗I am a sick person.‘
(32) Ɩ bɔbr =ɩ sɛlɛ.
1SG want.IPFV FOC black.PL
‗I want black ones.‘
(33) A zɩɩr aŋa vɩɛl =a
DEF red.PL DEM be:beautiful. PFV AFFR
‗These red ones are beautiful.‘
In example (31) and (32), the adjectival nouns baal (‗sick person‘) and sɛlɛ
(‗black ones‘) occur alone as Complements in their respective clauses. In (33),
zɩɩr (‗red ones‘) is the Head of the nominal group A zɩɩr aŋa (‗These red
ones‘), which is the Subject of the clause. As the example shows, adjectival
nouns can be modified by determiners such as the definite article a and the
demonstrative determiner aŋa (‗these‘). It should also be noted since
adjectival nouns are relatively abstract and non-specific, their referent is
normally assumed as given information, either as recoverable from the co-text
or identifiable from the material situational setting.
Another characteristic of adjectival nouns is that when they occur
73
together with a specific noun in the nominal group, as in kpar-sɛlɛ (‗black
shirts‘) in (34) and pɛn-vɩɛlɩ (‗beautiful cloths‘) in (35), the two nouns form a
nominal compound in which the adjectival noun indicate the number (i.e.
singular or plural) and the specific noun (e.g. kparɩ, ‗clothes‘; pɛnɛ, ‗cloths‘) is
normally clipped.
(34) Ɩ bɔbr =ɩ kpar-sɛlɛ.

1SG want.IPFV FOC shirt-black.PL
‗I want black (ones).‘
(35) Bɛ su =n pɛn-vɩɛlɩ zɩɩr
3PL.HM wear.PFV FOC cloth-beautiful.PL red.PL
Also, when more than one adjectival noun occurs in the nominal compound
such as in pɛn-vɩɛlɩ zɩɩr (‗red beautiful cloths‘) in (35) they always agree in
number.
3.4.2.2 Verbs
The word class to be discussed next is the verb. It is the domain of the
realisation of aspect, defined as the opposition between perfective and
imperfective meaning. Perfective aspect represents the process denoted by the
verb as bounded while imperfective aspect represents the process as
unbounded. While perfective aspect is realised by zero marking, imperfective
aspect is realised morphophonologically, by a complex system of prosodies.
Dagaare verbs can be classified into six groups in terms of the different
prosodies involved in the realisation of aspect. These prosodies are
summarised below:
(1) Tongue root prosody – a vowel harmony prosody where: (a) for verbs
with advanced tongue root, i.e. [+ATR] values, the final /i/ in the
perfective verb mutates to /e/ in the imperfective (e.g. píili, ‗start‘ /
píile) while (b) for verbs with retracted tongue root, i.e. [-ATR] values,
the final vowel /ɪ/ mutates to /ɛ/ (e.g. pɛlɩ ‗be white‘ /pɛlɛ).
(2) Tongue root plus nasal prosody: (a) a vowel harmony prosody where
the vowel in the imperfective suffix must agree with the vowel in the
root verb in terms of tongue root (±ATR), and (b) a simultaneous
74
vowel-consonant harmony where the consonant of the suffix must
agree with the vowel of the root verb in terms of nasality (e.g. kʋ̀, ‗kill‘
/kʋ̀-rɛ; kõ, ‗cry‘ / kõ-ne).
(3) Neutralisation of nasal prosody: a retrogressive consonant-vowel
harmony where a nasal vowel in the root verb loses its nasality when it
combines with the [-nasal] imperfective suffix -rɛ. (e.g. nyɛ, ‗see‘ /
nyɛ-rɛ)
(4) Palatal prosody where a short vowel in the root verb changes to a long
vowel in the imperfective form in anticipation of the [+aveolar]
imperfective suffix –r (su, ‗feed‘, ‗wear‘ /suu-r).
(5) Nasal prosody plus palatal prosody: (a) a vowel-consonant harmony
where the imperfective suffix, -r or –n, is chosen to agree with the
vowel of the root verb in nasality, and (b) a simultaneous consonant-
vowel harmony where the vowel changes from a simple vowel to a
closing glide in anticipation of the [+aveolar] suffix -r or -n (e.g. sɛ,
‗sew‘, ‗wear‘ /sɩɛ-r; sɛ, ‗roast‘ / sɩɛ-n).
(6) Palatal prosody plus labial prosody: two consonant prosodies where:
(a) a bilabial consonant (/b/ or/w/) ending the root verb is palatalised in
anticipation of the [+aveolar] imperfective suffix –r and, (b)
simultaneously, the suffix -r is labialised in anticipation of the
[+bilabial] ending of the root verb. (e.g. lɛb, ‗return‘ / lɛbr-rb).
The main groupings of verbs and their sub-divisions are further illustrated in
(36) to (40) in the order of the prosodies identified above.
The first group of verbs are those that differ in their realisation of
aspect based on the principle of vowel harmony, specifically that of tongue
root prosody (36). This category has to do with verbs whose perfective forms
end with /-l/ /-lɪ/, or /bɪ/ for [-ATR] vowels and /-l/, /-li/ or /bi/ for [+ATR]
vowels. Correspondingly, the imperfective for these verbs is realized by the
endings -lɛ,or -bɛfor [-ATR] verbs, and -le for [+ATR] verbs:
(36) Vowel harmony: [tongue root prosody]

[-ATR] [+ATR]
perf. imperf. perf. imperf.
lɩɛbɩ, ‗turn, become‘ lɩɛbɛ píili, ‗start‘ píile
75
maalɩ, ‗make, do well‘ maalɛ wul, ‗teach, show‘ wule
pɛlɩ, ‗be white‘ pɛlɛ yèl, ‗say‘ yèle
sɔwlɩ, ‗hide‘ sɔwlɛ yieli, ‗sing‘ yiele
The general principle underlining the realisation of aspect in this first group of
verbs is vowel mutation, where, for verbs with [-ATR] values, the final vowel
/ɪ/ in the perfective verb mutates to /ɛ/ in the imperfective and, for verbs with
[+ATR] values, the final /i/ mutates to /e/.
In the second group of verbs, tongue root prosody combines with nasal
prosody (or simply n-prosody) in the realisation of imperfective aspect. Here,
the prosody affects both vowels and consonants. This group can be divided
into four sub-categories as in (37) below:
(37) Vowel-&-vowel-consonant harmony: [tongue root plus nasal prosody]
[-ATR]; [-nasal] [+ATR]; [nasal] [-ATR]; [+nasal] [+ATR]; [+nasal]
perf. impf. perf. impf. perf. impf. perf. impf.
bɛr, bɛrɛ cen, cere dãw, dãwnɛ bì n, bì ne

‗release‘ ‗go‘ ‗be early‘ ‗put‘
kʋ̀, kʋ̀rɛ ír, íre nɔ̀ , nɔ̀ nɛ kõ, kõne
‗give‘ ‗get up‘ ‗love‘ ‗cry‘
'lɔr, 'lɔrɛ so, sore tʋ́ ɔ, tʋ́ ɔnɛ sõw, sõwne
‗immerse‘ ‗own‘ ‗be able‘ ‗help‘
tɛr, tɛrɛ yi, yire vɛ, vɛnɛ tõ, tõne
‗possess‘ ‗go out‘ ‗stop‘ ‗send‘
The realisation patterns illustrated above are identified as follows:
i. Verbs whose root forms have the features [-ATR] and [-nasal] select
the suffix -rɛ for their imperfective.
ii. Verbs whose root forms have the feature [+ATR] but [-nasal] selective
–re for the imperfective form.
iii. Verbs whose root forms have the feature [-ATR] but [+nasal], select -
nɛ for the imperfective form.
iv. Verbs whose root forms have both [+ATR] and [+nasal] select –ne for
76
the imperfective.
Thus, in each instance, there is tongue root harmony between the vowel in the
root morpheme and the vowel in the suffix and, simultaneously, a nasal
harmony between the vowel in the root morpheme and the consonant in the
suffix. As (37) shows, these harmony patterns construe a spectrum of a
combination of features, ranging from [-ATR] plus [-nasal] prosody to
[+ATR] plus [+nasal] prosody.
The third category is, in fact, the opposite of the progressive vowel-
consonant harmony illustrated by (37). The principle here is retrogressive
consonant-vowel harmony in which a nasal vowel in the root verb loses its
nasality when it combines with the [-nasal] imperfective suffix -rɛ (38). In
other words, the imperfective suffix -rɛ neutralises the nasality of the vowel in
the root verb.
(38) Consonant-vowel harmony: [neutralisation of n-prosody]
mɔr, ‗swell‘ mɔrɛ
nyɛ, ‗see‘ nyɛrɛ
ŋmã̀ a, ‗cut‘ ŋmàarɛ
ŋmʋr, ‗rush‘ ŋmʋrɛ
Further, the fourth group also involves a consonant-vowel harmony,

but the prosodies involved here is palatal prosody (or simply, r-prosody). A
short vowel in the root verb changes to a long vowel in the imperfective form
in anticipation of the [+aveolar] imperfective suffix –r, as example (39)
shows.
(39) Consonant-vowel harmony [r-prosody]

perfective impfective
kpi, ‗die‘ kpiir
tu, ‗follow‘ tuur
wa, ‗come‘ waar
su, ‗feed‘, ‗wear‘ suur
The fifth group of verbs is very much like the fourth. However, the
prosodies involved here are both r-prosody and n-prosody. We can divide this
77
category into two sub-groups (40):
(40) Vowel-consonant harmony: [r-prosody plus n-prosody]

[-nasal] [+nasal]
perfective impfective perfective impfective
de, ‗take‘ dier sɛ, ‗roast‘ sɩɛn
kɔ, ‗weed, farm‘ kʋɔr
nyɛ, ‗defecate‘ nyɩɛr
sɛ, ‗sew‘ sɩɛr
zɛ, ‗apply‘ zɩɛr
do, „climb‟ duor
In the first sub-group (e.g. de, ‗take‘ ~ dier), the perfective (or root) verbs
have the feature [-nasal] while those of the second sub-group (e.g. sɛ, ‗roast‘
~sɩɛn), have the feature [+nasal]. In both cases, imperfective aspect is realised
by r-prosody. The implication of the r-prosody in the realisation of the
imperfective aspect is that the vowel in the root verb undergoes a change from
a simple vowel such as /e/ or /ɛ/ to a glide such as /ie/ or /ɩɛ/, where the
initiating sound is a close vowel in anticipation of the [+alveolar] ending of
the imperfective verb. The only difference between the two groups then has to
do with the n-prosody, where:
i. a [-nasal] vowel in the root verb (e.g. de) requires the [-nasal] ending /-
r/ in the imperfective counterpart (e.g. dier) and
ii. a [+nasal] vowel in the root verb(e.g. sɛ) requires the [+nasal] ending /-
n/ in the imperfective counterpart (e.g. sɩɛn).
As the example (40) suggests, I could only identify one verb in the second
sub-category. Although I do not calim this is exhaustive (it could be), it is the
category with the least number of verbs in the language.
Finally, imperfective aspect in the six group of verbs is realised
simultaneously by both r-prosody and labial prosody. The verbs in this group
can also be divided into two sub-groups, namely those in which the labial
prosody is w-prosody and those where it is b-prosody (41):
78
(41) consonant prosody: [r-prosody plus labial prosody]
[+w-prosody] [+b-prosody]
perf. impf. perf. impf.
pàw, ‗close‘ pawrrw bɔ, ‗want‘ bɔbrrb
saw, ‗agree‘ sawrrw lɛb, ‗return‘ lɛbrrb
siw, ‗get down‘ siwrrw ɔb, ‗chew, be painful‘ ɔbrrb
sɔw, ‗respond‘ sɔwrrw sɔb, ‗be black‘ sɔbrrb
As the examples show, the imperfective suffix here is also realised as –r.
However, in the w-prosody category, there is a ‗mutual expectancy‘ effect
between the /-w/ ending of the root verb and the /-r/ ending of the imperfective
verb (cf. Firth 1957). This means that while the [-wr] is palatalised, the [-rw] is
labialised. Similarly, in the b-prosody, there is mutual expectancy between the
/-b/ and /-r/ sounds that end the imperfective verb, as in bɔbrrb (‗wanting‘). In
Chapter 6, aspect will be useful in our discussion of transitivity (cf. Sections
6.4.3 on mental clauses & Section 6.7 on relational clauses).
Apart from the different distribution of verbs based on their realisation
of aspect, Dagaare verbs can also be classified based on their semantic
properties and the different ways in which they contribute to meaning in the
clause. Based on this dimension, the verbs can broadly be classified into two
types, main verbs and catenative verbs. Main verbs can occur as Heads of the
verbal group in the clause and, ideationally, they represent the event, in a very
broad sense, realised by the clause. All the verbs that have been used for
illustrations in the discussion on aspect above are main verbs. Here, I will only
highlight one group of main verbs that need special attention, adjectival verbs.
Other main verbs will be discussed in detail in Chapter 6 on transitivity.
(1) Adjectival verbs: Adjectival verbs do not display any morphological
characteristics distinct from other Dagaare verbs (cf. Section 3.4.2.1 on
adjectival nouns). They however need a special statement due to the
typological variation in the realisation of qualities across languages. They are
illustrated in (42) below:
79
(42) adjectival verbs
cɩ, ‗be smart, be wise‘ cɩrɛ
gem, ‗be foolish‘ geme
pɛlɩ, ‗be white‘ pɛlɛ
pɔ̀l, ‗be beautiful‘ pɔ̀lɛ
mʋ̀ɔ, ‗be red, be ripe‘ mʋɔnɛ
sɔb, ‗be black, be dark‘ sɔbr
ʋ́lɩ, ‗be silver colour‘ ʋ̀lɛ
vɩɛl, ‗be good, be nice‘ -
yawmɛ, ‗be plenty‘ -
bɩɛrɩ, ‗be sick‘ bɩɛrɛ
Except for the verb vɩɛl (‗be good, be nice‘) and yawmɛ (‗be plenty‘),
adjectival verbs show distinction in aspect, patterned along the prosodic
realisation of aspect discussed above. These two verbs are inherently
perfective. Also, when the examples displayed in (42), especially verbs
representing colour, are compared with the examples provided for adjectival
nouns in (30), it can be realised that the same quality or property can be
represented either nominally or verbally. Adjectival verbs will also be useful
in the discussion of relational clauses of the attributive type in Chapter 6 (cf.
Section 6.5.2.2).
(2) Catenative verbs: Catenative verbs are verbs with an auxiliary status in
the verbal group. Thus, compared to main verbs, catenative verbs shade into
grammatical items in on the lexis-grammar continuum. Seven catenative verbs
are identified in this study and they are presented in Table 3.4 (see Creissels,
2000: 239 on auxiliaries in African languages). Catenative verbs seem to have
grammaticalised from main verbs apparently in their use in verbal group
complexes (‗serial verb constructions‘) (see Bendor-Samuel 1971: 160 on this
tendency in Gur languages). They generally have the following characteristics:
i. They are restricted in terms of the number of items in the class – i.e.
only seven verbs identified in this study.
80
Table 3.4 A List of Dagaare catenative verbs
gloss source (main) gloss

catenative verb verb
faa ‗be too late‘ Faa ‗to go waste/be
sploit‘, ‗to
escape‘
tʋ́ɔ ‗be able‘ tʋ́ɔ ‗overcome‘
dãw ‗be/do earlier‘ Dãw ‗be first, to come
first‘
cãa ‗still be‘ Cãa ‗remain/still
exist‘
maa ~ maalɩ be exact maalɩ ‗make, do‘
lɛb ~ lɛ ~ laa ‗do/happen again‘ lɛb ‗return‘
saw ‗agree‘ Saw ‗agree‘, ‗respond‘
ii. They are semantically eroded – i.e. losing some of their content
meaning (cf. Table 3.4).
iii. They have an auxiliary function in the verbal group – i.e. contributing
additional meaning to the main verb in the verbal group rather
specifying the event.
iv. They have lost their aspect distinction – i.e. they occur only in the
perfective form, which is the unmarked choice of aspect, even when
the main verb is imperfective.
Examples of the use of catenative verbs are given below, using the verb tʋ́ɔ
(‗be able‘):
(43) Workshop interview

A kɔb pɛr buor ʋ na tʋ́ɔ
DEF farming type which IDENT.SG POS.IND.FUT be:able.PFV
sow tɩ.
help.PFV 1PL
‗Which type of farming is the one that would be able to help us
81
(=‗Which type of farming would be able to help us).
||| Ŋmɩnŋmɩn na nyɩ ɩ || tʋ́ɔ
how IDENT.PL 2PL do.PFV be:able.PFV
gure? |||
sleep.IPFV
‗How is it that you are able to sleep?‘
(45) Political opinion interview
A nɩbɛ kʋ tʋ́ɔ cɛlɛ a
DEF people NEG.IND.FUT be:able.PFV listen.IPFV DEF
ʋ ƴɛrʋ ɛ.
3SG words NAFFR
‗The people cannot listen to his/her words.‘
As the examples show, the catenative verb tʋ́ɔ remains unchanged irrespective
of the aspectual value of the main verb in the clause. In (43), it co-occurs with
a main verb in the perfective aspect while in (44) and (45) it occurs with main
verbs in the imperfective aspect. Also, its function in the verbal group is to add
modal meaning to the clause, specifically ability.
3.4.2.3 Adverbs
As the discussion in the preceding two sections shows, Dagaare has a rich
class of nouns and verbs. Dagaare adverbs, on the other hand, are not a large
class although they do not constitute a close system (see Welmers 1973: Ch.
15 for an overview of adverbs across African languages). As a general
characteristic, they have no inflectional affixes. The sub-classes of adverbs are
adverbs of time, place, manner and adverbial particles. Each of these will be
discussed in turn.
(1) Adverds of time: Adverbs of time modify the verb in the clause in which
they occur by indicating the temporal location of the event construed by the
verb. As has been reported for many African languages, Dagaare time adverbs
mostly indicate depth of time, spaning days and years (cf. Welmers 1973: 447-
448). Together, there are nine depth-of-time adverbs in current use and they
82
are listed in (46) below:6
(46) Adverbs of indicating depth of time

dɩtaarɩ ‗two years ago‘
dɩya „last year‘
daar ‗two days ago‘
zà a ‗yesterday‘
zɩna ~dɩa ‗today‘
bio ‗tomorrow‘
deyere ‗two days from today‘
dɛtɛrɛ (not ‗three days from today‘
in frequent use)
vɩɛr ‗next year‘
As the list shows, regarding the day of speaking, speakers can make past
reference up to ‗two days ago‘and future reference up to ‗three days ahead‘
although dɛtɛrɛ is not in frequent use. With regards to years, past reference can
be made up two years ago while future reference is only limited to one year.
Thus, while future time has a deeper depth of time reference than past time in
terms of number of days, past time conversely has a deeper depth of time
reference than future time in terms of years. As in English, ‗this year‘ is not
expressed by a single word but by an adverbial group yuon na (‗this year‘).
The extreme points of reference also tend to have the potential for
extending their meanings. Thus, daar (‗two days), dɩtaarɩ (‗two years ago‘)
and vɩɛr (‗next year‘) can be used to mean a few days ago, a few years ago and
some unspecified year in the future respectively. But in this sense, they are
often modified by a demonstrative determiner (Compare (47) with (48) and
(49)):
6
Older speakers recall that some ‗depth-of-time adverbs‘ such as jamgboro (lit. a kind a
vegetable) and dɛnyamɛ (lit. ‗old men‘) were in use in the past. But it is most likely that these
were names of market days, that were also used as names of days of the week, rather than time
adverbs since market/week days were named after objects, towns, circumstances, etc. All
traditional names for week/market days have been lost and replaced by English/French and
Akan week days.
83
(47) Ʋ dɩtaarɩ wa =n ka.
3SG two years ago come.PFV FOC here.
‗S/he came here two years ago.‘
(48) Ʋ dɩtaarɩ kãw wa =n ka.
3SG two years ago some come.PFV FOC here.
‗S/he came here some years ago.‘
(49) Ʋ na wa =n vɩɛr kãw.
3SG POS.IND.FUT come.PFV FOC next year some
‗S/he will come some time in the years ahead.‘
For future time, neither deyere (‗two days from today‘) nor dɛtɛrɛ (‗three days
from today‘) can be used for unspecified time reference. It should also be
mentioned that, when daar is used for general time reference, it becomes
almost synonymous whith another adverb sãw, which tends to refer to a more
distant temporal location than daar (50) & 51).
(50) Ʋ wa =n ka daar kãw.
3SG come.PFV FOC here a few days ago some.

‗S/he came here some days ago.‘
(51) Ʋ wa =n ka sãw kãw
3SG come FOC here time some.
‗S/he came here some time ago.‘
Both daar and kãw are also used as generic nouns for time (52-53):
(52) Nyɩ bãw nɩ a daar nɛ [[ɩ na

2PL know.PFV FOC DEF time DEM 1SG REL
=a wa a]].
POS.IND.FUT come.PFV JUNC
‗You know the time [[that I will come]].

(53) Nyɩ bãw nɩ a sãw nɛ [[ɩ na
2PL know.PFV FOC DEF time DEM 1SG REL
=a wa a]].
POS.IND.FUT come.PFV JUNC
‗You know the time [[that I will come]].‘
84
In addition to depth-of-time adverbs, other time adverbs include among others
koroza (‗long time ago‘), and the compound adverbs dɩa-bio
(‗nowadays/today‘; lit. ‗today tomorrow‘), tew-kor-daar (‗in the olden days‘;
lit. earth-old-time), koroza sãw (‗in the past‘; lit. ‗old time‘), dãw-bio (‗next
day‘; lit. ‗early tomorrow‘), daar yɩɛw na (‗not long ago‘, ‗not long from
now‘; lit. ‗this near time‘), ƴɛrɛ na (‗now‘).
The position the time adverbs discussed above in the clause is variant
and mixed. There are two main positions involved: (i) before the verb and
after the Subject noun group (i.e. pre-verbal position) and (ii) clause final
position. First, with regards to depth-of-time adverbs, those with past time
reference, namely dɩtaarɩ (‗two years ago‘), daar (‗two days ago‘), dɩya (‗last
year‘) and zà a (‗yesterday‘), are versatile and can occur in both positions (see
(54) and (55)). On the other hand, zɩna~ dɩa (‗today‘) and adverbs with future
reference occur in clause final position only (compare (56) and (57)).
(54) Ɩ zã̀ a cen na.

1SG yesterday go.PFV AFFR
‗I went yesterday.‘
(55) Ɩ cen nɩ zã̀̃ a.
1SG go.PFV FOC yesterday
‗I went yesterday.‘
(56) Ɩ na cen nɩ bio.
1SG POS.IND.FUT go.PFV FOC tomorrow
‗I will go tomorrow.‘
(57) *Ɩ bio na cen na.
1SG tomorrow POS.IND.FUT go.PFV AFFR
‗I will go tomorrow.‘
For adverbs other than those of depth of time, only dãw-bio (‗next day‘) can
occur both preverbally and clause finally while the others occur only at clause
final position. However, all the time adverbs can be placed at clause intial
position, before the Subject of the clause, as Themes (see Chapter 5 on
Theme):
85
(58) Vɩɛr, tɩ na kɔ na.
next year 1PL POS.IND.FUT farm.PFV AFFR
‗Next year, we will farm.‘
Finally, the time adverbs can be used nominally as the Subject in the clause, as
in (59) below:
(59) Bio na ta =n fɔw lɛ.

Tomorrow POS.IND.FUT reach.PFV FOC quicklyADV
‗Tomorrow will come very quickly.‘
(2) Adverbs of place: Adverbs of place construe the spatial location of the
event represented by the verb in the clause. Dagaare place adverbs can be
grouped into three, namely (i) non-derived adverbs such as the demonstrative
adverbs ka (‗here‘) and be (‗there‘) as well as positional adverbs such as tew
(‗on the ground‘), sazu (‗up‘), pile (‗down‘) sɔwɔ (‗middle‘) and sɔsɔwlɩ sɔwɔ
(‗exactly in the middle‘) and (ii) the class of adverds derived from nouns with
thelocative suffix variously realised as -mɩ, -mi, -ɩ, -i (cf. Section 3.4.1). The
specific category of nouns that serve as the source of adverbs are body part
nouns such as pʋɔ (‗stomach‘), gbɛr (‗leg‘) and zukur (‗back‘) and locative
nouns such as kʋɔ (‗water‘) and salom (‗sky‘) (cf. Heine, Claudi &
Hünnemeyer 1991: 128-130; Heine 2011). In addition to the examples
provided in example (20), other examples are lombowri (‗beside‘; from
lombowr, ‗side‘), pɛrɩ (‗bottom‘, from pɛr, ‗anus‘), nyaamɩ (‗in the chest‘,
‗towards the top‘; from nyaa, ‗chest‘), sɩɛmɩ (‗under‘; from sɩɛ, ‗waste‘),and
nyʋɔrɩ (‗at the tail end‘; from nyʋɔr, nose). Other body parts related adverbs
such as gbèé (‗on the forehead‘; from gbè, ‗forehead‘), zùú (‗on the head‘;
from zù, ‗head‘) are derived by vowel mutation and a shift from level to
contour tones rather than adding the derivative suffix.
Adverbs of place occur at clause final position (60). Also, like time
adverbs, they can be used nominally as a Subject in a clause (61):
86
Bɛ tuori tɩ a be.
3PL.HM meet.PFV 1PL DEF there
‗They should meet us there‘
(61) A be vɩɛl =a.
DEF there be:good.PFV AFFR
‘There is good (= That place is good).‘
(3) Adverds of manner and ideophones: Adverbs of manner typically

indicate how the event represented by the verb in the clause is actualised.
Dagaare manner adverbs consist of a small class of descriptive adverbs such as
vla~vʋ̀la (‗well‘), yaga (‗plenty‘), fɔw (‗quickly‘) and paa (‗very much‘) and
ideophones. Vla has the same form as the adjectival noun vla (‗good‘) and paa
is borrowed from Akan.
Ideophones are common in African languages and they have been
discussed extensively in both the descriptive (e.g. Kunene 1965; Bodomo
2006; Dingemanse 2011a, b, c) and typological (e.g. Welmers, 1973; Watters,
2000; Blench 2010) literature. According to Welmers, the term ‗ideophone‘
was perhaps first introduced by Clement Doke (cf. Doke 1935), who
characterised it as: ―a vivid representation of an idea in sound. A word, often
onomatopoeic, which describes a predicate, qualificative or adverb in respect
of manner, colour, smell, action, state or intensity‖ (Doke 1935: 118; Welmers
1973: 461). Ideophones need not be onomatopoeic though and they are often
language specific. Also, across languages, they vary on the word class they
belong to. They can be nouns, verbs, adjectives or adverbs and their
classification is thus based on language specific analysis. (cf. Welmers 1973).
Another characteristic of ideophones that has been attested across languages is
that they typically involve reduplication and even triplication. In Dagaare, they
can be repeated up to four times to create strong intensification (e.g. zɩɛ ɓõɓõ
ɓõɓõ, ‗very very red‘). Examples of ideophones in Dagaare discourse are
highlighted in (62) and (63) below (cf. also Bodomo 2006):
87
(62) Bible.is (Rev. 18: 8b)
Vũu na dɩ ʋ nɩ kʋmɔkʋmɔ.
Fire POS.IND.FUT burn.PFV 3SG FOC IDEO
‗Fire will burn it completely.‘

(63) Bible.is (Rev 17: 3b)
Ɩ tɩ nyɛ nɩ pɔw a be, ʋ
1SG PST.REM see.PFV FOC woman DEF there 3SG
zɩ wɛ-dʋw-zɩɛ ɓõɓõ zu.

sit.PFV wild-animal-red.SG IDEO ADP
‗I saw a woman there, she was sitting on a very red wild animal.‘
These examples show that ideophones can be the Head of an adverbial group,
functioning as circumstance of Manner, such as kʋmɔkʋmɔ in (62). They can
also be used in the nominal group as modifiers of adjectival nouns such as
ɓõɓõ in (63). Due to their iconic nature, specific ideophones normally
collocate with particular words. For example, the following ideophones
collocate with particular adjectival verbs or nouns of colour: kpirkpir (‗black‘)
furututu (‗white‘) ɓõɓõ (‗red‘) and yololo (‗yellow‘). However, different
ideophones can be used with the same colour term to express different shades
of the colour such as the contrastive use of ɓõɓõ and wʋlʋlʋ in (64) below:
(64) Bible.is (Matie 16: 2-3a)

Ʋ tɩ sɔw bɛ na: “Zaanʋɔra wa
3SG PST.REM respond.PFV 3PL.HM AFFR evening EVT
ta a, nyɩ mɩ́ yèl =a: „A salom

reach.PFV JUNC 2PL HAB say. PFV AFFR DEF sky
ɩ =n zɩɛ ɓõɓõ, a zie na vɩɛl
COP.PFV FOC red IDEO DEF place POS.IND.FUT be:good
=a‟; ɛ a biɓaara nyɩ yèl: „Dɩa a,
AFFR CONJ DEF morning 2PL say. PFV today JUNC
a zie kʋ vɩɛl ɛ, a salom

DEF place NEG.IND.FUT be:good.PFV NAFFR DEF sky
na ɩ zɩɛ wʋlʋlʋ a ƴãw.‘‖
ADVLZ COP.PFV red IDEO JUNC sake
88
―S/he answered them: ‗When evening comes, you say: ‗The sky is
deeply red, the weather will be good‘; and in the morning, you say:
‗Today, the weather won‘t be good, since the sky is shady red.‘
As a general characteristic, Dagaare adverbs of manner occur in clause final

position.
(4) Adverbial particles: Adverbial particles are a closed system of adverbs

and they lie in the mid-region of the lexis-grammar continuum. Some of these
particles derive their specific meaning in the context in which they are used.
Table 3.5 presents fifteen adverbial particles identified in this study with
examples, and they perhaps constitute an exhaustive list. As Table 3.7 shows,
adverbial particles can be grouped according to their position in the clause.
Four groups can be identified: (i) those that are preverbal, occurring between
the Subject noun group and the verb; (ii) those that are post verbal, occurring
in clause final position; (iii) those that are versatile, having the ability to occur
either at preverb or clause final position; and (iv) one adverb intensifier lɛ,
which occurs after the adverb it modifies.
Adverbial particles are mostly grammaticalised items although their
specific lexical sources are not clear. However, all those that can occur
preverbally, which automatically include all the versatile ones, most likely
originate from verbs in the environment of verbal group complexes (‗serial
verb constructions‘) and they may have first been used as catenative verbs.
The following grammaticalisation pathways can be hypothesised: VERB >
CATENATIVE > ADV. PARTICLE. This hypothesis is supported by the fact that in
affirmative clauses, some of the particles, specifically dɛ, nιɛ, and sιrι, display
a syntactic characteristic similar to the first verb in a Dagaare serial verb
construction (67), namely they can immediately be followed by the affirmative
particle (66). Thus, both (65) and (66) below are both acceptable:
(65) Ʋ sιrι wa na.

3SG truly come.PFV AFFR
‗Sh/e truly came, as he promised.‘
89
Table 3.5. Dagaare adverbial particles
adverbial particle gloss example

post- adverb
pre-verb versatile
verb intensifier
Fʋ dɛ cen na.
dɛ just,
‗You definitely
definitely
went.‘
Ʋ wa na dɩɛ.
dɩɛ
anyway ‗S/he has come
anyway.
Ʋ wa na ende.
ende
anyway.‘
Ʋ ɛ wa na.
ɛ
anyway.‘
gba Ʋ gba wa na.
even
‗he has come even.‘
Ʋ nιɛ wa na bι?
nιɛ even ‗Has s/he even
come?‘
mɛ̀ like, Ʋ ɩ na mɛ̀ ana a.
seems ‗It is like this one.‘
mɩ̀ Ʋ mɩ̀ dɩ!
also, too ‗S/he should eat
too.‘
A fʋʋ mɔ, wa ka!
mɔ̀
like, even ‗You even, come
here!‘
A dɔɔ nɔw yarɛ na.
nɔw probably,
‗Maybe, the man is
maybe
mad.‘
A bɛ pãa ŋmɛ
pãa dawuro.
now, then
‗And then they beat
a gongon.‘
A ma yaa wa na.
yaa even,
‗Maybe, the mother
maybe
has come.‘
Ʋ wa na pɛɛ.
pɛɛ definitely/
‗S/he has surely
surely
come‘
Ʋ sɩrɩ na wa.
sɩrɩ truly ‗S/he came, true to
his/her words‘
Yi yow fɔw lɛ!
lɛ
very ‗Go out very
quickly!‘
90
(66) Ʋ sιrι na wa.
3SG truly AFFR come.PFV
‗S/he truly came, as promised.‘
(67) Ʋ zɔ na wa.
3SG run.PFV AFFR come.PFV
‗S/he ran and came (=He ran here).‘
While the affirmative particle can be placed either after the adverbial particle
or clause finally as the difference between (65) and (66) shows, with the serial
verb construction, it can only come after the first verb in the complex (67). A
plausible interpretation of the flexibility with the adverbial particle
construction is that sιrι has lost its verbal status. On the other hand, an
alternative analysis can classify it as a catenative verb or both an adverbial
particle (as in (65)) and a catenative verb (as in (66)). It is, however, evident
that particles such as sιrι and also dɛ and nιɛ are at an advanced stage of their
grammaticalisation pathway into adverbial particles, their only verbal trace
being the alternation exemplified in (65) and (66). For those particles that are
restricted to post-verbal position, dɩɛ (‗anyway‘) and pɛɛ (‗definitely‘) are
borrowed from Akan, where they occur in the same position.7 The particle mɛ̀
(‗like‘), like its English equivalent, is a comparative adverb as in A zebra is
like a horse.
Functionally, adverbial particles are generally attitude markers and
play an interpersonal role in enacting the clause as an exchange (cf. Chapter 4,
Section 4.4.4.2). However, dɛ has both a temporal sense (i.e. ‗just‘) and a
modal sense (i.e. ‗definitely‘) while pãa (‗now, then‘) has modal (68),
temporal (69) and conjuntive (70) senses in different grammatical
environemnts:
(68) Pãa ir!

ADV get up.
‗Now get up (‗You can get up now‘)!‘
7
It is most likely that the adverbial particle ende evolved from a fusion of the original
Dagaare ɛ and its synomous counterpart dɩɛ, borrowed from Akan, in the sequence ɛ dɩɛ. As a
result, ende is synonymous and interchangeable with any of the two particles (cf. Table 3.7).
91
(69) Tι pãa wa na.
1PL ADV come.PFV AFFR
‗We have now come.‘

(70) Tι cen na ɛ ʋ pãa wa.
1PL go.PFV AFFR CONJ 3SG ADV come.PFV
‗We went before s/he then came.‘
In (68), pãa indicates permission, in (69), it shows the time of the event
represented by the verbal group, and in (70) it has a conjunctive sense, making
a textual contribution to the relationship between the two clauses.
It should also be noted that the adverbial particles are distributed
differently across mood types. The different selections are summarised below:
(i) Dɛ (‗just‘; i.e. only the temporal sense), gba, mɛ̀, pãa, mɩ̀ and sιrι
occur in all mood types.
(ii) Yaa occurs in only indicative clauses (spefically declarative &
interrogative: polar).
(iii) Dɛ (‗definitely‘) pɛɛ and mɔ̀ occur in only declarative clauses.
(iv) Nιɛ and nɔw occur in only interrogative clauses.
(v) Dιɛ, ende, and ɛ occur in both declarative and imperative clauses.
The differences in their distribution reflects their different values in the system
of modal assessment (cf. Chapter 4, Section 4.5).
3.4.2.4 Pronouns
The next word class to be discussed in this section is pronoun and the
discussion will be limited to personal pronouns. Dagaare personal pronouns
can be classified based on several variables, including person, number,
emphasis and their function in the clause. Table 3.6 gives a list of the personal
pronouns (see Section 3.4.3 on identifying pronouns). As the table indicates,
all three persons show further distinction between singular and plural and, in
addition, third person pronouns distinguish between human and non-human
for plural. Thus, in Dagaare, unlike in English, it is the third plural rather than
the third singular which contrast between human and non-human. The third
92
plural pronoun a is also used for impersonal reference and it is alsothe pro-
form for mass nouns. The implication is that some instances of the use of ‗it‘
in English such as in reference to mass nouns and as an expletive pronoun
translate into Dagaare as the third plural a (e.g. in It has boiled, where it refers
to water).
In terms of function, Dagaare pronouns almost do not show case
distinctions in their role as subject, complement and possessive determiners. It
is only the first person singular which shows a formal contrast between
nominative (Ɩ) and accusative (mɛ) case. In addition, the first and second
person singulars can occur as enclitics in complement position. There is,
however, a systematic distinction between emphatic and non-emphatic
pronouns across functions. This distinction is textually motivated in the sense
that emphatic pronouns serve as resources for contrastive focus and will be
discussed in Chapter 5 (see Section 5.4.2). First and second plural emphatic
pronouns as well as third person emphatic pronouns take different forms when
they occur as Subjects in identifying clauses (identifying clauses are discussed
in Chapter 6, Section 6.5.2.1).
Table 3.6. Pronominal system of Dagaare
subject complement possessive

PERSON
NUMBER
HUMANES
non- emph. non- emph. clitic non- emph.

emph. emph. [after emph.
vowel]
non- identi-
ident. fying
1st sing. ɩ mãa mɛ mãa =m ɩ mãa
pl. tɩ tιιm tιmɛ tɩ tιιm tɩ tιιm

2nd sing. fʋ fʋʋ fʋ fʋʋ =b fʋ fʋʋ
pl. nyɩ ~ nyιιm nyɩmɛ nyɩ ~ nyιιm nyɩ nyιιm

nɩ ~ ~ nɩmɛ nɩ ~ nιιm ~ nɩ ~ nιιm
nιιm
3rd sing. ±human ʋ ʋl ʋlɛ ʋ ʋl ʋ ʋl
+human bɛ bɛl bɛlɛ bɛ bɛl bɛ bɛl

pl. -human a al alɛ a al a al
93
Further, reflexivisation in Dagaare is not realised by single word
pronouns. It is realised by a nomimal group in which the pronoun modifies a
reflexive noun, tʋɔra or mɩŋa (‗self‘):
(71) The story of Jesus

Dɔkta a, sanɩ fʋ tʋɔra!
doctor JUNC heal.PFV 2SG self
‗Doctor, heal yourself!‘
(72) Ɩ sanɩ nɩ ɩ̃ mɩŋa!
1SG heal.PFV FOC 1SG self
‗I have healed myself!‘
Only non-emphatic pronouns of the form of subject or possessive pronouns

are used for reflexivisation.
Finally, it is important to make a statement on the status of Dagaare
pronounsas separate words as opposed to pronominal affixes or subject
markers attached to verbs as it is the case in many West African languages,
especially Kwa languages (cf. Creissels 2000: 235-236, 238-239). Generally,
they do not bond together closely with the verb phonologically and native
speakers easily recognise them as separate words. The following
characteristics further support the interpretation of Dagaare pronouns as
separate words:
1. They retain their vowel harmony irrespective of the phonetic

characteristics of the adjacent verb (with the exception of ʋ, third
singular, only when it occurs in complement position; cf. Section
3.2.3).
2. Other words such as adverbs and particles can be placed between
subject pronouns and the verb (cf. Section 3.4.2.3 and examples
thereof).
3. Some pronouns in complement position (i.e. fʋ mɛ) can themselves be
reduced as enclitics (cf. Table 3.4).
3.4.2.4 Particles
The last word class to be considered is the class of particles. On the cline of
94
Table 3.7. List of Dagaare particles across metafunctions
interpersonal ideational textual
particle gloss particle gloss particle gloss
full clitic full clitic full clitic
na n=, affirmative tɩ temporal nɩ =n, focus

=a range =ɩ
(‗until‘)
ɩ, e, non- wa =a eventuality na binder
ɛ1 affirmative (adverb-
ializer,
nomina-
lizer,
relativer)
bɩ, polar tɩ remote past a#3 clause

wɛ, interrogative juncture
kpo,
ʋ,
na2 =a positive future #a3 contin-
indicative uative
kʋ2 negative future
indicative
bɛ2 negative non-future
indicative
ta, negative nɩ =n, agentive
taa imperative =ɩ
naa, modal tɩ distal
kʋʋ, (‗andative‘)
taa
wɛ, admonitive wa proximal
kɛ (‗ventive‘)
mɛ́ expectation mɩ́ habitual
mɔ counter-
expectation
dɛ emphasis
ka, insistent
kaka
na exhortative
yaa endearment
wɛ exclamative
Note: 1[ATR] variants; 2na, kʋ and bɛ simultaneously realise polarity (interpersonal) and tense
(primarily ideational); 3placed at clause boundary position.
lexicogrammar, particles lie at the grammatical end of the continuum. Dagaare

has a rich class of particles for realising grammatical meaning in the clause
95
(e.g. mood, attitude and focus) and verbal group (e.g. tense, modality and
polarity). Table 3.7 lists 35 unique particles, that is, not counting clitic forms
and [ATR] variants of the same particle.
As the table shows, the different particles realise meanings across the
metafunctions, namely interpersonal, ideational and textual functions. Many of
the interpersonal particles are placed prosodically at clause initial or final
position as mood and attitude markers:
(73) St. Maria play
Pɛw mɛ ka!
lend.PFV 1SG.ACC M.INS
‗Lend me (some money), I insist!
(74) Mɛ́ nyɩ dɩ na.

HST 2PL eat.PFV AFFR
‗I believe you have eaten.‘
In (73), the particle ka is placed at the end of the imperative clause to

modulate the proposal in an imploring tone and in (74) Mɛ́ is placed clause
initially to signal the speaker‘s uncertainty towards the proposition. On the
other hand, the particle na in (74) marks the clause as an affirmative clause.
What all these particles have in common is that they negotiate the proposal (as
in 73) or proposition (as in 74) realised by the clause. These interpersonal
resources will be discussed in detail in Chapter 4.
Ideational particles function in the verbal group and realise such
meanings as tense (75) and directionality (see 75 & 76), among others (verbal
group underlined):
(75) Dɛbɛlɛrɛ nʋ tɩ be be o!
spider IDENT.SG PST.REM EXIST.PFV there PRT
‗Spider was the one who lived (=There lived spider)!‘

(76) Cen tɩ dɩ!
go.PFV DIST eat.PFV
‗Go to eat!‘
96
(77) Wa wa dɩ!
come.PFV PROX eat.PFV
‗Come to eat!‘
In (75), for instance, the particle tɩ in the verbal group marks the temporal
frame of the clause as that of remote past, the typical tense of traditional
narratives. In (76), the distal marker tɩ indicates movement away from the
speaker and, in contrast, wa in (77) shows movement towards the speaker.
Ideational particles therefore contribute to the meaning of the clause as a
representation of experience. Ideational systems below the clause will not be
discussed in detail in this study.
Unlike interpersonal and ideational metafunctions, only few particles
realise textual meaning, mainly in relation to information focus and clause
combining. Information focus will be discussed in Chapter 5 while clause
combining is considered briefly in Section 3.4.3 below. As Table 3.7 also
shows, however, some particles have multiple functions. For instance, the
polarity markers bɛ (negative indicative, non-future) kʋ (negative indicative,
future) and na (positive indicative, future) also function as tense and mood
markers in the verbal group:
(78) Ɩ bɛ dɩ ɛ.
1SG NEG.IND.NFUT eat.PFV NAFFR
‗I have not eaten.‘

(79) Ɩ kʋ̃ dɩ ɛ.
1SG NEG.IND.NFUT eat.PFV NAFFR
‗I will not eat.‘

(80) Ɩ na dɩ na.
1SG POS.IND.FUT eat.PFV NAFFR
‗I will eat.‘
Thus, while (78) and (79) both mark the clauses in which they occur as
negative and indicative, they contrast in tense, with bɛ marking (78) as non-
future and kʋ marking (79) as future. In this sense, (78) is similar to (79) in
terms of tense but contrast with it in terms of polarity. Example (80) contrasts
with both (Polarity will be discussed in detail in Chapter 4 (see Section 4.5.1).
97
3.4.3 The Dagaare Clause
The clause is the highest-ranking unit serving as the point of origin of

grammatical systems and, as mentioned earlier, it is co-extensive with the tone
group in the unmarked case (cf. Section 3.2.4). Just as with the lower
grammatical units, the clause can be divided into classes based on its form and
function (see Figure 3.3). Two variables that tend to be universally relevant in
the classification of clauses are status and freedom (Halliday & Matthiessen
2004). On the variable of status, clauses can be divided into major and minor.
(1) Major and minor clauses: Major clauses in Dagaare, as in many other
languages, are those clauses that are predicated and embody metafunctional
diversity in the meanings they realise (i.e. simultaneously realising ideational,
interpersonal and textual meanings). They can be analysed into clausal
elements and normally have a Predicator element in their structure (see
Chapter 4 on interpersonal structure of the clause). As an illustration, example
(62) is analysed in the box diagram below in Figure 3.7.
Vũu na dɩ ʋ nɩ kʋmɔkʋmɔ
fire POS.IND.FUT burn.PFV 3SG FOC IDEO
Clause Subject Predicator Complement Adjunct
Actor Process Goal Manner

Theme Rheme
Given New
Group nom. verbal group nom. group adverbial
group group
Word noun particle verb pronoun particle adverb
Figure 3.7. Illustrating the structural analysis of a major clause
98
Minor clauses, on the other hand, are typically not analysable in terms of
clausal constituents. They are oriented towards the interpersonal function of
language and are realised by interjections, expletives and formulaic
expressions such exclamations and greetings as well as protolinguistic
remnants (cf. Halliday & Matthiessen 2014). Minor clauses will be discussed
and illustrated in Chapter 4 in relation to the system of MOOD (see Section
4.4.5).
(2) Free and bound clauses: The variable of freedom is a system of major
clauses. Major clauses can be either free or bound (cf. Matthiessen 2004: 612-
613 for a crosslinguistic account). Free clauses are those major clauses that are
syntactically independent while bound clauses consist of hypotactic clauses
and embedded clauses. An example of a free clause is analysed in Figure 3.4
above. A hypotactic clause is a ranking clause that is dependent on another
clause, a main clause, for its interpretation and an embedded clause is a
downrakned clause that is a constituent within another clause (cf. Hopper &
Traugott, 2003: Ch. 7). In order to make the distinction clear, I will first
highlight the characteristics of bound clauses below and illustrate them as the
discussion proceeds:
i. A bound clause cannot include end focus where one is required for a
corresponding free clause.
ii. A bound clause cannot occur with negotiation particles, such as the
affirmative (na) or non-affirmative (ɩ, e ɛ) particles, where they are
required for a corresponding free clause.
99
free
FREEDOM
major
+Pred.
bound
STATUS
minor
clause
+vocative
VOCATIVE
+vocative = nom. gp.
-vocative
Figure 3.8. Dagaare clause types presented as a system network
It should be noted that while all bound clauses have these characteristics, not
all clause without end focus or negotiation markers are bound clauses. Chapter
5 will discuss the conditions that motivate the absence or presence of
negotiation and end focus markers in free clauses (cf. esp. Section 5.4.3). The
following examples illustrate the difference between a free clause and a bound
clause (bound clause is underlined):
(81) Bible.is (Matie 16: 13)

||| A Yeezu na wa ta a Filib tẽw
DEF Jesus ADVLZ PROX reach.PFV DEF Phillip town
Sezaare a, || ʋ sowr =ɩ a ʋ
Caesarea JUNC 3SG ask.PFV FOC DEF 3SG
po-tuurbɛ […] |||

followers […]
―When Jesus got to Phillip‘s town Caesarea, he asked his followers …‘
100
(82) A Yeezu wa ta =n a Filib tẽw
DEF Jesus EVT reach.PFV FOC DEF Phillip town
Sezaare.
Caesarea
‗Jesus got to Phillip‘s town Caesarea‘
Example (81) is a clause complex comprising a hypotactic clause (the bound

clause) and an independent clause (the free clause). In the bound clause, the
binding particle na precedes the verbal group wa ta (‗reached‘). The nominal
group in Complement position, a Filib tẽw Sezaare, is not focused as would be
required for the corresponding free clause illustrated by (82). Another
observation in (81) characteristic of bound clauses is that they normally end
with a juncture prosody particle a simultaneously carrying a level-rising tone
(cf. Section 3.2.4).
Typical bound clauses are relative clauses, nominal clauses and
adverbial clauses. Relative clauses are bound clauses that modify the Head of
a nominal group (cf. Watters (2000: 225-228) for a brief account on African
languages). Comrie (1989) identifies three useful parameters on which
languages vary in their realisation of relative clauses, namely word order (i.e.
position of the relative clause in realtion to the Head of the nominal group);
status of the Head in the relative clause; and role of the Head in the main
clause. In terms of word order, the relative clause could be prenominal (as in
Korean), postnominal (as in English), or circum-nominal, where the head is
rather placed within the relative clause (as in Bambara). Dagaare relative
clauses are postnominal, which means they always follow the noun they
modify. The relative clauses in the following examples are underlined:

||| Erɔdɩ [[na ɩ a tew na-kpɛɛ
Herod REL be.PFV DEF town chief-big
lɩɛrɛ a Galɩle a]]
representative DEF Galilee JUNC
tɩ ƴaw na bɛ nyɔw
PST.REM make AFFR 3PL.HM catch.PFV
101
a Za || tɩ paw.|||
DEF John DIST jail. PFV
‗Herod [[who was the governor of Galilee]] had them arrest John and
jailed him.‘
Bɛ ɩ =n yéle yaga [[na vɩɛl
3PL.HM do.PFV FOC yéle plenty REL be:good
a]].
JUNC
‗They‘ve done many things that are good.‘
The nominal group consists of the Head noun Erɔdɩ followed by the relative
clause. Regarding the status of the Head in the relative clause, Comrie (1989)
notes that it can appear in the relative clause as a full noun (as in Hindi); it can
be substituted with a (personal) pronoun (as in Persian) or it can be retained in
the form of a relative pronoun (as in many English, Russian and many other
European languages). Dagaare relative clauses, however, belong to what
Comrie (1989: 151) labels the ‗gap–type‘, where the relative clause ―simply
does not provide any overt indication of the head‖. As example (83) and (84)
show, the bound status of the relative clause is simply indicated by preceding
the verb in the relative clause with the binding particle na. Comrie (1989)
relates the function of the noun relativised in the clause to the phenomenon of
attraction, where, in some case prominent languages, the case of the nominal
item in one clause conditions that in another clause. This parameter is not
relevant for Dagaare since it has no case system and has the gap-type of
relative clause.
However, one other parameter relevant for the Dagaare relative clause
is the identification of clause elements that can be relativized on, Comrie‘s
(1989: 155 - 160) notion of ‗accessibilty of noun phrase positions to relative
clause formation‘. The Head of the nominal group in the Dagaare relative
clause can either function as Subject or Complement in the relative clause.8 In
(83) and (84) above, for instance, the Head, Erɔdɩ and yéle, are Subjects of the
8
Complement here correspond to all participant roles associated with complement position
such as Goal, Scope, Recipient, Instrument, Phenomenon, etc. The exception is existential and
circumstantial relational clauses, where the Complement correspond to Place, typically
realised by adverbial groups (cf. Chapter 6 for details on participant roles)
102
respective relative clauses in which they occur. In the examples below, the
Head of the nominal group is Complement in the relative clause:

A dãa [[ɩ na da kʋ̀ fʋ
DEF beer 1SG REL buy. PFV give.PFV 2SG
a]], a wa?
JUNC 3PL.NHM be:where
‗The beer which I bought for you, where is it?
(86) St. Maria play
Fʋ wo =n a lɛ [[ʋ na yèl
2SG hear.PFV FOC DEF DEM 3SG REL say. PFV
a]]?
JUNC
‗You heard that which she said?‘
There is a characteristic difference between relative clauses in which the Head

is a Subject and those in which it is a Complement. When it is Subject, it is
followed by the relativiser (see (83) and (84)), but when it is Complement
such as A dãa (‗the beer‘) and alɛ (‗that‘) in (85) and (86) respectively, it is
followed by the Subject of the relative clause. On the other hand, a possessor
noun cannot be relativized in Dagaare and thus the English clause The man
whose yam you stole came here will be translated into Dagaare as (87) below:
(87) A dɔɔ nɛ nyuur [[nyɩ na zu a]],

DEF man DEM yam 2PL REL steal.PFV JUNC
wa =n ka.
come.PFV FOC here
‗That man‘s yam which you stole came here‘ (=The man whose yam
you stole came here).‘
Here, the noun representing the possessed entity, nyuur (‗yam‘), is what is
relativized. In such circumstances, a demonstrative determiner is normally
required to select the possessor as the Head of the nominal group. Thus, the
omission of the determiner nɛ in (87) will result in an absurd proposition like
103
The man‟s yam came here. Generally, circumstantial elements such as Place
and Time are normally not relativized.
In addition to the typology of relative clauses base of the three
parameters discussed above, another important distinction is that been
restrictive and non-restrictive relative clauses. While this distinction is clearly
marked in some languages, studies have shown that such a distinction is
absent in many languages (cf. Comrie 1989: 138-139), in general, and among
African languages (cf. Watters 2000: 225-228), in particular. Dagaare does not
make a formal distinction between the two types and the use of a proper noun
or common noun as Head of the nominal group is not enough criteria for
characterising a relative clause as restrictive or non-restrictive. The relative
clause in example (83) above, for instance, is restrictive and serves to single
out the Head, Erɔdɩ, among the family of Herods who ruled in several parts of
Judea. Thus, it is best translated in English as The Herod who was governor of
Galilee. The only signal of a non-restrictive relative clause are pauses before
and after the relative clause realised orthographically by commas, as in
English. An example is given in (88) below:

||| A dɩa nɩbɛ, <<na ɩ nɩ-faar || ɛ
DEF today people REL COP.PFV people-bad CONJ
bɛ tuur a Naaŋmɩn a>>, zɛlɛ

NEG.IND.NFUT follow.PFV DEF God JUNC beg.IPFV
nɩ bãwfʋ.|||
FOC knowledge
‗Todays people, who are evil and don‘t follow God, are pleading for
knowledge.‘
Restrictive relative clauses are embedded clauses while non-restrictive clauses

are hypotactic clauses serve as a logical extension of the main clause.
Like restrictive relative clauses, the nominal clauseis an embedded or
downranked clause and it typically functions as Subject or Complement in the
clause in which it occurs. In other words, it is a clause that has been
104
rankshifted to play a function typical of nominal groups. The form of the
nominal clause is like the relative clause. An example is given in (89) below:
(89) St. Maria play

Fʋ bɛ nyɛ [[ɩ na tɛr taya
2SG NEG.IND.NFUT see.PFV 1SG NMLZ possess.PFV catapult
a]]?
JUNC
‗Have you not seen that I have a catapult?‘
The category of clauses that attract nominal groups as Subjects and

Complements are limited. They typically occur as Complement/Phenomenon
inmental clauses as in (89) and as Subject in relational clauses as in (90) below
(cf. Chapter 6):
(90) Ɩ na tɛr a taya a

1SG NMLZ possess.PFV DEF catapult JUNC
bɛ vɩɛl ɛ.
NEG.IND.NFUT be:good.PFV NAFFR
‗My having the catapult is not good (=It‘s not good that I have the
catapult)‘
Adverbial clauses, on the other hand, are hypotactic clauses that

depend on a main clause for their interpretation. Adverbial clauses are quite
variant and only a few are highlighted here for illustration. One common type
has a similar form with relative and nominal clauses. An example is the clause
A Yeezu na wa ta a Filib tẽw Sezaare a (‗When Jesus got to Phillip‘s town
Ceaserea‘) in the illustration given at the beginning of this section (see
example (81)). Here, the Subject of the bound clause is followed by the
binding particle na (in this case interpretated as adverbialiser). Adverbial
clauses can however take other forms as (91) and (93) show:
(91) ||| Ʋ tõ mɛ na|| ɩ tɩ da

3SG send.PFV 1SG AFFR 1SG DIST buy. PFV
kparʋ. |||
shirt
‗S/he sent me to buy a shirt.‘
105
(92) Bible.is (Zã 10: 10b)
||| Mãa wa na|| kɛ́ a nɩsaalbɛ wa
1SG.EMP come.PFV AFFR CONJ DEF humans PROX
tʋɔ páw a nyɔ-vʋʋrʋ.|||

be:able.PFV get.PFV DEF nose-breath
‗I have come so that humankind can get life.‘
||| Fʋʋ bɛ wa lɩɛbɩ pawr a
2SG.EMP NEG.IND.NFUT COND change catch up with DEF
tew a,|| a tew na zɔ na

world JUNC DEF world POS.IND.FUT run.PFV AFFR
kɛ́ bɛr fʋ. |||

CONJ leave.PFV 2SG
‗if you don‘t change to catch up with the world, the world will run and
leave you.‘
In each clause complex in (91) and (92), the bound clause construes the
purpose for the actualisation of the main clause. In (91), there is no
morphological marker that explicitly marks it as a bound clause and it is
simply justaposed with the main clause. Its bound status is however indicated
by the fact that it does not and cannot receive unmarked or end focus, which,
in an agnate free clause, will be marked on kparʋ (shirt). In (92), in addition to
the absence of unmarked focus, the bound clause is introduced by the
conjunction kɛ́. On the other hand, in (93), the bound clause is a conditional
clause, and its conditional relationship with the main clause is
morphologically signalled by the particle wa. As a bound clause, it cannot take
the non-affirmative particle at clause final position, associated with free
negative clauses (cf. Chapter 4, Section 4.4.1.1 on non-affirmative clauses). It
is also prosodically marked by the juncture prosody marker a, and its spoken
version is realised by a level-rising tone.
3.5 Conclusion
In conclusion, this chapter has given an overview of the architecture of the
106
Dagaare language. It focused on both phonology and orthography (or
graphology) in the expression plane, and on lexicogrammar in the content
plane. On phonology, the chapter identified four phonological units, namely
tone group, [word], syllable and phoneme, and accounted for the major
phonological phenomena realised at each rank. In all, 62 phonemes have been
identified in the phonemic system, comprising twenty-nine (29) consonants,
eighteen (18) simple vowels and fifteen (15) diphthongs. It has been indicated
that a key feature relevant to all vocalic sounds is the opposition between
advanced tongue root [+ATR] and retracted tongue root [-ATR] quality. The
syllable is made up of at least one phoneme and it is the domain for the
realisation of syllabic tone and nasality. Further the [word] is composed of the
syllable and it is the point of originof harmony systems. The tone group is the
highest phonological unit and it realises intonation. Regarding orthography, it
has been indicated that Dagaare has two alphabet systems and both of which
are based on the Latin script.
On grammar, the chapter also identified four grammatical units,
namely clause, group, word and morpheme, and among these discussed the
morpheme, word and clause. Two derivational morphemes, the locative and
diminutive suffixes, were examined and a few inflectional suffixes were
identified as marking aspect in verbs and plurality in nouns. The word classes
discussed are noun, verb, adverb, pronoun and particle. Notably, the sub-
classification of nouns into count and non-count was highlighted and it has
been noted that plural marking in count nouns tends to group the nouns based
on the semantic features of humanness, kinship and social status, relational
body part nouns, collectiveness and discreteness. It has also been noted that
Dagaare has a rich sub-class of deverbal nouns. For verbs, a complex system
of aspect has been discussed, based on a range of harmony prosodies. Clauses
have been classified based on status, that is, whether they are major or minor;
and on the variable of freedom, that is whether they are free or bound. The rest
of the thesis will discuss in detail how these resources realise various
meanings in the language, especially at the rank of the clause.
107
CHAPTER FOUR
MOOD AND MODAL ASSESSMENT
4.1 Introduction
The preceding Chapter outlinedthe resources of Dagaare in terms of its

phonological, graphological and lexicogrammatical organisation. The present
chapter examines the interpersonal systems of the Dagaare clause. It focuses
on grammar as a resource for enacting roles and relationships in the speech
fellowship, in general, and for managing interaction and negotiating meaning
in verbal communication, in particular. The chapter first discusses the nature
of dialogue and then situates the semantic system of SPEECH FUNCTION within
this discussion (Section 4.2).It then proceeds to examine the interpersonal
structure of the Dagaare clause (Section 4.3). This is followed by an analysis
of the system of MOOD (Section 4.4)and finally the phenomenon of
grammatical metaphor in the MOOD system (Section 4.5).
4.2 Choices in Dialogue and the System of SPEECH FUNCTION
Since the 1970s, many discourse analysts have carried out extensive
investigations on the nature and functions of dialogue in different
communicative contexts (cf. Sacks, Schegloff & Jefferson 1974; Eggins &
Slade 1997; Matthiessen & Slade 2010). Among systemic
linguists,investigations on dialogue have been an integral part of language
description, in general, and accounts of interpersonal systems, in particular
(e.g. see Halliday 1984; Matthiessen 1995; Akerejola 2005; Teruya 2007;
Teruya et al. 2007; Bardi 2008; Kumar 2009; Quiroz 2008; Halliday &
Matthiessen 2014). The discussion in this section is based on the wealth of
knowledge scholars have produced on the nature and functions of talk-and-
interaction.
Dialogue is a universal socio-semiotic performance, and studies have
confirmed that the function of language as a resource for enacting roles and
relationships is most noticeable in dialogic interactions. When people engage
in talk, they position themselves relative to one another and to the talk itself.
108
One positioning strategy speakers adopt in dialogue is to assign different
speech roles to themselves and to other participants in the dialogue. The
primary roles conversation analysts have identified in dialogue are initiating
and responding to talk. When speakers initiate talk, they invite others to
respond, and these bidirectional roles are what carry the conversation forward.
Initiation or response has been identified in the extant literature on
conversation analysis as a move in dialogue, defined as one part of an
adjacency pair (cf. Matthiessen & Slade 2010; Slade et al. 2011).
Halliday (e.g. Halliday 1984; Halliday & Matthiessen 2014) has
identified two specific acts speakers engage in when they initiate dialogue:
giving and demanding. These two acts combine with one of two general
commodities that are traded in any speech situation, namely, information and
goods-&-services. Different combinations of the acts of giving and
demanding, on the one hand, and the commodity exchanged, on the other
hand, give four further delicate roles speakers perform in every dialogue. That
is, a speaker may give information, demand information, give goods-&-
services or demand goods-&-services.
Each of these speech roles is realised in language by the clause. The
clause is therefore the basic unit of language for enacting interpersonal
meaning in discourse; it is ‗the locus of interaction‘ (Thompson & Couper-
Kuhlen 2005: 811, original emphasis). Clauses realising demanding and
giving information are propositions, while those realising giving and
demanding goods-&-services are proposals (Halliday & Matthiessen 2014:
Ch. 4). The following dialogue, from an unscripted play, illustrates how the
different speech roles are realised by the clause in Dagaare. The interactants
comprise two customers (A & C) and a bar owner (B):
(1) St. Maria play

A(1): Fʋ nyɔw fʋ mιŋa, a zɑa dɑɑ
2SG hold.PFV 2SG self DEF yesterday pito
zɑa bɛ ι ɛ. Daι
yesterday NEG.IND.NFUT be:good.PFV NAFFR die
be =n be bι?
be.PFV FOC there INT
109
B (2): Ʋʋ-hʋ.
no
A (3): Mɩ́ bin daι! A aŋa bɛ mɩ́
HAB put.PFV die DEF DEM.PL NEG.IND.NFUT HAB
tɛr pɛr ɛ.
possess.PFV bottom NAFFR
C(4): Daι ya? Ʋ mɩ́ bin tιrιpʋl! Alɛ

die INT 3SG HAB put.PFV triple 3PL.NHM.EMP
na mɩ́ ι.
IDENT.PL HAB be:good.PFV
A (1): ‗Excuse me; yesterday‘s pito/beer was not good.‘ Is there two
days old pito/beer?‘
B (2) No!
A (3): ‗Always put two days old pito/beer. This is always useless.‘
C (3): ‗Two days old? S/he should put three days old pito/beer. That is
always good.‘9
Turn (1) has three clauses. The first is a proposal by which Speaker A
invites the bar owner to engage in talk. In the second clause, he makes a
statement (giving information) on the previous day‘s pito (a local corn beer
specific to the Dagaaba and other Mole-Dagbana communities). He follows
this with a question in the third clause (demanding information). The bar
owner, Speaker B, responds by giving information Ʋʋ-hʋ (‗No‘). Speaker A
demands a service in turn (3) and follows it with a statement (giving
information) to clarify his request. Speaker C reacts to A‘s request and
statement with three clauses, the first (i.e. Daι ya?,‗Two days old?‘)
demanding information, the next (i.e. Ʋ mɩ́ bin tιrιpʋl!, ‗S/he should always
put three days old!‘) demanding goods-&-services and the third (i.e. Alɛ na mɩ́
ι, ‗That‘s always good‘) giving information. The initiating roles and
commodity dimension of dialogue that have been demonstrated here illustrate
a general situational context for every dialogue. The semantic system that
realises this context in language is called SPEECH FUNCTION (Halliday, 1984;
9
Daι and tιrιpʋl are borrowed from the English words ‗die‘ and ‗triple‘ respectively. They are
used to refer to a two days old and three days old locally brewed corn beer called pito. They
represent the fact that the beer gets stronger in alcohol with time.
110
Halliday &Matthiessen 2014: Ch. 4). The speech functions corresponding to
the combinations of initiating roles and commodities in the extract are given in
Table 4.1.
Table 4.1. Analysis of speech roles and functions in Text 1

Speaker Turn Clause Initiation Commodity Speech function
role
1 demanding goods-&- proposal: offer
A 1 services
2 giving information proposition:
statement
3 demanding information proposition:
question
B 2 1 giving information proposition:
statement
A 3 1 demanding goods-&- proposal:
services command
statement
1 demanding information proposition:
C 4 question
2 demanding goods-&- proposal:
services command
statement
Cross-linguistically, four functions have been identified as the main

speech functions in communication (cf. Matthiessen 2004: 610-625; Teruya et
al. 2008; Matthiessen, Teruya & Wu 2008, and references therein). These are
statement, question, command and offer. As has been demonstrated in the
analysis above, giving information is realised by statements while demanding
for information is realised by questions. Command and offer correspond
respectively to demanding and giving goods-&-services. The term ‗command‘
is used broadly in this study to cover a number of related speech acts including
requests, demands, entreaties and suggestions, among others. Other umbrella
terms that have been used to cover these senses are directives and the
improvised morpheme mands (cf. Lyons 1977: 736-751; Bybee, Perkins &
111
Pagliuca 1994: 179ff). Generally, these speech acts, in the words of Lyons
(1977: 746), ―impose, or propose, some course of action or pattern of
behaviour and indicate that it should be carried out‖. Figure 4.1 presents a
network for the system of SPEECH FUNCTION.
Figure 4.1: The semantic system of SPEECH FUNCTION (Halliday &

Matthiessen 2014: 136)
As indicated earlier, this chapter will examine the lexicogrammatical

resources of the clause that Dagaare speakers use to realise the various speech
functions. The analysis here is, therefore, based mainly on dialogic texts. The
main lexicogrammatical system that corresponds to SPEECH FUNCTION is
MOOD. Thus, in this study, mood is not defined as a special verb or word form
but a grammatical system that is realisedin different ways across languages,
including but not limited to verbal morphology (cf. Bybee, Perkins & Pagliuca
1994: Ch. 6; Halliday & Matthiessen 2014: Ch. 4; Matthiessen 2004: 611-
112
635). Before MOOD is discussed, however, the following section will first
examine the interpersonal structure of the Dagaare clause.
4.3 Interpersonal Structure of the Clause

Work on different languages has revealed similarities across the interpersonal
grammar of languages (e.g. Matthiessen 2004; Teruya et al. 2007;
Matthiessen, Teruya & Wu 2008). For instance, similar functional elements,
such as Subject, Predicator, Complement and Adjunct have been identified
across languages. It has also been shown that, in every language, there are
some clausal elements that are more salient than others in realising
interpersonal meaning. On the other hand, these studies show that the
particular functional elements present in languages and their salience in the
interpersonal meaning of the clause vary cross-linguistically. For instance,
while Halliday and Matthiessen (2014) have identified the Subject and Finite
elements as most prominent in the interpersonal clause structure of English,
the Finite is absent in many languages as an element of the clause. 10There are
also differences in the particular order of the various elements in the clause
(see Matthiessen (2004) for a typological account).
In Dagaare, the general interpersonal structure of the clause is
(Negotiator) + (Subject) +(Adjunct) + Predicator + (Complement) + (Adjunct)
+ (Negotiator), where the brackets indicate optional elements. Examples (2)
and (3) are constructed to illustrate the full potential of the clause structure
while (4) to (7) illustrate other possibilities:
(2) Mɛ́ nyι zà a kɔ =n a wιɛ

HST 2PL yesterday weed.PFV FOC DEF farm
Nego. Subj. Adjunct Predicator Complement
vla wɛ?
well INT
Adjunct Nego.
‗I belive you weeded the farm well yesterday, right?
10
The system of finiteness is absent in the clause structure of many languages, as in Chinese,
Vietnamese, Thai, and also Dagaare. But if it is present, it is often reflected structurally in the
organisation of the ―verbal domain‖ (i.e. verbal group or the verb) rather than at clause rank
(Matthiessen p.c).
113
(3) Mɛ́ fʋ zà a su =n a bie
HST 2SG yesterday feed.PFV FOC DEF child
Nego. Subj. Adjunct Predicator Complement
a saab vla wɛ?
DEF book well INT
Complement Adjunct Nego.

‗I belive you fed the child the food well yesterday, right?
(4) Beyuo na kɔ na.
Beyou POS.IND.FUT weed.PFV AFFR
Subject Predicator Negotiator

‗Beyuo will weed.‘
(5) Nyι lɔb bɛr!
2PL throw.PFV leave.PFV
Subject Predicator
‗You throw (it) away!‘
(6) Na ìr ka!
EXH get up M.INS
Negotiator Predicator Negotiator

‗Please get up, I implore you!‘
(7) Ìr!
get up.PFV
Predicator
‗Get up!‘
While example (2) has one Complement, (3) has two Complements, a
characteristic of benefactive clauses (cf. Chapter 6, Sections 6.3 & 6.7.2).
Example (4) illustrates a situation where the Predicator is realised by a verbal
group complex (‗serial verb construction‘), where a single event is analysed
into phases realised by two or more verbs. Example (7) exemplifies the
minimum realisation of the interpersonal structure of the clause, where the
Predicator is the only element (cf. Section 4.4.2).The different functional
elements in the clause divide into two components, based on how prominent
they are in enacting interpersonal meaning. The most interpersonally salient
component is called the Mood base (or, simply, Mood) and it comprises the
114
Subject, Predicator and Negotiator. The other elements generally form the
Residue component of the clause. These components and the various elements
that compose them are discussed in the following sections.
4.3.1 The Mood Base

When we examine exchanges in dialogue among Dagaare speakers and
consider the questions, commands and statements in these exchanges, we
come to realize that the Subject, Predicator and Negotiator are salient in
distinguishing the grammatical realisation of speech functions. In order to
illustrate this, I will examine the flow of information in the following dialogue
between father (A) and daughter (B). In this extract, the father is seeking the
daughter‘s consent to a marriage proposal:
(8) St. Maria play
A (1): Ayɔɔ ɔ!
Ayour VOC
B (2): Baba a!
Dad VOC
A (3): Wa sɔr a bome aŋa

come.PFV count.PFV DEF things DEM
kaa!
check.PFV
|||Fʋ nyɛ a || a ŋmιn na? |||
2SG see.PFV3PL.NHM DEF how_much IDENT.PL
A lɛ [[ɩ na yèl a]], lɛ

DEF that 1SG REL say. PFV JUNC that
bɛ na ι? |||
NEG.IND.NFUT IDENT.PL NAFFR
B (4): Lɛ na.
That IDENT.PL
A (5): |||Fʋ zie na a pɑa be. |||

2SG place IDENT.PL 3PL.NHM ADV be.PFV
Fʋ saw na || kɛ ɩ de
2SG agree.PFV AFFR PROJ 1SG. take.PFV
115
a bome aŋa, Ayɔɔ? |||
DEF things DEM Ayour
B (6): Ɩ saw na.
1SG agree.PFV AFFR
A (7): |||Fʋ saw na || ɩ wob a

2SG agree.PFV AFFR 1SG pick.PFV DEF
bome ana?|||
thing.PL DEM
B (8): Ɩ saw na.

1SG agree.PFV AFFR
A(9): ||| Ayɔɔ ɔ, fʋ saw na || ɩι

Ayour VOC 2SG agree.PFV AFFR 1SG.PROJ
wob ι? |||
pick.PFV INT
B (10): Ɩ saw na.

1SG agree.PFV AFFR
A (11): ||| Fʋ saw na || ɩ de a

2SG agree.PFV AFFR 1SG take.PFV DEF
bome aŋa ƴaw a zʋbɛ pʋɔ || ɛcɛ a

things DEM put.PFV DEF pocket inside CONJ DEF
pobile ŋa kʋl fʋ? |||

gentleman DEM.PROX marry.PFV 2SG
B (12): Ʋʋ, ɩ saw na.

yes, 1SG agree.PFV AFFR
A: ‗Ayour!‘
B: ‗Dad!‘
A: ‗Come and count these things. How much is it? That [[which I
demanded]], is that not it?‘
B: ‗That is it.‘
A: ‗It is now your decision. Do you agree that I take these things, ``
Ayour?‘
B: ‗I agree.‘
A: ‗Do you agree I pick these things?‘
B: ‗I agree.‘
116
A: ‗Ayour, do you agree I pick these things?‘
B: ‗I agree.‘
A: ‗Do you agree I take these things and put them inside my pocket
and this gentleman marries you?‘
B: ‗Yes, I agree.‘
The text above is co-authored by father and daughter. The negotiation in

which they are engaged is incremental, building up turn by turn. The father
negotiates every action he undertakes in the marriage contract with the
daughter. More important to our discussion is the fact that there are particular
elements in the clauses that are carrying the negotiation forward. These
essential elements are the Subject (e.g. Fʋ, ‗you‘; and Ɩ, ‗I‘), the Predicator,
which is realized by the various verbs in the clauses (e.g. saw, ‗agree‘; wob,
‗pick‘; and kul, ‗marry‘) and finally the Negotiator element, realized by clause
final particles, notably the affirmative particle na, which punctuates many of
the clauses, particularly those in the daughter‘s turns. These three elements are
important in realizing the mood of a Dagaare clause.
One other element in a few of the clauses in the extract, specifically
those by speaker A, is the Complement. Instances are a bome aŋa (‗these
things‘) in turns (3), (5), (7) and (11). Unlike the Subject and the Negotiator,
the Complement can be assumed as taken for granted background and,
therefore, is not realised. For example, in turn (9), it is not realised in the
clause Ɩ wob ι? (‗I take?‘). This phenomenon is not an idiosyncratic instance
but rather a pervasive tendency in Dagaare, even with verbs such as de (‗take‘)
and wob (‗pick‘) that are associated with transitive clauses (see Chapter 6 on
transitivity). The implication is that the Complement is not relevant in
realising interpersonal meanings such as MOOD. Its importance lies in the role
it plays in information focus and this will be discussed in Chapter 5.
The point being made here is that in dialogic exchanges, the Subject,
Predicator and Negotiator function together to enact the clause as a move.
They are therefore normally maintained together as exchanges unfold while
other elements such as the Complement can be ellipted in the clause. Another
example of this phenomenon is given below:
117
A: Fιlιp ɓaw nι a sʋɔ a
Phillip pick.PFV FOC DEF matchet DEF
ka.
here
B: Ʋʋ, Ɩ bɛ ɓaw ɛ.
no 1SG NEG.IND.NFUT pick.PFV NAFFR
A: Fʋ ɓaw na. Nyɛ! Ʋlɛ nʋ fʋ

2SG pick.PFV AFFR see.PFV 3SG.EMP IDENT.SG 2SG
bin bɛr a fʋ puori.

put.PFV leave.PFV DEF 2SG behind.
A: ‗Phillip picked the matchet from here.‘
B: ‗No, I haven‘t picked (it).‘
A: ‗You picked (it). Look! It‘s the one you‘ve left behind you.‘
In the underlined clauses, the Subject, Predicator and Negogiator are deployed
together as the crucial elements in arguing about the polarity values of the
proposition presented in the first turn by Speaker A. They are the three most
important elements in Dagaare interpersonal grammar and, together, they form
the Mood base in the clause. The remaining elements in the clause such as
Complement and Adjunct will just be called Residue (but see Section 4.3.4.2
(a)
Ɩ nyɛrɛ na
1SG see.IPFV AFFR

Mood
‗I (can) see.‘ (The stroy of Jesus)
(b)
Ɩ wob a bome aŋa bɩ
1SG pick DEF thing.PL these INT
Subject Predicator Complement Negotiator

Residue
Mood
‗Should I pick these things?‘
Figure 4.2: The Modal structure of the Dagaare clause
118
on mood Adjuncts). The basic mood structure of the Dagaare clause is
illustrated in Figure 4.2.In example (a), the clause is made up of only the
Mood base while the clause in example (b) consists of both the Mood base and
a Residue element, realised by the Complement. It must be emphasised that
the prominence of the elements in the Mood base relative to other elements in
the clause is due to (i) the special interpersonal meaning they carry in the
clause (Section 4.4.2) and (ii) their salience in determining mood contrast
(Section 4.4). It should also be noted that, as with other systemic functional
accounts of grammar (e.g. Caffarel 2004; Teruya 2007; Halliday &
Matthiessen 2014), the notions Subject, Predicator and Negotiator are used
here as functional elements with semantic origins rather than as purely formal
categories. Before their semantic aspects are discussed in detail, however, we
will first consider the criteria for identifying them in the clause.
(i) Subject. The Subject is formally realised by a noun group or nominal
clause. In addition, its typical position is the beginning of the clause, where it
precedes the Predicator. The Subjects in the clauses in Figure 4.2 above are
both realised by the first person pronoun Ɩ (‗I‘). Figure 4.3 illustrates noun
groups in Subject position.
(a) The story of Jesus

A fʋ pɔw-yaa kpi na
DEF 2SG daughter die.PFV AFFR

nominal group verbal group
determiner pronoun noun verb particle
‗Your daughter is dead.‘
(b) The story of Jesus
Nɩbɛ yaga zie de ʋ na
people many place take.PFV 3SG AFFR

nominal group verbal nominal
group group
noun determiner adposition verb pronoun particle
‗Many people have accepted him/her‘
119
(c) The story of Jesus
A fʋ Naaŋmɩn sawfʋ sanɩ fʋ na
DEF 2SG God respond.NMLZ heal.PFV 2SG AFFR
Subject Pred. Compl. Negotiator
nominal group verbal nominal

group group
determiner pronoun noun noun verb pronoun particle
‗Your faith in God has healed you.‘

Figure 4.3: Noun group and nominal clause in Subject Position
In example (a) and (b), the nominal groups A fʋ pɔw-yaa (‗your daughter‘) and
Nɩbɛ yaga zie (‗Many people‘) function as Subjects. In example (c), the Head
of the nominal group (sawfʋ, ‗response‘) is a deverbal noun.
(ii)Predicator. The Predicator follows the Subject in the clause and is

typically realized by the verbal group (and also verbal group complexes). It is
normally marked for grammatical aspect. In many of the examples that have
been given in this section so far, the Predicator is realised by verbal groups
that consist of a single verb. Other realisations are illustrated in Figure 4.4
below:
(a) Sɛb-Sow Ƴɛr-bie 1996

A Kayɛ poru tι maalι vla
DEF Cain farms PST.REM make.PFV well
Subject Predicator Adjunct
nominal group verbal group adverbial group
determiner noun noun particle verb adverb
‗Cain‘s farms did well.‘
120
(b) Sɛb-Sow Ƴɛr-bie 1996
Nɩr za na tʋɔ na kʋ́ =m
Person all POS.IND.FUT be:able.PFV AFFR 1SG
kill. PFV
Subject Negotiator Complement
Predicator
nominal verbal group verbal nominal
group group group
particle verb particle verb pronoun
‗Any person can kill me.‘
(c)Constructed example
Nɩr za kʋ tʋɔ kʋ́ =m ɛ
Person all NEG.IND.FUT be:able.PFV 1SG NAFFR
kill. PFV
nominal group verbal group verbal nominal

group group
Noun determiner particle verb verb pronoun particle
‗Nobody can kill me.‘
Figure 4.4: Compound erbal group as Predicator
In example (a), the Predicator is realised by a verbal group, comprising the

verb maalι (‗do‘) and the tense particle tι, marking remote time. Examples (b)
and (c) are an agnate pair of positive and negative clauses respectively. In both
clauses, the Predicator is realised by a verbal group, comprising a tense-
polarity particle, a catenative verb, tʋɔ (‗be able‘), andthe main verb, kʋ
(‗kill‘). As example (b) shows, in affirmative clauses, the Negotiator element
splits the catenative verb and the main verb. In non-affirmative clauses such as
example (c), however, the Negotiator occurs in its typical position, at the end
of the clause. This phenomenon also also applies to verbal group complexes
(or ‗serial verb constructions‘). Particles associated with the Predicator are
listed in Table 4.2.
In verbless identifying clauses, the Predicator is realised by (i) either
an identifying pronoun (viz. nʋ and na) or (ii) the particle nɩ, a hybrid particle
121
combining copula and focus meanings. The predication status of both the
identifying pronouns and the hybrid particle nɩ in verbless clauses is indicated
Table 4.2. Grammatical categories that occur within the scope of the
Predicator
relevant system features realisation particle
pre- post-verb
verb
full clitic
Mood/Polarity/Tense indicative positive future na
negative future kʋ
negative non- bɛ
future
imperative: immediate ta
prohibitive non-immediate taa
Tense habitual mɩ́
remote past tɩ
Orientation duration range (‗until‘) tɩ
eventuality eventuality wa
directionality distant tɩ
proximal wa
Modal assessment conditional real wa
irrealis tɩ
modality positive naa
neutral taa
modality
negative kʋʋ
Transitivity agentive causative
instrumentality nɩ =n,=ι
comitative
by their ability to carry particles associated with the Predicator, such as tense
(9) and polarity (10) markers:
(10) Kʋɔra tɩ nʋ.

Farmer PST.REM IDENT.SG
Subject Predicator
‗He was a farmer.‘
(11) Ŋa bɛ nɩ kpɛɛ ɩ.
DEM NEG.IND.NFUT COP.FOC big NAFFR
Subj. Predicator Complement Negotiator

‗This is not the big one.‘
122
In (10), the remote tense particle scopes over the identifying pronoun nʋ
while, in (11), the particle nɩ is the locus of the negation. They therefore
display verbal characteristics. From a diachronic point of view, the identifying
pronouns nʋ (singular, neutral) na (plural, non-human) developed from a
fusion of the identifying copula verb nɛ and third person pronouns ʋ (singular,
neutral) a (plural, non-human) respectively, while nɩ developed from a further
fusion of the identifying pronouns and the attributive copula verb ɩ. This
diachronic development explains why they display verbal characteristics (see
Mwinlaaru & Yap 2017. for details; also see Section 6.5.2.1 on intensive
idenfification).
(iii) Negotiator: The Negotiator is realized by negotiation markers. As
mentioned earlier, itis realised as a juncture prosody, occurring at the
beginning or the end of the clause, that is, key points of the clause as a move.
Negotiation markers are a closed system of grammatical words and they are
listed in Table 4.10 in Section 4.5.3.
4.3.2 Meaning of Elements in the Mood Base

As mentioned above, the elements in the interpersonal structure of the clause
have semantic origins. The previous section focused on describing the formal
characteristics of the elements in the Mood base. This section will discuss the
different, although related, discourse meanings each of them contribute in
making the clause an interactive unit of grammar. The Negotiator will be
discussed first.
4.3.2.1 Negotiator
As its name suggests, it is that element which enacts the proposition or
proposal as a move. One strategy by which speakers make a proposition or
proposal negotiable is to take a subjective stance and invite the listener to
agree or disagree, confirm or contradict, challenge or respond to the stance or
invitation. The Negotiator grounds the proposition within the semantic space
open to speaker and listener as something that can be negotiated. It is also the
most salient element in showing delicate mood contrasts in the clause.
123
There are two types of negotiation particles: (a) those that are
obligatory for showing mood contrasts, specifically in indicative clauses
(example 12) and (b) those that are optional indicators of the attitude of the
speaker (example 13). Negotiator elements realised by the former category
will indicate whether the clause is a declarative or an interrogative and, in
declarative clauses, whether it is affirmative or non-affirmative (cf. Section
4.4.1). Negotiation particles that are soley attitudinal, on the other hand,
indicate the intersubjective stance of the speaker towards the proposition
and/or the listener. Illustrations are given below:
(12) St. Maria play

Host: Nι dι na wɛ?
2PL eat.PFV AFFR INT
Visitors: ʋʋ.
Host: ‗You have eaten, right?‘
Visitors: ‗Yes‘
(13) St. Maria play
Ɩ pɔw-yaa, na ƴɛrɛ ɩ zie.
1SG daughter EXH talk.IPFV 1SG place
‗My daughter, please talk to me.‘
In (12), the clause final particle wɛ mark the clause as an interrogative clause
and without it the clause will read as a declarative clause. At the same time, it
also shows that the speaker is biased towards a positive answer to the question,
thereby also encoding the speaker‘s attitude in the clause, as it were. The
clause initial particle na in (13), on the other hand, only modulates the
proposal realized by the imperative clause by exhorting the addressee. Unlike
the interrogative particle, it is not needed to mark the clause as an imperative.
In summary, the definition of the Negotiator element can be
summarised from what Halliday (e.g. Halliday, 1996; Halliday & Matthiessen,
2014) calls a trinocular perspective: (i) from below the clause, it is realised by
negotiation particles; (ii) around the clause, it occurs at the end or towards the
end of the clause and combines with the Subject and Predicator to show mood
contrast; (iii) semantically, it enacts a proposition as a negotiable unit either by
124
grounding the clause within the semantic space of speech function and/or by
marking the intersubjective stance of the speaker in relation to the proposition
or the listener. Negotiation will be discussed in detail in Section 4.5.3.
4.3.2.2 Predicator
This section proceeds to consider the Predicator. The Predicator is the locus of
the predicated clause, that which is stated, offered, demanded or questioned
about the Subject. As mentioned earlier, it is realised by a verbal group, and,
in verbless clauses, by an identifying pronoun (i.e. nʋ or na) or the hybrid
particle nɩ and contributes to establishing the validity and arguability of the
proposition. The specific function of the Predicator is that it is the domain for
the realisation of six experiential (i-vi) and three interpersonal (vi-viii)
meanings in the clause (Table 4.2):
i. It specifies the process that is predicated of the Subject; whether it is,

for example, an activity, a happening, a saying, thought or being (cf.
Chapter 6).
ii. It specifies the aspect of the Process, realised by a verb, whether it is
bounded as a unified event (i.e. perfective) or unbounded as a fluid,
incomplete event (i.e. imperfective) (cf. Chapter 3, Section 3.4.2.2).
iii. It specifies the temporal orientation of the speaker towards the speech
event, whether the event is regarded as an eventuality or ranged over
time (i.e. durative).
iv. It indicates directionality, whether the process is viewed as distant or
proximal to the speaker.
v. It indicates the agentivity (i.e. causation, instrumentality & comitation)
of the process; specifying, for instance, whether a participant is an
Initiator/Agent, Instrument or an Accompaniment in carrying out the
process (cf. Chapter 6, Section 6.3.2).
vi. It specifies the time of the event relative to the time of speech (i.e.
tense), including a primary tense distinction between future and non-
future and a secondary tense distinction between habitual and remote
past.
125
vii. It indicates the primary mood of the clause, whether it is indicative or
imperative (cf. Section 4.4).
viii. It specifies the polarity of the clause, whether it is positive or negative
(cf. Section 4.5.1).
ix. It indicates the speaker‘s assessment of the speech event through: (a)
modal probability and desirability (cf. Section 4.5.2); and (b) by stating
the contingencies for enforcing the event, whether it is real or unreal
(i.e. conditionality).
Whereas points (i) and (ii), on process type and aspect respectively, are
realised by the verb(s) in the verbal group realising the Predicator, the other
meanings are realised by particles in pre-verb and post-verb positions. These
particles are presented in Table 4.2 in Section 4.4.1 above. Also note that, as
indicated in Table 4.2, the realisation of polarity, primary mood (i.e indicative
versus imperative) and tense intersect in Dagaare, with the exception of
remote past and habitual tenses. Thus, points (v), (vii) and (viii) are
simultaneously realised by the same particles. MOOD is discussed in Section
4.4 while POLARITY and MODALITY are discussed in Section 4.5.
The picture presented by the various grammatical meanings realised by
the Predicator is that it carries much of the burden of making the proposition
valid and arguable. It brings the proposition down to earth by relating it to
accessible state of affairs or circumstances in the speech context. Listeners,
therefore, can evaluate these circumstances and confirm, deny, interrogate or
challenge the proposition. For instance, through tense, the speaker anchors the
proposition in a specific temporal circumstance which can be verified by
listeners to be valid or not.
It should be noted that interpreting the Predicator in terms of validity,
as has been done here, is not the same as interpreting it in terms of truth
conditions. As Halliday and Matthiessen (2014) note, interpreting the meaning
of the Predicator in terms of truth conditions will limit its interpersonal power
by restricting it to propositions (i.e statements and questions). In proposals, the
validity function of the Predicator is reinterpreted as indicating the efficacy of
the command or offer. Thus, if its meaning is restricted to truth value, as has
126
often been done in formal semantics, the speech functions of commands and
offers will be excluded.
In summary, the Predicator element is defined from a trinocular
perspective as follows: (i) From below, it is typically realised by the verbal
group and it is marked for aspect; (ii) within the clause, it combines with the
Subject and/orthe Negotiator in showing mood contrast; (iii) from above, in
the semantics, it contributes to the validity of a proposition by making it an
arguable unit of discourse. In commands and offers, it carries the efficacy of
the proposal.
4.3.2.3 Subject
The Subject is also an important interpersonal element of the clause. Together
with the Predicator and Negotiator it enacts the interpersonal meaning of the
clause as a move. It is the entity ―by reference to which the proposition can be
affirmed or denied‖ (Halliday & Matthiessen 2014:145). For example, in the
clause Ɩ saw na (‗I agree‘), while the Negotiator na affirms the proposition, the
Subject, realised by the pronoun Ɩ (‗I‘), shows the entity in respect to which
the affirmation is claimed to be valid. In other words, it is the speaker (Ɩ, ‗I‘)
who is modally responsible for the success of the proposition being affirmed.
She rests the speech event (saw, ‗agree‘) on herself (Ɩ, ‗I‘) and affirms this
(na).
There are several ways by which the role of the Subject as the modally
responsible entity in the clause can be recognised. One way is to examine the
use of personal pronouns in dialogue. In the dialogue between father and
daughter in extract (2), the Subject pronouns fʋ (‗you‘) and Ɩ (‗I‘), for instance,
serve as interpersonal deixis by which the interlocutors shift and negotiate the
modal responsibility assigned by the predication, from distant viewpoint (fʋ,
‗you‘) to proximal viewpoint Ɩ (‗I‘) (cf. Halliday & Matthiessen 2014: 144,
146). While the father assigns the modal responsibility of the proposition (in
this case, the agreement) to the daughter by use of the second person (fʋ,
‗you‘), the daughter claims this responsibility for herself (Ɩ, ‗I‘).
The second and perhaps the most noticeable way is to examine offers
and commands in discourse. The Subject is the entity that is recognised as the
127
one responsible for the execution of the proposal. The following
commandments from a Bible translation to illustrate this point:

Fʋ kʋ por a ɩ yo-sow
2SG NEG.IND.FUT mention.PFV DEF 1SG name-holy
zawla ι.
vain NAFFR
‗You shall not mention my holy name in vain.‘

Nyι ta dɔwrɛ a nι-saabɛ [[na
2PL NEG.IMP bully-IPFV DEF person-stranger.PL REL
be a nyι tew pʋɔ a ι!]]

be.PFV DEF 2PL town inside JUNC NAFFR
‗You should not maltreat the strangers that are in your town.‘
In the above clauses, the one responsible for the success of the command
realised in example (14) and in example (15) is the respective Subjects in the
clauses: fʋ (‗you‘, singular) and nyι (‗you‘, plural). These two elements
specify the one who is responsible for executing the command.
Apart from these semantic oriented properties of the Subject, some
grammatical mechanisms have been developed in the linguistic literature for
identifying more cryptic properties of the Subject within and across languages.
These include, among others, its accessibility to relativisation, the ability to
trigger reflexives and whether it can be presumed in continuing clauses in
paratactic sequences. These mechanisms imply a hierarchy of noun groups in
the clause and the Subject in a particular language is the nominal group that
occupies the highest rank in this hierarchy based on the grammatical
properties of the language in question (cf. Comrie 1989: Ch. 5; Creissels 2000:
232-236). Keenan and Comrie (1977; 1979), for instance, indentified a
hierarchy among elements in the clause based on their accessibility to
relativisation and noted a universal tendency where other elements in the
clause such as the Complement can only be relativized if the Subject can also
be relativised. In some languages such as Malagasy (Austronesian: Malayo-
128
Polynesian) only the Subject is accessible to relativisation (Comrie 1989:
156). As noted in Chapter 3 (cf. Section 3.4.3), in Dagaare, both the Subject
and Complement can be relativised. Thus, on this parameter, both occupy the
highest hierarchy among the elements in the clause. On the second parameter
however, only the Subject in Dagaare triggers reflexives. Thus, in the
following imperative clause, where there is no overt Subject, speakers of the
language will understand that the reflexive nominal group fʋ tʋɔra refers tothe
implicit Subject, the addressee, rather than a second party.

Dɔkta a, sanɩ fʋ tʋɔra!
doctor JUNC heal.PFV 2SG self
‗Doctor, heal yourself!‘
Also, only the Subject can be presumed in paractactic sequences. In other

words, as in English and many other languages, when two clauses, namely a
transitive clause (17) and an intransitive clause (18), are coordinated, it is only
the Subject that can be ellipted as an implicit element in the second clause (19)
(cf. Comrie 1989: 112).
(17) A pobile ŋmɛ =n a bie.

DEF young man beat.PFV FOC DEF child
‗The young man beat the child.‘
(18) A pobile zɔ na.
DEF young man run.PFV AFFR
‗The young man ran.‘

(19) ||| A pobile ŋmɛ =n a bie ||
DEF young man beat.PFV FOC DEF child
ɛ zɔ.|||
CONJ run.PFV
‗The young man beat the child and ran.‘
Thus, in (19), the addressee understands that it is the Subject of the first
clause, a pobile (‗young man‘) and not the Complement, a bie (‗the child‘),
who did the running construed by the second clause. As Comrie (1989: 112)
129
notes, in some languages such as Yidiny (Pama-Nyungan: Yidiny), the second
clause would be understood as ‗and the child ran‘. The implication of the
Dagaare example is that the Subjectis the nominal group with the highest
degree of bond with the Predicator and the element with the highest ability to
control cryptic mechanisms in the clause.
The privileged status of the Subject in the Dagaare clause is further
shown by the fact that, except in unmarked imperative clauses where the
addressee is a single interactant (see example (16)), the Subject must always
be present in every clause. In other words, typically, the Complement cannot
occur in a clause without the Subject. Thus, elliptical constructions in English
such as Cutting it for I‟m cutting it and Cutting it? for Are you cutting it? are
not possible in Dagaare. On the other hand, I‟m cutting and Are you cutting,
where the Complement is ellipted, are perfectly acceptable. As in (20), the
Subject can be a ―dummy‖ one, realised by the third person non-human
pronoun a (equivalent to empty-it in English), particularly with verbs of
necessity such as fɛr and sɛw:

A fɛr =a [...].
3PL.NHM be:necessary AFFR [...]
‗It is necessary ….‘
Here, the dummy Subject together with the Predicator enacts the proposition
as a need, a form of (impersonal) modulation. Modal responsibility is assigned
to a kind of ‗grammatical effigy‘, as it were; something not out there in world
of experience. This dummy Subject fulfils the grammatical requirement that
every Dagaare clause must take a Subject, except for an imperative clause
whose implied Subject is a single interactant (cf. Section 4.4.2). In short, the
point made here is that, while Dagaare lacks agreement systems between
between Subject and Finite, as in English and other Germanic languages, or
between Subject and pronominal prefixes in the verbal group, as in some
African languages such as those of the Bantu family (cf. Watters 2000: 201-
202; Matthiessen 2004), there are still cryptic properties of the Subject that
distinguishes it as a salient interpersonal element that is set apart from other
130
nominal groups. It can be defined from a trinocular perspective as follows: (i)
from below the clause, it is realised by a nominal group; (ii) from a
roundabout perspective, it occurs at the beginning of the clause and combines
with the Predicator and Negotiator to form the Mood base of the clause; (iii)
from above, that is, semantically, it is the nub of the argument in the
proposition or the modally responsible entity in both propositions and
proposals.
4.3.3 Function of the Mood Base

We now summarise the function of the Mood base as a unified block in the
interpersonal structure of the clause. From the discussion so far, it is clear that
the Dagaare clause has got a distinct interpersonal identity. In this
interpersonal structure, the Mood base is an essential semantic component. It
carries the burden of the clause as an interactive unit of exchange in three
ways: (i) by indicating the modal responsibility of proposals or of the verbal
event in a proposition through the Subject; (ii) by characterising the clause as
an arguable unit through the combination of Subject and Predicator; and (iii)
by specifying the negotiation value of the clause through the Negotiator
element (if any).
When we consider these meanings from above in the semantic stratum,
we observe that they are prosodic. That is, whether any of them is marked at
the beginning, medial or final position of the clause, it is diffused throughout
the clause to give it an interpersonal character. However, they can also be
considered from a roundabout perspective to see how the various elements in
the prosody of the clause relate to one another in order to realise these
meanings. From this point of view, it can be observed that, in Dagaare, both
the initial and the final positions of the clause are interpersonally prominent,
with the Subject and Predicator occurring at the beginning of the clause, and
the Negotiator at or towards the end of the clause, where the speaker is
potentially about to hand over the turn to the other interlocutor. Another
function of the Mood base is that it indicates mood contrast in the clause.
MOOD will be discussed in detail and illustrated in Section 4.4.
131
4.3.4 Other Elements in the Interpersonal Structure of the Clause
As mentioned earlier, the interpersonal structure of the clause is made up of
the Mood base and a Residue element. This section will discuss the realisation
and meaning of the elements in the Residue, namely, Complement and
Adjunct. The section will also discuss the form and function of Vocatives.
4.3.4.1 Complement
Similar to the Subject, the Complement is realised by nominal groups. These
different realisations are illustrated below in Figure 4.5:
(a) Noun as Complement

Bɛ wa =n nɩ pɔlɩ-kɩlɩnɩk
3PL.HM come.PFV CAUS FOC poly-clinic
Subject Predicator Complement
pronoun verb particle particle noun
‗They brought a poly-clinic.‘ (Political opinion interview)
(b) Adjectival noun as Complement

A tew tι ι =n vuo
DEF earth PST.REM COP FOC hollow
determiner noun particle verb particle noun
‗The earth was hollow.‘(Sɛb-Sow Ƴɛr-bie 1996)
(c) Noun group as Complement

Bɛ wa =n nɩ sukuul yaga
3PL.HM come CAUS FOC school many
pronoun verb particle particle noun determiner
‗They brought many schools.‘ (Political opinion interview)
(d) Pronoun as Complement

Ɩ nyɛ ʋ na
1SG see.PF 3SG AFFR

pronoun verb pronoun particle
‗I‘ve seen him.‘ (The story of Jesus)
Figure 4.5: Realisations of Complement
132
As in many other languages, clauses such as benefactive clauses can have up
to two Complements:
Bɛ kʋ̀ tι nι dιsιrιt
3PL.HM give.PFV 1PL FOC district
Subject Predicator Complement Complement
nominal group verbal group nominal group nominal group
pronoun verb pronoun particle noun
‗They gave us a district.‘ (Political opinion interview)
Figure 4.6: Example of two Complements in one clause
In addition, as mentioned earlier (see Section 4.4.1), the Complement is not an

interpersonally prominent element in the Dagaare clause and is part of the
Residue component of the clause structure. Unlike in English, it has no
potential of being elevated to the status of Subject in the clause in which it
occurs since Dagaare does not have the grammatical system of VOICE (but see
Chapter 6; Section 6.7.2.1 on middle clauses). However, when the
Complement is selected as marked Focus in the clause, it is placed clause
initially as the Subject of a thematic equative clause (see Figure 4.7). This
phenomenon is discussed in detail in Chapter 5 (cf. Section 5.4.2.1 on
thematic equatives).
Dιsιrιt nʋ bɛ kʋ̀ tι.

district IDENT.SG 3PL.HM give.PFV 1PL
Subject Predicator Subject Predicator Complement
nominal nominal nominal verbal nominal
group group group group group
noun ident. pronoun verb pronoun
pronoun
‗It is a district they gave us.‘
Figure 4.7: Complement as marked Theme
Thus, the Complement in the Dagaare clause is a textually prominent element.

In the unmarked case, it is normally placed at the end of the clause as
potentially new information. It is considered key in setting the background of
the discourse in any exchange. However, once the Complement becomes
given and established as a taken for granted background, speakers have the
133
choice of ellipting it in subsequent clauses to give prominence to negotiation
at the end of the clause (but see Chapter 6, Section 6.7.1 for exceptions). We
will discuss this phenomenon in detail in Chapters 5 and 6.
4.3.4.2 Adjunct
The Adjunct is realised by adverbial groups (including ideophones) and mood
particles (cf. Chapter 3, Section 3.4.2.3). Three types of Adjuncts can be
identified in the clause, based on their metafunctional meaning. These are
interpersonal Adjuncts, experiential Adjuncts, and textual Adjuncts. These are
illustrated in Figure 4.8 below:
(a) Experiential Adjunct of manner

A cere nι galmɛ
3PL.NHM go.IPFV FOC zigzag
nominal group verbal group adverbal group
pronoun verb particle adverb
‗It is going zigzag.‘ (Political opinion interview 2)
(b) Experiential Adjunct (of degree)
A nʋmɛ nι paa
3PL.NHM be.sweet.IPFV FOC very
pronoun verb particle adverb
‗It is very sweet.‘ (St. Maria play)
(c) Textual Adjunct (continuative)
Alɛ tι na ír a
then 1PL POS.FUT get up AFFR
Adjunct Subject Predicator Negotiator
adverbial group nominal group verbal group
adverb pronoun particle verb particle
‗Then, we will get up (= Well, we ask permission to leave).‘ (St. Maria play)
(d) Interpersonal Adjunct (of attitude)
Yél -mιŋa na ɛ̃
truth IDENT.SG ADV
Subject Predicator (mood) Adjunct

nominal group nominal group adverbial group
pronoun ident. pronoun adverbial particle
‗It is true anyway.‘ (St. Maria play)
Figure 4.8: Types of Adjuncts by metafunction
134
While circumstantial and textual Adjuncts function as circumstances in the
ideational structure of the clause, mood Adjuncts supplement elements in the
Mood base in realising negotiation and arguability in the clause.
Circumstantial Adjuncts are realised by adverbials of time, place, manner,
degree, etc. Textual Adjuncts are realised adverbial conjunctions such as alɛ, a
lɛ na (‗then‘), a puori, (‗afterwards‘), al puori (‗after that‘), tɔɔ (‗well‘), pãa
(‗then‘ ‗now‘) and mɩ̀ (‗also‘)(cf. Chapter 3, Section 3.4.2.3 on adverbs)
Interpersonal or mood Adjuncts are realised by two sub-classes of
adverbs. The first is the two synonymous modal adverbs kaapaw and a mɩnɛ
kàw (‗maybe‘, ‗probably‘):

Kaapaw, bɛ mɩ́ cen na tɩ zanɩ
maybe 3PL.HM HAB go.PFV AFFR DIST learn.PFV
bãwfʋ kàw zie kàw.
knowledge some place some
‗Maybe, they go to learn some knowledge somewhere.‘
The second group is modal adverbial particles such as ɛ in Figure 4.9 (example
d) and also nɔŋ (‗maybe‘, probably), yaa (‗even‘) and tɔɔ (‗well‘). Most of the
adverbial particles listed in Chapter 3 belong to this category (cf. Section
3.4.2.3).The difference between adverbial particles and attitude-oriented
negotiators is that the adverbial particles are versatile and are relatively
mobile, as example (22) shows:
(22a) Ɩ ende cere na.

1SG.EMP MOD go.IPFV AFFR
‗I am going, anyway.‘
(22b) Ɩ cere na ende.
1SG.EMP go.IPFV AFFR MOD
‗I am going, anyway.‘
The boundary between the two is, however, fuzzy rather than discrete and
particles such as yaa and mɔ̀ are identified among negotiation markers and
adverbial particles. Like attitude-oriented negotiators, mood Adjuncts indicate
135
the attitude of the speaker in the clause. Let‘s examine the use of mɔ̀ as a
mood Adjunct in the following extract, taken from a report by a female farmer
at a workshop:
(23) Workshop report 7

||| Soo a sɔn 'wɔb a a tɩ bɛw-kalsɩrɩ
so DEF rabbits chew AFFR DEF 1PL beans
baar a za kpɛlɛ. || Soo, a lɛ na
finish DEF all complete so DEF DEM IDENT.PL
tɩ bɔbr || ɛ sɛ mɔ̃̀ tɩ yèl ||

1PL want.PFV PROJ PROJ ADV 1PL say. PFV
mɔ̃̀ tɩɩ wa tɩ kʋɔr a al a, ||
ADV 1PL.EMP EVT 1PL farm.IPFV DEF DEM JUNC
tɩɩ ɩ mɔ̃̀ || tɩ ɩ a nɩ dabaala

1PL do.PFV ADV 1PL do.PFV 3PL.NHM FOC something
kàw. ||| Tɩɩ kàw ƴaw […]. |||
some medicine some put.PFV
‗So the rabbits completely chewed all our beans. So that is why we
want like to say like when we are cultivating those ones, we should
like do something about them. Some medicine should be put (in the
plants).‘
The preponderance of the modal adverb mɔ̀ in the extract enacts the report as a
highly modulated text. This reflects the tenor relationship between a rural
female farmer and the group of male agricultural experts she is addressing in a
male dominanted society.
4.3.4.3 Vocative
The last element of the interpersonal structure of the clause we will consider is
the Vocative. Unlike Complements and Adjuncts, the Vocative is peripheral to
the internal unity of the clause of which the Predicator is the nucleus (see also
Section 4.4.5.2 on minor clauses). Nonetheless, it is an important interpersonal
element.) It ―reinforce[s] the ‗you-and-me‘ dimension‖ of the exchange
(Halliday & Matthiessen: 2014: 160). In the Dagaare clause, the Vocative is
136
realised by nominal groups, including proper names (e.g. Ayɔɔ), nicknames,
kinship terms (e.g. baba, ‗dad‘; mama, ‗mum‘, saakʋm, ‗grandpa‘; makʋm,
‗grandma‘; saabile, ‗younger father‘/‗uncle‘; mabile, ‗younger mother‘/‗aunt‘;
pʋre, ‗aunt‘; ι dιɛm, ‗my in-law‘; yentaa, ‗rival‘ etc.), honorific titles (e.g.
Chief); descriptive terms (e.g. pɔwle, ‗young lady‘; pɔwsιra, ‗young woman‘;
pobile, ‗gentleman‘; dɔɔ, ‗man‘; pɔw, ‗woman‘; Seli ma, ‗Celestina‘s mother‘,
etc.), and emphatic pronouns (i.e. fʋʋ, singular; nyιιm, plural).
Some kinship or intimate familial terms used as Vocatives (or address
terms) may not necessarily reflect the kind of relationship between the speaker
and the addressee. They may just be used to show culturally conditioned
respect and endearment. For instance, a woman can address the bothers-in-
law, cousins-in-law and nephews-in-law as ι sιr (‗my husband‘) and, in turn,
receive the address ι pɔw (‗my wife‘). Also, a married woman and a male
friend or her husband‘s close friend will normally address each other as ι sɛn
(‗my boyfriend/girlfriend‘), while children address their paternal uncles as
Baba (‗Dad‘). This situation shows the high intimacy among members in
traditional Dagara communities. As Figure 4.9 shows, Vocatives can be
reinforced with negotiation through final vowel lengthening (see example 24)
or vowel insertion (see example 25).
Figure 4.9: A system network for negotiation in Vocatives
Examples of Vocatives are illustrated in the following extracts (Vocatives are

in bold):
(24) St. Maria play

Ayɔɔ ɔ, fʋ saw na ɩ wob ι?
Ayour VOC 2SG agree.PFV AFFR 1SG pick.PFV INT
137
‗Ayour, do you agree (that) I pick these things?‘
(25) The stroy of Jesus
Bibiir i, nyι wa yi yow
children VOC 2PL.EMP come.PFV go:out outside
fɔŋ lɛ!
quickly ADV
‗Children, you should go out very quickly!‘
Negotiation in Vocatives is an amplification strategy. However, to the extent

that Vocatives may be negotiated shows the pervasiveness of negotiation in
the interpersonal grammar of Dagaare. The importance of Vocatives lies in
building social roles and relationships in dialogue. As mentioned earlier, they
are outside the internal structure of the clause, composed of the Mood base
and Residue.
4.4 MOOD as a System

In Section 4.2, we discussed the semantic system of SPEECH FUNCTION and
identified that the main speech functions realised by the clause are statement,
question, offer and command. We also noted that statements and questions are
respectively oriented towards giving and demanding information (i.e.
propositions) while offer and command are respectively oriented towards
giving and demanding goods-&-services (i.e. proposals). In the preceding
section (Section 4.3), we noted that the core elements of the Mood base,
namely Subject, Predicator and Negotiator, are essential in determining the
particular speech function the clause is, by default, oriented towards. In other
words, these three elements generally determine mood contrast in the clause.
In this Section, we will discuss the mood contrasts in Dagaare in detail.
Although studies have revealed similar mood contrasts cross-linguistically,
there is often language specific variation as the contrasts extent in delicacy as
well as in the specific lexicogrammatical realisations of various mood types
(Matthiessen 2004; Teruya et al. 2007; Matthiessen, Teruya & Wu 2008). The
system network in Figure 4.11 shows the different mood types in the Dagaare
clause. As in other languages, the primary distinction is between indicative
and imperative.
138
These primary mood types are distinguished by the mood elements that
are obligatory in their basic realisation (see Figure 4.11). While the Subject
and Predicator are obligatory in every indicative clause, it is only the
Predicator that is obligatory in every imperative clause. In other words, every
indicative clause must have at least a Subject and a Predicator and every
imperative clause must have at least the Predicator element. The presence or
absence of the Negotiator in the indicative type and the specific particle that
realises it will show delicate mood contrasts. On the other hand, the presence
or absence of Subject and Negotiator in the imperative mood type and their
lexicogrammatical realisations will also show finer imperative mood
distinctions. It is important to reiterate that each mood type is defined from a
trinocular perspective: (i) from above by the speech function it typically
realises; (ii) from below the clause by its lexico-grammatical realisation and
(iii) from roundabout by how the elements in the clause relate with one
another. We will start our discussion with the indicative mood type. Delicate
distinctions in the imperative mood type are given in Section 4.4.
4.4.1 Indicative
As mentioned above, the indicative mood by default realises the speech
functions of statements and questions; that is, propositions. All indicative
clauses take a Subject and a Predicator, while the presence or absence of the
Negotiator element depends on the particular type of the indicative clause (cf.
Figure 4.10). The secondary distinction in the indicative mood is between
declarative (Section 4.4.1.1) and interrogative (Section 4.4.1.2) clauses. More
delicate contrasts in these two types are shown in Figure 4.10.
4.4.1.1 Declarative Clauses
As its name suggests, the declarative mood typically realises the general
speech function of giving information (i.e. statements). This mood type has
two further contrasts, namely the affirmative and non-affirmative clauses. The
distinction among these delicate mood types is indicated by the different
particles that realise the Negotiator element.
139
(i) Affirmative clauses
The affirmative clause is a declarative mood type that asserts the proposition
given by the speaker. It can be identified structurally as typically consisting of
Subject ^ Predicator ^ (Negotiator). The Negotiator element in the affirmative
clause is realised by the particle na, which occurs towards the end of the
clause. An example of the affirmative clause is given in the exchange below
(relevant Negotiator in bold):
(26) St. Maria play

A: Fʋ wõ =n a lɛ [[ʋ na
2SG hear.PFV FOC DEF DEM she REL
yèl a]]?
say. PFV JUNC
B: Ɩ wõ a na.
1SG hear 3PL.NHM AFFR
A: ‗You heard THAT WHICH SHE SAID (= ‗You heard WHAT SHE SAID?)‘
B: ‗I heard it.‘
140
Figure 4.10: A system network of MOOD
141
Here, Speaker B asserts the proposition realised by the clause by ending the
utterance with the affirmative particle na. It must be noted that the Subject and
Predicator elements combine with the Negotiator na to characterize the clause
as affirmative. It should be noted also that Speaker A‘s utterance is an indirect
question realised by an affirmative clause (cf. Section 4.4.4 on mood
metaphor). The absence of the affirmative particle, is however, conditioned by
the presence of end focus. In other words, when there is a focused
Complement or Adjunct in the clause, the affirmative particle does not occur.
The underlined clauses in (27) below further illustrate this phenomenon:
(27) St. Maria play

Father: Pɑa mɩ́ bɛr ɛ cɛlɛ a
now also stop.PFV CONJ listen.IPFV DEF
fʋ yéle !
2SG matter.
Son: Ʋʋ?
‗yes‘
Father: Fʋ mɩ́ ta na.
2SG also reach.PFV AFFR
Mother: Fʋ mɩ́ ta =n pɔw kulu.

2SG also reach.PFV FOC wife marry.NMLZ
Father: ‗Now, you stop (talking) too and listen to you matter.‘
Son: ‗Yes?‘
Father: ‗You too have reached.‘
Mother: ‗You too have reached MARRIAGE.‘
In the first underlined clause, the Complement is not realised and thus the
Negotiator is realised by the affirmative particle, na. In the last clause,
however, the Complement is realised, pɔw kulu (‗marrying a wife‘) and it is
focused by the enclitic from of the focus particle nι and, correspondingly, the
Negotiator is not realised. Mood contrast between the affirmative and non-
affirmative clause is, however, still maintained. The absence of the Negotiator
element is compensated for by the fact that the focus particle, nι, does not
142
occur in non-affirmative clauses and thus serves as a distinctive feature of the
affirmative clause. This phenomenon will be discussed further in Chapter 5
(cf. Sections 5.4.1 & 5.4.4).
(ii) Non-affirmative clauses
The non-affirmative clause contrasts with the affirmative clause. It shows that
the proposition it realises is non-assertive. Syntactically, it is realised by the
structure Subject ^ Predicator ^ Negotiator. Thus, unlike in the affirmative
clause, the Negotiator element is always present in the non-affirmative clause.
Characterised from below on the rank scale, the Negotiator is realised by one
of the following particles: e, ɛ, ι, depending on the advanced tongue root
[ATR] value of the preceding word. One dialectal difference between the Lobr
and Ngmere (‗Central Dagaare‘) dialects is that the Negotiator is not realised
in non-affirmative clauses in the Ngmere dialect. This difference is illustrated
by the following dialectal pair of clauses (Negotiator in bold):
(28a) Junior High School Book 3 (Central Dagaare)
O da ba nɔŋ a nembɛrɛ
3SG PST.REM NEG.IND.NFUT like.PFV DEF elders
yéle .
matter
‗S/he didn‘t like the elders.‘
(28b) author‘s translation (Lobr)
Ʋ bɛ tɩ nɔw a nɩ-bɛrɛ .
3SG NEG.IND.NFUT PST.REM like.PFV DEF elders
yéle ι
matter NAFFR
‗S/he didn‘t like the elders.‘
As indicated above, the selection of any of the non-affirmative particles

depends on the principle of [ATR] vowel harmony, based on the word that
immediately precedes the Negotiator (see Table 4.3). This implies that the
particle always forms phonological word with the adjacent word (cf. Chapter
3, Section 3.2.3). Grammatically, however, it is neither attached to the
143
preceding word as an enclitic nor to any word class such as the verb as an
affix. It always occurs as the final item of the clause irrespective of the class of
the preceding word. In (29) below, for instance, it occurs after a proper noun:
(29) Ʋ bɛ tɩ nɔw Peter ι.

3SG NEG.IND.NFUT PST.REM like.PFV Peter NAFFR
‗S/he didn‘t like Peter.‘
Table 4.3. Realisations of the Negotiator in non-affirmative clauses

word preceding particle realising Negotiator
Negotiator
final Ending [-ATR] [+ATR]
syllable with
phonetic orthographic phonetic orthographic
form form form form
plus /ɛ/ ɛ /e/ e
coda
(closed)
/ʊ/, /aʊ/, /ɛɪ/ ɛ
/ɪ/
minus /e/, /o/ /i/ ɩ
coda
(open) other /ɪ/ ɩ /ei/ e
vowels
The discussion proceeds to account for the specific phonological

environments that condition the different realisations of the non-affirmative
particle, as presented in Table 4.3. First, when the vowel of the immediately
preceding word is [-ATR] vowel, the Negotiator is realised by ɛ or ι.
Generally, ι occurs in instances where the preceding word ends with an open
syllable. Illustrations are given below (Negotiator is in bold):

Ɩ bɛ tɛr tɔ ι.
1SG NEG.IND.NFUT possess.PFV equal NAFFR
‗I don‘t have an equal.‘

Ʋ bɛ tι dι bom za ι.
3SG NEG.IND.NFUT PST.REM eat.PFV thing all NAFFR
‗S/he had not eaten anything.‘

144
(32) St. Maria play
Ʋ bɛ ƴawnɛ nιbɛ ι.
3SG NEG.IND.NFUT respect.IPFV people NAFFR
‗S/he doesn‘t respect people.‘
Phonologically, ɛ has two realisations: [ɛ] and [ɛɪ]. When the word preceding
the Negotiator ends with a closed syllable, it generally attracts [ɛ]:

Nyι kʋ bãw a pɛr ɛ.
2PL NEG.IND.FUT know.PFV DEF bottom NAFFR
‗You won‘t know the meaning.‘
(34) St. Maria Play
Fʋ bɛ tɛr pɛr ɛ.
2SG NEG.IND.NFUT possess bottom NAFFR
‗You are useless.‘
A ʋ naalʋ kʋ tɛr baarʋ
DEF 3SG kingship NEG.IND.FUT possess finishing
daar ɛ.
time NAFFR
‗His kingship will have no end.‘
On the other hand, for words with open syllables in which the final vowel is
/ʊ/, /ɑʊ/ or /ɪ/, the Negotiator is realised as [ɛɪ] but represented
orthographically as ɛ,as the following examples show:

A Farazιɛ bɛ nʋ ɛ.
DEF Pharisee NEG.IND.NFUT IDENT.SG NAFFR
‗It is not the Pharisee.‘

(37) St. Maria play
Ɩ bɛ bãw ɛ.
1SG NEG.IND.NFUT know NAFFR
‗I don‘t know.‘
145
(38) St. Maria play
Bɛ kʋ dι ɛ.
3PL.HM NEG.IND.FUT eat.PFV NAFFR
‗They won‘t eat.‘
Similarly, three choices are open to non-affirmative clauses with

lexical items ending with [+ATR] vowels. Here, when the preceding word
ends with a closed syllable, the Negotiator is realised by the particle e:
Ɩ bɛ kpɛ mimir e.
1SG NEG.IND.NFUT enter eye NAFFR
‗I am not mean.‘
Bɛ mὶ kʋ bãw a nyι dem
3PL.HM also NEG.IND.FUT know.PFV DEF 2PL own
e.
NAFFR
‗They will also not remember yours (your sins).‘

Nyι kʋ nyɛ bom e.

2PL NEG.IND.FUT see.PFV thing NAFFR
You will not see anything.
In non-affirmative clauses with final words generally ending with an open

syllable, the Negotiator is realised as [ei], but also represented
orthographically as e:

Bɛ bɛ tι zɔrɛ taar
3PL.HM NEG.IND.NFUT PST.REM fear.IPFV each other
vi e.
shy NAFFR
‗They were not shy of each other.‘
146
Fʋ bɛ cir kaa nyuu nyaw a ɩ
2SG NEG.IND.NFUT pour.PFV oil scent put DEF 1SG
zu e.
head NAFFR
‗You didn‘t pour perfume on my head.‘
Further, in non-affirmative clauses with final words ending with the vowels /e/
or /o/ in open syllables, the Negotiator is realised as [i]; represented
orthographically as ι:
(44) St. Maria play

Ʋlɛ bɛ so a tome ι.
3SG.EMP NEG.IND.NFUT own.PFV DEF work.NMLZ NAFFR
‗S/he doesn‘t own the work.‘

A nιsaalɛ bɛ tι ι ʋ yo
DEF human being NEG.IND.NFUT PST.REMCOP.PFV 3SG alone
ι.
NAFFR
‗The human being was not alone.‘

(46) St. Maria play
A zu bɛ tone ι.
DEF head NEG.IND.NFUT work.IPFV NAFFR
‗He is not intelligent.‘
As the examples in this section show, the non-affirmative clause intersects

with the negative declarative clause, just as the affirmative clause intersects
with the positive declarative clause (see Section 4.5.1 on polarity). Thus, in
each case, the polarity particle, which precedes the verb, and the mood
particle, which occurs at the end of the clause, form a prosody. The clause is
predicated by the polarity particle as either negative or positive and it ends
with an interpersonal punch, as the speaker potentially hands over the turn.
This phenomenon shows how elements of the Mood base are closely
connected in enacting the clause as an interactive unit. Such prosodies are also
147
a distinctive characteristic of interpersonal systems of language and they are
recurrent in the analysis in this chapter.
4.4.1.2 Interrogative Clauses

We turn our attention now to the interrogative mood type. This mood type
typically realises the speech function of questions, demanding information.
There are two main types of interrogative clauses, namely, the polar
interrogative and the elemental (or non-polar) interrogative clause. The main
distinction between them is that while the polar interrogative mood takes a
Negotiator element at the end of the clause, the elemental interrogative mood
does not require a Negotiator. Rather, it is marked by the presence of a
question word (Q-element). The polar interrogative type will be discussed
first.
4.4.1.2.1 Polar interrogative clauses
The polar interrogative clause, as the name suggests, realises a question about
polarity, offering an option to the interlocutor to affirm or deny a proposition.
A typical answer to questions realised by polar interrogative clauses is either
ʋʋ (‗yes‘) or ʋʋ-hʋ (‗no‘). It has the clause structure Subject ^ Predicator ^
Negotiator, which means that, like the non-affirmative clause, it always takes a
Negotiator. The particular particle realising the Negotiatior is important in
distinguishing among delicate types of polar interrogative clauses (cf. Figure
4.10). Dagaare has three kinds of polar interrogative clauses: (i) neutral polar
interrogative clause, (ii) biased polar interrogative clause and (iii) alternative
interrogative clause. Biased interrogative clauses clearly indicate the answer
the speaker expects from the listener while neutral interrogative clauses do not
explicitly encode the speaker‘s expectation.
The lexicogrammatical contrasts among sub-types of the polar

interrogative clause are indicated by the specific realisation of the Negotiator
element (see Table 4.4).
148
Table 4.4 Realisations of different polar interrogative types
Polar interrogative type Sub-type Realisation
‗Yes/no‘ neutral bɩ
initiative wɛ, kpo, ʋ1
biased
responsive ya
Alternative bɩɩ-(‗or‘) conjunction
Note: 1Wɛ and kpo are sub-dialectal variants
Before we proceed to describe the subtypes in more detail, one general

observation about negotiation in polar interrogative clauses is that the presence
or absence of the affirmative particle distinguishes between questions that are
oriented towards propositions (47) and those that are oriented towards
proposals (48):
(47) Bɛ wa na bι?
3PL.HM come.PFV AFFR INT
‗Have they come?‘

(48) Ɩ wa bι?
1SG come.PFV INT
‗Should I come?‘
As (47) shows, those clauses oriented towards propositions normally embody

the structure of declarative clauses. Thus, without the interrogative particle,
(47) has the form of a declarative clause, specifically an affirmative clause. On
the other hand, polar interrogative clauses oriented towards proposals, as (48)
embody the structure of an imperative clause. The reason is that, in Dagaare,
obligation is typically realised by the imperative clause rather than the modal
verbs we find in languages such as English (see the glossing in example 48).
(1) Neutral polar interrogative: This interrogative type is unbiased and the
typical realisation of the Negotiator is by the interrogator bι, as in the
underlined clauses in (49) and (50):
(49) St. Maria play

A: A dɔɔ nɔŋ ya-rɛ na bι?
DEF man possibly be:mad-IPFV AFFR INT
149
B: Ɩ bɛ bãw ɛ wɛ!
1SG NEG know.PFV NAFFR EXCL
A: ‗Is the man mad?‘

B: ‗I don‘t know!‘
(50) St. Maria play
A: Ãa bɛlι fʋ ɛ ɩ kuol
who deceive.PFV 2SG that 1SG be:drunk.PFV
a?
AFFR
B: Fʋ kʋ tʋɔ ar vla ι
2SG NEG be:able.PFV stand.PFV well NAFFR
wɛ!
EXCL
A: Gaa na ɩ ga bι?
lying:down IDENT.PL 1SG lie:down.PFV INT
A: ‗who deceived you that I‘m drunk?‘

B: ‗But you can‘t stand well!‘
A: ‗Am I lying down?‘
(2) Biased polar interrogative: This polar interrogative type seeks

confirmation from the addressee. That is, it suggests that the speaker expects
the listener to confirm the proposition realised by the clause. There are two
kinds of biased polar interrogative clauses: (a) initiative and (b) responsive.
These are defined based on the relative role of the speaker in the exchange.
(i) Initiative: For the initiative type, the speaker invites the listener to confirm
a proposition. The Negotiator in the biased initiative interrogative is realised
by one of two particles: wɛ or ʋ. Illustrations are given in the dialogues below:
(51) St. Maria play

A: Nι dι na wɛ?
B: Ʋʋ.
‗yes‘
A: ‗You have eaten, right?‘
B: ‗Yes.‘
150
(52) St. Maria play
A: Nι wõ =n a tɛwr wɛ?
2PL hear.PFV FOC DEF meaning INT
B: Ʋʋ.
‗yes‘
A:‗You understand, right?‘
B: ‗Yes.‘
(53) St. Maria play
A: Nι na wa nι nι libir ʋ?
2PL POS.IND.FUT come.PFV CAUS FOC money INT
B: Ʋʋ.
A: ‗Will you bring money?‘
B: ‗Yes.‘
As the illustrations suggest, this type of interrogative clause shows that the
speaker expects an affirmation to the question. This type of interrogative
clause is semantically a near equivalent of tag questions in languages such as
English, especifically, when the Negotiator is realised by wɛ. The particle ʋ is
realised phonologically as [ʊ] or [u] in [-ATR] and [+ATR] environments
respectively.
(ii) Responsive: In this biased interrogative type, the speaker responds to a

proposition with a confirmation seeking question (normally called an echo
question). Thus, the difference between the initiative and the responsive type
is that in the initiative the speaker is initiating an exchange while in the
responsive type, he is responding to a proposition. The Negotiator in the
responsive interrogative is realised by the particle ya as in the dialogue below:
(54) St. Maria play

A: Ɩ zà a wa na wa nyɛ
1SG yesterday come AFFR PROX see.PFV
a ι pɔw-yaa ʋ bιɛrɛ.
DEF 1SG daughter 3SG be:sick.IPFV
B: Ʋ bιɛrɛ ya?
3SG be:sick.IPFV INT
151
A: ‗Yesterday I came home to see that my daughter was sick.‘
B: ‗She was sick, you say?‘
Here, Speaker B shows surprise and seeks to confirm speaker A‘s observation
about her daughter‘s sickness. Thus, this interrogative type can also carry
mirative meaning (cf. Section 4.5.3).
(3) Alternative (or bιι) interrogative: The alternative or bιι interrogative is a

special kind of the polar interrogative type. As its name suggests, it realises a
question that presents alternative statements, and invites the addressee to make
a choice. It is structurally realised by a paratactic clause complex (55) or group
complexes (56) with a bιι conjunction (i.e. alternation). Examples are the
underlined clauses in the dialogues below:

A: Ŋmιŋmιn na a pɔlɩtɩks yéle cere a
how IDENT.PL DEF politics matter go.IPFV DEF
Nandɔm pʋɔ ka? A cere nι

Nandom inside here 3PL.NHM go.IPFV FOC
galmɛ bɩɩ a tuur =ι a

zigzag CONJ 3PL.NHM follow.IPFV FOC DEF
sɔr bie?
way seeds
B: Mmm. Tɔɔ, pɔlɩtɩks a Nandɔm ka
INTJ well politics DEF Nandom here
ende a, ɩ kʋ ŋmaa ziri
actually JUNC 1SG NEG.IND.FUT cut lie
mãa zie a cere nɩ galmɛ.
1SG.EMP place DEF go.IPFV FOC zigzag
A: ‗How is the politics going on at Nandom here?‘ Is it following the
right path or it‘s going zigzag?‘
B: ‗Mmm. Well, actually, politics at Nandom here, I won‘t tell lies …
for me, it is going zigzag.‘
152
(56) St. Maria play
A: Fʋ bʋɔ, nʋ́ɔ bιι baa nʋ?
2SG goat fowl or dog IDENT.SG
B: Ou! Ɩ pɔw nʋ.

INTJ 1SG wife IDENT.SG
A: ‗Is s/he your goat, fowl, or dog?

B: Oh! ‗S/he is my wife.‘
The alternative interrogative in (55) is realised by the conjunction of two

positive declarative clauses: A cere nι galmɛ (‗It is going zigzag‘) and A tuur ι
a sɔr bie ‗(It is following the path‘). The sense of the interrogation is added by
the bιι (alternative) conjunction. Since each of the clauses in the conjunction
semantically opposes the other in this interrogative type, one common way of
realising it is to conjoin a positive and negative clause. In (55), this semantic
opposition is realised by the antonymous relationship between the
Complements, a sɔr bie and galmɛ. Let‘s further illustrate this by adapted
versions of example (57):
(57) A tuur =ι a sɔr bie bιι a

DEF follow.IPFV FOC DEF way seeds CONJ DEF
bɛ tuur a sɔr bie?

NEG.IND.NFUT follow.IPFV DEF way seeds
‗Is it following the right path or it is not following the right path?‘
4.4.1.2.2 Elemental interrogative

The elemental interrogative clause realises content questions. A participant or
circumstance in the clause is queried by the speaker, and s/he expects the
addressee to supply information about that element. Its distinctive feature is
the presence of a Q-element. Table 4.5 lists the question words used in the
elemental interrogative clause and their associated grammatical features. As
the table shows, the selection of a Q-element in interrogative clauses is closely
linked with the potential function of the element in the clause. It may be
conflated with one of four interpersonal functions, namely, Subject,
Complement, Adjunct and Predicator. When it is conflated with the Subject
153
Table 4.5. Question words used in elemental interrogative clauses
lexicogrammatical categories Q-word gloss
potential humanness textual transitivity number

clause
element
thematic singular ãa who/whom/
whose
human focal singular anʋ who/whom
/whose
thematic/ plural a mɩnɛ who/whom/
Subject/ focal whose
Comple- identifying singular buor which
ment
thematic bʋʋ what
neutral
focal bʋnʋ, what
non-human
ŋmɩn
identifying singular buor which
identifying plural a bobe what
idea & bo ~ what
locution bʋnʋ
Adjunct reason neutral bʋʋ so why
time debor when
neutral place nyɩn ~ where
nyɩnɛ
quantity/ singular a ŋmɩn how much/
value, many
attribute
quantity/ plural a ŋmɩnɛ how much/
value (several many
entities)
manner, ŋmɩŋmɩn how
attribute
Predicator interrogative neutral wa be where
verb
or Predicator, it is part of the Mood base, and when it is conflated with

Complement or Adjunct, it falls within the Residue. Illustrations are given in
Figure 4.11 below:
154
(a)Q-element conflated with Subject
Bʋʋ wa =n nι a ka
what come.PFV CAUS 2PL DEF here
Subject Predicator Complement Adjunct
Mood Residue
nominal group verbal group nominal group adverbial group
Q-element verb particle pronoun determiner adverb
‗What brought you here?‘ (St. Maria play)
(b)Q-element conflated with Predicator
A ŋman wa a ka
DEF calabash be:where DEF here
Mood Residue
determiner noun interrogative verb determiner adverb

‗Where is the calabash here?‘
(c)Q-element conflated with Complement
Bʋʋ fʋ bɔbr
what 2SG want.IPFV
Complement Subject Predicator
Residue Mood
nominal group nominal group verbal group
Q-element pronoun verb
‗What do you want?‘ (St. Maria play)
Figure 4.11: Q-element conflating with clausal functions
(1) Subject and Complement related Q-elements: The Subject and

Complement related Q-elements show a distinction between human and non-
human referents. The exception is the singular form for identification (i.e.
buor, ‗which‘). In addition, the [+human] Q-elements distinguish between
singular (i.e. ãa and anʋ, ‗who‘) and plural (i.e. a mιnɛ, ‗who‘), while their
non-human counterparts (i.e. bʋʋ and bʋnʋ, ‗what‘) do not show number
distinctions. For non-human referents, however, the identification elements
buor and abobe (‗which‘) contrast as singular and plural respectively.
Again, bo (‗what‘) is restricted to the environment of projection (i.e
reportative and quotative constructions), where it interrogates the projection,
either locution (what was said; example 58) or idea (what was thought;
155
example 59). Bʋnʋ (‗what‘) is however, more general and can occur in both
projection and non-projection environments (example 60 and 61):

A Sɛb-sow yèl a kɛ bo?
DEF scripture-holy say. PFV AFFR PROJ what
‗What does the holy scripture say?‘
A ŋa pɛr =ι bo?
DEF this bottom FOC what
‗What does this mean?‘
(60) St. Maria play
Ɩ yèl kɛ bʋnʋ?
1SG say. PFV PROJ what
‗What do I say?‘
Fʋ tɩɛr kɛ bʋnʋ.
2SG think.PFV PROJ what
‗What do you think?‘
(2) Adjunct-related Q-elements: Q-elements that typically function as

Adjunct comprise bʋʋ so (‗why‘), debor (‗when‘), nyɩn or nyɩnɛ (‗where‘), a
ŋmɩn (‗how much/many‘, singular), a ŋmɩnɛ (‗how much/many‘, plural), and
ŋmɩŋmɩn (how). Reason is, arguably, the least specialised meaning in the
elemental interrogative clause. Thus, there are a range of strategies for
realizing it. As Table 4.4 shows, the most common Q-element for reason is
bʋʋ so (‗why‘; literally, ‗what owns‘) although bʋʋ can also be used alone
(Compare examples (62) and (63):
(62) St. Maria play

Bʋʋ so fʋ yèle a lɛ?
what own 2SG say.IPFV DEF DEM
‗Why are you saying that?‘
156
(63) St. Maria play
Bʋʋ fʋ mὶ lɛ ƴɛrɛ lɛ?
what 2SG too also say.IPFV DEM
‗Why are you too speaking like that?‘
In addition, reason is also interrogated by a combination of bʋʋ or bʋnʋ

(‗what‘) or ŋmιnŋmιn (‗how‘), on the one hand, and the postposition ƴaw (‗for
the sake of‘). Examples are below:

Bʋnʋ ƴaw na nyι bɔbr mɛ?
what sake IDENT.PL 2PL look.PFV 1SG.ACC
‗For the sake of what are you looking for me?‘

Ŋmιnŋmιn ƴaw na fʋ bɔbr kɛ fʋ
how sake IDENT.PL 2SG speak.PFV PROJ 2SG
zɛbr =ι a Naaŋmin?
quarrel FOC DEF God
‗It is for what sake that you want to quarrel with God?‘
Regarding a ŋmιn and a ŋmɩnɛ (‗how much/many‘), both are used to

interrogate quantity and value. However, in addition, a ŋmɩnɛ specifies that the
entities enquired about are several (66a, b), and this discreteness is often
emphasised by duplicating the Q-element (66b):
(66a) A simie ɩ =n a ŋmɩn?

DEF peanuts be.PFV FOC DEF how:much
‗How much is the peanuts?‘
(66b) A simie ɩ =n a ŋmɩnɛ ŋmɩnɛ?
DEF peanuts be.PFV FOC DEF how:much how:much
‗How much, how much are the peanuts?‘
In example (66a), the peanuts are conceived of by the speaker as a bulk of

entity while, in example (66b), they are conceived of as discrete packages or
piles.
157
The Q-element ŋmιn, without the definite marker a, is also used with
the sense of English ‗what‘ (example 67) and also ‗where‘in copula clauses
(example 68):
(67) St. Maria play

Ɩ pãa ι ŋmιn?
1SG ADV do.PFV what
‗What can I do now?‘
(68) St. Maria play
Ʋ ι =n ŋmιn?
3SG COP.PFV FOC what
‗Where is s/he?‘
Finally, the Q-element ŋmιnŋmιn (‗how‘) indicates manner, as in example (69)

below:

Ŋmɩnŋmɩn na nyɩ ɩ tʋɔ
how IDENT.PL 2PL do.PFV be:able
gure?
sleep.IPFV
‗How is that you are able to sleep?‘
(3) Interrogative verb wa: As has been illustrated earlier, Dagaare has one
interrogative verb, wa (‗be where‘) and it is used in interrogating existence
and locations. Illustrations are given below:
(70) A fʋ ma wa?
DEF 1SG mother be:where
‗Where is your mother?‘
(71) St. Maria play
A fʋ pɔw mὶ wa?
DEF 2SG wife also be:where
‗Where is your wife too?‘
The interrogative verb wa is not marked for aspect.
158
(4) Echo questions: Elemental interrogative clauses also realise echo
questions. This type of clause is similar to the biased responsive interrogative.
The difference is that elemental echo questions require the addressee to supply
a statement or repeat a proposition, rather than give a polar, ʋʋ (‗yes‘) or ʋʋ-hʋ
(‗no‘), answer. In such echo questions, the interrogative particle ya is used
together with the Q-element, and it is placed clause finally as a Negotiator
element:
(72) St. Maria play

A: Ɩ cen na ɛ ι tι ɔw a
1SG go.PFV AFFR CONJ 1SG DIST fetch.PFV DEF
kʋɔ o;
water PRT.
pobile kàw tι yèl kɛ ʋ bɔbr

gentleman some DIST say. PFV PROJ 3SG want.IPFV
mɛ na.
1SG.ACC AFFR
B: Ʋ ι =n ŋmιn ya?
3SG do.PFV FOC what INT
Subj. Pred. Complement Negotiator

A: ‗I went to fetch the water O; a gentleman said that he loves me.‘
B: ‗He what?‘
With this example, we can conclude that all five principal functions in the
interpersonal clause structure, comprising Subject, Predicator, Complement,
Adjunct and Negotiator, are engaged in realising different meanings in
elemental interrogative clauses. In addition, however, the Q-element can be a
modifier in a nominal group, as (73) shows:
(73) St. Maria play

Bʋʋ tome na fʋ to?
what work.NMLZ IDENT.PL 2SG work.PFV
‗What work have you done?‘
The distinctive character of the elemental interrogative is the presence of the
Q-element in the clause.
159
4.4.1.2.3 Placement of the Q-element in the clause
One important typological consideration in describing elemental interrogative

clauses is to examine the unmarked position of the Q-element in the clause.
Teruya et al. (2007) identify three motifs in this regard across several
languages. First, some languages assign the Q-element unmarked thematic
status in the clause, mostly in clause initial position (e.g. in English, German,
Danish, French and Arabic). In other languages, the default case is to place the
Q-element in unmarked focus position (e.g. Yoruba). Finally, some other
languages do not associate the Q-element with any special textual status of the
clause. Rather, it simply occurs in situ as it would in the corresponding
declarative clause (e.g. Chinese, Vietnamese, Tibetan, Telugu and Japanese).
Dagaare belongs to this last category. This principle is reflected in the
examples that have already been given on elemental interrogatives. The
following dialogue further illustrates this textual character of Q-elements in
the clause (Q-element is in bold and the corresponding referent in the agnate
declarative clause is underlined):
(74) St. Maria play

A: Dɔɔ, a fʋ your =ι bʋnʋ?
man DEF 2SG name COP.FOC what
B: Ɩ yuor =ι Gbιbgbãw.
1SG name COP.FOC Gbibgbaw.
A: Ʋ saa na.
3SG show.PFV AFFR
B: Alɛ! Bʋʋ na saa?

INTJ what IDENT.PL show.PFV
A: ‗Man, what is your name?‘
B: ‗My name is Gbibgbaw.‘
A: ‗It shows on him (= ‗The name matches his appearance).‘
B: ‗What! What is it (that) shows (on me)?‘
However, as Table 4.5 indicates, ãa (‗who‘) and bʋʋ (‗what‘), on the one
hand, and anʋ (‗who‘) and bʋnʋ (‗what‘), on the other hand, contrast as
160
thematic and focal Q-elements respectively. For an illustration, compare (75)
and (76), on the one hand, and (77) and (78), on the other hand:
(75) St. Maria play

Ãa nιɛ yèl kɛ a fʋ pɔw-yaa nʋ
who even say. PFV PROJ DEF 2SG daughter IDENT.SG
mɔ?
CE
‗Who even said that it is your daughter?‘

(76) St. Maria play
A fʋ pɔw nι anʋ a ka?
DEF 2SG wife COP.FOC who DEF here
‗who is your wife here?‘
(77) St. Maria play
Bʋʋ fʋ bɔbr a ka?
what 2SG want.IPFV DEF here
‗What do you want here?‘
(78) St. Maria play
Bʋrʋ yéle nι bʋnʋ
sowing matter COP.FOC what?
‗How does the issue of sowing come here?‘
Unlike languages such as English, where a focused Q-element is typically

used to realise echo questions (Teruya et al. 2007), the thematic and focused
Q-elements in the above clauses occur in situ, and thus are unmarked.
4.4.2 Imperative
The imperative clause is oriented towards the speech functions of offer and
command (i.e. proposals). It thus positions interlocutors within the semantic
space of the exchange of goods-&-services, offering them a resource for
expressing permission and obligations towards one another. As mentioned
earlier, its minimal realisation is the Predicator (see Figure 4.12). Generally,
an imperative clause may or may not take a Subject, depending on
grammatical person and number of the one responsible for executing the
161
command, whether it is an interactant or non-interactant, and whether it is
singular or plural. The only situation in which the Subject is absent is when
modal responsibility is assigned to a single addressee. The modal structure of
the imperative clause can be summarised as: (Subject) ^ Predicator ^
(Negotiator).
Figure 4.12: A system network of the IMPERATIVE TYPE
The primary contrast in the imperative clause is between prohibitive and non-
prohibitive clauses. The non-prohibitive imperative has the Predicator as the
only obligatory element and will require a Subject if the addressee is other
than the second person singular. Thus, in (79) and (80) below, the addressee is
a single interactant while in (81) and (82), the addressee is plural interactant:

Hɛɛ, pʋr bʋla ƴaw a Zan!
hey pour.PFV little put.PFV DEF Zan
‗Hey, pour a little for Zan!‘
Ƴɑw ʋ ɩ ƴɑw fʋ!
make.PFV 3SG do.PFV put.PFV 2SG
‗Let him/her give you some!‘
162
Nyɩ dɔwr paalɛ ɛ yɛrɛ!
2PL procreate.IPFV fill.IPFV CONJ spread.IPFV
‗You give birth to many, and multiply!‘
(82) Nɩ ír! Nɩ kpɛ mɔb!
2PL get up 2PL enter.PFV wrestling
‗You get up! You engage in wrestling!‘
On the other hand, the prohibitive clause takes one of two negative imperative
particles (ta and taa) and requires a Negotiator (realised by a non-affirmative
particle: ɩ e or ɛ), in addition to the Predicator element. Two types of the
prohibitive clause are distinguished, based on the temporal orientation of the
proposal: one with an immediate enforcement (example 83 to 85) and one that
prohibits a perceived future event (example 86). The former selects the
particle ta and can combine with a verb in either the perfective or imperfective
aspect (example 83 to 85). The latter, however, selects the particle taa and
occurs with only the imperfective aspect (86):

Ta tɛr =ɩ nɩr suur e!
NEG.IMP.IM possess.PFV COM person anger NAFFR.
‗Don’tharbor anger against a person!‘

(84) St. Maria play
Ta yɛrɛ nɩ mãa ɩ!
NEG.IMP.IM speak.IPFV COM me NAFFR
‗Don’t talk to me!‘

(85) St. Maria play
Ta mɩ́ yèl a lɛ ɩ!
NEG.IMP.IM HAB say. PFV DEF that NAFFR
‗Don’t be saying that!‘

Taa yarɛ yél-faa san ɛ!
NEG.IMP.NIM pay.IPFV matter-evil debt NAFFR
‗Don’t pay back evil! (= Never pay back evil!)‘
163
4.4.3 The Imperative Clause and MOOD PERSON
Another issue that need to be examined in relation to the imperative clause is

that of mood person. The Subject in the Dagaare imperative clause generally
conflates with all persons and numbers: first singular and plural, second
singular and plural and third singular and plural. The second person imperative
(e.g. Wa! ‗Come!‘; Nyι wa! ‗You come!‘) is the realisation of a command
proper, the demand for goods-&-services from the addressee. The first person
imperative is typically oriented towards offer (e.g. Ɩ tuo fʋ! ‗Let me help you
carry!‘) and suggestion (e.g. Tι ír ka! ‗Let‘s get up!‘).
The third person imperative mood covers the semantic space of
obligation, permission and suggestion, meanings that are often realised in
other languages such as English, as the by deontic modality (cf. Section
4.4.1.2.1). Let‘s consider example (87) below:

Bɛ voti kʋ̀ a NDC gɔvʋmɛnt!
3PL.HM vote.PFV give.PFV DEF NDC government
‗They should vote for the NDC government.‘
As the translation suggests, in English, the proposition in example (87) will be

realised by a modulated declarative clause.In Dagaare, the use of a non-
interactant Subject in the imperative enacts the clause as a modulated
proposal. The dialogue below highlights the distinction between second person
imperative and the third person imperative (imperatives underlined):
(88) St. Maria play

A: Mɩ́ bin daι! A ana bɛ mɩ́
HAB put.PFV die DEF DEM NEG.IND.NFUT HAB
tɛr pɛr ɛ.
possess bottom NAFFR
B: Daι ya? Ʋ mɩ́ bin tιrιpʋl!
die INT 3SG HAB put.PFV triple
A: ‗Always put two days old pito/beer! This is always useless.‘
B: ‗Two days old? S/he should put three days old pito/beer!‘
164
Speaker A‘s utterance is addressed to a bar owner while Speaker B‘s is a
reaction to A‘s command. The imperative clause in Speaker A‘s utterance is
addressed to a single interactantand is thus a command proper. In Speaker B‘s
reaction, on the other hand, the Subject of the imperative is a non-interactant,
thereby shifting the proposal away from the ‗you-&-me‘ dimension and
enacting it as a suggestion. Whether an imperative with a non-interactant
Subject realises an obligation, permission or suggestion depends on context.
4.4.4 A Note on MOODand Grammatical Metaphor
The discussion so far has focused on the congruent relationship between the
semantic system of SPEECH FUNCTION and the mood types that realise these
functions. It has been demonstrated that the lexicogrammatical resources of
mood evolved to realise speech functions. The system of MOOD is thus a
grammaticalisation of various discourse meanings which we have referred to
as speech functions. However, since language, including the system of MOOD,
is an evolved system rather than a designed one, there is sometimes
incongruity between the mood types identified above and the speech functions
they realise. This incongruity between semantics and lexicogrammar is what
Halliday (e.g. Halliday, 1985; Halliday & Matthiessen 2014) has identified as
grammatical metaphor. It theorises the fact that there is an interstratal tension
between semantics and lexicogrammar, as users of language constantly recruit
existing lexicogrammatical resources to construe or enact novel, innovative
meaning.
One type of grammatical metaphor is mood metaphor. It is a
decoupling of the congruent relation between the semantics of SPEECH
FUNCTION and the lexicogrammar of MOOD, and then rematching the two
systems in incongruent ways (Halliday 2008; Halliday & Matthiessen 2014).
Mood metaphor has been attested by the long tradition of research on
indirection, motivated by Austin and Searle‘s Speech Act Theory, (e.g. Austin
1962; Searle 1969) and Brown and Levinson‘s (1987) politeness framework.
In principle, while each of the mood types identified above for Dagaare
remains the default realisation of the speech function by which it is defined, it
can be recruited by a speaker to realise a speech function that is not typical of
165
it. This indirect realisation of speech functions expands the interpersonal
resources for negotiation of the relationship between speaker and addressee.
The following clausesrealise biblical commandments in the declarative clause
rather than the prohibitive imperative:

Fʋ kʋ kʋ́ nιr ɛ.
2SG NEG.IND.FUT kill. PFV person NAFFR
‗You will not kill a person.‘

Fʋ kʋ zʋ bom e.
2SG NEG.IND.FUT steal thing NAFFR
‗You will not steal things.‘

Fʋ kʋ ŋmaa ziri ƴaw fʋ tɔ ι.
2SG NEG.IND.FUT cut lie put 2SG fellow NAFFR
‗You will not lie against your fellow.‘
The sense of command in the above clauses is enacted partly by the

combination of the second person and future tense. The use of the future tense
for realising commands in the imperative has been noted in the typology
literature (cf. Bybee, Perkins & Pagliuca 1994). In our examples, above, the
metaphorical rendering of the commandments enacts them as definite and
strongly binding.
4.4.5 Mood and Elliptical and Minor Clauses

The discussion so far has focused on major and full clauses. However,
language in use is often characterised by minor and elliptical clauses and a
functional grammar must account for these constructions. It was noted in
Section 4.2 that the minimal structure of the indicative clause consists of
Subject + Predicator while an imperative can be realised by Predicator only,
depending on the mood person. In elliptical clauses, core elements of the
clauses are omitted and canbe recovered from co-text or situational context.
Such ellipsis, therefore, enact the here-&-now and you-&-me dimensions of
166
spoken discourse. One common type of ellipsis is exemplified by the
underlined clause in (92):
(92) St. Maria play

A: Dɔɔ kãw zà a yèl kɛ tɩ ƴɛr
man some yesterday say. PFV PROJ 1PL grind.PFV
sɩɓaan ƴaw a pɛrɩ.
pepper put DEF anus
‗Yesterday, a certain man said we should grind pepper and put it in the
anus.‘
B: St. Maria play
Sɩɓaan ya?
pepper INT
‗Pepper?‘
Here, the clause consists of only the Negotiator ya and the Complement
Sɩɓaan, and it enacts Speaker B‘s surprise at A‘s statement.
Minor clauses, on the other hand, are recurrent conventionalised forms
that are inherently reduced, or even unanalysable, in terms of constituent
structure (cf. Chapter 3, Section 3.4.3). They are typically characterised by the
absence of a Predicator element.Minor clauses often realise exclamations,
calls and responses, greetings and alarms (cf. Halliday & Matthiessen 2014).
Of these, exclamations and greetings need a special comment since they
embody many resources.
(1) Exclamations and calls: Exlamations, in Dagaare, are not realised as a

special clause type within ‗declarative‘ in the MOOD system, as it is in
languages such as English (e.g. How beautiful she is!). Rather, it can be
realised by a major clause of any mood type (except for polar interrogative)
and by minor clausessuch as interjections and expletives: wi! oyi! and ʋaw!
(see Mwinlaaru in press). Only minor clauses will be illustrated here. Major
exclamative clauses are discussed in Section 4.5.3. The extract below
illustrates exclamations and calls realised by minor clauses:
167
Oyi! Yeezu u! Oyi! Yeezu u!
INTJ Jesus VOC INTJ Jesus VOC
Daviir bi-dɛb ι, zɔ a ι nι-baalʋ wɛ!

David child-male VOC run.PFV DEF 1SG pity REQ
‗Oh, Jesus! Oh, Jesus! Son of David, have pity on me, please!‘
In (93), the exclamations relaised by Oyi! are juxtaposed with calls, and both
establish an interpersonal background for the proposition that follows. As the
example shows, calls are like Vocatives. However, Vocatives are meant to
alert and create a tenorrelation with the addressee, and they are normally
attached to the clause as functional element as in Daviir bi-dɛb ι (‗Son of
David‘). Calls, on the other hand, are expressed as isolated clauses and seek to
invite the addressee for interaction.
(2)Greetings: Like exclamations, greetings are realised by both major clause

and minor clauses. Temporally oriented daily greetings in Dagaare are set out
in Table 4.6, mostly in the form of time adverbs, although speakers often
embed them in major clauses.
Table 4.6. Daily greetings and responses in Dagaare

Turn Time of the day
morning afternoon evening night
Greeting zaamι, ŋmιna zà a-nʋɔr zι-ɓaan,
‗After ‗afternoon‘ ‗evening‘ ‗place-cool‘
yesterday‘;
a gaa be
=n sow?
‗Is the
sleep
well?‘
Response zaamι nʋ, ŋmιna tɛɛ, ‗ zà a-nɔ tɛɛ, zι-ɓaan tɛɛ,
‗After afternoon tɛɛ‘; ‗evening ‗place-cool tɛɛ‘
yesterday ŋmιna nʋ, tɛɛ‘; zι-ɓaan nʋ
it is‘; A be „Afternoon it is.‘ zà a-nuɔra ‗place-cool it is.‘
=n sow, ‗It nʋ, Evening
is well.‘ it is.‘
168
Greetings can also be in the form of nominal groups: Fʋ tome! (‗Your
work! / How is work?‘) Nɩ yaanɩ! (‗your hail!‘ / ‗you are hailed!‘), Tɩ bɛrʋ a
yɩ! (‗our two days!‘ / It‘s been a while‘), Tɩ kɔwrʋ! (‗our long time! / It‘s been
a long time‘), A zie! (‗The place!‘ / Is the place treating you well?). Except for
Nɩ yaanɩ! (Your hail! / You are hailed!‘), instances such as these can be
interpreted as reduced clauses and their corresponding major clause are
illustrated by (94) and (95):
(94) A fʋ tome be =n sow?

DEF 2SG work.NMLZ be.PFV FOC well
‗Is your work well? (How is your work?)‘
(95) A tɩ bɛrʋ ayi be =n sow?
DEF 1PL days two be.PFV FOC well
‗Is our two days well? (=How have you been since we last met?)‘
Greetings are also realised by minor exclamative clauses (cf. Section 4.5.3),
consisting of a time adverb (96) or a nominal group (97) and Negotiator (wɛ):
(96) St. Maria play

A: Zaamɩ wɛ!
yesterday EXCL
B: Zaamɩ nʋ!
yesterday IDENT.SG
A: ‗Yesterday! (=Good moring!)‘

B: ‗Yesterday it is! (=Good morning!)‘
(97) St. Maria play
A: A nɩ waa yaanɩ wɛ!
DEF 2PL coming hail EXCL
B: Ʋ be =n sow!
3SG EXIST.PFV FOC well.
A: ‗Hail your coming! (=You are welcome!)‘
B: ‗It is well!‘
In summary, although many of the clauses realising greetings have the

structure of major clauses, they are mostly semantically opaque; they are
169
formulaic expressions for the routine performance of sociality in the speech
fellowship.
4.5 Polarity and Modal Assessment

much of the discussion in the preceding sections has shown that the system of
MOOD is the grammar of speech functions, the main resource for enacting the
clause as a move. the system of mood is expanded by POLARITY (section 4.5.1)
and the modal assessment systems of MODALITY (section 4.5.2) and
NEGOTIATION (section 4.5.3), together with other resources such as mood
adjuncts, discussed in section 4.3.4.2 (also see chapter 3, section 3.4.2.3 on
adverbial particles). the lexicogrammatical items that realise the systems of
POLARITY and MODALITY form part of the predicator in the clause structure
(see table 4.2), while NEGOTIATION is realised by clause intial and clause final
particles.
4.5.1 POLARITY
Polarity is the semantic opposition between ‗positive‘ and ‗negative‘ in the

clause. In Dagaare, this opposition is grammaticalised differently across
indicative and imperative clauses, the grammar of propositions and proposals
respectively. In the imperative clause, polarity is realised by the opposition
between the prohibitive clause, realised by the negative particle ta (immediate)
or taa (non-immediate), and non-prohibitive imperative clause, which is
realised by zero-marking (see Section 4.4.2.1). The different realisations of
polarity across indicative and imperative clauses imply that polarity
contributes to primary mood distinction in the clause. This section will extend
the discussion on polarity introduced earlier in the account of imperative
clauses to the indicative clause.
Polarity in indicative clauses contributes to the validity and arguability
of the proposition by assessing it as either positive or negative. It intersects
with the ideational system of TENSE, specifically, FUTURITY, the opposition
between non-future and future. As Table 4.7 shows, for non-future tense,
positive polarity is realised by zero-marking while negative polarity is marked
by the particle bɛ. For future tense, on the other hand, both positive and
170
negative are overtly marked. Positive future is indicated by the particle na
while negative future is signalled by the particle kʋ.
Table 4.7 Realisation of polarity in the indicative clause
It should be noted that Dagaare does not make a distinction between

clausal negation and non-clausal negation. Thus, any negative meaning is
always realised in the Predicator. A constituent of the clause can however be
negated by focusing it in a cleft construction (cf.Watters 2000: 205-208 on
negation in African languages):
(98) Sukuul bɛ na a Dɛr cen e.

School NEG.IND.NFUT IDENT.PL DEF Der go.PFV NAFFR
‗It is not school that Der went to.‘
Thus, as example (98) shows, the negation of a constituent is still realised

clausally as a marked choice (see Chapter 5, Section 5.4.2.1 on cleft
construction and marked focus). Also, a non-assertive determiner, za (also
means ‗all‘), can be used in the nominal group to show that it is a central focus
of the negation in the clause (see examples 99& 100). The assertive
determiners that contrast with za are kãw (‗some‘, singular), bɛmɩnɛ (‗some‘,
plural human), amɩnɛ (‗some‘, plural non-human) (see example 101):

||| A tẽw tɩ ɩ =n vuo || ɛ
DEF earth PST.REM COP.PFV FOC hollow CONJ
bom za bɛ ka be ɩ.|||
thing all NEG.IND.NFUT EXIST.PFV there NAFFR
‗The earth was hollow and nothing was there.‘
171
Lit. ‗The earth was hollow and all things were not there.‘
(100) St. Maria play
Mãa bɛ tɛr ƴɛrʋ za ɩ.
1SG.EMP NEG.IND.NFUT possess.PFV speech all NAFFR
‗I don‘t have anything to say.‘

||| Fʋ pãa taa wa tɛr ƴɛrʋ kàw ||
2SG ADV MOD COND possess.PFV speech some
mɩ̀ na pãa kʋ̀ a cɛlcɛlbɛ? |||
also POS.IND.FUT ADV give.PFV DEF listeners
You maynow have something you want to also tell the listeners?
4.5.2 MODALITY
MODALITY is one of the common modal assessement systems across
languages (Matthiessen 2004: 631ff). Matthiessen (2004) uses the term modal
assessment as a general notion that encapsulates the different resources that
have been identified across languages for assessing the information or goods-
&-services being exchanged in discourse. In addition to modality, these
resources include evidentiality, honorification and negotiation. In Dagaare,
there are several grammatical resources dedicated to modal assessment. In the
discussion on the interpersonal structure of the clause, it was noted that mood
Adjuncts, realised by modal adverbs and adverbial particles, are used to show
the speakers attitude (cf. Section 4.3.4.2). Modality is also a key resource for
modal assessment and, in dagaare, it is defined as comprising the sub-systems
of PROBABILITY (Section 4.5.2.1) and DESIRABILITY (Section 4.5.2.2). Figure
4.13 presents these two modality resources in a system network.
Probability and desirability are a set of intermediate values between the
positive and negative poles of polarity; values that enact different
interpersonal meanings in propositions and proposals. The intermediate values
in propositions are degrees of probability and the intermediate values in
proposals are degrees of desirability. This systemic distinction shows that
there is more than one way of getting from positive to negative on the cline of
polarity: one through the exchange of information and the other through the
172
exchange of goods-&-services (cf. Halliday & Matthiessen 2014). As in the
modality systems of many languages across the world (cf. Bybee, Perkins &
Pagliuca 1994), both probability and desirability in Dagaare are marked by
essentially the same lexicogrammatical forms, which consist of three modal
particles (i.e. naa, taa and kʋʋ). These two sub-systems of modality are
discussed below.
4.5.2.1. Probability
Probability is an assessment of the positive or negative expectation of

situations and events. In other words, it tones down the degree of certainty
encoded by positive and negative clauses (See Figure 4.14). Three modal
categories can be distinguished in relation to probability: (i) positive
probability, (ii) negative probability and (iii) median probability. A positive
probability clause, realised by the particle naa, lowers the positive value of a
proposition, assessing it as less probable to occur, thus, shifting it closer to the
173
Figure 4.13: A system network of MODALITY in Dagaare
174
negative pole. A negative probability clause, on the other hand, reduces the
negative value of a proposition; thereby shifting it closer to the positive pole,
accepting the possibility of the actualisation of the otherwise negated
proposition. It is realised by the particle kʋʋ. Between these two poles of
modal probability lies the median probability, by which the speaker enacts an
uncommitted middle ground, neither indicating the positive nor the negative
value of the proposition. This modality type is realised by the particle taa.
Figure 4.14: Polarity and probability represented as a cline
Table 4.8 illustrates the systemic contrast among the three types of probability.
Further illustrations are given in examples (102) to (106).
Table 4.8. The realisation of probability by modal particles
175
(102) Yʋora na cen nɩ wɩɛ. (positive polarity)
Yuora POS.FUT go.PFV FOC farm
‗Yuora will go to farm.‘
(103) Yʋora naa wa cen nɩ wɩɛ. (modalised positive)
Yuora MOD.POS EVT go.PFV FOC farm
‗Yuora may go to farm.‘
(104) Yʋora taa wa cen wɩɛ. (median modality)
Yuora MOD EVT go.PFV farm.
‗Yuora might be going to farm/Yuora might have gone to farm.‘
(105) Yʋora kʋ̃ʋ wa cen wɩɛ. (modalised negative)
Yuora MOD.NEG EVT go.PFV farm
‗Yuora may not go to farm.‘
(106) Yʋora kʋ cen wɩɛ. (negative polarity)
Yuora NEG.IND.FUT go.PFV farm.
‗Yuora won‘t go to farm.‘
Unlike the positive (naa) and negative (kʋʋ) modal particles, the median
modal particle (taa) can co-occur with both positive and negative polarity
markers:
(107) Ʋ taa na zɩ a be.

3SG MOD POS.FUT sit.PFV DEF there
‗He might sit there.‘
(108) Ʋ taa bɛ zɩnɛ a be.
3SG MOD NEG.IND.NFUT sit.IPFV DEF there
‗He might not sit there.‘11 / ‗He might not have been sitting there.‘
(109) Ʋ taa kʋ zɩ a be.
3SG MOD NEG.FUT sit.PFV DEF there
‗He might not sit there.‘
This ability to co-occur with positive and negative particles owes to its neutral
sense on the positive-negative scale. In instances such as examples (107) to
(109), the median modal particle taa decreases the commitment of the speaker
11
The plausible future reading of example (108) is contributed by the imperfective verb zɩnɛ
and not the tense-mood-polarity bearing particle bɛ.
176
in relation to, as it were, the positive (as in 107) or negative (as in 108 & 109)
value of the proposition.
Also notable are instances of transferred or preposed modality;
defined as a situation where the modal particle is displaced from its logical
position in the secondary clause of a clause complex and placed in the primary
or initial clause:
(110) Tɩ taa cen bɛ wa nyɛ ʋ.

1PL MOD do.IPFV NEG.IND.NFUT EVT see.PFV 3SG
‗We might go and not see him.‘
(111) Ʋ naa zɔrɛ na wa lo.

3SG MOD.POS run.IPFV AFFR EVT fall. PFV
‗He may run and fall‘ (= ‗He may fall while running.‘)
In (110) and (111), the target of the modal meaning is on the process
‗perceiving‘ (i.e. it is possible that we will not see him) and ‗falling‘ (i.e. it is
possible that he will fall) respectively. However, the modal particles (taa and
naa) are transposed to the initial clauses. This phenomenon is in consonance
with the prosodic realisation of interpersonal meanings, they are diffusive in
their meaning.
4.5.2.2 Desirability
The system of DESIRABILITY offers resources for assessing the positive or
negative desirability of proposals, that is, whether the actualisation of the
proposal is desirable or undiserable (Matthiessen 2004). By using this kind of
modality, speakers enact their inability to bring about some desirable situation
or prevent an undesirable situation. On the other hand, they may simply be
laying down their ability to enforce a proposal in order to enact politeness.
Table 4.9 illustrates the various systemic options in this system. As in the case
of probability, desirability forms a cline of intermediate categories between
the positive and negative poles of polarity, ranging from highly desirable to
highly undesirable. High desirability is marked by the positive modal particle
naa, while high undesirability is marked by the negative modal particle kʋʋ.
177
Low desirability and low undesirability are marked by a compound of two
modal particles. Low desirability is marked by the compound particle taa naa,
Table 4.9. The realisation of desirablity by modal particles [change

examples]
modality: desirability particle example
high desirability naa Fʋ naa ƴɛr ƴɛrʋ
2SG MOD.POS talk.PFV talking
desirable kàw kʋ̀ bɛ.
some give.PFV 3PL.HM
‗I very much wish you tell
them something.‖
low desirability taa naa Fʋ taa naa ƴɛr ƴɛrʋ

2SG MOD MOD.POS talk talking
kàw kʋ̀ bɛ ʋ?
some give.PFV 3PL.HM INT
‗I wish you tell them
something, what do you think?‘
low undesirability taa kʋʋ Fʋ taa kʋ̃ʋ ƴɛr ƴɛrʋ
2SG MODMOD.NEG talk.PFV talking
undesirable za kʋ̀ bɛ ʋ?
all give.PFV 3PL.HM INT
‗I wish you don‘t tell them
anything, what do you think?‘
high undesirability kʋʋ Fʋ kʋ̃ʋ ƴɛr ƴɛrʋ
2SG MOD.NEG talk.PFV talking
za kʋ̀ bɛ ɩ.
all give.PFV 3PL.HM NAFFR
‗I very much wish you don‘t
tell them anything‖
uncommitted taa Fʋ taa na ƴɛr
2SG MOD FUT.POS.IND talk.PFV
ƴɛrʋ kàw kʋ̀ bɛ ʋ?
talking some give.PFV 3PL.HM INT
‗You may tell them something,
what do you think?‘
and low undesirability is marked by the compound taa kʋʋ (see example 113,
for instance). The use of the median particle taa, respectively, lowers the
positive (i.e. desirable) and negative (i.e. undesirable) values of naa and kʋʋ in
this context. The median modal particle taa itself, when used alone, enacts
uncommitted desirability (see (110) above & (114) below). In addition to the
illustrations in Table 4.9, other examples are given below from discourse:
178
(112) Bible.is (Fɩlɩmɔ 1: 13a)
Ɩ naa tɩ tɛr ʋ nɩ a ɩ
1SG MOD.POS PST.REM possess.PFV 3SG FOC DEF 1SG
zie ka.
place here
‗I should have liked to keep him here with me.‘
Nyɛ! zɩna mɩ̀ gba, ɩ taa naa yi
see.PFV today also even 1SG MOD MOD.POS go:out.PFV
cen tɩ nyu daa ʋ?

go.PFV DIST drink.PFV beer INT
‗Look! Today even, I feel like I should go out to drink beer, should I?‘
A nɩbɛ bɛŋa bɛ taa na wa
DEF people these 3PL.HM MOD FUT.POS come.PFV
tɩ ir ʋ?
1PL get up.PFV INT
‗These people, they may come so that we can get up, should they?‘
4.5.3 NEGOTIATION
The Negotiator is the grammatical element which realises the system of

NEGOTIATION in the clause, and, as its name suggests, it enacts the clause as a
negotiable unit of discourse.12 in other words, it grounds the proposition or
proposal realised by the clause within the semantic space open to speaker and
listener as something that can be negotiated. cross-linguistically, the system of
NEGOTIATION is found in a wide range of languages. It has been investigated
as part of various interpersonal linguistic resources under the notion of stance
(see Iwasaki & Yap (2015) and references therein). The semantic region of
negotiation is construed and realised differently across languages. In
intonation prominent languages such as English, it tends to be realised by
intonation (cf. Halliday & Greaves 123-128). In other languages such as
Chinese, Dagaare and other Mabia languages, negotiation is realised
12
The account on NEGOTIATION in this section has been published in Mwinlaaru (in press).
179
segmentally by clause final and/or clause intial particles, and many studies
have approached these grammatical resources for NEGOTIATION from below
the clause as a typology of clause final particles (see the contributions in
Hancil et al. (2015)). The account in this studyis a system-oriented approach to
Dagaare clause final particles and, indeed, clause initial particles, examining
them as constituting a clause rank system for modal assessment (see Halliday
& McDonald (2004: 341-342) on Chinese).
Until now only one kind of negotiation markers have been discussed in
this chapter, namely those that contribute to making mood distictions,
particularly in indicative clauses. In addition to these, however, there are
negotiation particles that are optional in the clause and only indicate the
speaker‘s attitude towards the proposition or proposal realised by the clause.
In this section, all negotiation particles, including those that contribute to
mood distinction, will be considered together as realising a system of
NEGOTIATION, a resource for indicating the speaker‘s stance or attitude in the
clause (see Table 4.10; see also Halliday & Mcdonald (2004: 341-342) on
chinese). as with the particles realising MODALITY, the meanings of negotiation
particles mostly vary across moods, but also based on the ‗affective loading‘
provided by different interactional contexts (Halliday & McDonald 2004:
342). Their categorisation and interpretation are therefore fuzzier, compared
with the account of modality above. Table 4.10 glosses their general
meanings. As the table shows, a few of the particles are restricted to indicative
clauses. Also, except forthe hesitative particle mɛ́ and the exhortative particle
na, which are clause initial particles respectively associated with indicative
and imperative clauses, the rest of the particles are clause final.
In spite of the fuzziness in the meanings of negotiation particles, it can
be generalised that negotiation in the indicative clause is mostly concerned
with espistemic stance, the degree to which the speaker is committed to the
knowledge claims of propositions and, in the imperative, negotiation mainly
modulates the proposal realised by the clause. The section will proceed to
discuss the different uses of negotiation particles in the indicative (Section
4.5.3.1) and imperative clauses (Section 4.5.3.2).
180
Table 4.10 Negotiation markers and their general meanings across moods
value particles meaning
indicative imperative
dɛ definitive
high ka assertive mild insistence
kaka strongly assertive strong insistence
kɛ, wɛ1 admonitive admonitive
ka mildly imploring
kaka strongly
imploring
median wɛ exclamative exclamative;
requestive
na (clitic forms: affirmative
n, a)
ɩ, e, ɛ non-affirmative prohibitive
bɩɩ opinative suggestive
na2 exhortative
bɩ neutral
interrogative
wɛ, kpo1 biased
interrogative
ʋ biased
low interrogative
ya biased
interrogative;
mirative
mɛ́2 hesitative
mɔ̀ counter-
expectation
yaa empathic adhortative
1
Particles that are sub-dialectal variants for realising the same meaning
2
Clause initial negotiation particle
4.5.3.1 Negotiation in the indicative mood: epistemic and affective stance
Attitudinal negotiation in the indicative mood can be divided into two main
types, based on the orientation of the negotiation, whether it is (1) proposition-
oriented or (2) interactant-oriented.
(1) Proposition-oriented negotiation: Proposition-oriented negotiation

indicates the speaker‘s attitude towards the propositional content of a clause.
Generally speaking, this type of negotiation represents different degrees of
assertiveness. For convenience and clarity of presentation they will be grouped
based on their semantic relatedness. The categories here comprise: (i) the
181
affirmative and non-affirmative, (ii) the various types of interrogative
negotiation, (iii) the definitive, assertive and strongly assertive, (v) the
exclamative, (vi) the opinative and (vii) the hesitative and counter-expectation
(see Table 4.10).
(i) Affirmative and non-affirmative: With affirmative and non-affirmative

negotiation, the speaker ends the clause by clearly assessing the polarity value
of the clause, which is signalled earlier in the verbal group by polarity
particles, as s/he is potentially about to hand over the turn to the addressee.
The Negotiator element for the affirmative is realised by the particle na
(example 115). For the non-affirmative, on the other hand, the Negotiator is
realised by one of phonologically variant particles, e, ɛ, or ι, depending on its
phonetic environment in terms of tongue root vowel harmony (example
116):13
(115) A: Fʋ wõ =n a lɛ [[ʋ na
2SG hear.PFV FOC DEF DEM she REL
yèl a]]?
say. PFV JUNC
‗You heard what s/he said?‘
B: Ɩ wõ a na.
1SG hear 3PL.NHM AFFR
‗I heard it.‘
(116) A: Fʋ bãw nι bom kãw?
2SG eat.PFV FOC thing some
‗You know something?‘
B: Ɩ bɛ bãw ɛ.
1SG NEG.IND.NFUT know NAFFR
‗I don‘t know.‘
As (115) and (116) show (both from St Maria play), the affirmative particle
asserts and negotiates the positive value of the clause while the non-
affirmative particle asserts and negotiates the negative value of the clause. In
13
Tongue root harmony is characteristic of West African languages. Morphemes constituting
one phonological word normally consists of only advanced tongue root ([+ATR]) vowels or
retracted tongue root ([-ATR]) vowels.
182
(116), for instance, polarity is realised by the particle bɛ in the verbal group
realising the Predicator. At the end of the clause, the speaker however
resonates the polarity value of the clause as an interpersonal punch, that is, to
establish the negotiatory value of the proposition, as s/he is potentially about
to hand over the turn to the listener. In (115), the polarity of the clause is zero-
marked and the affirmative particle na serves as a juncture prosody resonating
the positive value of the proposition.
(ii) Interrogatives: The various interrogative particles reduce the epistemic

force of the proposition and at the same time indicate the speaker‘s bias
towards the propositions for which they are seeking confirmation from the
addressee. The categories here comprise neutral interrogative (117), initiative
biased interrogative (118) and responsive biased interrogative (119). The
relevant particles are highlighted in bold in the dialogues below (all from St
Maria play). Other particles realising biased interrogatives are identified in
Table 4.10.
(117) A: A dɔɔ nɔŋ ya-rɛ na bι?

DEF man ADV be:mad-IPFV AFFR INT
‗Is the man possibly mad?‘

B: Ɩ bɛ bãw ɛ wɛ!
1SG NEG know.PFV NAFFR EXCL
‗I don‘t know!‘
(118) A: Nι dι na wɛ?
‗You have eaten, right?‘

B: Ʋʋ.
‗Yes.‘
(119) A: Ɩ zà a wa na wa nyɛ
1SG yesterday come AFFR PROX see.PFV
a ι pɔw-yaa ʋ bιɛrɛ.
DEF 1SG daughter 3SG be:sick.IPFV
‗Yesterday I came home to see that my daughter was sick.‘
183
B: Ʋ bιɛrɛ ya?
3SG be:sick.IPFV INT
‗She was sick, you say?‘
As the examples show, the interrogative clause is realised by simply adding an

interrogative particle to a declarative clause. The implication is that the
interrogative clause is a non-assertive proposition, a proposition where the
speaker relatively defers the epistemic claim to the listener. The biased
interrogative (example 118 & 119) signals some degree of commitment to the
epistemic claim of the proposition while the neutral interrogative (example
117) completely defers commitment to the listener. The two biased
interrogatives, initiative and responsive, show different orientations in the
exchange. The initiative flows from the speaker and thus indicates a higher
degree of commitment to the claim of the proposition while the responsive
realises an echo-question, often assessing the proposition affectively as a
surprise. Thus, in addition to its default interrogative uses, the particle ya has a
mirative meaning and both meanings are often enacted simultaneously (cf.
DeLancey (1997); Aikhenvald (2012) on mirativity).
Further, the interrogative particle ya is also used in ‗self-talk‘, where
the speaker signals to the listener that s/he is finding it difficult in cognitively
retrieving information from his/her consciousness. The following dialogue
from a casual conversation illustrates this phenomenon:
(120) A: Zan nʋ waar. […] Tɩ dɛ a dɛ

Zan IDENT.SG come.IPFV 1PL ADV AFFR ADV
fɩɩrɩ [laughter] […].

be:in trouble
‗Zan is the one coming … We are just in trouble …‘
B: Ʋ dɛ bɛ pawr a mɩ̀ ɛ. Ɛcɛ
3SG ADV NEG.IND.NFUT get.IPFV 3PL.NHM also NAFFR but
a a a… ãa lɛ nɩ a na ya?
DEF DEF DEF who COP FOC DEF DEM INT
‗S/he is just not getting some of it. And the the the … who is
this even?‘
184
C: A Gɔbr.
DEF Gobr
‗Gobr.‘
B: Ʋʋ.
‗Yes.‘
The material situation surrounding this exchange is a family get-together,
where a group of relatives are sharing a bottle of wine under a tree in front of
their house. A neighbour sees them and walks towards the scene to join the
celebration. Speaker A alerts the group of the additional party, indicating the
inadequacy of the wine for an extra mouth. Speaker B, on the other hand,
inquires of the whereabouts of his younger brother who is presently absent at
the scene, and for a moment, cannot remember his name. The particle ya
added to the clause, which is already a ‗wh‘-interrogative, simply indicates
that the proposition is not a genuine question but rather a self-reflection. Such
clauses are also commonly used as rhetorical questions showing surprise or
unexpectedness at seeing, for instance, a long ‗lost‘ friend. Thus, ya clearly
lies between the boundary of an interrogative marker, obligatory as it were,
and an optional attitudinal marker. Further research is needed to investigate
the full range of its uses.
(iii) Definitve, assertive and strongly assertive: First, definitive negotiation

gives a high degree of force to a proposition. It adds an emphatic punch to the
proposition realized by the clause, thereby showing a high epistemic
commitment on the part of the speaker. It is realised by the particle dɛ. An
example is the elliptical clause Ziri dɛ! (Lie!) in Speaker C‘s turn below. The
extract is an episode in a biblical drama (The Story of Jesus) where the Jewish
chief priest cautions his elders about Jesus‘s popularity:
(121) A: Nɩbɛ yaga zie de ʋ na ʋ

person.PL many place take 3SG AFFR 3SG
ɩ bɛ nàa baarɩ kɛ.

COP.PFV 3PL.HM king finish. PFV ADM
Many people have accepted him to be their king already, how

careless you are!‘
185
B: ||| Nàa ya? Nɩ-baalbɛ, pɔw-ƴɛmɛ, nyanyuuru
king INT person-sick.PL woman-barren.PL thieves
nàa? ||| Bɛ mɩ́ saa nɩ
king 3PL.HM HAB show.PFV FOC
dɩa, || dãw-bio bɔr || bɛ

today tomorrow disappear.PFV 3PL.HM
yiiri bɛ bɛr. |||

forget.PFV 3PL.HM leave.PFV
‗King?‘ King of sick people, barren women and thieves? They
appear today and disappear tomorrow (and) people forget about
them.‘
C: ||| Ziri dɛ! ||| A ʋ yuor yire
lie EMP DEF 3SG name come:out.IPFV
na dɔwlɛ bibie za || ɛ a
AFFR increase.IPFV days all CONJ DEF
nɩbɛ mʋʋrɛ ʋ mɩ̀. |||

people praise.IPFV 3SG ADV
‗Lie! He is getting popular every day, and the people are

praising him too‘.
A: ||| Siza na. ||| Nàa nʋ. ||| Ɛcɛ damnʋ
Truth IDENT.PL King IDENT.SG CONJ trouble
dɛ wa lɛ bere a tew pʋɔ
ADV COND be:again be.IPFV DEF town inside
a, || nyιmɛ so a sãwna. |||
JUNC 2PL.EMP possess DEF blame
‗It is true. He is king. But if any trouble just occurs in the town
again, you will have the blame.‘
This dialogue is rich in negotiation. The chief priest (A) admonishes his elders
for their inability to control the activities of Jesus, leading to his popularity
among the people. In the first clause, he uses the admonitive particle kɛ to
signal his disappointment and sound caution to the elders. In essence, the
particle kɛ prosodically enacts the whole statement as an admonition rather
than a bare statement of fact. In defence, one of the elders (B) tries to mitigate
186
the seriousness of the situation. Note his use of the biased interrogative
particle ya not only to query the proposition made by the chief priest‘s but also
to show surprise at his statement. Of particular interest here is the underlined
clause, where another elder (C) challenges the position of his colleague with a
definitive negotiation particle, signalled by the particle dɛ. This particle
colours the proposition it attaches to as conclusive. In other words, the speaker
lay a strong epistemic claim to the proposition. In general, the extract clearly
exemplifies negotiation in action in the flow of discourse.
The next proposition-oriented negotiation to be discussed consists of
those that have been labelled assertive and strongly assertive in Table 4.10. On
a scale of degree of force, they are less strong than the definitive negotiation
although they are also of high epistemic value. A mildly assertive negotiation
is realised by the particle ka while a strongly assertive negotiation is realised
by the particle kaka. Examples are given in the following constructed
examples:
(122) Ʋ wa na ka.
3SG come.PFV AFFR M.INS
He has come, I insist.

(123) Ʋ bɛ wa ɩ kaka.
3SG NEG.IND.NFUT come.PFV NAFFR S.INS
‗He has come, I strongly insist.‘
The interactional contexts where a speaker typically uses these two assertive
forms is to react to or challenge a proposition made by the addressee by
insisting on the truth value of the proposition the particle attaches to.
(iv) Exclamative: Another negotiation type in the proposition-oriented

category is the exclamative. Exclamation is realised in Dagaare only as a form
of attitudinal stance, unlike in many Indo-European languages, where it can be
realised by a sub-type of the declarative mood (e.g. How gracefully she
walks!) (cf. Halliday & Matthiessen (2014: Ch. 4) on English). With the
exception of the interrogative clause, any Dagaare clause can be turned into an
exclamative by adding the particle wɛ to it. An example is the utterance by
187
Speaker B in example (117). The dialogue below gives other illustrations
(from The Story of Jesus):
(124) Jesus: Pɔw ɩ, a fʋ baalʋ sanι

woman VOC DEF 2SG sickness heal.PFV
na.
AFFR
‗Woman, your sickness is healed.‘

Voice 1: Wi! Alɛ! Ʋ sanι na wɛ!
INTJ INTJ 3SG heal.PFV AFFR EXCL
‗What! She is healed!‘

Voice 2: Ʋʋ wɛ! Nɔ-ɓaa yél =ʋ. Nɩ
yes EXCL shocking matter IDENT 2PL
wa nyɛ a!
come.PFV see.PFV PRT
‗Yes! It is a miraculous incident. You come and see!‘
Example (124) is an extract from a movie episode on a healing miracle

performed by Jesus. Here, the exclamative particle occurs in two clauses, the
first being an affirmative clause (Voice 1) and the second being a minor clause
(Voice 2). The exclamative particle in these clauses shows the surprise and
awe feltby the speakers at the miraculous healing of the sick woman. As
Moutaouakil (1999: 7) observes of exclamation, it ―signals the speaker‘s
evaluation of [their] attitude towards the content of the linguistic expression
with the peculiarity that the source of the evaluation is the impression made on
the speaker by this content.‖ The realisation of exclamation as a modal
assessment in Dagaare rather than a mood type resonates with Moutaouakil
(1999) cross-linguistic characterisation of exclamation as a kind of modal
assessment. He notes that while declarative, interrogative and imperative
clauses have structural properties that distinguish them across languages,
exclamations can take variant forms even within the same language. As (124)
show, exclamation in Dagaare has no unique structural realisation, only being
indicated optionally as the speaker‘s attitude encoded in the clause (also see
Section 3.2 for instances of exclamation in the imperative clause).
188
(v) Opinative: The opinative is realised by the particle bɩɩ and enacts the
proposition realised by the clause as the speaker‘s opinion on an issue or as
his/her personal conviction. An example is given in the following dialogue
form (from St Maria play):
(125) Son: Mãa lɩɛbɛ nɩ faara o! [...]

1SG.EMP turn.PFV FOC priest PRT[…]
‗I am becoming a priest! …‘
Father: Bʋnʋ ya? A sukuul ɩ na yaw fʋ,
What INT DEF school 1SG REL put.PFV 2SG
fʋ bɛ zawrɩ ɛ? [...]
2SG NEG.IND.NFUT refuse.PFV NAFFR
‗What? The school I put you in, didn‘t you stop?‘

Fʋ manɛ a faara mɩnɛ na mɩ
2SG assume.PFV DEF priests PL REL HAB
zɩ a lɛ a, zu gbɩlɩ lɛ
sit.PFV DEF DEM PRT head.PFV knots COP
bɛ? […]
3PL.HM
‗You assume the priests who sit like that, they are uneducated?‘
Son: Ɩ na lɩɛb =a. Bɛ na
1SG POS.IND.FUT turn.PFV AFFR 3PL.HM POS.IND.FUT
de =m =a bɩɩ.
take.PFV 1SG.ACC AFFR OP
‗I will become (one). They will admit me, in my opinion.‘
In this extract, the father‘s discouraging response to his son‘s declaration of

becoming a priest implies that the son will not be admitted to the priesthood
since he is a school dropout and, for that matter, uneducated. Note again the
use of the biased interrogative particle ya by the father to show surprise or
unexpectedness at his son‘s statement. In reaction to the father‘s comments,
the son explicitly enacts his conviction in the final clause with the opinative
particle bɩɩ. It is also worth pointing out that the particle o in the first clause in
(125) has negotiatory value although, unlike the other particles, this particle is
paralinguistic juncture prosody and has no specific meaning. Its function is to
189
amplify the volume of the utterance. In fact, its use is not limited to Dagaare
and would be familiar to speakers of West African languages in general. 14
(vi) Hesitative and counter-expectation: Next, hesitative and counter-

expectation particles are associated with interrogative clauses. The hesitative
is indicated by the clause initial particle mɛ́. An illustration is given in (126)
and (127) below:

Mɛ́ tɩ be =n a ka a mʋtɔw
HST 1PL EXIST.PFV FOC DEF here DEF afternoon
za wɛ?
all INT
‗I believe (all of us) have been here since this afternoon, right?‘
(127) St Maria play

A: A libie nɩ ana.
DEF money FOC these
‗This is the money.‘
B: Na dɔm! Na dɔm =ɩ tɩ nyɛ
EXH squat EXH squat COM 1PL see.PFV
a! […]
PRT
‗Please squat! Please squat with it and let‘s see!‘

A: Mɛ́ a ta na wɛ?
HST 3PL.NHM reach AFFR INT
‗I believe it has reached (the agreed amount), right?

B: Ɩ bʋɔlɩ ɩ pɔw-yaa ʋ wa
1SG call.PFV 1SG daughter 3SG come.PFV
sɔr kaa nyɛ.
count.PFV check.PFV see.PFV
‗Let me call my daughter to come and count and see.‘
As the examples show, the hesitative particle is an epistemic downtoner. The

speaker hedges the proposition by hesitating in claiming full knowledge of it,
14
I have also observed the use of this juncture prosody among Cantonese speakers.
190
inviting the addressee to confirm the proposition. Thus, it often co-occurs with
the biased interrogative particle wɛ (and also ʋ or kpo) (see Table 1) as the
examples indicate. Semantically, the hesitative particle simultaneously
indicates both the speaker‘s bias and uncertainty towards the proposition.
Because of this, even if it occurs in a clause without an interrogative particle,
the clause mostly still carries the sense of a biased question (e.g. Mɛ́ tɩ be =n a
ka a mʋtɔw za, ‗I believe all of us have been here since this afternoon‘). The
extracts below, especially the dialogue in (129), illustrate this phenomenon
further:

||| Mɛ́ nyɩ taa na lɔb zʋkpar na
HST 2PL MOD POS.IND.FUT throw proverb DEM
kʋ =m: || “dɔkta a, sanɩ fʋ tʋɔra!” |||

give.PFV 1SG.ACC doctor PRT heal.PFV 2SG self
―I believe you will tell me this proverb: ‗doctor, heal yourself.‘‖
Agric Officer: […] A tɩ kɔb [[tɩ na mɩ́
[…] DEF 1PL farming 1PL REL HAB
kʋɔr a]], mɛ́ tɩntɛr =ʋ tɩ

farm.IPFV JUNC HST plains IDENT.SG 1PL
mɩ́ kɔ yaga.
HAB farm.PFV plenty
‗… Our farming we do, I believe it is plains we cultivate
more.‘
Host: mmm
INTJ
‗Yeah.‘
Agric Officer: Soo, bɛ kɔ =n tɩntɛr.
so 3PL.HM farm.PFV FOC plains
‗So they cultivated plains.‘
Generally, as Yang and Yap (2015: 54) characterised a Mandarin Chinese

stance marker, the hesitative particle is used when the speaker is quite certain
191
about the epistemic claims of the proposition, yet, at the same time, is
reluctant to come across as being assertive.15
Counter-expectation, on the other hand, is the grammatical encoding of irony,
as it were. What is negotiated in the clause could be either a situation the
speaker is reacting to or the propositional content of the clause itself. In other
words, it either indicates (i) whether the situation engendering a question is
opposite to the expectation of the speaker or (ii) that the speaker believes the
opposite of the propositional content of a question. It is realised by the particle
mɔ̀, illustrated in (130) below (from The Story of Jesus):
(130) Blind man: Alɛ ka! Bʋʋ nʋ? Bʋʋ nιɛ

wait M.INS what IDENT.SG what even
nʋ mɔ̃̀?
IDENT.SG CE
‗Wait, I implore you! What is even happening?‘

Man: Nazaratι tew Yeezu nʋ tɔlɛ a
Nazareth town Jesus IDENT.SG pass.IPFV DEF
ka.
here
‗It is Jesus of Nazareth who is passing here.‘
The use of mɔ̀ in (130) is an instance of an assessment of the situation

engendering the question as counter to the speaker‘s expectation. The extract
is from the episode on the healing of the blind Bartolomeo by Jesus. The blind
man hears an unusual commotion and movement of a large crowd along the
road where he sits in quest of arms. He approaches a fellow closer to him and
queries about the unusual drama unfolding before him. The particle mɔ̀, which
punctuates the interrogative clause signals the oddity of the situation. Another
example is given in (131) from the same text, but this time showing counter-
expectation to the propositional content of a clause (complex):
(131) ||| Yèl tɩ nyɛ! ||| A zuru yab [[tɩ na

say 1PL see.PFV DEF head-PL pay-NMLZ we REL
15
Yang and Yap (2015) examined the Mandarin verb kǒngpà (‗I‘m afraid‘) as a hedging
device. Comparatively, the Dagaare hesitative particle, which roughly translates as ‗I believe‘,
has a weaker epistemic force.
192
ya-rɛ kʋ-rɛ a nàa Sɩzaar a]], a
pay-IPFV give-IPFV DEF King Caesar JUNC 3PL.NHM
sɛw na || bɩɩ a
be.acceptable.PFV AFFR CONJ 3PL.NHM
bɛ sɛw mɔ̃̀? |||

NEG.IND.NFUT be:acceptable.PFV CE
‗Tell us!‘ The taxes we pay to Caesar, is it even acceptable or it‘s not
acceptable?‘
In this extract, the Jewish elders approach Jesus to inquire about the
appropriateness of the payment of taxes to Caesar. The use of the counter-
expectation particle in the alternative interrogative clause marks the
proposition as counter to the expectations of the speaker. However, given that
the question is realized by an alternative interrogative, the use of the counter-
expectation particle makes it confusing as to which of the two alternatives the
speaker is dissatisfied with. Thus, while the particle prompts the addressee that
the speaker disprefers one of the alternatives, it does not make the preference
explicit. It simply encodes the speaker‘s negative attitude in the complex.
(2) Interactant-oriented negotiation: Interactant-oriented negotiation, on the

other hand, indicates the speaker‘s inter-subjective attitude towards the
addressee. It comprises admonitive and empathic negotiation. The admonitive
is realised by the particle kɛ or wɛ, depending on the sub-dialect. An example
is the first clause in example (121), which has already been discussed: Nɩbɛ
yaga zie de ʋ na ʋ ɩ bɛ nàa baarɩ kɛ (‗Many people have accepted him to be
their king already, how careless you are!‘). Here, the admonitive particle kɛ
prosodically enacts the whole statement as an admonition rather than a bare
statement of fact.
Empathic negotiation, as the name suggests, signals that the speaker
empathises with the addressee or is emotionally involved in the proposition
realised by the clause in some way. It is realised by the particle yaa, illustrated
below by constructed examples:
(132) A mama na wa na yaa?

DEF mum POS.IND.FUT come.PFV AFFR EM
193
‗Mum will come, alright?‘
(133) A: Ɛ Zan wa?
CONT Zan be:where
‗And where is Zan?‘
B: Wi! Ʋ kpi na yaa.
INTJ 3SG die.PFV AFFR EM
‗Oh! S/he is dead, sorry for the shock.‘
Example (132) is a typical consoling statement to a child crying for his/her

absent mother while example (133) gives a scenario where a bereaved relative
responds to a question on the whereabouts of the deceased. In both clauses, the
empathic marker indicates that the speaker is emotionally involved in the
proposition by empathising with the addressee. While in (132), the speaker
signals endearment towards the worried child, in (133), the speaker anticipates
the shocking effect of the proposition to the listener and indicates his/her
empathy. Generally, the emphatic particle features prominently in care-taker-
talk, where is it used as an endearment marker (as in (133)).
4.5.3.2 Negotiation in the imperative mood: modulating proposals

This section proceeds to discuss negotiation in the imperative mood. As
mentioned earlier, negotiation here modulates the force of the proposal by
either toning it down, enforcing it or indicating disinterest on the part of the
speaker. This modulation strategy maps out a semiotic space consisting of
more delicate sub-types of the imperative. They can broadly be classified into
(1) hortative and (2) non-hortative types.
(1) Hortative negotiation: Hortative negotiation generally encourages or

exhorts the addressee to bring the proposal by the clause into effect. In a sense,
it modulates the force of the proposal by negotiating obligations. Two
hortative types are identified: (i) adhortative and (ii) exhortative. These are
semantically distinguished based on the degree to which the speaker can
enforce the proposal. With the adhortative, the speaker positions her/himself
as relatively less able to enforce the proposal. The addressee is empathised
194
with and encouraged or urged to bring about the goal of the proposal. The
Negotiator is realised by the particle yaa:
(134) Ta kone ɩ yaa!

NEG.IMP cry.IPFV NAFFR ADH
‗Don‘t cry, okay!‘
The exhortative, on the other hand, strongly encourages or exhorts the

addressee to bring about the goal of the proposal. It is marked by the clause
initial particle na, immediately preceding the Predicator. Examples are given
below (from St Maria play):
(135) Husband: Na bɔ bʋndɩrɩ kʋ̀

EXH search.PFV food give.PFV
mãa!
1SG.EMP
‗Please, get me food! (and let‘s stop talking about this

issue)‘
Wife: Anʋ bɔ bʋndɩrɩ kʋ̀ fʋ? [...]
who search.PFV food give.PFV 2SG
‗Who get you food?‘

Husband: Bʋʋ so fʋ yèle a lɛ?
what owns 2SG say.IPFV DEF DEM
‗Why are you saying so?‘

Wife: Fʋ kʋ̀ =m =ɩ lilir ɩ?
2SG give.PFV 1SG.ACC FOC money PRT
‗Have you given me money?‘
(136) Ɩ pɔw-yaa na ƴɛ̀rɛ ɩ zie!
1SG daughter EXH speak.IPFV 1SG place
‗My daughter, please speak to me! (and ignore your father)‘
As the English translations in (135) and (136) show, the exhortative particle is
used in contexts where the proposal the speaker is making is competing with
other topics for the attention of the listener (example 135) or where the
speaker is competing with other voices for the attention of the listener
(example 136). The particle na is thus used to exhort the listener to align with
195
the speaker‘s interest. An observation of interactions in casual conversation
reveal that the use of this particle can be offensive to the non-addressee
participant whose interest the listener is exhorted to ignore. A case in point is
(136), where a mother exhorts the daughter to continue their conversation and
ignore an interruption from the father.
(2) Non-hortative negotiation type: The non-hortative types comprise (i)

prohibitive, (ii) requestive / exclamative, (iii) admonitive, (iii) insistent /
imploring and (iv) suggestive negotiation. (i) The prohibitive negotiation
corresponds to the non-affirmative in the indicative mood and is realised by
the same set of phonologically variant particles (see Table 1). For want of
space, it will not be discussed further here.
(ii) Requestive / exclamative: The requestive entreats or requests the

addressee to ensure the success of the proposal realised by the clause. The
Negotiator is realised by the particle wɛ, as exemplified by the underlined
clauses in the extract below (from The Story of Jesus):
(137) A: Fʋ bɔbr kɛ tι ι ŋmιn?

2SG want.IPFV PROJ 1PL do.PFV what
‗What do you want us to do?‘
B: Faa tι wɛ!
save.PFV 1PL EXCL
‗Save us, please!‘

C: Yèl wɛ! Tι ι ŋmιn?
say. PFV EXCL 1PL do.PFV what
‗Tell (us), please! what do we do?‘
These utterances are made by a restless crowd following John the Baptist in a
biblical drama. The requestive clauses that realise Speaker B and C‘s
proposals entreat him to answer their questions. In the Bartolomeo extract
introduced earlier (example 130), the blind man (A below) employs the
requestive to entreat Jesus for healing. The extract is repeated below in an
extended form (The relevant clause is underlined):
196
(138) A: Alɛ ka! Bʋʋ nʋ? Bʋʋ nιɛ
wait M.INS what IDENT.SG what ADV
nʋ mɔ̀?
IDENT.SG CE
‗Wait! What is it? What is even going on?‘

B: Nazaratι tew Yeezu nʋ tɔlɛ a
Nazareth town Jesus IDENT.SG pass.IPFV DEF
ka
here
‗It is Jesus of Nazareth who is passing here.‘
A: Oyi! Yeezu u! Oyi! Yeezu u!
INTJ. Jesus VOC INTJ. Jesus VOC
Daviir bi-dɛb ι, zɔ a ι nι-baalʋ

wɛ!
EXCL
David child-male VOC run.PFV DEF 1SG pity

‗Oh, Jesus! Oh, Jesus! Son of David, have pity on me, please!‘
Supplicatory contexts such as these are typical environments of the requestive

Negotiator.
In addition to its requestive use, the particle wɛ is also used in the
imperative mood to realise exclamation (cf. Section 3.1). However, the use of
wɛ as an exclamation marker in the imperative is limited. It is only used with
the perceptive verb nyɛ (‗see‘) as in the following extract from a casual
conversation:
(139) Nyɩnɛ na fʋ dɛ ŋmʋrɛ a lɛ

where IDENT.PL 2SG ADV rush.IPFV DEF DEM.DIST
cere nɩ? Nyɛ wɛ!

go.IPFV COM see.PFV EXCL
‗Where is it that you are just dashing like that to? Look at that!
Here, Nyɛ wɛ! (‗See!‘) is not an invitation for the listener to engage in a
perceptive process, but rather an indication of the speaker‘s disapproval of the
listener‘s careless behaviour, signalled in the preceding clause. The perceptive
197
verb nyɛ (‗see‘) has been recruited for the realisation of exclamations and it
has almost lost its sense of perception in this environment. Below are other
exclamation clauses where it is used other than those with the exclamative
particle (from The Story of Jesus):
(140) Nɩ nyɛ a nɩɩ!

2PL see.PFV 3PL.NHM PRT
‗You look at that (= Look at what you have caused)!‘

(141) Fʋ bɛ nyɛ yɛrʋ!
2SG NEG see.PFV speech
‗What a speech!‘
Examples (140) and (141) are instances of recurrent formulaic expressions.

Example (140) is a depreciative exclamation while (141) is an appreciative
exclamation (cf. Moutaouakil, 1999: 8-10). With the exception of yɛrʋ
(‗speech‘) in (141), all the individual words in these two clauses are
semantically empty, they are non-referential. The pronouns are impersonal and
have no specific reference and the verb nyɛ (‗see‘) does not necessarily
indicate perception by sight in clauses such as (141). A more congruent
realisation of the process here would be wõ (‗hear‘) rather than nyɛ (‗see‘)
since what is perceived is sound rather than a physical object. Thus, while
(140) is used to indicate the speaker‘s frustration towards the listener, (141)
shows the speaker‘s admiration of the speech alluded to.
(iii) Admonitive: The admonitive negotiation in the imperative mood has a

similar meaning as in the indicative mood. But here it either cautions or urges
the addressee against the event or action represented by the Predicator/Process
(example 142) or indicates that the speaker is indifferent whether the proposal
is enforced or not. That is, it is the addressee‘s own business (see example
143). As mentioned earlier, there is a sub-dialectal variation in the realisation
of the Negotiator. Mostly, speakers in Burkina Faso use a distinctive particle
kɛ, while speakers in Ghana use the particle wɛ, which has the same form as
the exclamative particle. Thus, among Ghanaian speakers, the admonitive
imperative clause has the same form as the requestive and the two can only be
198
distinguished in context. Illustrations are given in the constructed examples
below:
(142) Zɩ wɛ/kɛ! Bɛ dɩrɛ na.

sit.PFV ADM 3PL.HM eat.IPFV AFFR
‗You sit! They are eating.‘ (You won‘t have any food to eat).‘
(143) A: Ɩ wa bɩ?
1SG come.PFV INT
‗Should I come?‘
B: Wa wɛ/kɛ!
come.PFV ADM
‗Come! (It‘s your business; I don‘t care).‘
(iv) Insistent / imploring: The particle ka (mild) or kaka (strong) in the

imperative clause can be interpreted as realising insistence or imploring,
depending on tenor and/or the nature of the demand imposed on the listener. In
the following clauses, ka is interpreted as imploring the listener to carry out
the command since, in each case, the listener is in a more privileged or
powerful position than the speaker.
(144) The Story of Jesus

Alɛ ka!
wait M.INS
‗Wait, I implore you!‘

(145) St Maria play
Pɛw mɛ ka! Ɩ dɩ =n san.
lend.PFV 1SG.ACC M.INS 1SG.NOM owe.PFV FOC debt.
Mãa cen, ʋ kʋ kʋ̀ mɛ
1SG.EMP go.PFV 3SG FUT.IND.NEG give.PFV 1SG.ACC
ɩ.
NAFFR
‗Lend me, I implore you! I have a debt.
Example (144) is from the blind Bartolomeo to a passer-by and (145) is from
an insolvent husband to his apparently well-to-do wife. A similar utterance
from father to son such as Wa kaka! (‗Come, I insist!‘) will be interpreted as
199
insistence. Thus, while in the declarative mood, ka and kaka boost the
speaker‘s commitment to the knowledge claims of the proposition, in the
imperative, they modulate the speaker‘s demand for goods-&-services.
(v) Suggestive: The final negotiation type to consider in the imperative mood
is the suggestive, realised by the clause final particle bɩɩ. It is illustrated by the
following constructed dialogue:
(146) A: Ɩ sɛ a nɛn?
1SG roast.PFV DEF meat
‗I (should) roast the meat?‘
B: Dʋw bɩɩ.
boil OP
‗Boil it, I suggest / I think!‘
As the exchange indicates, this particle is always used in reaction to a

preceding utterance and it signals the imperative clause as an alternative
course of action.
4.5.3.3 Negotiation concord

To recapitulate, the particles realising negotiation relatively have different
meanings in the indicative and imperative moods. In the indicative mood, they
mostly enact the speaker‘s assessment of the epistemic claim of propositions.
In the imperative, they indicate the speaker‘s assessment of tenor and the force
of proposals. These are, however, relative characterisations. For instance,
interactant-oriented negotiation in indicative clauses are also motivated by
tenor. Generally, the particles realising negotiation are juncture prosodies that
extends the indicative and imperative mood distinctions in delicacy. However,
more than one attitudinal marker can be found in a single clause, at most up to
three. In the indicative mood, this often happens in polar interrogative clauses,
which normally combine the affirmative or non-affirmative particle with the
selected interrogative particle (see Table 4.10). Also, a clause which already
has an obligatory Negotiator for mood contrast can take an optional Negotiator
element (see examples 132&133, for instance).
200
This section is particularly concerned with a situation where more than
one optional Negotiator are used in a clause to amplify negotiation. This
normally involve a clause initial particle and a clause final particle. I refer to
this phenomenon as negotiation concord to highlight the semantic resonance
between the co-occurring particles. In the indicative mood, such co-occurrence
consists of the clause hesitative particle mɛ́ and a variant of the initiative
biased interrogative particle (see Table 4.10). An example of this has already
been introduced in examples (126) and (127) in the preceding section. Both
particles indicate the speaker‘s uncertainty towards the epistemic claim of the
proposition (cf. Section 3.1). The following combinatory possibilities are
identified in the imperative mood:
(i) Na ………………………. ka. (only the imploring sense)

(ii) Na ……………………. kaka. (only the imploring sense)!
(ii) Na ……………………… wɛ.(the requestive sense)!
(iii) Na ……………………… yaa!
(iv) Na ……………………… bιι!
An illustration is given below by a dialogic musical interlude in a traditional

folktale about a wild duck and her paralytic duckling (Gbɛr-be-yeni Wʋba,
‗One-leg Paralytic‘; repeated lines are omitted):
(147) Duckling: Ɩ ma a, ɩ ma a, bɛ
1SG mother VOC 1SG mother VOC 3PL.HM
'wɔbr fʋ na wɛ!
chew.IPFV 2SG AFFR EXCL
‗My mother, my mother, they are eating you!‘

Duck: Ɩ bie e, ɩ bie e, na
1SG child VOC 1SG child VOC EXH
Ta kone ι wɛ!
NEG.IMP cry.IPFV NAFFR EXCL
‗My child, my child, please don‘t cry!‘

Mãa nιɛ pιrɛ yɛryɛryɛr ƴawnɛ
1SG.EMP ADV struggle.IPFV IDEO give.IPFV
201
bɛlɛ nɛ.
3PL.HM.EMP DEM.DIST
‗For I have earnestly been feeding those others.‘

Bɛlɛ nιɛ bɛ kone.
3PL.HM.EMP ADV NEG.IND.NFUT cry.IPFV
ι Ɛ fʋʋ pɑa nιɛ kone.
NAFFR CONJ 2SG.EMP ADV ADV cry.IPFV
‗They are not even crying and now you are crying.‘
Gbɛr-be-yeni wʋba a, na ta
leg-one paralytic VOC EXH NEG.IMP
kone ι wɛ!
cry.IPFV NAFFR EXCL
‗One legged paralytic, please don‘t cry!‘
The story relates that Mother Duck refused to feed her paralytic duckling due
to its deformity but took very good care of her healthy ducklings. When
Mother Duck was attacked and killed by a group of hunting boys, all the
ducklings scattered, except the paralytic one, who followed the boys wherever
the mother was taken. In this exchange, the duckling employs the exclamative
marker to indicate her desperation at her mother being eaten. Of particular
interest here is the combination of the exhortative particle na and the
requestive marker wɛ in Mother Duck‘s response. With this negotiation
concord, she negotiates the proposal at both ends of the clause to strongly
show her remorse and empathise with and entreat the duckling to stop crying.
This semiotic prosody established by negotiation is in harmony with
the use of other interpersonal resources in the interlude, namely the choice of
Vocatives, the use of emphatic pronouns for Subject person in Mother Duck‘s
turn (i.e. mãa, ‗I‘, bɛlɛ, ‗they‘, fʋʋ, „you‘) the foregrounding of mood Adjuncts
(i.e. nιɛ, pɑa ‗even‘), and the exclamative marker in the duckling‘s turn. This
conspiracy of interpersonal resources creates an emotionally charged semiotic
context, the social function of which is to evoke the right emotions in the
young audience and inject in them the moral import of the story (cf. Martin
(1988) on grammatical conspiracy).
202
4.6 Conclusion
This chapter has explored the clause as a unit for enacting interpersonal
meaning among Dagaare speakers, especially in dialogic interaction. It first
discussed the nature of dialogue and the general speech functions that are
enacted in verbal exchanges. It then examined the interpersonal structure of
the Dagaare clause, identifying the Mood base and Residue as two main
components of the clause. the chapter also examined the different mood types
that realise the system of SPEECH FUNCTION. the phenomenon of mood
metaphor has also been examined. Finally, the chapter discussed polarity and
the modal assessment systems of MODALITY and NEGOTIATION.
203
CHAPTER FIVE
THEME AND INFORMATION
5.1 Introduction
In Chapter 4, we discussed resources available to the Dagaare speaker for

enacting the clause as an arguable and negotiable move in discourse. In this
chapter, we will examine the textual systems of the Dagaare clause, that is,
grammatical resources that are dedicated to the organisation of the clause as a
message. The chapter first discusses the conception of the clause as a message
structure, drawing on insights from the extant literature (Section 5.2). This is
followed by a discussion of two key textual systems of Dagaare, the systems
of THEME (Section 5.3) and INFORMATION (Section5.4).
5.2 The Concept of the Clause as Message
The concern with the clause as a message structure is perhaps the earliest
recognition of constituent structure in language. In his study of the grammar of
ancient Greek, Plato divided the sentence (lógos) into parts of speech (mérē
lógou), comprising what he called ónoma and rhē̄̂ma (cf. Robins 1966;
Halliday 1977). This division is the antecedent of the terms Theme-Rheme or
Topic-Comment in contemporary analysis of the textual composition of the
clause. It was Aristotle, and later the Stoics, who began to expand Plato‘s
conception of parts of speech into what have come down to us as word classes.
The development of the notions of Theme and Rheme in contemporary
linguistic science is subsequently owed to the Prague School of linguists,
especially Vilém Mathesius and also Frantisek Daneš and Jan Firbas, who
developed them in the theory of Functional Sentence Perspective (FSP) in the
1950‘s (cf. Garvin 1964; Daneš 1974; Firbas 1992). Following his initial
research of Chinese in the 1950‘s, Michael Halliday, adopted a system-based
approach to thematic analysis as part of his systemic functional account of
English (Halliday 1967a, b, 1968), expanding this further to include an
account of information focus as a separate but complementary system to the
thematic orientation of the clause (see e.g. Halliday 1970; 1979; Halliday &
204
Greaves 2008; Halliday & Matthiessen 2014). This complementary
perspective has been usefully applied to the description of many languages,
and Matthiessen (2004; cf. Section 10.5) gives a typological overview of the
THEME and INFORMATION systems of the world‘s languages. in other
contemporary linguistic traditions, THEME and INFORMATION have not
beenclearly separated as complementary systems and have been investigated
under the general heading of information structure (e.g. Lambrecht 1994,
1996; Schwabe & Winkler 2007; Zimmermann 2010), and topic and focus
(e.g. Gundel 1974; Givón 1975a, 1983; Good 2010).
In African languages, studies on information structure have been
skewed to information focus systems (but see a recent study by Abdel-Hafiz
(2015) on Kunuz-Nubian); developing descriptive, experimental, typological
and grammaticalisation accounts of focus resources across languages (e.g.
Heine & Reh 1983; Rialland & Robert 2001; Hartmann & Zimmermann 2007;
Schwarz 2009; Zimmerman 2011; Güldemann et al . 2015). Investigations
have also been conducted on the role of prosody in information focus in some
languages (see Güldemann et al . (2015) for a recent review). The prominence
of information focus in studies on African languages is arguably because it is a
grammatical phenomenon that is often marked morphologically in many of
these languages. Thematic structure, on the other hand, is a covert
grammatical category across many languages. As the history of the science of
language has however taught us, a systematic analysis of discourse often
reveals striking grammatical reality of cryptic categories (cf. Whorf 1956;
Halliday 1967a, 197b, 1968; Fillmore 1968).
In this chapter, the discussion of the Dagaare clause as a message
structure will draw on insights from the conceptual and descriptive insights
noted above. Specifically, it takes a system-based (or a holistic) approach to
theme (Section 5.3) and focus (Section 5.4) in Dagaare, and shows how they
complement each other in the textual development of the clause.
5.3 Theme and Rheme

The first textual resource of dagaare grammar to be discussed is the system of
THEME. this is a system of textual prominence given to the initial element(s) in
205
the clause. in dagaare, as in many other languages, the system of THEME is
related to a distinct status of prominence that is assigned to the initial position
of the clause by postulating it as the orientation of the clause, a point of
departure in the flow of meaning (cf. Halliday & Matthiessen 2014). This
Table 5.1. Thematic analysis of a text [reporting: recounting – monologic

– written]
clause clause
no. Theme Rheme
[1] A Adama nɩ tɩ dɔw nɩ bibiir ayi:
PST.REM born FOC children two:
DEF Adam and Kayɛ nɩ Abɛl.
a Awa Cain and Abel
DEF Eve
[2.1] A Abɛl wa ɩ =n pi-cɩɩnɛ,
[2.2] ɛ a Kayɛ ɩ kʋɔra.
COP.PFV farmer
and DEF Cain
[3.1] A Kayɛ poru tɩ maal =ɩ vla
DEF Cain farms PST.REM do.PFV FOC well
[3.2] a ci yawmɛ.
be:plenty
DEF guinea corn
[4.1] Ʋ tɩ ir =ɩ a
3SG PST.REM remove.PFV FOC DEF
ci
guinea corn
[4.2] k‟ʋʋ pour =ɩ Naaŋmɩn barka.
CONJ.3SG greet.PFVFOC God gratitude
[5.1] A Abɛl tɩ nyɔw nɩ a

DEF Abel
PST.REM catch.PFV FOC DEF
ʋ pelé kãw
his lamb some
[5.2] maal =ɩ bawr kʋ̀
make.PFV FOC sacrifice give
a Naaŋmɩn.
DEF God
‗Adam and Eve gave birth to two children: Cain and Abel. Abel
became a shepherd and Cain was a farmer. Cain‘s farms did
well (such that) the guinea corn was plenty. He took some of the
206
guinea corn in order to offer thanks to God. Abel caught one of
his lambs and made sacrifice to God.‘ (Source: Sɛb-Sow Ƴɛr-bie
1996)
textual grounding of the clause is its Theme element. The Theme combines
with the remainder of the clause, the Rheme, to constitute a message, a
miniature text, as it were, and every clause, thus construed, creates a setting
for other clauses that potentially follow it. We will illustrate this textual
development of the Dagaare clause with an extract from the story of Cain and
Abel (Kayɛ anɩ Abɛl), which has been analysed in Table 5.1.
The analysis in Table 5.1 gives a picture of a text ―evolving in its context‖
(Halliday & Matthiessen 1999: 9). It clearly shows how the selection of
Theme contributes to the development of the clause in Dagaare. Each clause
begins with a Theme, which is then developed in the Rheme. The Theme
locates each clause in the environment of the unfolding text. In clause (1), the
Theme is A Adama nɩ Awa. The Rheme develops the Theme by telling us
about their offspring, Cain and Abel. In the remaining clauses, we observe that
the Theme of each clause relates the clause to the context of the preceding
text. In this text, as is characteristic of narratives, the Theme is typically
picked out from the Rheme of the preceding text. Two different kinds of
Themes can be observed in clauses (2.2) and (4.2). Each of these clauses
begins with a conjunction followed by a nominal group. Both the conjunction,
which in the case of (4.2) is aclitic form of the conjunctive particle kɛ́, and the
nominal group serve to orient the clause in two different ways. The
conjunction serves to connect the clause to the preceding one. In other words,
it clearly posits the preceding clause as the environment for the interpretation
of the following clause. This type of Theme is called a textual Theme while
those Themes in the text that are realised by nominal groups such as A Adama
nɩ Awa (‗Adam and Eve‘), A Kayɛ, (‗Cain‘) AKayɛ poru (‗Cain‘s farms‘) a ci
(‗the guinea corn‘) and Ʋ (‗s/he‘) are topical Themes, what the clause is about.
A topical Theme typically has an experiential function in the clause, that is, a
participant role, a process or a circumstantial element (but see below on
absolute Theme). All the topical Themes in the Cain and Abel text are
participants in the clause (see Chapter 6 for details on experiential structure of
the clause).
207
As can be observed from the interlinear glossing, the Rheme also
embodies another kind of textual prominence, that of information focus. These
two together, Theme and focus of New information, construe the textual
essence of the Dagaare clause. Information focus will be discussed in Section
5.4. In this section, we proceed to discuss the different types of Theme,
beginning with topical Theme in Section 5.3.1 and then textual and
interpersonal Themes in Section 5.3.2. Figure 5.1 presents a fragment of a
system network for THEME in Dagaare, specifically, a network on the selection
of topical Theme.
Figure 5.1: A system network of THEME SELECTION for Topical Theme
5.3.1 Topical Theme

Every language dedicates some of its lexicogrammatical resources for
signalling thematic prominence in the clause. Studies have revealed three
motifs of realisation across languages (see Matthiessen (2004) and references
therein). First, many languages, including Dagaare and, also, English, simply
place elements of thematic prominence at the initial position of the clause (cf.
Halliday & Matthiessen 2014). Second, some languages show thematic
prominence morphologically, typically by singling the Theme element out
with a special particle or even an affix. An example of such a language is
Tagalog, where the particle ang tags the topical Theme irrespective of its
placement in the clause (Martin 2004). Some other languages, such as
Japanese, combine these two strategies and exemplify the third motif across
208
languages. Japanese privileges initial position of the clause as thematic and at
the same time deploys the particle wa or ga to segmentally indicate thematic
prominence (Teruya 2007).
Although Dagaare does not indicate topical Theme morphologically, a
systematic study of discourse data clearly shows the initial position of the
clause as thematically loaded. We will elaborate this textual phenomenon by
identifying two main types of topical Theme in the Dagaare clause: absolute
Theme (Section 5.3.1.1) and non-absolute Theme (Section 5.3.1.2).
5.3.1.1 Absolute Theme

As indicated earlier, topical Themes normally have a function in the
experiential or transitivity structure of the clause as those we saw in the Cain
and Abel text. But it is also a very common characteristic of Dagaare for
speakers to clearly mark off the topical Theme as a separate element from the
rest of the clause. A Theme signalled this way is called absolute Theme (cf.
Matthiessen 1995: 552-554; Downing 2015: 214-142). Thus, we can define an
absolute Theme as a topical Theme that does not have a transitivity function in
the clause (cf. Chapter 6). Absolute Themes have often been discussed under
the topic of ‗left-dislocated constructions‘ in the study of African languages
(Abdel-Hafiz 2015; Güldemann et al . 2015). We illustrate their use in the
following report by a farmer at a workshop. Absolute Themes are marked in
bold and the clauses in which they occur are underlined:

[1] A dɩya, a tome nɛ
DEF last year DEF work.NMLZ DEM.DIST
tɩ na to a, a tome
1PL REL work.PFV JUNC DEF work.NMLZ
dɩya na vɩɛl, a kʋ̀
last year AFFR be good CONT give.PFV
=m.16
1SG.ACC
16
A kʋ̀ m (‗give me‘) is a fragmentadded as an afterthought. A more appropriate construction
will be as follows:
A tome dɩya vɩɛl a kʋ̀ =m.
DEF work.NMLZ last year be:good AFFR give.PFV 1SG.ACC
209
[2.1] Bʋʋ nʋ [2.2] so?
what IDENT.SG own
[3.1] A dãw niwn sob, a vɛ na a
DEF earlier face owner 3PL.NHMallow.PFV AFFR DEF
tɩ nɩbɛ ani nɛ, tɩ tɛr =ɩ

1PL people eight DEM.DIST 1PL possess FOC
nɔ-law
unity
[3.2] a tome cere.
DEF work.NMLZ go.IPFV
[4] ||| Ɛ wa a kɔb mɩŋa lombowri, ɩ
CONJ come DEF farm.NMLZ itself area 1SG
nyɛ n= a zie nɛ na siw

see.PFV FOC DEF place DEM.DIST REL lower.PFV
tɩ bʋla nɩ a be tɩ na
1PL small COP.FOC DEF there 1PL REL
bɛ ƴãw a nasa-kuolu a, ||
NEG.IND.NFUT put.PFV DEF European-fertilizer JUNC
ɛ bɛ ƴãw a Dagara kuolo

CONJ NEG.IND.NFUT put.PFV DEF Dagara fertilizer
mɩ yaga a. |||
also much JUNC
[5] Ɩ tɩɛ a be nʋ [[na

1SG think.PFV DEF there IDENT.SG REL
wa nɩ gbɛ-ŋmɛ a tɩ kʋɔbɛ
come CAUS shortcoming DEF 1PL farmers
zie a]].
place JUNC
[6.1] Ɩ tɩɛ a be bome na

IPL think DEF there things IDENT.PL
[6.2] bɛ maal vla kʋ̀ tɩ

NEG.IND.NFUT do:well.PFV good give.PFV 1PL
ɛ.
NAFFR
210
[7.1] ||| Ɛ wa, a zie nɛ [[tɩ na
and come DEF place DEM.DIST 1PL REL
maalɩ k= a ɩ mɛ
make.PFV so that 3PL.NHM COP.PFV like
a bowli na tɩ ƴaw a]],
DEF small:holes IDENT 1PL put.PFV JUNC
a be de nɩ ɓaar, || [7.2] a be
DEF there take.PFV FOC moisture DEF there
bome maal =a zuo a wɛrʋ a
things do:well.PFV AFFR more than.PFV DEF farms DEF
za [[tɩ na kɔ a]].|||
all 1PL REL farm.PFV JUNC
[8] ||| Waar mɩ̀, a zie nɛ mɩ̃̀ [[tɩ

come also DEF place DEM.DIST also 1PL
mɩ̃̀ na ƴaw a bɩwra a]], ɩ nyɛ

also REL put.PFV DEF ridges JUNC 1SG see.PFV
n= a be mɩ̀;||
FOC DEF there also
[9] a bɩwra nɛ [[tɩ na mɩ́ ƴaw a]],
DEF ridges DEM.DIST 1PL REL HAB put.PFV JUNC
ɩ tɩɛ a bɩwra bome mɩ̀ maal
1SG think.PFV DEF ridges things also do:well. PFV
=a zuo a tɩɩ-sɔw mɩŋa dem
AFFR be:more:than.PFV DEF land self own
[[tɩ na bʋr a]].|||
1PL REL sow.PFV JUNC
[10] Ɩ tɩɛ bom nɛ [[ɩ na

IPL think.PFV thing DEM.DIST 1SG REL
nyɛ yi nɩ a dɩya tome

see.PFV be:from.PFV FOC DEF last year work.NMLZ
pʋɔ a]] nɩ a lɛ.
inside JUNC COP.FOC DEF DEM.DIST
‘Last year, the work that we did, the work was good last year, for
me. Why? First, it let we those eight people, we had unity so that the
211
work progressed well. And come to the farming itself, I have seen
(that) the area [[that let us down a little]] is the place [[that we didn‘t
put the European fertilizer and also didn‘t put much of the Dagara
[=local] fertilizer]]. I think that place is the one [[that brought a
shortcoming on the part of we the farmers]]. I think the things [=crops]
there are the ones (that) didn‘t do well for us. And then, the place
[[that we made such that it seemed we had created small ponds]],
that place retained moisture, the things [=crops] there did better than
the rest of the farms that we made. Then also, the place too [[where
we made the ridges also]], I have seen that place too; those ridges
[[that we make]], I think the things [=crops] in the ridges also did
better than the ones [[that we sowed on the plain land]]. I think that is
the thing [=what] [[I found from last year‘s work]].‘
Each of the underlined clauses begins with an element, the absolute Theme,
set apart as the topic of the clause. This element has the potential of playing a
role in the transitivity structure of the clause, but it is not. Thus, the absolute
Themeis normally resumed later in the clause as a part of the experiential
composition of the clause. This is done in several ways, comprising the use of
repetition, rephrasing, a resumptive pronoun or some other pro-form. Clause
(1) has two absolute Themes: A temporal adverbial group A dɩya (‗last year‘)
and a nominal group a tome nɛ tɩ na to a (‗the work that we did‘) both of
which are rephrased later in the clause as circumstantial Adjunct and
Subject/Carrier respectively. In (5.1), the absolute Theme is resumed by the
pronoun tɩ again as Subject/Carrier; In (7), the absolute Theme is a locative
nominal group and it is resumed by the proform a be (‗there‘) in both (7.1) and
(7.2) as also Subject/Carrier; and in (8), it is resumed as a
Complement/Phenomenon, a be mɩ̀. In (9), it is paraphrased as a
Modifier/Classifier in the nominal group a bɩwra bome which functions as
Subject/Carrier in the clause.
These Themes contribute to the overall development of meaning in the
text. The first thematic element places the text within a particular temporal
context, A dɩya (‗last year‘) and within this temporal space, the farmer singles
out a particular activity, ‗the work that we did‘. Together, they set out the
212
aboutness of the text. Each of the following absolute Themes builds up the
topic of the text by specifying an aspect of it: first about us, the eight farmers
(tɩ nɩbɛ ani nɛ), then about the farming itself (a kɔb mɩŋa lombowri) and then
different aspects of it (see examples 7 to 9). Thus, within the text, there is a
division of labour between absolute Themes and the other types of Theme.
They are more oriented towards the global organisation of the text than the
non-Absolute ones such as a tome (‗the work) in (3.2). This is just one specific
instance of their use. But it is a function that is recurrent. The following
extract exemplifies a similar function of absolute Themes in the biblical
creation story, where they topicalise the different created things that populate
the empty world, one after the other (absolute Themes are in bold; relevant
clauses are underlined):

A pɛr tib daar Naaŋmɩn tɩ ir =ɩ
DEF bottom start time God PST.REM bring:forth.PFV FOC
a salom nɩ a tẽw. A tẽw tɩ
DEF sky and DEF earth DEF earth PST.REM
ɩ =n vuo ɛ bom za bɛ ka
COP.PFV FOC hollow CONJ thing all NEG.IND.NFUT EXIST.PFV
be ɩ. A lɛ na a Naaŋmɩn
there NAFFR DEF DEM IDENT.PL DEF God
tɩ yèl: “A cãa, ʋ be!” Ɛ a
PST.REM say. PFV DEF light 3SG EXIST.PFV CONJ DEF
cãa dɛ caalɩ […]

light ADV shine.PFV. […]
A Naaŋmɩn tɩ yèl a: “Zʋmɛ̃, a
DEF God PST.REM say. PFV AFFR fish 3PL.NHM
be duure a kʋɔmɩ […] Dʋn-bʋʋrɛ

EXIST.PFVswim.IPFV DEF in water [...] animal-kinds
bʋʋrɛ za, a bɛrɛ nɩ a pʋrmɛ, a
DUP all DEF big.PL and DEF little.PL 3PL.NHM
be a tew zu.”
EXIST.PFV DEF earth ADP
213
―In the beginning, God brought forth the sky and the earth. The earth
was hollow and nothing was there. Then, God said: ―The light, it be!‖
And the light just shone … God said: ‗Fish, let them be swimming in
the water … All kinds of animals, the big ones and the little ones, let
them be in the world!‖
It will be a useful exercise to examine the division of labour between absolute

Themes and other topical Themes where they occur together in a text.
Compared to non-Asolute Themes, they tend to be fairly frequent in Dagaare
discourse. In a sample of 375 clauses across different registers, they record
about 15% (i.e. 54 counts) instances while non-absolute Themes record about
72% (i.e. 271 counts), excluding minor clauses. We turn to non-absolute
Themes in the next section.
5.3.1.2 Non-absolute Theme

Non-absolute Theme is a cover term used for any Theme that has a function in
the transitivity structure of the clause. These are the kind of topical Themes
that were identified in the Cain and Abel text. As Figure 5.1 shows, they
divide into two sub-types, marked and unmarked Themes. Unmarked Theme
is what is chosen as Theme when there no good reason to choose another one
(Halliday & Greaves 2008) while marked Theme attracts attention to itself, it
adds a special effect or meaning to the clause.
(1) Unmarked Theme: In many languages such as English and other

Germanic languages, unmarked Theme varies across mood types (cf.
Matthiessen 2004; Teruya et al. 2007). In Dagaare, the typical unmarked
Theme in both the indicative and imperative is the Subject. A good example
for declarative clauses is provided by the Cain and Abel text we analysed at
the beginning of this Chapter, where all the topical Themes are unmarked. An
example is given below for the imperative, where the third person pronoun Nyɩ
is Theme:
Nyɩ dɔwr paalɛ ɛ yɛrɛ!
2PL procreate.IPFV fill.IPFV CONJ spread.IPFV
‗You give birth to many, and multiply!‘
214
The second person singular imperative, however, has a different selection of
unmarked Theme since it does not take the Subject in its unmarked usage. The
unmarked Theme here is the Predicator element, realised by the verbal group
(Predicator is in bold):

Faa tι wɛ!
save.PFV 1PL EXCL
‗Save us, please!‘

(5) St. Maria Play
Ta mɩ́ yèl a lɛ ɩ!
NEG-IMP HAB say. PFV DEF that NAFFR
‗Don’t be saying that!‘

(6) St. Maria play
Na bɔ bʋndɩrɩ kʋ̀ mãa!
EXH find.PFV food give.PFV 1SG.EMP
‗Please, find me food!‘
Studies have also shown that some languages, including English, give
thematic prominence to Q-elements in elemental interrogative clauses (cf.
Matthiessen 2004; Matthiessen, Teruya & Wu 2008; Teruya et al. 2007). In
Dagaare, Q-elements are often placed in situ, in the original position of the
clause which reflects their function in the clause structure (see Chapter 4,
Section 4.4.1.2.3; see also Bodomo 1997). This situation, however, seems to
be changing and there are many instances where the Q-element is placed at the
initial position of the clause irrespective of its function in the clause structure,
and without marked meaning. An example is given below:
(7) St. Maria play

Nyɩnɛ fʋ bãw a pobili bɛna?
where 2SG know.PFV DEF gentlemen DEM
‗Where do you know these gentlemen.‘
(2) Marked Theme: The difference between Marked Themes and absolute
Themes is that the former play a role in the transitivity structure of the clause.
215
However, the default position of elements that are realised as marked Themes
is not the initial position of the clause; they are only placed initially to give
them thematic prominence. In this sense, marked Themes carry more
prominence than unmarked ones. The most frequent elements that occur as
marked Theme in the Dagaare clause are circumstantial Adjuncts of time (see
also Halliday & Matthiessen (2014) on English). The following dialogue
highlights the use of an adverbial group as marked Theme:
(8) St. Maria play

Mother: Nyιnɛ na a fʋ sãa dambol
where IDENT.PL DEF 2SG father fool
nɛ nιɛ cen?
DEM ADV go.PFV
Son: Dambol ya?
fool INT
Mother: Zã̀̃ a za, ʋ bɛ kabe

yesterday all 3SG NEG.IND.NFUT EXIST.PFV
a ka ι.
DEF here NAFFR
Mother: ‗Where is that that foolish father of yours has gone to?‘
Son: ‗Foolish?‘
Mother: ‗Since yesterday, he has not been here.‘
The thematised adverbial group, Zà a za (‗since yesterday‘), functions as

circumstance of Duration in the transitivity structure of the clause and its
typical placement is clause final position, in this instance before the Negotiator
ι (non-affirmative). Its choice as Theme is motivated by the speaker‘s concern
with the duration of the absence of the husband, a situation which frustrates
her and motivates her question to the son in the first place. The status of
marked Themes is signalled by setting them off from the rest of the clause
prosodically with a pause or a comma in a written text. In this sense, they
behave like absolute Themes.
216
5.3.2 Textual and Interpersonal Themes
To recapitulate, we have already seen in our discussion on topical Theme that

the system of THEME interacts closely with both the ideational and
interpersonal structure of the clause. For instance, we have shown that the
identification of topical Theme and, subsequently, the distinction between
Absolute and non-absolute Theme is determined by the status of the element
in the transitivity or experiential structure of the clause. Also, we observe that
what counts as marked Theme and unmarked Theme varies according to
delicate mood types (e.g. second person singular imperative versus other
imperative types). This reflects the fact that the textual metafunction is an
enabling one; it organises the resources of the other two metafunctions,
ideational and interpersonal, into a coherent whole, a consumable semiotic
entity (see Chapter 2 on metafunctions). This enabling function is further
reflected in the fact that the principal types of Theme correspond to the three
metafunctional meanings of language. We discussed one of these, topical
Theme, in the preceding section. This section is concerned with the other two,
namely textual Theme and interpersonal Theme. While topical Theme
construes the aboutness of the clause, textual Theme connects the clause to the
preceding text, and interpersonal Theme orients the clause in relation to tenor,
the me-&-you dimension of discourse.
Textual Theme is realised by cohesive and structural conjunctions as
well as continuatives. Cohesive and structural conjunctions can be realised by
the same form, but while cohesive conjunctions connect the clause externally
to a preceding segment of the text, structural conjunctions link two clauses in a
clause complex. Let‘s illustrate this with an extract from the workshop report
discussed in Section 5.3.1.1 above:

||| Ɛ wa a kɔb mɩŋa lombowri, ɩ
CONJ come DEF farm.NMLZ itself area 1SG
nyɛ n= a zie nɛ na siw

see.PFV FOC DEF place DEM.DIST REL lower.PFV
tɩ bʋla nɩ a be tɩ na bɛ
1PL small COP.FOC DEF there 1PL REL NEG.IND.NFUT
217
ƴãw a nasa kuolu a, || ɛ bɛ
put.PFV DEF European-fertilizer JUNC CONJ NEG.IND.NFUT
ƴãw a Dagara kuolo mɩ yaga a. |||

put.PFV DEF Dagara fertilizer also much JUNC
Here, the conjunction Ɛ (‗and‘) is used at the beginning of the extract as a

cohesive element to connect the following discourse segment with the
preceding one and, also, at internal clause boundary to link the two clauses in
the clause complex. On the other hand, the two uses of ɛ have the same
function, that is, to anchor the clauses they introduce to their textual context.
Generally, the report, as we have seen, is rich in textual Themes and in its
textual development as a whole, carefully guiding the listeners through
different aspects of the report. We can sum up the overall textual development
of the text as follows, as a way of mapping the semiotic space in which the
textual Themes are located:

Macro-Theme [topical orientation of the text]
[1] A dɩya, a tome nɛ
DEF last year DEF work.NMLZ DEM.DIST
tɩ [[na tõ a]], a tome

1PL REL work.PFV JUNC DEF work.NMLZ
dɩya na vɩɛl, a kʋ̀
last year AFFR be good CONT give.PFV
=m.
1SG.ACC
‗Last year, the work that we did, the work was good for me last
year.‘
Hyper-Theme [textual signal of following points]
[2.1] Bʋʋ nʋ [2.2] so?
what IDENT.SG own
‗Why‘
Micro-textual Themes [functioning within the clause]
[3.1] A dãw niwn sob …
DEF earlier face owner
218
‗The first one …‘
[4] |||Ɛ wa a kɔb mɩŋa lombowri||
CONJ come DEF farm.NMLZ self area
ɛ … |||
CONJ
‗And come to the farming itself … and …‘

[7] Ɛ wa
and come
[8] Waar mɩ̀
come.IPFV also
‗Also …‘
Macro-New[closure]
[10] Ɩ tɩɛ bom nɛ [[ɩ na
1SG think.PFV thing DEM.DIST 1SG REL
nyɛ yi nɩ a dɩya tome

see.PFV be:from.PFV FOC DEF last year work.NMLZ
pʋɔ a]] nɩ a lɛ.
inside JUNC COP.FOC DEF DEM.DIST
‗I think that is what I have observed from last year‘s work.‘
When we analyse a text this way, it is striking to find out how a text is like the
clause in its development or, from the other perspective, how the clause is like
a text (cf. Halliday (1981, 1982, 2002: 219-260). In recognition of this fact,
Martin (1993) used the terms ‗hyper-Theme‘ to refer to the first level of
thematic orientation above the clause and ‗Macro-Theme‘ and ‗Macro-New‘
for further layers of thematic and also focal orientation in the hierarchical
structure of the text (see also Martin & Rose (2007)). What we have
demonstrated here is that the textual Themes of the clause interconnect with
the semantic progression of the text as a whole. It is in this sense that they
provide a point of departure, or a thematic context, for the clause.
Interpersonal Theme, on the other hand, orients the clause towards its
social context. It is a recognition of the interactants interacting in the
discourse, their presence, their roles and identities, and their attitudes and
sensibilities. In other words, it sets up a kind of social environment that is
219
local to the clause and within which the clause should be interpreted.
Interpersonal Themes have a wide range of realisations, including vocative
nominal groups, clause initial negotiation particles, and interjections or even
whole clauses, specifically projecting clauses (see Chapter 6, Section 6.7.1).
Let‘s consider some instances in the following extracts (interpersonal Themes
are in bold):
(11) St. Maria play
Mother: Ɩ̃ pɔw-yaa, na zɩ pɩɛnɩ!
1SG daughter EXH sit.PFV rest.PFV
Daughter: Mama, fʋ nyɛ nɩ yél kàw?
mama 2SG see.PFV FOC matter some
Mother: ‗My daughter, please, sit for a rest!‘
Daughter: ‗Mama, you see something?‘
Mɛ́ tɩ be =n a ka a mʋtɔw
HST 1PL EXIST.PFV FOC DEF here DEF afternoon
za wɛ?
all INT
‗I believe we have been here since this morning, right?‘

Tɔɔ, ok, a cɛlcɛlbɛ, a biɓaara ŋa,
well ok DEF listeners DEF morning DEM
tɩmɛ nɩ a Animal Research anɩ

1PL.EMP and DEF and
a agric dem [[=a law
DEF agriculture owners BIND be:together
taar a tɛr bɛ program kàw
each other CONT possess.PFV 3PL.HM some
a]], a bɛ yèl kɛ ɛ tɩ pãa a
JUNC CONT 3PL say. PFV PROJ PROJ 1PL ADV CONT
de a yéle ƴaw a sɛsɛb pʋɔ a nɩ

take.PFV DEF matter put DEF air inside CONT 2PL
cɛlɩ.
listen.PFV
220
‗Well, ok, listeners, this morning, we and the Animal Research and
agric people [[gathering here for a program they have]], and they say
that we should now put it on air for you to listen (to it).‘
In (11), interpersonal Theme is realised by vocatives and, in addition, the

negotiation particle na, marking exhortation in imperative clauses (see Chapter
4, Section 4.4.2.2). In each clause in the exchange, the vocative creates a role
identity in which the whole clause needs to be interpreted. The use of
negotiation marker in the mother‘s utterance further adds an attitudinal
orientation to the clause. Arguably, the positive social context created by the
interpersonal Theme is what motivates the daughter or at least sets the stage
for her to take her turn by wanting to share her experiences with the mother. In
(12), the interpersonal Theme is the negotiation particle Mɛ́ (modally
signalling expectation), which invites the listener to confirm the proposition. It
is prosodically resonated by another negotiation particle wɛ (biased polar
interrogator) at clause final position, thereby orienting the clause
interpersonally at both juncture positions. Example (13) is the opening
statement of a radio interview with an agricultural extension officer and it is
addressed to the listeners. In itself, it is an interpersonal orientation to the
interview as a whole. Its interpersonal orientation clearly resonates with the
multiple use of interpersonal Themes in the clause, a phenomenon which is
typical to this registerial context. The interpersonal Themes here serve to build
up a relationship with the virtual audience and invite their attention to the
content of the clause and by extension the discussion as a whole. Apart from
these lower rank realisations of interpersonal Theme, whole clauses do also
serve to interpersonally orient the clause. Such clauses are normally mental,
verbal and relational clauses that project the clauses in which they serve as
Theme (cf. Chapter 6 on transitivity). One example is given below from the
workshop report in (1) above:
(14) |||Ɩ ̃ tɩɛ a be bome na || bɛ

IPL think DEF there things IDENT.PL NEG.IND.NFUT
maal vla kʋ̀ tɩ ɛ.|||

do:well. PFV good give.PFV 1PL NAFFR
221
‗I think it is the things there that didn‘t grow well for us.‘
Here, the mental clause Ɩ tɩɛ (‗I think‘) is a modal assessment of the projecting
clauses in the complex, where it invites listeners to interpret the proposition as
the speakers own subjective reflection. In this sense, the mental clause has a
similar function as the negotiation particle Mɛ́ in example (12).
In summary, the general principle of the flow of information in the
clause is that any element is thematically prominent than the one it precedes.
Those elements that are singled out as Theme of the clause are those that are
placed at the initial position in the structural organisation of the clause. As we
have seen, in principle, there is no limit to the number of Themes that can
occur in a single clause. This will be determined by the practical exigencies
motivating their choice. The Theme element of the clause can be characterised
trinocularly as follows. Semantically, it locates the clause in context in the
textual flow of information and serves as its point of departure. Specifically,
textual Themes connect the clause to the preceding text; interpersonal Themes
orient the clause in relation to tenor; and topical Themes construe the
aboutness of the clause. From a roundabout perspective, that is, within the
clause itself, the Theme is that element (or those elements) which is (are)
given initial prominence in the clause, typically extending from the beginning
of the clause up to and including the first element with a function in the
transitivity structure of the clause (see Chapter 6). The exception is that when
the topical Theme is Absolute, the Subject of the clause is treated as part of the
Rheme, that part of the clause which develops the Theme. From below the
clause on the grammatical rank scale, textual Themes are realised by
conjunctions, both structural and cohesive ones; interpersonal Themes are
realised by interjections, negotiation markers, (modal) adverbial groups, and
also whole clauses. Finally, topical Themes are often realised by nominal
groups and adverbial groups, but also bound or ranked-shifted clauses.
5.4 Focus of information

In Dagaare, as in many other languages, the system of THEME combines with
another system, that of INFORMATION to give the clause its textual essence, as
mentioned above. In this section, we proceed to examine the contribution of
222
focus to the textual organisation of the Dagaare clause. Strictly speaking, the
unit of analysis here is not the clause but, rather, the information unit (cf.
Halliday & Greaves 2008; also see Section 5.6 for details). Like any other
grammatical unit, the information unit is difficult to define in explicit terms
without evoking circularity. But, as its name implies, it is a unit of information
in the flow of discourse, a quantum of information, as it were (Halliday &
Greaves 2008; Halliday & Matthiessen 2014).
When people engage in text-&-interaction, they often presuppose some
part of their messages to be known to their listeners and other parts to be
newsworthy. The information unit is thus typically made up of a structural
configuration of Given and New information. Within the information unit, one
element is singled out and given prominence as the Focus of New
information. It is this prominent element that locates the locus of news value
in the utterance. In Dagaare, Focus of New is signalled by a number of
lexicogrammatical resources, comprising the focus particle nι (or its enclitic
forms, n and ι), thematic equatives (or ‗cleft-constructions‘), emphatic
pronouns, and exclusive particles. These resources will be explained and
illustrated in detail in later sections. But let‘s first introduce and illustrate
different types of focus in the flow of discourse in Dagaare. These are
summarised in Figure 5.2 as a system network.
Figure 5.2: System network of INFORMATION in Dagaare

223
The default (or unmarked) type of focus, the one chosen in the absence of a
good reason to choose another, is indicated on a lexical item towards the end
of the clause by preceding this item with the focus particle nι.17 This focal
element is normally the Complement or Adjunct element in the modal
structure of the clause. Instead of a lexical item, the focal element could also
be a ranked shifted clause functioning as Complement. This unmarked
information focus will be called end focus since it occurs at or towards the
end of the clause (cf. Quirk & Greenbaum 1973). 18 The second type of focus is
contrastive focus. With this focus type, the speaker emphasises one element
in the information unit as being in contrast with some other information in the
context of the discourse. It is a marked choice because it adds special meaning
to the information by drawing particular attention to itself.
However, a speaker may treat the whole utterance as new information
and therefore does not need to mark focus on any segment of the information
unit. This focus type will be called broad focus, as opposed to end and
contrastive focus, the two of which we will classify as sub-types of narrow
focus (see Figure 5.2).19 The extract below illustrates the contribution of these
various types of focus to the flow of discourse in Dagaare:
(15) St. Maria play

Father: // Pãa mι bɛr // ɛ cɛlɛ a fʋ
now also stop.PFV CONJ listen.IPFV DEF 2SG
yéle . //
matter.
Son: Ʋʋ?
yes
Father: Fʋ mɩ̀ ta na.
2SG also reach.PFV AFFR
Mother: //Fʋ mɩ̀ ta =n pɔw kulu. //

2SG also reach.PFV FOC wife marry.NMLZ
17
In some other dialects of Dagaare such as Ngmere and Waali, the focus particle is la and
has the same form as the affirmativeparticle in these dialects (cf. Bodomo 1997, 2000).
18
In this case, Dagaare is like English in the position of unmarked focus.
19
In the typological literature, the kind of focus we are calling broad focus has also been
labelled as ―all -new-sentence‖ (cf. Güldemann et al . 2015), ―sentence focus‖ (Lambrecht
1994, 2000), ―thetic statement‖ (Sasse 1987; Güldemann et al ., 2015), and ―neutral focus‖
(Sasse 1981).
224
//A fʋ yɛ-bile kʋl sιr //
DEF 2SG sibling-small marry.PFV husband
ɛ fʋ caa zι ι? //
CONJ 2SG still sit INT
Son: Mãa // lιɛbɛ nι faara o! // Ɩ lιɛbɛ

1SG.EMP turn.IPFV FOC priest PRT 1SG turn.IPFV
nι faara. //
FOC priest.
Father: ‗Now, stop (talking) too and listen to you matter.‘
Son:‗Yes?‘
Father: ‗You too have reached.‘
Mother: ‗You too have reached MARRIAGE. Your younger sibling is
married and you still sit (home)?‘
Son:‗I‘m becoming a PRIESTo! I‘m becoming a PRIEST.‘
This exchange is co-constructed by father, mother and son. It follows the end
of a marriage negotiation for the daughter of the family and the discussion
now shifts to the marriage of the son. In the first turn, the father calls the son
to attention by the use of an imperative clause and this is followed by the son‘s
signal for further information. The father comments on the son‘s supposed
readiness for marriage while assuming the topic of marriage as shared
knowledge from the preceding marriage negotiation. The mother picks up the
turn, now treating the son‘s readiness as given information and introducing
marriage as the focus of information, marked by the enclitic focus particle n,
and follows this with a justification why her son should get married. The son‘s
response is in three information units. The first clause Mãa lιɛbɛ nι faara o!
(„I‘m becoming a PRIEST o!) consists of two information units. In the first,
focus is marked by the emphatic pronoun Mãa (‗I‘), by which the speaker
engenders a contrast between himself and his sister. This is an instance of a
marked or contrastive focus. In the second information unit, focus is placed on
faara (‗priest‘). The speaker repeats the utterance but, this time, treating the
initial part of the clause as given information, only marking end focus on the
last lexical item in the clause, faara (‗priest‘).
225
Thus, the flow of discourse in the dialogue is structured by different focus
types. In all, the exchange consists of ten information units. The first turn by
the father consists of two information units, followed by one each in the son‘s
question and the father‘s response. The mothers turn consists of three
information units and the final turn by the son, as mentioned earlier, comprises
three units. The information units in the initial turns by the father (i.e. Pãa mι
bɛr; ɛ cɛlɛ a fʋ yéle ) and the son (i.e. Ʋʋ?) are all in broad focus. Although
this is not the case in all dialogic contexts, but where a text opens with
information units in broad focus, they normally function to set the stage or
create a discourse setting for subsequent dialogic interactions.
As in the first turn in the dialogue above, imperative clauses, in the
unmarked case, take broad focus and, thus, never admit the focus particle nι.
This can be understood within the context that imperative clauses are
initializing devices in dialogic interactions. 20 But one interesting instance of
broad focus in the text is the the clause: A fʋ yɛ-bile kʋl sιr … (‗Your younger
sibling is married‘). Here, the speaker could have introduced the enclitic focus
particle (i.e. =ι) before sιr (‗husband‘) for end focus, but she chooses not to do
so in order to bring the whole information unit into focus as newsworthy. This
is not to suggest that these choices are made consciously by speakers. Rather,
they are unconscious choices that have been internalised by speakers as part of
their knowledge of enacting exchanges in the language.
The alternation of New information and Given information in the
information unit sets up a movement of non-prominence and prominence in
the clause in terms of newsworthiness (cf. Halliday & Greaves 2008). This
configuration of Given and New information will be referred to as focus
structure. It should be noted, however, that, there is only one obligatory
element in the information unit, and this is the New element, while the Given
element is optional. In other words, the minimal realisation of the information
unit is the New element, as we find in broad focus, and the addition of the
Given element is an expansion of it. An illustration of the focus structure of
some of the information units in the dialogue above is given in Figure 5.3.
20
The absence of the unmarked focus is common in Niger-Congo and, perhaps, other African
languages (cf. Heine & Reh 1983).
226
(a) broad focus
(b) end focus
(c) contrastive focus (plus end focus)
Figure 5.3: Illustration of information structure and focus types in

Dagaare
Two comments will be made about these illustrations. As mentioned

earlier, within the New element of the focus structure, the element that carries
focal prominence is technically called the Focus of New information (or just
Focus) (Halliday & Greaves 2008). The second observation is that focus
structure and, for that matter, the information unit, is not a kind of constituent
structure since the boundary between Given and New information is not
always clear cut (also see Halliday & Greaves (2008) on English). This tends
to be a characteristic of the focus systems of language in general. As Halliday
and Greaves (2008) note, they are organised by the principle of degree of
prominence and not by constituent structure. We are always sure of where the
227
locus is because this is normally signalled, but we may not be sure of where it
begins (cf. Halliday & Greaves 2008). The arrows in the box diagrams signal
this fluidity in the boundary between Given and New.
In the case of end focus, the most newsworthy information tends to be
at the end of the information unit and extends towards the beginning of the
clause. In other words, in the unmarked case, every element in the clause has a
higher news value than the element that precedes it. Unmarked focus
prominence is therefore a reversion of thematic prominence. In the case of
marked focus, however, the most newsworthy information is typically placed
at the left most end or the beginning of the information unit and may extend
towards the end. This means that, when the Focus element is marked or
contrastive, the speaker often alerts the addressee immediately the information
unit takes off (cf. Section 5.4.2). The various types of focus and the resources
that realise them will be discussed in detail in the following sections.
5.4.1 End Focus and the Focus Particle

End focus is the default focus type for positive indicative clauses in Dagaare.
As mentioned above, it is marked towards the end of the clause, that is, after
the verbal group realising the Predicator. The focal element is indicated by
preceding it with the focus particle nɩ or its weak variants n or ɩ (cf. Figure 5.3
above). The use of the weak forms of the focus particle is determined by the
phonetic environment in which they occur. That is, n is used when the focus
particle is preceded by an open syllable (16) and ɩ is used after a syllable that
ends with an alveolar consonant such as /r/, /l/, /n/ or a stop such as /b/ and /p/
(17):

A Abɛl wa ɩ =n pi-cɩɩnɛ.
‗Abel became A SHEPHERD.‘
Ʋ tɩ ir =ɩ a ci.
He PST remove.PFV FOC DEF guinea corn
‗He took some of THE GUINEA CORN.‘
228
As is the case in (17) above, the vowel quality of the enclitic particle ɩ, in its
phonetic realisation, may change from the [–ATR] vowel /ɪ/ to the [+ATR]
vowel [i] for vowel harmony. This difference is, however, not indicated in the
orthography.
In addition, when the word preceding the enclitic particle ɩ ends with
the front vowels /ɪ/ or /i/ as in –lɩ and –li, or –rɩand –ri, this vowel in the
preceding word is normally elided. An example is maalɩ (‗do, make‘) in (18)
and sowri (‗ask‘) in (19) below:21

A Kayɛ poru tɩ maal =ɩ vla a
DEF Cain farms PST do.PFV FOC good DEF
ci yawmɛ.
guinea:corn be:plenty
‗Cain‘s farms did WELL so that the guinea corn was many.‘
(19) Bible.is Matie (16: 13b)
Ʋ sowr =ɩ a ʋ po-tuurbɛ.
3SG ask.PFV FOC DEF 3SG followers
‗He asked his followers.‘
As mentioned earlier, the scope of the New information normally extends

from the Focus element to the end of the information unit. Thus, in example
(18) above, the scope of New information starts with vla (‗good‘) to the end of
the construction.
It was also mentioned in Chapter 4 that the Complement element in the
Dagaare clause structure has a textual relevance and not an interpersonal one
(also see Section 5.4). What this means is that the Complement is normally
realised in the clause structure as long as it has news value. Participants other
than Subject) are introduced in the discourse as new information and, when
they have been established as shared knowledge between speaker and listener,
they tend to be discarded in subsequent clauses in the exchange. It also follows
21
Here, we know that it is the vowel in words such as maalι (do, make) and sowri (‗ask‘) and
not the enclitic focus particleι that is elided because, in other environments where such words
precede the enclitic form of the affirmative marker na, the final vowel is also elided; as with
maalι in (b) below:
(a)A cι maalι na. (‗The guinea corn did well‘).
(b)A cι maal =a. (‗The guinea corn did well‘).
229
that when the Complement can be inferred from the context of situation, it
may not be stated at all. 22 An illustration of this phenomenon is given in the
following extract:

Baba: Ɛcɛ Zɩɛm ɩ, a dãa [[ɩ na da
but Ziem VOC DEF beer 1SG REL buy. PFV
kʋ̀ fʋ a]], a wa?
give.PFV 2SG JUNC 3PL.NHM be:where
Ɛ bɛ dɛ zɩ.
And 3PL.HM ADV sit.PFV
[…] […] […]
Naab: Nɔɔ, bɛ waar =ɩ na.
no 3PL.HM come.IPFV CAUS AFFR
Baba: A bɛ waar =ɩ ɛ bɛ
CONT 3PL.HM come.IPFV CAUS CONJ NEG.IND.NFUT
wa ɩ?
come.PFV NAFFR
Naab: […] Nɩ cãa bɛ ɩ a

[…] 2PL still NEG.IND.NFUT make.PFV DEF
dãa sɛr ɛ? [To ladies inside the compound yard]

beer still NAFFR
Baba: ‗But Ziem, the pito/beer I bought for you where is it? And they
(visitors) are just sitting.‘
Naab: ‗No, they are bringing (it).‘
Baba: ‗They are bringing (it) and they have not come?‘
Naab: … You have not prepared the beer/pito yet?‘
In the clauses Nɔɔ, bɛ waar =ɩ na (‗They are bringing (it)) and A bɛ waar =ɩ
(‗They are bringing (it)‘) the implied Complement is a dãa (‗beer‘, ‗pito‘), and
it is not realised because it can be inferred from the initial exchange between
Baba and Ziem. The Complement is, however, realised in Nɩ cãa bɛ ɩ a dãa sɛr
22
As will be indicated below, when the Complement is realised by a [+human] participant, it
is normally provided.
230
ɛ? (‗You are not done with the beer yet?‘) since the speaker now addresses a
different group of listeners away from the immediate context.
This phenomenon of not realising the Complement also resolves a
tension between the textual and interpersonal modes of meaning in the
Dagaare clause. Since both mood contrast and focus are often marked towards
the end of the clause, this situation sets up a metafunctional tension between
the interpersonal and textual modes of meaning. The result is that end focus,
signalled by the particle nɩ, cannot co-occur with the particle na, which
realises the Negotiator elementin the affirmative clause (see Chapter 4). 23
When there is the need to mark end focus, the textual mode of meaning always
takes precedence over the interpersonal one. But after focus has played its role
in the flow of discourse, it gives way to the Negotiator element.
It should be noted, however, that it is not always the case that the
Complement is unrealised in the flow of discourse after it has been established
as shared information. Other factors such as the humanness of the nominal
group and the process type realised by the clause play a role (see Chapter 6,
Section 6.8.1 for a detail discussion). Human participants, which have higher
experiential value, are normally kept track of in the discourse by the use of
pronouns after they have been introduced. This exception is, however, not a
textual one but an ideational one. It is another kind of metafunctional tension
motivated by the ideational meaning of the clause.
The grammar‘s way of dealing with this tension is that, in the
unmarked case, personal pronouns (specifically, non-emphatic pronouns) do
not receive focus. They are taken for granted as shared knowledge. In (21)
pronoun ʋ (third singular) in Complement position does not receive end focus
although it occurs in potentially focus prominent position:
23
In echoing Halliday (1984) and Konig and Siemund (2007), Matthiessen et al. (2008) note
that:―Of the three basic initiating speech functions that are grammaticalized in languages …
statements are the unmarked type in the overall system, and since they are typically realized
by ‗declarative‘ clauses, the ‗declarative‘ mood tends to be the unmarked type in the mood
system of languages‖ (p. 168).
231
Ʋ tõ ʋ na.
3SG send.PFV 3SG AFFR
‗He sent him.‘
All other things being equal, in (21), end focus would have been placed on the
Complement (realised here as ʋ, second singular). In that case, it would have
been realised by a noun or, in the marked case, an emphatic pronoun (i.e. ʋl;
see Section 5.4.2.2 below on emphatic pronouns). The absence of focus allows
the affirmative particle na to negotiate the clause.
Thus, the grammar of the Dagaare clause has evolved strategies for
managing the competing modes of meanings at the end of the clause. The
motivation to enact the clause as a unit of exchange through negotiation
(interpersonal), the motivation to signal newsworthiness (textual) and the
motivation to retain certain phenomena in the clause, in the case of human
participants, as valued fellows in the world of experience.
Another issue in relation to end focus and the experiential
metafunction is that, in benefactive clauses, there are two complementary
ways of realising end focus (see Chapter 6, Section 6.7.2 on beneficiary
clauses). In constructions such as (22), focus is placed on the participant which
is higher on the empathy hierarchy, first human and then animate, unless this
participant is realised by a non-emphatic pronoun, and thus out of focus (23)
(see Haspelmath (2015) for a typological account on the order of elements in
benefactive clauses).
(22) Ɩ kʋ̀ =n a pɔw-sιra libir?

1SG give.PFV FOC DEF woman-young money
‗I gave THE YOUNG WOMAN money.‘
(23) Ɩ kʋ̀ ʋ nι libir?
1SG give.PFV 3SG FOC money
‗I gave her/him MONEY.‘
The alternative realisation is by a verbal group complex (‗serial verb

construction‘) as shown in example (24). Here, the focal element is always the
232
Goal (‗affected‘) participant (or Verbiage in the case of verbal clauses) and it
normally follows the first verb in the complex:
(24) Ɩ de nι libir kʋ̀ a

1SG take.PFV FOC money give.PFV DEF
pɔw-sιra
woman-young
‗I gave MONEY to the young woman.‘
Finally, end focus is not marked in negative clauses (see Figure 5.3). This
follows that the focus particle nɩ does not occur in negative clauses. 24 This
phenomenon has been reported for the focus systems of many African
languages (see e.g. Heine & Reh 1983). As will be discussed in Section 5.4.2
below, however, particular segments of the negative clause can be singled out
and marked for contrastive focus.
5.4.2 Contrastive Focus

As mentioned earlier, end focus is the focus type that is chosen by Dagaare
speakers when there is no good reason to choose another type. But very often,
there is a good reason to mark focus on parts of a message other than a lexical
item (or rank-shifted clause) that is placed towards or at the end of the clause.
When an element other than the default one is chosen for focus, this is an
instance of contrastive focus, a marked focus choice.
Contrastive focus is defined here to include such meanings as ‗this and
not that‘, ‗this in particular‘, ‗this and this alone‘ and ‗this, excluding other
(contextual) alternatives‘ (see Van der Wal (2011) and Güldemann et al.
(2015) on ‗exclusion‘).25 Contrastive focus, therefore, does not always contrast
one element with another overt assertion, although this overt contrast is its
prototype meaning (for discussion on prototype, see Rosch (1983)). Whether
24
But in one kind of verbless (identifying) clauses, the focus particle nι can occur with the
negative particle. The particle in this context is, however, intermediate between a copula
construction and the focus particle, which evolved from this copula construction (cf.
Mwinlaaru & Yap 2017; see also Chapter 6, Section 6.5.2.1).
25
For convenience, ‗this‘and ‗that‘ are only used in the glosses here as a general reference for
focal items in contrastive focus. They encapsulate ‗me‘, ‗you‘, ‗s/he‘, they,‘ etc.
233
the item in contrast is implicit or explicit, however, this kind of focus always
signals some degree of contrast.
Dagaare has evolved a number of strategies for indicating contrastive
focus, namely the use of thematic equatives or ‗cleft-constructions‘, emphatic
pronouns and exclusive particles. Each of these resources will be discussed in
turn below.
5.4.2.1 Thematic Equatives (or ‘Cleft-constructions’)
The use of thematic equatives as a focus strategy in African languages has

been discussed widely in the literature (e.g. Schachter 1973; Heine & Reh
1983; Hartmann & Veenstra 2013, Güldemann et al. 2015). For instance,
Güldemann et al. (2015: 161), in their review of studies on information
structure in African languages, note that ―the most intensively studied strategy
involving marked, mostly contrastive, focus involves different types of cleft(-
like) constructions.‖
Specifically, in Dagaare, marked focus is often achieved through a bi-
clausal construction (i.e. a clause complex), consisting of a focus cleft
construction, which is the main clause, and a hypotactic clause, which contains
out-of-focus or Given information (see Heine & Reh 1983 on other African
languages). 26 Any clausal element which has a function in the transitivity
structure of the Dagaare clause can be focused in a thematic equative
construction. This phenomenon is illustrated in examples (25) to (29) below:
(25) Casual conversation [focus on Actor/Medium]

Zan nʋ waar.
Zan IDENT.SG come.IPFV
‗ZAN IS THE ONEcoming.‘
(26) St. Maria play [focus on Existent/Actor/Medium]
Fʋ taabɛ lɛ bɛ be wɩɛ pʋɔ kʋɔr […].
2SG fellows COP 3PL.HM be.PFV farm inside farm.IPFV
‗YOUR FELLOWS ARE THE ONES (who) are in the farm weeding.‘ (=YOUR
FELLOWS are in the farm weeding.)
26
Like all grammatical labels, ‗Given‘ does not always mean already known, although such
meaning is the prototype (cf. Halliday & Greaves 2008).
234
(27) workshop report 5 [focus on scope/range]
a bɛw-kãa na ɩ nyɛ.
DEF beans-oil IDENT.PL 1SG get.PFV
‗THE OILY BEANS ARE THE ONES I got.‘
(28) Workshop Interview [focus on circumstance, Q-element]
A nyɩnɛ na fʋ tone?
DEF where IDENT.PL 2SG work.IPFV
WHERE IS IT (that) you are working?
(29) The story of Jesus [focus on Process]
A cenu na ʋ cere bɩɩ?
DEF go.NMLZ IDENT.PL 3SG go.IPFV INT
IS IT GOING (that) s/he going? (= Is he LEAVING?)
The thematic equative construction is typically a verbless (identifying) clause,

which is composed of a Subject and Predicator, realised by an identifying
pronoun nʋ (singular) or na (plural). (Identifying clauses are discussed in
Chapter 6, Section 6.5.2.1; also see Chapter 3, Section 3.4.3).27 As (26) shows,
however, a copula verb is used in clauses in which the Subject is third person
plural and human.
As example (30) shows, when the verb is focused, a nominalised copy
of it is thematised in the cleft construction and the original verbal group
retained in situ.28 In verbal group complexes (‗serial verb constructions‘), only
one verb, often the initial verb in the complex, is focused. The constructed
example below illustrates this:
(30) Lɔbʋ na ʋ lɔb bɛr.

throw.NMLZ IDENT.PL 3SG throw.PFV leave.PFV
‗IT IS THROWING(that) he threw (it) away.‘
27
From a grammaticalisation perspective, the identifying pronouns nʋ (singular) and na
(plural, non-human) derive from a fusion of the identifying copula verb nɛ and the third
person pronouns ʋ (singular) and a (plural, non-human) respectively. For details on this
diachronic process, see Mwinlaaru & Yap (2017). Reduced forms of nʋ and na are ʋ and a
respectively.
28
With reference to other African languages, this phenomenon has been referred to as ‗verb
doubling‘ and is contrasted with ―term focus‖ (as against verb focus); i.e. focus on participant
elements of the clause (cf. Güldemann et al. 2015). Another term used for it is ‗predicate
focus‘ (cf. Heine & Reh 1983).
235
As mentioned in Section 5.4.1, negative clauses can also be marked for
contrastive focus. An example is given in the following bi-clausal construction
adapted from an interview transcript:
(31) A kamaan kɔb yéle bɛ na tɩ

DEF corn farm.NMLZ matter NEG.IND.NFUT IDENT.PL 1PL
tɩ wul ɛ.
PST.REM teach.PFV NAFFR
‗THE CULTIVATION OF CORN IS NOT THE ONE we taught.‘
In this situation, the scope of the negation is the identifying pronoun, which
together with the negative particle forms the predication of the clause (see
Chapter 4, Section 4.3.2.2).
An illustration of the use of thematic equatives for focus in the flow of
discourse is given in the dialogue in (32) below:
(32) Workshop report 5 (Q & A)

Agric Officer: Soo a tɩtal pʋɔ na fʋ dɛ
so DEF plain inside IDENT.PL 2SG ADV
bɛ nyɛ bom za ɩ?
NEG.IND.NFUT get.PFV thing all NAFFR
Farmer: //Al dɩɛ, // ɩ bɛ nyɛ

3PL.NHM.EMP EX 1SG NEG.IND.NFUT get.PFV
bom e.// Soo a bɛw-kãa na
thing NAFFR so DEF beans-oil IDENT.PL
ɩ nyɛ, // ɛcɛ a mɩ̀ bɛ

1SG get.PFV but 3PL.NHM also NEG.IND.NFUT
waar ɛ.//
come.IPFV NAFFR
Agric officer: So IT IS IN THE PLAIN (that) you didn‘t get anything?

farmer: as for that I didn‘t get anything. So THE OILY BEANS ARE THE
ONES I got but they are also not yielding. (workshop report 5)
this text is a question and answer interaction following a farmer‘s evaluation

of a number of faming methods he had used. The agricultural officer singles
236
out one of the farms, a tɩtal pʋɔ (‗in the plain‘), as the focus of his question,
using a thematic equative. The farmer picks this up as absolute Theme and, at
the same time, marking it for focus: al dɩɛ (‗as for that‘) (see below on
emphatic pronouns). Next, he contrasts the utter failure in this farm with a
relatively better crop, a bɛw-kãa (‗the oily beans‘), which is focused in a
thematic equative. Let‘s also consider at a dialogue from the opening of a
radio interview below:

Host: A nyɩnɛ na fʋ tone?
DEF where IDENT.PL 2SG work.IPFV
Agric Officer: Ɩ tone nɩ „Animal Research‟.
1SG work.IPFV FOC
Host: ‗WHERE IS IT (that) you work?‘

Agric Officer: ‗I work AT ANIMAL RESEARCH.‘ (Workshop Interview)
In this text, the interviewer places focus on the Q-element, nyɩnɛ (‗where‘),
thereby activating a presupposition of a number of places the interviewee
could be working. The interviewee, on the other hand, construes the content of
the question as Given information and indicates unmarked focus on his work
place, Animal Research.
The discussion will proceed to consider emphatic pronouns as another
realisation of contrastive focus in Dagaare.
5.4.2.2 Emphatic Pronouns29

Emphatic pronouns are inherent focal lexicogrammatical items (see Chapter 3
for Dagaare pronominal system). For a start, let‘s consider an instance of their
use in the dialogue below (emphatic pronouns are in bold):
(34) St. Maria play

Father: Ɩ kɔ =n a wιɛ za.
1SG farm.PFV FOC DEF farm all
29
Since personal pronouns are interpersonal deixis, emphatic pronouns, as the examples given
would suggest, also indicate interpersonal stance and attitude. The discussion here is,
however, limited to their function as contrastive focus markers.
237
Mother: A za?
DEF all
[…] […] […]
Son: //Mãa kɔ a. // A kpantole
1SG.EMP farm.PFV 3PL.NHM DEF mounds
nɛ fʋ na daa nyɛ a,
DEM.DIST 2SG REL three days ago see.PFV JUNC
mãa baa kɔ. //
1SG.EMP complete farm
Father: ‗I have weeded the whole farm.‘
Mother: ‗All?‘
Son: ‗I weeded it. Those mounds you saw three days ago, I finished
weeding (them).‘
In this exchange, father and son report to mother details of their activities in
the farm. It is interesting to compare the utterances by the father and the son.
The father reports his accomplishment to his wife, indicating unmarked focus
on the extent of his work. The woman follows this with an interrogation of the
New information. Their son then construes the preceding discourse between
father and mother as Given information and marks contrastive focus on the
Subject, using an emphatic pronoun, Mãa (‗me not him‘).
As pro-forms, emphatic pronouns can occur at the various nominal
positions in the clause. From the textual point of view, they can be absolute
Theme (35 & 36), and, from the interpersonal perspective, they can be Subject
(37) or Complement (38). With this versatile characteristic, they expand and
enrich the focus potential of Dagaare. The examples below illustrate the first
person pronoun in different positions:
(35) St. Maria play
//Mãa // Ɩ ciir =ι a lɛ
1SG.EMP 1SG dislike.PFV FOC DEF DEM.DIST
o! //
PRT
For ME I hate THAT o!
238
(36) St. Maria play
//Mãa na zι a na a, // dãa
1SG.EMP REL sit.PFV DEF DEM.PROX JUNC beer
bɛ ι ɩ yéle ι.//
NEG.IND.NFUT COP 1SG matter NAFFR
‗For ME SITTING LIKE THIS, beer is not my problem.‘
(37) St. Maria play
// Fʋ nιbɛ lɛ bɛ a. // Mãa bãw na? //
2SG people COP 3PL.HM JUNC 1SG.EMP know.PFV AFFR
‗They are your people. Do I know?‘

(38) St. Maria play
Pʋr kʋ̀ mãa! […] //Ɩ yèl kɛ fʋ
pour.PFV give.PFV1SG.EMP […] 1SG say. PFV PROJ 2SG
pʋr kʋ̀ =m [….]! // Mãa so

pour.PFV give.PFV 1SG.ACC […] 1SG.EMP own.PFV
a yir.//
DEF house
‗Pour (some) for ME! […] I say that you should pour (some) for me
before … I own the house.‘
In each of the uses of the emphatic pronoun, the speaker asserts himself in
contrast with presupposed ‗other(s)‘. In (35), the emphatic pronoun is absolute
Theme and it has been set aside as an independent information unit. In (36), it
is again an absolute Theme but is qualified by a relative clause. Again, it is an
independent information unit from the rest of the clause. This consistent
representation of marked focus also functioning as absolute Theme
corroborates Halliday and Greaves (2008: 106) observation on English that
―when a clause is mapped into two information units, in a clear majority of
cases the boundary coincides with that between Theme and Rheme.‖ In (37),
the emphatic pronoun is Subject of the clause (and unmarked Theme) and it is
set in contrast with Fʋ (‗you‘) in the preceding clause. In (38), it first occurs as
a Complement in an imperative clause and, in the final clause, as Subject in a
declarative clause. Also compare the pronoun mãa in Pʋr kʋ̀ mãa! with the
239
non-emphatic use in the clause pʋr kʋ̀ =m! In the second clause, the whole
clause is repeated in broad focus.
The versatility of the pronouns allows them to combine with other
focus constructions such as thematic equatives and, even, the focus particle,
when they occur as Complement in a positive declarative clause (also see
below on exclusive particles). In any environment, however, the emphatic
pronoun still maintains its contrastive meaning. In the underlined clauses
below, the emphatic pronoun fʋʋ (second singular) occurs in a thematic
equative:
(39) St. Maria play

A: Fʋ nyɛ na [[fʋ na diw bɛr
2SG see.PFV AFFR 2SG NMLZ drive.PFV leave.PFV
a]]?
JUNC
B: Fʋʋ bɛ nʋ diw bɛr

2SG.EMP NEG.IND.NFUT IDENT.SG drive.PFV leave.PFV
ɛ?
NAFFR
A: ‗You see that you have drove her away? (=Have you seen what you
have done?)‘
B:‗YOU ARE NOT THE ONE (who) drove her away? (=are YOU NOT THE
ONE …?)‘
(40) St. Maria play
//Fʋ mɔ na. // Fʋʋ nʋ tɛr a
1SG do:well. PFV AFFR 2SG.EMP IDENT.SG possess DEF
faw // na tʋɔ kul a ɩ

strength POS.IND.FUT be:able.PFV marry.PFV DEF 1SG
pɔw-yaa. //
daughter
‗You‘ve done well. YOU ARE THE ONE (who) has the strength to marry
my daughter.‘
In each of the underlined clauses, the emphatic pronoun will still be

contrastive without the thematic equative (e.g. Fʋʋ diw bɛr, ‗YOU drove her
240
away‘; Fʋʋ tɛr a faw na tʋɔ kul a ɩ pɔw-yaa, ‗YOU are strong enough to marry
my daughter‘).
Table 5.2. Illustration of emphatic pronouns in different clause positions

Participant Emphatic [contrastive focus] Non-emphatic [broad
in focus focus]
Beneficiary Ʋ kʋ̀ =n Ʋ kʋ̀ bɛ
3SG give.PFV FOC 3SG give.PFV 3PL
bɛl. na.
3PL.EMP AFFR
‗He has given (it) to THEM.‘ ‗He has given (it) to them.‘
(those have already been given,
but not us as yet).
Bɛlɛ lɛ bɛ ʋ
3PL.EMP COP 3PL 3SG
kʋ̀.
give.PFV
‗THEY are the ones he gave (it)
to.‘
Agent Bɛlɛ kʋ̀ =m Bɛ kʋ̀ mɛ
3PL.EMP give.PFV 1SG.ACC 3PL give.PFV 1SG.ACC
‗THEY gave (it) to me.‘ (they na.

can‘t say it‘s someone else) AFFR
‗They have given (it) to me.‘

Bɛlɛ lɛ bɛ kʋ̀
3PL.EMP COP 3PL.HM give.PFV
=m
1SG.ACC
‗THEY are the ones (who) gave

(it) to me.‘
However, the different positions available for deploying the emphatic

pronoun do sometimes construe subtle variations in meaning. We set up
constructed examples with the pronoun bɛlɛ (third plural, emphatic) in Table
241
5.2 to illustrate this phenomenon (bɛl is a phonetically reduced form).30As the
English glosses in the table show, different interpersonal meanings can be
evoked by placing emphatic pronouns at different clause positions or in
different construction types.
5.4.2.3 Exclusive Particles
The focusing effects of exclusive particles have been noted in typological

studies (Beaver & Clark 2008; Güldemann et al . 2015: 171-2). However, they
are often not treated as part of ‗canonical‘ markers of focus per se. Güldemann
et al. (2015), for instance, characterise them as ―focus-sensitive elements that
associate with focus‖ (p. 171). While in some languages such as English,
exclusive particles may not be centrally involved in the system of
INFORMATION, they are salient in the focus system of Dagaare.
Dagaare exclusive particles can be grouped into two main types, based
on their grammaticalisation sources. The first type derives from adverbial
particlesand the second type originates from emphatic pronouns, specifically
third person (cf. Table 5.3). 31 As Table 5.3 shows, each type displays
characteristics of its source. Those particles with adverbial origin can be
classified into two sub-groups according to their meaning. The particles ende,
ɛ, dee, and dɩɛ have the English sense as for, and gba and yaa, on the other
hand, can be glossed in English as even (see Chapter 3, Section 3.4.2.3 on
adverbial particles). Exclusive particles with pronominal origin contrast
between singular (ʋl) and plural (bɛl and al) and, among plural, human (bɛl)
and non-human (al). With the exception non-human al, these exclusive
particles have lost their person distinction (for details on this
grammaticalisation phenomenon, see Mwinlaaru & Yap 2017.; also see
Chapter 3, Section 3.4.2.4 on pronouns).
30
The reduced form bɛl is the typical realisation of the pronoun as Complement (or ‗Object‘).
Generally, in Dagaare, pronominal items are often reduced or even become clitics when they
are used in Complement position.
31
See Mwinlaaru & Yap (2017) on the grammaticalisation of the pronominal-exclusive
particles from reduced forms of third person emphatic pronouns. The grammaticalisation
pathway can be summarised as: pronouns > demonstrative determiners > exclusive particles. It
should also be noted that the exclusive particles share the same forms with their adverbial and
pronominal sources.
242
Table 5.3. Dagaare exclusive particles32
adverbial origin pronominal origin

as for type even type singular plural
human non-human
ende gba ʋl bɛl al
ɛ yaa
dee
dɩɛ
Exclusive particles mark focus on nominal and adverbial groups, and, in terms
of placement, they are postpositive:
(41) Ʋ sãa-mɩnɛ bɛl wa na.

3SG father-PL EX come.PFV AFFR
‗As for her/his fathers, they have come.‘
Generally, exclusive particles can focus an element either clause initially,

before the verbal group realising the Predicator, or clause finally, after the
Predicator. However, it is mostly placed in clause initial position on the
Theme element of the clause, either Absolute (see (42) - (44) below) or non-
absolute Theme (as in 41 above). This gives us a situation where the Theme
element conflates with New information. Instances of exclusive focus are
given in the dialogic exchanges below (Focus element in bold):

I: A nɩ kʋɔnɩ nɩ a nɩ vʋʋnɩ nɩ a
DEF 2PL waters and DEF 2PL lights and DEF
nɩ sukuul yéle a Nandom ka, fʋ

2PL school matter DEF Nandom here 2SG
tɩɛrɛ kɛ bɛ tʋɔ na ŋmɛlɩ

think.IPFV PROJ 3PL.HM be:able.PFV AFFR turn.PFV
kaa bɩ?
check.PFV INT
32
As mentined in Chapter 3 on adverbial particles, the particle dɩɛ is borrowed from Akan and
it is synonymous with ɛ and ende. It is most likely that ende evolved from a fusion of ɛ and its
borrowed counterpart dɩɛ (cf. n. 6).
243
R: //Ʋʋ. //Kʋɔ yéle ende a, // kʋɔ yéle cãa
yes water matter ADV JUNC water matter be:still.PFV
na fɛrɛ tɩ a bɛ cãa \
AFFR worry.IPFV 1PL CONJ 3PL.HM be:still.IPFV
mʋɔr =ɩ …//
struggle.IPFV COM
I: ‗Your water and your electricity and your school problems at

Nandom here, do you think that they have been able to resolve it?‘
R: ‗Yes.‘ AS FOR WATER, water is still worrying us and they are still
33
struggling with it …‘
Host: Ok, mɛ́ bɛ pãa na tɩ kɔ
ok ADV 3PLHM ADV ADVLZ PST.REM farm.PFV
a, a wa de a bome
JUNC CONT EVT take.PFV DEF things
a, a tɩ maal =a bɩɩ a
JUNC 3PL.NHM PST.REM do:well.PFV AFFR or 3PL.NHM
cãa na ŋmɛ a lɛ a
be:still.PFV AFFR be:like.PFV DEF DEM 3PL.NHM
na mɩ́ tɩ ŋmɛ a?
REL HAB PST.REM be:like.PFV JUNC
Agric Officer: //Aaah a maalʋ ɛ̃, // a

INTJ DEF do:well. NMLZ EX DEF
maalʋ gaw ... a

do.well.NMLZ be:more:than.PFV 3PL.NHM
gaw nɩ taar. //
be:more:than.PFV FOC each other
Host: ‗Ok, like when they then made the farm, and then harvested the
crops, they [the crops] did well or they are like how they used to be?‘
Agric Officer: ‗Aaah, AS FOR THE DOING WELL, the doing well some
did better ... some did better THAN OTHERS.‘
33
I use ‗for …‘ in the English translation to show absolute Theme and ―AS FOR …‖ to mean
exclusive marker‖. This is just an improvised way of distinguishing the two functions rather
than translation equivalents. In contrast, as for is a marker of absolute Theme in English
grammar (cf. Matthiessen 1995: 552-554).
244
(44) St. Maria play
// A saalʋ gba // fʋʋ wa dɩ // ƴa-gan
DEF okra EX 2SG.EMP EVT eat.PFV body
mɩ́ salɛ fʋ na. //
HAB bore.IPFV 2SG AFFR
‗Okra even, when you eat (it), you feel bored.‘34
In extracts (42) to (44), exclusive focus is marked on the Theme of the clause,
setting these thematic elements apart as separate information units from the
rest of the clause. In (42), the particle ende places focus on kʋɔ yéle (‗water
matter‘) while ɛ marks a maalʋ (‗the doing well‘) in (43) as focused and gba
in (44) focuses A saalʋ (‗The okra‘). The use of exclusive particles with
absolute Themes, in particular, those realised by items other than emphatic
pronouns, is an interesting phenomenon. With this situation, the focus
potential of the Dagaare clause is complete. As we observed earlier, thematic
equatives focuses elements that have experiential function in the clause. Since
absolute Themes, do not have experiential function in the clause structure,
exclusive particles become the focus resource in this domain (except when the
absolute Theme is an emphatic pronoun). As in the examples above, exclusive
particles are not indicators of the absolute Themes since these thematised
constructions such as in (42) to (44) can still occur as Themes without the
exclusive particles.
Exclusive particles, especially those of pronominal origin (i.e. ʋl, bɛl
and al) tend to give rise to ambiguities when some of the clauses in which they
occur are considered in isolation. Such ambiguity is normally between
interpreting the particle as a focus marker or a demonstrative determiner when
it occurs with the third person (cf. Mwinlaaru & Yap 2017). Let‘s compare
(45) and (46) below for an illustration.
(45) Ʋ sãa ʋl wa na.

3SG father EX come.PFV AFFR
‗As for her/his father, he has come.‘
34
In the example, the speaker plays on the similarity between the word for okra, saalʋ, and
boredom salɛ. Note that the noun for salɛ is salʋ.
245
(46) A bie ʋl wa na.
DEF child ? come.PFV AFFR
‗As for her/his fathers, they have come / That child has come.‘
In (45), since, the nominal item Ʋ sãa (‗his/her fathers‘) in Ʋ sãa ʋl has a
more specific reference, ʋl can clearly be identified as an exclusive marker.
On the other hand, in (46), A bie (‗the child‘) is less specific and ʋl is
ambiguous between a demonstrative determiner that is adding to the
specificity of the nominal group and a focus marker. This ambiguity is,
however, a trace of the the diachronic development of the exclusive marker
and is often clarified by context in actual usage. For the adverbially derived
exclusive particles, on the other hand, substituting an emphatic pronoun in
Subject position with a non-emphatic pronoun changes their meaning from
focus markers to adverbial particles:
(47) Fʋʋ ende cen na.

2SG.EMP EX go.PFV AFFR
‗AS FOR YOU, you went.‘

(48) Fʋ ende cen na.
2SG ADV go.PFV AFFR
‗You went anyway.‘
First, the phenomenon gives evidence that the exclusive particles evolved
from adverbial particles in post-Subject position through the mechanism of
functional split. Second, the shift in meaning from adverbial reading in (48) to
focus reading (47) due to the difference in information prominence of the
Subject pronoun supports the interpretation of the exclusive particles as focus
markers. That is, they are attractive to other focus resources. Further to this
second point is the observation that, like emphatic pronouns, exclusive
particles combine with other focus constructions such as emphatic pronouns
and the focus particle to bring interesting interaction between textual meaning
and interpersonal meaning. This is illustrated further below (example (49) is a
repetition of (32) above):
246
Agric Officer: Soo a tɩtal pʋɔ na fʋ
so DEF plains inside IDENT.SG 2SG
dɛ bɛ nyɛ bom za ɩ.
ADV NEG.IND.NFUT get.PFV thing all NAFFR
Farmer: //Al dɩɛ // ɩ bɛ nyɛ

3PL.NHM.EMP EX 1SG NEG.IND.NFUT get.PFV
bom e.// Soo a bɛw-kãa na
thing NAFFR so DEF beans-oil IDENT.PL
ɩ nyɛ, // ɛcɛ a mɩ̀ bɛ

1SG get.PFV CONJ 3PL.NHM also NEG.IDENT
waar ɛ. //
come.IPFV NAFFR
Agric Officer: ‗So IT IS IN THE PLAINS you just didn‘t get anything.
Farmer: ‗AS FOR THAT I didn‘t get anything. So IT IS THE OILY BEANS I
got, but it is also not producing.‘
(50) St Maria play
// Fʋʋ gba// ɩ bɛ bãw lɛ za
2SG.EMP EX 1SG NEG.IND.NFUT know.PFV DEM.DIST all
ɩ na ɩ tɩ kul fʋ a. //
1SG REL do.PFV PST.REM marry.PFV 2SG JUNC
‗YOU EVEN, I don‘t know how I came to marry you.‘
In these contexts, the exclusive particles dɩɛ and gba add attitudinal meaning
to the clause, while maintaining their textual meaning of contrastive focus.
With the addition of dɩɛ to the already focus prominent al (third plural, non-
human) in (49), the farmer sounds more emphatic, claiming epistemic
certainty to the proposition. In (50), which is an utterance from a wife to the
husband in an unscripted play, the addition of the exclusive gba to fʋʋ enacts a
negative attitude towards the husband.
5.4.3 Further on Broad Focus

In Sections 5.4 and 5.4.1, we noted that imperative and negative clauses, in the
unmarked case, receive broad focus (see also Figure 5.2). In this section, we
247
throw more light on these constructions and other issues in relation to broad
focus. The absence of focus particles in both imperative and negative clauses
has been reported for many African languages, including Nupe (Niger-Congo:
Kwa), Rendille, Boni and Somali (all Afroasiatic: Cushitic) (cf. George 1971;
Sasse 1981; Heine & Reh 1983).
The choice of broad focus as the default focus type in imperative and
negative clauses can be explained by their discourse functions, that is, from
above lexicogrammar – in the semantic stratum (see Chapter 2 on
stratification). As mentioned earlier, imperative clauses are typically discourse
initialising devices. This relates to their function as proposals – i.e. offers and
commands (including requests, suggestions and other directives; cf. Chapter 4,
Section 4.4.2). Thus, in making proposals in the imperative clause, the speaker
presents the whole proposal as New in the unmarked case.
As resources for negating propositions, negative clauses, on the other
hand, already presuppose the positive value of a proposition, either based on a
preceding assertion, a contextual cue or some other kind of assumed
knowledge. The focus of the negative clause is therefore on the arguability of
the proposition, its negative value, which, in the Dagaare, is marked by the
negative particle (viz.: bɛ, kʋ́) within the Predicator and resonates in the non-
affirmative particle at the end of the clause (see Chapter 4, Section 4.4.1.1).35
Thus, the negative meaning, like other interpersonal meanings, is a prosody
which diffuses over the whole clause. In other words, we encounter an
information unit where focus is on the arguability of the proposition, diffused
as it were. It should also be noted that the negative clause is a marked choice
in the system of POLARITY (see Chapter 4 on polarity). In a corpus of around
18 million words, Halliday and James (1993) found that the ratio of positive
clauses to negative clauses in English is 9: 1. The figure we encounter in
Dagaare is not much different. Across several texts from different registers,
the ratio of positive to negative is, on the average, 9.7: 0.3 per text. 36 The point
being made here is that negative polarity is a marked polarity choice and, thus,
35
As noted in Ch. 4, other dialects of Dagaare such as Ngmere (‗Central Dagaare‘) and Waali
do not have the non-affirmative particle. In these dialects, the negative clause is negotiated by
phonological prosody – a high tone on the last word.
36
Out of 375 clauses across different registers, positive is 362 (96.5%) and negative is 13
(3.5%).
248
when it occurs, the whole clause is brought into focus in the default case. As
we discussed in the preceding section, negative clauses as well as imperative
clauses do receive marked focus.37
Apart from negative and imperative clauses, end focus is sometimes
not indicated where it is expected. That is, end focus can be suspended to
place focus on the whole information unit (which is co-extensive with the
clause in this case). This phenomenon is common in narrative discourse,
where the speaker adds each information unit cumulatively as New as the
narrative builds up.38 As an example, we illustrate this phenomenon with the
Placement (i.e opening) of a traditional narrative text, a folktale (relevant
clauses underlined):
(51) Folklore
[1] Nɩ gure, gure o!
2PL be:alert.IPFV be:alert.IPFV PRT
[2] ||| A wɛ-dʋn za bɛ lɛ bɛ||

DEF farm-animals all NEG.IND.NFUT COP 3PL.HM
tɩ be be!|||
PST. REM EXIST.PFV there
[3] Kɔ kpɛɛ kpɛɛ kàw wa lo,
famine big big some EVT fall
[4] a bɛ ƴaw mɩra
CONT 3PL.HM put law
[4.2] ɛ fʋʋ nɛ za wa gã
PROJ 2SG.EMP.COND DEM.DIST all COND sleep.PFV
zanɩ nyɛ fʋ 'wɔb wɛ-dʋw
dream.PFV see.PFV 2SG chew.PFV farm-animal
za a,
all JUNC
[4.3] ɛ bɛ̃̌ ɛ nyɔw a wɛ-dʋw nɛ

PROJ 3PL.HM catch DEF farm-animal DEM.DIST
37
In the imperative clause, only emphatic pronouns and exclusives particles are used for
contrastive focus.
38
A similar phenomenon has been reported for Japanese narrative discourse (Teruya 2007).
249
kʋ̀ fʋ fʋ 'wɔb.
give.PFV 2SG 2SG chew
[5] Tɔɔ, a zie wa sɔb,
well CONT place EVT dark
[6] a bɛ kpɛ a gaa,
CONT 3PL.HM enter DEF sleep
[7.1] bɛ zuru gba bɛ tɔ tɩɩ-sɔw sɛr
3PL.HM heads ADV NEG.IND.NFUT touch ground yet
ɛ,
NAFFR
[7.2] ɛ Gbowgbor ír ŋmɛ cɛl-nyɛ […]

and wolf get up.PFV beat.PFV wail
‗Once upon a time, was it not all the animals in the forest that lived? (=
there lived all the animals in the forest!).39 Then, a very big famine
broke out. Then, they made a law that if anyone dreamt of eating any
animal, that they should give that animal to you to eat. Well, then night
fell. Then they went to sleep. Their heads had not even touched the
ground yet, and Wolf got up and alarmed …‘
In (4) and (5) in the extract above, the Complement elements mɩra (‗law‘) and
a gaa (‗sleeping‘) are potential candidates for end focus. However, these
clauses are presented as all-New-information in the continual flow of narrative
events. This textual status is indicated explicitly by the continuative marker a,
which always blocks focus whenever it occurs in a clause. The other clauses in
the extract receive broad focus in the unmarked instance. But, even here, there
is an interesting observation with regards to negotiation. All other things being
equal, clause (3) and (5) would take the affirmative particle na in clause final
position. Its absence clearly shows that the speaker is not ready to handover
the turn. As with focus, this phenomenon is a general principle; when the
continuative marker a introduces a clause, the affirmative marker is
39
The complex A wɛ-dʋn za bɛ lɛ bɛ tɩ be be! (‗Was it not all the animals that existed!‘) is an
instance of the use of a grammatically negative sentence to express a positive meaning, a kind
of ‗polarity metaphor‘ as it were. It should be noted that it does not end with the non-
affirmative Negotiator as is characteristic of true negative clauses and that it is said with a
question tone (i.e. rising tone).
250
automatically blocked (see clause (5)). As another illustration of this
phenomenon, compare the underlined clauses in the dialogue below (see
example (20) for a longer version):

wa ɩ?
come.PFV NAFFR

Baba: They are bringing (it) and (they) have not come?‘
It can be observed that the use of the continuative marker in the second clause
correspond with the absence of the Negotiator element (potentially, the
affirmative na). The speaker seems to be saying: ‗Look! I‘m still building up.
I‘m not done yet.‘ In this sense, negotiation is like focus; both are oriented
towards the listener as the speaker potentially hands over the turn, the former
enacting interpersonal stance and the latter indicating newsworthy information
(‗textual stance‘). In summary, constructions that typically take broad focus
and thus do not occur with the focus particle are negative clauses, imperative
clauses and clauses with the continuative marker a.
5.4.4 Information Unit and Clause
In concluding this section, we will briefly revisit the issue of the relationship
between the clause and the information unit. Both of them are grammatical
units, with the information unit uniquely dedicated to the realisation of focus.
In the unmarked case, the information unit is co-extensive with the ranking
clause (cf. Halliday & Greaves 2008; Halliday & Matthiessen 2014).40 What
this means is that when the focus type in the declarative clause is end focus
(indicated by the particle nɩ) as opposed to contrastive focus, the clause is
40
In this case, Dagaare is like English. We can assume that in languages where the clause and
the information unit are separate units, they will always be co-extensive in the unmarked case.
251
always equal to an information unit. This is the dominant textual organisation
in the Cain and Abel text introduced at the beginning of the chapter. The
following two clauses are repeated for illustration:

A Adama nɩ a Awa tɩ dɔw nɩ
DEF Adam and DEF Eve PST born FOC
bibiir ayi: Kayɛ nɩ Abɛl. A Abɛl wa

children two: Cain and Abel DEF Abel EVT
ɩ =n pi-cɩɩnɛ[…]
COP.PFV FOC shepherd […]
‗Adam and Eve gave birth to TWO CHILDREN: CAIN AND ABEL. Abel
became A SHEPHERD…
Here, each of the two clauses constitutes one information unit. As with end
focus, broad focus also presents the information unit as being co-extensive
with clause. An example is the following clause, which is a continuation of the
extract in (53) above:

[…] ɛ a Kayɛ ɩ kʋɔra.
[…] CONJ DEF Cain COP.PFV farmer
‗… and Cain was a farmer.‘
Since broad focus is the default focus type for the imperative and negative
clauses, this means that these clauses types are normally co-extensive with one
information unit if they are not marked for contrastive focus.
On the other hand, contrastive focus sometimes means that the
information unit does not correspond to the clause in either of two ways,
namely (i) one clause may map onto two information units and (ii) two clauses
may combine into one information unit. The first is associated with contrastive
focus realised by emphatic pronouns (55-56) and exclusive particles (57),
specifically when the focal element is also thematic:
252
(55) St. Maria play [contrastive focus + broad focus]
// Mãa// ɩ yire na o! //
1SG.EMP 1SG go:out.PFV AFFR PRT
‗For ME I‘m going out o!‘

(56) St. Maria [contrastive focus + end focus]
//Mãa // lιɛbɛ nι faara o! //
1SG.EMP turn.IPFV FOC priest PRT
‗I‘m becoming a PRIEST o!

(57) Workshop Report 5 [contrastive focus + broad focus]
//Al dɩɛ // ɩ bɛ nyɛ bom e. //
3PL.NHM.EMP EX 1SG NEG.IND.NFUT get.PFV thing NAFFR
‗AS FOR THAT I didn‘t get anything.‘
Each of these clauses embody two information units. While (55) and (57) are
composed of contrastive focus and broad focus, (56) consists of contrastive
focus and end focus. The nominal groups Mãa in (55) and (56) and Al dɩɛ in
(57) are the Themes of the clauses in which they occur. At the same time, they
are marked for contrastive focus and are presented as separate information
units from the rest of the clause. In other words, in both (55) and (56), for
instance, the speaker starts the clause with one quantum of information, saying
‗Look! This is about me‘, and then follows this with another quantum, ‗I am
going out‘ and ‗I am becoming a priest‘, respectively.
Instances where two clauses combine into one information unit are
typically where contrastive focus is marked by thematic equatives, as in (58)
below:

So a tɩtal pʋɔ na fʋ dɛ
so DEF plain inside IDENT.PL 2SG ADV
bɛ nyɛ bom za ɩ?
NEG.IND.NFUT get.PFV thing all NAFFR
‗So IT IS IN THE PLAIN you didn‘t get anything?‘
Here what would be a monoclausal construction in the unmarked case has

been configured as a bi-clausal construction, specifically, a hypotactic clause
253
complex, to place contrastive focus on the circumstantial element (a tɩtal pʋɔ,
‗in the plain‘). In this kind of bi-clausal construction, the equative clause,
which is the focal element, becomes the matrix clause while the out-of-focus
clause is dependent on it (cf. Heine & Reh 1983; Harris & Campbell 1995). 41
We can summarise the relationship between the information unit and the
clause diagrammatically as in Figure 5.4, with the broken line indicating that
the relationship is not always one to one.
Figure 5.4: Information unit in relation to clause on the rank scale

(adapted from Halliday & Matthiessen 2014: 115)
5.5 Conclusion
In conclusion, this chapter has presented a system-based account of THEME

and INFORMATION in Dagaare, drawing on theory and insights from the
descriptive and typological literature on information structure. The thematic
structure of the Dagaare clause, as in other languages, is identified as
consisting of Theme and Rheme. Theme has been defined and identified
trinocularly. Semantically, it is identified as the local context that serves as the
point of departure of the clause, orienting it to a particular interpretation.
Within lexicogrammar, and at clause rank, it is identified as that element that
is given initial prominence in the clause and that is developed by the Rheme.
41
See Mwinlaaru and Yap (2017) on how this focus strategy led to the development of the
focus particle nɩ (also see Chapter 6, Section 6.5.2.1 for a brief account).
254
Below the clause, different forms of realisation have been identified for three
types of Theme: textual, interpersonal and topical Themes. It has also been
discussed that topical Themes in Dagaare can be either Absolute or non-
Absolute and, if non-Absolute, they can be unmarked or marked. These
different types of topical Theme are identified based on their status or function
in relation to the transitivity structure of the clause. In addition, although
unmarked Theme is normally the Subject in the modal structure of the clause,
in first singular imperative clauses it is normally the Predicator.
On the other hand, the focus structure of the Dagaare clause is
minimally realised by the New element and it often combines with the Given
element, which is an optional element in the structure. New is identified
semantically as that element that is singled out in the information unit as
newsworthy. Three focus types have been identified, end focus, contrastive
focus and broad focus. End focus is the default choice for positive declarative
clauses and it is realised by the focus particle nι. Contrastive focus is realised
by thematic equatives, emphatic pronouns and exclusive markers. Broad focus
is identified as zero-realisation and it is the default choice for imperative and
negative clauses. It has also been shown that the domain of the realisation of
focus is the information unit rather than the clause although clause and
information unit are co-extensive in the unmarked instance. Theme and focus
together create the textual essence of the Dagaare clause. They reorganise the
resources of the other metafunctions, ideational and interpersonal (specifically,
mood), into ‗a processable syntax‘.
255
CHAPTER SIX
TRANSITIVITY: REPRESENTATION OF EXPERIENCE
6.1 Introduction
The two preceding chapters discussed the interpersonal and textual systems of
the Dagaare clause. Chapter 4 discussed the modal structure of the clause and
the system of MOOD, and has shown how it interacts with other interpersonal
systems, both verbal group, (i.e. POLARITY and MODALITY) and the clause (i.e.
NEGOTIATION). Chapter Five considered the textual metafunction,
investigating how the systems of THEME and INFORMATION combine to
organise the clause as a processable semiotic entity in the flow of discourse.
This chapter proceeds to examine a third major function of the Dagaare clause,
namely, its function as a representation of experience. In other words, the
chapter answers the question: How is the Dagaare clause organised to
represent different domains of human experience, in general, and the
experience in the folk life of the Dagara speech fellowship, in particular? The
major grammatical system relevant to this function is that of TRANSITIVITY.
The chapter will begin with an overview of the general function of language as
a reflection of experience, noting the various ways this function of language
has been theorised in linguistic science (Section 6.2). It then proceeds to
discuss one major sub-system of transitivity, the system of PROCESS TYPE
(Section 6.3 – 6.6). This will be followed by a discussion of some

generalisations that can be made across the various process types in Dagaare
(Section 6.7.1) and finally an account of the system of AGENCY, the second
major sub-system of transitivity (Section 6.7.2). The chapter concludes with a
summary of the grammar of experience in Dagaare (Section 6.8).
6.2 The Clause as a Representation of Experience

The recognition that language is a semiotic representation of human
experience is an old one in the history of linguistic science. It is a major motif
in the writings of Franz Boas, Edward Sapir, Benjamin Lee Worf, Ferdinand
de Saussure, Bronislaw Malinowski and J. R. Firth as well as Vilém Mathesius
andother linguists of the Prague school. Sapir (1921: 180), for instance, wrote
256
that: ―The latent content of all languages is the same – the intuitive science of
experience. It is the manifest form that is never twice the same‖ (original
emphasis). In contemporary linguistics, this representational function of
language has been a motif of functional linguists of different persuasions,
notably cognitive linguistics (cf. Lakoff & Johnson 1980; Lakoff 2007;
Langacker 2008).
In the study of grammar, the experiential function of language has
always been implied in the analysis of transitivity. Since the 1960s, however,
Michael Halliday and lingusits working with systemic functional linguistics
have taken a system based approach to tranisitivity and have developed a more
explicit account of it as the major system for representing experience in
language, both in the description of English (e.g. Halliday 1967a, b, 1968,
1985; Halliday & Matthiessen 1999, 2014; Matthiessen 1995) and other
languages (e.g. Cafferel, Martin & Matthiessen 2004; Teruya 1998, 2007) and
also from a typological perspective (e.g. Matthiessen 2004; Wang & Xu 2013;
Teruya & Matthiessen, 2015). Valuable contributions to developing the notion
of transitivity have also been made by other functional typologists, noting it as
a pervasive grammatical system in human languages (e.g. Hopper &
Thompson 1980; Haspelmath 2015). In their famous account on transitivity in
grammar and discourse, Hopper and Thompson (1980) divided the notion of
transitivity into different opposing features and shows how semantically
similar features conspire, as it were, to classify clauses into degrees of low and
high transitiviy. The experiential orientation of the clause has also been
theorised and investigated under topics such as case theory (cf. Fillmore 1968,
1977; Haspelmath 2009b, and references therein), valency (Tesnière 1959),
thematic or theta roles (Chomsky 1981), Aktionsart (Vendler 1957) and
aspectual classes (Comrie 1976). Differences in these accounts inter alia
stems from the perspective from which the linguistic data is viewed, whether
from a semantic, grammatical or lexical perspective. Together, however, they
robustly show that language is a theory of human experience; it is essentially
structured to construe our experience of reality and of our consciousness
(Halliday 1973, 2005, 2008; Halliday & Matthiessen 1999).
The account of Dagaare in this chapter follows the systemic functional
tradition, by taking a system based, holistic approach to transitivity. The term
257
TRANSITIVITY is used as the overarching lexicogrammatical system for
construing experience in the Dagaare clause (cf. Halliday & Matthiessen 2014:
Ch. 5). This grammatical system is manisfested in the language under two
models, the transitive model, where the underlining sub-system is that of
PROCESS TYPE (cf. Section 6.3 – 6.6), and the ergative model, where the
underlining sub-system is that of AGENCY (cf. Section 6.7.2). Both the
transitive and ergative models of transitivity are concerned with the grammar
of the clause nucleus of ‗process + participants‘, that is, elements that are
brought into ‗perspective‘, in the sense of Fillmore (1977: 72-80), in
construing a quantum of change in the flow of experience. They constitute the
grammar of NUCLEAR TRANSITIVITY. In Dagaare, and across languages in
general, process is typically realised by verbal groups while participants are
typically realised by nominal groups. The clause nucleus established by
‗process + participants‘ is, however, expanded and augmented by adverbial
groups construing circumstances such as manner place and time as well as
adpositional groups. This experiential configuration of the clause is illustrated
in the box diagram below.
A sãa kɔ =n ci dɩya
DEF father farm.PFV FOC guinea corn last year
participant process participant circumstance
nom. group verbal group nom. group adv. group
determiner noun verb particle noun adverb
‗The father cultivated guinea corn last year.‘
In this example, the process, kɔ, and the participants, a sãa and ci, form the
nucleus of the clause. They construe a semiotic figure which is located within
a particular temporal circumstance realised by the adverb dɩya (‗last year).
The discussion will proceed to examine the different experiential
domians represented in Dagaare and the different configurations of participant
roles each domain engender in the clause. Due to time limitation,
CIRCUMSTIATION will not discussed. Readers are, however, referred to the
discussion on adverbs, specifically those of place, time and manner in Chapter
258
3 (cf. Section 3.4.2.3), one of the main resources for realising circumstances in
the Dagaare clause. The account here begins with a discussion of the various
process types identified in Dagaare. These comprise three principal types,
namely material clauses, mental clauses and relational clauses, as well as three
minor types – behavioural, verbal and existential clauses (cf. Halliday &
Matthiessen 2014: Ch. 5). Together these process types construe the semantic
space for the representation of experience in Dagaare. As in previous chapters,
the accounts presented here are the results of a lengthy analysis of Dagaare
texts and supported by theoretical and typological guidance and transfer
comparison (cf. Chapter 2, Section 2.7.2).
6.3 Material Clauses: Processes of Doing-&-happening

The first process type to be discussed is material clauses. These are clauses of
doing-&-happening. They construe a quantum of change in the sequence of
events as being brought about through some input of energy (Halliday &
Matthiessen 2014). They are resources for construing experience of dynamic
change in the flow of events in the world around us, comprising the
happenings and activities of the physical world and our own participation in its
unfolding drama. To begin with, we will first examine an extract that contains
predominantly material clauses. In this extract, agricultural officer describes
farming practices that were adopted by a group of farmers in a rural farming
project and material clauses serve to construct the flow of actions, activities
and events he recalls from the project:

Soo bɛ kɔ =n tɩntɛr. A ayɩ sob,
so 3PL.HM farm.PFV FOC plain DEF two own
bɛ ŋmàa nɩ gàrɩ […] A gàrɩ aŋa, tɩ
3PL.HM cut.PFV FOC ridges […] DEF ridges DEM 1PL
nyɔw na po ziir ayi. A mɩnɛ tɩ ƴãw

hold AFFR divide places two 3PL.NHM some 1PL put.PFV
nɩ kuolu […] a bʋn-buli pʋɔ ɛ
FOC fertilizer […] DEF plant inside CONJ
259
a mɩnɛ, tɩ ƴãw nasa-kuolu, […]. A
3PL.NHM some 1PL put.PFV European-fertilizer […] DEF
ata sob nɩ: tɩ kɔ na, […] ɛ

three own COP.FOC 1PL farm.PFV AFFR […] CONJ
fʋ mɩ́ kpɛ bɩwrɩbɩwrɩ a sɔwɔ ʋʋ
2SG HAB enter.PFV cut.DUPL DEF middle CONJ.3SG
tɛr bowli bowli, a kʋɔ mɩ́
possess.PFV small:holes small:holes DEF water HAB
wa a tʋɔ tewli a be.

come.PFV 3PL.NHM be:able stand DEF there
‗So they cultivated plains. The second one, they made ridges […]
These ridges, we divided (them) into two. Some of them, we put
fertilizer (=manure) in the plants and some of them, we put European
fertilizer […] The third one is: we cultivated, and you enter the middle
(of the farm) and cut it such that it has small holes so that when the
water (=rain) comes it will stand there.‘
The material clauses in the extract construe the text as a sequence of concrete
changes in the land brought about by the farmers. The cohesion devices such
as a ayi sob (‗the second one‘), a mɩnɛ (‗some of them‘) and a ata sob (‗the
third one‘), organise these changes into a topology of farming practices. We
can divide the material processes in the text broadly into two kinds. The first is
doings such as kɔ (‗cultivate‘), ŋmàa (‗cut‘), po (‗divide‘), and ƴãw (‗put‘).
These dominate the extract and construe the activities of the farmers. The
second type are happenings, comprising a kʋɔ mɩ́ wa (‗the water comes‘) a tʋɔ
tewli (‗it‘s able to stand / remain stagnant‘). These are presented as not directly
brought about by the farmers although the flow of events construes them as
resulting from planned activities of the farmers. What is common among both
‗activities‘ and the ‗happenings‘ is that they are physical changes taking place
in the world of outer experience. It is in this sense that the clauses that
construe them are ‗material‘ clauses. As we will see below, the grammar of
Dagaare, and perhaps all languages, sets them apart from processes in other
realms of experience.
One way in which we can identify material clauses is by considering
260
the nature of the participants taking part in the process construed by the clause.
These participants will be introduced into the discussion as different types of
material clauses are examined. One key participant that, however, runs
through the different kinds of material clauses is the Actor, defined as the one
who brings about the unfolding of the process, the one who acts. In (1), this
participant is predominantly the farmers, and is realised by the pronouns bɛ
(‗they‘), tɩ (‗we‘) and fʋ (‗you‘). The use of these different pronouns to
construe the Actor is interpersonally motivated. It reflects the shifting
intersubjective orientation of the speaker. From an ideational interpretation, it
shows the fact that language construes our experience of the world –
constructs it into meaning – rather than directly reflecting it. As the text
shows, the Actor is not realised by only human participants. The last two
clauses contain non-human participants as Actor, namely a kʋɔ (‗the water‘)
and a (‗they‘). The Actor is realized by a nominal group, which may simply be
a noun or a pronoun such as a (‗they‘).
In addition to the Actor, another core participant associated with
material clauses is the Goal, which is defined as the participant that is
impacted upon by the activity of the Actor. In the extract above, the Goal
participant include tɩntɛr (‗plains‘), gàrɩ (‗ridges‘), kuolu (‗fertilizer‘) and a
sɔwɔ (‗the middle‘). In the next section, we will extend our discussion of these
participant roles and show how they configuration leads to delicate
distinctions among material clauses.
6.3.1 Transitive and Intransitive Material Clauses
The first distinction to make among material clauses is between transitive and
intransitive clauses. An intransitive clause is a material clause where the
outcome of the process engaged in by the Actor is confined to the Actor itself.
Such a material clause represents a happening. In a transitive clause on the
other hand, the unfolding of the process extends to another participant, which
is the Goal, and impacts it in some way. Such a material clause represents a
doing (or an action). Examples (2) and (3) below illustrate intransitive and
transitive clauses respectively:
261
(2) Saa wa na.
rain come.PFV AFFR
‗It rained /It has rained.‘

(3) Tɩ ƴãw nɩ kuolu a bʋn-buli pʋɔ.
1PL put.PFV FOC fertilizer DEF plant inside
‗We put fertilizer in the plants.‘
In these examples, the Actor primarily has the same function in both the
transitive and intransitive clause. In (2) and (3), both saa (‗rain‘) and tɩ (‗we‘)
are construed as bringing about the unfolding of the process. However, the
presence of the Goal participant in (3) brings out a difference in the transitivity
value of the clauses, a transitive one, in which the initial phase of Actor +
Process configuration goes through to impact on a Goal and an intransitive
one, in which the configuration, as it were, does not go through. We consider
the meaning of Actor and Goal in the Dagaare material clause in more detail
below.
(i) Actor: The prototypical meaning of Actor derives from transitive clauses,
where it inputs the energy needed for the realisation of the process and
impacting on the Goal participant. In intransitive clauses, the definition of
Actor is more complicated than this. The Actor in an intransitive clause
behaves differently when the clause is also a medio-passive, as in (4) to (6)
below (see Section 6.7.2.1 on middle clauses). Here, the process impacts on
the Actor. In other words, the process in medio-passive clauses is represented
as a happening and the Actor serves as the medium through which the
happening is actualised. The Actor normally represents an entity which
naturally, or in all logical possibility, does not have the capacity to bring about
the unfolding of the process by itself.
(4) A tɩntɛr kɔ na.

DEF plains cultivated AFFR
‗The plain has cultivated (=The plain is cultivated).‘

(5) A dãa wa na.
DEF beer come.PFV AFFR
The beer/pito has come (=The beer has arrived).
262
(6) Kuolu ƴãw n= a bʋn-buli pʋɔ.
fertilizer put FOC DEF plants inside
‗Fertilizer has put inside the plants (=Fertilizer has been put in the
plants).‘
As the glossing in (4) to (6) shows, translating Dagaare medio-passive clauses

into English presents a challenge. Although the best way to translate them is to
use the receptive (or passive) clause, this does not actually capture the
meaning of what is happening, it misses the mark! For instance, while we can
add an Agent to the passive in English with the by-phrase, there is no way we
can add an Agent to the Dagaare original (cf. Bodomo, 1997). In principle,
every transitive clause (an action) in Dagaare can be converted into a medio-
passive construction (a happening). It is a phenomenon that is however largely
limited to only material clauses and is a feature that distinguishes material
clauses from other process types (the exception is those behavioural clauses
that are oriented towards material processes; cf. Section 6.6.2). They evolve in
the language to serve a particular function, which is to construe activities as
though they were happenings.
Ideationally, they allow speakers to obscure the real originators of the
process. The use of the medio-passive in the clause underlined below allows
the speaker to off-set the agency of the process realised in the clause:
(7) Report 7
Soo, a lɛ na tɩ bɔbr ɛ
so DEF DEM IDENT.PL 1PL want.PFV PROJ
sɛ mɔ tɩ yèl mɔ tɩɩ wa
PROJ ADV 1PL say.PFV ADV 1PL.EMP EVT
tɩ kʋɔr a al a, tɩɩ ɩ
1PL farm.IPFV DEF DEM JUNC 1PL do.PFV
mɔ tɩ ɩ a nɩ dabaala kàw,
ADV 1PL do.PFV 3PL.NHM FOC something some
tɩɩ kàw ƴãw.
medicine some put.PFV
263
‗So that is why we want like to say like when we are cultivating those
ones, we should like do something about them. Some medicine should
be in / Some medicine should be put (in the plants).‘
This extract is highly modulated as is evident in the use of the adverbial

particle mɔ (cf. Chapter 3, Section 3.4.2.3). It is a proposal made by a female
farmer to a group of agricultural extension officers at a workshop. The use of
the medio-passive allows her to make her appeal a general one without
necessarily assigning agency to the officers. In other words, she construes the
clause as a happening as opposed to an action.
However, the likelihood of representing a material clause in the medio-passive
mode over the effective clause depends on the agency value of the entity
construed as Actor (see Section 6.7.2.2 on effective clauses). Let‘s compare
the pair of clauses in (8) and those in (9) in order to explain this phenomenon:
(8a) Ɩ ɓaw nɩ a gan.
1SG pick.PFV FOC DEF book
‗I picked the book.‘
(8b) A gan ɓaw na.
DEF book pick.PFV AFFR
‗The book is picked.‘
(9a) Ɩ ɓaw fʋ na.
1SG pick.PFV 2SG AFFR
‗I picked you.‘
(9b) Fʋ ɓaw na.
2SG pick.PFV AFFR
‗You picked (it)‘
In terms of agency, the (a) clauses are effective clauses while the (b) clauses
are middle clauses. (8a) and (8b) only differ in agency but similar in content
meaning. While (8a) construe the clause as action with Actor and Goal, (8b)
construes it as a happening impacting on the Actor. With (9a) and (9b), the
264
situation is different. They cannot be said to be agnate.They actually construe
different experiences since both Actors have the capacity to bring about the
process of picking.
(ii) Goal: The Goal participant is also realised by a nominal group and it is an
inherent participant in transitive clauses. It corresponds to the Complement
element in the interpersonal structure of the clause. However, the Goal is not
always realised in Dagaare transitive clauses although the clause still
maintains the sense of the process impacting on an entity (10). This
phenomenon is textually motivated in the sense that the Goal participant is
normally stated in the clauses as the culmination of new information and is left
unstated if it is known to the interactants from preceding the discourse or can
be inferred from the context of situation. This allows the rest of the clause to
be presented as New information, something the addressee does not know yet
about the Goal (cf. Section 6.7.1; Chapter 5, Section 5.4.1). The the following
clause from the text in (1) illustrates this point (relevant clause is underlined):

Bɛ ŋmàa nɩ gàrɩ […]. A gàrɩ aŋa,
3PL.HM cut.PFV FOC ridges […] DEF ridges DEM
tɩ nyɔw na po ziir ayi.

1PL hold AFFR divide places two
‗They made ridges. These ridges, we divided (them) into two.
In the first clause, bɛ ŋmàa nɩ gàrɩ (‗they made ridges‘), the Goal participant is
gàrɩ (‗ridges‘) and it is the focus of New information. The Goal is picked up in
the next clause as absolute Theme and the rest of the clause gives information
about this Theme, that is, the Rheme. In this second clause, the Goal
participant is left unrealised since it can be inferred from the surrounding text
as given information. The nursery rhyme in (11) illustrates this phenomenon
more clearly (relevant clauses are in bold):
(11) Nursery rhyme

Teacher: Bibiir e!
children VOC
265
Children: Yeeye!
INTJ
Teacher: A tea wa?

DEF be:where
Children: Tɩ nyu na.
1PL drink.PFV AFFR
Teacher: A bread wa?

DEF be:where
Children: Tɩ ɔb =a.
1PL chew.PFV AFFR
Teacher: A cup wa?

DEF be:where
Children: Ʋ ŋmɛr =a.
3SG break.PFV AFFR
Teacher: Ɩ na fɔb nɩ na o!
1SG POS.IND.FUT cane.PFV 2PL AFFR PRT
Children: Yaa yaa yaa.

pardon
Teacher: ‗Children!‘
Children: ‗Yeeye!‘
Teacher: ‗Where is the tea?‘
Children: ‗We have drunk (it).‘
Teacher: ‗Where is the bread?‘
Children: ‗We have eaten (it).
Teacher: ‗Where is the cup?‘
Children: ‗It has broken.‘
Teacher: ‗I will beat you o!‘
Children: ‗We beg for pardon.‘
In the underlined clauses above, the implied Goal is ‗tea‘ and ‗bread‘, which
are left unrealised because they can be inferred from the teacher‘s question.
However, it is not always the case that the Goal can be left unstated as
has been illustrated in the examples above. There is another factor affecting
transitivity in Dagaare apart from the textual motivation. This relates to the
266
experiential value of the entity realising the Goal, its value on the empathy
hierarchy; specificity whether it is animate or not and if animate whether it is
human or non-human. If we take this factor into consideration, we would have
to modify our general principle in this way: that non-human Goal participants
can be unrealised if they can be assumed from the context of discourse while
human Goal participants are normally realised (see Section 6.7.1 for further
details). Thus, if we were to replace A gàrɩ aŋa (‗these ridges‘) in example
(12) with a bie ŋa (‗this child‘), the Goal participant must be realized by a
pronoun:
(12) A bie ŋa, tɩ nyɔw ʋ na po ziir ayi.

DEF child DEM 1PL hold 3SG AFFR divide places two
‗This child, we divided him/her into two.‘
This principle of empathy hierarchy, however, opens up for a range of

linguistic manipulation and play for pragmatic effects. On the one hand, a non-
human Goal participant can be repeated in discourse for foregrounding, and,
on the other hand, a human participant may be left unrealised to construe some
special meaning in context. Let‘s illustrate the first point with with the
following extract (also cf. Chapter 4 for a longer version; on focus here is the
turns by the father):
(13) St. Maria

Father: ||| Fʋ saw na || kɛ ɩ de
2SG agree.PFV AFFR PROJ 1SG take.PFV
a bome ana a, Ayɔɔ? |||
DEF things DEM JUNC Ayɔɔ
Daughter: Ɩ saw na.
1SG agree.PFV AFFR
[…] […] […]

Father: |||Ayɔɔ ɔ, fʋ saw na ɩ||
Ayour, VOC 2SG agree.PFV AFFR 1SG
wob ι?|||
pick.PFV PRT
267
Daughter: Ɩ saw na.
1SG agree.PFV AFFR
Father: ||| Fʋ saw na || ɩ de a

2SG agree.PFV AFFR 1SG take.PFV DEF
bome ana ƴãw a zʋbɛ pʋɔ || ɛcɛ a

things DEM put.PFV DEF pocket inside CONJ DEF
pobile na kʋl fʋ ʋ? |||

gentleman DEM marry.PFV 2SG JUNC
Father: ‗Do you agree that I take these things, Ayour?‘

Daughter: ‗I agree.‘
Father: ‗Ayour, do you agree I should pick (these things)?‘
Daughter: ‗I agree.‘
Father: ‗Do you agree I take these things and put them inside my
pocket and this gentleman marries you?‘
This extract recreates traditional marriage negotiation in the Dagara society.

Within the overall exchange, only part of which is given here, the Goal
participant is left unstated in only one clause, ɩ wob ι? (‗should I pick?‘). The
repetition of the Goal serves to bring the pride price to the fore as the seal to
the marriage contract. The father intends here to absolve himself of any blame
by consistently drawing her daughter‘s attention to the marriage deed.
A human Goal participant may also be left unrealised. Let‘s examine
an instance in the dialogue in (14) below (relevant clauses are underlined):
(14) St. Maria

Gentleman 1: Fʋ bʋɔ, nʋ́ɔ bιι baa nʋ?
2SG goat fowl or dog IDENT.SG
‗Is s/he your goat, fowl or dog?‘

Gentleman 2: Ou! Ɩ pɔw nʋ.
excl. 1SG wife IDENT.SG
‗Oh! S/he‘s my wife.‘

[…] […] […]
268
Gentleman 1: ʋ̀ ʋ́, fʋ pɛrʋ nʋ.
no 2SG sheep IDENT.SG
‗No, she‘s your sheep.‘ (St. Maria Play)

Gentleman 2: Mãa, mãa ɩ ciir =ɩ a lɛ
1SG.EMP 1SG.EMP 1SG hate.PFV FOC DEF DEM
o! Fʋ bɛ nyɛ [[ɩ na
PRT 2SG NEG.IND.NFUT see.PFV 1SG NMLZ
tɛr taya a]]?

possess.PFV catapult JUNC
Gentleman 1: Hɛh, fʋ nyɛ nɩ kuur ɩ?

INTJ 2SG see.PFV FOC hoe JUNC
Lady: Nɩ bɛr a zɛɛb! Nɩɩm

2PL stop.PFV DEF fighting 2PL.EMP
wa bɔbr mɛ, nɩ wa ɩ
COND want.PFV 1SG.ACC 2PL come.PFV 1SG
saamɩnɛ zie!
Fathers‘ place
Gentleman 2: [lady already leaving] Cere! Fʋ nyɛ
go 2SG see.PFV
na [[fʋ na duw bɛr
AFFR 2SG NMLZ chase.PFV leave.PFV
a]]?
JUNC
Gentleman 1: Fʋʋ bɛ nʋ duw

2SG.EMP NEG.IND.NFUT IDENT.SG chase.PFV
bɛr ɛ?
leave.PFV NAFFR
1: ‗Is she your goat, fowl or dog?‘

2: ‗Oh! She‘s my wife.‘
1: ‗No, she‘s your sheep.‘
2: ‗As for me, as for me, I hate that o! Haven‘t you seen (that) I have a
hand catapult?‘
1: ‗Herh, have you seen a hoe? (=Have you seen that I‘m holding a
hoe?)‘
269
Lady: ‗Stop fighting! If you want (to marry) me, go to my fathers!‘
2: ‗Go! Have you seen (that) you drove (her) away?‘
1: ‗Are you not the one who drove (her) away?‘
In this scenario, two gentlemen quarrel over a lady they are in love with and
they blame each other for making her go away. In the clauses construing the
departure of the lady, the Goal is not realised even though the participant is
human (see underlined clauses). The interpretation is that the lady is
objectified as a lost opportunity, and this corroborates the animal references at
the beginning of the extract. Within the local context of the play, it contributes
to the comic effect of the play although it may have more general sociocultural
implications that are worth further investigating. The phenomenon of leaving
participants in complement position unrealised is, however, not limited to
material clauses but also applies to mental and behavioural clauses. We will
discuss this phenomenon in more detail in Section 6.7.1.
What we have demonstrated with the two extracts above is the
elasticity and evolving nature of language. Language evolves resources to
construe particular meanings and once certain norms and principles are
established, they are exploited to further create new meanings. Our Dagaare
scenario demonstrates the potential of the transitivity system to be exploited
for value-loaded meanings.
6.3.2 Extending Material Clauses in Delicacy: Additional Participants
We have introduced the principal types of material clauses as transitive and

intransitive and have noted the centrality of the Actor and Goal participants in
material clauses in general. Besides these participants, however, there are
other participants associated with material clauses. These additional
participants expand the core structure of the material clause and identify more
delicate sub-types. In this section, we will discuss these other participants,
comprising Scope, Recipient, Client, Instrument, Accompaniment and
Initiator.
(i) Scope: The Scope participant typically occurs in intransitive clauses. As the
examples below show, it occurs in a similar position as the Goal. From a
270
semantic point of view, the difference between Scope and Goal, however, is
that the Scope is not impacted by the process. Syntactically, while the Goal
participant can always be made the Subject of an agnate middle construction
as we discussed in Section 6.3.2, the Scope cannot. In other words, Scope
constructions do not have medio-passive counterparts. But, like Goal, Scope
may not be realised in the clause provided it can be inferred from context (But
see Section 6.3.3 below on abstract material clauses). In terms of its
realisation, the Scope is also realised by a nominal group. The following
clauses illustrate different manifestations of the Scope participant (Scope is in
bold):
(15) Report 5
Enti bom za [[ɩ na bʋr ƴãw a ɩ
so thing all 1SG REL sow.PFV put DEF 1SG
wɩɛ pʋɔ a dɩya a]], ɩ nyɛ nɩ

farm inside DEF last year JUNC 1SG get.PFV FOC
bom za yi a pʋɔ.
thing all be:from.PFV DEF inside
‗So everything that I sowed in my farm, I got everything from it.‘
Ʋ dɛ bɛ pawr a mɩ̀ ɛ.
3SG ADV NEG.IND.NFUT get.IPFV DEF ADV NAFFR
‗He is just not getting some of it.‘

[…] ɛ cãa sow a tɩ dʋn mɩ̀.
[…] CONJ still help.PFV DEF 1PL animals too.
‗and still help our animals too‘
A kʋɔbɛ bɛŋa za law nɩ
DEF farmers DEM all be:together FOC
taar a ka.
one another DEF here
‗All these farmers have met among themselves here.‘
271
|||Fʋʋ bɛ wa lɩɛbɩ pawr
2SG.EMP NEG.IND.NFUT COND change.PFV catch up with.PFV
a tew a || a tew na zɔ
DEF world JUNC DEF world POS.IND.FUT run.PFV
na || kɛ́ bɛr fʋ. |||
AFFR CONJ leave.PFV 2SG
‗If you don‘t change to catch up with the world, the world will run and
leave you.‘
Bɛ tuori tɩ a be.
3PL.HM meet.PFV 1PL DEF there
‗They should meet us there.‘
Bɛ tu =n sɔr.
3PL.HM go on.PFV FOC journey
‗They‘ve gone on a journey.‘
Mɛ́ nɩ laa tu a puor.
HST 2PL ADV follow.PFV DEF back
‗I believe you followed up on it again.‘
As these examples show, the Scope participant indicates the range of the
process rather than being impacted by it. The Actor is rather the participant
directly affected by the process. These clauses are, thus, intransitive because
the process does not go through.
(ii) Recipient and Client: Recipient and Client are two related participant
roles associated with the clause as a realisation of the transfer of goods-&-
services. As the names suggest, Recipient is the one who benefits from the
transfer of goods while Client is the one who benefits from a service carried
out by the Actor. From a view below the clause, both participants are realised
by a nominal group. One distinctive feature of these kinds of material clauses
is the predictive pattern of verbal groups associated with them. In Recipient
272
clauses, the process is normally realised by the verb kʋ̀ (‗give‘) or its negative
equivalent sib (‗deny‘, ‗refuse‘) as is illustrated in (23) to (25) (Recipient is in
bold):
(23) NANSU citation

A Naaŋmɩn kpɛw-za–sob [[na kʋ̀ fʋ a
DEF God power-all-owner REL give 2SG DEF
to-nʋɔr ŋa a]] na kʋ̀ fʋ nɩ

duty DEM JUNC POS.IND.NFUT give.PFV 2SG FOC
ya, faw anɩ kanyir.

wisdom strength and patience
‗God almighty who has given you this task will give you wisdom,
strength and patience.‘
Bɛlɛ lɛ bɛ kʋ̀ tɩ a dɩstɩrɩkt.
3PL.HM.EMP COP 3PL.HM give.PFV 1PL DEF district
‗They are the ones (who) gave us the district.‘
(25) A nàa sib =ɩ Dɛr a libir.
DEF chief deny.PFV FOC Der DEF money.
‗The chief denied Der the money.‘
In these constructions, the Recipient participant follows the Process and can be
followed by a Goal participant, which construes the entity that is transferred.
Here either the Recipient or the Goal can be the culmination of New
information. In (23), the locus of New information falls on the Goal
participant ya, faw anɩ kanyir (‗wisdom, strength and patience‘) while the
Recipient, realised by the pronoun fʋ, is construed as given information. When
the Recipient is realised by a noun or an emphatic pronoun, it carries the locus
of information focus (25).
Alternatively, the recipiency may be construed analytically with a
verbal group complex (‗serial verb construction‘) typically with the pattern:
de-verbal group, ‗take‘, (+Goal) + kʋ̀-verbal group, ‗give‘, + Recipient:
273
(26) Bɛ de nɩ dãa kʋ̀ a
3PL.HM take.PFV FOC pito/beer give.PFV DEF
dɔɔ.
man
‗They gave pito/beer to the man.‘
The motivation for this alternative analytical construction is to give relative

thematic prominence to the Goal participant over the Recipient.
Client-oriented clauses, on the other hand, are open to a wide range of
verbs. Two structural patterns can be identified for client-oriented
constructions as follows:
(i) verbal group denoting service (+Goal) + verbal group denoting

benefit + Client
(ii) de-verbal group (+Goal) + verbal group denoting service (+kʋ̀-
verbal group) + Client
We illustrate the first structure in (i) and the second structure in (ii) and (iii) in
Figure 6.1.
(i) Casual conversation

Hɛɛ, pʋr bʋla ƴãw a Zan!
hey pour.PFV small.SG putin.PFV DEF Zan
Goal Client
Process
verbal group nominal group verbal group nominal group
‗Hey, pour a little for Zan.‘
(ii) Casual conversation

Tɩ de bìn Gɔbr ɩ?
1PL take.PFV put.PFV Gobr JUNC
Actor Process Client

nominal group verbal group verbal group nominal group
‗We should keep (it) for Gobr?‘
(ii) Constructed example
274
Ʋ de nɩ a Bie gaalɩ kʋ̀ mɛ.
3SG take.PFV FOC DEF child lay.PFV give.PFV 1SG.ACC
Actor Goal Client

Process
nominal verbal nominal verbal verbal nominal
gp. gp. gp. gp. gp. gp
‗He has put the child to bed for me.‘
Figure 6.1. Illustration of the structure of clauses of cliency
In (i), the Process is realised by a verbal group complex consisting of one verb,
specifying the nature of the service, pʋr (‗pour‘), and another verb,
introducing the Client and denoting benefit, ƴãw (‗put‘). In (ii), the Goal
participant is implied and therefore not realised. As in (i), the Process is
realised by a verbal group complex, consisting of de (‗take‘) and bìn (‗put‘).
Example (iii), on the other hand, illustrates a complete structure, in which the
Goal is realised and is introduced by the verb de (‗take‘) while the verb kʋ̀
introduces the Client. The locus of the Process rest on the verbal group
realised by gaalɩ (‗lay‘). Generally, the verbal group that denotes the service
and which normally precedes the Goal immediately can be realised by any
verb of the material process while those that introduce the Client and denote
benefit are limited, the most frequent being: kʋ̀ (‗give‘), bìn (‗put‘), and ƴãw
(‗put in‘). Others include dɔwlɩ (‗place on‘), and tur (‗pass to‘).
As example (ii) shows, in many client-oriented clauses, the verb kʋ̀
(‗give‘), which is the most common benefit verb has lost its original meaning
of transfer of goods and has come to be used generally to mean benefit. Unlike
in other types of material clauses, the Goal and Recipient or Client cannot be
made Subjects in an agnate middle clause. In other words, benefactive clauses
do not have medio-passive counterparts.
In clauses of cliency, unmarked information focus is always placed on
the Goal participant, which normally precedes the Client, as in (i) in Figure
6.1. Also, as we saw in examples (23) and (24) for recipiency, cliency can be
construed in various degrees of abstractness (Client is in bold):
275
Tɩ na maalɩ nɔ-ɓaan yéle kʋ̀
1PL POS.IND.FUT do.PFV wonderful things give.PFV
taar nɩ pʋ-pɩɛlʋ.
one another COM stomach-whiteness
‗We will do wonderful things for one another with joy.‘

Bʋʋ nʋ fʋ paa ah tʋɔ voti
what IDENT 2SG ADV INTJ be:able.PFV vote.PFV
kʋ̀ Mahama[…] ɛ bɛ voti
give.PFV Mahama […] CONJ NEG.IND.NFUT vote.PFV
kʋ̀ a Kuubuor?
give.PFV DEF Kuubuor
‗What is it (that) you can vote for Mahama … and not vote for
Kuubuor?‘
Tɩ ar kʋ̀ a ʋl.
1PL stand.PFV give.PFV DEF DEM
‗We should stand for that one (=We should support him).‘
As the examples show, the verb kʋ̀ does not indicate a transfer of possession
but rather construes benefit.
(iii) Instrument, Accompaniment and Initiator: These are another related

set of participant roles. Defined from the point of view of semantics,
Instrument is the participant which serves as a means by which the Actor
brings the Process into actualisation and Accompaniment is a participant that
is construed as accompanying the Actor in bringing about the unfolding of the
process. Thus, both participants are construed as been brought into the
participation of the Process by the action of the Actor. The implication is that,
the Actor in such clauses is often an entity of a higher agency than the
Instrument and the Accompaniment. For instance, in clauses involving an
Instrument, the Instrument participant is always inanimate, while the Actor is
normally an animate entity (30). Similarly, in a material clause with an
276
Accompaniment, the Actor is normally realised by an entity of a higher
agency such as animate as opposed inanimate, human as opposed to non-
human, and adult as opposed to young (31). The Instrument participant and
Accompaniment participant in (30) and (31) respectively are in bold:
(30) Ʋ kʋɔr =ɩ nɩ a kuur.

3SG weed.IPFV INST FOC DEF hoe
‗S/he is weeding with the hoe.‘
(31) Ʋ kʋɔr =ɩ nɩ a bibiir.
3SG weed.IPFV COM FOC DEF children
‗S/he is weeding with the children.‘
The Initiator participant is the direct opposite of Instrument and

Accompaniment in terms of their relationship with the Actor. It is a participant
which is external to the Actor + Process configuration and which causes the
Actor to bring about the Process. Thus, it is experientially of a higher agency
than the Actor and the entity that it represents has a capacity to control the
Actor. For instance, in (32) below, the Complement of the clause, a bàa (‗the
dog‘), is construed as the Actor engaged in the Process while the Subject, Ʋ
(‗S/he‘) is the Initiator, the one that causes the dog to come:
(32) Ʋ wa nɩ nɩ a bàa.
3SG come.PFV CAUS FOC DEF dog
‗S/he has come with the dog (=has caused the dog to come).‘
As examples (30) to (32) show, the Initiator, Instrument and Accompaniment

participants occur in clauses with a very similar syntactic organisation.
However, while Initiator is normally the Subject in the interpersonal structure
of the clause and in which case the Actor assumes the Complement position
(32), Instrument and Accompaniment are normally Complement, as in (30)
and (31). With regards to their realisation from below the clause, the three
participant roles are morphologically signalled by the particle nɩ (or its enclitic
forms n= and =ɩ) indicated in the verbal group and following the verb. This
particle has been glossed variously as instrument (30), comitative (31) and
causative (32) respectively for clauses with Instrument, Accompaniment and
277
Initiator. The term ‗agentive marker‘ will be used as a general term for this
particle when we are referring to its various uses collectively. 42 When the flow
of discourse can enable addressees to recover the participant in Complement
position of the clause, this participant is left unrealised in which case only the
agentive marker encodes the sense of causation, instrumentality or
accompaniment in the clause. This is illustrated in the dialogue below
(relevant clause is in bold):

Baba: Ɛcɛ Zɩɛm ɩ, a dãa [[ɩ na da
fʋ a, a wa]]? Ɛ bɛ dɛ zɩ.
2SG JUNC 3PL.NHM be:where And 3PL.HM ADV sit.PFV
kʋ̀
give.PFV
Zɩɛm: Aa
INTJ

no 3PL come.IPFV CAUS AFFR
Ziem: ‗I don‘t know.‘
The use of the enclitic causative (or agentive) particle =ɩ within the verbal
group in the bɛ waar =ɩ na (‗they are bringing (it)‘) makes the Subject bɛ
(‗they‘) an Initiator rather than the Actor of the clause. The Actor is therefore
left implicit to be recovered from the preceding discourse.
The similarity in the syntactic organisation of clauses involving
Initiator, Instrument and Accompaniment participants also means that there is
a tendency for ambiguity in analysing these clauses. Such an ambiguity can
however, be partially resolved by assessing the context in which the clause
42
The common label is motivated by the close functional relationship between the three
different senses of nɩ. From a grammaticalisation perspective, we can posit a functional split
along the following development: FOCUS > COMITATIVE > INSTRUTMENT/CAUSATIVE.
See Heine & Kuteva (2002a: 79-88) for a similar pathway across languages.
278
occurs in the following way. Initiator oriented material clauses are effective
clauses while Instrument and Accompaniment oriented ones are middle
clauses. This means that it is only in Initiator oriented clauses that the
Complement can genuinely be reversed to become the Subject in an agnate
middle clause. Therefore, in the clauses in Figure 6.2, only the (a) clauses are
effective and the (b) clauses are their corresponding middle counterparts.
(1a) effective: causative (1b) middle

Ʋ kʋɔr =ɩ nɩ a bibiir A bibiir kʋɔr a
3SG weed.IPFV CAUS FOC DEF children DEF AFFR
children weed.IPFV
Initiator Process Actor Actor Process
noun group verbal group noun group noun group verbal group
‘He is making the children weed.’ ‘The children are weeding.’
(1c) middle; comitative

Ʋ kʋɔr =ɩ nɩ a bibiir
3SG weed.IPFV COM FOC DEF children
Actor Process Accompaniment
noun group verbal group noun group
‘He is weeding with the children.’
(2a) effective: causative (2b) middle

Ʋ kʋɔr =ɩ nɩ a kuur A kuur kʋɔr a
3SG weed.IPFV CAUS FOC DEF hoe DEF hoe weed.IPFV AFFR
Initiator Process Actor Actor Process

noun group verbal group noun group noun group verbal group
‘He is making the hoe weed.’ ‘The hoe is weeding.’
(2c) middle; instrumental

Ʋ kʋɔr =ɩ nɩ a kuur
3SG weed.IPFV INST FOC DEF hoe
Actor Process Instrument
‘He is weeding with the hoe.’
Figure 6.2. Illustration of the different meanings of the agentive particle
It should be noted that, both the Complement in the effective clause and the
Subject in the agnate middle clause has the same participant role, which is
Actor.
279
Another distinctive characteristic of the Initiator-oriented clause
among agentive clauses is that the Process can be construed analytically with a
verbal group complex where a general verb such as ƴãw (‗let‘), vɛ (‗let‘) and ɩ
(‗do‘, ‗make‘) construes the causation instead of the agentive particle.
Examples are given below (Initiator participant is in bold):
(34) Ʋ ƴãw na a bibiir kɔ a

3SG let.PFV AFFR DEF children weed.IPFV DEF wɩɛ.
farm.
‗He let the children weed the farm.‘
(35) Naaŋmɩn ɩ saa wa!
God do.PFV rain come.PFV
‗God should make it rain.‘
In each of these causative clauses, the verbal group complex realises one
Process and not two with ƴãw and ɩ simply encoding causation, while the
more specific verbs, kʋɔr (34) and wa (35) encodes the nature of the unfolding
Process.
Finally, with regards to accompaniment clauses, the Accompaniment
can represent a process as opposed to a participant. Let‘s compare (36) and the
reconstructed version in (37) below:

Nyɩnɛ na fʋ dɛ ŋmʋrɛ a lɛ
where IDENT.PL 2SG ADV rush.IPFV DEF DEM.DIST
cere nɩ?
go.IPFV COM
‗Where are you just rushing like that going with?

(37) Nyɩnɛ na fʋ cere nɩ a ŋmʋrʋ nɛ
where IDENT.PL 2SG go.IPFV COM DEF rushing DEM
bʋʋrɛ.
type
‗Where are you going with that kind of rushing?‘
280
In (36), the comitative particle nɩ signal that the first process, ŋmʋrɛ
(‗rushing‘), in the verbal group complex ŋmʋrɛ ... cere nɩ (‗rushing … going
with‘), accompanies the movement construed by cere (‗going‘). Example (37)
is an alternative construction. Here, the process is nominalised as ŋmʋrʋ
(‗rushing‘) and represented as a participant in the clause.
6.3.3 Concrete and Abstract Material Clauses

To recapitulate, the preceding sections discussed the different types of
material clauses, based on the nature of the unfolding of the process and the
participants associated with it. For each type, some grammatical critieria has
been identified as a signal of the kind of material clause construed and the
participant roles engendered. For instance, while a Goal participant can be
made a Subject in middle clause a Scope cannot. Also, Initiator is
distinguished from Instrument and Accompaniment based on the agency
characteristics of the entities these participant roles represent and by the fact
that while the Complement in an Intiator-oriented clause (i.e. the Actor) can be
made a Subject in a corresponding middle clause, Instrument and
Accompaniment do not allow this alternation. We also identified material
clauses as prototypically representing actions and happings. However, material
clauses are not always about concrete activities. Rather, they are sometimes a
metaphorical representation of abstract notions. This section will briefly
identify the characteristics of these abstract clauses.
The distinction between concrete and abstract material clauses in
Dagaare is relevant because abstract clauses provide special exceptions to
some of the characteristics of material clauses we have already discussed.
First, the more abstract a clause is the less likely it is that participants in
Complement position of the clause (typically, Goal and Scope, but also,
Instrument, Accompaniment and the Actor of a causative clause) would be
unrealised even if the textual context construct them as given information (cf.
Section 6.3.1 & 6.3.2). Second, the more abstract or metaphorical a material
clause is, the less likely it is that it can be converted into a corresponding
middle clause. In other words, the Goal participant in abstract clauses mostly
281
cannot become the Subject in a middle clause. Let‘s consider the following
examples:

[…] a bɛ wa =n a bɛ tɩɛrʋ,
[…] CONT 3PL.HM come.PFV CAUS DEF 3PL.HM thoughts
a tɩ mɩ̀ wa =n a tɩ tɩɛrʋ […]
CONT 1PL too come.PFV CAUS DEF 1PL thoughts
‗…then they will bring their ideas, then we too will bring out ideas …‘
Tɩ tɩɛ taar!
1PL push.PFV each other
‗We should push one another (=We should support one another).‘
In (38), the two clauses are causative clauses in which the Initiator is bɛ in the
first clause and tɩ in the second clause. The Actor is realised by abstract
nominal groups, bɛ tɩɛrʋ (‗their thoughts‘) and tɩ tɩɛrʋ (‗our thoughts‘), which
is a nominalisation of a mental process tɩɛrɛ (‗thinking‘). The clauses are
therefore metaphorical representations of a mental process. Example (39)
represents an abstract behaviour, providing social support. In this metaphorical
sense, the nominal group taar (‗each other‘, ‗one another‘) functions as a
Scope participant rather than a Goal participant since the Process does not
actually impact on the secondary participant. In this sense the clause is less
concrete than one in which the Goal is realised by a concrete noun such as Tɩ
tɩɛ a bie! (‗We should the child‘). Thus while we can convert this concrete
clause into a middle clause as in A bie tɩɛ na (‗The child has been pushed‘);
taar (‗each other‘, ‗one another‘) cannot become the Subject in a middle
clause.
The fact that abstract clauses provide exceptions to the characteristics
of material clauses indicates that they are on the border line between material
and immaterial processes such as sensing and being. These other processes are
the focus of the sections below, starting with mental clauses.
282
6.4 Mental Clauses: Processes of Sensing
Mental clauses construe processes of sensing such as thinking, wanting,
feeling, and perceiving. These are processes that construe our consciousness
and inner experiences. Mental clauses are therefore complementary to material
clauses, which construe our outer experience. Let‘s start by considering an
extract which contains a fair number of mental clauses (mental processes are
in bold):
(40) St. Maria

Father: Kʋɔ ɔwfʋ na ʋ cen tɩ nyɛ̃
water fetching IDENT.PL 3SG go.PFV DIST see.PFV
dɛb o! Ɛ wɩɛ pʋɔ ʋ tɩ nyɛ̃
man PRT CONJ farm inside 3SG DIST see.PFV
pobile, a bɔbr.
young man CONT court.IPFV
Daughter: Ɩ bɛ bɔbr ɛ.
1SG NEG.IND.NFUT court.IPFV NAFFR
[…] […] […]

Mother: […] bɛ buor sob =ʋ fʋ
[…] 3PL.HM which one IDENT.SG 2SG
pãa bɔbr?
ADV want.IPFV
Daughter: Ɩ bɔbr a sob nɛ [[na
1SG want.IPFV DEF one DEM REL
tɛr faw gaw ʋ

possess.PFV strength be:more:than.PFV 3SG
tɔ a]].
fellow JUNC
Father: […] Fʋʋ [[na bãw bɛ a]],

[…] 2SG.EMP REL know.PFV 3PL.HM JUNC
ãa nʋ kpɛmɛ …?
who IDENT.SG strong
283
Daughter: Ɩ bɛ bãw ɛ. Sɛ
1SG NEG.IND.NFUT know.PFV NAFFR unless
3PL.HM mɔ́.
bɛ wrestle.PFV
Father: Fʋ wõ =n a lɛ [[ʋ na
2SG hear.PFV FOC DEF DEM 3SG REL
yèl a]]?
say. PFV JUNC
Mother: Ɩ wõ a na.
1SG hear.PFV 3PL.NHM AFFR
Father: ‗It was water that she was going to fetch, then she saw a man;
and, in farm, she saw a gentleman and is courting (him).‘
Daughter: ‗I‘m not courting (him).‘
Father: ‗Which of then do you want now?‘
Daughter: ‗I want the one [[who is stronger than his fellow]].‘
Father: ‗You who know them, who is strong?‘
Daughter: ‗I don’t know. Unless they wrestle.‘
Father: ‗Have you heard that [[which she said]]?‘
Mother: ‗I heard it.‘
Let‘s examine how the text is construed in terms of the flow of events. The
extract starts with the father recounting processes his daughter has been
engaged in, beginning with a material process of motion (cen, ‗go‘), followed
by a mental state, specifically, perception (nyɛ, ‗see‘) and then wanting (bɔbr,
‗wanting‘). The text is a good example of how different configurations of
events conspire to construe the flow of experience through time. Regarding
mental clauses, the extract flows from the perception of a phenomenon,
through desiring it to cognising it. The mental clauses serve to construct these
inert experiences into meanings which can be negotiated and challenged in
fellowship with others.
One characteristic of mental clauses is the presence of a conscious
participant who does the sensing. This participant is called Senser, typically
realised by a nominal group. The range of entities that can be Senser are
limited to those that are conscious (i.e. human and other animate beings), or
284
that are endowed with consciousness by speakers. In the extract above, the
Senser is realised by the pronouns Ɩ (‗I‘), fʋ/fʋʋ (‗you‘) and ʋ (‗s/he‘). The
participant that is sensed is called the Phenomenon, and unlike Senser, it has
no limitation on the range of entities that realise it. It can be realised by a
nominal group or a rankshifted clause. In the grammar of Dagaare, therefore,
the Phenomenon participant has a wider scope of realisation compared to other
participant roles that correspond to the Complement since the realisation of
these other participants is restricted to group rank. In (40), nominal groups
realising Phenomenon are: dɛb (‗man‘), pobile (‗gentleman‘), bɛ buor sob
(‗which one of them‘), bɛ (‗third plural, human‘) and a (‗third plural, non-
human‘), and complex ones such as a sob nɛ na tɛr faw gaw ʋ tɔ a (‗the one
who is stronger than his/her fellow‘) and a lɛ ʋ na yèl a (‗that which s/he
said‘). A rankshifted clause realising Phenomenon is given below in (41) and
(42):
(41) Political interview

Ɩ dãw na cere tɔlɛ wone nɩbɛ
1SG ADV AFFR go.IPFV pass.IPFV hear.IPFV people
yèle […].
say.IPFV
‗I was passing by earlier on and heard people saying …‘
(42) St. Maria play
Gentleman 1: Fʋ bɛ nyɛ [[ɩ na tɛr
2SG NEG.IND.NFUT see.PFV 1SG NMLZ possess.PFV
taya a]]?
hand catapult JUNC
Gentleman 2: Hɛh! Fʋ nyɛ nɩ kuur ɩ?

INTJ 2SG see.PFV FOC hoe JUNC
[…] […] […]

Gentleman 1: […] Fʋ nyɛ na [[fʋ na duw
[…] 2SG see.PFV AFFR 2SG NMLZ chase.PFV
bɛr a]]?
leave.PFV JUNC
1: Haven‘t you seen that I have a hand catapult?‘
285
2: ‗Hey! have you seen a hoe?
1: ‗… Have you seen that you drove (her) away?‘
The size of the unit of realisation of the Phenomenon corresponds to different

kinds of phenomena that are sensed. Nominal groups realise phenomena that
can be described as ‗micro-thing‘ (including abstract things) such as those
realising the Phenomenon participants in (40). Phenomena that are realised by
a rankshifted clause, on the other hand, are higher-order phenomena,
consisting of ‗acts‘ and ‗facts‘. An act-clause objectifies an event as a ‗macro-
thing‘ or, more technically, a macrophenomenon (Halliday & Matthiessen
2014). An example of an act-clause is nɩbɛ yèle (‗people saying‘) in (41). A
fact-clause is a ‗meta-thing‘ or a metaphenomenon since it is construed as
belonging to the more abstract semiotic realm of experience as opposed to the
material realm. It is a proposition presented as existing by itself as a semiotic
phenomenon, ―without being brought into being by somebody uttering it‖
(Halliday & Matthiessen 2014: 253). Examples of metaphenomenal clauses
are underlined in (42). Macrophenomenal and metaphomenonal clauses are
commonly known in the typological literature as ‗complement clauses‘, a term
which highlights their function in the interpersonal structure of the clause (see
Chapter 5). 43 Following Halliday and Matthiessen (2014), I will use the
complementary term hyperphenomenal clauses to signify their ideational
meaning as higher-order phenomena.
In Dagaare, as in many other languages including English (see e.g.
Halliday & Matthiessen 2014; Teruya 2007; Lavid et al. 2010),
hyperphenomenal clauses are restricted to one sub-type of mental clauses,
those of perception (types of mental clauses are discussed in Section 6.4.2.2).
We do encounter fact clauses in emotive clauses but, unlike in perceptive
clauses, they function as a circumstance of Cause rather than as a participant
in the clause. I illustrate this in (43) and (44) below (circumstance of Cause is
in underlined):
43
The term ‗complement clause‘ has been used broadly in the typology and descriptive
literature to include projected clauses (which include but is not limited to reported and quoted
clauses). I depart from this tradition by restricting ‗complement clauses‘ to embedded clauses
that truly function as Complements in the structure of the clause (see Section 6.4.2.2) on
projection in mental clauses for the difference between complement (or hyperphenomenal)
clauses and projected clauses in Dagaare).
286
(43) Kpɩɛrʋ kpɛ mɛ na [[fʋ na zɔ a]].
pain enter.PFV 1SG AFFR 2SG NMLZ run.PFV JUNC
‗It pained me that you ran away.‘
(44) A kpɛ =m =a [[fʋ na zɔ a]].
3PL.NHM enter.PFV 1SG.ACC AFFR 2SG NMLZ run.PFV JUNC
‗It pained me that you ran away.‘
In (43), the Phenomenon, Kpɩɛrʋ (‗pain‘), conflates with the Subject of the
clause and the Senser conflates with the Complement mɛ (‗me‘) while a fact-
clause realises circumstance of Matter, that is what the pain is about. Example
(44) has a similar syntactic organisation with (43), but there is no Phenomenon
in the clause. Rather, the Subject position is filled by a dummy pronoun, a
(third plural, non-human), which has no function in the experiential grammar.
Its function lies in the interpersonal metafunction, where it combines with the
Predicator kpɛ (‗enter‘) to ground the clause as an arguable unit of discourse
(see Chapter 5). The fact-clause fʋ na zɔ a (‗that you ran away‘) also realises
circumstance of Cause in this clause. Mental clauses like (43) and (44) will be
discussed in Section 6.4.3 on types of mental clauses.
Before we proceed, let‘s consider one general characteristic of mental
clauses. They do not often occur in the imperative. Thus, when Jesus
commanded Bartolomeo in the following exchange to see, the imperative
mental clause comes out as highly marked and functions to highlight the
divinity of Jesus:

Yezu: Bʋʋ nʋ fʋ bɔbr kɛ ɩ
what IDENT.SG 2SG want.IPFV PROJ 1SG
ɩ kʋ =b?
do.PFV give.PFV 2SG.ACC
Batolemi: Ɩ bɔbr k'ɩɩ lɛ nyɛrɛ.

1SG want.IPFV PROJ.1SG ADV see.IPFV
Yezu: Nyɛrɛ! A fʋ Naaŋmɩn sawfʋ sanɩ
see.IPFV DEF 2SG God faith heal.PFV
fʋ na.
2SG AFFR
287
Batolemi: Ɩ nyɛrɛ na! Ɩ nyɛrɛ
1SG see.IPFV AFFR 1SG see.IPFV
na! Ɩ nyɛrɛ na!
AFFR 1SG see.IPFV AFFR
Jesus: ‗What is it that you want me to do for you?‘
Bartolomeo: ‗I want to see again.‘
Jesus: ‗See! Your faith in God has healed you.‘
Bartolomeo: ‗I see! I see! I see!‘
6.4.1 Direction of Sensing: Emanating and Impinging

Mental clauses construe sensing in two ways. One way is to construe it as
something emanating from the Senser and the other way is to construe it as an
outward phenomenon impinging on the Senser (cf. Halliday & Matthiessen
2014). Much of our discussion in the preceding section has focused on the
emanating type (but see 43 & 44). The impinging type clearly shows that the
Senser is like the Goal participant of a material clause in the sense that it is on
the Senser that the Process impacts upon. In Dagaare grammar, the impinging
type is a typical characteristic of emotive clauses such as those construing pain
and joy. The following extract is an instance from a concert advertisement:
(46) concert advertisement

[…] ɛ nʋ̀ɔ kpɛ nɩbɛ nɛ za [[na
[…] CONJ joy enter.PFV people DEM all REL
na wa a be || wa nyɛ a
POS.IND.FUT come.PFV DEF there PROX see.PFV DEF
bom nɛ tɩ [[na ɩrɛ a]]]].
thing DEM 1PL REL do.IPFV JUNC
‗… and all those people who will come there to see what we are doing
will feel joy.‘
Here, the Phenomenon, nʋ̀ɔ (‗joy‘), is construed as an entity that impinges the
Senser, which is realised by a nominalised clause complex following the
Process, kpɛ (‗enter‘). Verbs that typically realise the Process in this type of
mental clauses are given in Table 6.1 (note that the table does not give an
288
exhaustive list of these verbs). The Phenomenon is normally the Subject in
these clauses while the Senser is the Complement. The exception is that, when
the Process is realised by 'wɔb/'wɔbr (‗pain/paining‘), the Subject is realised
by a body-part noun which could either be represented as Senser or
Phenomenon (see example 4 in Table 6.1)
Table 6.1 Verbs that typically realise the Process in impinging mental
clauses
No. perfective imperfective example
1 kpɛ, ‗enter‘ kpɩɛr, Nʋ̀ɔ kpɛ mɛ na.
‗entering‘ ‗I feel joy (=I feel happy).‘
2 kʋ́, ‗kill‘ kʋ́rɛ, ‗killing‘ Vɩ kʋ́rɛ mɛ na.
‗I feel shy.‘
3 ɩ, ‗do‘ ɩrɛ, ‗doing‘ Nʋ̀ɔ ɩrɛ mɛ na.
‗I feel happy.‘
4 'wɔb; 'wɔbr; [Subject/Senser]:
‗pain‘, ‗paining‘, Ɩ zu 'wɔbr =a.
‗hurt‘ ‗hurting‘ ‗My head is paining (=My head
hurts).‘
[Subject/Phenomenon]:
Ɩ zu 'wɔbr mɛ na.
‗My head pains me.‘
5 fɛr, ‗worry‘ fɛrɛ, ‗worrying‘ Kʋɔ yéle caa na fɛrɛ tɩ.
‗(The) water issue is still worrying us.‘
6 tɔ tʋ́ɔr A ƴɛrʋ tɔ mɛ na.

‗The speech touched me.‘
Another characteristic of this type of mental clauses is that an additional

participant is sometimes introduced as an external causer of the Process, the
one who brings about the sensing. This participant is technically called the
Inducer and it is similar to the Initiator role in material clauses (cf. Halliday &
Matthiessen 2014). Clauses in which the Inducer occur are causative mental
clauses and we give examples in Figure 6.3.
a) subjective orientation
A fʋlaarʋ kʋ́rɛ mɛ nɩ suur.
DEF 1SG
laughter kill.IPFV 1SG.ACC FOC anger
Inducer Process Senser Phenomenon
noun group verbal group noun group noun group
‗Your laughter makes me angry.‘
289
A bʋndɩrɩ ɩrɛ mɛ nɩ vʋla.
DEF food do.IPFV 1SG.ACC FOC longing
‗The food makes me long for it (=I long for the food).‘
A dãa 'wɔbr mɛ nɩ zu.
DEF beer/pito pain.IPFV 1SG.ACC FOC head
‗The beer/pito gives me pain in the head.‘
b) objective orientation
A dãa 'wɔbr =ɩ zu.
DEF beer/pito pain.IPFV FOC head
Inducer Process Senser
‗The beer/pito makes my head painful.‘

Figure 6.3: Illustration of causative mental clauses
As Figure 6.3 shows, causative mental clauses are either construed from a
subjective orientation, where the speaker presents the information as his/her
subjective feeling or interpretation, or an objective orientation, where the
speaker presents the information as a fact, which may or may not encode
collective sensing. The pressure to represent sensing objectively sometimes
leads to a situation where it is totally phenomenalised in the sense that there is
no Senser participant in the clause. Table 6.2 compares clauses in which
sensing is phenominalised with those in which it is encoded subjectively.
Clauses that phenomenonalised sensing appear to be similar to
meteorological material clauses such as:
(47) Sa-bie ɩrɛ na.

rain-driblets do.IPFV AFFR
‗It is drizzling.‘
While these meteorological clauses construe observable phenomena in the

outer world, the mental processes are metaphorical and construe inert
sensation. They corroborate a common motif in language where concrete
objects and activities are pressed to construe abstract meanings.
290
Table 6.2. Phenomenonalisation of sensing versus subjective sensing
subjective/intersubjective objective
Nʋ̀ɔ ɩrɛ/kpɩɛr ʋ na. Nʋ̀ɔ ɩrɛ na.
joy do/enter.IPFV 3SG AFFR joy do.IPFV AFFR
‗S/he feels happy.‘ ‗There is joy.‘

Ƴawr ɩrɛ/kpɩɛr mɛ na. Ƴawr ɩrɛ na.
cold do/enter.IPFV 1SG.ACC AFFR cold do.IPFV AFFR
‗I feel cold.‘ ‗It is cold (=The weather is

cold).‘
Kɔ kpɩɛr tɩ na. Kɔ ɩrɛ na.
hunger enter.IPFV 1PL AFFR hunger do.IPFV AFFR
‗We feel hungry.‘ ‗There is hunger.‘

Phenomenon Process Senser Phenomenon Process
noun group verbal gp. noun noun group verbal
gp. gp.
6.4.2 Projection in Mental Clauses

Another major characteristic of mental clauses is their ability to project
another clause. That is, a mental clause may set up another clause (or a
combination of clauses) as a projection of the content of consciousness. We
have seen instances of this in the preceding sections, as in example (45). Other
instances of projection are provided below:

Fʋ bɔbr a kɛ tɩ ɩ ŋmɩn?
2SG want.PFV AFFR PROJ 1PL do.PFV what
‗You want that we do what? (=What do you want us to do?)‘
Tɩ bãw kɛ a yél ŋa a, ʋ
1PL know.PFV PROJ DEF matter DEM JUNC 3SG
na tɔl =ɩ ni-daa.
POS.IND.FUT pass.PFV FOC forward
291
‗We know that, this matter, it will go far (= … that this programme
will be successful).‘
(50) Workshop Report 1

Ɩ tɩɛ a be bome na bɛ
IPL think DEF there things IDENT.PL NEG.IND.NFUT
maal vla kʋ̀ tɩ ɛ.

do:well. PFV good give.PFV 1PL NAFFR
‗I think the things (=crops) there are the ones that did not do well for
us.‘
(51) Sɛb-Sõw ƴɛr-bie (1996)
Ʋ tɩ la na kɛ́ tɩɛrɛ: ―A
3SG PST.REM laugh.PFV AFFR CONJ think.IPFV DEF
sãan bɛ maal bãw a ɩ yome

visitor NEG.IND.NFUT ADV know.PFV DEF 1SG years
nʋɔr ɛ."
number NAFFR
―S/he laughed and was thinking: ‗The visitor doesn‘t know my age
well.‘‖
In each of these examples, a mental clause projects another clause. The

projected clause is the content of consciousness, namely wanting as in (48)
and thinking as in (49) to (51). A projected clause is distinguished from
hyperphenomenal clauses based on the following characteristics.
Semantically, while hyperphenomenal clauses represent acts and facts, the
projected mental clauses are thoughts and desires. Mental projection is
therefore associated with clauses that represent cognition and desideration (cf.
Section 6.4.2.2). Second, a projected clause forms a clause nexus with the
mental clause and is normally introduced by the conjunctive particle kɛ. Thus,
one way of testing whether a clause is a projected or hyperphenomenal clause
is to insert the particle kɛ between the Process of the mental and the following
clause. For instance, the absence of the particle kɛ in (50) is an instance of
ellipsis and it can be inserted after the Process tɩɛ.
292
In addition, projection comes in two modes, reporting and quoting.
Projected reporting is exemplified by examples (48) to (50) while (51) is an
instance of a projected quote. Also, a projected clause also maintains its full
status as a negotiable and arguable unit of discourse by retaining all the
characteristics of its mood type. It is thus a free clause (cf. Chapter 3, Section
3.4.3). In (48), for example, the projected clause tɩ ɩ ŋmɩn? (‗what should we
do?‘) maintains its characteristic as an interrogative clause although the
projecting clause is a declarative. Similarly, the projected clauses in (49) to
(51) maintain their characteristics as declarative clauses that can be negotiated
and argued about. A hyperphenomenal clause (‗complement clause‘), on the
other hand, is an embedded clause, and its non-finite status is indicated by the
absence of mood markers or end focus, as has been shown in (41) and (42) (cf.
Section 6.4).
6.4.3 Principal Types of Mental Clauses

In the discussion in the preceding sections, the principal types of mental
clauses were alluded to. As Halliday and Matthiessen (2014) note for English,
four sub-types have been identified for Dagaare, comprising cognitive,
desiderative, emotive and perceptive clauses, although the realisation patterns
are not exactly the same as in English. In this section, we identify their
distinctive characteristics more systematically as a way of summarising our
account of mental clauses in general (see Table 6.4). Cognitive clauses
construe thinking; desiderative clauses construe wanting; emotive clauses
construe feeling and perceptive clauses encompass seeing, hearing and other
perceptual processes. 44
The grammar of Dagaare sets them apart as distinctive types of mental
clauses. Cognitive and desiderative clauses are set apart from the emotive and
perceptive clauses by their ability to project other clauses. Thus, even when
the verb nyɛ (‗see‘), which is originally associated with perception, projects, it
gives the clause a cognitive interpretation. In (52), for example, nyɛ (‗see‘)
represents inference as opposed to its perceptive sense in the reconstruction in
(53):
44
‗Thinking‘, ‗wanting‘, and ‗feeling‘ are used here as general labels (or prototypes) for other
processes within these domain of sensing.
293
A Naaŋmɩn tɩ nyɛ na k'aa al

DEF God PST.REM see.PFV AFFR PROJ.DEF 3PL.EMP.NHM
za [[ʋ na maalɩ a]] vɩɛl =ɩ

all 3SG REL make.PFV JUNC be:good FOC
yaga.
plenty
‗God saw (=realised) that all that he made was very good.‘
(53) A Naaŋmɩn tɩ nyɛ nɩ a al
DEF God PST.REM see.PFV FOC DEF 3PL.EMP.NHM
za [[ʋ na maalɩ a]], a vɩɛl

all 3SG REL make.PFV JUNC 3PL.NHM be:good
yaga.
plenty
‗God saw all that he made to be very good.‘
Examples (52) and (53) are cognitive and perceptive clauses respectively. In
(52), the clause k'aa al za ʋ na maalɩ avɩɛlɩ yaga (‗that all that he has made
was very God‘) is projected as the content of the consciousness of God. The
projection is indicated morphologically by the clitic form of the particle kɛ.
While the projected clause in (52) is free, as indicated by the presence of end
focus onthe adverbial group yaga (‗plenty‘), the metaphenomenal clause in
(53) is embedded and, therefore, has no obligatory (or end) focus.
Table 6.3. Contrast between cognitive and desiderative clauses on aspect

selection
mental perfective imperfective
clause type
cognitive Ɩ bãw a na. Ɩ bãwnɛ a na.
1SG know.PFV 3PL.NHM AFFR 1SG know.IPFV 3PL.NHM AFFR
‗I know it.‘ ‗I am knowing it.‘
desiderative Ɩ bɔbr a na.
- 1SG want.IPFV 3PL.NHM AFFR
‗I want it.‘
294
Between cognitive and desiderative clauses, on the other hand, they
show a difference in their selection of aspect. The unmarked choice of aspect
for cognitive is perfective while the only choice of aspect available for
desiderative clauses is the imperfective. This contrast is presented in Table
6.3.
Regarding perceptive clauses, their unique characteristic is that they
are the only sub-type in which the Phenomenon participant can be realised by
a hyperphenomenal clause, as in (53). However, although perceptive clauses
typically do not project, as has been mentioned earlier, there is a metaphorical
context in which the verb wõ (‗hear‘) projects to enact evidentially,
specifically hearsay. Compare, for instance, (54) and (55):
(54) Ɩ wõ =n gɔmɛ.

1SG hear.PFV FOC noise
‗I heard noise.‘
(55) Ɩ wõ na kɛ fʋ kul =ɩ
1SG hear.PFV AFFR PROJ 2SG marry.PFV FOC
sɩr.
husband
‗I hear that you got married.‘
This use of wõ (‗hear‘) in (55) is an incongruent realisation of a verbal

process. A more congruent realisation would be the verbal clause below,
where the Subject Bɛ (‗They‘) lacks specific reference and the projection
serves to enact the proposition as hearsay:
(56) Bɛ yèl kɛ fʋ kul =ɩ sɩr.

1SG say. PFV PROJ 2SG marry.PFV FOC husband
‗They (=people/someone) say you got married.‘
Strictly speaking, therefore the use of wõ (‗hear‘), as in (55), is not as much to
represent perception as to enact evidentiality.
Finally, on emotive clauses, they are pervasively construed in the
impinging mode of sensing compared to the other sub-types, which are
typically construed in the emanating mode. In perceptive clauses, in particular,
295
Table 6.4 Characteristics of the sub-types of mental clauses
Criteria Cognitive Desiderative Emotive Perceptive
phenome- phenome- Ʋ bãw ʋ Ʋ bɔbr a na Ʋ nɔnɛ ʋ na. Ʋ nyɛ ʋ na

nality nal na. S/he wants S/he likes S/he saw
S/he it. him/her. him/her.
knows
him/her.
macro- Ʋ nyɛ na bɛ
phenomenal waar.
S/he saw
them coming
metaphe- Ʋ nyɛ na a
nomenal vɩɛl.
S/he has
seen it is
good.
projection Ʋ bãw kɛ Ʋ bɔbr kɛ fʋ –
a vɩɛl a. wa. [see below
S/he S/he wants on
knows you to come. evidentiality]
that it is
good.
directionality emanating    
impinging Zʋ́mɛ kpɛ Vʋla kpɛ mɛ [pervasive] [restricted]
mɛ na. na. Nʋ̀ɔ kpɩɛr Ƴawr ɩrɛ mɛ
Intuition Longing mɛ na. na.
entered entered me. Joy is Cold entered
me. entering me. me.
metaphorical evidentia- Ɩ tɩɛ ʋ wa [hearsay]
modal lity na Ɩ wo na kɛ ʋ
assessment I think kpi na.
s/he has I hear that he
come. died.
desirability Ɩ na bɔbr kɛ ɩ
(cf. Chapter sowri fʋ.
4 on I will want
modality) to ask you.
unmarked perfective imperfective imperfective perfective
aspect
the impinging type is restricted to bodily perception such as feeling cold or

hot, which, in fact, lies on the borderline between perception and emotion
(‗feeling‘). The unmarked aspect for emotive clauses is the imperfective
aspect as opposed to cognitive and perceptive clauses. We summarise the
296
differences among the four sub-types of mental clauses in Table 6.4 (compare
with Halliday & Matthiessen (2014: 256) on English).
6.5 Relational Clauses: Processes of Being & Possessing
Relational clauses are clauses of ‗being‘. They characterise and identify

phenomena in the world. In this sense, they are like material clauses since they
both primarily represent the outer world of experience. While material clauses,
however, represent dynamic experience, relational clauses represent inert
experience, that is, relations among phenomena. Two types of relational
clauses can be identified, based on the mode of participation in the Process.
These sub-types are identifying and attributive clauses. The following clauses
from the opening of a radio interview with a Dagara chief show a good
contrast between identifying (57a) and attributive clauses (57b, c):
(57) Interview with Nandom nàa

(a) Mãa nɩ a Nandɔm nàa, Nàa Dr.
1SG.EMP COP.FOC DEF Nandom chief Chief Dr.
Charles Pʋɔ-ure Pʋɔbɛ-ciir ayɔpoi sob.
Charles Puo-urePuobe-chiir seven owner
‗I am the chief of Nandom, Nàa Dr. Charles Puo-ure Puobe-
chiir VII.‘
(b) Ɩ ɩ =n a paramount chief of
1SG COP.PFV FOC DEF
Nandom Traditional Area […]

(c) Ɩ lɛ ɩ =n a president of the
1SG ADV COP.PFV FOC DEF
Upper West Regional House of Chiefs.

‗I am the (paramount) chief of Nandom traditional Area … I
am also the president of the Upper West Regional House of
Chiefs‘
The chief‘s introduction of himself opens with the identifying clause in (57a)
and proceeds with a series of attributive clauses, two of which are given in
(57b) and (57c). In an identifying clause, one participant, the Identifier, serves
to identify another, the Identifier. In example (57a), the Identified participant
297
is the pronoun Mãa (‗I‘), which represents the speaker, while the Identifier is
the specification of the speaker, and it is construed at two levels: first, his role
identity (i.e. a Nandɔm nàa, ‗the chief of Nandom‘), which is then further
specified appositively by his name and titles (Nàa Dr. Charles Pʋɔ-ure Pʋɔbɛ-
ciir ayɔpoi sob). The relationship between the Identified and the Identifier in a
relational clause is therefore that of specification.
In the attributive clause, on the other hand, the relationship between
the participants is that of ascription of a class membership, where an Attribute
is ascribed to a Carrier. In (57b) and (57c) the first person pronoun Ɩ is the
Carrier, and the roles of the speaker are represented as Attributes. Thus, while
the speaker represents his identity as the traditional seat upon which his stool
name is based (57a), he construes his other roles as class membership, as one
of the paramounts chiefs in Ghana (57b) and one of the ten presidents of the
various Regional Houses of Chiefs in the country (57c). In other words, while
there can be only one chief of Nandom, there are many other paramount chiefs
and presidents of Regional Houses of Chiefs in Ghana. It must be noted,
however, that the two modes, identifying and attribution, are generally not
given by the state of the world, as it were. Rather, they are selections made by
speakers as a reflection of their perspective on experience. The two
participants in both attributive and identifying clauses are realized by nominal
groups. Attributive clauses are also associated with two more participants, the
Attributor, the external participant that ascribes the Attribute to the Carrier,
and Beneficiary, the one who benefits from the Attribute. In the following
extract, the Attributor and Beneficiary are tɩ (‗we‘) and a bʋn-bul (‗the plant‘)
respectively:

|||Ba a lɛ [[[tɩ na =a tʋɔ
but DEF DEM 1PL REL POS.IND.FUT be:able
ɩ, || a tɩɩ-sɔw, ʋ tʋɔ tɛr
do.PFV DEF land 3SG be:able.PFV possess.PFV
buulu, ||ʋ tʋɔ tɛr ɓaaru kʋ̀
silt 3SG be:able.PFV possess.PFV moisture give.PFV
298
a bʋn-bul]]].|||
DEF plant
‗But what we can do (so that), the land, it can have silt, it can have
moisture for the plant.‘
As in other benefactive clauses, the Beneficairy participant in the attributive

clause is introduced by the verb kʋ̀ (‗give‘) in a verbal group complex.
6.5.1 Extending the types of relational clauses in delicacy

The preceding section identified two types of relational clauses based on the
mode of participation in the process, identifying and attributive clauses.
Relational clauses in Dagaare and in many other languages, if not all, also
divide up according to the nature of the process (see Caffarel et al. (2004) and
contributions therein; Halliday & Matthiessen 2014). This dimension gives us
three types of relational clauses, namely intensive, possessive and
circumstantial clauses. When these sub-types are combined with identification
and attribution, we get a complex sub-types of relational clauses, as Table 6.5
shows.
Table 6.5. Illustration of the principal types of relational clauses
These six sub-types are discussed in the following sections.
299
6.5.2 Intensive Clauses
Our discussion on relational clauses in the preceding section has been focused
on the intensive type. The reason is that intensive clauses are the prototype of
relational clauses and offer the clearest examples for the distinction between
the identifying and attributive type in Dagaare. They are also the most
frequent in discourse compared to the possessive and circumstantial types. In
this section, intensive clauses will be discussed in more detail, beginning with
the identifying mode (Section 6.5.2.1) and then the attributive mode (6.5.2.2).
6.5.2.1 Intensive Identification

Intensitive identification is realised in three ways. The first is where the
Process is realised by a unique identifying copula, nɛ or lɛ (‗be‘), depending
on idiolectal and sub-dialectal variation. This realisation is illustrated in the
underlined clause in (59) and the constructed dialogue in (60):
(59) Bible.is (Ebire 5: 5b)

Ɩ bie nɛ =b. Ɩ dɔw fʋ nɩ
1SG child COP 2SG.ACC 1SG give birth 2SG FOC
a dɩa.
DEF today
‗You are my child. I give birth to you today.‘
(60) A: Ãa lɛ =b.
who COP 2SG.ACC
B: Dɛr lɛ =m.
Der COP 1SG.ACC
A: ‗Who are you?‘

B: ‗I am Der.‘
In (59), the Indenfier is Ɩ bie (‗my child‘) and the Identified is the enclitic
pronoun =b. In (60), The Identifier participant in Speakers A and B
propositions are Ãa (‗who‘) and Dɛr respectively while the enclitic pronouns
=b and =m are the Identified participants. The second type of intensive
identifying clauses is verbless clauses, exemplified in (61) and (62) below:
300
(61) Bible.is (Matie 16: 14a)
Zã-Bati nʋ.
John-Baptist IDENT.SG
‗John the Baptist is the one (=It is John the Baptist).‘

(62) Kamaan na.
maize IDENT.PL
‗It is maize.‘
Example (61) will be a response to a question like Who is s/he? while (62) will
be a response to the question What is this? Thus, the function of clauses such
as these in the linguistic system is to identify an entity as opposed to assigning
a class membership. As these examples show, this type of identifying clauses
has no Process. The Identified participant is realised by one of two identifying
pronouns, nʋ (singular) or na (plural, non-human), and the Identifier is
realised by a noun group or a nominal clause (cf. Chapter 3, Section 3.4.3).
Diachronically, these kinds of clauses evolved from constructions with the
identifying verb nɛ such as those in (59), but where the Complement was
originally realised by the personal pronouns ʋ (third singular) or a (third
plural, non-human). In such an environment, the identifying verb nɛ fused
with the personal pronouns to derive the identifying pronouns nʋ (third
singular) and na (third plural, non-human), and the resulting in a verbless
clause (see Mwinlaaru & Yap (2017) for details). The third realisation of
intensive identification is also a verbless clause. An example is given in (63)
below:

I: A fʋ yuor nɩ bʋnʋ?
DEF 2SG name COP.FOC what
R: Ɩ yuor =ɩ George.
1SG name COP.FOC George
I: What is your name?
R: My name is George.
The structural pattern of this type is the reverse of the clauses in (61) and (62).
That is, the Identified element precedes the Identifier. In (63), the Identified
301
elements are A fʋ yuor (‗your name‘) and Ɩ your (‗my name‘) in the
interviewer‘s question and the respondent‘s answer respectively. The
corresponding Identifier elements are the Q-element bʋnʋ (‗what‘) and
‗George‘. Of note is the fact that the Identifier in these clauses is always
introduced by the focus particle nɩ (or the clitic form =n or =ɩ), which also
displays some verbal characteristics such as the ability take polarity paprticles
and other verbal group particles (e.g. Ɩ yuor bɛ nɩ George ɛ, ‗My name is not
George‘). This situation can also be explained by grammaticalisation. That is,
such clauses evolved from a further fusion of the identifying pronoun nʋ (and
also na) with the attributive copula ι to derive the focus marker nɩ (cf. Heine &
Reh 1983; Harris & Campbell 1995; Diessel 1999; Mwinlaaru & Yap 2017).
The focus particle in the resulting verbless clause is unique in the overall
system of the language since it is only in these kinds of clauses that it can co-
occur with the negative particle.
Together, the three types of clauses introduced above, namely the
copula construction and the two types of verbless clauses, realise intensive
identification in Dagaare. The extract below illustrates how they can combine
to construe identity in text (relevant clauses underlined):
(64) Bible.is (Matie 16: 13-16)45

A Yeezu na wa ta a Filib
DEF Jesus ADVLZ EVT reach.PFV DEF Phillip
tẽw Sezaare a, ʋ sowr =ɩ a
town Caesarea JUNC 3SG ask.PFV FOC DEF
ʋ po-tuurbɛ: “A nɩbɛ zie a, ãa

3SG followers DEF people place JUNC who
nɩ a Nɩsaal Bie?” Bɛ tɩ sɔw na:
COP.FOC DEF human child 3PL.HM PST.REM respond.PFV AFFR
“Bɛ mɩnɛ zie, Zã-Bati nʋ, bɛ mɩnɛ

3PL.HM some place John-Baptist IDENT.SG 3PL.HM some
zie ʋ ɩ Eli, bɛ mɩnɛ zie ʋ
place 3SG COP.PFV Elijah 3PL.HM some place 3SG
45
Punctuation has been modified. e.g. n‟ɩand n'ʋ in the original are changed to nɩand nʋ
respectively (see n. 4). Source: http://www.bible.is/DGIABB/Matt/16.
302
ɩ Zeremi bɩɩ Naaŋmɩn-ƴɛr-manɛ kãw.‖ Ʋ
COP.PFV Jeremiah or God-prophet some 3SG
lɛb sowri bɛ: “Ɛ a nyɩɩm, nyɩ yèl

again ask.PFV 3PL.HM and DEF 2PL.EMP 2PL say
kɛ ãa nɛ =m?” Sɩmʋ Pɩɛr tɩ sɔw
PROJ who COP 1SG.ACC Simon Peter PST.REM respond.PFV
na: “Fʋʋ nɩ a Naaŋmɩn Nɩ-kabra,
AFFR 2SG.EMP COP.FOC DEF God person-most powerful
a Naaŋmɩn [[na vʋʋrɛ a]] Bie.‖
DEF God REL live.PFVJUNC child
―When Jesus got to Phillip‘s town Caesarea, he asked his followers:
‗For the people, who is the Son of Man? They answered: For some,
John the Baptist is he; for some, he is Elijah; for some, he is Jeremiah,
or one of the prophets of God. He asked them again: ‗And you, you say
that I am who?‘ Simon Peter answered: ‗You are the Most Powerful
Man of God, the Child of the God who lives.‘‖
Jesus‘s question to his disciple borders on his identity, a Nɩsaal Bie (‗the Son
of Man‘). In their response, the disciples first used a verbless clause to
construe his perceived identify, in which the Identified element is the pronoun
nʋ and the Identifier is Zã-Bati (‗John the Baptist‘). Subsequent clauses are
attributive, where Jesus is characterised based on the qualities of the Old
Testament prophets. The second question still construes his identify with an
identifying clause and this time, it is not a verbless clause, but rather, it has a
Process, realised by nɛ, and his perspective simultaneously shifts from third
person to first person. The reponse which follows is again an identifying
clause realised by the nɩ-type of verbless clause.
When we compare the choices in the realisation of the identifying
clauses, we see that they are also influenced by the organisation of the clause
as a message. Jesus‘s first question places thematic prominence on the Q-
element, ãa. The disciples pick this up by thematising the requested
information. In the second question, Jesus again thematises the Q-element.
303
Peter‘s response, however, is now more influenced by the nature of the
information provided, the end-weight principle, the tendency to place longer
stretches of constructions in final position of the clause (cf. Quirk et al. 1985;
Halliday & Matthiessen 2014). Thus, the use of a complex noun group
constrains the choice of a more flexible verbless clause, which allows a switch
between the elements in the clause. This choice also allows the writer to give
focus prominence to the specification of Jesus identity, thereby given a textual
force, as it were, to this historic declaration.
It seems that identifying clauses across languages tend to be closely
related to systems of information packaging in the clause (i.e. THEME and
INFORMATION). For, instance, there is robust cross-linguistic evidence that
support the diachronic development of focus and theme markers from
identificational clauses (cf. Heine & Reh 1983; Heine & Kuteva 2002; Harris
& Campbell 1995; Mwinlaaru & Yap 2017) and Halliday and Matthiessen
(2014: 276-289; 298-300) show the impact of information structure on
patterns of identifying clauses in English. Thus, we see a tension between, on
the one hand, identifying clauses as specificational resources in the ideational
metafunction, and resources for signalling textual prominence, on the other
hand. This tension reflects that, at a more abstract level, referential
identification and focus of information belong to the same semiotic space, that
of pointing (see Mwinlaaru & Yap 2017). In this sense, identifying clauses are
unique in the overall system of TRANSITIVITY. They provide resources for
language to organise itself into a consumable entity.
6.5.2.2 Intensive Attribution

In text (64), intensive attributive clauses combine with identifying clauses in
construing Jesus‘ identity. The attributive clauses in the text are isolated
below:46
(65) Bible.is (Matie 16: 14b)

[…] bɛ mɩnɛ zie ʋ ɩ Eli.
[…] 3PL.HM some place IDENT.SG COP.PFV Eli
46
These clauses are in broad focus. Thus, the end-focus particle nɩ, whose usual place is after
the Process, is absent (cf. Chapter 5, Section 5.4.1 & 5.4.4).
304
‗… for some (people), he is Elijah.‘
(66) Bible.is (Matie 16: 14c)
[…] bɛ mɩnɛ zie ʋ ɩ Zeremi bɩɩ
[…] 3PL.HM some place 3SG COP.PFV Jeremiah or
Naaŋmɩn-ƴɛr-manɛ kãw.
God-prophet some
‗… for some (people), he is Jeremiah or one of the prophets of God.‘
As with identifying clauses, intensive attribution also has a unique copula verb
that realises the Process, which is ɩ (perfective) or ɩrɛ (imperfective). Other
verbs of attribution will be identified in the course of the discussion in this
section. Generally, clauses of intensive attribution can be characterised along
the following lines: (a) the class the Carrier is ascribed to can either be an
entity or a quality (b) the process of attribution can either be neutral or phased
(cf. Halliday & Matthiessen 2014).
(i) Class membership specification: The Attribute ascribed to the Carrier

can be construed as either an entity or quality. In the extracts below, the
Attribute particpants represent entities (Attribute is highlighted).

A Abɛl wa ɩ =n pi-cɩɩnɛ, ɛ a
DEF Abel EVT COP.PFV FOC shepherd CONJ DEF
Kayɛ ɩ kʋɔra.
Cain COP.PFV farmer.
‗Abel became a shepherd and Cain was a farmer.‘
(68) NANSU citation
Fʋ ɩ =n nɩr [[na mɩ́ ɓaa a tɩ
2SG COP.PFV FOC person REL HAB cool.PFV DEF 1PL
suur a, || ɛ kʋ̀ tɩ pʋpɩɛlʋ a]].

anger JUNC CONJ give.PFV 1PL joy JUNC
Fʋ ɩ =n sãa ɛ lɛ ɩ pi-cɩɩnɛ a
2SG COP.PFV FOC father CONJ also COP.PFV shepherd DEF
tɩ zie.
1PL place
305
‗You are someone [[who cools down our temper || and gives us joy]].
You are a father and also a shepherd to us.‘
Example (67) represents the occupation of Cain and Abel as Attributes. Thus,
while Abel is characterised among the class of shepherds, Cain is
characterised among farmers. Example (68) is from a doing text, a tribute by a
student association expressing gratitude to their mentor and teacher. The
orientation of the text is to praise the qualities of the mentor. These qualities,
are however, represented in the text as entities by use of metaphor, that both
grammatical (underlined) and lexical (in bold) metaphor.
When the attribute is realised as quality, there are two alternatives
involved. The first represents the quality as an Attribute participant, realised
by a nominal group whose Head is an adjectival noun. The second encodes the
attribute in the Process, which is realised by an adjectival verb (See Chapter
3.4.2.1 & 3.4.2.2 on adjectival nouns and verbs respectively). Table 6.6
illustrates the two alternatives with colour terms. Nominal groups and varied
in number and modification are verbal groups are varied in aspect in order to
show different dimensions of the examples.
Table 6.6. Options in representing quality in intensive clauses
Further examples of encoding the Attribute in the Process are given in (68)
below:
306
A kɔb, dɩya ʋ vɩɛl a
DEF farming last year 3SG be:good AFFR
kʋ̀ mãa. Ɩ pʋɔ pɛlɩ na […].

give.PFV 1SG.EMP 1SG stomach be:white.PFV AFFR […]
Al bome sãw na. Al

3PL.EMP.NHM things be:spoiled. PFV AFFR 3PL.EMP.NHM
bɛ maal ɛ.
NEG.IND.NFUT do:well NAFFR
‗The farming last year, it was good for me. I‘m happy … The crops
there are spoiled. The crops there did not do well.
In (69), there are four attributive clauses and in all the Process is realised by an
adjectival verb. The first begins with an Asolute Theme a kɔb (‗the farming‘),
and a marked Theme, dɩya, (‗last year‘). The Carrier participant is ʋ (third
singular) and the Process is vɩɛl (‗be good‘) and it encodes the Attribute
ascribed to the Carrier. There is also a Beneficiary participant, mãa (first
singular, emphatic) and it is introduced by the verb kʋ̀ (‗give‘). The second
clause (Ɩ pʋɔ pɛlɩ na, ‗I am happy‘) represents emotion. The Carrier is a body
part, Ɩ pʋɔ (‗my stomach‘), and the Process, pɛlɩ (be white) encodes the
Attribute. As the analysis shows, this is a metaphorical representation of
emotion. The verb realising the Process in the third clause (i.e. sãw, ‗be
spoiled‘) represents a negative attribute while the verb in the fourth clause (i.e
maal, ‗do well‘) represents a positive attribute although the verbal group in
which it is Head is negative, as the negative particle bɛ indicates. Thus, in all
these clauses, the burden of attribution lies with the Process.
(ii) Neutral versus phased attribution: This relates to one characteristic

distinction between attributive and identifying clauses in relation to aspectual
contrast. As was mentioned above, the Process in identifying clauses only
occurs in perfective form, which reflects that it represents static experience.
Attributive clauses, however, can occur in either perfective or imperfective
aspect. Many of the examples that have been considered so far, such as in
(69), are in the perfective aspect. In these instances, the Attribute ascribed to
307
the Carrier is bounded as a static quality. The imperfective aspect, on the other
hand, construes attribution as a dynamic process. Examples are given below:
(70) St. Maria play

Ɩ lιɛbɛ nι faara.
1SG turn.IPFV FOC priest.
I‘m becoming a priest.‘
(71) Ɩ ɩrɛ nɩ faara.
1SG COP.IPFV FOC priest
‗I am becoming a priest.‘
In both (70) and (71) the Attribute faara (‗priest‘) ascribed to the Carrier is
represented as being progressive as opposed to a completely attained quality.
The imperfective can also represent semblativity, the sense of semblance in
quality as opposed to exactness:
(72) A kparʋ ɩrɛ nɩ pʋla.

DEF shirt/dress COP.IPFV FOC white
‗The atire seems white‘.
6.5.2.3 Projection in intensive clauses
Projection is unique to intensive clauses among relational clauses and within

intensive clauses, it is restricted to processes realised by a small set of verbs.
In the attributive mode, these verbs are sɛw (‗become necessary‘, ‗be
appropriate‘) and fɛr (‗be necessary‘, ‗be crucial‘). Clauses in which they
occur can only project imperative clauses:

||| A sãa nɩ bie, a sɛw na || ɛ
DEF father and child 3PL.NHM be:appropriate AFFR PROJ
bɛ ƴɛrɛ wõne taar. |||

3PL.HM speak.IPFV hear.IPFV each other
‗The father and son, it is necessary that they are agreeable.‘
308
||| A fɛr =a || kɛ nɩr za
3PL.NHM be:necessary.PFV AFFR PROJ person all
mɩ̀ lɛb kul a ʋ saamɩnɛ
also return.PFV go.home.PFV DEF 3SG fathers
tew-dɔwra-pʋɔ . |||
town-birth-inside
‗That it is necessary that everybody also goes back to the birth town of
his fathers.‘
As the examples show, these relational clauses have no participants. The

Subject, which is always realised by the pronoun a (third plural, non-human),
is a dummy Subject and has no role in the transitivity structure of the clause.
The function of the projecting clause as a whole is to modulate the proposal
realised by the projected clause. It characterises the projected clause as a
necessity, a metaphorical strategy speakers use to distance themselves (and
others) from the projected proposal.
With regards to the identifying mode, the Process in the projecting
clause is always wul (‗show‘, ‗means‘) and it can project either a declarative
clause (see examples (75) & (76)) or an imperative clause (77):

||| A wul || ɛ a kʋɔra za, […] bɛ
3PL.NHM show.PFV PROJ DEF farmer all […] 3PL.HM
ɩ =n a za nyɛ wɛ? |||

do.PFV FOC DEF all see.PFV INT
‗It means that all the farmers, they tried all?
||| A wul || kɛ a tɩntɛr tɛr =ɩ
3PL.NHM show.PFV PROJ DEF plain possess.PFV FOC
ziir ata. |||

places three
‗It means that the plain has three areas.‘
309
(77) A fʋ ɩb wul kɛ tɩ cere.
DEF 2SG behaviour show.PFV PROJ 1PL go.IPFV
‗Your behaviour means that we (should) go.‘
As with the identifying mode, the projecting clause enacts an impersonal

stance towards the proposition or proposal that is realised by the projected
clause. In other words, they construe a locution as exisitng out there in the
semiotic realm without being uttered by anybody, that is, as a fact (Halliday &
Matthiessen 2014).
6.5.3 Possessive Clauses

In possessive clauses, the Process is typically realised by a verbal group with
the verb tɛr (‗possess‘) or law (‗possess together‘) as Head:
(78) St Maria
(a) Ãana bɛ mɩ́ tɛr pɛr
these NEG.IND.NFUT HAB possess.PFV bottocks
ɛ.
NAFFR
‗These ones have no use.‘

(b) Fʋ tɛr =ɩ libir bɩ?
2SG possess.PFV FOC money INT
‗Do you have money?‘
(79) Tɩ law nɩ a libir.
1PL possess:together.PFV FOC DEF money
‗The money belongs to both of us.‘
The distinction between the attributive and identifying modes in possessive

clauses is based on the definiteness of the nominal group functioning as
Complement (i.e. Identifier or Attribute). In the identifying mode, the nominal
group realising the Identifier is always definite while the one realising the
Attribute is always indefinite. Thus, while the clauses in both (78a, b) are
attributive(79) is an identifying clause, it borders on identifying the status or
belonginess of the money.
310
There are two other characteristics that are unique to attributive
possession and that set it apart form the identifying mode. The first is the
general ability of attributive clauses to occur in the imperfective or represent
dynamic relation (80):
(80) Ʋ tɛrɛ nɩ ya.

3SG possess.IPFV FOC wisdom
‗S/he is getting wise.‘
Second, as in intensive attribution, the Attribute possessed by the Carrier can

be represented as either an entity or quality. Examples (78a) and (80) are
instances where the Attribute is an abstract noun pɛr (‗bottocks/usefulness‘)
and ya (‗wisdom‘) respectively. Another example is given in (81):
(81) Wrokshop report 1

A tɩ nɩbɛ ani nɛ tɩ tɛr =ɩ
DEF 1PL people eight DEM 1PL possess.PFV FOC
nɔ-law.
mouth-together
‗We those eight people, we have unity.‘
In (78b) and (79), on the other hand, the Attribute is an entity, libir (‗money‘).
In addition to qulaities, however, possessive clauses also represent

other abstract phenomena such as verbalisation ((82) & (83)) and emotion (84)
as Attributes:

Fʋ pãa taa wa tɛr ƴɛrʋ kàw.
2SG ADV MOD COND possess.PFV speech some.
‗You may have something to say.‘
(83) St. Maria
Mãa bɛ tɛr ƴɛrʋ za ɩ.
1SG.EMP NEG.IND.NFUT possess.PFV speech all NAFFR
‗I don‘t have anything to say.‘
311
(84) NANSU citation
Tɩ tɛr =ɩ pʋ-pɩɛlʋ.
1PL possess.PFV FOC stomach-whiteness
‗We have joy.‘
This characteristic expands the potential of possessive clauses in the sense that
they can represent processes typical of other domains of experience such as
those construed by verbal and emotive clauses.
6.5.4 Circumstantial Clauses

Circumstantial relational clauses are by far the least common in discourse
among the three sub-types of relational clauses. They are typically realised by
a clause in which the Head of the verbal group is a locative verb, such as be
(‗be at, ‗be in‘), yi (‗be from‘, ‗be out‘),47pʋɔ (‗be among‘), and law (‗be
together‘). The following extract illustrates the use of be (‗be at‘, ‗be in‘) in
circumstantial clauses:

R: A yéle a, bɛ wa =n nɩ sukuul
DEF things JUNC 3PL.HM come CAUS FOC school
yaga. A sow. Bɛ wa =n polɩkɩlɩnɩk
many 3PL.NHM help.PFV 3PL.HM come CAUS polyclinic
anɩ eerh ...
and INTJ
I: Nyɩnɛ na a polɩkɩlɩnɩk nɛ be
where IDENT.PL DEF polyclinic DEM be:at
a?
JUNC
R: Kàw be =n Ko, kàw be eerh …

one be:at FOC Ko one be:at INTJ
Lawbusie, kàw be bom ŋa eerh ...

Lambusie one be:at thing DEM INTJ
47
It is worthwile to make a distinction between the locative verb yi (‗be from‘) and its
homonymous counterparts, comprising the motion verb yi (‗come/go out‘) and the adjectival
verb yi (‗washed clean‘, ‗to fade‘). The latter two do occur in the imperfective, yire
(‗coming/going out‘; ‗washing clean‘, ‗fading‘).
312
Baabili.
Baabili
R: ‗The things, they brought many schools. (And) they helped. (And)
they brought (a) polyclinic and eerh …‘
I: ‗Where is it that polyclinic is? (=Where is that polyclinic?)‘
R: ‗One is in Ko, (and) one is in Lawbusie, (and) one is in this thing
eerh … Baabili.‘
Dagaare circumstantial clauses are also the least distinct between the two
modes, attributive and identifying. The only distinguishing characteristic is the
ability of attributive clauses to construe dynamic process, that is, their ability
to occur in the imperfective aspect, while identifying clauses cannot. This also
implies that the unique ablility of attributive clauses to construe dynamic
relation is the most consistent distinguishing characteristic between the
identifying and attributive modes in all three sub-types of relational clauses.
Instances of attributive circumstantial clauses are illustrated by examples (86)
and (87) while (88) and (89) illustrate identifying circumstantial clauses:

Tɩ gari mɩ́ pemperi =n a lɛ [[a
1PL ridges HAB cross.PFV FOC DEF DEM DEF
kʋɔ na cere a]].

water REL go.IPFV JUNC
‗Our ridges normally cross where the water passes.‘

(87) Ʋ pʋɔrɛ nɩ tɩɩm pʋɔ.
3SG be:among.IPFV FOC 1PL.EMP inside
‗He is going to be among us‘ (=He will be in our group).‘
Ɩ yi =n ah a Nandɔm West.
1SG be:from.PFV FOC INTJ DEF Nandom
‗I come from uhm Nandom West.‘
313
Ʋ yi na ƴãw pʋɔ a tɩ
3SG come:out.PFV AFFR put.PFV be:among DEF 1PL
ƴãw ɛ nɩbɛ bɛ nyɛ ʋ ɛ.

self CONJ people NEG.IND.NFUT see.PFV 3SG NAFFR
‘He has come out among us, but people have not seen him.‘
As the examples show, the Process in circumstantial clauses is locative in

meaning and, likewise, the Attribute participant denotes place. In
circumstantial clauses, therefore, the boundary between circumstances and
participants is neutralised since the locative adverbial group that realises the
Attribute is obligabtory in the structure of the clause (cf. Halliday &
Matthiessen 2014; see also Section 6.6.3 on existential clauses).
Two points need to be made of the interaction between circumstantial
clauses and interpersonal grammar. The first relate to polarity and specifically
has to do with the semantically related verbs be, ka, kabe and bebe (‗be at‘).
Among these, the verb be is neutral in terms of polarity, occurring in both
positive and negative clauses. The others ka, kabe and bebe, on the other hand,
are restricted to negative clauses. However, they do not carry the negative
meaning of the clause by themselves since the presence of a negative particle
is still required for negation. 48 Thus they illustrate a linguisitc phenomenon
where the negative feature has been experientialised (as in English lack for
‗not have‘) and a negative polarity marker is still needed as an interpersonal
expression of polarity. The second point regards mood choice. There are two
ways of realising the non-polar interrogative in circumstantial clauses. The
first is to encode the question in the Process by using an interrogative verb
(wa, ‗be where‘). The second option is to enact the question by using a Q-
word, nyɩnɛ or nyɩn (‗where‘). As an illustration, let‘s compare the clause
highlighted in (90) with a reconstructed version in (91) below (interrogative
item is in bold):
48
This is the case in the Lobr dialect, which is our focus here. In some dialects (e.g. Central
Dagaare) the near equivalent of the copula ka, is a true negative copula realised as kyɛ (also
true of other Mabia languages such as Dagbani). It carries negative meaning by itself. Just like
the Lobr ka Central Dagaare kyɛ grammaticalises from the locative adverb kyɛ(‗here‘) in this
dialect.
314
Table 6.7 Examples of verbs realising the Process in relational clauses 49
process type verb gloss
perfective imperfective gloss
identifying nɛ, lɛ - ‗be‘
dɩ - ‗be called‘ (i.e. naming)
Attributive ɩ ɩrɛ ‗be‘
ya, sãw, sɔb, yarɛ, sãwnɛ, ‗be mad‘, be sploit, be
nʋmɛ,ŋmɛ, sɔbr ŋmɛrɛ,etc. black, ‗be sweet‘, ‗be like‘
etc. (‗resemble‘),etc.
Possessive tɛr, so, law tɛrɛ, sore, lawnɛ ‗possess‘, ‗own‘,
‗possess/own together‘
be bere
bebe -
ka karɛ ‗be at‘
kabe -
circumstantial law lawnɛ ‗be together‘
pʋ́ɔ pʋ́ɔrɛ ‗be among‘
yi - ‗be from‘
pemperi - ‗lie/be across‘
wa - ‗be where‘

A dãa [[ɩ na da kʋ̀ fʋ
a]], a wa?
JUNC 3PL.NHM be:where
(91) A dãa [[ɩ na da kʋ̀ fʋ

a]], a be =n nyɩnɛ?
JUNC 3PL.NHM EXIST.PFV FOC where
The pito/beer I bought for you, where is it?
49
The ‗be at‘locative verbs evolved from the locative adverbs ka (‗here‘) and be (‗there‘) in
two layers, first the verbs ka and be (through renalysis) and then kabe and bebe (through
fusion of the already evolved copulas and the locative adverbs). See Heine & Kuteva (2002b:
203-204) for examples of this grammaticalisation pathway (see also fn. 4). Also, the adjectival
verb nʋmɛ has no imperfective form.
315
It should be noted that the use of an interrogative verb in Dagaare is unique to
circumstantial clauses (cf. Chapter 4, Section 4.4.1.2.2on elemental
interrogative clauses).
The verbs realising the Process in the different types of relational
clauses are summarised in Table 6.7. Except for attributive processes realised
by adjectival verbs (see also Chapter 3 on adjectival verbs), the rest are only
realised by a limited number of verbs. Also, as the table shows, some verbs do
not have imperfective forms.
6.6 Other Process Types

In the preceding sections, three principal process types in Dagaare have been
identified and discussed, comprising material, mental and relational clauses. It
has been shown that each of these process types occupy a point in a semantic
region that construes different domains of experience. Material clauses are
more towards outer experience, typically construing our experience of the
physical word, that is, actions or doings and happenings. Mental clauses, on
the other hand, are more towards inner experience, typically construing
sensing, defined as cognition, desideration, emotion and perception. Relational
clauses represent being and having across two modes: identification and
attribution. Although they are more towards outer experience such as material
clauses, they also embody characteristics of abstract clauses since they
embody forms of symbolic representation, specifically intensive clauses that
project. It has also been shown that each process type comes along with a
unique configuration of participants.
Between these principal types of process types, however, lie other
forms of experience, each of which combines the characteristics of at least two
of the principal types. These are also three types, namely verbal, behavioural
and existential clauses. Verbal clauses lie between relational and mental
clauses. Behavioural clauses lie between mental and material clauses and
embody some characteristics of verbal clauses as well. Existential clauses lie
between material and relational clauses. Each of these intermediate process
types are discussed below.
316
6.6.1 Verbal Clauses
These are clauses of saying. One discourse function of verbal clauses in
conversations and other dialogic contexts is to frame and position different
voices in the text. This function is illustrated in (92) with the opening of an
interview (verbs realising verbal processes are in bold):

Tɩ ƴɛr a sɔw nɩbɛ yaga a
1PL speak.PFV AFFR respond people plenty CONT
bɛ za yèl kɛ […] bɛ lɛ lɛb

3PL.HM all say. PFV PROJ […] 3PL.HM ADV return.PFV
voti kʋ̀ a NDC ɛ bɛ
vote.PFV give.PFV DEF NDC CONJ 3PL.HM
cãa tɛr a tew [...]

be:still possess.PFV DEF country […]
ɛ tɩ mɩnɛ mɩ̀ yèl kɛ "ai, ɛ bɛ
CONJ 1PL some also say. PFV PROJ no PROJ 3PL.HM
voti sɩkɛt an blaʋs.” […] Ɛcɛ sɛw

vote.PFV skirt and blouse […] CONJ be:appropriate
fʋ na ƴɛr a, ɩ bɔbr ɛ ɩ
2SG POS.IND.FUT speak JUNC 1SG want.PFV PROJ 1SG
de niwn sowri fʋ: A fʋ yuor nɩ bʋnʋ?

take face ask.PFV 2SG DEF 2SG name COP.FOC what
―We have spoken to many people and they say that it is appropriate
that they vote for NDC again so that they continue to govern the state
… And some of us say that ‗no, that they should vote skirt and blouse.‘
... But in order for you to speak, I want to first ask you: What is your
name?‖
As the extract shows, there is a range of verbs that realise verbal processes.
These include verbs of speaking or talking, asking, naming and calling. The
most common and general verb is, however, yèl (‗say‘).
317
Two general characteristics can be identified for verbal clauses,
namely the key participants associated with them and their ability to project
other clauses. The participants are first described below:
(i) Sayer: The Sayer is the one to whom the saying is attributed, the one who
says. It is the most salient participant and may be the only one in a verbal
clause. In the extract above, the Sayer participants are realised by the pronouns
Tɩ (‗We‘) bɛ (‗they‘) fʋ (‗you‘) in that order, reflecting shifts in the
interpersonal orientation of the flow of discourse and how the interviewer
positions herself relative to different voices. The Sayer is realised by a
nominal group and, as our examples suggest, it is typically a conscious
participant. Only a restricted class of non-conscious entities such as gan
(‗letter‘, ‗book‘), Sɛb-sow (‗Scripture‘), taam (‗time‘), waalɛsɩ (‗wireless‘) and
TV can occur as Sayer:
(93) A Sɛb-sow yèl a kɛ bo?

DEF writing-holy say. PFV AFFR PROJ what
‗What does the Scripture say?‘
These are nouns that are semiotic in nature by serving as modes of

signification in society. Given that these modes of meaning are modern to the
Dagara society, it can be inferred that the use of verbal clauses with these non-
concious entities as Sayer is a new development in the language largely
through contacts with western civilisation.
(iii) Receiver: Another participant associated with verbal clauses is the

Receiver, the one who receives the saying. The Receiver is always a conscious
participant and is realised by a nominal group. In the extract above, the
Receiver participants are nɩbɛ yaga (‗many people‘) in Tɩ ƴɛr a sɔw nɩbɛ yaga
(‗We spoke to many people‘) and fʋ (‗you‘) in ɩ de niwn sowri fʋ (‗I first ask
you‘). This participant is like the Recipient and Client participants in material
clauses in the sense that it is the beneficiary of the saying and it is also often
introduced by the verb kʋ (‗give‘), particularly in clauses of saying as opposed
to asking (see Section 6.7.2 on benefactive clauses). In the following extract,
318
the Receiver is realised by the enclitic pronoun m (‗me‘) while the Sayer is
realised by the second plural nyɩ (‗you‘):

Mɛ́ nyɩ taa na lɔb zʋkpar na
EXP 2PL MOD POS.IND.FUT throw proverb DEM
kʋ̃ =m: “dɔkta a, sanɩ fʋ tʋɔra!”

give.PFV 1SG.ACC doctor JUNC heal.PFV 2SG self
―I believe you will tell me this proverb: ‗doctor, heal yourself.‘‖
(iv) Verbiage: This participant indicates what is said. Specifically, the

Verbiage may either be (a) the name of what is said (e.g. zʋkpar ‗proverb‘ in
example 94) or (b) the content of the saying. We will call the former
manifestation of Verbiage the elaborating type since it elaborates the
verbalisation by specifying it. The latter manisfestation is an extending type.
That is, it adds further information to the verbalisation realised by the Process.
The elaborating type is underlined in (95) and the extending type is underlined
in (96):

Fʋ na de nɩ niwn bʋɔlɩ a
2SG POS.IND.FUT take FOC face mention.PFV DEF
fʋ your anɩ a lowbowr nɛ fʋ na

2SG name and DEF area DEM 2SG REL
yi a Nandom ka a.
be:from DEF Nandom here JUNC
‗You will first mention your name and the area where you come from
in Nandom here.‘
Ɩ mɩ̀ kʋ yèl kʋ̀ nyɩ sob nɛ
1SG too NEG.IND.FUT say. PFV give.PFV 2PL person DEM
[[na kʋ̀ =m a sɔr a ]] ɩ.

REL give.PFV 1SG.ACC DEF way JUNC NAFFR
‗I too will not tell you the one who give me the permission.‘
319
Regarding, the elaborating type, there is a general pattern of collocation
between the verbs realising the Process and the noun group realising the
Verbiage participant: yèl yél (‗say something‘); manɩ yél (‗explain
something‘); lɔb zʋkpar (‗tell a proverb‘); bʋɔlɩ yuor (‗mention name‘); suoli
suolu (‗narrate a story‘); ƴɛr ƴɛrʋ (‗speak speech‘); manɩ manʋ (‗narrate
events/happenings‘), manɩa tɛwr (‗explain the meaning‘). As these examples
suggest, the nominal group realising the Verbiage in such instances is very
often a nominalisation of the verb realising the Process in the clause. Also, just
as the verb yèl (‗say‘) is semantically general, its nominal counterpart, which
has been transalted here as ‗something‘, is the least in experiential specificity
among the deverbal nouns. Compared to material clauses, therefore, the
Verbiage participant is like the Scope participant, which also display similar
nominalisation tendency as the examples given here (cf. Section 6.4.1 &
6.7.2).
(v) Target: Target is the participant at whom a verbal activity is directed. The
Target participant can be distinguished from the Receiver participant in the
following terms: while Target is like Goal in a material clause, Receiver, as
mentioned earlier, is like Recipient and Client. In other words, the verbal
process impacts on the Target, as it were, while the Receiver is a beneficiary
of the Process. The Target in the following extract is highlighted (lengthy ones
underlined):

Ɩ bʋɔlɛ n= a nɩbɛ nɛ za [[na
1SG call.IPFV FOC DEF people DEM all REL
ɩ a tɩ ba-taabɛ]] anɩ a tɩ yɛbr

COP.PFV DEF 1PL friends-mates and DEF 1PL brothers
tɩ sãakʋm mɩnɛ, tɩ makʋm mɩnɛ, a tɩ
1PL grandfather PL 1PL grandmother PL DEF 1PL
sãa mɩnɛ […]. Ɩ puore bɛ nɩ barka […].

father PL[…] 1SG greet.IPFV 3PL.HM FOC gratitude […]
320
‗I am inviting/calling all those people [[who are our colleagues]]; and
our brothers, our grandfathers; our grandmothers; our fathers … I
thank them …‘50
The Target participant is restricted to a particular sub-type of verbal clauses,

namely ‗targeted verbal‘ clauses, as opposed to non-targeted clauses (Halliday
& Matthiessen 2014: 307). In the terms of Hopper & Thompson (1980),
targeted verbal clauses have a higher transitivity value than those that are no-
targeted and are therefore more towards the material end of the spectrum of
process types (cf. Hopper & Thompson (1980) on ‗high‘ and ‗low‘
transitivity).
Another major characteristic of verbal clauses besides the unique
configuration of participant roles is their ability to project a proposition or a
proposal. They share this characteristic with mental clauses, specifically,
cognitive and desiderative ones (Section 6.4.2.2) and a sub-type of intensive
relational clauses (cf. Section 6.5.2.3). As in mental clauses, projection in
verbal clauses can be in the form of quoting (98) or reporting (99). Projected
clauses are underlined:

Mɛ́ nyɩ taa na lɔb zʋkpar na
EXP 2PL MOD POS.IND.FUT throw proverb DEM
kʋ =m: “dɔkta a, sanɩ fʋ tʋɔra!”

give.PFV 1SG.ACC doctor JUNC heal.PFV 2SG self
ɛ lɛ yèl mɛ ƴãw pʋɔ: “maalɩ a fʋ
CONJ ADV say. PFV 1SG.ACC add.PFV be:among do.PFV DEF 2SG
tew-kpɩɛra-pʋɔ ka a lɛ [[fʋ na maalɩ a

hometown here DEF DEM 2SG REL do.PFV DEF
Kafanawɔm a]]!”
Capernium JUNC
―I believe you will tell me this proverb: ‗doctor, heal yourself!‘ and tell
me again: ‗do here in your hometown what you did in Capernium.‖
50
Here, tɩ yɛbr (‗our brothers‘) and tɩ sãakʋm-mɩnɛ (‗our grandfathers‘) are appositively
related. In the Dagara kinship system, one‘s grandfather is also one‘s brother, in principle.
321
(99) St. Maria
Ãa nɩɛ yèl kɛ a fʋ pɔw-yaa nʋ
who ADV say. PFV PROJ DEF 2SG daughter IDENT.SG
mɔ? ʋ yèl kɛ ʋ wa =n ʋ
CE 3SG say. PFV PROJ 3SG come.PFV FOC 3SG
pɔw ƴãw. Fʋ pãa kʋ nɩɛ sowr ʋ

woman sake 2SG ADV NEG.IND.FUT ADV ask.PFV 3SG
ʋ pɔw buor [[na nʋ || be a
3SG woman which REL IDENT.SG EXIST.PFV DEF
ka a]]?
here JUNC
‗Who even said that your daughter is the one? He said that he came for
his wife. Won‘t you now ask him which wife of his is here?‘
There is a difference between a projected clause and the Verbiage participant.

This difference is exemplified clearly in (98), where zʋkpar (‗proverb‘) in the
first clause is a Verbiage participant and the quoted clause is a projected
clause. While Verbiage is a participant within the verbal clause, a projected
clause is a separate proposition or proposal represented as the content of the
verbalisation.
Projecting clauses serve several discourse functions (see also Section
6.4.2.2on mental clauses) of an interpersonal kind such as evidentiality and
negotiation.
(i) Evidentiality: The evidential use of projecting clauses is already evident in

the interrogative clause in (99) above. Here, the verbal clause realises a
rhetorical question whose function is to focus the listener‘s attention on the
evidence of the proposition realised by the projected clause. The projecting
clause, therefore, often serves to attribute a proposition to a source. As we
discussed in Section 6.4.3, hearsay is often construed by speakers with a
projecting clause in which the Subject is realised by the the third person plural
which, in this context, lacks specific reference. We repeat the example from
Section 6.4.3 for the sake of convenience:
322
(100) Bɛ yèl kɛ fʋ kul =ɩ sɩr.
3PL.HM say. PFV PROJ 2SG marry.PFV FOC husband
‗They (=people/someone) say you got married.‘
The following example also shows the use of self-quoting strategy to enact a
proposition as first-hand information:

Ʋʋ, mãa yèl nyɩ: Ŋmɩn ƴɛr-manɛ bɛ
yes, 1SG.EMP say. PFV 2PL God prophet NEG.IND.NFUT
dãw nyɛ ƴawfʋ a ʋ tew e.

ADV see.PFV respect DEF 3SG town NAFFR
‗Yes, I say to you: A prophet of God has never got respect in his
hometown.‘
Self-quoting enacts a proposition as having more evidential weight than

hearsay, for instance.
(ii) Negotiation: This is the use of self-reporting as a strategy to manage

interaction in dialogue. This point can be illustrated with the following
dialogues (projecting clauses underlined):

Host: Ɩ na yèl kɛ fʋ barka yaga,
1SG POS.IND.FUT say. PFV PROJ 2SG thanks plenty
cɛ fʋ paa taa wa tɛr ƴɛrʋ
CONJ 2SG ADV MOD COND possess.PFV speech
kàw mɩ̀ na paa kʋ̀ a
some too POS.IND.FUT ADV give.PFV DEF
cɛlcɛlbɛ.
listeners.
Agric Officer: Mmmm, tɩ na yèl kɛ, a
INTJ 1PL POS.IND.FUT say. PFV PROJ DEF
kʋɔbɛ, yelmɩŋa na, a tew ɛ᷈

farmers truth IDENT.PL DEF world ADV
323
lɩɛbɛ na.
change.IPFV AFFR
Host: ‗I will say that thank you very much, but you might have
something to say to our listeners.‘
Agric Officer: ‗Mmmm, we will say that, the farmers, it is true, the
world is actually changing.‘
Host: Nyɩnɛ dem a nɩ tɩ wul a tome a
where people IDENT.PL 2PL PST.REM teach DEF work DEF
ŋa taabɛ?
DEM kind
Agric Officer: Tɩ yèl kɛ Ɔrbili deme, kʋɔbɛ
1PL say. PFV PROJ Orbili people farmers
anɩ [[na yi Ɔrbili a]], tɩ
eight REL be:from Orbili JUNC PST.REM
pʋɔ na.
be:among AFFR
Host: ‗Which people did you teach this kind of work?‘

Agric Officer: ‗Let‘s say that Orbili people, eight farmers who are
from Orbili, were among.‘
The underlined verbal clauses are used by the speakers to negotiate the
propositions in the projected clauses. One characteristic of projecting clauses
in the indicative, as in (102), is that they often include a future tense particle,
na (‗positve indicative‘) or kʋ (negative indicative), as a modulating strategy.
6.6.2 Behavioural Clauses
Behavioural clauses construe physiological and psychological processes.

Generally, they cover outward or visible manifestations of consciousness. In
other words, the processes they represent are observable phenomena that
emanate from consciousness. In this sense, behavioural clauses interface
between material clauses and immaterial ones such as mental clauses and they
share the characteristics of these two domains of experience. All the three
324
clauses in the following extract are behavioural clauses (Processes are in
bold):
(104) Agric workshop

||| A kʋɔbɛ, tɩ mɩ́ zɛlɛ nɩ
DEF farmers 1PL HAB beg.IPFV 2PL
na || […] nɩ mɩ́ ɩ kokere || ɛ mɩ̀

AFFR […] 2PL HAB do.PFV effort CONJ also
mɩ́ tʋɔ cɔwlɩ kaa a lɛ
HAB be:able copy.PFV check.PFV DEF DEM
[[tɩ na wa nɩ a]]. |||

1PL REL come CAUS JUNC
‗The farmers, we always beg you … you should always do well and
try out [[what we bring]].‘
As with material clauses, behavioural clauses normally denote an input of

effort in the actualisation of the process. Compared to material clauses,
however, they are of a lower transitivity value since this effort is relatively
abstract and the Process does not impact upon a participant as it is in an action
clause (cf. Hopper & Thompson 1980). The participants associated with
behavioural clause are Behaver and Behaviour.
Behaver: The Behaver participants the conscious entity engaged in the

Process. In the extract in (104), the pronoun tɩ (‗we‘) in tɩ mɩ́ zɛlɛ nɩ na (‗we
always beg you‘) is the Behaver and the behaviour it is enaged in is construed
by the Process mɩ́ zɛlɛ (‗always begging‘). The Behaver is typically a concious
being such as humans and animals as in (104). However, this includes plants
with verbs denoting growth and life (105):

||| A bʋn-bul, ʋ ta kpi e, || ɛcɛ
DEF plant 3SG NEG.IMP die.PFV NAFFR CONJ
cãa tʋɔ a baa. |||

be:still.PFV be:able.PFV PRT grow.PFV
‗The plant, it should not die, but should still be able to grow.‘
325
Here, the plant is conceived of as animate, a living thing. The Behaver is ʋ
(‗third singular) and the behaviour is represented by the Processes, which are
underlined.
Behaviour: The behaviour is sometimes not encoded in the Process but in a

participant in the clause. An example is kokere (‗effort‘) in nɩ mɩ́ ɩ kokere
(‗you should always put in an effort‘) in the extract in (104). This participant
is called the Behaviour. In clauses such as nɩ mɩ́ ɩ kokere (‗you should always
put in an effort‘), the verb realizing the Process is normally a general one,
typically ɩ (‗do‘), which lacks specificity. In other contexts, the the verb
realising the Process may be that of a material kind (and tends to be vague)
while the behaviour is encoded in the participant: Bɛ ŋmaa nɩ ziri (‗They told a
lie‘, literally, ‗They cut a lie‘); Ʋ sãw nɩ niwn (‗S/he frowned‘, literally, ‗S/he
distorted the face‘).
Like the Verbiage participant, in verbal clauses (cf. Section 6.6.1), the
Behaviour may, semantically, be an elaboration of the Process such as bɛ̀lʋ
(‗looking‘) in Tɩ bɛ́l =ɩ bɛ̀lʋ (‗We did looking‘, literally, ‗We looked looking‘),
and vʋʋrʋ in Ɩ vʋʋr =ɩ vʋʋrʋ (‗I took a breath‘, literally, I breathed breath). In
these constructions, the Behaviour functions to amplify the Process, and it is a
common intensification strategy in folktales and conversational narratives. It is
always possible in this kind of behavioural clauses to omit the Behaviour.
Almost all Dagaare behavioural clauses can be construed with this
amplification strategy, the exceptions being the metaphorical type in which the
Process is realised by a verb denoting action (e.g. Bɛ ŋmaa nɩ ziri, ‗They told a
lie‘ or literally, ‗They cut a lie‘). This phenonmenon is, however, also present
in material clauses (Tɩ cen nɩ cenu; ‗We walked a lot‘, literally, ‗We walked
walking‘) and verbal clauses (ʋ man =ɩ manʋ; ‗S/he narrated a lot of stories‘,
literally, ‗S/he narrated narrations‘).
As Halliday and Matthiessen (2014) note for English, behavioural
clauses are also the most indeterminate process type in Dagaare. As is
suggestive from our discussion so far, they share some of the characteristics of
material, mental and verbal clauses. Table 6.8 list examples of verbs realising
the Process across various sub-types of behavioural clauses (compare with
Halliday & Matthiessen (2014: 302) on English).
326
Table 6.8. Examples of verbs used in different types of behavioural
clauses
characteristics examples of verbs
(i) [near processes of cɛlɩ (‗listen‘), kaa (check), nyɛ

mental] consciousness (‗check‘/ ‗look‘), bɛ́l (‗look‘) zanɩ
represented as (‗dream‘; ‗learn‘), tɩɛrɩ (‗think‘), bɔ
forms of behaviour (look for/date), cɛlɩ (‗listen‘), saw
(‗agree‘), fɛr (‗worry somebody‘)
(ii) [near verbal processes as yɛr (‗talk/speak‘), ŋmaa (‗cut‘ as in

verbal] forms of behaviour ‗cut/tell a lie‘), sɔw (respond to), saw
(‗respond to a call‘)
(iii) physiological cɩr (‗spit‘) ko (‗cry‘), la (‗laugh‘), sãw

processes: (‗distort‘, as in ‗distort one‘s face‘),
volitional gɔrɩ (‗nod‘), siwni (‗sob‘)
(iv) physiological kpi (‗die‘), baa (‗grow‘), vʋʋrɩ

processes: non- (‗breathe‘), kɔr (‗cough‘), cɩr
volitional (‗sneeze‘) gbew (‗hiccup‘) ti (‗vomit‘),
gúr (‗sleep‘), maw (‗shiver‘)
(v) [near cɔwlɩ (‗mimic‘ / ‗try‘), zɛlɩ (‗beg‘), mŋɛ

material] ƴãw ã (‗curse‘), ɩ (‗do‘), mɔ (‗do
well‘), kaa (‗take care of‘), ga (‗lie
down‘), sɛb (‗dance‘), zɛb (‗quarrel‘),
zɩ (‗sit‘), ar (‗stand‘), yieli (‗sing‘),
ŋmarɛ (‗writing‘), cɛlɩ (‗wait‘)
We will briefly outline the characteristics of behavioural clauses that interface

with material, mental and verbal clauses below.
(i) Near material: These types of behavioural clausesrepresent activities or

actions as behaviour. In other words, they construe processes that are
otherwise material as processes of consciousness. They tend to be descriptive
or evaluative. Examples are given below (Process in bold):

Bɛ za ayi ende mʋɔr a.
3PL.HM all two ADV do:well AFFR
‗The two of them are doing their best anyway.‘
327
||| Kʋɔ yéle caa na fɛrɛ tɩ […] || A
water matter be:still AFFR worry.IPFV 1PL […] CONT
bɛ cãa mʋɔr =ɩ. |||

3PL.HM be:still.PFV struggle.IPFV COM
‗The issue of water is still worrying us … And they are still

struggling with it.‘
When there is a Behaviour participant in these type of clauses, it shares the

characteristics of the Scope participant in material clauses as in koreke in the
following clause from example (104): nɩ mɩ́ ɩ kokere (‗you should always do
well‘).
(ii) Near mental: These are behavioural clauses that construe mental
processes as psychological activity. In other words, they represent processes of
consciousness are behaviour. Examples are given below (Where there are two
clauses, the behavioural clause is underlined):

Bɛ mɩ́ cen na tɩ zanɩ bãwfʋ
3PL.HM HAB go.PFV AFFR DIST learn.PFV knowledge
kàw zie kàw.
some place some
‗They go to learn some knowledge somewhere.‘

A nɩbɛ kʋ̃ tʋɔ cɛlɛ a ʋ
DEF people NEG.IND.FUT be:able listen.IPFV DEF 3SG
ƴɛrʋ ɛ.
words NAFFR
‗The people cannot listen to his/her words.‘

Bɛ ƴɛrɛ wõne taar.
3PL.HM talk.IPFV hear.IPFV each other
‗They should be agreeable with each other.‘
328
Tɩ bɔbr kɛ tɩ paa gɛl nyɛ̃.
1PL want.IPFV PROJ 1PL ADV search.IPFV see.PFV
‗We now want to find out.‘
Examples (108) represent cognitive behaviour while (109) and (110) represent
perceptive behaviour. Example (111) combines the senses of both cognitive
and perceptive activity. Due to their hybrid nature, these types of behavioural
clausescan take a Phenomenon participant,as do mental clauses. In (108), the
Phenomenon is bãwfʋ (‗knowledge‘); in (109), it is a ʋ ƴɛrʋ (‗his/her words‘)
and, in (110), it is taar. Example (111) has one participant, the Behaver,which
is realised by tɩ (‗we‘).
Some verbs can be used in either mental or behavioural clauses
although they display different characteristics across the two process types.
For example, although ‗thinking‘ can be represented as either a behavioural
process or a mental process, in behavioural processes, the unmarked aspect is
the imperfective while in the mental clause, the unmarked aspect is the
perfective:
(a) thinking as a mental process

Ɩ tɩɛ kɛ fʋ wa na.
1SG think.PFV PROJ 2SG come.PFV AFFR
Senser Process: Actor Process:

mental material
nominal verbal gp. conj. nominal verbal gp.
gp. group gp.
‗I thought/think that you have come.‘

(b) thinking as a behavioural process
Ɩ tɩɛrɛ na.
1SG think.PFV AFFR
Senser Process: mental

nominal gp. verbal gp.
‗I am thinking.‘
Figure 6.4. ‘Thinking’ as a mental and as a behavioural process

329
Also, when ‗wanting‘ is represented as a behavioural process it can occur in
either perfective or imperfective aspect (bɔ; ‗look for‘, ‗date‘; bɔbr; ‗looking
for‘, ‗dating‘), but, in mental clauses, only the imperfective form (bɔbr;
‗want‘) can be used.
(a) perfective (behavioural)

Ɩ bɔ =n a gan.
1SG look.PFV FOC DEF book
Behaver Process: behavioural Behaviour
nominal gp. verbal group nominal gp.
‗I looked for the book.‘
(b) imperfective (mental/behavioural)
Ɩ bɔbr =ɩ a gan.
1SG want.PFV FOC DEF money
Senser Process: mental Phenomenon
Behaver Process: behavioural Behaviour
nominal gp. verbal group nominal gp.
‗I want the book.‘ / ‗I am looking for the book.‘
Figure 6.5. ‘Wanting’ as a mental and as behavioural process
As (b) in Figure 6.5 shows, out of context, the imperfective may lead to an
ambiguous reading between mental and behavioural clauses. There is also
lexical ambiguity with the verb bɔ since it can mean either ‗look for‘ or ‗date‘
although both senses are behavioural. Let‘s illustrate below how these
ambiguities is intentionally exploited in a comic play:
Daughter: Ɩ na cen ɩɩ tɩ ɔw a
1SG ADVLZ go.PFV 1SG.PUR DIST fetch.PFV DEF
kʋɔ a o, pobile kàw o, tɩ

water JUNC PRT gentlemen some PRT DIST
yee ʋ bɔbr mɛ na.

say. PFV.PROJ 3SG want.IPFV 1SG.ACC AFFR
[…] […] […]
330
Father: Ʋ bɔbr fʋ na ya? Fʋ bɔ
3SG want.IPFV 2SG AFFR INT 2SG look.PFV
na bɩ?
AFFR INT
Daughter: ‗I went to fetch the water O, and a certain gentleman O, said

he wants (=loves) me.‘
Father: ‗He wants you?‘ / ‗He is looking for you?‘
‗Have you dated (him)?‘ / ‗Did you look for (him)?‘
The underlined clauses evoke ambiguity. On the one hand, we can say that the
father intentionally misinterprets the statement of the gentleman‘s desire
(wanting or loving you) to mean behaviour (looking for you) in order to create
a comic effect. A superficial and a more plausible interpretation of the first
clause (i.e. Ʋ bɔbr fʋ na ya?) is, however, that of wanting and the second
clause (i.e. Fʋ bɔ na bɩ?) is that of courtship.
(iii) Near verbal: These are behavioural clauses that construe saying as
behaviour. Examples are below (Process is in bold):

Ɩ kʋ̃ ŋmaa ziri e.
1SG NEG.IND.NFUT cut.PFV lie NAFFR
‗I will not tell a lie.‘

Ah ɩ puore bɛ nɩ barka bɛ
INTJ 1SG greet.IPFV 3PL.HM FOC gratitude 3PL.HM
na na saw sɔw a lɛ tɩ
ADVLZ POS.IND.FUT agree respond DEF DEM 1PL
[[na yèl a]].

REL say. PFV JUNC
‗Ah I thank them for (the fact that) they will respond positively to
[[what we said]].‘
The Behaviour participant in these type of behavioural clauses shares

characteristics of the Verbiage in verbal clauses. Also, the verb yɛr can be
331
used in both behavioural and verbal clauses with different meanings. Compare
the following, for instance:
(115) A bie ƴɛrɛ nɩ paa.

DEF child talk.IPFV FOC a lot
‗The child talks a lot.‘
(116) St. Maria

Ɩ pɔw-yaa na ƴɛrɛ ɩ zie.
1SG daughter EHX talk.IPFV 1SG place
‗My daughter, please talk to me!‘
While (115) represent the behaviour or character of the child, and is thus a
behavioural clause, (116) represents saying, and is a verbal clause.
6.6.3 Existential Clauses

Existential clauses represent that something exists (or do not exist). They are
not very common in Dagaare discourse and are, perhaps, the least frequent
process type across languages where they can be rcognised as a distinct
process type (see Halliday & Matthiessen (2014) on English). The following
extract from the biblical creation story (from Sɛb-Sõw ƴɛr-bie, 1996) illustrates
existential clauses in discourse (existential clauses are underlined; Process is
in bold):
(117) A pɛr tib daar Naaŋmɩn tɩ ir =ɩ

DEF bottom start time God PST.REM bring:forth.PFV FOC
a salom nɩ a tẽw. A tẽw tɩ
DEF sky and DEF earth DEF earth PST.REM
ɩ =n vuo ɛ bom za bɛ ka
COP.PFV FOC hollow CONJ thing all NEG.IND.NFUT EXIST.PFV
be ɩ. A lɛ na a Naaŋmɩn tɩ
there NAFFR CONT DEM IDENT.PL DEF God PST.REM
yèl: “A cãa, ʋ be!” Ɛ a cãa

say. PFV DEF light 3SG EXIST.PFV CONJ DEF light
dɛ caalɩ.
ADV shine. PFV.
332
‗In the beginning, God brought forth the sky and the earth. The earth
was hollow and there was nothing.Then, God said: ―The light, it be!‖
And the light just shone.‘
The key participant in the existential clause is the Existent, the one whose
existence is represented by the clause, and it is normally realised by a nominal
group. In the existential clauses underlined in (117), the Existent is the only
participant, realised by bom za (‗everything‘) and ʋ (‗third singular‘). It should
be noted that, in the first clause, ɛ bom za bɛ ka be ɩ (‗and nothing was there‘),
the locative adverb be is a dummy Complement and lacks reference or
experiential specificity. It fulfils a requirement that the Complement position
in an abstract clause in Dagaare must be filled (see Section 6.7.1 for details).
Thus, it has no role in the transitivity structure of the clause. Existential
clauses such as A caaʋ be! (the second underlined clause in (117)) lie on the
borderline between existential and creative material processes (i.e. bringing
into existence), realised by verbs such as maalɩ (‗make‘)and ir (‗bring forth‘).
This use of existential clauses is, however, rare in Dagaare discourse.
The lack of reference in the complement of existential processes
realised by be and ka is important in distinguishing an existential clause with
these verbs from a circumstantial relational clause with the same verb (i.e. ‗be
at‘) (cf. Section 6.4.5). Thus, while example (118) is an existential clause,
(119) is a circumstantial clause, with be as locative Attribute:
(118) Bom za bɛ ka be ɩ.
thing all NEG.IND.NFUT be:at.PFV there NAFFR
‗There is nothing.‘
(119) Bom za bɛ ka a be ɩ.
thing all NEG.IND.NFUT be:at.PFV DEF there NAFFR
‗Nothing is there (=where you are standing).‘
As mentioned earlier, in (118), the Complement be is dummy. In (119), the

definite article a (‗the‘), however, gives specific reference to the locative
adverb be. It is this referential meaning that renders the clause a circumstantial
clause rather than an existential clause.
333
Registrially, existential clauses are characteristic of traditional
narratives and they are always used in the opening of folktales in introducing
characters as the following example shows (also see example (51) in Chapter
5, Section 5.4.3):
(120) Folklore
Nɩ gure gure o! Badɛr nʋ
2PL be:alert.IPFV be:alert.IPFV PRT Spider IDENT.SG
tɩ be be. Nàa pɔw wa kpi.

PST.REM EXIST.PFV there king woman EVT die.PFV
‗Once upon a time, there lived Spider. Then the King‘s wife died.
This registerial use of existential clauses is a strong typological tendency that

has been reported across languages (cf. Teruya 2007; Halliday & Matthiessen
2014).
Further, in addition to representing material existence as has been
illustrated so far, existential clause can be used metaphorically to represent the
existence of abstract phenomena (121) and emotions (122):
A kɔb aŋa, a tome aŋa [[tɩ na

DEF farming DEM DEF work.PL DEM 1PL REL
to a]], tonɛ tʋɔ na be

work.PFV JUNC profit be:able .PFV AFFR EXIST.PFV
be bɩ?
there INT
‗This farming, this project [[that we have done]], is there profit in it?‘
Pʋ-pʋla wa be be a, tɩ bãw
good will COND EXIST.PFV there JUNC 1PL know.PFV
kɛ a yél ŋa a, ʋ na tɔl
PROJ DEF matter DEM JUNC 3SG POS.IND.FUT move.PFV
=ɩ ni-daa.
FOC forward
334
‗If there is good-will, we know that, this matter, it will move forward
(=this programme will be successful).‘
Examples (121) and (122) are instances of grammatical metaphor. A more

congruent realisation of the meaning they represent would be relational
clauses. Specifically, (121) is a realisation of attribution, where a more
congruent realisation would be a possessive relational clause as in (123) below
(see Section 6.5 on relational clauses). Example (122), on the other hand, is an
indirect realisation of emotion of the kind typically encoded in the Process of
an attributive clause as has been reconstructed in (124) below (cf. Section 6.5):
(123) A tome aŋa tɛr =ɩ tɔnɛ bɩ?

DEF work.PL DEM possess.PFV FOC benefits INT
‗Does this project have benefits?‘

(124) Tɩ pʋɔ wa pɛlɩ a…
1PL stomach COND be:white.PFV JUNC
‗If we are willing / happy …‘
Again, all the examples that have been given so far illustrate typical
existential clauses, those with the verbs be and ka. These verbs neutrally
encode that something exists (including the negation of existence). However,
the Process in existential clauses may also encode the manner of existence (cf.
Halliday & Matthiessen 2014). Here, the Process is realised by a wider range
of verbs that denote posture or manner (see Table 6.9). One characteristic of
this second sub-type of existential clauses is that they normallyrequire two
participants, the Existent and Place of existence. Exceptions are processes
realised by the verbs cãa (‗still be‘) and cɛ (‗remain‘), where the Existent is
always the only participant in the clause. Let‘s illustrated this with a
constructed example in Figure 6.6:
(a)
Kaa laarɩ nɩ a kʋɔ zu.
oil be:on surface.PFV FOC DEF water ADP
Existent Process Place
‗There is oil on the surface of the water.‘
335
(b)
Kaa cãa na
oil be:still.PFV AFFR
Existent Process
nominal group verbal group
‗There is still oil.‘
Figure 6.6. The structure of existential clauses
Again, except for cãa (‗still be‘), cɛ (‗remain‘), the verb functioning as Head
in the verbal group that realises the Process in manner-encoding existential
clauses is always in the perfective (Compare with Halliday & Matthiessen
2014: 310). Imperfective forms of these verbs (cf. Table 6.9) are restricted to
their use in material clauses, as illustrated in Figure 6.7 in contrast example (a)
in Figure 6.6.
A Naaŋmɩn Vʋʋrʋ tɩ laarɛ nɩ a kʋɔnɩ zu

DEF God spirit PST.REM hover.IPFV FOC DEF waters ADP
Actor Process: material circumstance: Place

‗The Spirit of God was hovering over the water.‘
Figure 6.7. The structure of a material clause
Thus, aspectual contrast does contribute to the classification of process types

in Dagaare (see also Hopper & Thompson 1980 on the role of aspect in
transitivity).
Table 6.9. examples of verbs associated with existential clauses

Type examples of verbs
neutral be bere
ka karɛ
cãa (‗still be‘), cɛ (‗remain‘) cãanɛ, cɛrɛ
laarɩ (‗be on surface‘), pemperi (‗lie across‘), yawlɩ -
+manner
(‗be hung‘), ga (‗lie‘), tewli (‗be against‘), sawlɩ
(‗be in‘), ba (‗be pinned‘)
336
6.7 Generalising across Process Types: The Transitive Model versus
the Ergative Model
As mentioned earlier, the system of TRANSITIVITY in Dagaare, as in many

other languages such as English, comprises two simultaneous sub-systems,
PROCESS TYPE and AGENCY (cf. Section 6.7.2). This means that every clause in
the language can be analysed from two perspectives in terms of transitivity.
The first perspective is to examine the nature of the process realised by the
clause and the participants participating in it. The different configurations that
accrue from this analysis constitute the system of PROCESS TYPE, and this
differentiation of clauses into grammatically distinct process types is based on
the transitive model of transitivity. The second perspective is to examine how
the process is brought about, whether there is an external agent causing the
process or it is represented as a self-engendered process. This is the ergative
model of transitivity, and the underlining system here is AGENCY. In the
preceding sections, process types have been discussed in detail. Six process
types, material, mental, relational, verbal, behavioural and existential clauses,
are identified as construing different domains of experience. The system of
AGENCY, on the other hand, is a generalisation across the six process types.
Before we discuss this system, however, we will first make some
generalisations about the transitive model. Thus, this section is concerned with
two kinds of generalisations about the system of TRANSITIVITY, one from the
transitive point of view (Section 6.7.1) and the other from the ergative point of
view (Section 6.7.2).
6.7.1 The Transitive Model

The discussion proceeds with the transitive model. Here, the generalisation
across the process types will be made based on two variables. The first is the
ability of the clause to allow an omission of the Complement (which includes
the notion of ‗Object‘), and the second is the ability of the clause to project
another clause. In the discussion on material clauses in Section 6.3, it was
mentioned that the Complement element in the clause may sometimes be
unrealised (see also Chapter 5, Section 5.4.1). This phenomenon was
important in making a distinction between transitive and intransitive clauses
337
and, in particular, in characterising the Goal element. However, it is a motif
that permeates the overall system of process types. The general principle
underlining the omission of the Complement in the Dagaare clause can be
stated as follows:
Principle (1): The Complement is realised if it is new information in the flow

of discourse or undentifiable to the listener, but can be left unrealised if it is
given information or identifiable to the listener (cf. Matthiessen (2004: 640-
642) for a typological discussion on textual statues and implicit elements in
the clause). This principle is illustrated in the following dialogue, a short
version of which was presented in Chapter 5 (cf. Section 5.4.1):

Baba: Ɛcɛ Zɩɛm ɩ, a dãa [[ɩ na dà

kʋ̀ fʋ a]], a wa? Ɛ
give.PFV 2SG JUNC 3PL.NHM be:where And
bɛ dɛ zɩ.
3PL.HM ADV sit.PFV
Zɩɛm: Aa
INTJ

wa ɩ?
come.PFV NAFFR
Naab: A ɩ pɔwbɛ lɛ bɛ mɩ́ ɓaa

DEF 1SG woman.PL COP 3PL.HM HAB cold
zuo.
be:too much
[Shouting to ladies inside the compound yard]: Nɩ cãa
2PL be:still.PFV
bɛ ɩ a dãa sɛr ɛ?
NEG.IND.NFUT make.PFV DEF beer still NAFFR
338
Ziem: ‗I don‘t know.‘
Baba: ‗They are bringing (it) and they have not come?‘
Naab: ‗My wives are the ones who are too slow. [shouting to ladies
inside the compound]: You have not prepared the beer/pito yet?‘
In the first two of the clauses underlined in the extract, the implied
Complement is a dãa (‗beer‘, ‗pito‘), and it is left unstated because it can be
inferred from the initial exchange between Baba and Ziem. The Complement
is, however, introduced in Nɩ cãa bɛ ɩ a dãa sɛr ɛ? (‗You are not done with the
beer yet?‘) although the clause would be equally grammatical without it. The
Table 6.10. Transitivity components in relation to realisation of the

Complement
transitivity feature information status of

Complement
Given/ New/
identifiable unidentifiable
Human  
non- concreteness in concrete clause x 

human
in abstract clause  
[seeing]: phenomenal  
phenomenality macrophenomenal x 
motivation here is a shift in addressee from Baba to the women, who are
distant away in the compound. Its function here is therefore to set a textual
context for a new exchange. The main principle regulating the realisation of
the Complement in the Dagaare clause is, therefore, a textual one, as
339
mentioned earlier. However, there are other factors of an experiential kind that
motivate the retention of the Complement even when it is given information.
These experiential factors are summarised in Table 6.10 in relation to the
textual one (tick = realised; cross = unrealised). As the table shows, the
primary experiential features that motivate the realisation of the Complement
are those of (i) humanness and (ii) concreteness. Specifically, the first
opposition is between whether the nominal group realsing the Complement is
human or non-human, and, if non-human, whether it occurs in a concrete
clause or an abstract clause. The first of these principles is stated below:
Principle (2): The Complement is normally realised if the nominal group

realising it is human.
Thus, the referent of the pronoun ʋ in the following clause will automatically
be assumed by a Dagaare speaker to be human because, as a non-emphatic
pronoun, it is represented as given information and is only retained in the
clause due to its referential value (cf. Chapter 5, Section 5.4.1):
(126) Bɛ waar =ɩ ʋ na.

3PL.HM come.IPFV CAUS 3SG AFFR
‗They are bringing him/her.‘
As was mentioned earlier, despite this pervasive principle in Dagaare, a

human Complement is still sometimes left unrealised. This phenomenon,
however, enacts attitudinal meanings. It is only limited to certain contexts
such as those of linguistic play and emotionally loaded settings (cf. Section
6.3.1). The third principle derives from the second one and it is stated below.
Principle (3): In case the Complement is non-human, it must be realised if the

clause in which it occurs is an abstract clause, otherwise it can be omitted.
This third principle, divides the six process types into two groups, comprising
concrete and abstract clauses as follows:
concrete clauses abstract clauses
concrete material clauses abstract material clauses
behavioural clauses mental clauses
340
verbal clauses existential clauses
relational clauses
[
(intensive identification)
The notions of ‗concrete‘ and ‗abstract‘ in this context are relative

approximations in the sense that a concrete material clause will be more
concrete than a behavioural one. They should also be understood as global
properties of an entire clause rather than properties of the the (potential)
Complement. The abstractness of a clause can either be determined by an
abstract entity functioning as Complement (as in abstract material clauses; as
in example 127) or by the abstract nature of the process itself (as in existential
clauses such as example 128). The Complement in the following clauses are
underlined:
(127) A be wa nɩ nɩ gbɛ-mŋɛ.
DEF there come.PFV CAUS FOC leg-hit
‗That place brought a fault (=We made a mistake with regards to this
issue).‘
(128) Pʋ-pʋla be nɩ be.
good will EXIST.PFV FOC there
‗There is joy.‘
The requirement that the Complement must be realised in abstract clauses

(provided the clause takes a Complement in the first place) explains why
existential clauses such as example (128) still needs be (‗there‘) as a dummy
Complement.
The location of behavioural and verbal clauses under the concrete
category also needs some comments due to their relative abstract nature
compared to concrete material clauses. Two alternative reasons may account
for the omission of the Complement in these clauses. First, they represent a
semantic region between abstract experience and concrete experience in the
sense that they represent conciousness as a form of activity. In this regard, the
grammar of Dagaare seems to push them more towards the concrete pole than
the abstract pole on the variable of the realisation of the Complement.
Alternatively, we can interpret the tendency to omit the Complement in
341
behavioural and verbal clauses as due to the nature of the nominal group that
functions as Complement in these clauses, that is, the Behaviour and Verbiage
participants respectively. As we mentioned earlier, the nominal group realising
these participants is often an elaboration of the process and is ―not so much of
a participant as a self-refinement of the process itself‖ (Halliday &
Matthiessen 2014: 347) (also see Section 6.7.2 below on Range). In this case,
the Complement in these clauses is often an optional element in the clause (see
Sections, 6.6.1 and 6.6.2). Nevertheless, compared to the more inert process
types listed under abstract clauses, behavioural and verbal clauses are
construed by the grammar as relatively concrete.
Further, relational clauses are the most indeterminate category in terms
of classifying processes into concrete and abstract clauses. The only consistent
sub-type in this regard is intensive identifying clauses. These construe static
relations between entities and do not allow an omission of the Complement.
Other sub-types of relational clauses are not consistent in the realisation or
omission of the Complement. Nonetheless, the same principle of concrete
versus abstract clauses is at play in these other sub-types. For example,
possessive clauses that represent concrete possession allows an omission of
the Complement as in (129) while those that represent abstract relation as in
(130) or metaphorically construe emotions as possessions as in (131) generally
do not (Complement is in bold):
(129) A: Fʋ tɛr =ɩ libir bɩ?
2SG possess.PFV FOC money INT
‗Do you have money?‘
B: Ʋʋ, ɩ tɛr =a.
yes 1SG possess.PFV AFFR
A: ‗Do you have money?‘

B: ‗Yes, I have.‘
(130) Ʋ tɛr a na.
3SG possess.PFV 3PL.NHM AFFR
‗S/he is rich.‘
(131) Tɩ tɛr =ɩ pʋ-pɩɛlʋ.
1PL possess.PFV FOC stomach-whiteness
‗We have joy.‘
342
In the underlined clause in (129), it is possible to omit the Complement
because it represents concrete possession. In (130) and (131) such an omission
is not possible if the original sense of the clause is to be maintained.
Again, circumstantial clauses that represent concrete locations of
entities allow an omission of the Complement (132) while those that represent
abstract relations such as identity in terms of place of origin do not (133):
(132) Ɩ pʋɔ na.

1SG be:among AFFR
‗I am among.‘
1SG be:from.PFV FOC INTJ DEF Nandom West
In mental clauses, there is another principle at play in relation to the

phenomenality of the Phenomenon participant, which can be stated as follows:
Principle (4): If the potential Complement in a perceptive clause representing

‗seeing‘ is a macrophenomenal clause (i.e. an act clause), it can be omitted;
otherwise it must be realised (see Section 6.4 on mental clauses).
This principle is illustrated in the constructed dialogues below (relevant clause

is underlined; Complement is in bold):
(134) a. A: Fʋ nyɛ na a lile na ƴaw

2SG see.PFV AFFR DEF bird NMLZ fly
a?
JUNC
B: Ɩ bɛ nyɛ ɩ.
1SG NEG.IND.NFUT see.PFV NAFFR
A: ‗Have you seen the bird fly(away)?‘ / ‗Did you see the
bird fly (away)‘
B: ‗I haven‘t‘ / ‗I didn‘t see.‘
343
b. A: Fʋ nyɛ nɩ a lile?
2SG see.PFV FOC DEF bird
‗Have you seen the bird?‘ ‗Did you see the bird?‘
B: Ɩ bɛ nyɛ ʋ ɩ.
1SG NEG.IND.NFUT see.PFV 3SG NAFFR
‗I haven‘t seen it.‘ / ‗I didn‘t see it.‘

*Ɩ bɛ nyɛ ɩ.
1SG NEG.IND.NFUT see.PFV NAFFR
*‗I haven‘t seen / ‗I didn‘t see it.‘
In clause (134a), where the complement is a nominal clause, the most

acceptable answer is for Speaker B to omit the complement in his response. In
(134b), on the other hand, an omission of the complement in Speaker B‘s
response will render the clause unacceptable since the complement, a lile (‗the
bird‘) is a micro-Thing and the grammar of the language requires it to be
stated.
The principles stated above can be interpreted in relation to Hopper
and Thompson‘s (1980) transitivity hypothesis, which, in itself, can be
conceived of as a generalisation across process types from a typological
perspective. They divide the notion of transitivity into nine parameters with
each component having a high and a low transitivity variable and hypothesise
that categories of high transitivity will co-select each other. Our findings here
generally support this hypothesis in terms of the clustering of process types
under concrete and abstract clauses. As mentioned earlier, material clauses
represent doings-&-happenings and behaviourial and verbal clauses share
these characteristics by representing concious processes as physiological (or
psychological) and verbal activities. These three process types, in relative
terms, are therefore activity-oriented compared to mental, existential and
identifying clauses, which construe inert processes. This characteristic of
material (i.e. concrete), behavioural and verbal clauses classify them as high in
transitivity as opposed to mental, existential and relational clauses, which are
of low transitivity. This relative dichotomy between high and low transitivity
corresponds with how they cluster into concrete and abstract clauses based on
the realisation of the Complement. Based on this single transitivity variable,
344
clauses that are more towards the activity pole cluster together as concrete
clauses, and those that construe inert processes cluster together as abstract
clauses.
However, on face value, it would seem that the Dagaare data do not
support Hopper and Thompson‘s (1980) assumption that clauses with two or
more participants (‗Agent‘ and ‗Object‘) are high in transivity while those
with one participant are low in transitivity since Dagaare allows the omission
of the Complement in concrete clauses but not in abstract clauses. As has been
mentioned, however, this tendency in Dagaare is textually motivated and the
Complement is always implied and recoverable. Nonetheless, this study
reveals that a clause with one participant can be very high in transitivity while
a clause with two participants can be very low in transitivity.
The second variable on which the process types in Dagaare can be
generalised is that of projection. Different process types cluster together based
on their ability or inability to project another clause. For lack of better terms
(particularly for those that do not project), we identify these groupings as
direct representation of experience and symbolic representation:
direct representation symbolic representation

(– projection) (+ projection)
mental clauses (emotive & mental clauses (cognitive &
perceptive) desiderative)
behavioural clauses verbal clauses
relational clauses (others) relational clauses (intensive:
existential clauses necessity/fact-type)
material clauses
The difference in orientation of the two types of clauses can be illustrated by

the following passage (Projecting clauses are underlined):

|||Galile tew dem za a, nyɩ bãw nɩ
Galilee town owners all VOC 2PL know.PFV FOC
a:|| a na-kpɛɛ Sezaare Ogusiri yèl

3.NHM DEF chief-big Ceasar Augustus say. PFV
345
a || kɛ bɛ sɔr a nɩsalbɛ za [[na
AFFR PROJ 3PL.HM count.PFV DEF people all REL
kpɩɛr a Galile nɩ a Zude tew

live.IPFV DEF Galilee and DEF Judah town
pʋɔ a]]; ||| ɛ a fɛr =a || kɛ
inside JUNC PROJ 3PL.NHM be:necessary.PFV AFFR PROJ
nɩr za mɩ̀ lɛb kul a ʋ

person all also return.PFV go:home.PFV DEF 3SG
sãa mɩnɛ tew dɔwra pʋɔ bɛ tɩ sɔr ʋ

fathers PL town birth inside 3PL.HM DIST count 3SG
a be.|||
DEF there
‗All the people of Galilee, you should know this: The big chief Ceasar
Augustus says that they should count all the people living in Galilee
and the state of Judah. That it is also necessary that everybody goes
back home to the birth city of his fathers for them to count him/her
there.‘
In this extract, the non-projecting clauses represent activities (e.g. bɛ sɔra
nɩsalbɛ za, ‗they should count all the people‘; nɩr za mɩ̀ lɛb kul ‗everybody
goes back home‘) and states (e.g. a nɩsalbɛ za na kpɩɛr a Galile … ‗all the
people living in Galilee …‘). These are observable processes that are more
towards our world of experience. The projecting clauses, on the other hand,
are symbolic. They generally represent the interpretation of the other
processes, their organisation as knowledge (as in nyɩ bãw nɩ a, ‗you should
know this‘) and their evidential status (e.g. Sizar Ogusiri yèl a ‗Ceasar
Augustus says‘; a fɛr a, ‗it has become necessary‘). In other words, as has
been demonstrated throughout the discussion of projection across process
types, projecting clauses assess propositions and proposals and intrepret them
through the resources of evidentiality and modality. There is therefore, a clear
division of labour between projecting and non-projecting clauses in discourse
and in the system of language.
346
6.7.2 The Ergative Model
This section proceeds to discuss the ergative model, the second generalisation
across process types, as mentioned above. While the tranisitve model is
concerned with whether or not the process extends and impacts on an entity,
the ergative model is concerned with whether the process is construed as being
self-engendered or brought about by an external causer. Just as the transitive
view of the clause has an underlining system, the system of PROCESS TYPE, the
ergative perspective also has an underlining system, the system of AGENCY
(see Figure 6.8). Before, the categories associated with this system are
examined, however, we will first discuss the participant roles associated with
the ergative perspective of the clause. Four participant roles have been
identified as follows (cf. Halliday & Matthiessen 2014): Medium, Agent,
Beneficiary, and Range.
(i) Medium: This participant is ―the medium through which the process is
actualised‖ and, together with the Process, they form the nucleus of the clause
(Halliday & Matthiessen 2014: 336). In material clauses, the Medium is the
Goal in transitive clauses (136) and the Actor in intransitive (137) and
causative (138) clauses. Thus, a bie (‗the child‘) is the Medium in each of the
following clauses:
(136) Ɩ ŋmɛ =n a bie.

1SG beat.PFV FOC DEF child
‗I beat the child.‘
(137) A bie wa na.
DEF child come.PFV AFFR
‗The child has come.‘

(138) Ɩ wa nɩ nɩ a bie.
1SG come.PFV CAUS FOC DEF child
‗I have caused the child to come (= I have brought the child).‘
In (136), the nominal group a bie (‗the child‘) is the Goal participant in a
transitive clause; in (137) it is Subject and the Actor participant in an
intransitive clause; and, in (138), it is the Actor and Complement in a
causative clause. What is common among the use of a bie (‗the child‘) in all
347
these instances is that, it is the participant that is impacted by the process, the
one through which the process is actulised or brought into being. The Medium
thus subsumes the senses of ‗affected‘ and ‗patient‘ participants that are
commonly used in the linguistic literature. Medium is, however, a more
inclusive participant role and includes processes that are not of the material
type. In mental clauses, the Medium is the Senser; in relational clauses, it is
the Identified, in the identifying mode, and the Carrier, in the attributive mode;
in verbal clauses, it is the Sayer; in behavioural clauses, it is the Behaver; and
in existential clauses, it is the Existent participant (see Table 6.11). The
Medium is thus ―the nodal participant throughout the system … the one that is
critically involved, in some way or other according to the particular nature of
the process‖ (Halliday & Matthiessen 2014: 343).
Unlike in English, however, the Medium can be unrealised in clauses
where it is non-human and is presupposed as given information (cf. Section
6.7.1). Also, in verbless clauses, the Process is lost due to diachronic changes
(cf. Section 6.5), as in the examples below (Medium is in bold):

Zã-Bati nʋ.
John-Baptist IDENT.SG
‗John this Baptist is the one (=He is John the Baptist).‘

Ɩ̃ yuor =ɩ George.
1SG name COP.FOC George
My name is George.
These clauses form an exception in the language so far as the ergative

interpretation of the clause is concerned. As with the copula clauses from
which they originated, the Identified participant is the Medium.
348
Table 6.11. Generalising across participant roles in process types
(ii) Agent: The Agent is the external participant that brings about the process.
In other words, it is the external agency participanting in the nucleus
established by the Process + Medium. In material clauses, the Agent is the
Initiator participant (138) or the Actor in transitive clauses (136). In examples
(136) and (138), for instance, the Subject pronoun Ɩ is the Agent. In (138), Ɩ wa
nɩ nɩ a bie (‗I have brought the child‘, lit. ‗I have caused the child to come‘),
the Agent is represented as the Initiator, the one who brought about the
movement or change in position of the child. This external causation is
morphologically indicated by the causative particle nɩ. In (136), Ɩ ŋmɛ =n a bie
(‗I beat the child‘), the Agent is the one who brought about the process of
beating, the external participant causing the process of which the
Medium/Goal is the undergoer. In mental clauses, the Agent is generally the
Inducer, but, in clauses of the impinging type (see Section 6.4), it is the
Phenomenon, the one who brings about the sensing. In attributive clauses, on
the other hand, it is the Attributor participant (cf. Section 6.5) and in
identifying clauses, it is the Assigner.
(iii) Beneficiary: This participant is associated with benefactive clauses,

comprising material clauses of cliency and recipiency (cf. Section 6.3.2),
verbal clauses (cf. Section 6.6.1), and a sub-type of attributive clauses (cf.
349
Section 6.5). The Beneficiary is the one to whom or for whom the process is
construed as taking place, although the benefit may be represented as a denial
or a negation of it. This participant is typically introduced by the verb kʋ
(‗give‘). In attributive clauses, it occurs in contexts where the process is
relaised by adjectival verb of a positive evaluation (cf. Section 6.5.2.2):

Al bome maal =a kʋ =m.
3PL.EMP.NHM things do:well AFFR give.PFV 1SG.ACC
‗The things (=crops) there did well for me.‘
The Beneficiary does not occur in mental, behavioural and existential clauses
(see Halliday & Matthiessen 2014 on English). Thus, the clause below is,
strictly speaking, a (benefactive) material clause although the Process is
realised by a verb typical of behavioural clauses, bɛl (‗look‘):
(142) Bɛlɛ a bie kʋ =m!

look.IPFV DEF child give.PFV 1SG.ACC
‗Keep an eye on the child for me!‘
The process here is represented as a service done for the speaker rather than
simply an inert physiological behaviour.
(iv) Range: As its name suggests, the Range participant is the element that
specifies the range or scope of the process. Compared with the other
participant roles, Range is less of an entity participating in the process
(Halliday & Matthiessen 2014). It occurs in all six process types (see Table
6.11). In mental clauses of the emanating type, the Range is the Phenomenon
as opposed to the impinging type, where the Phenomenon is Agent. In material
clauses, it is the Scope, Instrument and Accompaniment; in identifying
clauses, it is the Indentifier and in attributive clauses, it is the Attribute; in
behavioural clauses, it is the Behaviour; in verbal clauses, it is the Verbiage an
in existential clauses, it is the Place.
In many instances, the Range participant is represented as a
restatement of the Process itself (Halliday & Matthiessen 2014). That is, it
may be specifying a particular variety of the process, as in (143) below:
350
Mɛ́ nyɩ taa na lɔb zʋkpar

HST 2PL MOD POS.IND.FUT throw.PFV proverb
na kʋ =m […]
DEM give.PFV 1SG.ACC […]
‗I believe you will tell me this proverb …‘
Or, second, this element may be an entity, but it marks the domain of the
Process rather than being an active participant or one that is impacted by the
Process. Examples of this are highlighted in (144) to (146) below:
(144) Workshop interview (abstract material clause, +Scope)

A kɔb pɛr buor =ʋ na tʋɔ
DEF farming type which IDENT.SG POS.IND.FUT be:able.PFV
sow tɩ.
help.PFV 1PL
‗Which type of farming would be able to help us.‘

1SG be:from.PFV FOC INTJ DEF Nandom West
Bɛ za be =n a ka.
3PL.HM all EXIST.PFV FOC DEF here
‗All of them are here.‘
As mentioned earlier participant roles discussed above are a

generalisation across process types and they are engendered by the system of
AGENCY, that is, the opposition between middle and effective clauses. A
system network for AGENCY is presented in Figure 6.8 and we proceed to
discuss middle and effective clauses in the sections below.
351
Figure 6.8. The system of AGENCY in Dagaare
6.7.2.1 Middle Clauses
Middles clauses are those clauses without the feature of agency and thus the
Process + Medium configuration is represented as internally engendered. It
construes a process as a happening as opposed to an action, and the Medium is
always the Subject of the clause. As has been mentioned in Section 6.3.1, two
sub-types of middle clauses can be identified based on the variable of whether
or not the Medium is capable of engendering the process by itself.
(i) Medium as capable of the process: This is the unmarked case of middle
clauses, where, the Medium is construed as acting on itself:
(147) A bie lo na.

DEF child fall. PFV AFFR
‗The child fell.‘
(148) Ɩ kʋɔr =a.
DEF weed.IPFV AFFR
‗I am weeding / I am farming.‘
Here, the grammar represents the effect of the process as limited to the
Medium, it does not go through.
352
(ii) Medium as incapable of the process (non-volitive): This type of middle
clauses are medio-passive, and they are associated with only material clauses
(see Section 6.3.1). 51 Examples are given below:
(149) A wɩɛ bɛ kɔ ɩ.
DEF farm NEG.IND.NFUT weed.PFV NAFFR
‗The farm is not weeded / The farm is not cultivated.‘

(150) A dãa dà na.
DEF beer buy.PFV AFFR
‗The beer has been bought / The beer sold well.‘

(151) A dãa dàar =a.
DEF beer buy.IPFV AFFR
‗The beer is selling (well).‘
Here, the Medium represents an entity that is incapable of engendering the

process by itself. Although, in reality, the process is engendered by an external
participant, the grammar of the language represents it as taking place without
the causer. They have agnate effective counterparts such as Nɩbɛ daar =ɩ a
dãa (‗People are buying the beer‘). Let‘s consider another example in the
dialogue below on a healing miracle of Jesus:

Jesus: Pɔw ɩ, a fʋ baalʋ sanι
woman VOC DEF 2SG sickness be:heal.PFV
na.
AFFR
Voice: Wi! Alɛ! Ʋ sanι na wɛ!

INTJ INTJ 3SG be:heal.PFV AFFR EXCL
Jesus: ‗Woman, your sickness is healed.‘

Voice: ‗What! S/he is healed!‘
The major clauses in this extract construe the healing of the woman as a
happening rather than an action brought by Jesus. The nominal groups a fʋ
baalʋ (‗your sickness‘), in the fist clause, and Ʋ (‗s/he‘), in the second clause,
51
Since Dagaare has no system of voice, this type of middle clauses does not cover the same
semantic region as the ―medio-passive‖ (oe medio-receptive) in English.
353
are construed as the Medium through which the Process is actualised, and the
process of healing itself is represented as unfolding in spite of a causer.
6.7.2.2 Effective Clauses
Effective clauses, on the other hand, are those clauses with the feature of
agency and the Agent is the inherent participant here. Generally, they tend to
be less common in Dagaare discourse, compared to middle clauses, and they
are mostly material clauses. In a sample of about 500 clauses across different
registers (including minor clauses), effective clauses account for only 84
(16.8%) instances, 78 (92.9%) of which are material clauses, while middle
clauses record 360 (72%) instances. 52 In effective clauses, the process is
represented as being caused by an external participant, the Agent. The basic
structural configuration is Agent + Process + (Medium); where the Medium
may be ellipted as given information (cf. Section 6.8.1). In principle, all
transitive material clauses (cf. Section 6.3.1), mental clauses of the impinging
type (cf. Section 6.4) and targeted verbal clauses (see Section 6.6.1) are also
effective clauses. As mentioned earlier, in material clauses, agency may also
be explicitly marked by the particle nɩ. This strategy is, however, restricted to
processes of motion such as wa (‗come‘), cen (‗go‘) zɔ (‗run‘) do (‗climb / go
up‘), siw (‗descend / ‗come down‘) etc. Agency can also be construed
analytically as in the underlined clause in the extract below (also see Section
6.7.2):
(153) NANSU citation

A tɩɩm Nandɔm bibiir […] puore
DEF 1PL.EMP Nandom children […] greet.IPFV
nɩ Naaŋmɩn barka nɩ a ʋ
FOC God gratitude CONJ DEF 3SG
sownu na ɩ ka fʋ tɩ lɩɛbɩ
help ADV do.PFV JUNC 2SG RNG turn.PFV
52
The predominance of ‗material‘ clauses in the ‗effective‘ category again corroborates
Hopper and Thompson‘s (1980) transitivity hypothesis that features of high transitivity co-
select each other.
354
a Cape Coast University wulwulbɛ nɩ-kpɛɛ
DEF Cape Coast University teachers leader
a.
JUNC
‗We the children of Nandom thank God for his assistance that made
you became the president of the University of Cape Coast.‘
Here, the Agent is a ʋ sownu (‗his assistance‘) while the Medium is fʋ (‗you‘)
and a Cape Coast University wulwulbɛ nɩ-kpɛɛ is Range. Causation is
indicated analytically by the general verb ɩ (‗do‘) as a process engaged in by
the Agent.
If we push the account further in delicacy to the lexical end of the
lexicogrammar continuum, we can identify a few verbs that oppose each other
in their unmarked selection of agency type (see Table 6.12). Those verbs that
are typically associated with effective clauses such as sanɩ (be healed) can also
be used ergatively (i.e. in middle clauses), as is illustrated in (152), in which
case they become marked.
Table 6.12. Examples of verbs that typically realise processes in middle

and effective clauses53
No. + middle  effective
verb process type Verb unmarked
process type
[1] ga, ‗lie behavioural gaalɩ; ‗put down material:
down‘ in a lying position‘ transitive
[2] ir, ‗wake behavioural sɩw, ‗cause to material:
up‘ wake up‘ transitive
[3] lo, ‗fall‘ material: lɔb, ‗make fall‘, material:
intransitive ‗drop‘ transitive
[4] kpi, ‗die‘ behavioural kʋ́, ‗kill‘ material:
transitive
[5] sa, ‗get behavioural sanɩ, ‗cause to material:
well‘ heal‘ transitive
[6] zɩ, ‗sit‘ behavioural zɩɩlɩ, ‗make to sit‘ material:
transitive
53
This table is based on a similar one by Bodomo (1997: 91)
355
In addition, we can identify pairs of effective clauses that contrast between
benefactive and non-benefactive type based on specific verbs realising the
(1a) nyu, ‗drink‘ [effective: non-benefactive]

A bie nyu =n buulu
DEF child drink.PFV FOC porridge
Actor material: transitive Goal
Agent Process Medium
‗The child drank porridge.‘
(1b) tuuli, ‗feed liquid (put to mouth)‘ [effective: benefactive]

Ɩ tuuli nɩ a bie buulu
1SG feed:liquid.PFV FOC DEF child porridge
Actor material: transitive Client Goal
Agent Process Beneficiary Medium
‗I feed the child porridge.‘
(2a)dɩ, ‗eat‘ [effective: non-benefactive]

A bie dɩ =n saab
DEF child eat.PFV FOC food type
Actor material: transitive Goal
Agent Process Medium
‗The child ate tuo-zaafi (TZ).‘
(2b) su, ‗feed solid (put to mouth)‘ [effective: benefactive]
Ɩ su =n a bie bʋndɩrɩ
1SG feed:solid.PFV FOC DEF child food.PL
Actor material: transitive Client Goal
Agent Process Beneficiary Medium
‗I feed the child food.‘
Figure 6.9. Benefactive versus non-benefactive effective clauses
356
process, such as nyu (‗drink‘) verus tuuli (‗feed someone something liquid‘)
and dɩ (‗eat‘) versus su (‗feed someone something solid‘) (see Figure 6.9).
6.8 Conclusion
In conclusion, this chapter has examined the major grammatical system for
construing experience in the Dagaare clause, the system of TRANSITIVITY. The
chapter first gave an overview of the clause as a representation of experience
and summarised how this function of language has been theorised in the extant
literature on linguistic science to provide a conceptual background for the
analysis of Dagaare.
The chapter proceeded to discuss one sub-system of transitivity, that of
PROCESS TYPE. As in English and many other languages (cf. Caffarel et al.
2004; Matthiessen 2004), six process types have been identified, comprising
material, mental and relational clauses as the principal types, and behavioural,
verbal and existential clauses as minor types. Material clauses typically
represent outer experience and their principal sub-types are transitive clauses,
which construe doing or action, and intransitive clauses, which construe
happening or events. Mental clauses represent consciousness, comprising
cognition, desideration, emotion and perception. Relational clauses construe
being and having by characterising and identifying entities in the outer world
of experience but also by representing emotion as attributes. The principal
sub-types identified are intensive, possessive and circumstantial clauses, each
of which manifest in two modes, attribution and identification. Behavioural
clauses construe consciousness as outer experience, thereby embodying the
characteristics of material and mental clauses. They are the most indeterminate
process type and shades also into verbal clauses. Verbal clauses in themselves
are clauses of saying and represent consciousness as a kind of relation between
Sayer and what is said. Existential clauses represent that something exists.
Semantically, they are like material clauses by typically representing outer
experience, but grammatically, they are like relational clauses by representing
existence as a relation between the Existent and the Place of existence.
Throughout the discussion, it has been shown that there is clear evidence in
the grammar of Dagaare that differentiates clauses into different process types.
357
The chapter, however, continued to identify principles that are general
across the different process types. First, it has been shown that the various
process types cluster into smaller groupings based on features such as
concreteness and abstractness and their ability to project another clause.
Second, it has been discussed that clauses can be classified as middle or
effective, depending on whether or not they embody the feature of agency.
This latter generalisation is based on the second major sub-system of
transitivity, that of AGENCY. Middle clauses represent the process as being
self-engendered while effective clauses represent it as being caused by an
external participant, the Agent.
The two systems, PROCESS TYPE and AGENCY are underlying systems
of two models of transitivity, namely the transitive model and the ergative
model respectively. The transitive perspective is concerned with the internal
nature of the process, whether it extends beyond the participant engendering it
and impacts on a second participant or not. The ergative model is concerned
with how the process is brought about, whether it is self-engendered or it is
caused by an external agent. The two views give a fuller picture of the
experiential organisation of the language.
358
CHAPTER SEVEN
SYNOPSIS: DAGAARE AND BEYOND
7.1 Introduction
The last four preceding chapters have given a detailed discussion of the
interpersonal, textual and experiential systems of the Dagaare clause. This
chapter will summarise and conclude the study. It will first recap the aims of
the study and the theoretical and methodological procedures adopted (Section
7.2). It will then proceed to summarise the descriptive categories presented in
the preceding chapters (Section 7.3) and finally draw implications for the
description, in terms of theory, research and practical applications (Section
7.4).
7.2 Summary of Aims and Procedures

The general objective of the study has been to contribute to studies in
functional language typology, in general, and systemic functional typology, in
particular. Specifically, the aim is to give a comprehensive description of
clause rank systems of Dagaare, comprising experiential, interpersonal and
textual systems. It has been indicated that research on Dagaare dates back to
the 1930s. Studies on the language can roughly been categorized into three
stages. Early research, dating from the late 1880‘s, was within the context of
language classification and dialect studies, typically based on word lists
collected by European traders and colonial administrators. The second stage of
research begins in the 1950‘s and was led by missionary linguists, whose
interest was to develop written materials for religious teaching. Since the
1980‘s, however, Dagaare has been explored by many professional linguists,
including native speakers, in areas such as dialect variation, phonology and,
more especially, lexicogrammar. These descriptions have provided material
for language education, the development of orthography and bilingual
lexicons. One limitation of these descriptions, however, is that they are not
discourse-based and they reveal little information on the rich meaning
potential of Dagaare. The present study has been designed to fill this lacuna.
359
In order to achieve these objectives, the study used the semiotic map
provided by systemic functional theory as a guide. The cline of instantiation
provides a complementary view of system and text, grounding the description
in discourse. The hierarchy of stratification provides a trinocular vision on
lexicogrammar and the cline of lexicogrammar itself allows for a delicate
description of grammatical features. The spectrum of metafunction broadens
the scope of the study to cover three different modes of meaning and provides
a theoretical blue print for a more systematic analysis. The dimension of rank
guides towards the recognition of clause systems and helps the analysis to
account for the functions of groups and word classes in clause structure. The
dimension of axis provides a powerful analytical tool in the form of the system
network and aids in mapping grammatical meanings (i.e. deep grammar) to
their realisations by grammatical form and/or structure (i.e. surface grammar).
The dimension of semogenesis brings in perspectives from grammaticalisation
and helped clarifying the many fuzzy forms and meanings in the language.
Regarding methodology, the study adopted an ethnographic approach,
following the traditions established by American anthropological linguists,
such as Franz Boas, Edward Sapir, Benjamin Whorf, Dell Hymes and Joseph
Greenberg, and European functionalists, exemplified by Bronislaw
Malinowski, J. R. Firth, and M. A. K. Halliday. The corpus for the study
comprises naturally occurring texts, collected from Dagaare speakers in Ghana
and Burkina Faso. Techniques employed in analyzing the texts include
discourse analysis, the construction of system networks and the development
of paradigms. These were supported by theoretical and typological guidance,
transfer comparison, dialect comparison and interviews with language
consultants.
7.3 Summary of Description

The description spans across four chapters, Chapters 3 to 6. Chapter 3 gave an
overview of the architecture of the Dagaare language. It focused on both
phonology and orthography (or graphology) in the expression plane, and on
lexicogrammar in the content plane. On phonology, the chapter identified four
phonological units, namely tone group, [word], syllable and phoneme and
360
accounted for the major phonological phenomena realised at each rank. In all,
62 phonemes have been identified in the phonemic system, comprising
twenty-nine (29) consonants, eighteen (18) simple vowels and fifteen (15)
diphthongs. It has been indicated that a key feature relevant to all vocalic
sounds is the opposition between [+ATR] and [-ATR] quality. The syllable is
made up of at least one phoneme and it is the domain for the realisation of
lexical tone and nasality. The [word] is composed of at least one syllable and
it is the unit where different kinds of harmony systems are realised. The tone
group is the highest phonological unit and it realises intonation. Regarding
orthography, it has been indicated that Dagaare has two alphabet systems, one
developed in Burkina Faso by the Sous-Commission Nationale du Dagara and
the other developed by the Ghana Alphabet Committee as a general
orthography for indigenous Ghanaian languages.
On grammar, the chapter identified four grammatical units: the clause,
group, word and morpheme, and among these discussed the morpheme, word
and clause. Morphemes have been divided into free and bound, and, with
regards to bound morphemes, two derivational morphemes and a range of
inflectional morphemes were identified. The derivational morphemes are the
locative suffix –mi (or –mɩ) and diminutive suffix –le while the inflectional
morphemes consist of plural suffixes in nouns and aspectual suffixes in verbs.
Also, 35 particles were identified as realizing crucial grammatical meanings
across the three metafunctions of language, interpersonal, ideational and
textual. In addition to particles, eight other word classes were identified,
comprising nouns, verbs adverbs, pronouns, adpositions, determiners,
conjunctions and interjections. Out of these, noun, verb, adverb and pronoun
were examined systematically. Nouns have been sub-classified into count and
non-count and it has been highlighted that plural marking in count nouns
establish a noun class system based on semantic classification, including but
not limited to the features humanness, loaned nouns, kinship and social status,
relational body parts, collectiveness and discreteness. It has also been noted
that Dagaare has a rich sub-class of deverbal nouns. Further, a complex system
of verbal ASPECT has been discussed, based on a range of harmony prosodies.
At clause rank, clauses have been classified, based on their status, whether
they are major or minor, and their freedom, that is whether they are free or
361
bound.
Next, Chapter 4 examined the clause as a unit for enacting
interpersonal meaning, especially in dialogic interaction. It first discussed the
nature of dialogue and the general speech functions that are enacted in verbal
exchanges, namely statement, question, command and offer. It then examined
the interpersonal structure of the Dagaare clause, identifying the Mood base
and Residue as two main components of the clause. Three interpersonal
functions, comprising Subject, Predicator and Negotiator, have been identified
as forming the Mood base and as the most salient elements in enacting the
clause as a unit of exchange. These three elements show mood contrast,
establish the validity of the clause and characterise it as an arguable and
negotiable unit. Other elements such as Vocative and mood Adjuncts augment
these three elements in enacting interpersonal meaning. While mood Adjuncts
are part of the Mood base of the clause, Vocatives are peripheral to the
internal integrity of the clause and thus fall out of both the Mood base and
Residue. The chapter also examined the different mood types that realise the
system of SPEECH FUNCTION. It has been shown that the Negotiator combines
with the Predicator and the Subject to indicate delicate mood types in the
clause. The phenomenon of mood metaphor has also been considered,
exemplifying some of the incongruent realisations of speech functions.
Finally, polarity and modal assessment have been examined. Polarity
in the indicative mood interacts with the ideational system of TENSE, showing,
specifically a contrast between positive future (realised by na), and negative
future (realised by kʋ). Positive non-future has zero-realisation while negative
non-future is realised by the particle bɛ). There are distinctive negative
particles (ta, immediate; and taa, non-immediate) for the imperative, resulting
in a systemic contrast between prohibitive and non-prohibitive sub-types of
imperative clauses. Two modal assessment systems have been discussed. First,
the system of MODALITY comprises ‗probability‘ and ‗desirability‘. Probability
consists of intermediate categories between the positive and negative pole in
propositions while desirability modulates the positive and negative value of
proposals. Both sub-systems of modality are realised by the particles naa
(modal positive), kʋʋ (modal negative) and taa (median modality). The system
of NEGOTIATION, on the other hand, enacts the speaker‘s attitude in the clause
362
by assessing the knowledge claims of propositions and modulating the force of
proposals.
Chapter 5 proceeded to present a system-based account of THEME and
INFORMATION. The thematic structure of the Dagaare clause, as in other
languages, was identified as consisting of Theme and Rheme. Theme has been
defined and identified from a trinocular perspective. Semantically, it is
identified as the local context that serves as the point of departure of the
clause, orienting it to a particular interpretation, in the unfolding text. Within
lexicogrammar, and at clause rank, it is identified as that element which is
given initial prominence in the clause and that is developed by the Rheme.
Below the clause, different forms of realisation have been identified for three
types of Theme: textual, interpersonal and topical Themes. It has also been
noted that topical Themes in Dagaare can be either Absolute or non-Absolute
and, if non-Absolute, they can be unmarked or marked. These different types
of topical Theme are identified based on their status or function in relation to
the transitivity structure of the clause. In addition, although unmarked Theme
is normally the Subject in the modal structure of the clause, in first singular
imperative clauses the Theme is normally the Predicator.
On the other hand, the focus structure of the Dagaare clause is realised
minimally by the New element and it often combines with the Given element,
which is an optional element in the structure. New is identified semantically as
that element which is singled out in the information unit as newsworthy. Three
main focus types have been identified, end focus, contrastive focus and broad
focus. End focus is the default choice for positive declarative and it is realised
by the focus particle nι. Contrastive focus is realised by thematic equatives,
emphatic pronouns and exclusive markers. Broad focus has zero-realisation
and it is the default choice for imperative and negative clauses. It has also
been shown that the domain of the realisation of focus of information is the
information unit rather than the clause although clause and information unit
are co-extensive in the unmarked case.
Finally, Chapter 6 examined the major grammatical system for
construing experience in the clause, the system of TRANSITIVITY, comprising
the sub-systems of PROCESS TYPE and AGENCY. As in English and many other
languages (cf. Caffarel et al. 2004; Matthiessen 2004), six process types have
363
been identified, comprising material, mental and relational clauses as the
principal types, and behavioural, verbal and existential clauses as minor types.
Material clauses typically represent outer experience and their principal sub-
types are transitive clauses, which construe doing or action, and intransitive
clauses, which construe happening or events. Mental clauses represent
consciousness, comprising cognition, desideration, emotion and perception.
Relational clauses construe being-&-having by characterising and identifying
entities in the outer world of experience but also by representing emotion as
attributes. The principal sub-types identified are intensive, possessive and
circumstantial clauses, each of which manifest in two modes, attribution and
identification. Behavioural clauses construe consciousness as outer
experience, thereby embodying the characteristics of material and mental
clauses. They are the most indeterminate process type and shades also into
verbal clauses. Verbal clauses in themselves are clauses of saying and
represent consciousness as a kind of relation between Sayer and what is said.
Existential clauses construe something as exising. Semantically, they are like
material clauses by typically representing outer experience, but grammatically,
they are like relational clauses by representing existence as a relation between
the Existent and the Place of existence. Throughout the discussion, it has been
shown that there is clear evidence in the grammar of Dagaare that divides
clauses into these different process types.
The chapter went on to identify principles that are general across the
different process types. First, it shows that the various process types cluster
into smaller groupings based on features such as concreteness and abstractness
and their ability to project another clause. Second, clauses can be classified as
middle or effective, depending on whether or not they embody the feature of
‗agency‘. Middle clauses represent the process as being self-engendered while
effective clauses represent it as being caused by an external participant, the
Agent.
The two systems, PROCESS TYPE and AGENCY are underlying systems
of two models of transitivity, namely the transitive model and the ergative
model respectively. The transitive perspective is concerned with the internal
nature of the process, whether it extends beyond the participant engendering it
and impacts on a second participant or not. The ergative model is concerned
364
with how the process is brought about, whether it is self-engendered or it is
caused by an external agent.
7.4. Implications of the Study

The study makes contributions to research into linguistic theory and
description, and has potential for application in various professional contexts.
7.4.1 Theoretical Contributions

First, the study contributes to the interface between theory and language
description. It is one of the few comprehensive descriptions of language that
are systematically based on a metatheory of language. The advantage is that it
contributes directly to theoretical issues on language while, at the same time,
describing Dagaare in its own terms. Thus, the study narrows the dichotomy
between theoretical and descriptive linguistics that has unfortunately
characterised modern linguistics. Following the introduction of generative
linguistics, there has been a proliferation of linguistic theories since the 1960s.
Many of these theories are, however, not suited for comprehensive description
of languages in their own right, that is, if they are suited for description at all
(cf. Haspelmath 2009a). While generative theories are essentially theories for
theory‘s sake, functional theories such as Functional Discourse Grammar
(Hengeveld & Mackenzie 2008) and Role and Reference Grammar (Van Valin
2000, 2007), respond to theoretical questions in generative theory and
provides descriptive frameworks that are constraining to account for the rich
complexity, fuzziness, and variability of language (for details, see Matthiessen
& Nesbitt 1996; Haspelmath 2009a). The result is that descriptive grammars
are often couched as theory-free accounts of language. This situation often
creates a dichotomy between theories in want of data and descriptions in want
of theoretical explanation. The present study, on the other hand, places the
description of language within a general metatheory of language and shows
how theory and description can enrich each other. While it adopts the general
theory of language systemic linguistics offers, it makes direct connections
with other functional perspectives on language such as speech act theory (cf.
Chapter 4), Fillmore‘s (1968, 1977) case theory, and Hopper and Thompson‘s
365
(1980) transitivity hypothesis (cf. Chapter 6). The study shares this
characteristic with previous descriptions of language based on systemic
functional theory (e.g. Matthiessen 1995; Halliday & Matthiessen 2014;
Caffarel 2006; Teruya 2007).
7.4.2 Contributions to Research
The study also makes some contributions to research. First, it contributes to

studies in functional language typology, in general, and systemic functional
typology, in particular. Language typology and language description have
always moved hand in hand (cf. Caffarel, Martin & Matthiessen 2004;
Haspelmath 2010). As noted in Chapter 2, language description provides
reference materials for typological generalisations and typology, which, in
turn, provide guidance to the description of new languages. By implication,
whenever a new language is described, it contributes to the ongoing
investigation of the regularities and generalities in languages. The present
study particularly used theoretical and typological guidance as well as transfer
comparison in investigating Dagaare data. Thus, although the categories
described are defined relative to Dagaare, the description is sensitive to
typological accounts of language and can easily be used as a reference
material for typological and comparative research. Again, it shares this
characteristic with and adds to previous systemic description on languages
such as English (e.g. Halliday & Matthiessen 2014), Japanese (e.g. Teruya
1998; 2007), O̩ ko̩ (Akerejola 2005), French (e.g. Caffarel 2006), Chinese (e.g.
Tam 2004; Li 2007), Bajjika (Kumar 2009) and Spanish (e.g. Lavid et al.
2010; Quiroz 2013), just to mention a few. One advantage of developing a
description based on the dimensions of systemic functional theory is that the
description can be easily compared systematically with these similar previous
studies for typological generalisations.
In addition, the study contributes to studies on African linguistics.
Since the 1960s, African languages have featured prominently in linguistic
theories and studies in language typology. Joseph Greenberg worked
extensively with African languages in his language classification project (e.g.
Greenberg 1955, 1963), subsequently leading to many related historical and
366
comparative linguistic studies (e.g. Swadesh et al. 1966; Givón 1971, 1975b;
1979). In recent decades, African languages have also featured prominently in
grammaticalisation research (e.g. Heine, Claudi & Hünnemeyer 1991; Heine
2011) and studies on linguistic notions such as information structure (cf.
Güldemann 2015 for a review) and serial verb constructions (cf. Haspelmath
2016). However, many African languages remain either undescribed or
partially described, and there is still need for comprehensive grammars (Heine
& Nurse 2000, Ch. 1). Among Niger-Congo languages, including Dagaare,
only the Bantu is well described. Given that the present study is the most
comprehensive description of Dagaare, it fills a modest gap in the ongoing
study of African languages. In addition, by deploying systemic functional
theory, it adds a new dimension to the many studies on African languages. For
instance, it brings a new perspective to issues such as information structure
(cf. Chapter 5) and accounts for many topics that have previously not been
analysed systematically in African languages, particularly, those of the niger-
congo phylum (but see, for instance, Akerejola (2005)). These include mainly
interpersonal systems of the clause such as MOOD and NEGOTIATION and the
ideational systems of PROCESS TYPE and AGENCY. The description of these and
other systems takes the account from above, in the semantics, showing how
they fit into the overall system of the language, how they interact with one
another, and showing the fuzzy boundaries between the meanings they realise.
This way, the study avoids the fragmentation of categories and greatly reduces
the danger of inaccuracies that are characteristic of the description of many
African languages (cf. Heine & Nurse 2000). The study is therefore a model
for subsequent description of other African languages.
Another area where the study contributes to African linguistics and
Dagaare studies, in particular, is its systematic analysis of discourse data. For
studies on Dagaare, this approach is unique and new. By taking an
ethnographic and a text-based approach, the study provides an account of the
language that resonates with the spirit of the Dagara society, an account that
shows how the language works in the folk life of its speech fellowship. It
maps out how the language is organised to perform various functions, as a
resource for negotiation and interaction, for construing reality and
consciousness and for organising text for semiotic processing. The findings of
367
the study will also be useful to scholars and students of language typology,
anthropology, sociology and ethnolinguistics who are interested in the nature
of African societies. Besides, it provides a metalanguage for talking about
Dagaare. Categories such as negotiation particles, identifying pronouns,
adjectival nouns, adjectival verbs, adverbial particles, just to mention a few,
have been identified together with the different functions they perform in the
clause. These are helpful metalanguage for analysing aspects of Dagaare that
have been elusive in previous scholarly engagements with the language. They
will also give non-native scholars a clearer picture of the organisation of the
language.
7.4.3 Practical Applications
The study is also relevant for language education, discourse analysis for
specific purposes and to scholars interested in modelling natural language
computationally. First, the description is oriented towards supporting language
education. In both Ghana and Burkina Faso, Dagaare is one of the indigenous
languages selected for mother tongue education in basic schools. It is also
studied as an elective subject in senior high schools, colleges and universities.
However, one challenge of Dagaare education, especially at higher education
level, is the absence of a comprehensive grammar. In addition, since almost all
students of Dagaare are native speakers, there is the quest for a practical and
advanced grammar that would motivate and stimulate interest in the subject.
The absence of such a material has resulted in a continual disinterest in the
subject and a reduction in enrolment for Dagaare studies in universities. The
description in this study directly provides material for teachers and students
for studying Dagaare and engaging with it in advanced scholarship. It can also
be adapted by curriculum developers for teaching the language at pre-
university levels. For example, it will be interesting for students to analyse
how resources such as attitudinal negotiation, focus of information and
transitivity contribute to meaning and the aesthetics of different Dagaare
folklore such as folktales, dirges and praise songs.
It will also give impetus to discourse analysis in various critical
contexts in the Dagaare society, including healthcare, media, folk culture,
368
forensics, and religion. For instance, one challenge of educated Ghanaians and
professionals is the ability to translate their academic knowledge on topics
such as health, agriculture, family planning, micro-business, just to mention a
few, to the large uneducated population they serve during interactions at
workshops, on radio and other communicative contexts. The solution to this
problem must involve linguistics and the role of linguists here is to analyse
texts in these contexts to identify patterns of both effective communication
and failed communication for language training. The present study has
provided a grammar for a systematic analysis of this kind. It is itself based on
a systematic analysis of texts in their social contexts and it identifies
grammatical and semantic regularities across these texts that are appliable to
discourse analysis.
The study is useful to scholars and professionals interested in modeling
natural language in computational contexts. Since the 1980s systemic
descriptions have been implemented in computational text generation and
natural language processing. Examples of these are the the PENMAN project,
including its NIGEL grammar, developed by Bill Mann and Christian
Matthiessen, the KPML system by John Bateman and the COMMUNAL
project by Robin Fawcett, Gordon Tucker and their team of researchers (cf.
Henrici 1981; Matthiessen & Bateman 1991; O‘Donnell & Bateman 2005).
Currently, a multilingual natural language generation project, KPML, is being
developed in the University of Bremen in Germany. 54 It involves a variety of
languages that have been described in systemic functional terms, including
English, German, Dutch, Chinese, Spanish, Russian, Bulgarian, and Czech.
The present study can be applied in this and similar computational contexts.
7.5 Final Remarks
Despite the the contributions outlined above for the study, there are many gaps
in the description that need to be explored in further research. The first is the
detailed account of the phonological and grammatical units outlined in
Chapter 3. Due to time limitations, word rank resources such as conjunctions,
54
For details on the KPML project, see http://www.fb10.uni-
bremen.de/anglistik/langpro/kpml/README.html
369
determiners, adpositions and interjections have not been discussed in any
detail and neither have classes of group rank. Also, the account of transitivity
does not include CIRCUMSTANTIATION. Other aspects that need further research
include experiential deixis such as tense and directionality, compounding in
the nominal group, complex constructions such as verbal group complexes
(‗serial verb constructions‘) and clause complexes. Table 7.1 summarises the
resources of Dagaare in a function-rank matrix, showing gaps and as well as
areas that have been discussed in this study. It serves as a blue print for a long-
term project on the meaning potential of Dagaare. This description would
serve its purpose if it sparks further scholarly discussions on the language.
370
Table 7.1 Function-rank matrix for Dagaare Lexicogrammar
metafunction Ideational interpersonal textual
rank class Logical experiential
clause/ TAXIS; LOGICO- TRANSITIVITY: MOOD; MOOD ADJUNCT THEME CONJUNCTION;
SEMANTIC [PROCESS TYPE/ NEGOTIATION; VOCATIVE (Chapter 5) ELLIPSIS/
TYPE AGENCY] (Chapter 4) SUBSTITUTION
(Chapter 6)
1 2 3 …;  (Agent) + Process (+Medium) Mood (Subject • Predicator • Negotiator Theme ^ Rheme Adjunct
(+Range) (+Beneficiary) • Adjunct [modal]) + Residue [conjunctive]
(+circumstances) (Complement, Adjunct [circumstantial]) (Chapter 4)
+ Vocative
information INFORMATION
FOCUS (Chapter 5)
unit
Given + New
group nominal LOGICO- MODIFICATION; THING TYPE PERSON (Chapter 3) DETERMINATION ELLIPSIS/
SEMANTIC COMPOUNDING SUBSTITUTION;
TYPE REFERENCE
emphatic & non-emphatic personal

pronouns
verbal TAXIS; LOGICO- EVENT TYPE; TENSE; POLARITY, MODALITY
SEMANTIC EVENTUALITY; AGENTIVITY; (Chapter 4)
TYPE DIRECTIONALITY
1 2 3 …;  (Finite) + (Orientation) + Event + (Finite) + (Orientation) + Event +
(Agentive) (Agentive)
adverbial LOGICO- MODIFICATION CIRCUMSTANCE TYPE MODAL ASSESSMENT (i.e. MODAL
SEMANTIC (Chapter 3) ADJUCT TYPE)
TYPE
word noun DERIVATION NUMBER, DENO- CONNOTATION LEXICAL
(Chapter 3) NOMINAL-ISATION, TATION COHESION
CLASS (Chapter 3)
verb ASPECT (Chapter 3)

adverbs DERIVATION
(Chapter 3)
conjunction CONJUNCTION
TYPE
complexes simplexes
371
APPENDIX
This appendix illustrates the close systemic analysis of selected texts. It is

organised into three main sections, reflecting the steps involved in the
analysis, namely transcription (step 1), clause chunking (step 2), and tabulated
systemic analysis in Excel spread sheet (step 3). Morpheme-by-morpheme
interlinear glossing and English translation has been provided in Section 2 to
guide readers. The numbering in the clause chunking are the same as those
used in the tables in Section 3. Each number stands for a ranking clause as
opposed to a downranked (or embedded) clause (i.e. clauses functioning as
constituents within other clauses). Downranked clauses are indicated by
double square brackets [[ …]]. For clause complexes, levels of numbering
(e.g. 1.1, 1.2, 1.3, etc.) have been used to indicate that they belong to the same
complex.
1. Text – Recommending: promoting, a concert advertisement

[spoken]
Intervention de BEYOUONE SOMDA par rapport au concert du samedi 05

mars 2016 à 20h au théâtre de l' amitié à bobo. ESCOM DISTRIBUTIONS.
―BEYOUONE SOMDA‘S talk about the concert held Saturday 5 March 2016
at 8: 00 pm at the Théâtre de l' Amitié [Theatre of Friendship] in Bobo [a city
in Burkina Faso]. ESCOM DISTRIBUTIONS.‖ [Recorded February 6, 2016]
In this text, the speaker advertises a concert that aims to showcase Dagara
music and culture. The singer and his team have been committed to the
promotion of the Dagara culture. Much of the advertisement below focuses on
promoting the singer and motivating the Dagara youth who have migrated
from their traditional homeland to attend the concert and learn their tradition.
The advertisement had been video recorded and uploaded on Facebook
(French words are in single quotation marks).
Ɩ na puori a tɩ yɛ-bɛrɛ bɛl na ar a sãakʋmʋ yéle puor-tub bɛ bɔbr kɛ a

sãakʋmʋ bɔr a. Bɛ kpɛɛ za mɩ be n= a Guo. Ɛ ƴɛrɛ ŋa a, tɩ ɩ =n nɩ-yawa na
372
bɔbr kɛ tɩ tɩɛ taar tʋɔ maalɩ yél kaw a. A za kɩ yɛ-kpɛɛ lɛ yi na a na ɩ a tɩ kpɛɛ a.
Tɩɩm za kpɛɛ nʋ. Ʋ yi na ƴaw pʋɔ a tɩ ƴaw ɛ nɩbɛ bɛ nye u e [nyɛ ʋ ɛ], bɛ bɛ
bãw ʋ ɛ. Bɛ tu =n sɔr ɛ a nɩbɛ bɛmɩnɛ bɛ mɩ nye u [nyɛ ʋ]. Ŋmɩŋmɩn n= ʋ ɩrɛ
bɩɩ bʋʋ nʋ ʋ mɩ ɩ ɛ yiele. Tɩ za na tu sɔr a, ɛ tɩ tuori taar a be, „à Bobo, a cinq
mars‟. „Sàmédí‟ bibir ɩ a lɛ; „au théā̂tre de l‟amitié ‟, Vɛntɛɛr vuo pʋɔ. Ɩ bʋɔlɛ
n= a nɩbɛ nɛ za na ɩ a tɩ ba-taabɛ a nɩ a tɩ yɛbr tɩ sãa kʋm-mɩnɛ, tɩ makʋm-
mɩnɛ, a tɩ sãa mɩnɛ. Bɛ tuori tɩ a be a yél taa wa ɩ zawla ɩ. Tɩmɛ b‟a [bɛ wa] ɩ
tɩɩ na na sow taar a, nɩr kʋ tʋɔ wa sow tɩ ɛ. Bɛ tuori tɩ a be tɩ maal a yél nɛ ʋ tɔl
nɩ ƴa-ɓaarʋ ɛ nʋ̀ɔ kpɛ nɩbɛ nɛ za na na wa a be wa nyɛ a bon nɛ tɩ na ɩrɛ a. Tɩ
na maalɩ nɔ-ɓaan yéle kʋ̀ taar nɩ pʋ-pɩɛlʋ. Tɩ na ɩ sãa-been bibiir ɛ wa faa ta
nɩbɛ a ŋmɩn yèl kɛ tɩ tɩɛ taar a, pʋ-pʋla wa be be a, tɩ bãw kɛ a yél ŋa a, ʋ na
tɔl ɩ ni-daa ɛ a tɩ bibiir bɛl na be a mʋɔ pʋɔ a, bɛ ya mɩ na lɛ lɩɛbɩ na. Ah, ɩ
puore bɛ nɩ barka bɛ na na saw sɔw a lɛ tɩ na yèl a.
2. Clause chunking
[1] Ɩ na puori a tɩ yɛ-bɛrɛ bɛl

1SG POS.IND.FUT greet.PFV DEF 1PL sibling-big DEM
[[na ar a sãakʋmʋ yéle puor-tub ||

REL stand.PFV DEF tradition matters back-follow
bɛ bɔbr kɛ a sãakʋmʋ bɔr
NEG.IND.NFUT want.IPFV PROJ DEF tradition lose.PFV
a]].
JUNC
‗I will greet those our elder siblings [[who have stood for the
continuity of tradition]].‘
[2] Bɛ kpɛɛ za mɩ̀ be n= a Guo.
3PL elder all ADV be:at.PFV FOC DEF Guo
‗The eldest of them all too is in Guo.‘
[3] Ɛ ƴɛrɛ ŋa a, tɩ ɩ =n nɩ-yawa
and now JUNC 1PL COP.PFV FOC people-many
[[na bɔbr || kɛ tɩ tɩɛ taar ||
REL want.PFV PROJ 1PL push.PFV one another
tʋɔ maalɩ yél kaw a]].
be:able make.PFV matter some JUNC
373
‗And now, we are many people [[who want to push one another to be
able to do something]].‘
[4] A za kɩ yɛ-kpɛɛ lɛ yi na
DEF all PRT sibling-big ADV be:from PRT
[[a na ɩ a tɩ kpɛɛ a]].

3PL.NHM ADVLZ COP.PFV DEF 1PL elder JUNC
‗The most important thing too comes from [[the fact that it (he) is our
elder (sibling)]].‘
[5] Tɩɩm za kpɛɛ nʋ.
1PL.EMP all elder IDENT.SG
‗He is an elder sibling of us all.‘

[6.1] Ʋ yi na ƴaw pʋɔ a tɩ
3SG come out AFFR put.PFV be:among DEF 1PL
ƴaw
body
‗He has come out among us‘
[6.2] ɛ nɩbɛ bɛ nyɛ ʋ ɛ,
and people NEG.IND.NFUT see.PFV 3SG NAFFR
‗and people haven‘t seen him‘

[6.3] bɛ bɛ bãw ʋ ɛ.
3PL NEG.IND.NFUT know.PFV 3SG NAFFR
‗they don‘t know him.‘

[7.1] Bɛ tu =n sɔr
3PL travel.PFV FOC road
‗They have travelled‘
[7.2] ɛ a nɩbɛ bɛ-mɩnɛ bɛ mɩ̀ nyɛ ʋ,
so that DEF people 3PL-some 3PL ADV see.PFV 3SG
‗so that some of the people, they too (will) see him‘
[8.1] ŋmɩŋmɩn n= [8.2] ʋ ɩrɛ
how IDENT.PL 3SG do.IPFV
‗how is he doing‘
[8.3] bɩɩ bʋʋ nʋ [8.4] ʋ mɩ́ ɩ
or what IDENT.SG 3SG HAB do.PFV
‗or what is it that he does‘
374
[8.5] ɛ yiele.
and sing.IPFV
‗and sings.‘
[9] Tɩ za [[na tu sɔr a]], ɛ tɩ
1PL all REL travel.PFV road JUNC PROJ 1PL
tuori taar a be, ‗à Bobo, à

meet.PFV one another DEF there à Bobo à
cinq mars‟.
cinq mars
‗All of us [[who have sojourned]], that we should meet one another
there, at Bobo, 5 March.‘
[10] ‗Sàmédí‟ bibir =ɩ a lɛ;
Sàmédí day FOC.COP DEF DEM
‗That is Saturday‘
[11] ‗au théā̂tre de l‟amitié‟, Vɛntɛɛr vuo pʋɔ.
au théā̂tre de l‟amitié Venteer area inside
‗at the Theatre of Friendship. Around Venteer area.‘
[12] Ɩ bʋɔlɛ n= a nɩbɛ nɛ za [[na
1SG call.IPFV FOC DEF people DEM all REL
ɩ a tɩ ba-taabɛ]] a nɩ a tɩ
COP.PFV DEF 1PL friend-mates DEF and DEF 1PL
yɛbr tɩ sãakʋm-mɩnɛ, tɩ makʋm-mɩnɛ,

brothers 1PL grandfather-PL 1PL grandmother-PL
a tɩ sãa-mɩnɛ.
DEF 1PL father-PL
‗I am inviting all those people [[who are our mates]]; and our brothers,
our grandfathers; our grandmothers; our fathers. 55
[13.1] Bɛ tuori tɩ a be
3PL meet.PFV 1PL DEF there
‗They should meet us there.‘
55
‗Grandfathers‘ and ‗brothers‘ in this example have the same referent. In Dagara social
structure, one‘s grandfather is also one‘s ‗brother‘. The implication is that grandsons can
inherent grandfathers.
375
[13.2] a yél taa wa ɩ zawla ɩ.
DEF matter NEG EVT COP.PFV empty NAFFR
‗so that the matter (= goal) will not be useless.‘

[14.1] Tɩmɛ bɛ wa ɩ [14.2] tɩɩ na
1PL.EMP NEG.IND.NFUT COND do.PFV 1PL.EMP IDENT.PL
na sow taar a,
POS.IND.FUT help.PFV one another JUNC
‗If we do not try to help one another.‘

Lit. ‗If we do not make (it possible such that) we help one another‘
[14.3] nɩr kʋ tʋɔ wa sow tɩ
person NEG.IND.FUT be:able come.PFV help.PFV 1PL
ɛ.
NAFFR
‗somebody cannot come and help us.‘

[15.1] Bɛ tuori tɩ a be
1PL meet.PFV 1PL DEF there
‗They should meet us there‘
[15.2] tɩ maal a yél nɛ
1PL make.PFV DEF matter DEM
(so that) we make that matter‘

[15.3] ʋ tɔl nɩ ƴa-ɓaarʋ
3SG pass.PFV COM body-cool
‗for it to pass with peace.‘
[15.4] ɛ nʋ̀ɔ kpɛ nɩbɛ nɛ za [[na na
and joy enter people DEM all REL POS.IND.FUT
wa a be || wa nyɛ a bom
come.PFV DEF there PROX see.PFV DEF thing
nɛ tɩ [[na ɩrɛ a]]]].
DEM 1PL REL do.IPFV JUNC
‗and for all those people [[who will come there to see that thing which
we are doing]] to be happy.‘
[16] Tɩ na maal =ɩ nɔ-ɓaan yéle
1PL POS.IND.FUT make.PFV FOC mouth-cold matters
376
kʋ̀ taar nɩ pʋ-pɩɛlʋ.
give.PFV one another COM stomach-whiteness
‗We will do wonderful things for one another with joy.‘
[17.1] Tɩ na ɩ sãa-been bibiir
1PL ADVL COP.PFV father-one children
‗Since we are descendants of one father‘
[17.2] ɛ wa faa ta nɩbɛ a ŋmɩn
and EVT ADV reach.PFV people DEF how much
‗and have now reached a considerable number of people‘
[17.3] yèl kɛ [17.4] tɩ tɩɛ taar a,
say. PFV PROJ 1PL push.PFV one another JUNC
(and) say we should push one another‘

[17.5] pʋ-pʋla wa be be a,
stomach-white COND EXIST.PFV there JUNC
‗if there is good-will‘

[17.6] tɩ bãw kɛ [17.7] a yél ŋa a,
1PL know.PFV PROJ DEF matter DEM JUNC
ʋ na tɔl =ɩ ni-daa
3SG POS.IND.FUT pass.PFV FOC face-ahead
‗we know that this matter, it will move forward‘ (= ‗… it will achieve
greater heights‘).
[17.8] ɛ a tɩ bibiir bɛl [[na be a
and DEF 1PL children DEM REL EXIST DEF
mʋɔ pʋɔ a, bɛ ya mɩ̀ na lɛ

bush inside JUNC 3PL mind ADV POS.IND.FUT ADV
lɩɛbɩ na.
change AFFR
‗and those our children [[who have sojourned]], their minds too will
change again‘
[18.1] Ah, ɩ puore bɛ nɩ barka
INTJ 1SG greet.PFV 1PL FOC thanks
Ah, I thank them‘
377
[18.2] bɛ na na saw sɔw a lɛ tɩ
1PL ADVLZ POS.IND.FUT agree respond.PFV DEF DEM 1PL
[[na yèl a]].

REL say. PFV JUNC
for the fact that they will respond to that [[which we have said]].‘
378
3. Analysis
3.1 Tabulated Polysystemic Analysis

# THEME Theme MOOD DEITICITY ASPECT POLARITY Mood PROCESS Clause
SELECTION (TELICITY) TYPE (ranking)
(TOPICAL)
Textual interpersonal topical Subj. Pred. Nego.
[1] unmarked Ɩ declarative: temporal: perfective positive Ɩ puori verbal Ɩ na puori a
affirmative future tɩ yɛ-bɛrɛ
bɛl [[na ar
a sãakʋmʋ
yéle puor-
tub || bɛ
bɔbr kɛ a
sãakʋmʋ
bɔr a]].
[2] unmarked Bɛ kpɛɛ declarative: temporal: perfective positive Bɛ be relational: Bɛ kpɛɛ za
za affirmative non-future kpɛɛ za identifying mɩ be n= a
Gǔo.
[3] unmarked Ɛ ƴɛrɛ tɩ declarative: temporal: perfective positive tɩ ɩ relational: Ɛ ƴɛrɛ ŋa a,

ŋa a affirmative non-future attributive tɩ ɩ =n nɩ-
yawa [[na
bɔbr || kɛ tɩ
tɩɛ taar ||
tʋɔ maalɩ
yél kaw
a]].
[4] unmarked A za kɩ declarative: temporal: perfective positive A za kɩ yi relational: A za kɩ yɛ-
yɛ-kpɛɛ affirmative non-future yɛ-kpɛɛ identifying kpɛɛ lɛ yi
na [[a na ɩ
a tɩ kpɛɛ
379
a]].
[5] unmarked Tɩɩm za declarative: temporal: perfective positive Tɩɩm nʋ relational: Tɩɩm za
affirmative non-future za identifying kpɛɛ nʋ.
[6.1] unmarked Ʋ declarative: temporal: perfective positive Ʋ yi … na relational: Ʋ yi na

affirmative non-future ƴaw identifying ƴaw pʋɔ a
tɩ ƴaw
[6.2] unmarked ɛ nɩbɛ declarative: temporal: perfective negative nɩbɛ bɛ ɛ mental: ɛ nɩbɛ bɛ
non- non-future nyɛ perceptive nyɛ ʋ ɛ,
affirmative
[6.3] unmarked bɛ declarative: temporal: perfective negative bɛ bɛ ɛ mental: bɛ bɛ bãw ʋ
non- non-future bãw cognitive ɛ.
affirmative
[7.1] unmarked bɛ declarative: temporal: perfective positive bɛ tu material: Bɛ tu =n
affirmative non-future ranged sɔr
unmarked ɛ nɩbɛ declarative: temporal: perfective positive bɛ nyɛ mental: ɛ a nɩbɛ
[7.2] affirmative non-future perceptive bɛmɩnɛ bɛ
mɩ nyɛ ʋ.
[8.1] marked Ŋmɩŋmɩn n= Ŋmɩŋmɩn interrogative: temporal: imperfective positive ʋ ɩrɛ material: Ŋmɩŋmɩn
n= elemental non-future intransitive n= [8.2] ʋ
ɩrɛ
[8.3] marked bɩɩ bʋʋ nʋ bʋʋ nʋ interrogative: temporal: perfective positive ʋ ɩ material: bɩɩ bʋʋ nʋ
elemental habitual transitive [8.4] ʋ mɩ ɩ
[8.5] ɛ bound temporal: imperfective positive yiele behavioural ɛ yiele.
non-future
380
[9] Absolute Tɩ za imperative perfective positive tɩ tuori material: Tɩ za [[na
[[na tu ranged tu sɔr a]], ɛ
sɔr a]], tɩ tuori taar
a be, à
bobo, a san
maas
<<cinq
mars>>.
[10] unmarked Sàmédí declarative: temporal: positive Sàmédí =ɩ relational: Sàmédí
bibir affirmative non-future bibir identifying bibir =ɩ a
lɛ;
[11] minor <<au
théâtre de
l‟amitié>>,
Vɛntɛɛr vuo
pʋɔ.
[12] unmarked Ɩ declarative: temporal: imperfective positive Ɩ bʋɔlɛ material: Ɩ bʋɔlɛ n=

affirmative non-future ranged a nɩbɛ nɛ za
[[na ɩ a tɩ
ba-taabɛ]]
a nɩ a tɩ
yɛbr tɩ
sãakʋm-
mɩnɛ, tɩ
makʋm-
mɩnɛ, a tɩ
sãamɩnɛ.
[13.1] unmarked Bɛ imperative perfective positive Bɛ tuori material: Bɛ tuori tɩ
ranged a be
[13.2] unmarked a yél declarative: modality: perfective negative a yél taa relational: a yél taa
non- desirability wa ɩ attributive wa ɩ zawla
381
affirmative ɩ.
[14.1] unmarked Tɩmɛ declarative: temporal: perfective negative Tɩmɛ bɛ wa material: Tɩmɛ bɛ wa
non- non-future ɩ intransitive ɩ
affirmative
[14.2] unmarked tɩɩ na declarative: temporal: perfective positive tɩɩ na material: tɩɩ na na
affirmative future sow ranged sow taar a,
[14.3] unmarked nɩr declarative: temporal: perfective negative nɩr kʋ ɛ material: nɩr kʋ tʋɔ
non- future tʋɔ ranged wa sow tɩ ɛ.
affirmative wa
sow
[15.1] unmarked Bɛ imperative: perfective positive Bɛ tuori material: Bɛ tuori tɩ
affirmative ranged a be
[15.2] unmarked tɩ imperative perfective positive tɩ maal material: tɩ maal a
transitive yél nɛ
[15.3] unmarked ʋ bound perfective positive ʋ tɔl material: ʋ tɔl nɩ ƴa-
intransitive ɓaarʋ
[15.4] ɛ nʋ̀ɔ bound perfective positive nʋ̀ɔ kpɛ mental: ɛ nʋ̀ɔ kpɛ
emotive nɩbɛ nɛ za
[[na na wa
a be || wa
nyɛ a bon
nɛ tɩ [[na
ɩrɛ a]]]].
[16] unmarked Tɩ declarative: temporal: perfective positive Tɩ na material: Tɩ na maal
affirmative future maal transitive =ɩ nɔ-ɓaan
… kʋ yéle
kʋ taar nɩ
pʋ-pɩɛlʋ.
[17.1] unmarked Tɩ bound temporal: perfective positive Tɩ ɩ relational: Tɩ na ɩ sãa-
non-future attributive been bibiir
[17.2] unmarked ɛ [Tɩ] bound temporal: perfective positive [Tɩ] wa relational: ɛ wa faa ta
non-future faa ta attributive nɩbɛ a ŋmɩn
382
[17.3] unmarked [Tɩ] bound temporal: perfective positive [Tɩ] yèl verbal yèl kɛ
non-future
[17.4] unmarked [tɩ] bound [tɩ] tɩɛ material: tɩ tɩɛ taar a,
ranged
[17.5] unmarked pʋ-pʋla bound temporal: perfective positive pʋ- be existential pʋ-pʋla wa
non-future pʋla be be a,
[17.6] unmarked tɩ declarative: temporal: perfective positive tɩ bãw mental: tɩ bãw kɛ

affirmative non-future cognitive
[17.7] Absolute a yél ŋa declarative: temporal: perfective positive ʋ na tɔl material: a yél ŋa a,
a, affirmative future ranged ʋ na tɔl =ɩ
ni-daa
[17.8] Absolute ɛ a tɩ declarative: temporal: perfective positive bɛ ya na na material: ɛ a tɩ bibiir
bibiir bɛl affirmative future lɩɛbɩ intransitive bɛl [[na be
na be a a mʋɔ pʋɔ
mʋɔ pʋɔ a]], bɛ ya
a, mɩ na lɛ
lɩɛbɩ na.
[18.1] unmarked ɩ declarative: temporal: imperfective positive ɩ puore verbal Ah, ɩ puore
affirmative non-future bɛ nɩ barka
[18.2] unmarked bɛ bound temporal: perfective positive bɛ na behavioural bɛ na na
future saw saw sɔw a
sɔw lɛ tɩ [[na
yèl a]].
383
3.2 Tabulated Transitivity Analysis
# clause
AGENCY PROCESS TYPE (ranking)
Agency realised realised realised
type Agent Process Medium Process type Participant 1 by Participant 2 by Participant 3 by
[1] a tɩ yɛ-
bɛrɛ bɛl
Ɩ na puori a
[[na ar
tɩ yɛ-bɛrɛ
a
bɛl [[na ar
sãakʋmʋ
a sãakʋmʋ
yéle
yéle puor-
puor-tub
tub || bɛ
|| bɛ
bɔbr kɛ a
bɔbr kɛ
sãakʋmʋ
a
bɔr a]].
na sãakʋmʋ
middle puori Ɩ verbal Sayer Ɩ Target bɔr a]].
[2] Bɛ kpɛɛ za
Bɛ kpɛɛ relational: Bɛ kpɛɛ za mɩ be n= a
middle be za circumstantial identifying Existent mɩ Place a Guo Gǔo.
[3] nɩ-yawa
Ɛƴɛrɛŋa a,
[[na
tɩɩ =n nɩ-
bɔbr ||
yawa [[na
kɛ tɩ tɩɛ
bɔbr || kɛ tɩ
taar ||
tɩɛ taar ||
tʋɔ
tʋɔ maalɩ
maalɩ
yél kaw yél kaw
relational:
a]].
middle ɩ tɩ intensive attributive Carrier ɩ Attribute a]]
[4] A za kɩ yɛ-
[[a na ɩ kpɛɛ lɛ yi
A za kɩ relational: A za kɩ yɛ- a tɩ kpɛɛ na [[a na
middle yi yɛ-kpɛɛ intensive identifying Identified/Token kpɛɛ Identifier/Value a]] ɩa tɩ kpɛɛ
384
a]].
[5]
Tɩɩm za
relational: Tɩɩm za
kpɛɛ nʋ.
middle nʋ intensive identifying Identifier/Value kpɛɛ Identified/Token nʋ
[6.1] Ʋ yi na
yi … relational: pʋɔ a ƴaw pʋɔ a
middle ƴaw Ʋ circumstantial Identifying Identified/Token Ʋ Identifier/Value tɩƴaw tɩƴaw
[6.2] ɛ nɩbɛ bɛ
bɛ nye nye u e
middle [nyɛ] nɩbɛ mental cognitive Senser nɩbɛ Phenomenon ʋ [nyɛʋɛ],
[6.3] bɛ bɛ bãw
middle bɛ bãw bɛ mental cognitive Senser bɛ Phenomenon ʋ ʋɛ.
[7.1] Bɛ tu =n
middle tu bɛ material intransitive Actor bɛ Scope sɔr sɔr
[7.2] ɛ a nɩbɛ
bɛmɩnɛ bɛ
nye mɩ nye u
middle [nyɛ] bɛ mental cognitive Senser bɛ Phenomenon ʋ [nyɛʋ].
[8.1]
Ŋmɩŋmɩn
relational:
n=
middle n= intensive Identifier/Value Ŋmɩŋmɩn Identified/Token n=
[8.2] ɩrɛ ʋɩrɛ
middle ʋ material intransitive Actor ʋ
[8.3]
relational: bɩɩ bʋʋ nʋ
middle nʋ intensive identifying Identifier/Value bʋʋ Identified/Token nʋ
[8.4] ɩ ʋ mɩɩ
middle ʋ material transitive Actor ʋ Goal [bʋʋ]
[8.5] [near
ɛ yiele.
middle yiele [ʋ] behavioural verbal] Behaver [ʋ]
[9] Tɩ za [[na
tu sɔr a]], ɛ
middle tuori tɩ material intransitive Actor tɩ Scope taar tɩ tuori taar
385
a be, à
Bobo, a
cinq mars.
[10] Sàmédí
Sàmédí [Saturday]
relational: [Saturday] bibir =ɩ a
middle a lɛ intensive identifying Identifier/Value bibir Identified/Token a lɛ lɛ;
[11] <<au
théâtre de
l‟amitié>>,
Vɛntɛɛr vuo
minor minor pʋɔ.
[12] a nɩbɛ
nɛ za
[[na ɩ a Ɩ bʋɔlɛ n=
tɩ ba- a nɩbɛ nɛ za
taabɛ]] [[na ɩ a tɩ
a nɩ a tɩ ba-taabɛ]]
yɛbr tɩ a nɩ a tɩ
sãa yɛbr tɩ sãa
kʋm- kʋm-mɩnɛ,
mɩnɛ, tɩ tɩ makʋm-
makʋm- mɩnɛ, a tɩ
mɩnɛ, a sãa mɩnɛ.
tɩ sãa
middle bʋɔlɛ Ɩ material Actor Ɩ Scope mɩnɛ
[13.1] Bɛ tuori tɩ
middle tuori Bɛ material intransitive Actor Bɛ Scope tɩ a be
[13.2] a yél taa
taa wa relational: wa ɩ zawla
middle ɩ a yél intensive attributive Carrier a yél Attribute zawla ɩ.
[14.1] Tɩmɛ b‟a
[bɛ wa] ɩ
b'a ɩ … tɩɩna na
effective Tɩmɛ na sow tɩɩ material transitive Initiator Tɩmɛ Actor tɩɩ Scope taar sow taar a,
386
[14.2] kʋ tʋɔ nɩr kʋ tʋɔ
middle wa sow nɩr material transitive Actor nɩr Scope tɩ wa sow tɩɛ.
[15.1] Bɛ tuori tɩ
middle tuori Bɛ material intransitive Actor bɛ Scope tɩ a be
[15.2] tɩ maal a
effective tɩ maal a yél nɛ material transitive Actor tɩ Goal a yél nɛ yél nɛ
[15.3] Accompa- nɩƴa- ʋ tɔl nɩƴa-
middle tɔl ʋ material intransitive Actor ʋ niment ɓaarʋ ɓaarʋ
[15.4] nɩbɛ nɛ
ɛ nʋ̀ɔ kpɛ
za [[na
nɩbɛ nɛ za
na wa a
[[na na wa
be || wa
a be || wa
nyɛ a
nyɛ a bon
bon nɛ tɩ
nɛ tɩ [[na
[[na ɩrɛ
ɩrɛ a]]]].
middle kpɛ nʋ̀ɔ mental emotive Phenomenon nʋ̀ɔ Senser a]]]]
[16] Tɩ na maalɩ
nɔ-ɓaan
na yéle kʋ
maalɩ nɔ-ɓaan nɔ-ɓaan Client; taar; nɩ taar nɩ pʋ-
effective Tɩ … kʋ yéle material transitive Actor Tɩ Scope yéle Accompaniment pʋ-pɩɛlʋ pɩɛlʋ.
[17.1] sãa-
Tɩ na ɩ sãa-
relational: been
been bibiir
middle ɩ Tɩ intensive attributive Carrier Tɩ Attribute bibiir
[17.2]
ɛ wa faa ta
wa faa relational: nɩbɛ a
nɩbɛ a ŋmɩn
middle ta [Tɩ] intensive attributive Carrier [Tɩ] Attribute ŋmɩn
[17.3] yèl kɛ
middle Yèl [Tɩ] verbal Sayer [Tɩ]
[17.4] tɩ tɩɛ taar a,
middle tɩɛ tɩ material transitive Actor tɩ Scope taar
[17.5] pʋ-pʋla wa
middle wa be pʋ-pʋla existential Existent pʋ-pʋla be be a,
[17.6] tɩ bãw kɛ
middle bãw tɩ mental cognitive Senser
387
[17.7] a yél ŋa a,
ʋ na tɔl ɩ
middle na tɔl ʋ material intransitive Actor ʋ Scope ni-daa ni-daa
[17.8] ɛ a tɩ bibiir
bɛl na be a
mʋɔ pʋɔ a,
relational: bɛ ya mɩ na
middle na lɩɛbɩ bɛ ya intensive attributive Carrier bɛ ya lɛ lɩɛbɩ na.
[18.1] Ah, ɩ puore
middle puore ɩ verbal Sayer ɩ Target bɛ Verbiage barka bɛ nɩ barka
[18.2] bɛ na na
a lɛ tɩ saw sɔw a
na saw [[na yèl lɛ tɩ [[na
middle sɔw bɛ behavioural Behaver bɛ Phenomenon a]] yèl a]].
388
REFERENCES
Abdel-Hafiz, Ahmed-Sokarno. 2015. Focus constructions in Kunuz Nubian.

Dotawo: A Journal of Nubian Studies 2(5). 109-132.
Aikhenvald, Alexandra Y. & R. M. W. Dixon (eds). 2003. Studies in
evidentiality. Amsterdam & Philadelphia: John Benjamins.
Aikhenvald, Alexandra Y. 2012. The essence of mirativity. Linguistic
Typology, 16. 435-485.
Akerejola, Ernest. 2005. A text-based lexicogrammatical description of Ò̩ ̀ kò̩: A
systemic functional approach. Sydney: PhD Thesis, Macquarie
University.
Ali, Mark. 2006. Coordination in Dagaare. Journal of Dagaare Studies 6. 1-
19.
Andersen, Thomas, Uwe Helm Petersen, &Flemming Smedegaard. 2001.
Sproget som ressource: Dansk systemisk funktionel lingvistik i teori og
praksis. Odense: Odense Universitetsforlag.
Andersen, Thomas, Uwe Helm Petersen, & Flemming Smedegaard. 2002.
Metafunctional profile: Danish. A paper presented at the 14th Euro-
International Systemic Functional Linguistics Workshop, University of
Lisbon, Portugal, July 4-27, 2002.
Antilla, Arto &Adams Bodomo. 2000. Tonal polarity in Dagaare. In Carstens,
Vicki & Frederick Parkinson (eds.), Trends in African linguistics 4:
Advances in African linguistics, 119-134.Trenton, NJ: Africa World
Press.
Atoyebi, Joseph Dele. 2010. A reference grammar of Ò̩ ̀ kò̩: A West-Benue
Congo language of North-Central Nigeria. Köln: Rüdiger Köppe
Verlag.
Austin, John L. 1962. How to do things with words. Oxford: OUP.
Bache, Carl. 2010. Hjelmslev‘s glossematics: A source of inspiration to
systemic functional linguistics? Journal of Pragmatics 42. 2562–2578.
Bardi, Mohammed A. 2008. A systemic functional description of the grammar
of Arabic. Sydney: PhD Thesis, Macquarie University.
Bateman, John & Mick O‘Donnell. 2015. Halliday and computational
389
linguistics. In Webster, Jonathan J. (ed.), The Bloomsbury companion
to MAK Halliday, 453-466. London: Bloomsbury Publishing, 2015.
Beaver, David I. & Brady Z. Clark. 2009. Sense and sensitivity: How focus
determines meaning. Chichester, West Sussex: Wiley Blackwell.
Bemile, Sebastian K. 1983. Dagara phoneme contrasts, Vol. 1 & 2.
Saarbrücken: Universität des Saarlandes.
Bemile, Sebastian K. 1990. Dagara orthography. Papers in Dagara Studies
1(2). 1-31.
Bendor-Samuel, John T. 1971. Niger-Congo, Gur. In Thomas A. Sebeok
(eds.), Current trends in linguistics: Linguistics in Sub-saharan Africa,
vol. 7, 141-178. The Hague: Mouton.
Blench, Roger. 2010. The sensory world: Ideophones in Africa and elsewhere.
In A. Storch (ed.), Perception of the invisible: Religion, historical
semantics and the role of perceptive verbs, Sprache und Geschichte in
Afrika (275–296). Cologne: Köppe.
Bodomo, Adams. 1993a. Complex predicates and event structure: An
integrated analysis of serial verb constructions in Mabia languages of
West Africa. Trondheim: University of Trondheim.
Bodomo, Adams. 1993b. Dagaare syntax: A two-level X-bar account.
University of Trondheim Working Papers on African Linguistics 19.
20-51.
Bodomo, Adams. 1994. The noun class system of Dagaare: A phonology-
morphology interface. University of Trondheim Working Papers on
African Linguistics 22. 106-131.
Bodomo, Adams.1996. Linguistic relativity and the Mabia temporal system:
Evidence from Dagaare and Dagbane. Gur/Cahiers Voltaiques 1. 95-
103.
Bodomo, Adams. 1997a. The structure of Dagaare. Stanford: Stanford
University, CSLI Publications.
Bodomo, Adams. 1997b. Paths and pathfinders: Exploring the syntax and
semantics of complex verbal predicates in Dagaare and other
languages. PhD Thesis, The Norwegian University of science and
Technology, Trondheim.
Bodomo, Adams. 1998. Serial verbs as complex predicates in Dagaare and
390
Akan. In I. Maddieson & T. Hinnebusch (eds.). Language history and
linguistic decription in Africa.195-204. Trenton: Africa World Press.
Bodomo, Adams. 2000. Dagaare. Munchen, Germany: Lincom Europa.
Bodomo, Adams. 2006. The structure of ideophones in African and Asian
languages: The case of Dagaare and Cantonese. In J. Mugane et al.
(eds.), Selected proceedings of the 35th Annual Conference on African
Linguistics, 203-213. Somerville, MA: Cascadilla Proceedings Project.
Bodomo, Adams & Hiraiwa, Ken. 2004. Relativisation in Dagaare. Journal of
Dagaare Studies 4. 53-75.
Bresnan, Joan & Bodomo, Adams.1997. A note on Dagaare action
nominalisations as fixed categories. Stanford: Standford University
Brown, Penelope & Levinson Stephen C. 1987. Politeness: Some universals in
language usage. Cambridge: CUP.
Bureau of Ghana Languages. 1995. Concepts: English-Dagaare. Tamale,
Ghana: BGL.
Butler, Christopher. 2003. Structure and function: An introduction to three
major structural-functional theories, Part 1: Approaches to the
simplex clause. Amsterdam: John Benjamins.
Bybee, Joan, Revere Perkins &William Pagliuca. 1994. The evolution of
grammar: Tense, aspect and modality in the languages of the world.
Chicago: University of Chicago Press.
Caffarel, Alice. 2004. Metafunctional profile of the grammar of French. In A.
Caffarel, J. R. Martin & C. M. I. M. Matthiessen (eds.), Language
typology: A functional perspective,77-138. Amsterdam & Philadelphia:
John Benjamins.
Caffarel, Alice. 2006. A systemic functional grammar of French. London &
New York: Continuum.
Caffarel, Alice, J. R. Martin. & Christian M. I. M. Matthiessen (eds.). 2004.
Language typology: A functional perspective. Amsterdam &
Philadelphia: John Benjamins.
Callow, John C. 1969. Pre-survey report on Dagaare dialects. Accra: Institute
of African Studies.
Chomsky, Noam. (1965). Aspects of the theory of syntax. Cambridge, MA:
MIT Press.
391
Chomsky, Noam. 1981. Lectures on government and binding: The Pisa
lectures. Dordrecht & Cinnaminson, N.J: Foris Publications.
Christaller, Johannes G. 1889. Sprachproben aus dem Sudan von 40-60
Sprachen und Mundarten hinter der Gold-und Sklavenkuste. Zeitschrift
für Afrikanische Sprachen 3. 133-154.
Christie, Frances & Len Unsworth. 2005. Developing dimensions of an
educational linguistics. In Ruqaiya Hasan, Christian Matthiessen &
Jonathan Webster (eds.), Continuing Discourse on Language: A
Functional Perspective, Vol. 1, 217-250. London: Equinox Publishing.
Comrie, Bernard. 1976. Aspect. CUP.
Comrie, Bernard. 1989. Language universal and linguistic typology: Syntax
and morphology, 2nd edn. Oxford: Blakwell Publishing Ltd.
Creissels, Denis. 2000. Typology. In Bernd Heine & Derek Nurse (eds.),
African languages: An introduction, 231-258. Cambridge: CUP.
Croft, William. 1990. Typology and universals. Cambridge: CUP.
Crof, Willaim. 1991. Syntactic categories and grammatical relations.
Chicago: University of Chicago Press.
Dakubu, Mary E. Kropp. 1982. The tones of Dagaare. MS,Institute of African
Studies, University of Ghana, Accra.
Dakubu, Mary E. Kropp. 1989a. The Dagaare verbal phrase. MS, Department
of Linguistics, University of Ghana, Accra.
Dakubu, Mary E. Kropp. 1989b. Dagaare clause structures. MS, Department
of Linguistics, University of Ghana, Accra.
Dakubu, Mary E. Kropp. 2005. Collected language notes on Dagaare
grammar. Accra: Institute of African Studies.
Dakubu, Mary E. Kropp & James A Saanchi. 1997. Constituent focus in
Dagaare. A paper presented at the 9th Niger-Congo Syntax and
Semantics Workshop, University of Ghana, Accra, June 30-July 2.
Daneš, Frantisek, (ed). 1974. Papers on functional sentence perspective.
Prague: Academia.
Dansieh, Solomon Ali. 2008. Relevance of linguistic markers: Implications for
translating from English into a Gur language. Nordic Journal of
African Studies 17(4). 228-246.
de Saussure, Ferdinand. 1916. Cours de linguistique generate, Paris: Payot.
392
DeLancey, Scott. 1997. Mirativity: The grammatical marking of unexpected
information. Linguistic Typology1. 33–52.
DeLancey, Scott. 2012. Still mirative after all these years. Linguistic Typology
16. 529-564.
Delplanque, Alain. 1998. Focalisation et identification en Dagara. Gur
Papers/Cahiers Voltaiques 3. 19-26.
Delplanque, Alain. 2000. Perspective temporelle et valeurs modales des
marques de conjugaison: Le cas de wa et ti en Dagara-Lobr, Gur
Papers/Cahiers Voltaiques 5. 67-74.
Diessel, Holger. 1999. Demonstratives: Form, function and
grammaticalisation. Amsterdam & Philadelphia: John Benjamins.
Dimmendaal, Gerrt J. 2000. Morphology. In Bernd Heine & Derek Nurse
(eds.), African languages: An introduction, 161-193. Cambridge: CUP.
Dingemanse, Mark. 2011a. Ideophones and the aesthetics of everyday
language in a West-African society. The Senses and Society 6(1). 77–
85.
Dingemanse, Mark. 2011b. The meaning and use of ideophones in Siwu.
Nijmegen: PhD thesis, Radboud University, Retrieved from
http://thesis.ideophone.org/
Dingemanse, Mark. 2011c. Ezra pound among the Mawu: Ideophones and
iconicity in Siwu.In P. Michelucci, O. Fischer, & C. Ljungberg (eds.),
Semblance and signification, iconicity in language and literature, 39–
54. Amsterdam: John Benjamins.
Dixon, R. M. W. 1977. Where have all the adjectives gone. Studies in
Language 1.19-80.
Dixon. R. M. W. 1984. The semantic basis of syntactic properties. Berkeley
Linguistics Society Proceedings 10. 583-595.
Dixon, R. M. W. 2010a. Basic linguistic theory, Vol. 1: Methodology.
Oxford: OUP.
Dixon, R. M. W. 2010b. Basic linguistic theory, Vol. 2: Grammatical topics.
Oxford: OUP.
Dixon, R. M. W. 2012. Basic linguistic theory, Vol. 3: Further grammatical
topics. Oxford: OUP.
Dixon, R. M. W. & Alexandra Y. Aikhenvald (eds.). 2006. Complementation:
393
A cross-linguistic typology. Oxford: OUP.
Dong, David. A.1981. The verb phrase in Dagaare. Accra: Language Centre,
University of Ghana.
Downing, Angela. 2015. English grammar: A university course, 3rd edn.
London & New York: Routledge.
Dryer, Matthew S.& Martin Haspelmath (eds). 2013. The world atlas of
language structures online. Leipzig: Max Planck Institute for
Evolutionary Anthropology. Available online at http://wals.info,
Accessed on 2016-03-04.
Durand, J. B. 1953. Dagaare-English dictionary. Navorongo: St. John Bosco‘s
Press.
Eggins, Suzanne. 2004. An introduction to systemic functional linguistics, 2nd
edn. NewYork & London: Continuum.
Eggins, Suzanne & Diana Slade. 1997. Analysing casual conversation.
London & Oakville: Equinox.
Fawcett, Robin. 1981. Generating a sentence in systemic functional grammar.
In Michael A. K. Halliday & James R. Martin (eds.). 1981. Readings in
systemic linguistics. 146-183. London: Batsford Academic &
Educational Ltd.
Fillmore, Charles J. 1968. The case for case. In E. Bach & R. T. Harms (eds.),
Universals in linguistic theory, 1-88. London: Holt, Rinehart &
Winston.
Firbas, Jan. 1992. Functional sentence perspective in written and spoken
communication. Cambridge: CUP.
Firth, J. R. 1957. Papers in linguistics 1934-1957. London: OUP.
Garvin, Paul L. (ed.). 1964. A Prague School reader on esthetics, literary
structure, and style. Washington, DC: Georgetown University Press.
George, Isaac. 1971. The á-construction in Nupe: Perfective, stative, causative,
or instrumental. In Kim Chin-Wu & Herbert Stahlke (eds.). Papers in
African linguistics, 81-100. Berlin: Edmonton.
Givón, Talmy. 1971. On the verbal origin of the Bantu verb suffixes. Studies
in African Linguistics 2.145-164.
Givón, Talmy. 1975a. Focus and the scope of assertion: Some Bantu evidence.
Stud. Afr. Linguist 6.185–205.
394
Givón, Talmy. 1975b. Serial verbs and syntactic change: Niger-Congo. In
Charles Li (ed.), Word order and word order change, 47-112). Austin:
University of Texas Press.
Givón, Talmy. 1979. Language typology in Africa: A critical review. JALL 1.
199-224.
Givón, Talmy, ed. 1983. Topic continuity in language. Amsterdam: John
Benjamins
Girault, Louis. 1964. Note sur la particule post-verbale na en Dagara. BIFAN
26. 499-504.
Girault, Louis. 1967. Description phonologique de langue Dagara. Dakar:
Institut fondamental d'Afrique noire, Universite de Dakar.
Good, Jeff. 2010. Topic and focus fields in Naki. In I. Fiedler & A.
Schwarz(eds.), The expression of information structure: A
documentation of its diversity across Africa, 35–67.Amsterdam:
Benjamins 
Goody, Jack. 1954. The ethnography of the Northern Territories of the Gold
Coast. London: Colonial Office.
Goody, Jack. 1972. The myth of the Bagre. Oxford: OUP.
Greenberg, Joseph H. 1955. Studies in African linguistic classification. New
Haven: Compass Publishing Co.
Greenberg, Joseph H. 1966. The languages of Africa, 2nd edn. Bloomington:
Indiana University.
Greenberg, Joseph H. 1966. Some universal of grammar with particular
reference to the order of meaningful elements. In J. H. Greenberg (ed.)
Universals of language, 2nd edn., 73–113. Cambridge, MA: MIT Press.
Güldemann , Tom. 2008. Quotative indexes in African languages: A
synchronic and diachronic survey. Berlin & New York: Mouton de
Gruyter.
Güldemann, Tom, Sabine Zerbian &Malte Zimmermann. 2015. Variation in
informationstructure with special reference to Africa. Annual Review
of Linguistics, 1:155–178.
Gundel, Jeanette Marie. 1974. The role of topic and comment in linguistic
theory. Austin: PhD thesis, University of Texas.
395
Halliday, M. A. K. 1961. Categories of the theory of grammar. Word 17(3).
242-92.
Halliday, M. A. K. 1966. Some notes on ‗deep‘ grammar. Journal of
Linguistics 2 (1). 57- 67.
Halliday, M. A. K. 1967a. Notes on transitivity and theme in English, Part 1.
Journal of Linguistics 3(1). 37-81.
Halliday, M. A. K. 1967b. Notes on transitivity and theme in English, Part 2.
Halliday, M. A. K. 1968. Notes on transitivity and theme in English, Part 3.
Halliday, M. A. K. 1970. Language structure and language function. In John
Lyons (ed.), New horizons in linguistics, 140-165. Harmondsworth:
Penguin.
Halliday, M.A.K. 1973. Explorations in the functions of language. London:
Edward Arnold.
Halliday, M. A. K. 1977. Ideas about language. Occasional Papers I, 32-55.
Halliday, M. A. K. 1979. Modes of meaning and modes of expression: Types
of grammatical structure and their determination by different semantic
functions. In D. J. Allerton, Edward Carney & David Holdcroft (eds.),
Function and context in linguistic Analysis: A festschrift for William
Haas, 57-79. Cambridge: CUP.
Halliday, M. A. K. 1981. Text semantics and clause grammar: Some patterns
of realisation. In James E. Copeland & Phillip W. Davies (eds.),The
Seventh LACUS Forum, 31-59. Columbia, SC: Hornbeam Press.
Halliday, M. A. K. 1982. How is the text like a clause? In Sture Allen (ed.),
Text Processing: Text analysis and generation, text typology and
attrition, 209 – 247. Stockholm: Almqvist & Wiksell.
Halliday, M. A. K. 1984. Language as code and language as behaviour: A
systemic-functional interpretation of the nature and ontogenesis of
dialogue. In Robin Fawcett, M.A.K. Halliday, Sydney M. Lamb &
Adam Makkai (eds.), The semiotics of culture and language, 1-35.
London: Frances Pinter.
Halliday, M. A. K. 1985. Introduction to functional grammar. London: Arnold.
Halliday, M. A. K. 1992. How do you mean? In Martin Davies & Louise
396
Ravelli (eds.), Advances in systemic linguistics: Recent theory and
practice, 20-35. London: Pinter.
Halliday, M. A. K. 1996. On grammar and grammatics. In Hasan, R. & C.
Butt (eds.). Functional descriptions: Theory into practice,1-38.
Amsterdam: John Benjamins.
Halliday, M. A. K. 2002. Volume 1 in the collected works of M. A. K.
Halliday: On grammar. London & New York: Continuum.
Halliday, M. A. K. 2005. On matter and meaning: The two realms of human
experience. Linguistics and the Human Sciences 1(1). 59-82.
Halliday, M. A. K. 2008. Complementarities in language. Beijing: The
Commercial Press.
Halliday, M. A. K. 2009. Methods – techniques – problems. In Michael A. K.
Halliday & Jonathan J. Webster (eds.). Continuum companion to
systemic functional linguistics. 59-86. London & New York:
Continuum.
Halliday, M. A. K. & Greaves, William. 2008. Intonation in the grammar of
English. London & Oakville: Equinox.
Halliday, M.A.K. & Jonathan Webster (eds.). 2009. Continuum companion to
systemic functional linguistics. London & New York: Continuum.
Halliday, M. A. K. & J. R. Martin (eds.). 1981. Readings in systemic
linguistics. London: Batsford Academic & Educational Ltd.
Halliday, M. A. K. & Christian M. I. M. Matthiessen. 1999. Construing
experience through meaning: A language-based approach to cognition.
London: Continuum.
Halliday, M. A. K. & Christian M. I. M. Matthiessen. 2014. Halliday‟s
introduction to functional grammar, 4th edn. London & New York:
Routledge.
Halliday, M. A. K. & Edward McDonald. 2004. Metafunctional profile of the
grammar of Chinese. In A. Caffarel, J. R. Martin & C. M. I. M.
Matthiessen (eds.), Language typology: A functional perspective, 305-
393. Amsterdam & Philadelphia: John Benjamins.
Halliday, M. A. K. & Ruqaiya Hassan. 1976. Cohesion in English. London:
Longman.
Halliday, M. A. K. & Zoe L. James. 1993. A quantitative study of polarity and
397
primary tense in the English finite clause. In John M. Sinclair, Michael
Hoey & Gwyneth Fox (eds). Techniques of description: Spoken and
written discourse. 32–66. London & New York: Routledge.
Harris, Alice C. & Lyle Campbell. 1995. Historical syntax in cross-linguistic
perspective. Cambridge: Cambridge University Press.
Hartmann, Katharina & Malte Zimmermann. 2007. In place—out of place?
Focus in Hausa.In Kerstin Schwabe, & Susanne Winkler (eds). On
information structure, meaning and form: Generalisations across
languages, 365-403. Amsterdam: John Benjamins. 
Hartmann Katharina, Tonjes Veenstra (eds.). 2013. Cleft structures.
Amsterdam: Benjamins. 
Hasan, Ruqaiya, Christian M.I.M. Matthiessen & Jonathan Webster (eds.).
2005. Continuingdiscourse on language: A functional perspective, Vol.
1. London: Equinox Publishing.
Hasan, Ruqaiya, Christian M.I.M. Matthiessen & Jonathan Webster (eds.).
2007. Continuing discourse on language: A functional perspective,
Vol. 2. London: Equinox Publishing.
Haspelmath, Martin. 2009a. Framework-free grammatical theory. In Heine,
Bernd & Narrog, Heiko (eds.) The Oxford handbook of grammatical
analysis. Oxford: Oxford University Press, 341-365.
Haspelmath, Martin. 2009b. Terminology of case. In Andrej Malchukov &
Andrew Spencer, eds. The Oxford handbook of case. 505-517. Oxford:
OUP.
Haspelmath, Martin. 2015. Ditransitive constructions. Annual Review of
Linguitics 1. 19-41.
Haspelmath, Martin. 2016. The serial verb construction: Comparative concept
and cross-linguistic generalisations. Language and Linguistics,17(3).
291-319. 
Hancil, Sylvie, Alexander Haselow & Margie Post (eds.). 2015. Final
particles. Berlin & Boston: Mouton de Gruyter.
Hengeveld, Kees, and J Lachlan Mackenzie. 2008. Functional Discourse
Grammar: A typologically-based theory of language structure. Oxford:
OUP.
398
Heine, Bernd. 2011. Grammaticalisation in African Languages. In Bernd
Heine & Heiko Narrog (eds.), The Oxford handbook of
grammaticalisation, 696-707. Oxford: Oxford University Press.
Heine, Bernd, Ulrike Claudi & Friederike Hünnemeyer. 1991.
Grammaticalisation: A conceptual framework. London: The
University of Chicago Press.
Wray (ed.). 2002. The transition to language. 376-397. Oxford: Oxford
University Press.
Heine, Bernd & Tania Kuteva. 2002a. World Lexicon of grammaticalisation.
Cambridge:Cambridge University Press.
Heine, Bernd & Tania Kuteva. 2002b. On the evolution of grammatical forms.
In Alison Wray (ed.). The transition to language. 376-397. Oxford:
Oxford University Press.
Heine, Bernd & Tania Kuteva. 2007. The genesis of grammar. Oxford: Oxford
University Press.
Heine, Bernd & Heiko Narrog, eds. 2009. The Oxford handbook of linguistic
analysis. Oxford: Oxford University Press.
Heine, Bernd & Derek Nurse. (eds.) 2000. African languages: An
introduction. Cambridge:CUP.
Heine, Bernd & Derek Nurse. 2000. Introduction. In Bernd Heine & Derek
Nurse (eds.). In Bernd Heine & Derek Nurse (eds.), African languages:
An introduction, 1-10. Cambridge: CUP.
Heine, Bernd & Mechthild Reh. 1983. Diachronic observations on completive
focus marking in some African languages. Sugia: Sprache und
Geschichte in Afrika 5. 7-44.
Henrici, Alick. 1981. Some notes on the systemic generation of a paradigm of
the English clause. In Michael A. K. Halliday &J. R. Martin (eds.).
Readings in systemic linguistics. 74-98. London: Batsford Academic &
Educational Ltd.
Hiraiwa, Ken &Adams Bodomo. 2008. Object-sharing as symmetric sharing:
Evidence from Dagaare. In C. B. Chang & H. J. Haynie (eds.).
Proceedings of the 26th West Coast Conference on Formal Linguistics.
243-251. Somerville, MA: Cascadilla Proceedings Project.
Hjelmslev, Louis. 1954. La stratification du langage. Word 10. 163-188.
399
Hjelmslev, Louis. 1969. Prolegomena to a theory of language. Translated by
Francis J. Whitfield. Madison, Milwaukee & London: The University
of Wisconsin Press (Original: Omkring Sprogteoriens Grundlaeggelse,
1943).
Hockett, Charles F. 1942. A system of descriptive phonology. Language 18(1).
3-21.
Hopper, Paul. 1991. On some principles of grammaticalisation. In Traugott,
Elizabeth C. & Bernd Heine (eds.), Approaches to grammaticalisation
(vol. 1): Focus on theoretical and methodological issue, 17-35.
Amsterdam & Philadelphia: John Benjamins.
Hopper, Paul J. & Sandra A Thompson. 1980. Transitivity in the grammar of
discourse, Language 56(2). 251-299.
Hopper, Paul J. & Elizabeth C. Traugott. 2003. Grammaticalisation.
Cambridge: CUP.
Huddleston, Rodney D. 1981. Systemic features and their realisation. In
M.A.K. Halliday & J.R. Martin (eds.), Readings in systemic linguistics,
58-73. London: Batsford Academic & Educational Ltd.
Hymes, Dell. 1974. Ways of speaking. In R. Bauman & J. Sherzer (eds.).
Explorations in theethnography of speaking, 433-451. London:
Cambridge University Press.
Iwasaki, Shoichi & Foong Ha Yap. 2015. Stance-marking and stance taking in
Asian languages. Journal of Pragmatics 83.1-9.
Keenan, Edward & Bernard Comrie. 1977. Noun phrase accessibility and
universal grammar. Linguistic Inquiry 8. 63-99.
Keenan, Edward & Bernard Comrie. 1979. Data on the noun phrase
accessibility hierarchy. Language 55. 333.51.
Kennedy, John. 1966. Collected field reports on the phonology of Dagaari.
Accra: Institute of African Studies, University of Ghana.
Kumar, Abhishek. 2009. Mood, transitivity and theme in Bajjika in a
typological perspective: A text-based description. Sydney: PhD Thesis,
Macquarie University.
Kunene, Daniel P. 1965. The ideophone in southern Sotho. Journal of African
Languages 4. 19-39.
Kunz, Kerstin, Elke Teich, Silvia Hansen-Schirra, Stella Neumann, Peggy
400
Daut (eds.). 2014. Caught in the middle – Language use and
translation: A festschrift for Erich Steiner on the occasion of his 60th
birthday. Saarbrücken: Saarland University Press.
Ladefoged, Peter & Keith Johnson. 2011. A course in phonetics, 6th edn.
Boston: Wadsworth.
Lakoff, George & Mark Johnson. 1980. Metaphors we live by. Chicago:
University of Chicago press.
Lakoff, George. 2007. Ten lectures on cognitive linguistics. Beijing: Foreign
LanguageTeaching and Research Press.
Lamb, Sydney M. 1966. Outline of stratificational grammar. Washington DC:
Georgetown University Press.
Lambrecht, Knud. 1994. Information structure and sentence form. Cambridge:
CUP.
Lambrecht, Knud. 1996. Information structure and sentence form: Topic,
focus, and themental representations of discourse referents.
Cambridge: CUP.
Lambrecht Knud. 2000. When subjects behave like objects: an analysis of the
merging of S and O in sentence-focus constructions across languages.
Studies in Language 24.611–82. 
Langacker, Ronald W. 2008. Cognitive grammar: A basic introduction.
Oxford: OUP.
Lavid, Julia, Jorge Arús & Juan Rafael Zamorano-Mansilla. 2010. Systemic
functional grammar of Spanish: A contrastive study with English.
London & New York: Continuum.
Lemke, Jay L. 1984. Semiotics and education. Toronto: Toronto Semiotic
Circle.
Lewis, M. Paul, Gary F. Simons, and Charles D. Fennig (eds.). 2016.
Ethnologue: Languages of the world, 19th edn. Dallas, Texas: SIL
International. Online version: http://www.ethnologue.com.
Li, Charles N. & Sandra A. Thompson. 1976. Subject and topic: A new
typology of languages. In Charles N. Li, (ed.), Subject and topic, 457-
490. New York: Academic Press.
Li, E. Sum-hung. 2007. A systemic functional grammar of Chinese. London &
401
New York:Continuum.
Luke, Kang. K. & Adams Bodomo. 2000. A comparative study of the
semantics of serial verbs constructions in Dagaare and Cantonese.
Languages in Contrast 3(2). 165-180.
Lukin, Annabelle & Jonathan J. Webster. 2005. SFL and the study of literature.
In Ruqaiya Hasan, Christian Matthiessen & Jonathan Webster (eds.),
Continuing discourse on language: A functional perspective, Vol. 1,
413-456. London: Equinox Publishing.
Lyons, John. 1977. Semantics, vol. 2. Cambridge: CUP.
Malinowski, Bronislaw. 1923. The problem of meaning in primitive
languages. In C. K. Ogden & I. A. Richards (eds.), The meaning of
meaning, 146-152. London: Routledge & Kegan Paul.
Malinowski, Bronislaw. 1935. Coral gardens and their magic: A study of the
methods of tilling the soil and of agricultural rites in the trobriand
islands, vol. 1. Hamburg: Severus. 
Martin, J.R. 1987. The meaning of features in systemic linguistics. In M. A. K
Halliday & Robin P. Fawcett (eds.), New developments in systemic
linguistics, 14-40. London: Frances Pinter.
Martin, J.R. 1988.Grammatical conspiracies in Tagalog: Family, face and
fate — with regard to Benjamin Lee Whorf. In J. D. Benson, M. J.
Cummings & W.S. Greaves (eds), Linguistics in a systemic perspective. 243-
300. Amsterdam: Benjamins.
Martin, J. R. 1993. Life as a noun: Arresting the universe in science and
humanities. In M. A. K. Halliday & J. R. Martin (eds.), Writing
science: Literacy and discursive power, 221-267. Pittsburgh:
University of Pittsburgh Press.
Martin, J.R. 2004. Metafunctional profile of the grammar of Tagalog. In Alice
Caffarel, J. R. Martin & Christian Matthiessen (eds.), Language
typology: A functional perspective, 255-304. Amsterdam &
Martin, J.R. 2010. Semantic variation – Modelling realisation, instantiation
and individuation in social semiosis. In Monika Bednarek & J. R.
Martin (eds.), New discourse on language: Functional perspectives on
402
multimodality, identity and affiliation, 1-34. London & New York:
Continuum.
Martin, J. R. 2013. Systemic functional grammar: A next step into the theory –
Axial relations. Beijing: Higher Education Press.
Martin, J. R, Michele Zappavigna, Paul Dwyer & Chris Cleirigh. 2013. Users
in uses of language: Embodied identity in Youth Justice Conferencing.
Text & Talk 33(4–5). 467-496.
Martin, J. R.& David Rose. 2007. Working with Discourse: Meaning beyond
the clause, 2nd edn. London: Continuum.
Matthiessen, Christian M.I.M. 1987. Notes on Akan phonology: A systemic
interpretation. Mimeo.
Matthiessen, Christian M. I. M. 1995. Lexicogrammatical cartography:
English systems. Tokyo: International Language Sciences Publishers.
Matthiessen, Christian M. I. M. 2004. Descriptive motifs and generalisations.
In Alice Caffarel, J. R. Martin & Christian Matthiessen (eds.),
Language typology: A functional perspective,537-674. Amsterdam &
Matthiessen, Christian M. I. M. 2007a. The ―architecture‖ of language
according to systemic functional theory: Developments since the 1970s.
In Ruqaiya Hasan, Christian M.I.M. Matthiessen & Jonathan Webster
(eds.), Continuing discourse on language: A functional perspective,
vol. 2, 505-562. London & Oakville: Equinox.
Matthiessen, Christian M. I. M. 2007b. Lexicogrammar in systemic functional
linguistics: descriptive and theoretical developments in the ‗IFG‘
tradition since the 1970s. In Ruqaiya Hasan, Christian M.I.M.
Matthiessen & Jonathan Webster (eds.), Continuing discourse on
language: A functional perspective, vol. 2,765-858. London &
Oakville: Equinox.
Matthiessen, Christian M. I. M. 2013a. Appliable discourse analysis. In Fang
Yan & Jonathan J. Webster (eds.). Developing systemic functional
linguistics: Theory and application, 135-205. London: Equinox.
Matthiessen, Christian M. I. M. 2013b. Applying systemic functional
linguistics in healthcare contexts. Text & Talk 33(4–5). 437-467.
Matthiessen, Christian M. I. M. 2015. Register in the round: Registerial
403
cartography. Functional Linguistics 2(9). 1-48.
Matthiessen, Christian. M. I. M. &John A. Bateman. 1991. Text generation
and systemic-functional linguistics: Experiences from English and
Japanese. London &New York: Frances Pinter Publishers & St.
Martin's Press.
Matthiessen, Christian M. I. M. & Diana Slade. 2010. Analysing conversation.
In Ruth Wodak, Barbara Johnstone, Paul E. Kerswill (eds.),The Sage
handbook of sociolinguistics, 375-395. London: Sage.
Matthiessen, Christian M. I. M. & M. A. K. Halliday. 2009. Systemic
functional grammar: A first step into the theory. Beijing: Higher
Education Press.
Matthiessen, Christian M. I. M. & Abhishek K. Kashyap. 2014. The construal
of space in different registers: An exploratory study. Language
Sciences 45. 1-27.
Matthiessen, Christian M. I. M., Kazuhiro Teruya & Canzhong Wu. 2008.
Multilingual Studies as a multidimensional space of interconnected
language studies. In Jonathan Webster (ed.), Meaning in context:
Strategies for implementing intelligent applications of language
studies, 146-220. London: Continuum.
Matthiessen, Christian M. I. M. & Kazuhiro Teruya. 2015. Grammatical
realisations of rhetorical relations in different registers. Word 61(3).
232-281.
Matthiessen, Christian C. M. I. M. & Christopher Nesbitt. 1996. On the idea of
theory-neutral descriptions. In Ruqaiya Hasan, Carmel Cloran and
David G. Butt (eds.), Functional descriptions: Theory in practice, 39-
84. Amsterdam: Benjamins.
Miehe, Gudrun. 2012. Dagara cluster. In Miehe, et al. pp. 250-268.
Miehe, Gudrun, Brigitte Reineke & Kerstin Winkelmann (eds.). 2012. Noun
class systems in Gur languages (Vol. 2 ): North Central Gur
languages, (Gur Monographs 11). Köln: Rüdiger Köppe Verlag.
Moutaouakil, Ahmed. 1999. Exclamation in Functional Grammar: Sentence
type, illocution or modality? (Working Papers in Functional Grammar
69). Amsterdam: University of Amsterdam.
Mwinlaaru, Isaac N. 2014a. Style, character and the theme of struggle and
404
change: Chinua Achebe‘s Anthills of the Savannah. Research in
African Literatures 42(2). 103-121.
Mwinlaaru, Isaac N. 2017. Bridging across genre traditions: A systemic
functional account of biodata. Functions of Language 24(3). in press.
Mwinlaaru, Isaac N. in press. Grammaticalising attitude: Clause juncture
particles and NEGOTIATION in Dagaare. In Lise Fontaine & Akila S.
Baklouti (eds.), Perspectives from systemic functional linguistics.
London & New York: Routledge.
Mwinlaaru, Isaac N. & Foong Ha Yap. 2017. A tale of two distal
demonstratives in Dagaare: Reflections on directionality principles in
grammaticalisation. Language Sciences, 64. 130-151.
Mwinlaaru, Isaac N. & Winfred Wenhui Xuan. 2016. A survey of studies in
systemic functional language description and typology. Functional
Linguistics, 3(8): 1-41.
Mensah, Emmanuel N. A. & Zakari Tchagbale (eds.). 1983. Atlas des langues
gur de Cote d‟Ivoire. Abidjan: ACCE/ILA
Nakuma, Constancio. 1990. Etudes contratives des systems phonologies
Dagaare-Francais,avec proposition d‟une orthographie Dagaare
tenant du compte des Faits Tonals. Paris: PhD Thesis, University of
Sorbonne.
Nakuma, Constancio. 1999. Phone et graphie tonale du Dagaare: Langue
Voltaique. Paris: Editions de l‘Harmattan.
Narrog, Heiko & Bernd Heine. 2011. The Oxford handbook of
grammaticalizaion. Oxford: OUP.
O‘Donnell, Mick & John Bateman. 2005. SFL in computational contexts: A
comtemporary history. In Ruqaiya Hasan, Christian M. I. M.
Matthiessen & Jonathan J. Webster (eds.),Continuing discourse on
language: A functional perspective, vol. 1, 343-382. London: Equinox
Publishing.
Painter, Clare. 2009. Language development. In Halliday, Michael A. K. &
Jonathan J. Webster (eds.), Continuum companion to systemic
functional linguistics, 87-103. London & New York: Continuum.
Pustet, Regina. 2003. Copulas: Universals in the categorisation of the lexicon.
Oxford: OUP.
405
Quirk, Randolph & Sydney Greenbaum. 1973. A university grammar of
English. New Delhi: Pearson Education India.
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech & JanSvartvik. 1985. A
Comprehensive Grammar of the English Grammar. London: Longman.
Quiroz,Beatriz. 2008. Towards a systemic profile of the Spanish MOOD.
Linguistics & the Human Sciences 4(1). 31-65.
Quiroz, Beatriz. 2013. The interpersonal and experiential grammar of Chilean
Spanish:Towards a principled systemic-functional description based
on axial argumentation. Sydney: PhD thesis, Universidad de Sídney.
Rialland, Annie & Stéphane Robert. 2001. The intonational system of Wolof.
Linguistics 39(5). 893-940.
Rattary, Robert S. 1932. The tribes of the Ashanti hinterland. Oxford: OUP.
Robins, Robert H. 1966. The development of the word class system of the
European grammatical tradition. Foundations of language 3-19.
Rosch, Eleanor. 1983. Prototype classification and logical classification: The
two systems. In Ellin Kofsky Scholnick (ed). New Trends in cognitive
representation: Challenges to Piaget's theory, 73-86. Hillsdale, NJ:
Lawrence Erlbaum Associates.
Rose, David & J.R. Martin. 2012. Learning to write, reading to learn: Genre,
knowledge and pedagogy in the Sydney School. Sheffield & Bristol:
Equinox.
Saanchi, James A. 1980. The nominal phrase in Dagaari. Accra: B.A. Thesis,
University of Ghana.
Saanchi, James A.1997. The vowel system of Dagaare. Gur Papers/Cahiers
Voltaiques 2. 129-135.
Saanchi, James A. 2003a. Aspect and the Dagaare verb. Gur Papers/Cahiers
Voltaiques 6. 101-106.
Saanchi, James A. 2003b. Spatial and locative constructions in Dagaare.
Legon Journal of the Humanities 14. 29-36.
Sacks, Harvey, Emanuel A. Schegloff & Gail Jefferson. 1974. A simplest
systematic for the organisation of turn-taking for conversation.
Language 50(4). 696-735.
Sakurai, Kazuhiro 2014. The syntax and semantics of focus: Evidence from
Dagaare. Hong Kong: PhD Thesis, The University of Hong Kong.
406
Sapir, Edward. 1921. Language: An introduction to the study of speech.New
York: Harcout, Brace & Company.
Sasse, Hans-Jürgen. 1981. Basic word order and functional sentence
perspective in Boni. Folia linguistica 15(3-4). 253-290.
Sasse, Hans-Jürgen. 1987. The thetic/categorical distinction revisited.
Linguistics 25(3). 511-580.
Schachter Paul. 1973. Focus and relativisation. Language 49.19-46 
Schwabe, Kerstin & Susanne Winkler (eds.). 2007. On information structure,
meaning and form: Generalisations across languages. Amsterdam:
John Benjamins.
Schwarz, Anne. 2009. Tonal focus reflections in Buli and some Gur relatives.
Lingua 119.950-72. 
Searle, John. 1969. Speech acts: An essay in the philosophy of language.
Cambridge: CUP.
Shopen, Timothy (ed.). 2007a. Language typology and syntactic description
(vol. 1): Clause structure. Cambridge: CUP.
Shopen, Timothy (ed.). 2007b. Language typology and syntactic description
(vol. 2): Complex constructions. Cambridge: CUP.
Shopen, Timothy (ed.). 2007c. Language typology and syntactic description
(vol. 3): Grammatical categories and the lexicon. Cambridge: CUP.
Simpson, Paul. 2014. Stylistics: A resource book for students. London & New
York: Routledge.
Sinclair, John. 1991. Corpus concordance, collocation. Oxford: OUP.
Slade, Diana, Hermine Scheeres, Marie Manidis, Christian M. I. M.
Matthiessen, Rick Iedema, Maria Herke, Jeannette McGregor, Roger
Dunston & Jane Stein-Parbury. 2008. Emergency communication: The
discursive challenges facing emergency clinicians and patients in
hospital emergency departments. Discourse & Communication 2(3).
289-316.
Slade, Diana, Marie Manidis, Jeannette McGregor, Hermine Scheeres, Jane
Stein-Parbury, Roger Dunston, Nicole Stanton, Eloise Chandler,
Christian Matthiessen & Maria Herke. 2011. Communicating in
407
hospital emergency departments: Final report. Sydney: University of
Technology.
Slade, Diana, Marie Manidis, Jeannette McGregor, Hermine Scheeres, Eloise
Chandler, Jane Stein-Parbury, Roger Dunston, Maria Herke, &
Christian M. I. M. Matthiessen. 2015. Communicating in hospital
emergency departments. Berlin: Springer.
Somé, B. A. 2003. Brève rétrospective sur l‘orthographe dagara. In Joachim
D. Somé (ed), Actes de la Table Ronde organisée sur l‟orthographe
moderne Dagara. Diébougou,Burkina Faso: ABBF.
Somé, Coubabaon. 1985. Elements de grammaire Dagara. Mémoiré de
Maîtrise, University of Ouagadougou.
Somé, Joachim D. 2004. Dagara orthography. Journal of Dagaare Studies 4.
15-51.
Somé, Pé nou-Achille. 1962. Systématique du signifiant en Dagara: Variété
Wúlé. Paris: Harmattan.
Stassen, Leon. 1997. Intransitive predication. Oxford: Clarendon.
Steiner, Erich. 2002. Grammatical metaphor in translation - some methods for
corpus-based investigations. In Hasselgard, Hilde, Johansson, Stig,
Behrens Bergljot, & Fabricius-Hansen, Cathrine (eds), Information structure
in a cross-linguistic perspective, 213-228. Amsterdam: Rodopi.
Steiner, Erich. 2005. Halliday and translation theory: Enhancing the options,
broadening therange and keeping the ground. In Ruqaiya Hasan,
Christian Matthiessen & JonathanWebster (eds.), Continuing discourse
on language: A functional perspective, vol. 1, 481-500. London:
Equinox Publishing.
Swadesh, Morris, Evangelina Arana, John T. Bendor-Samuel & W. A. A
Wilson,1966. Apreliminary glottochronology of Gur languages.
Journal of West African Languages 3(2). 27-66.
Tam, Hoi-sang Roy. 2004. A systemic functional interpretation of Cantonese
clause grammar. Sydney: PhD thesis, University of Sydney.
Tench, Paul. (ed.) 1992. Studies in systemic phonology. London: Pinter. 
Teruya, Kazuhiro. 1998. An exploration into the world of experience: A
408
systemic-functionalinterpretation of the grammar of Japanese. Sydney:
PhD thesis, Macquarie University.
Teruya, Kazuhiro. 2007. A systemic functional grammar of Japanese, vol. 1 &
2. London & New York: Continuum.
Teruya, Kazuhiro, Ernest Akerejola, Thomas H. Andersen, Alice Caffarel,
Julia Lavid, Christian Matthiessen, Uwe Helm Petersen, Pattama
Patpong & Flemming Smedegaard. 2007. Typology of MOOD: A text-
based and system-based functional view. In Hasan, Ruqaiya, Christian
M.I.M. Matthiessen & Jonathan J. Webster (eds.), Continuing
discourse on language: A functional perspective, Vol. 2. 859-920.
London: Equinox.
Teruya, Kazuhiro & Christian M.I.M. Matthiessen. 2015. Halliday in relation
to language comparison and typology. In Jonathan J. Webster (ed.),
The Bloomsbury companion to M.A.K. Halliday, 427-452. London:
Bloomsbury Academic.
Tesnière, Lucien. 1959. Éleménts de syntaxe structurale. Paris: Klincksieck.
Thompson, Sandra A. & Elizabeth Couper-Kuhlen. 2005. The clause as a
locus of grammar and interaction. Language and Linguistics, 6(4).
807-837.
Traugott, Elizabeth C. & Bernd Heine (eds.). 1991. Approaches to
grammaticalisation (vol. 1): Focus on theoretical and methodological
issues. Amsterdam & Philadelphia: John Benjamins.
van der Wal, Jenneke. 2011. Focus excluding alternatives: Conjoint/disjoint
marking in Makhuwa. Lingua 121. 1734-1750. 
van Valin, Robert D Jr. 2000. A concise introduction to Role and Reference
Grammar. Fluminesia 12(1-2). 47-78.
van Valin, Robert D. Jr. 2007. The Role and Reference Grammar analysis of
three-place predicates. Prihvaćeno za tisak 14(2). 31-63.
Vendler, Zeno. 1957. Verbs and times. The Philosophical Review 66(2). 143-
160.
Wang, Yong, & Jie Xu. 2013. A systemic typology of existential and
possessive constructions. Functions of Language 20(1). 1-30. 
Watters, John R. 2000. Syntax. In Bernd Heine & Derek Nurse (eds.), African
409
languages: An introduction, 194-230. Cambridge: CUP.
Webster, Jonathan. 2015. Understanding verbal art: A functional linguistics
approach. Heidelberg, New York, Dordrecht & London: Springer.
Welmers, Wm. E. 1973. African language structures. Berkley, Los Angeles &
London:University of California Press.
Whorf, Benjamin Lee. 1956. Language, thought and reality, Selected writings
of Benjamin Lee Whorf. Ed. John Bissell Carroll. Cambridge &
Massachusetts: The MIT Press.
Williamson, Kay & Roger Blench. 2000. Niger-Congo. In Heine, Bernd &
Derek Nurse(eds.), African Languages: An introduction, 11-42.
Cambridge: CUP.
Wilson, W. A. A. 1962a. Notes on Dagaari: Kaleo-Jirapa dialects. MS.,
Institute of African Studies, University of Ghana, Accra.
Wilson, W. A. A. 1962b. The vowel system of Dagaare and Lobiri. MS.,
Institute of African Studies, University of Ghana, Accra.
Yang, Ying & Foong Ha Yap. 2015. ‗‗I am sure but I hedge‘‘: Fear expression
kǒngpà as aninteractive rhetorical strategy in Mandarin broadcast talk.
Journal of Pragmatics 83. 41-56.
Zimmermann, Malte & Caroline Féry (eds.). 2010. Information structure:
Theoretical, typological and experimental perspectives. Oxford: OUP.
Zimmermann, Malte. 2011. The grammatical expression of focus in West
Chadic: Variationand uniformity in and across languages. Linguistics
49. 1163-213. 
410

Mwinlaaru - PHD Thesis - Final

Uploaded by

Copyright:

Available Formats

Mwinlaaru - PHD Thesis - Final

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mwinlaaru - PHD Thesis - Final

Uploaded by

Copyright:

Available Formats

A SYSTEMIC FUNCTIONAL DESCRIPTION OF THE GRAMMAR OF

ISAAC NUOKYAA-IRE MWINLAARU

The Hong Kong Polytechnic University

A Systemic Functional Description of the Grammar of Dagaare

Isaac Nuokyaa-Ire Mwinlaaru

_______________________________________________ (Name of student)

In memory of the Late Rev. Fr. Linus Zan Mwinlaaro

And to my ancestors: Mwinlaaru and Kognepuor

Since the 1960‘s, the systematic description of languages has gained

INFORMATION. Theme is defined as the element that is given initial

Book Chapter (refereed)

 Mwinlaaru, Isaac N., Christian M. I. M. Matthiessen & Ernest

Journal Articles (refereed)

 Mwinlaaru, Isaac N. & Foong Ha Yap. 2017. A tale of two

Conference presentations (refereed)

 Matthiessen, Christian M.I.M & Isaac N. Mwinlaaru. 2017. Sayerhood

 Matthiessen, Christian M. I. M., Abhishek Kumar Kashyap,

 Mwinlaaru, Isaac N. 2014. Towards an interpersonal grammar of

 Mwinlaaru, Isaac N. 2015. Projection as a fractal motif in Dagaare. In

 Mwinlaaru, Isaac N. 2015. The system of NEGOTIATION in the

 Mwinlaaru, Isaac N. 2015. A systemic functional description of

 Mwinlaaru, Isaac N. 2015. Grammaticalisation in Dagaare: From

 Mwinlaaru, Isaac N. 2015. Polarity and modality in the lexicogrammar

 Mwinlaaru, Isaac N. 2016. The structure and functions of the Dagaare

 Mwinlaaru, Isaac N. 2017. A prosodic-systemic account of ASPECT in

 Mwinlaaru, Isaac N. 2017. POLARITY and MODALITY in the grammar

 Mwinlaaru, Isaac N. 2017. Demonstratives and the division of semiotic

CERTIFICATE OF ORIGINALITY ............................................................................... ii

DEDICATION ............................................................................................................... iii

RELATED OUTPUT .................................................................................................... vii

LIST OF FIGURES .................................................................................................... xvii

LIST OF TABLES ........................................................................................................ xx

ABBREVIATIONS AND SYMBOLS ........................................................................ xxii

CHAPTER ONE ............................................................................................................. 1

CHAPTER TWO .......................................................................................................... 18

THEORY AND APPROACH ....................................................................................... 18

AN OVERVIEW OF THE DAGAARE LINGUISTIC SYSTEM.................................. 50

CHAPTER FOUR ....................................................................................................... 108

MOOD AND MODAL ASSESSMENT ...................................................................... 108

4.5.2 MODALITY ................................................................................................ 172

4.5.2.1. Probability ........................................................................................... 173

THEME AND INFORMATION ................................................................................. 204

CHAPTER SIX........................................................................................................... 256

TRANSITIVITY: REPRESENTATION OF EXPERIENCE ....................................... 256

CHAPTER SEVEN..................................................................................................... 359

SYNOPSIS: DAGAARE AND BEYOND .................................................................. 359

APPENDIX ................................................................................................................ 372

REFERENCES ........................................................................................................... 389

Figure 1.1 Locating the Dagara Speech Community on the Map of

// - tone unit boundary IDEO – ideophone

1.2 Geolinguistic Landscape of Dagaare

1.2.1 Typological Context of Dagaare

Figure 1.2 locates Dagaare in the Niger-Congo phylum. 1 Genetically, it

Figure 1.2: Locating Dagaare in the Niger-Congo Phylum (Adapted from

1.2.2 Varieties of Dagaare: A Dialect Continuum

The name Dagaare designates a variety of dialects with relative mutual

1.2.3 Dagaare in Language Contact

The evolution of the various Dagaare dialects is partly explained by contact