PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or boron date duo. ‘
a

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution

 

LINEAR ORDER IN HAYA VERBAL MORPHOLOGY:
THEORETICAL IMPLICATIONS

By

Yunus Ismail Rubanza

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Linguistics and Germanic,
Slavic, Asian and African Languages

1988

r
0'“

be H"?

ABSTRACT

LINEAR ORDER OF THE HAYA VERBAL MORPHEMES:
THEORETICAL IMPLICATIONS

By

Yunus Ismail Rubanza

This study examined the factors which determine the
linear order and the cooccurrence restrictions of the Haya
verbal. morphemesn .Also the study' discussed these
determining factors SUI relation 1x) Bybee’s (1985) theory,
which claims a correlation of a degree of semantic
relatedness of the inflectional categories to the verb-stem,
and the formal structure of the verb.

The study is divided into four parts: (1) A
phonological sketch of the Haya language, (2) identification
of the verb morphemes and their functions, (3) an account,
where possible, for the cooccurrence restrictions of verbal
morphemes, and (4) the relation of the linear order and the
cooccurrence restrictions to B’s theory of RELEVANCE and
GENERALITY. .

The data for this study comes from Haya books,l
newspapers, transcripts of tape recordings and from the
author, a first language speaker of Haya.

The major findings of this study are: (1) the Haya

language has twenty three verbal morphemes, of which only

Yunus Ismail Rubanza
twenty two are examined in this study. However, not all of
these morphemes do exhibit the maximum degree of
cooccurrence with each other. (2) Four factors are

identified as responsible for the cooccurrence restrictions

of the verbal morphemes. These factors are: phonological,
syntactic, semantic and pragmatic. (3) The linear order of
the Haya verbal morphemes , supports B’ s predictions

concerning hierarchy and order and more generally RELEVANCE
and GENERALITY.

The findings of this study ‘bring to the fore the
observation that a model of the linear order of the verbal
morphemes of Haya (and this may be true of other languages)
must consider not only the semantic relevance of the
morphemes to the verbs but also other aspects of

linguistics, for example, phonology, syntax, and'pragmatics.

This work is dedicated to my late parents:

Ismail Rubanza and Halima Abdallahmani-Rubanza.

iv

ACKNOWLEDGMENTS

I would like to express my sincere gratitude to Dr.
Carolyn Harfordq vdu) as guidance Committee Chairperson and
dissertation director provided support, ~ insight, and
challenge. Dr. Harford was there when needed, and provided
perspectives which kept this study within manageable limits.

I would also like to thank my doctoral committee: Dr.
Ruth Brend, Dr. Barbara Arbbott, Dr. David Dwyer, Dr. Seok
C. Song and Dr. Denise Troutman-Robinson. Acknowledgements
must also be extended to former guidance committee
Chairpersons: Dr. Carol Scotton and Dr. Grover Hudson who
provided guidance during coursework, comprehensive exams and
the early stages of this dissertation.

I would also like to thank the Chairpersons: Dr. D.
Mkude, Prof. J. Wang and Dr. T. Juntune for support and
encouragement. A dept of gratitude is owed to Dr. D. Wiley,
Dr. Y. Fisseha and the staff of the African Studies Center
and Department of Linguistics and Germanic, Slavic, Asian
and African Languages for financial support and supplies.

I also wish to express my deep appreciation to
many who contributed to this project: to University of Dar—
es-Salaam, Michigan State University for tuition and fees,
Smithsonian Institution and Sage foundation for research and
producing dissertation respectively; to Dr. Peter Seitel for
data and guidance, Dr. and Mrs Rugumamu S. for offering me a

place to stay during research implementation in Washington;

V

to Dr. Kenneth Ebert of International Students and Scholars
Office for administrative support; Fr. Deogratias Mutefunya,
Fr and Dr. Peter Rutoyongororwa, Dr. John Rutayuga for
providing me with Haya books, newspapers and guidance; my
friends Robert Mabagala, Dr. Kulikoyela Kahigi, Julie
Chamberlain, Steve Burghess, my students; and the Michigan
State Community who helped make my difficult days seem
bearable.

Finally, a very special dept of gratitude is owed to my
wife Amida and my children Halima, Hashimu and Farida who
lived tolerantly through the long days of studies and
writing. I greatly aknowledge a dept of love, support and

caring.

vi

TABLE OF CONTENTS

LIST OF TABLES D O O O O O C O O I I 0

LIST OF SYMBOLS AND ABBREVIATIONS. . .

CHAPTER 1:

BACKGROUND INFORMATION.

CHAPTER 2:

1.1. Goals. . . . . . . . . . . . . . . . .
1.2. The Language . . . . . . . . . . . . .
1.2.1. Classification . . . . . . . . . . .
1.2.2. Dialects of Haya . . . . . . . . . .
1.3. The People . . . . . . . . . . . . . .
1.4. Publications and data. . . . . . . . .
1.4.1. Published material . . . . . . . . .
1.5. Organisation of dissertation . . . . .
Chapter 1: Endnotes . . . . . . . . . . . .

PHONOLOGICAL SKETCH OF HAYA . . . . .
2.1. Introduction . . . . . . . . . . . . .
2.2. Phonological Inventory . . . . . . . .
2.2.1. Inventory of Consonant Phonemes. . .

2.2.1.1. Distinctive Features of Consonant
Phonemes . . . . . . . . . . . . . . .
2.2.1.2. Redundancy Rules for Consonant
Phonemes . . . . . . . . . . . . . . .
2.2.'. Inventory of Vowel Phonemes. . . . .
2.2.2.1. Distinctive Features of Vowel

Phonemes . . . . . . . . . .

vowels O O O 0 O O O O O 0 O O O 0 O 0

vii

2.2.2.2. Redundancy Rules for Vowel Phonemes.
2.3. The Tones. . . . . . . . . . . . . . . .
2.3.1. Lexical Tone . . . . . . . . . . . . .
2.3.2. Grammatical Tone . . . . . . . . . . .
2.4. Distribution of Consonant and Vowel
Phonemes . . . . . . . . . . . . . . . .
2.4.1. Distribution of Consonants . . . . . .
2.4.1.1. Distribution of Consonants Occurring
Singly . . . . . . . . . . . . . . . . .
2.4.1.2. Consonant Clusters . . . . . . . . .
2.4.2. Distribution of Vowels . . . . . . . .
2.4.2.1. Short Vowels . . . . . . . . . . . .
2.4.2.2. Long Vowels. . . . . . . . . . . . .
2.4.3. Distribution of Consonant Before

Page

xii

xiii

Hood-dozmwan-a

H

12

13
13

16

Page

2.5. Phonological Rules . . . . . . . . . . . 33
2.5.1. Phonological Rules Affecting

Consonants . . . . . . . . . . . . . . . 33
2.5.1.1. Nasal Assimilation . . . . . . . . . 33
2.5.1.2. Consonant Alternation. . . . . . u . 35
2.5.1.2.1. Strengthening. . . . . . . . . . . 35
2.5.1.2.2. Spirantization . . . . . . . . . . 37
2.5.1.2.3. Palatilization . . . . . . . . . . 39
2.5.1.2.4. Morphophonemic Rules . . . . . . . 40
2.5.2. Phonological Rules Affecting Vowels. . 42
2.5.2.1. Glide Formation. . . . . . . . . . . 42
2.5.2.2. Vowel Deletion . . . . . . . . . . . 44
2.5.2.3. Vowel Lengthening. . . . . . . . . . 45
2.5.2.3.1. Lengthening of Identical Vowel

Sequence . . . . . . . . . . . . . . . . 46
2.5. 2.3. 2. Lengthening before Nasal Clusters. 46
2.5. 2.3. 3. Lengthening before n . . . . . . . 47
2.6. A Word on Syllable Structure . . . . . . 48
2.7. General Remarks and Summary. . . . . . . 51
Chapter 2: Endnotes . . . . . . . . . . . . . 53

CHAPTER 3: THE MORPHEMES OF THE VERB FORMS AND

FUNCTIONS . . . . . . . . . . . . . . . . . . 54
3.1. Introduction . . . . . . . . . . . . . . 54
3.2. The Root . . . . . . . . . . . . . . . . 56
3. 3. The Non-Stem Morphemes . . . . . . . . . 60
3.3.1. The Indicative Morpheme. . . . . . . . 60
3.3.2. The Infinitive Morpheme. . . . . . . . 61
3.3.3. The Subject Marker . . . . . . . . . . 63
3.3.3.1. The Noun-Class System. . . . . . . . 63
3.3.4. The Subjunctive Morpheme . . . . . . . 68
3.3.5. The Tenses . . . . . . . . . . . . . . 70
3.3.5.1. The Present Tenses . . . . . . . . . 71
3.3.5.1.1. Present Universal Habitual . . . . 72
3.3.5.1.2. "Still" Tense. . . . . . . . . . . 73
3.3.5.2. The Past Tenses. . . . . . . . . . . 74
3.3.5.2.1. Immediate Past . . . . . . . . . . 74
3.3.5.2.2. Near Past. . . . . . . . . . . . . 75
3.3.5.2.3. Recent Past. . . . . . . . . . . . 76
3.3.5.2.4. Remote Past. . . . . . . . . . . . 77
3.3.5.2.5. Far Remote Past. . . . . . . . . . 79
3.3.5.2.6. Immediate Past Habitual. . . . . . 80
3.3.5.3. The Future Tenses. . . . . . . . . . 81
3.3.5.3.1. Immediate Future . . . . . . . . . 82
3.3.5.3.2. Near Future. . . . . . . . . . . . 85
3.3.5.3.3. Remote Future. . . . . . . . . . . 86
3.3.5.3.4. Immediate Future Habitual. . . . . 88
3.3.5.4. The "Imaginary" -ku- Tense . . . . . 90
3.3.6. The Subject relative morpheme. . . . . 91
3.3.7. The Conditional Morpheme . . . . . . . 95
3.3.8. The Focus Morpheme . . . . . . . . . . 97‘
3.3.9. The Negative Morphemes . . . . . . . . 103

viii

1. NEGl . . . . . . . . . .
2. NEG2 . . . . . . . . . .
3. NEG3 . . . . . . . . .
. The Suffixal Morphemes. .
0.1. The Locative Morphemes.
0.2. The Manner Morpheme . .
0.3. The Intensifier Morpheme
0.4. The Question Morphemes.
The Stem Morphemes . . . . .
. The Object Marker. . . .
. The Reflexive Marker . .
The Verbal Extensions. . .

The Causative Extension.
The Applied Extension. .
The Passive Extension. .
The Reciprocal Extension
The Repetitive Extension
The Stative Extension. .
. Conclusion . . . . . . . .
Chapter 3: Endnotes . . . . . .

wwwwwmwwwwwwawwwwww
C C. O. O .0 O O . O O 0.. C C .

CHAPTER 4: THE MORPHEMES OF THE VERB:
RESTRICTIONS. O O I O O O O I O O O O
4.1. Introduction . . . . . . . .‘. .

4.2. The Cooccurrence Restrictions of

COOCCURRENCE

Prestem

O O O O Q

Morphemes. . . . . . . . . . . .
4.2.1. The Subject Marker . . . . .
4.2.2. The Infinitive Morpheme. . . .
4.2.3. The Tense Morphemes. . . . .
4.2.4. The Subject Relative Morpheme.
4.2.5. The Conditional Morpheme . . .
4.2.6. The Focus Morpheme . . . . . .
4.2.7. The Negative Morphemes . . . .
4.3. The Cooccurrence Restrictions of

Poststem Morphemes . . . . . . .
4.3.1. The Indicative Morpheme. . . .
4.3.2. The Subjunctive Morpheme . . .
4.3.3. The Tense Suffix ~11g. . . . .
4.3.4. The Locative, Manner, Intensifier and

Question Morphemes . . . .'. . .

4.3.4.1. Additional Restrictions on the

Question Morphemes . . . . . . .

4.3.4.2. Additional Restrictions on the

Manner Morpheme. . . . . . . . .

4

Morphemes. . . . . . . . . . . .
4.4.1. The Object Marker. . . . . . .
4.4.2. The Reflexive Marker . . . . .
4.4.3. The Verbal Extensions. . . . .
4.4.3.1. The Causative. . . . . . . .
4.4.3.2. The Applied. . . . . . . . .
4.4.3.3. The Passive. . . . . . . . .

ix

.4 The Cooccurrence Restrictions of Stem

Page

103
104
107
109
109
113
113
114
115
115
125
126
127
129
132
134
135
138
139
140

152
152

154
154
156
162
164
167
170
173

176
176
177
178

180
180
181

182
182
185
188
188
190
193

mmmmmmmm
hOONt-a

hhhtfihhhfi

1
2
3

4.7.

and Non—Stem Morphemes . . .

Cooccurrence Restrictions and Verb Types

4. The Recriprocal.

.5. The Repetitive . . . . . . . . . . .
.6. The Stative. . . . . . . . . . . . .
Cooccurrence Restrictions between Stem

The Subjective . . . . . . .
The Subject Marker . . . . . . . .

The Question Marker. . . .

The Manner Marker. . . . .

Intransitive Verbs . . . . . . . . .

Fossilized Intransitive Verbs. . . . .
. Transitive Verbs . . . . . . . . . . .
Summary and Conclusion . . . . . . . . .

Chapter 4: Endnotes . . . . . . . . . . . . .

CHAPTER 5:
5.1.
5.2.
5.2.1

GOOQNNNNNN

OIUIUIOIUIOIOIUICHOIUTUIO'I

NH

Ocmcnowmcnowmcn
wcnoswcnoaucn

:. . .

3
3
3
3
3
3
3
3
t

ap

CHAPTER 6: CONCLUSIONS AND RECOMMENDATIONS

6.1.

6.2. Suggestions for further research

Summary. . . . . . . . . .

6.3. Contribution . . . . . . .

PRINCIPLES OF VERBAL MORPHOLOGY . . . .
Introduction . . . . . . . . . . . . . .
B's theory . . . . . . . . . . . . . . .
. Predictions arising from Relevance and
Generality . . . . . . . . . . . . . . .
.1. Relevance. . . . . . . . . . . . . .
.1.1. Order. . . . . . . . . . . . . .
.1.2. Frequency. . . . . . . . . . . . .
.1.3. Degree of fusion . . . . . . . . .
.2. Generality . . . . . . . . . . . . .
.2.1. Order. . . . . . . . . . . . . . .
.2.2. Frequency. . . . . . . . . . . . .
.2.3. Degree of fusion . . . . . . . . .
. Synthesis. . . . . . . . . . . . . . .
. Comments on B’s theory . . . . . . . .
Haya as a target of study. . . . . . . .
. Typological position of Haya . . . . .
. Appropriateness of Haya as a test of
8’s theory . . . . . . . . . . . . . . .
. Application of the theory to Haya. . .
.1. The Prestem Morphemes (Non-Stem) . .
.2. The Stem Morphemes . . . . . . . . .
.3. The Poststem Morphemes (non-stem). .
.4. The Subject Marker and Object Marker
.5. The Split between sets of morphems .
.6. Castairs theory. . . . . . . . . . .
.7. Summary and conclusion . . . . . . .
er 5: Endnotes . . . . . . . . . . . . .

Page

194
197
197

199
199
200
201
201
202
203
205
208
209
213

216
216
216

219
221
221
223
224
227
228
229
230
231
232
233
235

239
242
242
244
245
246
267
274
275
277

281
281
283
283

Page

APPENDICES I I I I I I I I I I I I I I I I I I I I 285
APPENDIX A I I I I I I I I I I I I I I I I I I 285
APPENDIX B. . . 299

BIBLIOGRAPHY I I I I I I I I I I I I I I I I I I I 306

xi

Table

Table

Table

Table

Table

Table

Table

Table

Table
Table
Table

Table

Table

10

12

13

Distinctive Features of Consonant and

Glides. .

Distinctive Features with Redundancy
Consonant Phonemes

Rules:

Distinctive Features of Vowel

Phonemes.

Distinctive Features with Redundancy
Vowel Phonemes

Rules:

Distribution of Consonants before

Vowels. .

The Internal Structure of Conjugated

LIST OF TABLES

Verb in Haya.

Inventory of Morphemes and Allomorphs

List of Noun Classes and Subject

Morpheme Markers. .

Haya Affirmative and Negative Tenses.

Subject Relative Morphemes.

Direct/Indirect Objects . . . . . . .

Summary of Cooccurrence Restrictions

of Prestem Morphemes.

Morphemes

xii

'Cooccurrence Restrictions of Stem

Page

17

19

21

22

32

146

147

148
149
150

151

214

215

Alv.

Alv. Pal.

ant

bk

C

cf

00118

cont

001‘

DA

DE

fn

FRP

GC

GEN

hi

IF

IFH

INFL

IP

In Vb

LIST OF SYMBOLS AND ABBREVIATIONS

Alveolar
Alveolar Palatal
Anterior

Back

Consonant

Confer ’compare’
Consonant(al)
Continuant
Corona

Doer of Action
Distinctive Features
Foot Note

Far Remote Past
Glide Consonant
Generality

Human

High

Immediate Future
Immediate Future Habitual
Inflectional
Immediate Past

Inflected Verb

.xiii

IVb
Lab. Dent.
lo
MHV
nas
NC
NH
NF
NPH
NPS
NP(s)
NPSu
NPt
OSuV
OVSu
PAT
pl(p)
PUH
R1--R13
RLV
RMF
RMP
RP
Sg(s)
son

SuOV

Instransitive Verb.
Labio-Dental

Low

Morphemic Formula of the Haya Verb
Nasal

Nasal Consonant
Non-Human

Near Future

Near Past Habitual
Near Past

Noun Phrase(s)

Noun Phrase Subject
Noun Phrase Topic
Object Subject Verb
Object Verb Subject
Patient

Plural

Present Universal Habitual
Rules (1-13)
Relevance

Remote Future
Remote Past

Recent Past
Singular

Sonorant

Subject Object Verb

xiv

SuVo'
STL
Str
Syl
TOP

UNR

Vcd
VP
VOSu
VSuO

WLT

Subject Verb Object
Still (tense)
Strudent

Syllabic

Topic

Unreal (immaginary)
Verb

Voiced

Verb Phrase

Verb Object Subject
Verb Subject Object

Weak Lexalist Hypothesis

XV

 

CHAPTER 1

BACKGROUND INFORMATION

1.1. Goals

In this dissertation we discuss the linear order and
co-occurrence restrictions of the Haya verb morphemes. The
theoretical basis employed in this study is that of Bybee
(henceforth B) (1985), which is discussed in chapter 5.

This study has three related goats. The first is to
identify the Haya verb morphemes and their linear ordering
and to answer the question "is the linear ordering of the
Haya verb morphemes predictable?" This study will attempt
to make generalizations that capture the predictions of
linear ordering.

If the answer to the question above is yes, which our
data seem to suggest, then the second goal is to find out
the factors which determine this ordering. In this study we
will argues that the linear ordering of the Haya verb
morphemes is determined by phonological, syntactic, semantic
and pragmatic factors. The third goal which is closely
related to the above, is to determine whether these factors
follow B’s theory, if not, we shall seek how B’s theory can

be modified to incorporate these factors? Finally, this

2
work, as far as we'know, will be the first detailed study of
Haya verb morphoplogy.

The comparison of B’s work and the present study reveal
both significant similarities and major differences, a point
we will further explore in. detail in chapter 5. Our
procedure is as follows:

a. outline B’s main hypothesis and her arguments.

b. show the possibility of reinterpreting B’s work.

0. present the Haya data, and how it could be
interpreted with regard to B’s claims.

d. Finally, give the discrepancies between the data and
B’s claims we propose a framework in which to view the data.

In the course of our discussion we show how some of B’s
arguments are supported by our data. This leads us to our
major goal of exploring and developing a theoretical
framework in terms of which the linear ordering of the Haya
verb morphemes can be understood and in terms of which the
Haya verb structure in general can be described.

That there is usually a fixed linear order of morphemes
particularly in agglutinating languages, is not an issues,
but rather how to comprehend, them. There are several
proposals, both specific and general which state how such a
linear order can be determined. In view of the limitations
of these proposals, particularly with regard to Haya this
dissertation presents another such proposal, one that

differs from previous proposals mainly in that it attempts

to be highly restrictive, and, in particular, it deals
specifically with the Haya language.

Unlike> many’ previous studies of languages including
that. of B, that have been based on a wide variety of
languages (50 languages in B’s study), our work focuses on a
single language. As further argued in chapter 5, this
approach has an important advantage of depth over a
comparative approach, as Chomsky (1985:6) argues: "A valid
observation that has frequently been made (and often,
irrationally denied) is that a great deal can be learned

about Universal Grammar from a study of a singlewlanguage,

if such study achieves sufficient depth to put forth rules
or principles that have explanatory force. . . ." (emphasis
mine). Our study undertakes an in-depth study of Haya verb

structure in order to uncover sufficient detail about the
nature of Haya linear ordering so that we can offer an

evaluation, in light of B’s work.

1.2. The Language
Haya is a Bantu language spoken in the Northwestern
part of Tanzania to the west of Lake Victoria. It is spoken

in the Bukoba and Muleba districts in the Kagera region.

1.2.1. Classification
Guthrie (1970:13) classified Haya under zone E.20 Haya—

Jita group which includes:

E.21 Nyambo (Karagwe)

E.22 Ziba (erroneously called Haya)

E.23 Dzindza (Jinja)

E.24 Kerebe (ukerewe)

E.25 Jita (Kwaya)
However, scholars including Copes (1971), Meeussen (1976),
and Nurse (1969) seem to disagree with him in his separation
of some languages in zones D and E.

The more convincing classification of the Haya language
is by Meeussen (1976) whose classification is composed of
Guthrie’s zone D and E. This classification is composed of

the languages most chsely related to Haya:

 

 

 

Meeussen’s classification Guthrie’s Zone Country
Nyankore E:13 Uganda
Ciga E:14 Uganda
Ganda E:15 Uganda
Nyoro E:11 Uganda
Nyaruanda 'D:61 Ruanda
Rundi D:62 Burundi
Subi D:64 - Tanzania
Ha D:66 Tanzania
Vinza D:67 Tanzania
The classification issue is not settled yet. In a

recent study, Nurse (1979:2) quoted Mosha as saying that

5

"The Ugandan survey indicated that several of the
'interlacustrine languages (which included. Haya) form. one
large group, which is called Rutara, . . . . The languages
involved. were .Nyankore, Kiga, Nyoro, and. Tooro."l Nurse
further quoting Mosha further notes that "From the Tanzanian
point of 'view, Rutara should be expanded to include Haya
(with. ‘Nyambo), Zinza, and. IKerewe." Regarding the
relationship between these languages he points out that,
"All these languages are almost equally close: a speaker of
any of them listening to another on the radio would have
little brouble understanding; they all operate variants of
the same tone system. . . .” Nurse (1979:2) also notes that
"Rutara is also linked to:

a. Ha, Hangaza, Shubi, Ruanda, Rundi, Vinza;

b. Jita, Kwaya, Ruri, Regi;

c. Gusii, Kuria, Ngurumi, etc;

d. Luhya;

e. Ganda, Soga, Gwere, etc. . . .

and these together form what is called the Lacustrine

group."2

1.2.2. Dialects of Haya
Haya has eight dialects, corresponding to the location

of former chiefdoms:

 

 

Dialect Chiefdom
Ziba Kiziba
Hamba Kianja
Hangiro Ihangiro
Nyakisaka Misenye
Yoza Kyamtwara
Endangabo Bugabo
Bumbira Bukara
Mwani Karagwe

 

The dialects are mutually intelligible.

A speaker of any of

these dialects has little trouble understanding any of the
others, as the author can testify such as he is a native

speaker of the Ziba dialect.

1.3. The People

The speakers of the language known as AbaHaya, ’the
Haya people’ call their language OluHaya. The Haya speakers
mainly inhabit Bukoba and Muleba districts in' the Kagera
region in Tanzania.3 They are surrounded by speakers of
other Bantu languages: Nyaruanda (Ruanda) to the west, Ciga,
Ganda (Uganda) in the north; Zinza and Hangaza (Tanzania) to
the south. On the eastern side is lake Victoria. However,

Hayas are currently scattered all over Tanzania mainly in

7

the cities of Dar-es-Salaam and Mwanza working in government
and private offices, and some engaging in commerce.

The population of the Haya people is estimated at about
a million or more. This is merely an estimate because the
1977 Tanzanian census abolished the identification of the
population by their ethnic groups.‘

Haya living in the rural areas are primarily farmers;

they grow bananas and plantains which constitute their

staple food. Beans, sweet potatoes, yams, millet, corn,
tomatoes and cassava are supplements. of their diet. The
principle cash crops are coffee and tea. Other Haya people

who live near lake Victoria and along the banks of river
Kagera do some fishing in addition to farming. The Haya
also raise cattle, goats and other livestock. They also

engage in small commercial activities.

1.4. Publications and data

1.4.1. Published material

As far as literature is concerned, there are few
published materials on the Haya language in general and no
extensive work on the morphology of the Haya verb. General
works on Haya include that of Betbeder and Jones (1949);
Rascher (1958) on the grammar of the Haya language; Bona-
Baisi (1950), a Haya dictionary; Byarushengo (1975), a
phonological sketech of Haya; Byarushengo et al. (1977)

Haya grammatical structure; Hyman and Byarushengo, a model

8
of Haya verb tonology in Clements and Goldsmith (1985).5

Other related works on Haya include: Rubanza (1979), the
relationship between Kiswahili and other African langauges:
the case of Haya and Seitel (1980) a collection of Haya
stories, proverbs and heroic poetry. There is also

literature on the history of Haya such as that of Lwamgira’s

(1949) "History of Kim," one of the former chiefdom of
Hayaland, and various Bible, song and religious primers. In

addition, there exists a monthly' newspaper, RUMULI ’the
torch,’ published by the Bukoba Roman Catholic Diocese.

Some of the data of this dissertation comes from the
above citations and specifically from the following
articles: Byarushengo (1977:205-213), a word list; Dalgish
(1977) ’Past tense formation in OruHaya’ and Seitel’s Haya
epic collection (1970’s-1’980’s).6 The author of this
dissertation, a native speaker of the language, also

provides a major source of the data.

1.5. Organisation of the dissertation

In chapter 2 we will outline a phonological sketch of
Haya. Particular attention is paid to the aspects of
phonology that are of relevance to the analysis of the Haya
verb in general and the linear order of the Haya verb
morphemes in particular. The phonological outline is
provided to assist the reader and preclude discussion every
time phonology comes up in the course of the morphological

analysis. Moreover, the provision of the phonological

9

outline allows the handling of a particular alternation as
morphophonemic if and when it ‘cannot be explained
phonologically. These reasons provide the justification for
having a chapter on phonology when, as a matter of fact, the
present work deals mainly with morphology, a point further
taken up in chapter 2.

Chapter 3 contain an inventory and provides a detailed
description of the morphemes that constitute the Haya verb.
The functions of each of these morphemes is given. The
order of the morphemes is not discussed a task which is the
basis of chapter 4.

Chapter 4 specifically deals with the linear order of
the verbal morphemes and their co-occurrence restrictions.
In this chapter two claims are made and substantiated:

a. that there is a interrelationship between the
location of the stem morphemes and the root,
depending on the number of arguments a verb can
take.

b. that four related factors determine the order of
occurrence and co-occurrence; restrictions of the
Haya verb morphemes: (i) a phonological factor,
(ii) a syntactic factor, (iii) a semantic factor,
and (iv) a pragmatic factor.

Chapter 5 lays out the essential points of B’s theory.

Her two major principles Relevance and Generality are
discussed and exemplified especially with respect to the

predictions they make concerning ordering, frequency and

10

degree of fusion. Chapter 5 also discusses the implications
of the Haya data to B’s theory. This exercise sorts out
those claims that are supported by the Haya data and those
that are not. In the light of these developments, we offer
modified version of B’s theory regarding linear ordering.
Chapter 6 summarizes of the main findings of the
dissertation, suggestions for future research. These
suggestions are drawn from the findings of the Haya data,
which have relevance to other Bantu languages in particular

and other languages of the world in general.

11
CHAPTER 1 ENDNOTES

1 Nurse (1979:146) says in a footnote that Mosha’s
remarks are from Ladefoged (1972:69).

3 Nurse (1979:2) claims that he used what he termed a
Lexico-statistical method which is not defined at least in
this paper.

3 Bukoba and Muleba have been two separate districts
since the mid 1970’s. Otherwise it was one district
(Bukoba).

4 The current population data on Tanzania is that of
1977, in which the population of the residence of Bukoba and

Karagwe districts were.

Karagwe 185,016
Bukoba rural 256,349
Muleba 217,493
Bukoba urban 77,519

738,277

Rate of growth from the 1977 census was + 42.5%. But as
it was mentioned elsewhere, Hayas are scattered all over the
region. and the» nation, particularly' in big cities (see
Bureau of Statistics (1982)).

5 Hyman and Byarushengo’s (1984) work is the only
detailed study on Haya tone we know of. The work is
enriched with many examples at the end of the chapter.

5 I am deeply grateful to Dr. Peter Seitel who allowed
me to use part of the data he collected from Bukoba. (see

appendix 1).

CHAPTER 2

PHONOLOGICAL SKETCH OF HAYA

2.1. Introduction

The presence of a chapter on phonology in the present
study of morphology is explained in this introductory
section.1 In addition to accurately representing
distinctions in information needed to understand the surface
morphological. representations a jphonological. study should
reveal an interrelationships between the study of phonology
and the study of morphology, as well as other parts of
grammar. This is especially true of morphology as Selkirk
(1982:59) who, when dealing with affixation argues that "Any
theory of affixation must allow for grammars that represent
explicitly and perspicuously the grammatically relevant

information that is idiosyncratically associated with a

particular affix morpheme." Selkirk further notes that this
information is of three varieties namely: syntactic,
semantic and phonological. Regarding the phonological

variety, Selkirk (ibidz63) notes that the first property to
be represented is information concerning pronunciation of
the affix. "This will include; minimally a distinctive
feature matrix representing the composition of the affix."

Finally, Selkirk mentions what he called the idiosyncratic

12

13

phonological property of affixes, which is "their
phonologically unpredictable effect on the pronunciation of
surrounding morphemes, for example, certain affixes trigger
certain types of allomorphy in other morphemes" (Cf Dell and
Selkirk (1978), Aronoff (1976)).

In the current study, the description of phonological
processes has been limited to those that have impact on the
Haya verb morphology. The basic features of phonemes of the
language are first given followed by redundancy and

phonological rules.

2.2. Phonological Inventory

Haya phonemes are consonants (including true consonants
and glides) and vowels distributed segmentally and
sequentially in the syllable. Vowel phonemes are syllabic
and tone bearing and obligatory in every word.
Haya has:

19 consonant phonemes

10 vowel phonemes

2 tones.

2.2.1. Inventory of Consonant Phonemes.
There are nineteen consonant phonemes in Haya, given in

(1) and Illustrated in (2):2

l4

(1) Bilabial Lab.Dent Alv. Alv.pal. Pal. Velar Glottal
Stops p,b t,d k.g
Affricate 5,!
Fricative f s,z s h
Nasal m n ‘3
Lateral l
Glide . y w
The alveopalatal phonemes will henceforth be

orthographically written as follows:

/é/ = ch
/5/ = J

/§/ = sh
/K/ = ny

The phoneme 1 has two allophones: [l] and [r]. In this work
the same symbol 1 is used for both [1] and [r]. All of the
above phonemes have the phonetic value usually associated
with these phonemic representations, except for /l/, which

has free variants [l] and [r].

~(2) /P/ /t/
juguta ’blow’
paikuka ’to be uncontrolled’ tiila ’borrow’

e.g. child’

paapala

empako
/b/
bona
shoba
gamba
lk/
lekela
ziika
kuula
lg/
gaga
yoga

ngoma

/f/
enfuka

faakala

15
’to get lost e.g.

in unknown land’

’a cave’

’see’

’30 bad’

’say’

’cease’
’bury’

’uproot’

’go bad’
’take a shower’

’drum’

’a hole’
’make a mistake

(sin)’

fafalikana ’become unhappy’

kuntila ’foot fall’
/d/
tonda ’create’
dindind ’deep inside’
dini ’religion’
/é/
chumba ’cook’
chuchula ’comb hair’
okuchwa ’to spit’
/5/
juba ’fish’
olubaju ’side’
janjaba ’help e.g. a
sick person’
/8/
sinda ’moan’
musigazi ’a youth’
lemesa ’cause someone
to be tired’
/z/
zimba ’swell’

gendeleza ’continue,
develop’

nzilo ’dirt’

lé/
shashula
mushanju
gasha

lm/

manya

omwongu
lemala
/3/

nyiikila

enyenje
okumanya

ly/

yanika

iya

hooya

2I2I1I1I

The distinctive features for these consonant phonemes

are given in table 1.3

and

[-consonantal].

(see 2.2.2.1).

16

’pay back’
’seven’

’be useful’

’know’

’pumpkins’

’become lame’

’put more
effort’
’mane’

’to know’

’put on the sun
to dry’
’uproot’

’sing’

[+consonantal]. The

Vowels are

glides

/h/
hanuula
bihuka
omuguha

/n/

inuka

igana

nigaala
/l/

laalika

shaalila
ilagula
/W/

wetole

iwe

[+syllabic]

are

and

’give advise’
’insects’

’r0pe’

’go home from
work’

’be equal’
’get anger’

’invite’

’become bitter’

’become black’

’steal’

ryour

Distinctive Features of Consonants Phonemes

All true consonants are {-syllabic]

l-syllabic]

{-consonantal]

17

Table 1. Distinctive Features of Consonant and Glides.

 

 

syl.
SOU- ‘ - - - - - - - - - - - - + + + + + +

cons.+ + + + + + + + + + + + + + + + + - -

cont.- - - - - - - - + + + + + - - - - + +
nas. - - - - - - — — - - — — — + + + - - -
ant. + + + + - - - - + + + '- - + + _ + - _
cor. - - + + + + - - - + + + - - + - + - -
vcd. - + - + - + - + - - + - - + + + + + +
str - - - - + + - - - + + + — - - - - - -

 

2.2.1.2. Redundancy Rules for Consonant Phonemes
Important redundancy rules pertaining to consonants and

glides are:

 

 

(3) R1 F .
-syl --> -nas
-cons ~ant
+cont
+son
. J
R2
+cons
+cont --> [-vcd]
-cor

R3 [+son] --> [+vcd]

R4

R5

R7

R8

R9

R10

18

‘80!)

[+cons] —-> [-syl]

[-son] --> {-nas}

[+nas] -—> -cont.
+son

-cor --> [-str]

+ant

-cont

-son --> {-cor}

-cont

-ant

-son --> [+str]

+cont

+cor

Table 2 is a repeat of table

accounted for by redundancy rules in

1

10 refer to redundancy rules.

1

(3).

with some spaces

The numbers e.g.

19

 

Table 2. Distinctive Features with Redundancy Rules:
Consonant Phonemes.
pbtdéjkgtszéhmnxlyw

 

syl 4 4 4 4 4
son - - - - -

cons + + + + +

cont - - - - -
nas 5 5 5 5 5
ant + + + + -
cor - - + + +
vcd - + - + -
str 8 8 - - +

- - — + + + + - 7 7 7 1 1
5 5 5 5 5 5 5 5 + + + - 1 1
- - - + + + - - + + - + 1 1
+ 9 9 - + + + - - + + + + -
+ - + 2 10 10 10 2 3 3 3 3 3 3

 

2.2.2.

Haya has ten phonemic vowels,

(of.

illustrated in (5).

(4) i ii

also Byarushengo

Inventory of Vowel Phonemes

five short and five long
1977: 3-4) as given in (4) and
u uu
o 00
a
aa

These have the phonetic values usually associated with their

phonemic representations.

20

(5)
i sima ’knock ii siima ’thank
something’ (someone)’
i bika ’announce ii biika ’keep’
death’
e tega ’set up a ee teega ’curse’
trap’ I
lega ’trap water’ leega ’strech’
u balula ’burst, open’ uu baluula ’tell’
chuchula ’comb hair’ chuuchuula ’run fast’
0 kolola ’cough’ oo kooloola ’bring things
improperly’
yola ’get poor’ yoola ’get things
from the ground’
a laza ’set fire wood aa laaza ’extinguish’
_ extensively’
shalila ’take off leaves shaalila ’become bitter’

of the banana

plant’

2.2.2.1. Distinctive Features of Vowel Phonemes

The Haya vowels are categorized in a distinctive
feature) (DF) matrix in table 3. All vowels share the
feature [+syllabic] which distinguishes vowels from
consonants. Phonetically, short and long vowels occur, but
phonologically the long vowels will be analyzed as sequences

of two vowels.

21

(7) Va Va --> lV:]

Table 3. Distinctive Features of Vowel Phonemes.

 

 

1 e a o u
syllabic + + + + +
high + - - - +
low - — + - -
back - - + + +

 

2.2.2.2. Redundancy Rules for Vowel Phonemes.
Certain redundancy rules allow us to predict some
features in the DF matrix. Certain feature values are

abstracted from table 3 by redundancy rules R11-R13

(3)
R 11 [+hi] --> {-101
R 12 [+10] --> [-hi, +bk]
R 13 [-bk] --> {-10}

Table 4 include numbers that correspond to the

redundancy rules in (8) above.

22

Table 4. Distinctive Features with Redundancy Rules: Vowel

Phonemes.

 

 

i e a o u
syllabic + + + + +
high + - 12 - +
low 11 13 + - 11
back - - 12 + +

 

2.3. The Tones

Two levels of tones are distinguished: High (H) and Low

(L) as exemplified below:‘ Note that debate of whether or

not tone is a feature of vowel or syllable is

scope of this study:

(9)

’pregnancy’
’louse’
’break’

’her husband’
’steal’
’finish’
’intestines’
’pull’

’inherent’

beyond the

23

The two functions of tone are of two types:
(1) Lexical

(2) Grammatical

2.3.1. Lexical Tone
Lexical tone in Haya as in other Bantu language serves
as a feature distinguishing one lexical item from another as

shown in (9).

2.3.2. Grammatical Tone

Grammatical tone in Haya distinguishes number, and
tense/aspect. For example a high tone pattern in the
present universal habitual tense, distinguishes the first

person singular from the third person singular:

(10)
n -5 -gend-‘
SM-TNS-go -IND
’I am going’
n -a -gé'nd-A’?
SM-TNS-go -IND

’is he going?’

In what follows, tone will not be marked unless it

distinguishes lexical and/or grammatical categories.

 

24

2.4. Distribution of Consonant and Vowel Phonemes.
The distribution and frequency of occurrence of
consonant and vowel phonemes in Haya words are discussed

and exemplified below.

2.4.1. Distribution of Consonants

2.4.1.1 Distribution of Consonants Occurring Singly

All consonants phonemes occur singly word initially,
and between vowels.5 One example of each occurrence is given
for each consonant phoneme. See more examples in appendix

B.

(11) Initial Position

/p/ pima ’measure’

/b/ byaala ’plant’

/t/ telela ’slip’

/d/ dura ’to uplift-oneself’
/k/ kinga ’close e.g. door’
/g/ gangabula ’split’

/5/ chunda ’shake’

/3/' juna ’help’

/f/ fuha ’blow nose’

/s/ sindika ’ ’push’

/z/ zana ’play’

/§/ shona ’sew’

/h/ hooya ’sing’

/m/
/n/
NH
/1/
/y/
/w/

(12) Between
/p/
/b/
/t/
/k/
/3/
/5/
/3’/
/f/
/s/
/z/
/§/
/h/
/m/
/n/
/?f/
/1/
HI
AM

25

myama
nunka

nyyama
lobeka
yoleka

wazambuka

Vowels
paapala
babiilila
yetonge
ikala
bugangana
chachalika
yajula
fufuza
sisiimuka
bazangana
shashula
hohahoba
lemala
nanura
monyoka
guluka
gayangana

wawalikana

’sleep’

’smell’

’sleep’

’soak’

’show’

’come out of sleep

suddenly’

’to run out of fear’
’itch’

’deny’

’dwell’

’meet’

’talk nonsense’
’crow’

’to go in a wrong direction’
'wake up’

’ask each other’

’pay back’

’to be frightened’

’become lame’

’cure’

’break off’

’jump’

’despice each other’

’to be unclear of what one

is doing’

26

The phoneme (1 does not appear between. vowels (see
2.5.1.2.1).

There are no word final consonants.

2.4.1.2. Consonant Clusters
There are two types of consonant clusters: (a) a nasal
consonant followed by another consonant (henceforth

abbreviated as NC),

and

(b) a glide followed by another

(true) consonant (henceforth GC).

/Np/
/Nb/
/Nt/
/Nd/

/Nk/

/N8/
/N6/
/N3/
/Nf/
/N8/
/N2/
/N§/

-mpako
-mbuzi
-ntale
-ndwana
-nkilabinu
-nkeremeke
-ngoma
-nchumita
-njoka
-mfuzi
-nseiso
-nzilo

-nshanani

’cave’

’goat’

’lion’

’1 fight’

’the middle finger’
’baby’

’drum’

’1 spear’

’snake’

’orphan’

’a stone used to grind’
’dirt’

’gap (tooth)’

27
(13.b) NC: between vowels

/Np/ -mpolampola ’slowly’

/Nb/ -kambula ’tear down’

/Nt/ -muntu ’a person’

/Nd/ -muhanda ’path'

/Nk/ -mukankaburo ’a banana which is not
mature’

/Ng/ -omungemu ’in the banana field’

/N6/ -okuchinja ’to slaughter

/N3/ -mwekubanjuro(s) ’tiredness’

/Nf/ -enfuka ’a hole’

/Ns/ -sinsimuko ’dirt’

/Nz/ -enkwanzi ’bead’

The following examples show the GC consonant clusters
in medial position which is the only position in which they
occur. Although. they' are <orthographically' presented as
sequence of two vowels, they are phonologically consonant

clusters:

(14) a. GC: geija --> [geyja] ’become fat’
boigola --> [boygola] ’bark’

chaichiliza --> [6ayéiliza] ’tease’

28

2.4.2. Distribution of Vowels

2.4.2.1. Short vowels

Short vowel phonemes appear word initially, between

consonants and word finally as in (15)-(17).

(15) Initial

/i/ ikiliza ’agree’
inama ’bend’
ija ’come’
/e/ eya ’sweep’
enda ’break’
ebembela ’lead e.g. way’
/u/ ulila ’listen’
umuza ’put down’
uma ’become blind’
/o/ oloola ’answer’
ondela ’follow’
oba ’be frightened’
/a/ ababikana ’come out with a torrent of
words’
abya ’demolish e.g. a house’
ahula ’clean the house’

In dialects of Haya other than the Ziba dialect the vowel
phoneme u does not occur in word initial position, instead

we find hm in those dialects.

(16) a

(17) Between Consonants

(18) Final

/i/

/e/

/u/

/o/

la/

/i/

bika
hika
lila
leka
lela
tema
juma
kula
bula
bona

yoga

shomba

baza

shaba

kama

mulimi
muzaani

mwombeki

29

’announce death’
’arrive’

’cry’

’leave (it)’
’bring up child’
’cut down’
’abuse’

’grow’

’get lost’

’see’

’bathe’

’carry things’
’ask’

’as for’

’milk cow’

’farmer’
’player’

’builder’

30
/e/ a-lagil—e ’he should give message’

SM-give-SUBJ

ba-kul-e ’they should grow’
SM—grow-SUBJ
tu-jun-e ’We should help’

SM-help—SUBJ

/u/ ichumu ’spear’
ndimu ’lime’
bizimu ’ghosts’

/o/ koyo ’a small heart’
kalimo ’a small job’
mulalilo ’promise’

/a/ lagila ’give message’
kula ’grow up (in age)’
kulima ’farming’

The short vowels i, _u_, g, and a appear word finally in
nouns, a appearing only in infinitival nouns. Verbs always
end in a, and only end in g in subjunctive and perfective

forms.

2.4.2.2. Long Vowels
Long vowel phonemes appear between true consonants and

glides:

31

(19) /i/ biiya ’become bad’
ziika ’bury’
biika ’keep’

O/e/ leeba ’look’
teega ’curse’
teema ’play around’

/u/ juula ’undress’
baluula ’tell’
shuula ’greet’

/o/ goola ’ashame’
yoola ’get things from the ground’
shooba ’go slowly’

/a/ zaana ’play’
shaaga ’remainfl
shaabuka ’cross over’

Long vowels appear neither word-initially nor word finally.

2.4.3. Distribution of Consonants Before Vowels.

Table 5 shows the distribution of consonant phonemes
before vowel phonemes word initially. It shows that the
consonants j, g and the semi-vowels x and w are not followed

by the vowel i. The vowel e does not follow the consonants

n and d. The consonant phoneme d is also not followed by
the phoneme vowel a. Consonants s, g and the semi-vowel g
are not followed by the vowel u. These cases are regarded

as accidental gaps in the language (except w before u which

32

Table 5. Distribution of Consonants before Vowels.6

 

 

/i/ /e/ /a/ /o/ /u/
/p/ + - + + +
/b/ + + + + +
/t/ + + + + +
/d/ + - - + +
/k/ + + + + +
/g/ + + + + +
/e/ + + + + +
/j/ - + + + +
/f/ + + + + +
/s/ + + + + -
/z/ + + + + -
/§/ - + + + +
/h/ + + + + +
/m/ + - + ' + +
/n/ ~ + + + + +
/3/ . I + + + + +
/1/ + + + + +
/y/ 1 - + ‘ + + +

/w/ - . + + - -

 

33

is principled). This needs historical explanation, which is

beyond this study.

2.5. Phonological Rules
The following phonological rules affecting vowels and

consonants, are observed in Haya:

Consonants: Nasal assimilation and consonant
alternation.

Vowels: gliding, vowel deletion, and vowel
lengthening.

2.5.1 Phonological Rules Affecting Consonants
The following are phonological rules affecting

consonants:

2.5.1.1. Nasal Assimilation
A nasal shares place of articulation with the following

consonant.

(20)

B cor

[+ nas] --> [a ant] / ---- [- syl]

34
(21)

n-bul -a --> m-bula --> mbula
I-get lost-IND

’I get lost’

n-b -a -—> m-ba --> mba

I-be-IND

’I become’

n-f —a --> m-fa --> mfa
I-die-IND

'I die’

The data show that the phoneme n always assimilates to
the place of articulation of the following consonant within
a word" The n- above represents the first person subject
marker in a verb. As mentioned in section 2.3.2. this rule
only concerns the only prefix n of the language (the first
person).

In addition, the nasal becomes syllabic word initially

if followed by a non-nasal consonant.

(22) [+nas] --> [+syl]/# --- -syl
-nas
(23) n-pa --> [mpa] ’I give'
n-bika --> [mbika] ’I announce death’

n-kunga --> [nkunga] ’I shout’

35

However, if the nasal consonant is followed by another

nasal consonant, the nasal cluster consonant is shortened.

(24) [+nas] [+nas] --> [+nas]/# ---

(25) n-manya --> manya ’I know’
n-mila --> mila . ’I swallow’
n-mala --> mala ’I finish’

2.5.1.2. Consonant Alternation

The following alternation occur in Haya.7

2.5.1.2.1. Strengthening
The contrast of I and d is neutralized in favor of g,

after a nasal g.

(26) C
+cor
+ant --> [-son]/[+nas] ---

-nas
-cont

(27) a. o -n -lim -ila --> ondimila
SM-OM-STEM-
you-me-farm-for-BEN—IND

’You farm for me’

36

b. a —n -lum -a --> anduma
SM-OM-STEM-IND
he-me-bite-IND

’he bites me’

o. n -leet -a --> ndeeta
SM-bring-IND

’I bring’

Another example of strengthening is exhibited in the

neutralization of h, and p in favor of p after the nasal

phoneme n.

(28) C
-cor
-vcd --> [-cont]/[+nas] ---
-str '
(29)
a. n-hango --> m-pango --> mpango

it (e.g. cow) big

’it is big’

b. a -n -hiig -a --> a-m-piiga --> ampiiga
SM-OM-STEM -IND
he-me-look for-IND

’he looks for me’

37

c. n -hum -a --> m-puma --> mpuma
SM-STEM -IND
I -blind—IND

’I become blind’

2.5.1.2.2. Spirantization
The contrast between the three consonants I, d and yg

is neutralized in favor of z before the high front vowel i.

(30) v
+COI‘
+vcd --> [+contl/ ---- +hi
-cont -bk
(31) a.

(i) a -kol -a
SM-work-IND
a -kol -ile --> akozile
SM-work-TNS
’he/she has worked’
(ii) ba-laa-shol -a
SM-TNS-pay tax-IND
ba -laa-shol -ile --> balashozile
they-TNS-pay tax-TNS

’they have already paid tax’

38
b.(i)

mu-lond-a
SM-pick-IND
mu -lond-ile --> mulonzile
you(pl)-pick-TNS
’you(pl) picked’
(ii) ba-lind -a
SM-guard-IND
ba-lind-i --> balinzi
SM-guard

’guards’

c.(i)
ba-enj -a
SM-search-IND
ba-enj-i --> baenzi
’searchers’

(ii) mu-ganj -a
SM-be loved-IND
mu-ganj—i --> muganzi

’lover’

Finally, the contrast of sh and s is neutralized in

favor of s before i.

~39

 

 

(32) r “ V
-son
+cont --> [+ant]/ -—-- +hi
+cor -bk
-vcd

(33)

a. shash-a
STEM —IND
’get sick’
shash -il -a --> shasila

get sick-APL-IND

’get sick/hurt for . . .’
b. lash -a
throw-IND

lash -il -a --> lasila
throw-APL-IND

’throw at/for . . .’
See also 2.5.1.2.4 (37) on page 42).

2.5.1.2.3. Palatalization
The contrast between the nasal phonemes n

neutralized in favor of n before the high vowel i.

(34) v

+nas --> [-ant]/ ---- +hi
+cor -bk

and :1 is

40

(35)
a. ni -n-0 -ikiliz-a --> ni-ny-ikiliza --> ninyikiliza
FOC I-TNS-agree -IND

’I am agreeing’

b. n-O -inam-a --> ny-inama --> nyinama
I-TNS-bend-IND

’I bend’

c. ba -n -0 -iluk-a --> ba—ny-iluka --> banyiluka
they-me-TNS-run -IND

’they run away from me’

2.5.1.2.4. Morphophonemic Rules

Consonant alternations involving t, g, g, and s are

phonologically unnatural, and are morphologically
conditioned.
(36) t

a —--> s/ ---- -ile

recent past

where -ilg is the suffix recent past and the discontinuous
morphemes of the near tense, explained fully in chapter 3

page 76.

41.
(37)

a.(i) a -it -a
SM-kill-IND
a -it -ile --> aisile
he—kill-TNS
’he has killed’
(ii) mu-jugut-a
SM-blow -IND
mu -jugut-ile --> mujugusile
you(pl)-blow -TNS

’you have blown’

b.(i) n-chaich -a
I-disturb-IND

(ii) n-chaich -ile --> nchaisile
I-disturb-TNS

’I got disturbed’

c.(i) tu-laa-shash -a
SM-TNS-get sick-IND
"we-will get sick’

(ii) tu-la -shash -ile --> tulashaasile
SM-TNS-get sick-INS

’we have already got sick’

As evidence of the above claim, we cite examples in (38)

where no alternations of 111.5: 5!; occur in the phonological

42

context —ilg. The s s alternations do appear, (38.0) as we

explained in (32).

(38)
a.(i) ba-it -a
SM-kill-IND
’they kill’
(ii) ba -it —il -a --> baitila

They kill-APL-IND

’they kill for/at . . . ’

b.(i) n-chaich -a
I disturb-IND
’I disturb’

(ii) n-chaich ~iliz-a -—> nchaichiliza
I-disturb-RPT -IND

’I become disturbed repeatedly’

2.5.2. Phonological Rules Affecting Vowels

The phonological rules affecting vowels in Haya are
documented in Byarushengo (1975, 1977). A few examples and
illustrations are shown here. The rules are: glide

formation, vowel deletion and vowel lengthening.

2.5.2.1. Glide Formation

The contrasts of i and e with y and of u with 31 are

neutralized in favor of glide before vowels.

43

(39) a.
V
{-bk] --> -syl -—-- V
[+hi J
b. V
[+bk ] --> {-syl] / --- V
+hi
(40) a. i. ku —li -a --> kulya
INF-eat-IND
’to eat/eating’
ii. ba-laa-li-end -a —-> balyenda
SM-TNS-OM like-IND
’They will like it’
iii. ba-li -ki-ongel-a --> bakyongela
SM-TNS-OM-add IND
’they will add to it’
iv. tu-O -bi-umuz -a --> tubyumuza

SM-TNS-OM-put down-IND

’we put them down’

As was pointed out in rule 16, in most of the other Haya

dialcts, the last example would be realized as tu-bi-umuza

—-> tubihumuza ’we put them down,’ because the phoneme u is

realized as hu verb root initially.

44

(40) b. i. ku -igan -a -—> kwigana

INF-be equal-IND
’to be equal’

11. ku -ey -a --> kweya
INF-sweep-IND
’to sweep’

iii. ku -og -a --> kwoga
INF-swim-IND
’to swim’

iv. ku -at -a --> kwata
INF-break-IND

’to break, to catch’

Rule (7) repeated here as (41), which forms a long vowel
from a phonological cluster of like vowels applies before

(39) to prevent 11 --> 2;, up --> HQ etc.
(41) V¢V¢ --> [V:]

2.5.2.2. Vowel Deletion

In Haya, there are exceptions to rule (39.a.b.).
Byarushengo (1977:5) correctly indicates that although both
the negative marker ti- and the pmesent universal habitual’
marker ni- end in i, this vowel does not undergo glide
formation. Instead, it undergoes vowel deletion. However,
this process is morphologically conditioned, appearing only

after negative (NEG) and the present universal habitual.

45
(42) v
NEG; --> g / ---- - hi
FOC + bk

(43)
a. ti -a -gend-a --> tagenda
NEG SM-go -IND

’he doesn’t go’

b. ni -a -gend-a --> nagenda
FOC-SM-go -IND

’he is going’

c. ni -o -shom -a --> noshoma
FOC-SM-study-IND

’you are studying’

2.5.2.3. Vowel Lengthening

Vowels are predictably long in Haya as in Kinyaruanda,
a language closely related to Haya (Kimenyi (1979)) in the
following environments: (1) when a morpheme ending in a
vowel precedes another morpheme that begins with an
identical vowel, (2) when a.vowel precedes a nasal cluster

within stems, and (3) before the morpheme n - ’I/me.’

46
2.5.2.3.1. Lengthening of Identical Vowel Sequences

When a morpheme ending in a vowel precedes another
morpheme that begins with an identical vowel, the sequence
of the two identical vowels are realized as long as was

shown in rule (7) repeated here as (44).

(44)

v.v._--> [v:]

(45)
a. ku -li -ik-a --> kuliika
INF-eat-st-IND

’to be edible’

b. ku -ulil -a --> kuulila
INF-listen-IND

’to listen’

0. ba-a -lim —a --> baalima
SM-TNS-cultivate-IND

’they are cultivating’

2.5.2.3.2. Lengthening before Nasal Clusters
Vowels are lengthened if they occur before nasal

consonant clusters.

47

(46)

linda --> liznda ’wait’

tunga --> tuznga ’get rich’

kenga --> keznga ’suspect’

mu ’ -shondo --> musho:ndo ’gift’

SM(-¢)-8ift

lamba --> lazmba ’lick’
(47)a.

[+ syl J
[+syl] --> + long / ----- NC, where NC is nasal +

cosonant clusteru But because we .have regarded the .DF
[+long] as redundant in Haya, (47.a) can be formulated as
(47.b).

b. V --> V:/ ---- NC

2.4.2.3.3. Lengthering before n
A ‘vowel is lengthened if it occurs preceeding the

morpheme n- ’me’

(48)
a. ku -n -mil -a --> ku:mila
INF-OM-swallow IND

’to swallow me’

48

b. ku -n -mal -a --> ku:mala
INF-OM-finish-IND

’to finish me’

c. ku -n -nag -a --> ku:naga
INF-OM-throw-IND

’to throw me’ (not take care of me)
Lengthening can also be interpreted as in (49).

(49)

+ syl
[+syll] --> + long / --- [+nas] [+nas]

Rule (49), however, is just a special case of (48), with the
simultaneous nasal assimilation (rule (20)), and the nasal

cluster shortening (rule (24)).

2.6. A Word on Syllable Structure

Haya has two types of syllable structures: Closed and
open. The closed syllables, are of two kinds: (a) those
ending in glides (see also glide formation page 42) and (b)
those ending in nasals (see consonant clusters, section
2.4.1.2). Examples of closed sylables ending in glides are

in (14) repeated here as (50).

49

(50)
geija --> [geyja] ’become fat’
boigola --> [boygola] ’bark’
chaichiliza --> [éayéiliza] ’tease’

Examples of closed syllables ending in nasals are in (13)

repeated here as (51).

(51)

kambula ’tear down’

sinsimuko ’dirt’

mpolampola ’slowly’
Otherwise Haya has open syllables. The syllable final
vowels are either short or long. An abstract notion of the

syllable structure, containing one obligatory vowel is
exemplified in (52), in which G represents glides and N

represents nasals and C represents all consonants.
(52) (N)(C)(G)V(=)
The representation in (52) is partially illustrated in (53)

immediately followed by two examples for each

representation.

5O
(52) (N)(C)(G)V(2)

(a) NCGV:
(b) NOV:
(0) NV:
(d) NV
(e) CGV:
(f) CGV
(8) GV:
(h) GV
(i) V:

(J) V

(53)
a. NCGV: o -mu-mbyaazi --> omumbyazzi
IV-SM-planting field

’in the planting field’

b. NGV: a ~myaam-e -—> amxaame
SM-sleep-SUBJ

’he should sleep’

0. NV: amaani --> ama:ni
’strength’
d. NV tu-mal -a --> tumala

SM-finish-IND

'we finish’

e. CGV:
f. CGV
g. GV
h. GV
1. V:
j. V

51

kxaala --> kyazla

’take a walk’

luala --> lwala

’get sick’

ba-yggI-a --> bayoola
SM-pick-IND

’they pick things from the ground’

mu-ta -xasham -a --> mutayashama
SM-NEGz open mouth-IND

’you should not open your mouth’

ta -leeb-a --> taleeba
NEG-look-IND

’he/she does not look’

ichumu

’spear’

2.7. General Remarks and Summary

In

this chapter selected. aspects

of phonology

applicable to the understanding of the morphology of the

Haya word have been discussed and exemplified, namely: the

phonological inventory,

(including distinctive features and

52

redundancy rules), the distribution of Haya phonemes,
certain phonological and morphophonemic rules, and, briefly
Haya syllable structure. These various aspects of phonology
will be assumed in later chapters which deal with

descriptive and theoretical issues of Haya verb morphology.

53
CHAPTER 2 ENDNOTES

1 For a detailed sketch of Haya phonology, see
Byarushengo (1975).

3 There are few lexical items in Haya beginning with
the phonemes ./p/, /d/ and /w/. There are for instance 30
entries of #P - in the Haya dictionary, Bona-Baisi (1957), 7
entries of #d-, and 14 entries of #w-.

3 We have selected features that capture the necessary
distinctions in this language. Some features are acoustic
while others are articulatory. (Jakobson et al. 1951,
Jakobson and Hale 1956, Chomsky and Halle 1968). Note that
there is not DF specification for phonemes /m/and/n/on syl
because they could either be syllabic or non-syllabic (see
for example rule (22) page 34).

‘ Haya appears to be among Bantu Languages which have
undergone partial reduction. in tonal distinctiveness and
tonal morphologization.

5 In the examples given a verb is regarded as the
simple stem and an indicative marker. Where no such
examples exist a noun has been given.

' In the rest of this work tables will be placed at the
end of each chapter.

7 For a detailed analysis of consonant alternations,

see Lee Trithart in Byarushengo et al. (1977).

CHAPTER 3

THE MORPHEMES OF THE VERB: FORMS AND FUNCTIONS

3.1. Introduction

This chapter describes the morphemes of the Haya verb
in terms of their meaning and their linear order within the
verb. The cooccurrence restrictions which hold among these
morphemes are the topic of chapter 4.

The twenty three verbal morphemes to be described are
listed in table 6 at the end of this chapter followed by an
inventory of all morphemes with their allomorphs in table 7.
We represent the twenty three morphemes in. a morphemic
formula in (1) (henceforth PHD”. .As will be seen in this
chapter, the Haya Verbal morphemes may be divided into two

groups: (a) The Root, and (b) The Non-Stem Morphemes.

54

.«toaaago c. cummsom_u ma conga

ox“. golgs manna .ao.mo_o=o=a =-LLuo u.“ «.93. Laooc-oo mocogaco. .clzv u>.xs_.o. «:2 can .zm. casca- sooaasm as. cogs .o
.o._a a La .aoaaa >~¢ mu: was >_=o .m

.uaa .o usulxa. a mo~_a.~ Ea..om.oa=m a .o mocomaa o=_ ..

“MW .Amm..a : amo- Ea L=ooo >a. x "ex ..o... .Loa.== oz. .o ..-._ as. o. a: do moo:o..=ooo o.a...=. a mean. .~=_. ..oa. .n
.zma. Luxcac .ooansm o>~za.oa a cue. m._=mo. gm uca >_ .c =o_8~=.a.oo o=_ .N

.amco, 00., .0: .sn an .o oo_ogo as» mo.a-mmooo= >_ as «usage «4. .—

.musoz

2-3-3-2- 2::- :- 2-25-3-3: : s 5...; m A A - o - m $-13:
com

a :2: as 3..

:E-a-Eaa- 2% l a..- :c- «31:22 12.7521 . 5.. . I:£51.:rag-32:;€5-22-2.8
a2. mm¢a mace >—

..>==. m¢u> ¢»c= ux_ me «Jagged e.guzamcz ._.

56
3.2. The Root

We may begin the description of the verbal morphemes in
Haya with a description of the root, which may be considered
to be the minimal lexical unit of the verb. In this study,
Driever’s (1976: 23) definition of root in Swahili, another
Bantu language, may be adopted for Haya: "the part of the
verbal expression which carries the lexical meaning of the
verb and cannot be morphologically analyzed into smaller
meaningful units." IExamples of such. minimal meaningful

units are given as follows:

Iroot

’die’

b. -lum-
root

’bite’

root

’fury’

d. -zal-
root

’give birth’

57

The presence of hyphens on either side of' root forms
indicates the absence of any other morphological material.

A verb root may coocur with a large number of verbal
morphemes, whose description is the purpose of this chapter.
First, a root may cooccur with morphemes which modify its
lexical meaning or alter its argument structure: the verbal
extensions.2 One example, which will be discussed in more
detail later, is an extension which, among other functions

adds a beneficiary argument:3

(3) -zal-il-
ROOTrAPL-

’give birth for’

Second, a root may occur with morphemes which encode direct
(DO) and indirect objects (IO): the valence morphemes.‘ The
previous example may take a morpheme encoding the

beneficiary argument:

(4) a. mu-zal -il-
DO-give birth-for

’give birth for him’

b. mu-ba-jun -il-
° DO-IO-help for

’help him for them’

58

At this point we define another term: the stem. The term
stem will be used to refer to the set of all monomorphemic
roots plus the set of all roots with verbal extensions
and/or valence morphemes.5 It should be noted however that
no other source (”1 any other Bantu language includes the
valence morphemes as part of the stem.

Third, a root may cooccur with morphemes indicating
tense/aspect, mood, finiteness, negation, conditionality,
focus, status as a relative clause and agreement for person,
gender and number of the subject of the verb. One example,
whose details are again to be discussed later, may be given

as follows:

(5) a. a -ka -zal-
3Psg-TNS-give birth

’she gave birth’

where 3Psg is the third person singular (future
representation will be modified), and TNS refers to tense to

the specifics of which will be discussed later.

b. 0 -li -bon-
2Psg-TNS-see

’you(sg) will see.’

Three examples of apparently maximal verbs generated by the

morphemic formula of the Haya verb (MHV) are:

(6) a. ka -ba —ta -li-ku -mu-tel -el -a -ha -ge -kwo.
COND-SM -NEG2-NF-INF-OM-ROOT-APL-IND-LOC-MNR-INT
if -they-not- -ing-him-hit well-indeed

’If they are not indeed hitting him well here’

b. ti -tu-ka -ba -lim -is -e -ki-
NEG1-SM-RMP-OM -ROOT-CAUS-SUBJ Q
Not-we- -them-farm- - -what-

’What haven’t we caused them to farm?’

0. ka -mu -ta -ba -mu-leb -angan-il -a -ha -ge -kwo
2 -5 -3 -9a -9b-10 -16 -13 -18 -20 -21 -23
COND-SM-NEG -DO -IO-ROOT-RCP -APL-IND LOC-MNR-INT.
’If you are not indeed taking care of him well for

one another here.’

From these examples it becomes apparent that it is
impossible to get all the 23 morpheme in one verbal
cOnstruction. There are certain co-occurrence restrictions
on a number of them the details of which are discussed in
chapter 4. We will later present criteria to show that
these two groups of morphemes, the morphemes constituting
the stem as the first group and all the other morphemes as
the second group, may be reasonably regarded as constituting

separate sets.

60

The remainder of the chapter is organized as follows.
The second group of morphemes (the non-stem morphemes) are
presented first. They are described in relation to the
stem, which will be referred to simply as "the stem" without
reference to its internal structure. Each morpheme will be
presented as part of one or more verb forms. Next, the stem
morphemes are presented” The reason for this order of
presentation is that the stem morphemes are mostly morphemes
which affect the argument structure of roots. For this
reason, they are best illustrated with sentences rather than
isolated verb forms, as in the case of the non-stem
morphemes. Prior description of the non-stem morphemes will

make these examples easier to understand.

3.3. The Non-Stem Morphemes

3.3.1. The Indicative Morpheme

We may begin the description of the verbal morphemes of
Haya with the indicative morpheme :a, sometimes called the
imperative. This morpheme occurs directly following the

verb stem, schematized as follows:

(7) STEM - IND

The indicative morpheme in combination with the verb stem

forms an imperative:

61

(8) a. abazaile ba-laa-gul-a ebitabo
parents SM -IF -buy—IND books

The parents will buy the books.

b. iba ya-leet —a ebilaatwa
husband-SM-bring-IND shoes

’her husband brought shoes’

c. nyowe nyw -a amaizi gangi bulikilo
I drink-IND water plenty every day

’I drink plenty of water every day.’

The indicative also occurs in other forms of the verb as

well, as will be discussed as the dissertation develops.

3.3.2. The Infinitive Morpheme

The infinitive morpheme kg; realized as kg and kg (see
chapter 2 page 44), is attached to the stem, with final
vowel :a, without tense markers (It may be followed by the

(NEGz) if there is one).

(9) INF - (NEGz) - STEM -
(10) a. ku -shom-a
INF-read-IND

’to study/studying"

62
b. ku -ib -a --> kwiba

INF-steal-IND

’to steal/stealing’

c. ku -ta -gab -a
INF-NEG: give-IND

’not to give/not giving’

The infinitive form of the verb has both nominal and verbal
syntactic functions, the description of which is beyond the
scope of this study. When the infinitive is nominal, it

controls the agreement of class 15.

(11) a. O -ku -shom-a ku-tin -is -a abantu
IV-INF-read—IND 15-be afraid-CAUS-IND people.

’Reading frightens people’

b. O -ku -lim -a ku-jun -a
IV-INF-farm-IND 15-help-IND

’farming helps’

The SM(INF) agrees with the nominal infinitive which in the
examples above is preceded by an initial vowel (IV). We
will at a later stage discuss and exemplify the noun classes

in Haya in section 3.3.3.1.

63
3.3.3. The Subject Marker

The subject marker is a verbal morpheme which shows
agreement with the subject of the verb. It may be
exemplified with the present universal/habitual tense (see
section (3.3.5.1.1.)), of the verb, in which it is prefixed

to the verb stem fellowed by the indicative morpheme:

(12) a. ba-lim -a
SM-farm-IND

’they farm (usually)’

b. a -guz -a
SM-sell-IND

’he sells (usually)’

The precise linear order of the subject marker with respect
to other pre-stem verbal morphemes will become clear below

when these morphemes are discussed.

3.3.3.1. The Noun-Class System

More specifically, the subject marker agrees in noun-
class with the subject of the verb. In order to understand
the degree or variation found in the subject marker, we will
briefly discuss the noun-class system in Haya.

Nouns in Haya, as in many other Bantu languages, belong
to concordial classes, which are referred to as noun-

classes. Nouns which belong to the same noun-class control

64

the same set of agreement affixes on verbs and modifiers
(cf. Bennett 1982). In addition, the nouns belonging to a
single noun-class tend to be marked with morphologically
identical noun-class prefix, although this generalization is
not absolute. There are eighteen noun-classes in Haya. In
this dissertation, we will use the numbering system for the
noun-class system of Bantu languages referred to as Welmers
system (1973). Under this numbering system, the first ten
noun classes form singular/plural pairs, as shown in table 8
(page 148). Odd-numbered classes have their plurals in the
following even-numbered. classes. 131 addition, Class 11
nouns have their plurals in Class 10.‘ Class 12 (ka-) is
predominantly the class of diminutives (Ashton et al. 1954
in Luganda). The prefix ka- which usually follows prefixed
to the initial vowel a- (aka-), may replace any noun in
order to indicate the diminutive form sometimes slightly

changing the original meaning as exemplified below:

(13) a. O -mu-ntu a -ka-ntu
IV-lZ-STEM— IV-lZ-STEM
’a person’ ' ’a small thing’
b. O -mu-ti a -ka-ti
IV-3 -STEM IV-12-STEM

’a tree’ ’a small stick’

65

c. e -li-sho a -ka-isho
IV-5 -STEM IV-lZ-STEM
’an eye’ ’a small eye’

whereas example (13.0) shows the diminutive form of an eye,
in (13.a) the meaning attached to the diminutive form is not
of a person but of a ’small’ thing.

Class 13 (tu-) may function as a plural form of class

12.
(14) a. a —ka-ntu o -tu-ntu
IV-lZ-STEM IV-13-STEM
’a small thing’ ’small things’
b. a -ka-isho o -tu-isho
IV-lZ-STEM IV-13-STEM
’a small eye’ ’small eyes’

Class 14 (bu-) is a class of the augmentative of the
plural diminutive class 13 and may as well function as the

plural augmentative form of class 12 (ka-).

(15) a. a -ka-ntu o -bu-ntu
IV-12-STEM IV-14-STEM

’a small thing’ ’little bigger things’

66

b. o -tu-ntu o —bu-ntu
IV-13-STEM IV-14-STEM
’small many things’ ’little bigger many
things’

Class 15 consists entirely of infinitive forms and some

body parts:

(16) a. o -ku-andika
IV-15—STEM

’writing/to write’

b. o -ku-gulu
IV-15-STEM

’a leg’

when class 15 is used as a body part, it takes the plural

form of class 6 (ma-).

(17) o -ku-gulu a -ma-gulu
IV-lS-STEM IV-6 -STEM
’a leg’ ’legs’

Noun classes 16-18 are traditionally referred to as locative
classes (Rascher ibid.: 65-69).7 These classes are
indicated by the morpheme marker ha-(16), ku-(17) and mu-

(18) with the underlying idea of definite place, indefinite

67

place and. withinness (inside) respectivelyu These noun
classes will become more clearer in the course of the
discussion.

In classes 1 and 2, the classes of singular and plural
human beings, respectively, the verbal subject markers vary
for lst, 2nd and 3rd person as well as for class, as seen in

table 8 and exemplified in (18) and (19).

(18) a. n -chumb-a buli kilo
ls-cook -IND every day

’I cook every day’

b. 0 -lim -a ekilo
Zs-farm-IND night

’you(sg) farm at night’

0. a -yombek-e amaju
3s-build -SUBJ houses

’he should build houses’

(19) a. tu-guz -a emotoka
lp-sell-IND cars

’we sell cars’

b. mu-bug -a emotoka eziango '
2p-drive-IND cars big

’you drive big cars’

68

c. ba-s -e omugusha
3p-grind—SUBJ millet

o’they should grind the millet’

Since nearly every Haya verb example in this dissertation
contains a subject marker, they will not be exemplified

further here.

3.3.4. The Subjunctive Morpheme

The subjunctive form of the verb is formed with the
subject marker prefixed directly to the verb stem, which is
followed by the vowel :e in place of the indicative vowel

:a, schematized in (20) and exemplified in (21):°

(20) SM - STEM - SUBJ

(21) a. ni -bend-a ba-lim -g ekibanja
FOC-STEM-IND 2 -farm-SUBJ garden

’they want to farm the garden'

b. a—gend-e Bukoba?
I-STEM-SUBJ

’should he go to Bukoba?’

69

c. ba-shom-e ebalua ba-many-e agalimu
2 -STEM-SUBJ letter 2 -STEM-SUBJ content
’they should read the letter in order to know the

content’

d. a-tel -g . ente ezi?
I-STEM-SUBJ cows these

’should he hit these cows?’

The subjunctive morpheme is used in different ways including
but not limited to: following verb like want (21.a), when
permission is sought (21.b), when the intention is to
express purpose (21.0). Other functions not exemplified
above include: serving as a more polite imperative, used in
indirect commands; with the question morpheme :ki; it
questions the possibility. With the locative morpheme it
questions the location. Both of these possibilities are

exemplified below:

(22) a. ba-lim -e -ki
2 -farm-SUBJ—Q

’what should they farm?’

b. a-shuntam-e -ha
I-sit -SUBJ-LOC

’where should he sit’

70

The question and locative morphemes will be discussed in

detail in section 3.3.10.

3.3.5. The Tenses

As far as the present author can determine, there are
twelve tenses in Haya (see also section 3.3.8), which are
listed in table 9 (page 146).9 It is not the purpose of this
work to provide an exhaustive description of the semantics
and usage of each tense. Such a description is probably not
possible for the tense system of any Bantu language at the
current state of our knowledge. The following descriptions
of tenses in Haya are intended to be working definitions
only, to facilitate an understanding of tense morphology.
They' will be based on ‘broad conceptions of ’a present
moment,’ ’time before the present time’ and 'time after the
present moment’ developed in Moshi (1988) (for Chaga, a
Bantu language related to Haya.) Within these three
categories, there are further distinctions made, which will
be brought up in the course of the discussion of each tense.
It will also be apparent that several of the tenses involve
elements of meaning which are more properly described as
aspectual, but while we will note such elements, we have not
attempted to systematize them, again because they fall
beyond the scope of this work, which is intended to focus on

the organization of verbal morphology.

71
3.3.5.1. The Present Tense

Moshi’s conception of present time probably is used in
referring to situations which occupy a much longer period of
time than the present moment, that is, those which include
the present moment (see Comrie, 1985: 37).10 If such a
conception is true, then Moshi’s concept of the present
tense fits into the description of the Haya present tense.
Present’ tense in Haya includes the actions, states and
processes which hold at the present moment, but which began
before the present moment and which may well continue for a
short period of time after the present moment as exemplified

below:

(23) a. Omwalimu mukulu n -a -nyegez -a abagenyi
teacher old Is-IR-welcome-IND visitors

’the principal is inviting visitors’

b. Abajaji ni -ba-lamul -a omushango guango
judges FOC 2 judging-IND case big

’the judges are judging the big case.’

In each of these examples, it is true that the actions of
welcoming, and judging hold’ at the present moment.
However, these actions may have started two or more hours
before and may continue for some time. Therefore it is
accurate to say that the basic meaning of the present tense

in Haya as in many other languages, locates a situation at

72

the present moment, but the situation may continue beyond
the present moment, or may also hold in the past-~depending
on the event. We may now discuss the formation of the
present tenses in Haya beginning with the present universal

habitual (PUH).

3.3.5.1.1. Present Universal Habitual

The present universal habitual tense is formed by
attaching the subject marker directly to the verb stem,
which is followed by the indicative morpheme, schematized as

follows:
SM - STEM - IND

This tense refers to habitual, ongoing, repetitive actions,

as seen in the following examples:

(24) a. ba-lim -a
3 -farm-IND

’they farm’

b. mu-it -a
2p-kill-IND

’You(pl) kill’

73
3.3.5.1.2. "Still" Tense

The "still" tense is formed by attaching the morpheme
:kza; directly preceding the verb stem. The morpheme :kya;
is itself directly preceded by the subject marker, and the
indicative morpheme is suffixed to the verb stem,

schematized as follows:

SM - STL - STEM - IND

The "still" tense is a present tense which also carries the
more aspectual implication that the action or state is still

in progress. It is illustrated in the following examples:

(25) a. ba-kya-lim —a
2 -STL-cultivate-IND

’they are still cultivating’

b. zi-kya-nyw -a
10-STL-drink-IND

’they(e.g. cow) are still drinking’

In example (25.b) the class 108M refers to ente ’cows,’

while in (25.0) the class 148M refers to obunaku ’poverty’

0. bu-kya-ij -a
14-STL-come-IND

’poverty is still coming’

74

3.3.5.2. The Past Tenses

The past tenses are used to describe a situation
holding sometime before the present moment. Moshi (ibid:2)
mentions that in Kivunjo the past represents a three way
distinction: P1(-a-) within few minutes, P2(-am-) within few
hours and P3(-le-) furthest point in the past. In Haya
there are separate forms of the verb referring to six
periods of the past time, the immediate past (to-day), the
near past (very early to-day) the recent past (yesterday),
the remote past (day before yesterday), and the far remote

past (long ago). These will be described in more detail in

the following sections. In addition, the near past has a
habitual form. These tenses are listed in table 9 on
page 147.

3.3.5.2.1. Immediate Past.

The immediate past tense is formed by attaching the
immediate past tense morpheme :a; (IP) directly preceding
the verb stem. This tense morpheme is in turn directly.
preceded by the subject marker. The indicative morpheme is
suffixed to the verb stem. The formation of the immediate
past is therefore parallel to the formation of the "still"
tense, and all tense morphemes to be described in the
chapter will be seen to occupy the same prestem position
except for the near past (NPS), the far remote past (FRP),

and the near past habitual (NPH) all of which, in addition,

75

have morphemes suffixed to the verb-stem. The recent past
(yesterday) only has the morpheme following the stem. This

tense may be schematized as in (24):
(24) SM - IP - STEM - IND

This tense is used to refer to an event or state which
occurred or was in force at some very immediately previous
point on the day of speaking and is now finished. We may
now’ propose that this tense refers to events that have
occurred within the past few hours (three to four hours).

It may not refer to a time other than to-day.

(27) a. gu-a -gw —a
3 -IP¥fall-IND

’it (tree) fell (to-day)’

b. tu-a -guz —a
Ip-IP-sell-IND

’we sold (to-day)’

3.3.5.2.2. Near Past

.The near past is marked by prefixing the immediate
tense morpheme :_a__-_ (NPS) to the verb stem and by suffixing
the stem with the morpheme :ilg (which replaces the

indicative morpheme).11 This tense may be schematized as:

76

ooooo

The near past tense is used to refer to an event or state
which occurred or was in force at some point early on the
day of speaking. The event might have happened five to six
hours previously, such that this tense is usually used to
demarcate the day into morning, afternoon, or evening.

Examples of the recent past tense are given below:

(29) a. tu-a-gul-ilg (--> twaguzile) ebitooke bwanka
1p- -buy- bananas morning

’we bought bananas this morning’

b. kya—g-endek-ilg nyemishana kyashubamu
7- -break- afternoon repeat
kuendeka mbwenu aha
to break to-day now
’It (the chair) broke this afternoon and now it has

repeated breaking again’

3.3.5.2.3. Recent Past

The recent past is marked only by the suffix _;_i_l_g,
attached directly‘to the verb stem, in the place of the
indicative morpheme._ The verb stem is directly preceded by
the subject marker, and there is no present tense Omarker
parallel to the tense marker :3; of the near past. The

recent past may be schematized as follows:

77

(30) SM - STEM - RP

This tense is used to refer to an event or state which
occurred or was in force at some point the day before the
day of speaking and is now finished. Examples of the recent

past tenses are given in (31).

(31) a. a-lim -ile
I-farm-RP

’he/she farmed (yesterday)’

b. li-endek-ile
5 -break-RP

’it (the tooth) broke’

The suggested referent in (31.b) (the tooth) in the gloss
belong to noun class 5 (see noun classes 144). The recent
pastin Haya is cognate to a Kamba recent or previous day's
past tense (Farnsworth 1952:26) where the tense morpheme is

pa with -ig suffixed to the verb-stem. Probably the gig

now-0‘...-

3.3.5.2.4. Remote Past
The remote past has a morphological structure which is
parallel to the near past rather than the recent past.

There is a tense marker :ka; (RMT) which is prefixed

78
directly to the verb stem. and preceded by the subject

marker. The indicative morpheme is suffixed to the stem.

The remote past may be schematized as follows:

(32) SM - RMT - STEM - IND

This tense is used to refer to an event or state which
occurred or was in force at some point before on the day
before, usually one year or more, and is now finished. It
is not easy to delineate the exact time in the past to which
this tense refers, nor is it possible to state how long ago
the action took place or the state was in force. The
reference made above is only a guiding definition of this
tense. The remote past tense in Haya seems to be the same
as that in Runyankore, another Bantu language closely
related to Haya (Morris and Kirwan 1957). As in Haya, to
form this tense in Runyankore -ka- is inserted between the
subject prefix and the stem. Examples of the remote past

tense in Haya are:

(33) a. n -ka -tel-a
Is-RMT-hit-IND

’I hit (last week)’

b. tu-ka -bandam -a
Ip-RMT-crounch-IND

’we crounched (e.g. last year)’

79

3.3..5.2.5. Far Remote Past

The far remote past is also a combination of a prefixal
tense morpheme and a suffixal one. As in the near past
tense, the suffixal morpheme is :ile. attached directly to
the verb stem. The prefix, again preceded by the subject
marker and attached directly to the verb stem, is the
morpheme :_I_a;. The far remote past may be schematized as

follows:

(34) SM - Ia - STEM - lie

This tense is used to refer to an event or state which
occurred or was in force at some point a long time
previously, usually more than two years. Again this time is
not an absolute but simply a relative time. Nurse (1979:12)
calls this tense the"have already’ (long ago), while both
Rasher (ibid) and Byarunshengo et al. (ibid) fail to make a

note of this tense. This tense is exemplified as follows:

(35) a. e-la-SW-ilfi
4 fell

’they (trees) have already fell (a long time ago)’

where the referent trees belong to noun class 4.

80
‘b. ba-Ia-yombek-ilg

2 build

’they have already build (e.g. three years ago)’

3.3.5.2.6. Immediate Past Habitual

The formation of the immediate past habitual is nearly
identical to that of the immediate past. The tense morpheme
:a; (IPH) directly precedes the verb stem and is in turn
directly preceded by the subject. marker. The indicative
morpheme is suffixed to the verb stem. Unlike the present
universal habitual, the immediate past habitual contains a
morpheme specifically indicating habitual action: the
habitual morpheme _-__a_g_-_, which is suffixed directly to the
verb stem preceding the indicative morpheme.12 The

immediate past habitual may be schematized as follows:
(36) SM - a - STEM - ag - IND

The immediate past habitual indicates actions or situations
that hold at all times, or repetitive actions or events.13
The immediate past habitual differs from the immediate past
in the’time slot. Whereas the immediate past refers to
events that have taken place on the day of speaking, the.
immediate past habitual includes both the actions of the
immediate past and near past which occur repeatedly. In

other contextual situations, the immediate past habitual may

81

also includes the concept of the possibility of not

repeating the action again as shown in examples below:

(37) tu-a-lim-ag—a (bilebi tu-ka-Iekel—a)
Ip- -farm- IND these days SM- stop IND
’we were habitually farming (we have in these days

stopped)’
Other examples of the near past habitual are:

(38) a. ba-a-beiy-ag-a
2- -lie- -IND

’they were usually lying’

b. n —a—yeseS-§£‘a
Is- -teach- -IND

’I was usually teaching’

3.3.5.3. The Future Tenses

The future tenses are used to describe an event or
state which will occur or hold sometime subsequent to the
present moment. In other words, the most typical uses of
the future tense involve events or states that are planned
to happen or to be implemented sometime subsequent to the
present time. However, there are certain future categories

which do not involve the prediction of an action or state if

82

there is no belief or intention of implementing these

actions as exemplified below:

(39) n-laa-tem -a ente (ndaatema)
Is-IF-slaughter-IND cow

’1 will slaughter the cow’

The prediction cu? ’slaughtering the cow’ cannot be
implemented if the speaker believes that the slaughtering
will actually not take place. Again as in earlier
definitions of tenses, the definition of the future tense is

given as a guideline, but the actual meaning is context as

well as intentional sensitive. As for the phonological
changes shown in example (39), see chapter 2 section
(2.5.1). Haya has separate forms of the verb referring to

three periods of future time parallel to those we have seen
for the past tenses: the immediate future, the near future
and the remote future, which. will. be described. in. more
detail in the following sections. In addition, as in the
past tense series, the near future has a habitual form.

These tenses are listed in table 9 on page (147).

3.3.5.3.1. Immediate Future

The three future tenses (apart from the habitual) are
formed exclusively with prefixal morphemes parallel to those
we saw in the formation of the past and perfect tenses.

There is no suffixal morpheme involved in the formation of

83

the future tenses. The immediate future (IF) is formed by
attaching the morpheme :laa; directly to the verb stem, in
the same position relative to the stem as the past and
perfect prefixal morphemes, preceded, as before, by the
subject marker. The indicative marker is suffixed to the

stem. The immediate future may be schematized as follows:
(40) SM - Ian - STEM - IND

This tense is used to refer to an event or state which will
take place or be in force sometime today or tomorrow. Again
the immediate future tense is relative, ”where the reference
point for location of a situation is some point in time
given by the context, not necessarily the present moment"
(Comrie ibid:56). Comrie’s explanation is made clear by the

following Haya examples:

(41) Yesu a-laa-lind -a abakristu bona
Jesus I-IF -protect-IND Christians all
aba-laa-ondel -a obugolozi bwe
who-IF -follow-IND leadership his
’Jesus will protect all those who will follow his

leadership’

The interpretation which can be given to example (41) is
that the present moment is relative to the time when the

statement was uttered. For someone who heard it ten years

84

ago this statement is not taking place on the day it was
spoken (for the first time) or on the day after the day of
speaking but it has been taking place for ten years prior to
ten years ago and will take place after the present moment
for an indefinite time (as long as he/she is alive).

Compare example (41) to (42).

(42) ba-laa-gend-a Tanzania omwezi ogulaija
2 -IF -go -IND month next

’they will go to Tanzania next month’

In this example, the month in which this utterance is made
is taken as a ’reference point to-day.’ If it is uttered in
July, then their going to Tanzania is in August, regardless
of when in August. Other examples which exhibit the use of

the immediate future tenses in Haya are:

(43) a. ga-laa-yatik-a
10-IF -break-IND

’they (eggs) will break’ (to-day or to-morrow).

Here again the suggested referents ’eggs’ belong to noun

class 10.

b. tu-laa-gul-a ebitabo
Ip-IF -buy-IND books

’we will buy (to-morrow) books’

3.3.5.3.2. Near Future
The near future (NF) is formed by prefixing the
morpheme 3,1,1": directly to the verb stem, preceded by the

subject marker. The indicative marker is suffixed to the

stem. The near future may be schematized as follows:

(44) SM - I; - STEM - IND

This tense is used to refer to an event or state which will
take place or be in force the day after tomorrow. Again the
near future tense is relative depending on context as

illustrated in the following Haya examples.

(45) a. Omwana ogu a-li-bug -a endege
child this-I-NF-drive-IND planes

’this child will be a pilot (lit:driver of planes)’

The time reference to this sentence, although it is
basically sometime in the future, is not clear, because it
depends on the age of the child and the laws of countries
for one to become a pilot. If the child is four years old
and the minimum permit age for a pilot is 25 years, then the
so called ’near future’ tense is 21 years from the time of
the utterance of this statement. Compare example (45.a) to

(45.b):

86

(45) b. a-li-zal -il-a omuhosbitali ya Sparrow
I-NF-give birth-at-IND hospital of Sparrow

’She will give birth at Sparrow hospital’

If this sentence is uttered to-day and the person talked
about is due in two weeks, then the ’near future’ period
from to—day as a reference point is two weeks. Thus, again
’the near future tense’ is only realized contextually.

Other examples of the near future tense in Haya are as

follows:

(46) a. enjula e-li-lekel-a kugwa kemulitema emiti
rain 4-NF-stop -IND to rain when you cut trees
yona
all

’the rain will stop raining when you will cut down

all the trees’

b. Omuleke, a-li-shwel-a olwo a-li-yend-a
leave him alone, I-NF-marry-IND when I-NF-like-IND
’leave him alone he will get married when he will

like’H

3.3.5.3.3. Remote Future
The remote future (RMF) is formed by prefixing -lika-

directly to the verb stem, preceded by the subject marker.

87

The indicative marker is suffixed to the stem. The remote

future may be schematized as follows:
(47) SM - RMF - STEM - IND

The remote future is used to refer to an event or state
which will take place or be in force at a point in time
farther in the future than that referred to by the near
future. We will not discuss the relativeness of this tense
as we did in other tenses, except to point out that as the
previous tenses, this tense too depends on the context in
which it is uttered. Like other future tenses however, it
too refers to future events or states without an absolute
time reference.

Examples of the remote future are as follows:

(48) a. O -lika-jum -a
Is-RF -abuse-IND

’You will (in the remote future) abuse’

b. zi-lika-ij -a
4 -RF -come-IND
’they (e.g. Monkeys) will (in the remote future)

come’

The remote future marker is the only bisyllabic prestem

tense marker in the past, present and future series of

88

tenses in the Haya verbal system. This observation raises
the question of whether the remote future marker constitutes
a unitary morpheme, or whether it is composed of smaller
morphemes. It is suggestive that the first syllable of this
marker, :1;;, matches the near future marker :11;,
segmentally, and the second syllable, :kagJ matches the
remote past marker ;k_a_;. It is possible that the remote
future marker is composed of the near future marker followed
by the remote past. marker, with the semantic effect of
indicating a future event or state which also holds at the

farthest point in the future.

3.3.5.3.4. Immediate Future Habitual

The formation of the near future habitual is patterned
on that of the immediate future in the same way that the
formation of the near past habitual is patterned on that of
the immediate past. In the immediate future habitual, the
tense morpheme zlaa; (IFH) directly precedes the verb stem
and is in turn directly preceded by the subject marker. The
indicative morpheme is suffixed to the verb stem. As in the
immediate past habitual, the habitual morpheme gag; is
suffixed directly to the verb stem preceding the indicative
morpheme. The immediate future habitual may be schematized

as follows:

(49) SM - Laa - STEM - ag - IND

89

This tense may be generally described as being used for all
actions or states which are to take place repeatedly the
next. day and also for' any future action for which no
definite time is specified (see immediate future 3.3.5.3.1).
It differs from the immediate future only in frequency.
Whereas actions may take place in the immediate future only
once» and. for all, in the immediate future habitual the
actions may happen at all future times or be repeated
actions. This tense also differs from the present universal
habitual in the sense that in the Present habitual the
actions are currently taking place repeatedly whereas in the
immediate future habitual' the actions are to take place
habitually in the future time reference relatively speaking.

This tense is illustrated in the following examples:

‘(50) a. tu-Iaa-gul-ag-a
Ip- -buy— -IND
’We will (habitually) be buying (e.g. in the

evenings)’

b. ba-laa-shom -ag-a
2- -study- -IND
’they will (habitually) be studying (e.g. every

night)’

90
3.3.5.4. The "Imaginary" -ku; tense

p I..- an

The "imaginary" :ku; (UNR) tense is a conditional tense
formed by inserting the morpheme :ku; directly preceding the
verb stem in either the immediate past or the remote past,
one of the two subject marker. (Note that it may not appear
in. the near past :Qggp) . The term "imaginary" 353;, is

borrowed from Salone (1977). The "imaginary” :ku; tense may

be schematized as follows:

(51) SM - g - kg - STEM - IND
._-...J.'._l.e

The ”imaginary" 3,1531; tense may be exemplified as follows
(note that the tense occurs in both clauses of the

conditional):

(52) a. ka -n -a -ku -gul-a e—ndege na-a -ku -gy-a
COND-Is-IP-UNR—buy-IND IV-plane Is-IP-UNR-go-IND
Bulaya
Europe

’If I buy a plane, I would go to Europe"

b. ka -n -a -ku -mu-bon-ile n -a-ku -mu-siim-ile
COND-Is- ~UNR-3s-see -Is- —UNR-Ss-thank-
(boine)

’If I would have seen him, I would have thanked him’

91
It also seems likely that the :ku: of this form is to be

identified with the infinitive Ell: rather than one
"imaginary” -ku-, since infinitive km; is not preceded by a

subject marker, parallel to the -ku- of this form, and

unlike "imaginary" :kuz. However, it is also possible that
Haya has three distinct kg; morphemes. (We will leave the
matter open here. The ”immaginary" :ku; tense is the last
tense to. be discussed in this work-15 The discussion of

tenses discussed thus far are summarized in table 9 (147).

3.3.6. The Subject relative morpheme

The subject relative morpheme is a morpheme which
transforms a non-relative verb into a relative verb. Only
verbs which have a prestem tense marker are relativized.
The subject relative morpheme takes the form of a vowel
prefixed to the subject marker. This vowel will be referred
to as the initial vowel, following the usage of Ashton et
al. for Luganda (1954), Taylor for Nkore-Kiga (1985)
languages of Uganda, Morris and Kirwan for Runyankore 1957,
and Rascher, 1958 for Haya.1° The subject relative verb may

be schematized as follows:

(53) IV - SM - TNS - STEM - IND
rile

IND and _-__i_l._e_ alternate with each other and either is an

option in a relative verb as represented above by curly

92

brackets. As its name implies, a subject relative verb is

used only when a subject is relativized:

(54) a. omuti o -gu-gw -a
tree IV-3 -fall-IND

’the tree which fall.’

b. eliino e -li-laa-shash-a
tooth IV-5 -IF -hurt -IND

’the tooth which will hurt’

(See table 10 subject relative morphemes page 148).

Note that since the head noun of a subject relative is also
the subject of the relative verb, the subject marker of the
relative verb will agree with it in noun-class.” The
identity of the initial vowel is determined by the noun
class of the subject marker of the verb: each noun class has
its own initial vowel. A list of initial vowels for all of
the noun-classes is given in table 10. Note that classes 17
and 18 do not relativize because the relativization made in
class 16 covers the other two classes (the locative classes
as traditionally known-see fn 7). Although in noun classes
1, 4 and 9, we have included both vowels (the IV and SM),
vowel coalescence takes place as was discussed in chapter 2
section 2.5.2. This initial vowel may be compared to the
initial vowel of nouns not discussed in this dissertation

but may briefly be exemplified here:

93

(55) a. O —mu-ti o —gu-laa-kul -a tu- laa-gu-tem -a
IV-3 -STEM IV-3 -IF -STEM-IND 2p(pl) IF -3 -STEM-IND
Noun

’The tree (it) which will grow we will cut it’

b. e -mi—ti e -e-laa-kul -a tu-laa-gi-tem -a
IV—4 -STEM IV-4-IF -grow-IND Ip-IF -4 -STEM-IND
Noun

’the trees (they) which will grow we will cut them’

(56) a. e -ki-tabo e -ki-li-gul—w -a ba-li-ki-shom-a
IV-7 -STEM IV—7 ~NF—buy-PASS-IND 2 -NF-7 cread-IND
Noun
’the book (it) which will be bought they will read

it’

9b. e -bi-tabo e -bi-li-gul-w -a ba-li-bi-shom-a
IV-8 -STEM IV-8 -NF-buy-PASS-IND 2 -NF-8 -read-IND
Noun.
’the books (they) which will be bought they will

read them’

As exemplified in (55) and (56), the initial vowel in both
nouns and verbs omuti/emiti , ’ tree/ trees , ’ ekitabo/ebitabo
’book/books,’ are the same, prefixed to the noun class

prefix. However, both the nouns and verbs may not have an

94

IV to which no relativization is realized as in (57) and
(58). The passive form (PASS) used in example (54.a) will

be discussed in section 3.5.3.

(57) a. mu-ti gu-laa-kul-a tu-laa-gu-tem-a

’a tree (it) will grow we will cut it’

b. miti e-laa-kul-a tu-laa-gi-tem-a

’they are trees they will grow we will cut them’

(58) a. kitabo ki-li-gul-w-a ba-li-ki-shom-a

’it is a book it will be bought they will read it’

b. bitabo bi-li—gul-w-a ba-li-bi-shom-a
’they are books they will be bought they will read

them.’

The parallelism between the use of IVs with nouns and that
with verbs suggests that the addition of the IV to a verb
form has the effect of nominalizing it. In support of this
suggestion, note that the subject relative verb may function
by itself as a nominal, as in the following headless

relative construction.

95

(59) a. e -ki-laa-zok -a ki-laa-gul-w -a
IV-7 -IF -be found-IND 7 -IF -buy-PASS-IND
’(it) which will be found will be bought ’e.g. a

book”

3.3.7. The Conditional Morpheme

The conditional morpheme in Haya is the marker kg;
which is prefixed directly to the subject marker. When the
subject marker has an initial vowel, then coalescence takes
place as discussed in chapter 2 section (2.5.2). A verb

which contains the conditional morpheme may take a number of

meanings, which may be translated using "when," "as," "if"
and "let":
(60) a. ekitoke ka -ki-kul -a ba-ki-tem-a

banana COND-7 -get ripe-IND 2 -7 -cut-IND

’when.the bananas becomes ripe they cut it’

b. ka -ba-laa-b -a ba-laa-ij -a
COND-2 -IF -be-IND 2 -IF -COME-IND
tu -ba-chumb-il -e
Ip(pl)-2 -COOK -APL-SUBJ

’if they are coming let us cook for them’

Note that the APL morpheme -il- which occurs in (60.b) will

be discussed fully in section 3.5.2.

96

0. ka -zi-gend-e ku -nyw -a amaizi
COND-lO-go -SUBJ INF-drink-IND water

’let them (e.g. cows) go to drink some water’

The conditional morpheme is also used with a special
form of the verb, in a construction which Salone (1977:156)
calls the "imaginary conditional." Writing on conditions in
Haya, Salone argues that there is an imaginary conditional
different from other types of conditionals in that the
imaginary conditional takes the suffix :ilg as well as the
marker :kuz. The suffix :11; as discussed earlier is also
used with the near past, recent past, and far remote past.
The marker :ku; is placed in the tense marker slot directly
preceding the verb stem. Salone calls this marker :53; a
marker of "imaginariness" perhaps derived historically from
the infinitival marker (see also 3.3.5.4). Its usage is
different from its usage as an infinitival marker in several
ways. (1) an infinitiver morpheme never occurs with a
subject prefix (2) an infinitive morpheme never carries a
tense morpheme, both of which do occur with this imaginary

indicator:

(61) a. n -a -ku -chumb-ile kyonka t -a -li-mu majuta
Is-IP-UNR-cook -RP but NEG-IP-be-LOC oil

’I would cook but there is no oil.’

97
b. ente y-a -ku -za -ile kyonka
cow 9-IP-UNR-give birth-RP but
y-a -bul -w -a obujuni

9-IP-loose-PASS-IND help

’the cow would have given birth but it lacked help’

0. abana ka -ba-ku —many-ile ti -ba-ku -las -ile
children COND-Z -UNR—know-RP NEG-2 -UNR-throw-RP
mabale
stones
’if the children would have known, they wouldn’t
have throw stones.’

3.3.8. The Focus Morpheme

The focus morpheme in Haya is the marker 31; which is
prefixed directly to the subject marker without the initial
vowel. We are using the term "focus" to refer to this
morpheme because it appears to be cognate to morphemes in
other Lacustrine and Thaagicu Bantu languages which have
been referred to as a "focus" morphemes, (cf. Clements 1984,
Bergvall 1987, Perez (Harford) 1985 for .Kikuyu, Dalgish
1978, Moshi 1986 for Kivunjo, Harford 1987 for Kitharaka;
cf. also Bennett 1982 for further discussion and comparison
to other Bantu groups). However, the use of this morpheme
in Haya is much more restricted than the use of the focus
morpheme in these other languages. Moreover, its range of

meaning and usage is not ‘well understood in. any Bantu

98

language, although time works just cited represent initial
attempts to arrive at such an understanding.

In particular, the focus morpheme in Haya is restricted
to three verb forms: the present universal habitual, the

"still" tense and the Subjunctive:

(62) a. Kato na Kwezi ni -ba-lip-a amabanja omuduuka
Kato and Kwezi FOO-2 -pay-IND debts store

’Kato and Kwezi are paying debts in the store.’

b. Omwisiki ogu n -a-iluk-a kukila abaisiki bona
girl this FOC-I-run -IND more than girls all

’this girl is running faster than all the girls’

0. Omwishulukale ni -ba-kya-mu-pim -a emyenda
the soldier, FOC-2- -OM-measure-IND clothes

’the soldier, they are still measuring him clothes’

In example (62.b) there are phonological changes which take
place when the FCC n1; is followed by a vowel subject marker

as discussed in chapter 2 section 2.5.2.2.

a. abana ni -ba-iluk-e enjula y-a-ij -a
children FOC-2 -run -SUBJ rain 9-IP-come-IND
ku -gw -a
INF-fall-IND

’let the children run it is almost raining’

99

Example (62.0) show the use of the FOC ni; with the :kya;
’still’ tense. The focus morpheme may be omitted in all of

these examples as in (63):

(63) a. Kato na Kwezi ba—lip-a amabanja omuduka
Kato and Kwezi 2 -pay-IND debts store

’Kato na Kwezi pays debts in the store’

b. Omwisiki ogu a -iluk-a kukila abaisiki bona
girl this is-run -IND than girls all

’this girl runs faster than all the girls’

0. Omwishulukale ba-kya-mu -pim -a emyenda
soldier 2 -STL-3s(OM)-measure-IND clothes
’the soldier, they are ’still’ measuring him’

clothes.’

d. abana bai -luk-e enjula yaija kugya
children 2-run-SUBJ rain is coming to rain

’the children should run, it is almost raining’

The contrast in meaning between these examples with and
without the focus morpheme provides evidence concerning its
meaning in Haya. While example (62.a) indicates that the
action of paying is still taking place; i.e., a present

progressive meaning, example (63.a) indicates that the

100

action of paying debts is habitual, as described above in
the section (3.3.5.1.1.) on the present universal habitual.
The same meaning contrast may be observed between examples
(62.b) and (63.b).

Note that our analysis of the role of the focus
morpheme differs from that of earlier work on the verbal
system of Haya. Previous authors have characterized the
present universal habitual with and without the focus
morpheme as two separate tenses (Byarushengo 1977, Nurse
1979, Rascher 1958, Betbeder and Jones 1949, Barlow 1951)18
We suggest that there is in fact only one tense, whose
meaning may be altered by the addition of this morpheme as
described in this sectionl There are two further,
observations that may be made in support of this point.
First, the focus morpheme may also be attached to nouns in
Haya, not just verbs. Although a complete description of
this phenomenon is outside of the scope of this work, the

following examples may be cited:

(64) a. e -ki-shusi ki-a-yat -ik-a
IV—7 -guard 7 —break-ST-IND

’the guard broke’

b. n -e -ki-shusi e -ki-a-yat -ik-a
FOC-IV-7 -guard IV-7 -break-ST-IND

’it is a guard which broke

101

Example (64.b) indicate that the main function of ni- also
in nouns is a "focus marker" as correctly observed by Moshi
(1986:2) who stated that Bl: "is relegated to pragmatics by
dealing with discourse, intentions and emphasis," (emphasis
provided). The use of the focus morpheme with nouns is
subject to the same phonological changes noted earlier in
its use with verbs. The use of the focus morpheme with both
nouns and verbs suggests that it is not a specifically
verbal morpheme used in the formation of a single tense. An
analysis which posits a separate "present progressive" would
be forced to claim that there are two morphemes, one nominal
and one verbal.

Second, recall that there are other Lacustrine and
Thaagicu Bantu languages in which the focus morpheme is used
with a much wider range of tenses than in Haya. If the
focus morpheme were analyzed as creating new tenses in these
languages as well, the number of tenses which they possess
would be nearly doubled, and the description of each of
these new tenses would be forced to repeat information about
the role of focus.

Returning to the data, the contrast between examples
(62.d) and (63.d) may be described in terms of emphasis:
whereas (62.d) is a suggestion, (63.d) is more of a command.
Similarly the contrast between (62.0) and (63.0) with and
without the focus morpheme may also be described in terms of
emphasis: whereas (62.0) may be translated as "the soldier,

they are indeed still measuring him clothes", (63.0) is

102

merely stating the facts, that ”they as still measuring the
soldier clothes."

The meaning of the focus morpheme may also be brought
out in more detailed examples which clarify the sorts of
contexts in which this morpheme is used. We will cite just

one, as follows:

(65) "Muliba baberwa, K’abantu balibanoba,
’You will be lucky, when people will avoid you’
ka balibabinga, ka balibajangura, ka
’when they will kick you out, when they will harass
you, when’
balibashaima eibala lyanyu okw’obona
’they will throw your name as if’
oti libi, nibabatura Omwana w’omuntu.”
’it is bad, because of a person’s child.’

(Lanctot 1959: 218 omut. Luka 6 (emphasis mine)).

From the above description and examples we may generalize
that the focus morpheme is restricted; to those verb forms
(apart from the infinitive in its verbal uses) which lack a
tense marker directly preceding the. stem, apart from the
"still" tense, in which the prestem tense marker is arguably

aspectual.

103
3.3.9. The Negative Morphemes

Haya has three) morphemes indicating negation, which
will be referred to as NEGl, NEG2 and NEG3. These three
morphemes take different forms , appear in di fferent
positions within the verb and are used in different

syntactic contexts.

3.3.9.1. NEGl

NEGl takes the form ti; (subject to phonological
changes; of. chapter 2, section 2.5.2.2) and directly
precedes the subject marker. NEGl forms are used
exclusively in independent clauses; that is, all clauses
which are not relative, conditional, temporal or imperative,
the domain of NEGZ (cf. the next section). It may be used
with any indicative tense apart from the remote future.

NEGl may be exemplified as follows:

(66) a. Omushaija ogu a -nyw -a malwa
man this Is-drink-IND beer

’this man drinks beer’

b. Omushaija ogu t -a -nyw -a malwa
man this NEGl-Is-dink-IND beer

’this man does not dink beer.’

104
(67) a. Abantu ba Kanyigo ba-laa-yombek-ile eshule

people of Kanyigo 2 -IF -build -RP school

’Kanyigo people have already build the school’

b. Abantu ba Kanyigo ti -ba-ka -yombek-ile shule
People of Kanyigo NEG1-2 -RMP-build -RP school

’Kanyigo people have not build the school.’

In order to negate the remote future the negative of the
near future is used, with an adverbial phrase denoting the
appropriate amount of time, as follows:
(68) abantu ti -ba-li-lekel-a kuiga maela neila

people NEGl-Z -NF-stop -IND to find money forever.

’people will never stop finding money for ever.’

3.3.9.2. NEGZ

NEG2 takes the form ;t§;_ and directly follows the
subject marker. As mentioned in the previous section, NEG2
is used to negate verbs occurring in relative, conditional,
temporal and imperative clauses. (Cf. Barlow 1951: 88 for
the use, in Kikuyu, of a morpheme probably cognate to this
one, with a similar if not identical distribution.) Within
these limits, it may be used with any verb form, again
excepting the remote future (RMF). NEG2 may be exemplified

as follows:

105
(69) Relative

a. a -ba—laa-shom -a ba-laa-sing-a e -mi-tiani
IV-2 -IF -study-IND 2 -IF -pass-IND IV-4 —exam

’those who will study will pass the exams’

b. a -ba-ta -a —shom -e ti -ba-a -sing-e mitiani
IV-2 -NEG2-IP-study-SUBJ NEGl-Z -IP-pass—SUJB exams

’those who will not study will not pass the exams’

(70) Conditional
a. ka -mu-laa-shom -a mu-laa-sing-a emitiani
COND-Zp-IF -study-IND Zp-IF -pass-IND exams

’if you study you will pass the exams’

b. ka -mu-ta -a -shom -e ti -mu-sing-e mitiani
COND-Zp-NEGZ-IP-study-SUBJ NEGl-Zp-pass-SUBJ exams

’if you don’t study you won’t pass the exams’

(71) Imperative: (i) indirect
a. mugambile a -it -e embuzi elikwilagula
tell him Is-slaughter-SUBJ goat black

’tell him to slaughter the black goat’

b. mugambile a -ta -it -a mbuzi elikwilagula
tell him Is-NEGZ-slaughter-IND goat black

’tell him not to slaughter the black goat’

106

(ii) direct
a. kol-a omutiani kileki
do -IND exam to-day

'(you) take the exam to-day’

b. o -ta -kol-a mutiani kileki
25-NEG2-do -IND exam to-day

’you should not take the exam to-day.’

'As before, in order to negate the remote future (RMF), the
negative of the near future is used, with an adverbial

phrase denoting the appropriate amount of time, as follows:

(72) a. ba-lika-ib -a empia za ishebo
2 -RMF -steal-IND money of their father
’they will (in the far future) steal their fathers

money.’
b. ba-ta -li-ib -a mpia za ba ishebo neila
2 -NEG2-NF-Steal-IND money of their father forever.

’they should never steal their fathers money ever.’

NEG2 is also used to negate a nominal infinitive:

107

(73) O -ku -ta -mal -a shule ku-a-ba-koz -a
IV-INF-NEGZ-finish-IND school 15-2-do-CAUS-IND
emilimo e -yo-mu-ta -li -kwend-a
work IV-4 -2p-NEG2 -like -IND
’not finishing school has caused you to do the work you

did not like’

An infinitive negated in this way may also be used as an

imperative:

(74) o -ta -ly -a!
Zs-NEGZ-eat-IND

’(you) don’t eat’

3.3.9.3 NEG3

NEG3 has the form Lag. It: appears in only one form,
prefixed to «i, which in turn is attached to the stem,
which is followed by the indicative morpheme. This verb

form may be schematized as follows:

NEG3 - kg - STEM - IND

As illustrated in the following sets of examples, this form

is used as a negative imperative:

108

(75) a. chumb-a amajuta galimu
cook -IND oil be there

cook! there is oil

b. Ka -ku -0humb-a (amajuta ti-ga-limu)
NEG3-UNR—cook -IND oil NEGl be in

’do not cook there is no oil’

(76) a. pim -a abana amaisho
examine-IND children eyes

examine the children’s eyes

b. ka -ku -pima abana amaisho
NEG3-UNR-examine children eyes
’do not examine the children’s eyes (e.g. when you

know for sure you are not a doctor).’

It seems likely that the morpheme k_a__:_ is a shortened and
cliticized form of the word leka, which may be translated as
’do not’ and is also used in negative imperatives. The
following sets of examples contrast this negative imperative

with the shortened form:

(77) a. leka ku -lim -a alwazi
do not INF-farm-IND rock

’do not farm on the rock.’

109

b. ka ~ku ~lim -a alwazi
NEG3—UNR-farm-IND rock

’do not farm on the rock’

(78) a. leka ~ku -tah -is -a omwana amaizi
do not INF-draw water-CAUSE-IND child water

'do not make the child to draw some water’

b. ka -ku -tah -is -a omwana amaizi
do not-UNR-draw-CAUS-IND child water

’do not cause the child to draw some water.’

3.3.10. The Suffixal Morphemes

In the preceding sections, we have seen that a verb
stem in Haya may be inflected for mood, subject agreement,
tense/aspect, finiteness, status as a relative,
conditionality, focus and negation. The suffixal morphemes
to be discussed in the next four sections constitutes a set
of morphemes which are suffixed to inflected verb as a unit,

without regard to its internal morphological content.

3.3.10.1. The Locative Morphemes
There are three locative morphemes which are suffixed
to the inflected verb. The first, :ha, indicates a location

which is near, within the context of discourse:

110
(79) a. abashaija balaabileta-ha ebitebe

men In Vb -LOC chairs

’the men will bring the chairs here (where I am

standing)’

b. obagambile baije bashuntama-ha
tell them to come In Vb -LOC

’tell them to come (and) sit here (e.g. where there

are empty chairs).’

The second, -ku indicates, within the context of discourse,
a location which is far. It is appropriate here too to
mention that originally the locative kg is added to the verb

from a separate word "oku" here, in which the initial vowel

o- is deleted.

(80) a. gambila abana bashuntama-oku (bashutamoku)
tell children In Vb -LOC

’tell the children to sit there’

b. nababona mbagenda-oku (mbagendoku).

I saw them In Vb‘ -LOC

’I saw them going there (lit: this side).’

In both examples there are phonological changes that take
place as discussed in chapter 2 section 2.5.2. In order to

clarify the qualifier "within the context of discourse," the

111

following examples may be cited in which, depending on the

context, the same location may be referred to using either

:11; or :1101:

(81) a. AbaTanzania babili nibashomela-ha Marekani
Tanzanians two In Vb -LOC America

’Two Tanzania are studying here in America.’

b. AbaTanzania babili nibashomela-oku (nibashomeloku)
Tanzanians two In Vb -LOC
Merekani
America
’Two Tanzanians are studying ’here/there’ in

America.’

The two examples above use two different locative morphemes
within the same repeated sentence. The contextual meaning
of the two differentiates the use of one form from the
other.. In example (81.a), the speaker might be somebody who
also lives in the United States, whereas example (81.b) may
be uttered by a visitor from Tanzania who is presumably
contextualizing the student residents in terms of distance
of Tanzania to United States. However both the situations
show some ambiguity in terms of distance depending on where
the sentence was uttered and the location of these students
in the U.S. The third locative morpheme, _:_m_u indicates

location inside:

112

(82) abana bashomele-mu omukiyumba eki
children In Vb-LOC in the room this

’the children should study in this room’

These three suffixal locative morphemes are cognate to
the noun-class and agreement morphemes of classes 16, 17 and
18 in other Bantu languages.19 In Haya, only one of these
classes, class 16, exists as a noun class (cf table 8).
Some other Bantu languages such as Shona (Fortune 1955),
Sotho (Doke and Mofokeng 1957) Tswana (Cole 1955) Zulu (Doke
1927) only to mention a few have all three classes. There
is no trace of the other two locative noun-classes, apart
from these three locative suffixal morphemes. This class is
illustrated in the following examples, in which a Class 16

noun occurs 0

(83) a. Ha-elu ha-li-yo Omushana
16-out lG-be-REL sunshine

’there is sunshine outside’

b. ha-nu ha-in -a eibugumo
16-in-16—have-IND warmth

’it is warm in here’

113

3.3.10.2. The Manner Morpheme

The manner morpheme :ge is suffixed to the inflected

verb immediately following the locative morpheme, if there
is one. In a statement, this morpheme indicates that the
action of the verb is performed well, or that the person or
thing existing in the state denoted by the verb is also

<ioing well” The manner morpheme is illustrated in the

following examples:

(84) a. baabishoma-ge (ebitabo)
In Vb -MNR (the books)

’they have read them well (the books).

b. twaalimaga-ge
In Vb -MNR

'We used to farm well’

This morpheme might be cognate to the Swahili Manner

Inorpheme -je ’how.’ (Ashton 1944, Polome 1957 and Maw 1984).

£3.3.10.3. The Intensifier Morpheme
The intensifier morpheme :kgg is suffixed to the

inflected verb immediately following the manner morpheme,

and hence also the locative morpheme, if these appear. As

its name suggests, the intensifier morpheme adds the meaning

tl‘lat the action denoted with the verb is performed with more

t311an usual intensity. It may be translated into English as

114

"indeedfl or "truly." The intensifier' morpheme is

exemplified as follows:

(85) a. mushune-kwo
In Vb -INT

’pinch him/her hard’

b. tulaalipa-kwo
In Vb -INT

’we will truly pay’

Note that in example (85.a) mu is a direct object morpheme
as discussed in (3.4.1). The occurrence of the intensive
morpheme is associated particularly with the use of the

repetitive verbal extension; (cf. sec 3.5) below.

3.3.10.4. The Question Morphemes

There are two suffixal question morphemes in Haya: :Ii_
"when" and :31 "what." The question morphemes are suffixed
to the inflected verb immediately following the intensifier
morpheme and the other suffixal morphemes preceding it, if
these are present. ‘These morphemes have the effect of
transforming a sentence in which the verb they are suffixed
to appears into a question, as seen in the following

examples:30

115

(86) a. nibachumba-ki
In Vb -Q

’what are they cooking?’

b. a yegesa-ki
In Vb -Q

’what does he/she teach?’

(87) a. olaazituza-li ente
In Vb -Q cows
’when are you taking out the cows to the pasture?’
b. nashoma-li olufalansa
In Vb -Q French

’when is he/she studying French?’

3.4. The Stem Morphemes

In the next two sections, we will discuss the stem
morphemes which are: the object marker (OM) and the
reflexive marker. These morphemes do not co-occur. They

are immediately attached to the verb-stem.

3.4.1. The Object Marker

The first stem morpheme we will discuss is the object
marker. The object marker directly precedes the root and
encodes the nOun-class of the object of the verb, just as
the subject marker discussed earlier encodes the noun-class

of the subject. (In spite of this identity of morphological

116

function, we will argue in chapter 5 that subject and object
markers have significantly different syntactic functions.)
A list of object markers for all of the noun-classes may be
found in table 11. Note that subject and object markers
have the same form, except for those in classes 1, 4, and 9.

Object markers may be exemplified as follows:21

(88) a. Paulo a -gend-e a-mu -ba-twal-il -e omuka
Paulo Is-go -SUBJ Is-3s-2 -take-APL-SUBJ home

"Paul should go (and) take him for you (pl) home’

b. a -ba-Hehe ba-ka -ga-ba-las -il -a amachumu
IV-2 -Hehe 2 -RMP-6 -2 -throw-APL-IND spears

’the Hehe threw the spears at them’

(89) a. bonene ba-gu-tem-e b-onka
they 2 -3 -cut-SUBJ 2-alone

’they should cut it alone’

~b. Omugulusi a -ta -zi-mil -a a -zi-kanj-e
the old man Is-NEGZ-lO-swallow-IND Is-lO-chew-SUBJ
’the old man should not swallow them, he should chew

them’

Note that the OM -zi- refers to things in class 10, such as
coffee beans. In Haya, a particular verb may contain up to

three object markers.“ When more than one object marker

117

appears within the verb, they refer to different objects of
the verb. The following examples illustrate two object
markers within a single verb, one referring to the direct

object of the verb and the other to the indirect object:

(90) a. Omusigazi a-laa-ba-mu -leb —el -a
the youth l-IF -3s-take-care-APL-IND

’the youth will take care of them’

b. embwa a -ka -gi -ba -lum —is -a
the dog Is-RMP-OM(9)-OM(2)-bite-CAUS-IND

the dog, he caused it (dog) to bite them (people).

In examples (90.a) and (90.b) we have glossed the applied
and causative morphemes which we will discuss in details in
section (3.5.2.) and (3.5.1.) respectively. Verb forms with
three object markers are not common, but they do exist. In
the following example, there are three object markers on the
verb, one referring to a direct object (DO), the other two

to indirect objects:

(91) tu-lika-gi -mu -ba -chumb-il -il -a -mu
Ip-RMP -OM(9)-OM(3s)-OM(2)-cook -APL-APL-IND-LOC
’we will in the (far remote future) cook for him in it

on their behalf.’

118

When. a verb has more than one object marker, the
relative order of these object markers with respect to the
stem depends on a number of features possessed by their
references, which may or may not be marked on the object
markers themselves. These features are: person and number,
thematic roles and animacy. The effect of all of these
features in determining the order of object markers has been
analyzed in previous studies of those Bantu languages which
permit multiple object markers, (including Haya). These
studies include: Duranti (1979) for Haya and Shambala, Hyman
and Duranti (1982) on Bantu in general, Hawkinson and Hyman
(1974) on Shona, Sabimana (1986) on Kirundi and Kidima
(1987) on Kiyaka (and a section on Haya).

In all of these studies, values for these features are
arranged into hierarchies, and whether or not one object
marker occurs closer to the verb stem than. another is
determined by whether or not it possesses a value for a
feature which is higher on the hierarchy than the value for
that feature possessed by the other object marker(s). The
writers cited in the previous paragraph propose four
hierarchies, all of which are applicable to Haya: a person
hierarchy, an animacy hierarchy, a case hierarchy, and a
number hierarchy. ’These hierarchies may be schematized as
follows, the feature farthest to the right being the closest
to the verb stem (Duranti 1979 ehd Kidima 1987):

Person hierarchy: 3rd > 2nd > lst

Animacy hierarchy: Non-human > Human

119

Number hierarchy: Plural >.Singular

Case hierarchy: Instrumental/Patient > Goal/Benefactive

The relevance of each of these hierarchies to the order of
Haya object markers will be discussed in turn.

The operation of the person hierarchy in Haya may be
seen in the following examples. In the (a) example of each
pair, the object marker referring to the person hierarchy
appears closer to the stem; the (b) example shows that the

reverse order is ungrammatical:

(93) a. Ba-ku-m -p -el -a ebitooke.
2 ~23—Is-give-APL-IND bananas
’They give you bananas for me.’

’They give me bananas for you.’

b. *Ba-m-ku-p-el-a ebitooke.

(94) a. Ba-mu-ku-m -p —el -a ebitooke.
2 —3s-Zs-Is-give-APL-IND bananas
’They give him bananas for you.’

*’They give you bananas for him.’

b. *Ba-ku-mum-p-el-a ebitooke.

In these examples, the object. markers contrast only in

person. but are constant in terms of animacy, case and

number. In this way, only the effects of the person

120

hierarchy have been tested. However, it can also be shown
that when the requirements of the person hierarchy conflict
with those of the other hierarchies, the person hierarchy
takes precedence. The following two sets of examples

indicate that the person hierarchy takes precedence over the

number hierarchy, as in (95), and also over the case
hierarchy, as in (96) (note some phonological changes):
(95) a. Ba-mu-tu-h -el -a ebitooke.

2 -3s-Ip-giveeAPL-IND bananas
’They give him bananas for us.’

’They give us bananas for him.’

b. *Ba—tu-mu-h -el -a ebitooke.

2 -Ip-3s-give-APL-IND bananas

(96) a. A-ka -mu-ku-leet -el -a.
I-RMP-3s-23-bring-APL-IND
’He brought him to you.’

’He brought you to him.’

b. *A-ka -ku-mu-leet -el -a

I-RMP—Zs-3s-bring-APL-IND

The conclusion that the person hierarchy takes precedence
over the case hierarchy may be drawn from the observation

that the order of the object markers in examples (95) and

121

(96), (the set contrasting person. and. case) may' not be
varied to reflect case contrasts. Instead, the single
possible order is ambiguous between both case

interpretations, so that either object marker may refer to
either case.

It would also be desirable to contrast the person and
animacy hierarchies in this way. However, this contrast is
not testable, because the relevant example would require an
object marker referring to an inanimate object which is
ranked higher on the person hierarchy than an object marker
referring to an animate object: i.e., the inanimate object
would also have to be first or second person, and inanimates
cannot be referred to in either the first or second person
in Haya.

Nevertheless, with some complications, it can also be
shown that the animacy hierarchy takes precedence over the
number hierarchy in the same way as the person hierarchy
does. The following set of examples contrasts the animacy

hierarchy with the number hierarchy:

(97) a. A -ki-ba-gul-il -a
Is-7 -2 -buy-APL-IND
’He buys it for them.’

#’He buys them for it.’

Note that in this example, the second interpretation is

ruled out for pragmatic reasons (an inanimate object is

122

unlikely to have anything bought for it), contrary to in
example (96) above. However, reversing the order of the
object markers, in violation of the animacy hierarchy,

produces an ungrammatical sentence:

b. * A -ba-ki-gul-il -a.

Is-Z -7 -buy-APL-IND

We will say for this example that it is ungrammatical
because it violates the animacy hierarchy. However, it
could also be ungrammatical because of the case hierarchy,
since the object marker referring to the patient is closer
to the verb stem than the object marker referring to the
applicative (benefactive). The opposite interpretation,
which does not violate the case hierarchy is, as we have
seen, ruled out on pragmatic grounds. The evidence is
indeterminate in this case, and such indeterminacy cannot be
eliminated from this discussion. In any case, the example
indicates that the number hierarchy may be overridden.

Total indeterminacy is found when the animacy hierarchy

is contrasted with the case hierarchy.

(98) a. ba-ka -gi-mu-twal-a -mu
2 -RMP-9 -3s-take-IND-LOC
’They sent him to it (9; i.e., the cave)’

#’They sent it to him.’

123

As in the above examples, the second interpretation is ruled
out for pragmatic reasons (a cave cannot be sent somewhere).
However, as before, reversing the order of the object
markers, in violation of the animacy hierarchy, produces an

ungrammatical sentence:

b. * ba-ka -mu-gi-twal-a -mu

2 -RMP-3s-9 -take-IND-LOC

This example could be ruled. out by either the animacy
hierarchy or the case hierarchy, since both are violated and
this example does not distinguish the two.

The examples we have seen so far indicate one hierarchy
taking absolute precedence over another: an. ordering of
object markers which contradicts the hierarchy is
ungrammatical. However, when the number of case hierarchies
are compared, the effects of each are relative rather than
absolute. The number hierarchy has the effect of placing an
object. marker (which refers to) a singular* noun phrase)
closer to the verb stem than an object marker referring to a

plural noun phrase, as seen in the following example:

(99) a. A -ka -ba-mu-leet -el -a
3s-RMP-2p-33-bring-APL-IND
’He brought them to him’

’He brought him to them’

124

In this example, the singular object marker occurs closer to
the verb stem than the plural object marker, in accordance
with the number hierarchy. Note that these object markers
refer to two thematic arguments of the verb: the patient
(what is being brought) and the goal (who the patient is
being brought to). As Shown ill the translations, either
object marker may refer to either thematic argument.

Now consider the following example, in which the object
markers of the previous example appear in the opposite

order, in violation of the number hierarchy.

b. A -ka-mu-ba-leet -el -a
he-P3-3s—Zp-bring-APL-IND
’He brought him to them’

*’He brought them to him’

Here only one interpretation is possible, in. which the
plural object marker closer to the stem is the goal. This
example indicates that the number hierarchy may be violated
when the results conform to the requirements of the case
hierarchy. The interpretation of ’example (99.b) in which
both the number and the case hierarchies are violated is not
possible.

Note as well that the case hierarchy is violated in
(99.a) but this violation is possible only when the number

hierarchy has been obeyed.

125

These examples indicate that both the number and the
case hierarchies are relevant to the order of object markers
in Haya, but only in a relative sense. If number were
completely irrelevant, then (99.a), in which number is
obeyed but case violated, should be ungrammatical, but it is
not. If case were completely irrelevant then the second
interpretation of (99.b) should be possible, but it is not,
it is precisely when both number and case are violated that
complete unacceptability results.

The data we have presented in this section suggest
that, of the four hierarchies, the effects of person and
animacy are absolute, whereas the effects of number and case
are cumulative. This observation in turn suggests that when
the animacy and case hierarchies are in conflict it is in
fact the animacy hierarchy that takes precedence, even

though this effect is not directly observable.

3.4.2. The Reflexive Marker

The reflexive marker in Haya occupies the same position
immediately preceding the verb stem as the object markers.
It signifies coreferentiality between the subject and the
object of a sentence. Unlike ‘the object. markers,. the
reflexive marker takes the invariant form :e;, subject to
phonological changes described in chapter 2 (section 2.5.2).
Also, a verb stem may contain only one reflexive marker, and

no object markers when the reflexive marker is present. The

reflexive marker may be exemplified as follows:

126

(100) a. abana ba-e -shal-a
children 2 -RFM-cut -IND

’the children have cut themselves’

b. abojo ba-li -e -mois —a
boys 2 -RFM-shave-IND

’the boys will shave themselves.’

3.5. The Verbal Extensions

As mentioned above in section 3.2, a verb stem may
consist of a verb root plus one or more morphemes which
modify its lexical meaning and/or alter its argument
structure. These morphemes are usually referred to as the
verbal extensions, and this section will briefly describe
and exemplify their use.23 (See Byarushengo et al. 1977,
Scotton 1967, Khamisi 1983, Maw 1984). .A single root may
contain more than one extension, each of which typically
contributes to the total meaning of the stem, with certain
cooccurrence restrictions which will be discussed in the
next chapter. In this chapter, the meaning of each
extension will be illustrated in isolation. In the absence
of other extensions, each extension is suffixed to the root,

directly preceding the final vowel, schematized as follows:

(101) ROOT - EXT - IND
SUBJ

127

When more than one extension is present, all are suffixal,
but they are ordered with respect to each other in ways that

will be described in the next chapter.

3.5.1. The Causative Extension

The causative extension in Haya takes one of the
following forms: gig-z—es- or -iz-/-ez- (subject to vowel
harmony, described in chapter 2, section 2.5.2.), 3;; or
:y;.3‘ In order to describe the meaning which the causative
extension contributes to the verb stem, we adopt Ashton’s
(1944: 232) description for Swahili, which is "to" express
causation." Nevertheless, as. will be seen below, while a
definition such as this one may cover most of the uses of
the causative, it does not cover all of them. The causative
extension. may .be exemplified in the following pairs of
examples, in which the first example contains an unextended
verb and the second contains the same verb plus the

causative extension:23

(102) a. Abana ba-a -bon-a ente
children 2 -IP-see-IND cow

’the children have seen the cow’

b. Abana ba-a -bon-es -a nyina-bo ente
children 2 -IP-see-CAUS-IND mother their cow
’the children have shown their mother a cow’

(lit. caused their mother to see a cow)

128

(103) a. Omulilo gu-a -yak -a
fire 3 -IP-blaze-IND

’the fire has blazed’

b. abana ba-a -yak —y -a omulilo
children 2 -IP-blaze-CAUS-IND fire

’the children have set a fire’

(104) a. Abakozi ba-laa-kol-a emilimo mingi
workers 2 -IF -do -IND work many

’the workers will do a lot of work’

b. Abebembezi ba-laa-kol -z +a abakozi emilino
leaders 2 —IF -work-CAUS-IND workers work
mingi
many

’the leaders will make the workers do a lot of

work’

The following use of the causative has the effect of adding

an object referring to the instrument of the action:

(105) a. Abafungwa ba-li-lim-a eshamba
Prisoners 2-NF-cultivate-IND farm
’the prisoners will cultivate the farm (e.g. to-

morrow)’

129

b. Abafungwa ba-li-lim -is -a eshamba
Prisoners 2 -NF-cultivate-CAUS-IND farm
etarakita
tractor
’the prisoners will cultivate the farm with a

tractor’

Note that if the causative extension were not present, the
instrument could only appear in the sentence as a

prepositional phrase:

(106) Abafungwa balilima eshamba nametargkiti

’prisoners will cultivate the farm withma tragtgr’

muc-

3.5.2. The Applied Extension

The applied extension has also been referred to as

benefactive extension (Byarushengo et al. 1977 for Haya) and

as prepositional extension (Ashton 1944 for Swahili). In
Haya, it takes the form -' - - -, depending on vowel
harmony. It adds an object to the argument structure of the

verb; without the extension, this object would have to be
expressed in a prepositional phrase, or else be
unexpressable. This object may represent a number of

thematic roles, three of which are mentioned here:

130

(1) An applied verb may denote an action in relation to a

person, normally for his/her advantage:

(107) a. abaalimu ba-ka -leet -a amaua
teachers 2 -RMP-bring-IND flowers

’the teacher brought flowers’

b. abaalimu ba-ka -leet —el -a abanafunzi babo
teachers 2 -RMP-bring—APL-IND students their
amaua
flowers
’the teachers brought some flowers for/to their

students’

Note that there is no example possible in which the verb is

unextended and the beneficiary expressed in a propositional

'phrase. The applied object may also represent an "anti-
beneficiary." This can be exemplified by a sentence like.
(108) a. Omushuma owo bakwete a -ka -ba-f -el -a

the thief whom they caught Is-RMP-Z -die-APL-IND
omumikono yabo
hands theirs

’the thief whom they caught died in their hands’

131

This sentence may be interpreted as "anti-beneficiary"
unless the catchers wanted their thief to die before taking

him to court, a propasal which is highly unlikely.

(ii) An applied object may denote a place.

(109) a. ente e-ka -f -a
cow 9-RMP-die-IND

’the cow died’

b. ente e -ka -f -el -a omulweya
cow IV-RMP-die-APL-IND pasture

’the cow died in the pasture’

(iii) An applied object may' denote a cause, reason or

motive:

(110) a. Richard, ba-lika-mu-fung-a
Richard 2 -RMF -2p-jail-IND

’Richard, they will jail him’

b. Richard ba-lika-mu-fung-il -a obushuma
Richard 2 -RM -23-jail-APL-IND robbery

’they will jail Richard for robbery’

132

3.5.3. The Passive Extension

.The passive extension in Haya takes the form ;E; (with
one irregular exception, to be noted below). In a sentence
in which the verb contains the passive extension, the object
of the unextended verb is the subject, and the subject of
the unextended verb appears optionally in a prepositional

phrase, exemplified as follows:

(111) a. embwa y-a -lum -a .abashuma
dog 9-Ip-bite-IND robbers

’the dog bite the robbers’

b. abashuma ba-a -lum -w -a embwa
thieves 2 -IP-bite-PASS-IND dog

’the robbers were bitten by the dog’

(112) a. Omushaija y -a -mu-tel-a
man Is-Ip-3s-hit-IND

’the man hit him’

b. n -a -tel-w -a omushaija
Is-IP-hit-PASS-IND man

’I was hit (by) the man’

Note that when an agent is human there is a preference not
to have a propositional phrase even if it is possible as

exemplified in 113.

133

(113) a. abashuma baalumwa na embwa

’the robbers were bitten by the dog’

b. natelwa na omushaija

’I was hit by the man’

The following exemplifies an irregular form of the passive

extension:

(114) abana bato ba-h -abw -a ebitoke na
children young 2 —give-PASS-IND bananas by
nyinenkuru bo
grandmothers their

’the children are given bananas by their grandmother’

In addition, there are verbs which are used exclusively in
the passive voice with active meanings, as observed by

Rascher (1958:39):

(115) a. abalimu ba-yeb -w -a omukutano
teachers 2 -forget-PASS-IND meeting

’the teachers forgot the meeting’

b. omukazi ogu a-tamw-a abagenyi
woman this hate visitors

’this woman hates visitors’

134

3.5.4. The Reciprocal Extension

The reciprocal extension in Haya takes the form

-a -—. This extension changes the meaning of'ti verb to
indicate that the action is reciprocal or mutual; i.e., done
to "one 'another." The reciprocal extension may be

exemplified as follows:

(116) a. Halima n -a -chuchul-a Farida eishoke
Halima Is-IP-comb -IND Farida hair

’Halima is combing Farida’s hair’

b. Halima na Farida ni -ba—chuchul-angan-a eishoke
FOO-2 -comb -RCP -IND hair

’Halima and Farida are combing one another’s hair’

Because of its meaning, a sentence containing a verb with
the reciprocal extension is well-formed only if the verb‘
denotes an action which may be performed mutually and has a
plural subject. The following pair of examples are
ungrammatical because, in (a), the intransitive verb does
not denote an action which may be performed mutually and, in

(b) the verb has a singular subject:

(117) a. *Abaisiki ba-laa -shabuk-angan-a. omwiga
girls 2 -IF -cross -RCP -IND river

’the girls will cross one another the river.’

135

b. *Omwisiki n -a -chuchul-angan-a eishoke
girl Is-IP-comb -RCP -IND hair

t’The girl is combing one another the hair.’

Even if a verb is transitive, and has a plural subject, its

use as a reciprocal may be barred pragmatically.25

(118) abashaija aba ba-a -chumb-angan-a
men these 2 -Ip-cook -RCP -IND

’these men are cooking ’one another”

Example (118) is acceptable only when it is enterpreted

idiomatically as exemplified in (117).

(119) Abashaija aba ba-a-chumb-angan-a kukola ebyashala
men these 2-IP MOTIVATE-RCP-IND to-do business

’these men have motivated one another to do business.’

3.5.5. The Repetitive Extension

The term "repetitive extension" is used in this work to
refer to three extensions which express various forms of
intensity, including repeated, frequent, continuous or
forceful action. These extensions are: 1) -' - - 2-
(subject to vowel harmony), 2) :aij; and 3) :ul;, and may be

exemplified as follows:

(120)

(121)

(122)

136

Koku a -shon-a emyenda
Koku Is-sew -IND clothes

’Koku sews clothes’

Koku a ~shon-elez-a emyenda
Koku Is-sew -RPT -IND clothes

’Koku sews clothes repeatedly’

Bojo ba-gamb-ile ba-angu —e tugende
please 2 -tell-RP 2 -quick-SUBJ (so that) we go

’please tell them to be quick so that we go’

Bojo bagambile ba-angu -iliz-e tugende
please tell them 2 -quick-RPT -SUBJ (so that) we go

’please tell them to travel fast so that we go’

Omwana ogu a—li-nag -a ebyakula omunju
child this I-NF-throw-IND food in the house

’this child will throw the food in the house’

Omwana ogu a-li-nag -aij-a ebyakulya omunju
child this I-NF-throw-RPT-IND food in the house
’this child will throw the food continuously/

carelessly in the house’

137

(123) a. abashuma ba-ka -sik -a ebiliibwa
thieves 2 -RMP-pull-IND cassava

’the thieves pulled the cassava’

b. abashuma ba-ka -sik -ul -a ebiliibwa
thieves 2 -RMP-pull-RPT-IND cassava
’the thieves pulled forcefully/hurriedly the

cassava’

Since these three morphemes have a common strand of meaning,
we have decided to treat them as a single extension in this
work. In this respect, we follow Rascher (1958:84), who
uses the word "intensive" to refer to this set of morphemes
for Haya, and Doke (1943:22) who uses the term "extensive"
to refer to different forms appearing in ‘various Bantu
languages. 131 Zulu (Doke 1927:148) the form is -isis-
suffixed to the verb-stem. However, it is possible that
future research will suggest that each of these morphemes is
associated most often with a particular aspect of intensity,
and. such. an .association would. provide justification for
assuming that they constitute separate verbal extensions.
For example, -' ' - - - may turn out to be the only one
of the three which consistently denotes repeated action (as
opposed to forceful action). For now, however, we will

assume the present analysis.

138
3.5.6. The Stative Extension

The stative extension in Haya takes the form -ik-/ek-,
subject to vowel harmony. Its usage is similar to that of
the passive: in a sentence in which the verb contains'the
stative extension, the object of the unextended verb is the
subject of the stative verb. However, there are two
significant differences. First, in a sentence containing a
stative verb, the subject of the unextended verb may not
appear, in a prepositional phrase or in any other form.
Second, another nuance of meaning which distinguishes the
stative from the passive is that of potentiality: the
subject of the stative verb is implied to be capable of
undergoing the action denoted by the unextended verb. For
this reason, the stative extension is readily translatable
using the English adjectival suffix .1832.qu The stative

extension may be exemplified as follows:

(124) a. e -nsi e-gi ° t -a -11m -ik-a (telimika)
IV-land IV-this NEG-IP-cultivate-ST-IND

’this land is uncultivatible’

b. e -bi-kombe bi-ona bi-a -yat -ik-a (byatika)
IV-8 cup 8 -all 8 -IP-break-ST-IND

’all cups are breakable’

139

c. a -ma-banga ti -ga-temb —ek-a bwangu
IV-6 -mountain NEG-6 -climb-ST-IND quick

’the mountains are not quickly climbable’

Other scholars have used the term Neuter (e.g. Doke 1927,
1943, Doke and Mofokeng 1957, Cole 1955 to mention a few) to
refer to the stative extension (a term used by both Rascher

(1958) and Byarushengo (1977)).

3.6. Conclusion

This chapter’ has described the .meanings and linear
order of the 23 morphemes which occur in the various forms
of the Haya ‘verb. The next. chapter will describe the

cooccurrence restrictions holding between these morphemes.

140
CHAPTER 3 ENDNOTES

1 The reversive morpheme will not be discussed further.
It is in this work treated as unproductive in Haya (see
appendix B) and thus regarded as lexical. See also Khamisi
(1983:94) on the treatment of reversive (called conversive)
as a lexical group in Swahili also a Bantu language.

Also in order to help readers’ confusion in the use of
terms, compared to other writers, we list the terms we used
along with other different terms in different studies on the
grammar of African languages. This list contains only the
frequently used terms in this work (individual authors will
be mentioned as we discuss the concepts).

Reversive: Conversive (Ashton 1954, Khamisi 1983)
Benefective: Applied, Prepositional

Repetitive: Augmentative

Reciprocal: Associative

Stative: Neuter (Rascher 1958)

4 Several writers including Inn; not limited to
Byarushengo 1977, Maw 1985, Perez (Harford) 1986, have used
the term verbal extensions.

4 In Haya society a woman gives birth for the man; that
is the society practises the patriarch system.

4 This term is not used for these morphemes in Haya or
in any other Bantu language in any source that we know of.
We have taken the term from Bybee; of chapter 5 for

justification.

141
5 See Ashton et al. (1954:18) quoting the definition

from Doke’s Bantu Linguistic Terminology in footnote 1.

4 From now on, subject morphemes will be indicated by
numbers corresponding to their noun classes. As regards
noun class 1 and 2, first and second person singular will be
indicated by a number e.g. 23 where 2 is for the second
person and s for singular while p will be for plural.

7 Sabimana 1986 argues convincingly that noun classes
15, 16 and 17 denoting place in Kirundi (locatives), a
language closely related to Haya language are not noun
classes (class markers) for the following reasons:

i noun. class markers as Sabimana. calls them are

selected by at least one stem, but not locatives

ii locatives never trigger agreement as do other

markers

iii Normal class markers attach to stems but locatives

can attach to proper names and personal pronouns

iv They do not show the same alternations in the same

environments as other class markers do.
Thus Sabimana suggests they be treated as prepositions, a
suggestion which we did not incorporate in this study.

4 The indicative and subjunctive morphemes have also
been referred to as mood suffixes with reference to Swahili
(Vitale 1981:17), and also as final vowel (Perez (Harford)
1985:10)) for Kikuyu and Shond.

4 The number of tenses will be modified after

discussing the focus morpheme in section (3.3.9).

142

1° We were unable to obtain Moshi’s full version of her
paper presented at the 19th African Linguistics at Boston
University April 14-17, 1988. We only had access to her
outline distributed at the conference and a personal
conversation with the author.

11 In his treatment of Kikuyu, Clements (1984:293)
treats -ire as having a final vowel 1;, so that the morpheme
itself is -dJ~u Such an analysis has not been adopted in
this work. Note also that [l] and [r] are allophones of the
same phoneme /l/ in most Bantu languages. See chapter 2
(14). Kikuyu :13; is cognate to Haya :ilg.

13 The habitual morpheme -ag- in Haya is cognate to
~ag- in Kikuyu (Barlow 195) where it is used to form three
past imperfect tenses: "immediate past imperfect," "near
past imperfect” and "remote past imperfect." Barlow notes
(ibid:142) that when used in. past time in. general, it
expresses (like in Haya) an action done habitually, or
recurrently . . . .Other Bantu languages which use this form
as a habitual morpheme are among others Sukuma, Nyamwezi
(personal conversation with Kahigi). Furthermore, recently
Wilt (1988:58 first draft) notes a very suggestive-looking
verbal morpheme -ak- in Bukavu Swahili, which may be a loan
from ethnic Bantu languages.

13 We also note Polome’s 1976:77 suggestion that the
:31; nominal agentive suffix in standard Swahili is cognate
to ‘;ag; expressing a constantly repeated or a habitual

state.

143

14 In Haya society a man marries and a woman is
married, therefore if example 46.b were to be interpreted in
the Haya context, it would have read "leave him alone, he
will marry when he will like."

15 Compound tenses are not discussed in ‘this work.
However, they have mostly been covered in the discussion of
the simple tenses. For details of these tenses refer to
Nurse 1979, Byarushengo 1976, Hyman and Byarushengo 1984 in
Clements and Goldsmith (eds) p. 53-104.

14 The IV has also been called variously a "double
prefix" (Guthrie 1974), a "noun augment" (de Blois 1970),
and pre-prefix (Masamba 1976).

17 When an object is relativized, the relative clause
is marked with a separate relative marker, and the form of
the verb is not changed from its non-relative form except
for tonality which is realized as log (L) on the object (See
Hyman and. Byarushengo 1984:79 in. Clements and Goldsmith
(eds) 1984).

14 Barlow 1951:117 calls the focus morpheme the initial
marker n;;.

1’ Classes 16, 17 and 18 are numbers assigned to the
three locative noun classes in the Welmers (1973) numbering
system. See also section (3.3.10.1).

4° The question morpheme :ki constitutes an exception
to the generalization mentioned in the introduction to
section 3.3.10, that suffixal morphemes are attached to the

inflected verb as a unit, without regard to its internal

144

morphological content. This morpheme can, in fact, only be
suffixed to transitive verbs, an observation that will be
taken up in the next chapter (section 4.6). Therefore, in
order to attach this morpheme correctly, it is necessary to
know the argument structure of the root and how this
argument structure has been affected by verbal extensions.)

31 Note that object. markers may’ co-occur ‘with post
verbal object NPS in Haya:

a. a -laa-gi-mu-gul-il -a ishe emotoka

Is-IF -4 -3s-hit-APL-IND father car

’he will buy it for him the father a car’

b. embwa zi-lika-ba-lum -a Amina na Fatuma

dogs 10-RMF -2 -bite-IND Amina and Fatuma

’the dogs will (in the far future) bite them Amina

and Fatuma.’
The co-occurrence of the OM and object NPs is unusual but
when it happens, the object NP(s) function as a topic, a
point further discussed in chapter 5.

3* Bantu languages vary as to how many object markers
are permitted within a single form. Swahili for instance
permits only one OM as is the case of Kikuyu (Barlow 1951).
Other Bantu languages such as Nkore and Kiga (Taylor 1985)
languages of Uganda permit more than one OMs. '

33 Only productive extensions are considered: writers
on Bantu languages tend to take anything that appears on a

couple of verbs with roughly the same meaning and elevate it

to the rank of extensions, but we will not do this here.

145

34 This is not meant to be the maximal forms of
causatives in Haya, nor are all uses discussed.

*5 Literal translations may be acceptable in folktales,
like the Haya folktale where the Hare and the Tortoise are
supposed to cook each other. This suggests that verbs with
ridiculous meanings may nonetheless exist under exceptional

contexts of usage.

146

 

Table 6. The Internal Structure of Conjugated Verb in Haya

1 Initial Vowel IV

2 Conditional COND
3 Negative NEG

4 Focus FOC

5 Subject Marker SM

6 Tense TNS

7 Infinitive INF

8 Reflexive RFM
9a Direct Object (Accusative) DO
9b Indirect Object (Dative) IO
10 VERB ROOT ROOT
11 Reversive RV
12 Causative CAUS
13 Applied APL
14 Stative ST
15 Passive PASS
16 Reciprocal RCP
17 Repetitive RPT
18 Indicative IND
19 Subjunctive SUBJ
20 Locative 9 LOC
21 Manner particle MNR
22 Interrogative (Question) Q

23 Intensifier INT

 

147

 

 

Table 7: Inventory of Morphemes and Allomorphs.
Morpheme Allomorphs

1. IV a/-012, 016, 0112, 0116
e/-cl4, 015, 017, 018, 019, 0110
o/-cll, 013, 0111, 0113, 0114, 0115, 0118

2. COND ka-

3. NEG ti/-c{main‘clause}
ta/-{other}
ka/-INF

4. FOC ni-

5. SM a/-cll, ba/-012, gu/-cl3, e/-cl4, 019,
li/-015, gal-0116, ki/-cl7, bi/-018,
zi/-cllO, lu/-clll, ka/-0112, tu/-0113,
bu/-cll4, ku/-cll5, 17, ha/-0116,
mu/-cll8.

Numbers correspond to noun-class
numbering system.

6. TNS puH/-, STL/-kya-, IPl-a, NPS/—a--ile,
RP/-ile, RMP/-ka-, FRP/-la--ile,
IPH/-a--ag-, IF/-laa-, NF/-li-,
RMF/-lika-.

7. INF ku-

8. RFM —e-

9. DO/IO (OM): m/-cll:1, ku/-cll:2,15,17,
mu/-cll:3, 2:2, 18, tu/-2:1,13,
ba/-012:3, gu/-013, gi/-cl4,9, li/-cl5,
gal-016, ki/-cl7, bi/-cl8, zi/-cllO,
lu/-0111, ka/-0112, bul-cll4, ha/-0116.

10. ROOT (STEM) varies

11. RV -ulul-/-olol-

12. CAUS -is-/-es, -y-, -z-, -s-.

13. APL -il-/-el

14. ST -ik-/—ek

15. PASS -w—

16. RCP -angan-

17. RPT -iliz/-elez, -aij-, -ul—

18. IND -a-

19. SUBJ - -e

20. LOC -ha-/-ku—/-mu-

21. MNR -ge '

22. Q -ki-/-li-

23. INT -kwo

 

This list includes vowel harmony as discussed in chapter 2.

148
Table 8: List of Noun Classes and Subject Morpheme Markers.

 

 

Class Noun prefix Person Number Subject
Marker

1 Mu- 1 SG n-

2 SG 0-

3 SG a-
2 BA- 1 PL tu-

2 PL mu-

3 PL ba-
3 Mu- SG gu-
4 M1- PL e-
5 i-/Li SG li-
6 MA- PL 8a-
7 KI- SG ki-I
8 BI- PL b1-
9 N- SG e-
10 Lu- SG lu-
11 N- PL zi
12 KA- SG ka-
13 Tu- SG/PL tu-
14 Bu- 7 PL bu-
15 Ku- SG ku-
16 HA- - - ha-
17 - - -

18 - - -

 

Table 9: Haya Affiraative and Negative Tenses

1.4E3

 

 

 

 

Affiraative

Tense Harker Exanple Gloss
Present Universal Habitual (PUH) - ba-lua-a ’they usually bite’
Still Tense (STL) -kya- zi-kya-nyw-a ’they (cows) are still drinking’
Iaaadiate Past (IP) -a- ba-a-ching-a ’they have closed’
Near Past (NPS) -a- -ile tu-a-zin-ila ’we have sung’
Recent Past (RP) - -ila gu-gw-ile ’it (tree) fell’
Reaota Past (RHP) -ka- ba-ka-ib-a ’they stole’
Far Renata Past (FRP) -la- -ila e-la-gw-ile ’they (trees) have already fell

(long ago)’
Iaaadiata Past Habitual (IPH) -a- -a9- ba-a-tea-ag-a ’they were habitually cutting’
Twaadiata Future (IF) -laa- ga-laa-yatik-a ’thay (eggs) will break’
Near Future (NF) -li- a-li-lua-a ’ha will bita’
Relate Future (RHF) -lika- o-lika-jua-a ’You will in the resale future
abusa.’
Negative

Tense Barker Exaapla Gloss
Present Universal Habitual (PUH) - ti-ba-lua-a ’they do not usually bita.’
Still Tense (STL) -kya- ti-zi-ki-nyw-a ’they (cows) are no longer drinking’
laaadiata Past (IP) -a- ti-ba-a-ching-a ’they have not closed’
Near Past (NPS) -a- -ila ti-tu-ka-zin-ila ’wa have not sung’
Recent Past (RP) - -ila ti-gu-gw-ila ’it (tree) did not fall’
Reacts Past (RHP) -ka- ti-ba-ka-ib-ag-a ’they have never stola’
Far Raaota Past (FRP) -la- -ile ta-Ra-gw-ag-a ’they trees have never fall’
Iaaadiata Past Habitual (IPH) -a- -ag- ti-ba-a-teI-ag-a ’they were habitually not cutting’
laaediata Future (IF) -laa- ti-ga-yatik-a ’thay (eggs) will not break’
Near Future (NF) -11- t-a-li-luI-a ’ha will never bite’
Raaota Future ' (RHF) -lika- t-o-li-qu-a neila ’you will never ever abusa’

 

Table 10: Subject Relative Horpheees

1.5()

 

 

Class IV SH Exaeple Gloss
1 a- a a-a-gend-a (he) who does e.g. ean
2 a- -ba aba-gend-a (they) who go e.g. een
3 o- -gu o-gu-laa-gw-a (it) which will fall e.g. tree
4 e- -e e-e-laa-gw-a (they) which will fall e.g. trees
5 e- -li e-li-kya-lil-a (it) which is still crying e.g. eye
6 a- -ga a-ga-kya-lil-a (they) which are still crying e.g. eyes
7 e- -ki e-ki-a-bul-a (it) which got lost e.g. book
8 e- -bi e-bi-a-bul-a (they) which got lost e.g. books
9 e- -e e-e-li-bul-a (it) which will grow e.g. cow
10 e- -zi ezi-li-kul-a (they) which will grow e.g. cows
11 o- -lu o-lu-laa-ij-a (it) which will coee e.g. death
12 a- -ka a-ka-li-chueb-w-a (it) which will be cooked e.g. little banana
13 o- -tu o-tu-laa-gul-w-a (they) which will be bought e.g. little things
14 o- -bu o-bu-a-sheshek-a which has poured e.g. flour
15 o- -ku o-ku-laa-gue-a (it) which will becoee hard e.g. cooking
16 a- -ha a-ha-laa-sheeel-a (it) which will be good e.g. place
17 - ku -
18 - eu -

 

151
Table 11: Direct/Indirect Objects.

 

Noun Class Person Number Objects: (DO/IO)

1 1 86 -m-

2 SG -ku-

3 SG -mu-
2 1 PL -tu-

2 PL -mu-

3 PL -ba-
3 SG -gu—
4 PL -81-
5 ' SG -li-
6 ~ PL -ga-
7 SG A -ki-
8 PL -bi-
9 SG -gi-
10 SG -lu-
11 PL -zi-
12 SG -ka-
13 . SG -tu-
14 ’ PL -bu-
15 ' SG -ku-
16 - _-ha-
17 - ' -ku-

18 - -mu-

 

CHAPTER 4

THE MORPHEMES OF THE VERB: COOCCURRENCE RESTRICTIONS

4.1. Introduction

Chapter 3 described the linear order of the verbal
morphemes with little comment on cooccurrence restrictions
which hold between these morphemes. This chapter will focus
on two topics: 1) what cooccurrence restrictions there are,
and 2) what, if any, reasons (syntactic, semantic,
pragmatic, phonological or arbitrary) can be given for these
restrictions.

As observed at the end of chapter 3, the total number
of morphemes which constitute the various forms of the Haya
verb come to 23. If all of these morphemes were freely
combinable with each other, the total number of forms in the
Haya verbal paradigm would come to (1,334,832). However, as
was apparent in the previous.chapter, the number of actually
occurring forms is a small fraction of the possible forms.
(This ratio of possible to actual forms will be discussed in
the next chapter, section 5.4.)

This observation raises the problem of determining
which of the possible forms are actual, and thereby
determining the cooccurrence restrictions which hold among

the verbal morphemes. Two methods were used to derive the

152

153

generalizations cited in this chapter. First, some of these
generalizations were discovered by introspection by the
author, a first-language speaker of Haya” Others were
discovered by application of the following method: the
morphemes of the Haya verb were numbered sequentially, from
left to right. Then, random sets of these numbers were
selected and verb forms composed using just the morphemes
corresponding to these numbers in the appropriate order.
The resulting forms were then evaluated to uncover
generalizations about cooccurrence restrictions that may
have escaped introspection.

The presentation in the remainder of the chapter is
organized. as follows. First, the verbal morphemes are
divided into three groups, based on their position wdthin
the verb: 1) prestem: preceding the stem; 2) poststem:
following the stem; and 3) stem: stem-internal. Second, the
cooccurrence restrictions for the prestem morphemes are
discussed, then, the cooccurrence restrictions for the
poststem morphemes, then, the cooccurrence restrictions for
the stem morphemes. The discussion of the non-stem
morphemes presupposes that they may cooccur with an
unanalyzed stem, and the discussion of the stem morphemes
presupposes that the occurrence of non-stem morphemes is
irrelevant to the internal structure of the stem. The few
exceptions to these generalizations will be discussed last,

and commented on further in chapter 5.

154

4.2. Cooccurrence Restrictions of Prestem Morphemes

4.2.1. The Subject Marker

The subject marker is required in every verb form of
the Haya verbal paradigm except the indicative imperative
(cf. section 3.3.1.) and the infinitive. The following
examples indicate that it may not appear in the two latter

forms:

(1) * tu-ku -shun -a
Ip-INF-pinch-IND

x ’We to pinch you’

The infinitive morpheme kg in example (1) has the same shape
and occupies the same slot as the object morpheme 3,5,3:
second person singular as discussed earlier. When $2 in (1)
is regarded as the OM second person singular then the form

in (1) repeated here as (2) is well formed.

(2) tu-ku-shun -a
Ip-Zs-pinch-IND

’We pinch you’

Similarly example (3) is ungrammatical as an imperative;

otherwise possible as a present universal habitual (24).

155
(3) * ba-shom-a

2 -read-IND (imperative)

(4) ba-shom-a
2 -read-IND

’they (habitually) study’

Example (3) is well formed only as an imperative without the

subject morpheme (2) as shown in (5)

(5) shom-a
read-IND

’study’

Furthermore example (3) is also well-formed only when used
with the subjunctive morpheme to be discussed in detail in

section (4.3.2.). We however give an example here:

(6) ba-shom-e
2 -read-SUBJ

’let them study’

We now, discuss the cooccurrence restrictions of the

infinitive (INF) with other prestem morphemes.

156
4.2.2. The Infinitive Morpheme
The infinitive morpheme may cooccur only with the

poststem indicative marker and NEG2. It may not cooccur

with the subject marker:

(7) a. ku -zin -a
INF-sing-IND

’to sing’

b. * a -ku -zin -a
Is-INF-sing-IND

’he to sing.’

Example (7.a) indicates the. use' of the INF ‘without the
subject marker and (7.b) shows that the two may not cooccur.
As was the case with the subject marker above in section
4.2.1, the same reason applies here. That is, the marker -
ku- occupies the same slot and has the same shape as the
object marker (OM) second person singular. This might be
regarded as a phonological identity restriction, a term we
are using here to mean the cooccurrence restrictions of two

or more morphemes which have similar phonological values.

The infinitive morpheme may not occur with any tense marker:

(8) a. t e -li—ka -ku -gw -a

IV-9 -RMP-INF-fall-IND

157

......

2p- -INF pierce- -IND

c. * zi-ku -li-gomok -a

IO-INF-NF-become fat-IND

d. * ba-ka -ku -ib -a

2 -RMP-INF-steal-IND

Examples (8.a) and (8.b) indicate that no tense marker
precede the INF morpheme and examples (8.0) and (8.d)
indicate that no tense marker may follow the INF morpheme.
However, in other Bantu languages such as Shona (9.a) and
Kikuyu (9.b) (Harford (Perez) 1986: 68) infinitives have
"overt future markers as illustrated in the following

examples from Shona and Kikuyu:

(9) a. Ku -cha-ziva '
INF-fut-know

’to know still’ (Fortune 1980: 3.60) (Shona)

b. Menya u -gok -a gu -ka negenia.
know 23 fut. come ind. inf. fut. Is obj unbraid
(kikuyu) (Barlow 1951: 127).
’Mind you do not come to upbraid me (i.e., at some

future time).’

158

The infinitive morpheme cannot co-occur with the

subjunctive morpheme as shown below:

(10) a. * ku -gend-e

INF-go -SUBJ

b. * ku -lis -e ente
INF-graze-IND cattle

c. * ku -ly -e omuchele
INF-eat-SUBJ rice

The cooccurrence restriction holding between the INF, and
SUBJ is because of the cooccurrence requirement of the SUBJ
morpheme with the SM (see 4.3.2 for details). We have also
seen that SM and INF do not co-occur. Thus there is a clash
of the occurrence requirements of SUBJ with SM and

cooccurrence restrictions of INF with SM.

The infinitive morpheme k1; cannot co-occur with NEGI
and NEG3. As for NEGl, (as discussed in chapter 3
(3.3.9.1)), it is prefixed to the SM which is at the same
time ruled out by the cooccurrence restrictions between SM

and INF as discussed in (4.2.1):

159
(11) a. * ti -ku -gul-a

NEGl-INF-buy-IND

’not to buy’

b. * ti -ku -fumol -a
NEGl-INF-converse-IND

’not to converse’

c. * ti -ku -lind-a
NEGl-INF-wait-IND

’not to wait’

However, as noted in chapter 3 (3.3.2), when examples (11.a-
c) are regarded as nominal infinitives they are acceptable

as shown in (12).

(12) a. o -ku -lwan -a ku-abo ti -ku-li-ba-gu(l)-z -a
IV-INF-fight-IND 15-their NEGl-lS-NF-Z -buy-CAUS-IND
bi-ntu
8 -thing

’their fighting won’t make them buy things’

b. o -ku -boigok-a
IV—INF-shout -IND
ti -ku-li-ba-fumo(l) -z -a -ge
NEGl-15-NF-2 -converce-CAUS-IND-MNR

’shouting won’t make them converce well’

160

c. o ~ku -lind-a ku-ngi ku-a —mu—laa(l)-z -a
IV-INF-wait-IND 15-much 15—IP-3s-sleep -CAUS-IND
omwilungu
on the way
’his waiting too much has caused him to spend a

night on the way (in the wood).’

Note that consonant l is neutralized in favor of _z in the
causative form.

On the other hand, NEG3 seems as if it co-occurs with
an INF morpheme. However, as we discussed in chapter 3
(3.3.9.3), the morpheme ,jgu; which co-occur with NEG3 is
what we have referred to as an imaginary marker (UNR). Thus

NEG3 and INF do not co-occur as exemplified below:

(13) a. ¥ ka -ku -shom -a 'e -bi—o o -ta -bas-e
NEG3-INF-study-IND IV 2 -REL Zs NEG2 be able SUBJ
* ’do not to study that which you wont be able

to understand’

b. 3 ka .-ku -sig -a a -b-ana -b-onka
NEG3-INF-leave-IND IV-2-child Z-alone

’do not to leave the children alone’

161

c. * ka -ku -imuk —a bwanka muno
NEG3-INF-wake up-IND morning very

’do not to wake up early (in the) morning’

When in examples (13.a-c), the imaginary marker 3,513”; is
substituted for the INF marker, then all the examples are

well formed, as shown in (14.a-c).

(14) a. ka -ku -shom -a ebyo otabase
NEG3-UNR-study IND which you won’t be able

’do not study (things) which you cannot understand’

b. ka -ku -sig -a abana bonka
NEG3-UNR-leave-IND children alone

’do not leave the children alone’

0. ka -ku -imuk -a bwanka muno
NEG3-UNR—wake up-IND morning very

’do not to wake up early in the morning’

Examples (13) and (14) therefore, indicate that in Haya you
cant get a negative infinitive, instead it would ‘be a
negative imperative as in (14). Note that there are
phonological realizations that take place in the surface
realizations of forms in the examples above as discussed in

chapter 3. For example, the vowels i and g are neutralized

162

in favor of the glide 1. Therefore ebig in (14.a) is

...,.... ........

4.2.3. The Tense Morphemes

Tense must co-occur with the SM and IND, except for the
near past tense, the remote past and the future remote past
‘which have the past-stem marker —ile: The morphemes -ile and

IND may not co-occur:

(15) a. ba-lika-gamb—a a -maano.
2 -RFM -tell-IND IV-forbidden

’they will in the far future tell the forbidden’

b. tu-la-mu-tang -ilg ku -zan -a na kato
Ip- 3s refuse' INF-play-IND and Kato

’we have refused him to play with Kato.’

c. * mu-a-sig -ile-a busha

um

2p- leave -IND empty

Example (15.c) is ill formed because the remote past (RP)
has a post stem marker igl__e_ and therefore cannot co-occur
with IND.

Moreover, all tense morphemes co-occur with NEGl and

NEG2 as exemplified partially in (16). However, NEG2 cannot

follow the infinitive morpheme (as shown in 17), it only co-

163

occurs with only one tense: the present universal habitual

(PUH) in a nominal infinitive form as shown in (18):

(16) a. ti -zi-l -ile bunyasi bungi
NEGl 10-eat-RP grass many

’they (e.g. cows) did not eat a lot of grass’

b. ba-ta —li-gend-a ki—lo
2 -NEGZ—NF-go -IND night

’they should not go at night’

(17) a. * ku—ta -kya-hiig-a ente
INF-NEGZ-STL-find-IND cows

’to not still find the cows’

b. * ku-ta -ka -yombek-a enju

INF-NEGZ-RMP-build -IND house

Examples (17.a, b) show the co-occurrence restrictions of
NEGZ with other tenses when. preceded by’ the infinitive
(’still’ tense and remote past). As pointed out above NEGZ
only cooccurs with the present universal habitual (PUH) as

shown below.

(18) a. o -ku -ta —ly -a ku-teb -es -a ente
IV-INF-NEGZ-eat-IND 15-become thin-CAUS-IND cow

’not eating causes cows to be thin’

164

b. o -ku -ta —lim -a ku-lum -is -a enjala
IV-INF-NEGZ-farm-IND 15-bite-CAUS-IND hunger

’not farming causes hunger’

Furthermore, the tense morphemes cannot cooccur with
the imaginary 1kg; except the recent past (19.b).
(19) a. t ka -ku -laa ~jum -a a -ba-alimu
NEG3-UNR-IF -abuse-IND IV 2 -teacher

’do not abuse the teachers’

b. ba-ku -shuntam-ile ba-ly -a
2 -UNR-sit 2 ~eat-IND

’they could sit and eat’

4.2.4. The Subject Relative Morpheme

The subject relative morpheme takes the form of an
initial vowel. The initial vowel (IV) occurs in relative
verbs; which always contain a subject morpheme marker. It
may not cooccur with COND, NEGl, NEG3, or FOC. Otherwise,
it may co-occur with any other prestem, poststem or stem
morphemes. Furthermore, the subject relative morpheme may
cooccur with any tense. The following are some examples to

illustrate these cooccurrence restrictions:

165

(20) a. * e-ki-ka-many-a
IV-7-COND-know-IND

’which when know'

b. * e-li-ti -shash-a
IV 5—NEG1-hurt -IND

’which not hurt’

c. i a -ba-ka -shom-a
IV-2 -NEG3—read-IND

’who not read’

d. * e -bi-ni -lum -a

IV-8 -FOC-bite-IND

Examples (20.a-d) show the co-occurrence restrictions of IV
with COND (20.a), NEGl (20.b), NEG3 (20.c) and FOC (20.d).
We would attribute the ill-formed form of (20.a) as being a
clash of two morpheme the class 12 (diminutive class) and
the condition morpheme both of which have the same shape and
may be interpreted as occurring in the same slot unless
other morphemes entervene. The ill-formed ferm.cfif (20.b)
may be explained as due to the position of NEGl, which only
appears verb initially, and only negates independent
clauses. Dependent clauses like the relative are negated by

NEGZ as shown in (21.b) below:

(21) a. e -li-shash-a
IV 5 -hurt

’which (e.g. eye) hurts’

b. e -li-ta -shash-a
IV 5 NEGZ-hurt IND

’which (e.g. eye) do not hurt’

Example (20.0) is ill-formed because NEG3 only cooccurs with

the imaginary _-ku- (UNR), or the INF ku- in nominal forms

 

initially. Similarly (21.d) is ill-formed because FOC only
appears verb initially and not in any other position.
Therefore it cannot cooccur with the IV. (relative) form.
This may hypothetically be called ’positional clash.’
Example (20.a), however, may be interpreted as acceptable
only when the marker :kg; (COND) is regarded as that of noun
class 12, sometimes called the diminutive class (e.g. Doke
and Mofokeng for the Sotho language, Byarushengo 1977 for
Haya). COND and class 12 share the marker which has the

same shape as exemplified below:

(22) a -ba-ka-many-a
IV-2 -12-know-IND

’who knows it (e.g. a small dog).’

Although we have attributed the ill formed forms to the

position they occupy in the verbal paradigm (which is

167

plausible), the major reason is the function these morphemes
perform in the verb. As discussed in detail in the next
section, IV. (subject relative morpheme), NEGI, NEG3 and FOC
introduce dependent clauses, such that the cooccurrence of
any two of these morphemes are restricted because of what we
might call "identical functional clash." Furthermore, IV.
is followed by SM, but not vice versa, a point we have
discussed at length when discussing the linear order in

chapter 3.

(23) a. e -zi-li-gul-w -a
IV 10-NF-buy-Pass-IND

’which (e.g. cows) will be bought’

b. * zi-e -li—gul-w -a

10-IV-NF-buy-PASS-IND

4.2.5. The Conditional Morpheme
The COND morpheme may not co—occur with IV, NEGl, NEG3
or FOC; it may co-occur with any tense except the remote

past (RMP) and the remote future (RMF).

(24) a. 3 a -ba-ka —li-shutul-a
IV 2 COND NF carry IND

’who if will carry’

b.
(25) a.
b.
(26) a.
b.
(27) a.

N'

96

168

ka -a -ba-li—shutul-a

COND-IV-Z -NF-carry -IND

ka -ti -ba—shek —a
COND-NEGl-Z -laugh-IND

’if not they laugh’

ti -ka -ba-shek -a

NEGl-COND-Z -laugh-IND

ka —ni -ba-many-a
COND—FOC-Z -know—IND

’if they’

ni -ba—ka -many-a
FOC-2 -COND know-IND

’they (indeed) if know’

ka -ka -ku -zin -a
COND-NEG3-UNR-sing-IND

’if not sing’

ka -ku -ka -zin -a
NEGS-UNR-COND-sing-IND

’not if sing’

169

In examples (24-27) various cooccurrence restrictions of
COND with other morphemes are shown in pairs following and
preceding the COND. The cooccurrence restrictions of COND
with IV (example 24), NEGl (example 25) and FOC (example 26)
may be explained again as ’similar functional clash.’ All
the three morphemes introduce the dependent clause and
therefore they cannot cooccur. Similarly, although NEGB
does not occur in the same position as the other morphemes
(IV, NEG, and FOC), its function is similar to them, as

exemplified below:

(28) ka -ku -ly -a o -ta -nab -ile mi-kono
NEG3-UNR-eat-IND Zs-NEGZ-wash-RP 4 -hand

’do not eat without washing hands (your)’

Moreover, COND does not cooccur with the remote past

(RMP) and remote future (RMF) tenses.

(29) * ka -ba-ka -it -a

COND-2 -RMP-kill-IND

Example (29) is ill-formed because COND does not co-occur
with (RMP) tense. The RMP 35;; has the same shape as the OM
(kg) which agrees with class 12. Thus example (29) may be
regarded as well formed when the kg, of the RMP tense is

substituted for the ka of OM (class 12) as illustrated:

tut-m

170
(30) ka-ba-ka-it-a

COND-Z-IZ-kill-IND

’when they kill it (e.g. a small goat).’

The same reasons apply in the co-occurrence
restrictions of COND and the remote future (RMF) where the
marker -lika— clashes with the marker _;l_i__; of near past
tense and the marker :kg; clashes with the 3kg; of the OM

(class 12) (see also chapter 5 (5.3.3.2)).

(31) * ka -ba-lika-gul-a
COND-2 -RMF -buy-IND

’? if they (in the remote future) buy’

However, example (31) is regarded as well formed when the
marker -lika- is split into two markers: -li- with the role
of the near future tense marker (NF) and :5a; with the role
of. the OM (class 12), as discussed in chapter 3

(3.3.5.3.3.).

(32) ka -ba-li-ka-gul-a
COND 2 -NF-12-buy-IND

’when they will buy it’

4.2.6. The Focus Morpheme
FOC occurs only in the present universal habitual (PUH)

and the "still" tense (-kya—). It may only co-occur with

171
the SM, the IND and the SUBJ and the ’still’ (-kya-) tense

morphemes as exemplified:

(33) a. ni -ba-kya—lil-a
FOC-Z -STL-cry-IND

’they are still crying’

b. ni -ba-lil-e
FOC-2 -cry-SUBJ

’let them cry’

(34) a. * ni -mu-laa-lil-a

FOC-2p-IF -cry-IND

b. * ni -ba—a-lil-ag-a

FOC-2- -cry- -IND

Examples (34.a) and (34.b) indicate that the FOC morpheme
marker may not cooccur with the immediate future morpheme
nor with the immediate past habitual morpheme. . These are
examples of certain tenses which do not cooccur with FOC;
otherwise all except the (PUH) and STL do not cooccur. The
FOC morpheme also does not co-occur with any NEG morpheme,
IV, or COND. We will only show NEG morpheme examples here.

Other examples have been covered in the previous sections.

172

(35) a. * ni -tu-ta -jum -a

FOC-Ip-NEGZ-abuse-IND

b. * ni -ka -ku -log —a

FOC-NEGB-UNR-witch-IND

c. * ti -ni -ba-hend -a

NEGl-FOC-2 -break-IND

Again (35.b) can only be interpreted as well formed when the
NEG3 marker is regarded as the SM (class 12) and the -ku- of

UNR is realized as the OM.

(36) ni -ka-ku-log -a
FOC-12-2s-witch-IND
’it is bewitching you’

Example (35.a) whose affirmative form is:

(37) ni -tu-jum -a
FOC-ls-abuse-IND

’we are abusing’

can only be negated by deleting the FOC marker and prefixing

the NEGl to the SM as indicated below:

173
(38) ti--tu—li-ku-jum-a

NEGl-Ip-NF—INF-abuse-IND

’we are not abusing’

This restriction is also observed in Kikuyu
On the other hand Tharaka permits F00 and

(Perez (Harford) 1987).

4.2.7. The Negative Morphemes
NEGl may co-occur with any tense

,therefore, (cf. discussion of tense markers

(Bergvall 1987).

NEG to co-occur

of the verb,

for more detail)

since it only negates main verbs (cf. chapter 3), it may not

cooccur ‘with IV, COND, or FOC as we mentioned in the

previous sections.

(39) a. e -bi-kol-w -a
IVé8 -do -PASS-IND

’which (things) are done’

b. ti -e -bi-kol-w -a

NEGl-IV-8 -do -PASS-IND

(40) a. ka -ba-laa-kul -a
COND-2 -IF -grow-IND

’when they grow’

174
b. * ti -ka -ba-laa-kul -a

NEGl-COND-Z -IF -grow-IND

’not when they grow’

(41) a. ni -tu—shom -a
FOC-Ip-study-IND

’we are studying’

b. * ti-ni-tu-shom-a
NEGl-FOC-Ip—study-IND

’not we are studying’

In examples (39) to (41) the second pair of each example is
ill-formed» illustrating the cooccurrence):restrictions of
NEGl with IV, COND and FOC, respectively. These morphemes
introduce dependent clauses, so that they cannbt co-occur
with NEGl, which, as stated above, negates only main verbs
(independent clauses). ,

NEG2 co-occurs with the SM, all tenses except remote
future and the infinitive morpheme. When NEG2 follows the
infinitive morpheme it only' cooccur with the present
universal habitual (PUH) as discussed above in section
(4.1.3). NEGZ may cooccur with (IV), and (COND) only when

following them.

175
(42) ba-ta —li-galuk -a hanu lundi

2 -NEGZ IF-return-IND here again

’they should not return here again’

(43) a. ku -ta -mu-jun -a kiki
INF NEG2 3s-help-IND why

’what is restricting you from helping him’

b. * ku —ta -lika-mu-jun —a kiki

INF-NEGZ RMF 3s help—IND why

NEG3 only co~occurs with the ’imaginary’ morpheme

marker (see 4.2.2 above).

(44) ka -ku —it -a amani gawe busha
NEG3-UNR-’kill’-IND energy yours for nothing
’don’t waste your time’ (lit. don’t kill your energy

for nothing).

Finally, the NEG morphemes may not co-occur with each other

as exemplified in (45.a-c).

(45) a. 3 ti -ba-ta -1i -shom-a

NEGI-SM-NEGZ-TNS-read-IND

b. * ba-ta -ka -ku -songok -a

SM-NEGZ-NEG3-INF-descend-IND

176

c. X ti —ya-ta -ka -ku -yet -a

NEGl SM-NEGZ-NEG3-INF-call-IND

The three examples in (45.a-c) indicate that Haya does not
allow "double" or "triple" negatives in a verbal
construction. This is semantically plausible because Haya
verbs usually carry one assertion and this on; assertion is
negated by gng negative morpheme.

Thus far, the co-occurrence restrictions of the prestem
morphemes have been discussed. These restrictions are .

summarized in a table form as in table 12 at the end of the

chapter on (page 214).

4.3. Co—occurrence Restrictions of Poststem Morphemes:

4.3.1. The Indicative Morpheme

The only significant cooccurrence restrictions in this
group hold for IND and SUBJ, which may not cooccur with each
other. IND must occur with all tenses, in infinitives and

certain imperatives.

(46) a. yanguh-a
’hurry-IND’

’hurry’

177
b. * yanguh-a-e

hurry-IND-SUBJ

c. * yanguh—e-a

hurry-SUBJ-IND

As illustrated in examples (46.b-c), the IND morpheme cannot
cooccur with the SUBJ mor‘pheme whether the SUBJ morpheme
follows (46.b) or precedes the IND (46.c0. The indicative
morpheme is obligatory, used when no other is called for.
However, the subjunctive and .:..i.....1._.<i3. suffix may be used as

described in the next sections.

04.3.2. The subjunctive Morpheme

SUBJ must occur with SM and cannot occur with any tense
marker, except the present universal habitual (PUH). When
the SUBJ occurs in an imperative form with an OM, the SM may

be optionally omitted.

(47) a. 0 Qba-tang -e
23-2 -forbid-SUBJ

’You should forbid them’

b. ba-tang -e
2 forbid-SUBJ

’forbid them’

178
(48) a. * a -li-ba-tang -e

33-NF-2 -forbid-SUBJ

b. * a -kya-ba-tang -e

3s-STL-2 -forbid-SUBJ

Examples (47 .a) and ( 47 .b) show the optionality of (SM).
Examples (48.a) and (48.b) show the cooccurrence
restrictions of the near future tense and ’still’ tense, the

illustration of the remaining tenses being omitted.

4.3.3. The Tense Suffix :ile

The tense suffix -ile occurs in three tenses: the near
past (NPS), the recent past (RP), and the far remote past
(FRP). As was pointed out in chapter 3 (77), the near past
and the far remote past tenses appear as a combination of a
prefixal tense morpheme and a suffixal one. We are,
however, concerned with the suffix marker -ile. The suffix

.11; must cooccur with the SM, and may cooccur with COND.

(49) a. ba-a-zin-ile
2- -sing- (NPS)

’they have sung’

3 -fall-RP

’it (e.g. tree) fall’

179

c. e-la-sw 4.12
4- fall (FRP)

’they (e.g. trees) have fell (a long time ago)’

d. ka -a-zin -ile ba-mu-shua -ile
COND-l-sing- 2 -3s-give a tip-

’when he sung, did they give him a tip’

e. ka -ba—mu-kwas -ile ba-ka -mu-fung-a (kabamukwete)
COND-2 -3S-catch RP 2 -RMP-3s-jail-IND

’when they caught him did they put him into custody’

Examples (49.a-c) show the cooccurrence requirement of -ile
with the subject morpheme marker. Those in (49.d-e) show
the cooccurrence of :ile with the conditional morpheme ka-
while examples (50.a-b) show the cooccurrence of :ilg

without the conditional morpheme.

(50) a. a -zin -ile, ba-mu-shua-ile
Is-sing-RP 2 -3s—give a tip-RP

’he sung, did they give him a tip’

b. ba-mu-kwas -ile, ba-ka -mu-fung-a (bamukwete)
2 -3s-catch-RP 2 -RMP-33-jail-IND

’they caught him, did they put him into custOdy’

180

4.3.4. The Locative, Manner, Intensifier and
Question Morphemes.

There are no cooccurrence restrictions on LOC/MNR/
INT/Q (except those mentioned in the following two
sections): they may all cooccur with each other and with any

verb form including the infinitives and subjunctives:

(51) a. ba-ta -li-ij -a -ho -1i
2 -NEG3 IF-come-IND-LOC-Q

’when shouldn’t they come here’

b. ti -ba-mu-chumb-il -il -al —ha -ge -kwo-ki
NEGl-Z -3s-cook -APL-APL-IND-LOC-MNR-INT-Q

’what don’t they indeed cook well for him here?’

0. ku -ta -ba-mu-leb -el -a -ha -ge -kwo
INF-NEG2-2 -3S-’look’ APL-IND-LOC-MNR-INT

’not indeed well taking care of him for them here’

d. ba-shom -e —ge -kwo-ki
2 -study-SUBJ-MNR-INT-Q

’what should they indeed study well’

4.3.4.1. Additional Restrictions on the Question Morphemes
Q may not appear in a sentence which does not contain a
verb with an argument position which Q corresponds to (see

also chapter 5 (5.3.3.2))

181
(52) a. mu-a -ly -a -ki

2p-Ip-eat-IND-Q

’what did you eat’

b. * mu-a —f -a —ki
2p-IP-die-IND-Q

’what did you die’

Example (52.b) is ill-formed because the verb -f- ’die’ has
no argument (intransitive verb), thus the question is

blocked by non-correspondence of an argument.

4.3.4.2. Additional Restrictions on the Manner Morpheme.

MNR may not be attached to a verb that does not denote

an action (or is not transitive).

(53) a. ba-laa-andik-a -ge
2 -IF -write-IND-MNR

’they will write well’

b. * mu-a -f v—a -ge
2p-IP-die-IND-MNR

’did you(pl) die well’

(53.b) is ill-formed because the verb -f- is intransitive.

Compare (53.b) with (53.c) which has a transitive verb; and

therefore well-formed.

182

c. mu-a -chumb-a -ge
2p—IP-cook -IND-MNR

’you have cooked well’
4.4. The Cooccurrence Restrictions of Stem Morphemes

4.4.1. The Object Marker

The linear ordering of the OMs has been fully discussed
in. chapter’ 3 (3.4.1). In this section we discuss the
cooccurrence restrictions of the OM with the other stem
morphemes. As mentioned in the introductory section at the
beginning of this chapter, we will ignore the non-stem
morphemes here.

The OM cannot cooccur with the reflexive morpheme
(RFM). It may cooccur with verbal extensions, except the

passive, stative and reciprocal morphemes.

(54) a. ba-ka -mu-tem-a
2 -RMP-3s-cut-IND

’they cut him/her’

b. * ba-ka -mu-e -tem-a
2 -RMP-3s-RFM-cut-IND

’they cut him themselves’

183
Examples (54.b) indicate that OM cannot cooccur with RFM.

The cooccurrence restriction of the OM with the (RFM) is
because of the internal contradiction resulting from the
dual specification of the object. Whereas RFM claims an
identical reference of the agent, that is, that the actor
and the patient are the same individual, OM also refers to
the patient. Thus there is a double reference to the same
patient which is not permitted. In other words RFM and OM
have the same function. The presence of one makes the other
redundant.

There is also a cooccurrence restriction between the OM

and the passive morpheme, exemplified as follows:

(55) a. mu -holol—a
OM(3s)-reply-IND

’(I) reply him’

b. X mu -holol-w -a

OM(3s) reply-PASS-IND

OM and PASS work in an opposite way. Whereas the OM implies
an object, the passive removes it. This means that when an
object is present in the verb (that is the OM) then the
passive morpheme -w- cannot be attached to the verb-stem.
If the agent is absent in the verb then the passive morpheme

-w- applies as shown in (56).

184
(56) holol-w -a

reply-PASS-IND

Moreover, OM cannot cooccur with the stative morpheme,
again because of the ’internal contradiction.’ Whereas the
OM morpheme represents an object argument of the verb the ST
morpheme like the passive, removes an object. That is it
does not imply agentivity (Perez 1986:11). Thus, the ill-
formed forms of the (b) examples in (57) and (58) are due to
the cooccurrence of an object (OM) and an extension
entailing objectlessness (ST) which are contradictory and

therefore do not cooccur.

(57) a. bi-li -ik-a
8 -eat-ST-IND

’they (the bananas) are edible’

b. X bi-ba-li -ik-a

8 -2 -eat-ST-IND

(58) a. e—kom-ek-a
4-tie-ST-IND

’they (loads) are tiable’

b. X e-ku-kom-ek-a

4-Zs-tie-ST-IND

185

Finally, the OM cannot cooccur with the reciprocal morpheme;

they are mutually exclusive.

(59) a. X ba-laa-mu-kom-angan-a
2 -IF -3s-tie-RCP -IND

’they will tie him one another’

b. X o -lika-ba-tel-angan-a
Zs-RMF -2 -hit-RCP -IND

’you will in the far future hit them one another’

Examples (59.a) and (59.b) show that the OM cannot cooccur
with the RCP because the RCP morpheme refers to reciprocity
of actions of the subject and NP object. Thus, examples

(60.a-b) are well formed without an OM:

(60) a. ba—laa-kom-angan-a
2 IF -tie-RCP -IND

’they will tie each other’

b. o -lika-tel-angan-a
23—RMF -hit-RCP -IND

’you will in the far future hit one another’

4.4.2. The Reflexive Marker
The Reflexive morpheme cooccurs with all the stem

morphemes except the (”1 (see 4.3.1), stative, passive and

186

reciprocal. The discussion of these restrictions follow.
However, we will not discuss the cooccurrence restriction of
RFM and OM a task we have done in the previous section
(4.3.1)

The cooccurrence restrictions of RFM with the stative
morpheme is again due to the function of RFM which indicates
the coreferentiality of the actor or director (agent) as a
subject and the patient as an object, and the stative
morpheme which indicates the absence of an object. Thus
reference to objecthood shown by RFM and non-reference to
objecthood shown by stative causes an internal contradiction

which prohibits the cooccurrence of the two.

(61) a. y-e —shal-a
I RFM cut IND

’he/she cut himself/herself’

b. X y-e -shal-ik-a

I RFM cut ST-IND

RFM and Passive do not cooccur for the same reason.
RFM indicates coreference of the agent subject and NP object
while at the same time passive extension entails the absence
of an object. This contradiction does not cooccur in Haya

as exemplified below:

187
(62) a. a-li—e ~tel-a

I-NF-RFM-hit-IND

’he will hit himself’

b. X a-li-e -tel-w -a
I-NF-RFM-hit-PASS-IND

’he will hit himself by (x)’

Moreover RFM does not cooccur with the RCP morpheme.
Although these two morphemes occupy different slots of the
verb, that is, RFM occupies the slot of an OM, whereas RCM
is formed by suffixing the verb with the ROM morpheme -
angan-, they both involve some identities between the
subject and the object. Thus the two morphemes cannot

cooccur because of the incompatible reference to the object.

(63) a. ba-lika-e -chumit-a
2 -RMF -RFM-spear -IND

’they will spear themselves’

b. X ba-lika-e -chumit-angan-a
2 -RMF -RFM-spear -RCP -IND

’they will spear themselves one another.’

(63.b) indicates that in the presence of RFM, RCP is

redundant, and vise versa.

188

4.4.3. Verbal Extensions

In the present section we present the various
cooccurrence restrictions holding between verbal extensions.
As we did in the previous sections we will discuss the
morphemes that must cooccur, those that may cooccur, and
those that cannot cooccur. We first begin with the

causative.

4.4.3.1. The Causative
The causative morpheme may cooccur with the OM, RFM,
passive, reciprocal and repetitive morphemes exemplified by

one example for each:

(64) a. a-mu-shun -is -a e -bi-aly -a
I-3s-pinch-CAUS-IND IV-2 -finger—IND

’he pinches him with fingers’

b. y-e -tem-es -a e -ki-so
I-RFM-cut-CAUS-IND IV-7 -knife

’he cut himself with a knife’

0. baag -is -angan-a e -nte
slaughter-CAUS-RCP -IND IV cow

’cause to slaughter a cow for one another’

189

shun -is -ibw -a e- bi-oma
pinch-CAUS-PASS-IND IV 2 iron

’be pinched by (use of) iron’

kom -es -iliz-a
tie CAUS-RPT IND

’cause to tie repeatedly’

Examples (64.c-d) indicate that the CAUS may be followed by

the RCP,

PASS and RPT. The CAUS morpheme seems as if it

cooccur with the APL morpheme only when preceding the

(CAUS), but not with the stative or repetitive morphemes.
(65) a. ?sindik-il -iz -a
send -APL-CAUS-IND
’cause to send for someone’
b. X sindik-ik-is -a
send -ST-CAUS-IND
c. X sindik-iliz-is-a
send-RPT-CAUS-IND
However, a close observation of example (65.a)

indicates that the applicative and causative form together

is as a matter of fact the repetetive form which is split

into two as shown in (65.d) see also 4.4.3.5.

190

d. sindik-iliz-a
send -RPT -IND

’send frequently’

An attempt will be made to discuss the reasons for the
cooccurrence restrictions of the morphemes as we discuss

each of them individually.

4.4.3.2. The Applied
The applied morpheme usually cooccurs with the (W1 or
RFM. It may also cooccur with all other extensions except,

the CAUS and the stative morphemes, exemplified as follows:

(66) a. mu-shuntam-il -a
3s-sit -APL-IND

’(I) sat on him (for his benefit?)’

b. ba-e -baag -il —a
2 -RFM-slaughter-APL-IND

’they slaughter for themselves’

Note that in (66.b) RFM must cooccur with the subject

morpheme SM(Z), a point we will turn to later.

191
(67) a. shuntam-il -w -a

sit -APL-PASS-IND

’be sat on’

b. X shuntam-w -il -a

Sit -PASS-APL-IND

(68) a. lamul-il -angan-a
judge-APL-RCP -IND

’judge the case for one another’

b. lamul-angan-il -a (owo omukama)
judge-RCP -APL-IND at chief’s place

’judge one another (at the chief’s place)’

(69) a. kol -el -elez-a
work-APL-RPT -IND

’work frequently for someone’

b. X kol-elez-el-a

work-RPT-APL-IND

In the above paired examples, it is noted that the applied
may cooccur with the passive and repetitive only in the
proper sequence. That is, the applied may not be preceded
by the passive or repetitive. On the other hand APL may be

both followed and preceded by the RCP, although the forms

192

that result have a slightly different meaning as shown in
example (68.a) and (68.b). We will discuss the cooccurrence
restrictions of PASS and APL and RPT with APL in sections
(4.4.3.3) and (4.4.3.5) respectively. Example (70-71) show
the cooccurrence restrictions of APL with CAUS and ST

morphemes.

(70) a. X many-il -is -a

know-APL-CAUS-IND

b. X many-is -il -a

know-CAUS-APL-IND

(71) a. X shuntam-il -ik-a

sit -APL-ST-IND

b. X shuntam-ik-il -a

sit -ST-APL-IND

Examples (70) and (71) indicate that not only is there a
cooccurrence restriction holdings between APL followed by
CAUS or ST, but also between APL preceded by these
morphemes. APL and ST cannot cooccur, again because the
applied morpheme requires an agent whereas the stative does
not require an agent. Thus, the cooccurrence of the two
morphemes APL and ST is due to the internal contradiction of

their argument requirements.

193
Note also that the applied morpheme may be reduplicated

as exemplified below:

(72) a. shom-a

study-IND

b. shom-el-a
study-APL-IND

’study for/at’

c. shom—el-el-a
study APL-APL-IND
study for x for y

e.g. ’read for Juma for his son’

4.4.3.3. The Passive

The passive morpheme may not cooccur with any other
stem morphemes, that is, the passive followed by other
extensions. The following examples illustrate the
cooccurrence restrictions holding between the passive

morpheme and the other extensions:

(73) a. X lim -ik-w -a

.farm-ST-PASS-IND

b. X tel-angan-w -a

hit-RCP -PASS-IND

194

c. X yombek-elez-w -a

build -RPT -PASS-IND

However, APL and CAUS, where the meaning allows may cooccur

with the PASS following them:

(74) a. chumb-is -ibw -a
cook -CAUS-PASS-IND

’be caused to cook by (someone)’

b. kwat -il -w -a
catch-APL-PASS-IND

’be caught at a certain place’

4.4.3.4. The Reciprocal

The reciprocal morpheme cooccurs with all the
extensional morphemes except the passive and stative
morphemes. The following examples illustrate the

cooccurrence possibilities of some of these extensions:

(75) a. hiig -angan-is -a
search-RCP -CAUS-IND

’search for one another’

195

b. hiig —is —angan-a
search-CAUS-RCP -IND

’cause to search for one another another’

(76) a. lamul-angan—il -a
judge-RCP -APL-IND

’judge one another (for one another’s benefit)’

b. lamul-il -angan-a
judge-APL-RCP -IND

’judge for one another’

(77) a. tel-angan-iliz-a
hit-RCP -RPT -IND

’hit one another frequently/repeatedly’

b. tel-elez-angan-a
hit-RPT -RCP -IND

’hit repeatedly for one another’

The paired examples in (75-77) show the possibilities of RCP
with CAUS, APL and RPT following RCP in examples (a) and
preceding the RCP in examples (b). The following examples

show the cooccurrence possibilities of RCP with OM and RFM:

196

(73) a. b-e -kom-angan-a
2-RFM-tie-RCP -IND

’they tie one another’

b. tu-e -hanuul-angan-a
Ip—RFM-advise RCP -IND

’we advise one another’

However as mentioned above, the RCP does not cooccur with

the passive and stative morphemes, as shown below:

(80) a. X tel-angan-w -a

hit-RCP -PASS-IND

b. X tel-w -angan-a

hit-PASS-RCP ~IND

(81) a. X shal-angan-ik-a

cut -RCP -ST-IND

b. X shal-ik-angan-a

cut -ST-RCP -IND

Examples (81.a) and (81.b) indicate the cooccurrence
restrictions holding between the RCP morphemes and the

stative (ST) morpheme, both preceding and following RCP. We

197

discuss the reasons for this cooccurrence restriction in

section (4.4.3.6).

4.4.3.5. The Repetitive

The repetitive morpheme cooccurs with all the stem
morphemes except, again, the passive and stative morphemes.
We only show these two impossible combinations with tensed
verbs for clarity. These combinations will be shown in
pairs with the passive and stative morphemes following and

preceding the RPT morpheme:

(82) a. X a-li-chumb-iliz-w -a
I-NF-cook -RPT -PASS-IND

’he will repeatedly cook by (someone)

b. X ba-laa-chumb—w -iliz-a

2 -IF -cook -PASS-RPT -IND

(83) a. X tu-lika-shom-elez-ek-a

Ip-RMF-read-RPT-ST-IND'

b. X mu—ka-shom-ek-elez-a

2p-RMP-read-ST-RPT-IND
4.4.3.6. The Stative

The stative morpheme (ST) does not cooccur with any

other stem morpheme as shown in the examples below. The

1 /_ '-

198

examples also indicate that the reverse order of other
extensions with the stative is not permissible either. .At

this point, we cannot account for these restrictions.

(84) a. X a-lika-many-ik—w -a

I-RMF -know-ST-PASS-IND

b. X a-lika-many-w -ik-a

I-RMF -know-PASS-ST-IND

(85) a. X mu-ka -jum -ik-angan-a

2p-RMP-abuse-ST-RCP -IND

-b. X mu-ka -jum -angan-ik-a

2p-RMP-abuse-RCP -ST-IND

(86) a. X ba-kya-shon-ek-elez-a

2 STL-sew -ST-RPT -IND

b. X ba-kya-shon-elez-ek-a

Z STL-sew -RPT -ST-IND

Different reasons account for the cooccurrence restrictions
of ST with other extensions as discussed in the previous
sections. For example, passive and stative don’t cooccur
because they perform the same function, that is, removing

the agentivity.

199

The cooccurrence restrictions of the stem morphemes
discussed thus for are summarized in Table 13 at the end of

the chapter (212).

4.5. Cooccurrence Restrictions between Stem and
Non-stem Morphemes

There are certain cooccurrence restrictions between the
stem morphemes and non-stem morphemes which we will now

discuss, one at a time.

4.5.1. The Subjunctive Marker
The subjunctive must cooccur with either SM or OM. But
when it occurs with OM, there is always the option of an SM

as examples (87.b) and (88.b) show.

(87) a. ba-mu-jun —e
2 3s help-SUBJ

’they should help him’

b. mu-jun -e
38 help SUBJ

’help him’

(88) a. tu-ba-kom-e
Ip-2 -tie-SUBJ

’we should tie them’

200

b. ba-kom-e
2 tie SUBJ

’tie them’

4.5.2. The Subject Marker

The cooccurrence of RCP with RFM requires the
cooccurrence of SM. This is because RCP essentially
requires the coreferentiality of SM and the NP.object, as

seen in the following examples:

(89) a. X ki-kol-angan-a

7 -do -RCP -IND

b. ba-ki-kol-angan-a
2 -7 -do -RCP -IND

’they do it for one another’

(90) a. X gi-guz -angan-a

10-sell-RCP -IND
b. tu-gi-guz —angan-a
Ip-9 -sell-RCP -IND

’we sell it (e.g. cow) to one another’

Note that (a) examples do not contain SMs.

201
4.5.3. The Question Marker

The question marker must cooccur with a transitive
verb; or may represent the demoted subject of a passive

verb. Further discussion will be made in the next section

on verb types.

(91) a. ba-a -gul-a -ki
2 -IP-buy-IND-Q

’what did they buy’

b. X ba—a -gob -a -ki
2 -IP-arrive-IND Q

’what did they arrive’

c. mu-a ely -a -li
2p-IP-eat-IND-Q

’when did you eat’

(91.b) is ill formed because the verb :ggb; ’arrive’ is an

intransitive verb.

4.5.4. The Manner Marker

The manner morpheme marker must cooccur with an action

verb as shown:

202
(92) a. lim -a -ge

farm-IND-MNR

’farm well’

be-IND-MNR

’be well/good’

Example (92.b) is ill-formed because the verb b; ’be’ is not
an action verb. Compare (92.b) and (92.a). The (a) form is

well formed because the verb lim- ’farm’ is an action verb.

4.6. Cooccurrence Restrictions and Verb Types

Certain verb stem types resist the suffixation of stem

morphemes, that is, there are verb stem cooccurrence
restrictions with certain stem morphemes. This is a
discussion we now turn to. We will divide Haya verbs into
three groups (see also Khamisi 1983 for Swahili):
intransitive, fossilized intransitive, and transitive
verbs.2 Our rational in dividing these verbs into three

categories is not based only on morphology but also on their
syntactic behavior, the details of which are beyond the
scope of this study.

Gardner (1985:89), for example, has used feature
arguments to categorize verbs; he determined that verbs are
classified "on the basis of the number of the arguments and

of the semantic category to which these arguments belong."

203

For English, for example, Gardner gives the following

sentences with their arguments:

(93) a. with three arguments: The tggghgr teaches grammar to

the mile-
b. with two arguments: The tggghgr teaches grgmmar

c. with two arguments: The teggher teaches the pupils

d. with one argument: The tggchgr teaches.

In. the examples above it. is safe to say that the
sentence subject and direct and indirect objects are

regarded as arguments. Thus

(93) a. contains th§_t§gghgr, gggmmar, and pupils as three
arguments
b. The teacher and grammar, as two arguments
c. The teacher and pupil, as two arguments

d. The teacher, as one argument.

4.6.1. Intransitive Verbs (IVb)

There are cooccurrence restrictions which affect the'
suffixation of the PASS, RCP and RPT morphemes to
intransitive verbs, while the CAUS, APL 'morphemes may

cooccur with IVb as indicated below:

204
(94) a. -f -a

die-IND

’die’

b. -f -is -a
die-CAUS-IND

’cause to die’

c. -f -el -a
die-APL-IND

’die for/at’

die-APL-IND

e. X -f -angan-a

die-RCP -IND

f. X -f -elez-a

die-RPT -IND

Again, as discussed in the previous sections, the
cooccurrence restriction holding between IVb :1; and the
passive morpheme is because intransitive verbs cannot be
passivized due to their lack of any objects. Moreover,
intransitive verbs do not cooccur with the RCP as in (94.e)

because the RCP morpheme implies an object, whereas IVb lack

205

objects. The same reasons apply to the
restrictions holding between IVb and RPT.

cooccur with transitive verbs.

4.6.2. Fossilized Intransitive Verbs

cooccurrence

Repetitive

Fossilized intransitive verbs refer to verbs "which no

longer maintain independent productive verb roots" (Khamisi

1983:103 for Swahili). That is, the root is no longer used.

Fossilized intransitive verbs are illustrated below:

(95) Free Base
8. 1y "8
eat-IND

’eat’

eat-ST-IND

’edible’

(96) Fossilized Base
a. X kond-a

- IND

b. kond-ek-a
ST IND

’bend’

206

Whereas /ly-a/ ’eat’ in (95.a) has a meaning and a
separate entry in the Haya dictionary (Bona-Baisi 1957),
/kond-/ does not have a meaning of its own and does not have
an entry in the above dictionary. It can only be used with
'the ST suffix as shown above. These fossilized intransitive
verbs not only cooccur with ST but also other stem morphemes

such as the applied morpheme:

(97) a. X hond-a

- IND

b. hond-el-a
APL

follow IND

Let us now test the fossilized intransitive verbs with the

stem morphemes:

(98) a. kondek-es -a
‘bend -CAUS-IND

’cause to bend’

b. kondek-el -a
bend -APL-IND

’bend for (someone)’

207

c. kondek-w -a
bend —PASS-IND

’be bend by (someone)’

d. ?kondek-angan-a3
bend -RCP -IND

’bend one another’

e. X yombek-ek-a

build -ST-IND

f. yombek-elez-a
build -RPT -IND

’build repeatedly’

The examples in (98) show that the fossilized intransitive
verbs do .not co-occur with the stative morpheme. ‘This
pattern suggests that even though the form :ek; is not used
statively when attached to the fossilized IVb -kond-, a
productively attached stative morphemes is interpreted as
redundant. A similar observation may be made for other
fossilized intransitive verbs such as example (99.b) which
is in the causative form and therefore does not cooccur with

the CAUS morpheme.

(99) a. lam-y -a

- caus-IND

208

b. X lamy -is -a
salute-CAUS-IND

’salute/greet’

4.6.3. Transitive Verbs.

There are also certain transitive verbs that do not

cooccur with a number of stem morphemes. The verbs chumb
’cook,’ -h- ’give,’ and gul- ’buy’ are examples of. such
verbs:

(100) a. chumb-a
cook -IND

’cook’

b. X chumb-angan-a
cook -RCP -IND

’cook one another’

(101) a. h -a

give—IND

b. X h -w -a

give-PASS-IND

209
(102) a. gul-a

buy-IND

b. X gul-is-a

We would like to attribute the ill-formedness of examples
(100.b), (101.b) and (102.b) to the following reasons. a)
As was clearly explained in chapter 3 (section 3.5.4) the
ill-formedness of (100.b) is basically semantic/pragmatic.
No one cooks another in this society and also it is
impractical to be cooked. b) Example (102.b) is ill-formed

because Haya, like Swahili for example, has two different

lexical items for ’buy’ and ’cause to buy, ’ ggg; ’sell , ’
which blocks the occurrence of Xgul;i§;. We are unable to
account for the ungrammaticality of (b) beyond the

possibility that it presents an accidental lexicography, or

that the form itself is phonologically impossible.

4.7. Summary and Conclusion

This chapter has discussed a number of factors which
determine the co-occurrence restrictions of the Haya verbal
morphemes. These factors are (1) phonological; (2)
semantic; ( 3) syntactic; and (4 ) pragmatic factors .
Phonologically we have observed that some morphemes do not
cooccur because they are phonologically identical. One of
the examples which we came across frequently was the kg;

which waspident'ified as COND, NEG3, RMP, and the class 12

210

OM. The case was also seen of the infinitive kg; which is
phonologically identical to ku- which we referred to as an
’imaginary marker.’ Semantically, the cooccurrence of the
verb morphemes is restricted if the meaning resulting from
these combinations is not acceptable in Haya society. On
the other hand, certain meaningful combinations are
impossible because these particular meanings have already
been assigned to other morphemes. The other observation of
the cooccurrence restriction were observed among the.
extensions which had "functional clash" (thematic role
clash). For example an extensional morpheme which removes
an object, such as passive -w-, cannot cooccur with one
which also removes an object, such as stative -ik-. Other
cooccurrence restrictions are based on pragmatics. Certain
morphemes cannot cooccur because their combinations resulted
in a meaning that was not pragmatically acceptable.
Finally, it was also pointed out in this chapter that the
argument structure of the verb root is a controlling factor
of what morphemes may precede or follow it.

The factors discussed above support Halle’s (1973:24)
observation (Scalise 1984:24) that "if a grammar is a formal
representation of a native speaker’s knowledge of his
language, then there must be a component some place in the
grammar that accounts for the speaker’s lexical knowledge."
A speaker knows (a) what the words of his language are (for

example, nibalima ’they are farming’ is a word in Haya, but

211
banimali is runn, He) that words have internal structure,

that is a Haya verb for example is composed of morphemes:

bi-a -shaalil-a —-> byashaalila
8 -IP -ROOT -IND

’they (e.g. cassava) have become bitter’

(c) that the internal structure respects a specific order of

morphemes involved. For example

(110) ba-laa-lu -ba-ching-ul-il -a
2 -IF -IONH-2H close-RV-APL-IND

’they will open (unclose) it for them (e.g. door)’

is a possible word but

(111) X a ~laa-ba-ching -ba -ul -lu -il-

IND-IF-Z-close-ZH-RV-IONH-APL

is not possible.

Furthermore, the discussion made in this chapter
strongly supports the hypothesis by morphologists such as
Aronoff (1981), Scalise (1986), Bybee (1985) and others,
that morphology interacts with other parts of grammar,
namely: syntax, semantics, phonology (as shown in chapter 2)
to mention a few. In the next chapter (chapter 5) we take

up one such theory, by B (1985) whose theory is semantically

212
based and apply her theory to the linear order of the Haya

verbal morphemes.

213
CHAPTER 4 ENDNOTES

1 Khamisi (1983:94) calls what we have named stem
morphemes as syntactic lexical suffixes.

3 Fossilized intransitive is a term we have borrowed
from Khamisi (ibid:103).

3 Eggggk;angan; ’bend one another’ sounds a little odd
because usually, for bending human beings another lexical

item is preferred, which is:

inam-y -angan-a
bend-CAUS-RCP -IND
’cause one another to bend’

where there is a causative morpheme involved.

21111

Table 12. Sueeary of Cooccurrence Restrictions of Prestee and Poststee Horpheees

 

(IV) (COND) (NEGI) (FOC) (SH) (NEGZ) (TNS) (MEGS (INF (IND)/(SU8J)

UNR

(IV) ‘ - - + + - - + +
(COND) - - - + + t - - f -
(NEGI) ‘ - - + - + - - + -
(FOC) ’ - - - + - t - - + / +
(Sh) + + + + + + - - + / +
(N662) + + - - ' + 1 - + - -
(TNS) t 1 + g + t - - 1 -
(Recs - - - - - - - . , -
(INF/UNR) - - - - - - - + + -
(IND) + + + + + + + + i + -
(SUBJ) + + - + + + + + + -

 

1. As for tense (j) leans that only certain tenses eay cooccur with that particular eorpheee. Refer
to specifics in the text.

2. +/+ eeans either indicative or subjunctive eay be applied

215

Table 13. Cooccurrence Restrictions of the Stem Morphemes.

 

OM RFM CAUS APL PASS RCP ST RPT

 

OM + — + + - - - +

RFM - + + - - - +

CAUS + + — +’ + - +

APL + + - + + + - +

PASS - - + + - L -

RCP + + + + - - +

ST - - — - - - -

RPT + + + + - + -

1. The ’—’ marker indicate the cooccurrence restriction of

the morpheme involved.

2. ’+’ marker indicate the possibility of cooccurrence.

CHAPTER 5

PRINCIPLES OF VERBAL MORPHOLOGY

5.1. Introduction

In this chapter we lay out the essential points of B’s
theory tn) account for' the recurrent. properties of
morphological systems including fusion and allomorphy.1 We
also discuss the application of the Haya data to B’s theory.

The chapter is presented as follows: in the second
section we discuss B’s theory, followed by a discussion of
why we chose Haya as a target of study in section three. In
section four we discuss the application of B’s theory to
Haya. Note that we do not intend to discuss all of the
verbal morphemes, but will discuss a few of them and will
make some general observations about linear order. Two sets
of morphemes are discussed at length; the subject, and
object morphemes. We conclude this chapter in section five
by commenting on a theory of Paradigm Economy presented by
Carstairs (1983) and its relevance to the present study,

followed by the summary and conclusion of this chapter.

5.2. B’s theory
The major assumption behind B’s theory is that the

formal properties of morphological expressions correlate

216

217

directly with aspects of the meaning expressed; the relation
between meaning and form is not entirely arbitrary (4) (Note
that numbers in brackets are page numbers). By "form," B
refers specifically to the ways in which meaning elements
may be related to each other formally (11). She gives three
such ways:2

1. Two meaning elements may be combined in a single
nondecomposable lexical item, which she refers to as a
"lggicalngxprgssign. An example is the English verb "kill,"
which combines the meaning elements "die" and "cause" (11).

2. Two meaning elements may be seen to be expressed by
separate units, but these units are combined into a single
words B refers to a combination such as this as an
”inflggtignaLwexprgsgigg." An example is the English word
"walked," which combines the meaning element "go on foot by
taking steps" (13), which is expressed by the unit "walk"

and the meaning element "past tense,’ which is expressed by
the unit "—ed" (11).

3. Two meaning elements are expressed by separate

words, referred to by B as a mtagtigwexpressign." An
example is the English expression "come to know," which
combines syntactically the meaning elements of "inchoative"
and "know" (11). B points out fthat the same meaning
elements also find lexical expression as the verb "realize"
(12).

B proposes two principles predicting what kind of

formal relation will hold between two meaning elements and,

218

in particular, which meaning elements are most likely to be

expressed as inflectional categories: i.e., falling under
(2) above. These principles are called Relevance and

Generality. B defines Relevance as follows: ”A meaning
element is relevant to another meaning element if the.
semantic content of the first directly affects or modifies
the semantic content of the second" (13; italics omitted).
B makes the prediction that two meaning elements which are
relevant to each other may find expression lexically or
inflectionally, whereas elements which are irrelevant to
each other will only combine syntactically.

B gives an example of the predictive power of Relevance
from lexicalization in English (13). There is an English
verb walk, which encodes the semantic notion "to go on foot
by ‘taking steps." She considers the Relevance of two
additional semantic notions to this one: that of doing
something "through water" and that of doing something ”on a
cloudy day." According to her intuitions, "through water"
is highly relevant to "go on foot by taking steps," and
hence we find the two combined in the English verb wade. ‘Cn
the other hand, "on, a. cloudy day" is not particularly

relevant to "go on foot by taking steps,’ and hence we find
no verb which means "go on foot on a cloudy day" (14).

B defines Generality as follows: "(For) an inflectional
category (to be general, it) must be applicable to‘all stems

of the appropriate semantic and syntactic category and must

obligatorily' occur' in the .appropriate syntactic context"

219
(17; material in parentheses added). She then suggests that

a morphological process must have relatively low semantic
content to be general. As an example of high semantic
content restricting Generality, she cites motion verbs in
Latin, which take a number of prefixes indicating direction:
Since these prefixes have high semantic content, they are
restricted to a subclass of verbs, namely motion verbs, and
consequently lack Generality, since they are not applicable
to all verbs. These prefixes therefore cannot be considered
to be an inflectional category (17).3 From: these two
principles, B make a number of predictions which are now

discussed next.

5.2.1. Predictions arising from Relevance and Generality
Before we discuss the predictions, note that based on a

survey of 50 of the world’s languages, B came up with a list

of categories which she regards as most likely to be

inflectional, which we reproduce below. without details or

examples (20). We will give and discuss examples when we
discuss the application of the Haya data; These categories
are:

(a) Valgngg refers to differences in number or roles of
arguments that the verb stem can take. (B:20) Examples are
tritransitive (four arguments), ditransitive (three
arguments), transitive (two arguments) and intransitive (one
argument), and causative forms for ‘which the number 'of

arguments varies.

220

(b) Vgige. distinctions change the relation that the
surface subject bears to *the verb (B:20, citing Barber
1975). Examples include active, passive, reflexive,
reciprocal.

(c) Aspggt, refers to the ‘way iJi which the internal
temporal constituency of the situation is viewed. Examples
include perfective and imperfect. Imperfective verb forms
indicate that the action has not been completed while
perfective verb forms indicate the completion of the action.

(d) Lens; locates the situation with respect to an
established point in time, either the moment of speech, or
some other point in time (B:22).

(e) Mpgd refers to the way the speaker presents the
truth of the proposition in the context of discourse and the
real world, including probability, possibility and certainty
(22) B points out that this category includes, but is not
limited to, expressions of assertion (indicative), non-
assertion. (subjunctive), command (imperatived and. warning
(admonitive).

(f) Agreemen; as a verbal inflectional category
formally encodes the relationship between a verb and its
arguments in terms of ‘features such as person, number,
gender' and. other' characteristics of participants in the
situation.

These categories will be mentioned frequently in the

discussion of B’s theory and its application to Haya. We

next turn to the predictions.

221

5.2.1.1. Relevance

The principle of Relevance makes three major
predictions: order, frequency, and fusion. of the verbal
forms to the stem, the details of which we will get into
momentarily. At the moment it is in order to mention that B
discussed other independent predictions which we were unable
to apply to Haya.‘ Moreover, it is appropriate to mention
here that this study will test the predictions which B draws
from her principle, but will not evaluate the degree to

which the predictions actually follow from the principles.

5.2.1.1.1. Order

This prediction, which will now be referred to as
linear order, has to do with the order of categories with
respect to a stem. B claims that the categories that are
more relevant to the verb stem will occur closer to the stem
than those that are less relevant (24). By this she means a
greater effect on the semantic content of the stem.
Conversely, the less releVant the category is to the stem,
the further away it is located from the verb stem. B argues
that this principle applies because~ it affects elements
before they are bound and are still movable in the clause.
This argument implies that the more a concept has to do with
the content of the verb, the closer it will occur to the
verb stem (211). Therefore when these elements become bound

(affixes) to the verb stem, their order will be the same.

222
That is, the more relevant affixes are closer to the verb-
stem, and therefore the less relevant affixes are further

away from the stem.5 B’s argument can be illustrated using

the following schema:

(1) x-y—z-stem-a-b—c

According to B’s hypothesis, relevance increases from x to g
and from g to a. More specifically B proposes that the
following inflectional categories common on verbs (20ff) may
be arranged on the following scale of relevance (24) where
(+) indicates the most relevant end of. the scale, and (-)

the least relevant.

(2) Valence-V0ice-Aspect-Tense-Mood-Agreement

m --------------- ‘ -------------------- (-)

Let us assume that the verb stem may be arranged in two
ways: (a) left of the stem and (b) right of the stem as in

(3.a) and (3.6) (using B’s categories).

(3) a. AGREEMENT-MOOD-TENSE-ASPECT-VOICE-VALENCE-VERB
STEM

(3) b. VERB-VALENCE-VOICE-ASPECT-TENSE-MOOD-AGREEMENT
STEM

223

As shown in (3) B’s theory predicts a mirror-image order of
morphemes on the left of the verb-stem and on the right of
the verb-stem, given that "the most relevant inflectional
categories occur closest to the verb-stem, and the least
relevant occur at the greatest distance from the verb—stem.”

An important point which B makes is that these categories

form. a continuum. The continuum hypotheses is a very
important point, for unlike the traditional approach
(Chomsky and Halle 1968, Siegel 1974, Aronoff 1976 . . .1 B

does not make a sharp division between categories that fall
under derivational or inflectional categories as has been
the case within the traditional approach but they are
relevant to the concepts of Relevance and Generality, as
will be discussed shortly.

Thus, B’s approach may be summarized as follows:

(1) the more relevant the categories are, the more
interior they appear with respect to the verb stem.

(2) the less relevant the categories are, the more
exterior they appear with respect to the verb stem.

Let us now discuss the second prediction: frequency.

5.2.1.1.2. Frequency

B tested the frequency of morphological categories for
verbs in 50 languages. She made two predictions: that there
were (1) few inflections among the most highly relevant
categories (29) and (2) more inflections among the least

relevant categories. These statements can be schematized as

224
in (4) where 1+) refers to a more of a category and (-) less

of a category and INFL means inflectional and RLV means

relevance.

(4) a. [~INFL]-[+RLV]

b. [+INFL]-[-RLV]

However, frequency predictions can only be tested
comparatively across languages. Nevertheless, as noted
above, Haya has all of the verbal inflectional categories
(see chapter 3). Haya, canr be added as another sample

language to B’s data.

5.2.1.1.3. Degree of fusion

B’s third prediction concerns fusion. She predicts
that the more relevant the morphemes are to the stem, the
more likely there is to be fusion; the less relevant the
morphemes the less likely there is to be fusion. She claims
there are two reasons for this: (1) that relevant categories
occur closer to the verb in the syntactic string (2) the
psychological restructuring of two words into one depends on
the relatedness of the semantic elements being joined, and
their ability to form a coherent semantic whole (38). These

predictions are hierarchically represented in this order.

( 5 ) Lexical-derivational-inflectional-free graeeatical-syntaotic
(+) < H

 

225

Again, B states that these expressions form a continuum that
ranges ”from the most highly fused means of expression,
lexical. expression, to the most. loosely joined. means of
expression, syntactic or periphrastic expression” (12). In
(51 the (+1 marker indicates the most highly fused and (-1
indicates the most loosely joined. Again, in order to have
a clear discussion of these categories their meanings are
briefly given without getting into details.

(a) lexigal egpressigg. In a lexical expression, two or

 

more semantic elements are expressed :Ui a. monomorphemic
lexical item. B mentions some examples from various
languages including English. In English as mentioned

earlier, the lexical items:

(6) a. kill combines the semantic elements ’die’ and ’cause’

(to die).
b. drop combines the semantic elements ’fall’ and

’cause’ (to fall).

(b) Der'vatiggul Exprmmsigg. B points out that a

 

"derivational expression resembles lexical expression in
that derivational morphemes are often restricted in
applicability and idiosyncratic in formation or meaning. It
resembles inflectional expression in that two distinct

morphemes are combined in a single word" (B:12). Three of

226

B’s verbal categories falls under derivational expression.
These categories are valence, voice and aspect (99). As an
example, B discusses the category of aspegt as a lexical,
derivational and inflectional category (100). Let us review
what she says about the aspectual derivational categories.
"Inceptive and iterative tend to be restrictedto stative
verbs, so in languages with a morphological inceptive (or
inchoative) marker, this marker is derivational.” She also
points out that iterative markers are derivational because
they are applicable only to some verbs which describe events
that take place in a very short period of time, such as
ggwgh, and .blink, but not. to stative verbs (100). The
example given is the morpheme wade. in the Kiwai language
(Ray 1933) which indicates "the repetition of an action or
its frequent or regular performance"; such a morpheme
resembles the Haya repetitive (RPT) morpheme discussed in
chapter 3 (3.5.51.

(c1 IatLeetiQrabexnrelssaen . In an inflectional
expression, each semantic element is expressed in an
individual unit, but these units are bound into a single
word. Examples are affixes added to a stem or a change of
stem, such as the English regular past tense, walkgd vs
irregular bggwght. Inflectional morphemes obligatorily
accompany the verb-stem. They are generally unrestricted in
applicability.

(d) Free grammaticgl. Free grammatical morphemes

belong to a closed class and occur in a fixed position.

227

They "resemble inflections in that they make up contrast
sets that are obligatory in certain environments and they
have positional restrictions” (12). B adds that they
"resemble periphrastic expressions in that they are not
bgwnd to lexical stems." Unfortunately B does not give an
example of these morphemes, but simply states that they are
not studied in her work, "although their existence is
recognized by the theory to be proposed" (12).

(e) Syntactic expression. By syntactic expression B

no... no.

means that semantic elements are expressed in separate

words . For example ’ come to know ’ contains the semantic
notions ’ inchoative ’ and ’ know . ’ These notions are also
expressed by a single word ’realize.’ The syntactic

expression is sometimes called a "periphrastic" expression
(ibid:12).

We now turn to the second principle Generality and
discuss how the predictions which follow from this

principle.

5.2.1.2. Generality

According to the determinant of generality "an
inflectional category must be applicable to all stems of the
appropriate semantic and syntactic category and must
obligatorily occur in the appropriate syntactic context”
(17). Simply stated, generality "involves the degree to

which a particular process is applicable within a particular

domain" (Walker 1986:4941.

228

As was the case with Relevance, Generality is proposed

to express the same three major predictions: order,
frequency and fusion. These predictions are the reverse of
those predicted by the Relevance prin‘ciple. We briefly

discuss these predictions based on Generality.

5.2.1.2.1. Order
This prediction concerns the proximity of verbal
categories to the verb-stem. Unlike Relevance, the most
general categories occur at the greatest distance from the
verb-stem. Using the linear order of the verbal morphemes

as proposed by B,

(7) VERB STEM-VALENCE-VOICE-ASPECT-TENSE-MOOD-AGREEMENT
1-1 ----------------------------------- (+1

Generality increases from valence to agreement, where ((+1
indicates more general and (-1 less general. In other words
agreement is more general to the verb-stem than valence. As
also was the case with Relevance, Generality forms a

continuum from the less general to the more general

categories . In this case, valence is less general than
voice, voice is less general than aspect . . . and mood is
less general than agreement. This ordering may be

represented as in (8) where ) indicates is more general than

229

(8) AGREEMENT>MOOD>TENSE>ASPECT>VOICE>VALENCE-VERB
STEM

B’s argument may be summarized as follows: the more general
the category is the more exterior it appears with respect to
the verb—stem and conversely the less general the category
is the more interior it is located with respect to the verb
stem. This generalization is the reverse of that of the
Relevance principle with respect to linear order.6 Let us
now discuss the second prediction in relation to Generality,

which is frequency.

5.2.1.2.2. Frequency

B’s test of 50 languages indicates that there are
larger .numbers of inflections among the least relevant
categories than there are among the most highly relevant
categories. In Haya (see chapter 3) there are thirteen
least relevant categories and ten highly relevant
categories, which support B’s prediction.7 Again, as
earlier mentioned, -frequency and order coincide. The
further away the categories are from the verb-stem, the more
likely they are to be highly General and the higher the
frequency. The closer the categories are to the verb-stem
they more likely they are least General and the lower the
frequency. Thus, the frequency prediction is summarized as‘
follows: (a) there are many inflectional categories among
the most highly General categories and conversely (b) there

are few inflectional categories among the least General

230
categories. These statements can be schematized as in (9)
where (+1 indicates more of a category and (-) indicates

less of a category and GEN means General, and INF means

Inflectional.
(9) a. [+GEN) - [+INFL]
b. {-GEN] — l-INFL]

Now let us move on to the third prediction, which is the

degree of fusion.

5.2.1.2.3 Degree of Fusion

Again, as was the case with frequency, the degree of
fusion also coincides with the order of categories. B, as
mentioned earlier, predicts that the more relevant the
morphemes are to the verb-stem the closer they are to the
verb-stem and the more likely it is that fusion will take
place. Conversely, therefore, the less relevant‘ the
morphemes are to the verb-stem the further away they are to
the verb-stem and the less likely there is to be fusion. As
discussed above, it is predicted that Relevant categories
are more likely to fuse than General categories. i

As was the case with other predictions, B represents

the fusion hierarchy as follows:

( 1(1) Lexical-derivational-inflectionaI-free graeeatical-syntactic.

231

A syntactic expression is the most loosely joined expression
and therefore more General, while the lexical expression is
the most highly fused expression and therefore less General
(more relevant). These predictions can be represented in a
continuum hierarchy as shown in (11) where > indicates more
General than and (+1 a more fused category and (-1 a less

fused category.

( 1 1 ) syntactic > free graeeatical > inflectional > derivational > leXical
(fl ‘ (fl

 

5.2.2. Synthesis

The discussion thus far has shown the intertwining of
B’s two major principles, Relevance and Generality, and the
predictions that are derived from these principles, which
are: order, frequency and fusion. The two principles and
their predictions coincide in the sense that while Relevance
predicts closeness of the category or categories to the
stem, fewer categories which are closer to the stem, and
more likelihood of categories which are closer to the stem
to fuse, Generality predicts the opposite of Relevance. It
predicts that categories which are General are farther away
from the stem, that more of these categories are
inflectional and that they are less likely to fuse.

This analysis may be schematized into a redundancy

feature matrix as suggested in (12.a) and (12.b)

232

(12) a.
+ CLOSE TO STEM
- INFL ---- [RELEVANCE]
+ FUSION

- CLOSE TO STEM
INFL ---- [GENERALITY]
- FUSION

+

5.2.3. Comments on B’s theory

It is not the purpose of this dissertation to review
B’s work (see Aronoff 1987, and Walker 1986). However, a
few comments are in order here. B mentions three
predictions based on the two principles Relevance and
Generality: Order, frequency, fusion. Of the three
predictions, order received. more emphasis throughout her
work than any of the others. From this order, the degree of
fusion is also predictable - with those categories near the
stem most likely to fuse with it than those further away
from the stem. While such a claim might be appropriate with
fusional languages, it might be a problem with agglutinative
languages like Haya. Also, the hierarchy of order of
inflectional categories has been questioned. For example,
Aronoff (1987) questions "why should voice, which relates
argument structure to syntactic subject, be more relevant
than aspect, which is confined semantically to the verb
itself? (118). We shall not attempt to answer this sort of

question; but rather, attempt to apply B’s theory as she

233
gives it to the Haya data. Interestingly, this application

sheds light on the analysis of two problematic morphological
categories, inflectional and derivational (see her chapter
4: 81-110).

In her work B reviewed a number of linguistic
definitions offered thus far (Greenberg 1954, Mathews 1974,
Anderson 1982, Kurytowicz 1964, Bloomfield 1933, Nida 1946 .
. . ) and pointed out some major problems, with each of them
(which are not of immediate relevance here). She then
proposed to treating Inflection and derivation as single
category constituting a continuum from inflection~ to
derivation. Such an approach, she claims fits her general
principles of Relevance and Generality as has been discussed
above. This is a claim that forms the discussion of the
applicability of Haya data to her theory in the next
section. But before that it is pointed out why Haya was

chosen as target of study.

5.3. Haya as target of study

As seen in the previous section, B’s theory is intended
to cover a wide, if not the entire, range of morphological
phenomena found in human language and has been tested on a
diverse sample of 50 languages. However, it is by no means
obvious that .B’s specific predictions necessarily follow
from her very generally formulated principles, given the
potential for wide interpretation of such terms as 7directly

affects" or "modifies" in the definition of Relevance.

234
Aronoff (1987:117) notes the problem of circularity of B’s

definitions. For example, he questions how we know what is
relevant. Even the justification provided by B for relative
ranking of individual cases leaves a lot to be desired.
Such cases as ’walking through water’ is lexicalized as wgdg
because ’whether one has one’s feet on dry land in water is
quite relevant to the act of walking’ (13), however, does
not seem to suggest that B has developed any specifically or
generally applicable criteria for Relevance.

We will apply the predictions B makes concerning the
ordering of specific categories to some of the data we have
assembled on Haya verbal morphology in chapters 3 and 4, and
confine our attention to those morphological issues that
this theory touches_upon.

The discussion will focus on inflectional expression to
the exclusion of the other two types of formal expression,
lexical expression and syntactic expression, because this
work has focussed on the verbal morphemes of Haya, which are
inflectional according to 3’3 definitions. Also, the
discussion will concentrate on linear order, with a few
comments on frequency and no mention of fusion.

This approach may be justified by first considering the
position of Haya within a morphological typology of the

world’s languages.

235

5.3.1. Typological position of Haya

Languages are classified into structural types, three
of which are relatively common: isolating, agglutinating and
fusional (Lyons 1968, Comrie 1981, Bybee 19851.3 An
isolating language is defined as "one in which all words are
invariable" (Lyons (1968:1891). This means that all words
are without internal morphological structure. Examples of
such languages include Chinese and Vietnamese, as can be

illustrated below in a Vietnamese sentence (Comrie 1981:40).

(14) Khi toi den nha ban toi, chung toi bat
when I come house friend I PLURAL I begin
dau lam bai.
do lesson.

’When I came to my friends’ house, we began to

' do lessons.’

This sentence shows that each "of the words in this sentence
is invariable, there being no morphological variation. . ."
(ibid:40). For example the word gig}; indicates both the
tense ’come/came,’ and bat dgw indicate the tense
begin/began. This language also has to; for both "I" and

"my,' and ”plurality is indicated, in the case of pronouns,
by the addition of a separate word rather than by
morphological means, so that the plural of toi "I" is chung

toi ’we’" (ibid:40).

236
Agglutinating languages are languages in which words
are composed of a sequence of morphemes with each morpheme
representing more or less one strand of meaning. B cites

Care as an example of an agglutinating language; and gives

the following sentence as an example (1985:451.9

(14) Garo /sok -ba -ku -ja -ma/=/sokbakujama/
arrive-toward-yet-neg-int

’has he not yet arrived?’

Example (141 shows that the word jsokbakujama/ is made of
morphemes each with a particular meaning. Haya is also a
highly agglutinating language, the details of which we will
come to momentarily.

Fusional languages (sometimes called inflecting e.g.
Lyons 1977:72) display no clear cut boundary between
morphemes. Expression of (different categories within the
same word is fused together. That is unlike AGGLUTINATIVE
languages, there is 1K) one-to-one~ correspondence between
these morphemes and the linear sequence of MORPHS (Comrie
1981). In other words, a single morpheme typically
represents several strands of meaning. Examples of fusional
languages include Latin, Greek, Hebrew and Arabic, to
mention a few. The following is an example from classical

Latin (Bybee 1985) and Hebrew, respectively (ibid:45).

(15) Latin: amo ’I love’

237

The morpheme —0 simultaneously represents the following:
1 present tense
ii active voice
iii lst person singular agreement

iv indicative mood.

(16) Hebrew (the verb ’to write’)

Past 33 masc: katav
Present, 33 masc: kotev
Future, 33 masc: yixtov

Imperative, masc. sg: ktov

"the consonants of the stem interdigitate with the vowels

that signal the inflectional categories, so that IE) clear

segmentation is possible" (45).
The three terms ’isolating,’ ’agglutinating’ and
’fusional’ refer to idealized language types. In reality,

languages can rarely be characterized as being exclusively
of one of these types. For example, Haya, as well as other

Bantu languages is highly agglutinating:

(17) Haya ba-laa-mu-jun -a
SM-TNS-OM-STEM-IND

’they will help him’

238

Example (17) shows that, the morphemes'occur sequentially,

one per syllable. However, it also shows a sign of fusion:

(18) tu-ka -mu-taigy -a
SM-TNS-DO-say good bye-IND

’we said good-bye to him/her’

the morpheme tu- ’we’ simultaneously represents
(19) i gender - first person
ii number - plural

iii agreement - agrees with noun

class 2 abantu "people"

And Haya also shows signs of isolation, such as in the

following morphemes:

kygnkg ’but,’ Qlwekyg ’therefore,’ . u"

’daily’ as

P.

1
1

contextualized in (20.a-c).

(20) a. Nyakato ya-a -chumb-a kxgwkg
. Nyakato SM-TNS cook -IND but
ti -ya-a -yogy -a byombo
NEGl-SM-TNS-wash -IND dishes

’Nyakato cooked but she didn’t wash the dishes’

239

b. Kamuzora a -la-yombek-ile glwgkyg na iwe
Kamuzora SM- -build - ~therefore and
o —yombek-e
you farm ~SUBJ
’Kamuzora has already built (a house), you should

also build (a house).’

c. Nyamwiza a -shom-a bwli
Nyamwiza SM-read-IND daily

’Nyamwiza reads daily’

The words kmnka ’ but , ’ olwekyo ’ therefore , ’ 121-1.1.11. ’ daily ’
have no internal morphological structure, unlike many other
words in Haya as shown in the above sentences . But in

general Haya shows overwhelmingly agglutinative morphology.

5.3.2. Appropriateness of Haya as test of B’s theory

The highly agglutinating nature of Haya makes it
particularly appropriate to evaluating a theory concerning
the linear order of inflectional categories:

a. First of all, Haya has a lot of inflectional
categories: most, if not 'all of the categories on B’s
hierarchy are expressed inflectionally in Haya as verbal
affixes. In addition, there are also verbal inflections
which do not fit any of her categories particularly well.
Before we get into the details of these discussions, it is

in order here to mention that, in evaluating B’s predictions

240

using the Haya data, a number of problems were encountered,
similar to those encountered by B in her analysis. Three
problems are discussed here:

(1) determining relative order, (2) distinguishing in
the light of current knowledge, categories like Tense and
Aspect, and (3) identifying Haya morphemes which represent
B’s categories. Concerning Relative Order, there are a
number of categories which cannot cooccur in a verb and
which occur in the same position, for example, the focus
marker (FOC), the initial vowel (IV), which signifies the
relative morpheme, the conditional morpheme (COND) and the
negative morpheme (NEGl). We have represented such mutually

exclusive categories with bracket notation:

(21) FOC

IV

COND

NEG
(cf. chapter 3:53). Second is the problem of indeterminacy
of categories such as tense and aspect. In Haya as far as

we know, there has been relatively little work on tense and
aspect and as of now it is impossible to state for sure
whether a particular morpheme belongs to the category TNS
or ASP. The relevant morphemes have therefore been
categorized as occurring in a single TNS/ASP slot (27).
Third, is the problem that, certain Haya morphemes do not
readily fit B’s categories.1° These morphemes occupy the

slots in the model (2211abelled xl-xa.

241

Thus far the Haya model can be represented as:

STEM
(22) Xl-AGR-TNS/ASP-Xz RFM - ROOT - X3
OM ExtensionsilNDi}LOC-MNR—Q-INT
SUB

X1 slot includes the mutually exclusive morphemes FOC ,
IV
COND
NEG

and X2 is an ‘{INF}_ morpheme.
UNR
As we have mentioned in chapter 3 section ~(3.21 the root
occurs in the middle of the verb so that morphemes that are
more relevant appear closer on both sides of the root as
shown in (22). Again looking at the Haya data, B’s
predictions are borne out, the most relevant categories, are
closer to the root on both sides, and the least relevant
morphemes are further away from the root, which indicates
they' are most general. This order of the Haya ‘verbal
morphemes based on the order of morphemes predicted by B
raises a number of questions: (1) do all the morphemes as
they appear in the Haya verb follow B’s predictions? (2) if

not how do we modify B’s theory or account for the ’misfit’

of these morphemes with respect to B’ 3 ‘model . These are
some of the questions we raised in chapter 1 ( 1 . 1'1 . An
attempt to answer these questions, leads us into the

theoretical part of this study. We will shortly discuss the
status of the object. morpheme ‘marker in respect to 18’s

linear order prediction in (5.3.3.4) below.

242

b. Secondly, the more agglutinating and less fusional
nature of the inflectional morphemes of Haya make linear
order much easier to evaluate. What fusion there is (of.
chapter 2) does not obscure the relative order of the
morphemes which undergo it.

c. Thirdly, we would also like to point out that B’s
original 50 language survey did not include any Bantu
languages, although at one point in the text she discusses

causativization in Luganda (18).

5.3.3. Application of the theory to Haya

We begin the discussion with a classification of the
Haya verbal morphemes (apart from the SM and OM, which are
discussed in the next section) according to B’s verbal
inflectional categories, with comments on the extent to
which their position in the Haya verb matches B’s

predictions.

5.3.3.1. The Pres-stem Morphemes (non—stem)

a. Initial Vowel

As we discussed in section (3.3.6) and (4.2.4) IV is
prefixed to the subject marker to form the subject relative
morpheme. This kind of morpheme is not discussed in B’s
work. Maybe B would have categorized it under agreement, if
anything, because its form varies according to the noun

class of the subject of its verb (see table 4). If so, the

243

IV occurs in precisely the position predicted for it, at the

periphery of the remaining verbal morphemes.

b. Conditional

In Haya the conditional morpheme £813., which may be
translated as the English if, states the condition to be met
in order for a certain event to take place. B (ibid:1985)
suggests that COND should be categorized under MOOD
following agreement. Such a suggestion fit into B’s order
of morphemes according to Relevance and Generality, with
only the change of position between the agreement and COND
morphemes in Haya. Whereas agreement is the farthest
category in B’s hierarchy and MOOD the last but one, the
Haya agreement and conditional morphemes exhibit the reverse

order.

0. Negative and Focus

B (ibid:176) admits that negation does not quite fit in
any of the proposed categories but in some of its "uses
negation can resemble (what she is calling) mood, in that it
can have the whole proposition in its scope" (words in
brackets mine).

On the other hand, the category focus is not mentioned
by B. This would possibly be categorized under mood also.
B defines mood as "how the speaker chooses to put the
proposition in the discourse context" (165). As we

discussed in section 3.3.8 one of the functions (known thus

244

far) of focus in Haya is emphasis. Such a function of focus
fits B’s categorization of rmmxi. If the proposal that
negative and focus go under mood go through, then it is
appropriate to claim that both NEG and FOC fit B’s
prediction. They are located farther away from the stem

than most of the non-mood categories in Haya.

d. Tense
Tense is one of B’s categories. Tense in Haya follow
B’s prediction that a tense morpheme will occur

approximately in the middle of the other verbal morphemes.

e. Infinitive

The infinitive is not discussed at all in B’s work. We
do not know where B would have placed it. Probably INF
would have been categorized as MOOD, but we do not know for

sure.
5.3.3.2. The Stem Morphemes

a. The Reflexive, Passive and Reciprocal

These morphemes fit B’s categorization of voice. B
(ibid:20) defines voice as a category "that changes the
relation. that the surface subject has to the verb." For
example, in the reflexive the subject both performs the
action and is affected by the action. These morphemes

occurs closer to the stem as predicted by B.

245

b. Causative and Applicative

These two morphemes are categorized under valence.
According 1x) B such morphemes are highly relevant to "the
situation described in the verb stem, since the situation
expressed changes according to the number and role of
participants in the) situation" (20). Such a description
fits the two morphemes in Haya, as described in chapter 3.
Moreover these .morphemes occur immediately following the
verb-root as discussed in (4.4), which fits the prediction

that valence morphemes will occur close to the stem.

c. The Reversive and Repetitive
Reversive and. Repetitive .morphemes have no apparent

classification in B’s list of categories.
5.3.3.3. The Poststem Morphemes (non-stem)

a. The indicative/subjunctive morphemes

The indicative and subjunctive morphemes are included
in B’s mood category. These two morphemes are described as
the expression of assertion and non assertion, respectively.
Their position in Haya fits B’s prediction in so far as they

are further away from the stem.

246

b. Locative, Manner, Question and Intensifier
These morphemes have no apparent counterparts in any of
B’s categories, although the manner morphemes may possibly

be regarded as mood.

5.3.3.4 The Subject Marker and Object Marker

In. chapter' 3 section (3.4.1), we proposed, without
justification, that the subject markers and object markers
in Haya have significantly different syntactic functions.
We will discuss this proposal further in this section.

Both the SM and the OM encode noun-class features, of
the subject and the object in a sentence, respectively.
Both of them are, in this sense, agreement markers.
However, when we look at their position with respect to the
root, it can be seen that the OM is close to the root, at
the periphery of the stem, whereas the SM is farther from
the root, at the periphery of the. other inflectional
morphemes. If both the SM and the OM are agreement markers,
then their different positions constitute a counterexample
to 8’8 theory. Is there any evidence that the SM and OM in
Haya represent different categories, and thus that their
different positions might be predicted by B’s theory?

In order to answer this question, we will apply ideas
developed for the analysis of object markers in Chichewa, a
Bantu language related to Haya, by Bresnan & Mchombo (1987;
henceforth referred to as B & M1.1l Although B & M develop

their analysis within the framework of Lexical-Functional

247

Grammar (Bresnan 1982; this is given in B &.NVs references
as Bresnan 1982d), we believe that the arguments they give
may stand on their own without extensive discussion of the
assumptions underlying this theory.

B 8: M distinguish between grammatical and anaphoric
agreement. According to 13 & M, morphemes marking
grammatical agreement redundantly express agreement features
of arguments of the verb. .Morphemes marking anaphoric
agreement are incorporated pronominals, and the NPs with
which they agree are non-arguments.

B & M give two criteria for distinguishing the two, one
involving evidence from sentence grammar, the other
involving evidence from discourse. We will describe and
apply these two to Haya, to argue that the OM in Haya shows
syntactic behavior parallel to that of the OM in Chichewa,
which constitutes our argument that the OM in Haya is a
marker of anaphoric agreement, in B & M’s terms.

The first criterion is that of word order. B & M show
that in simple transitive sentences, where the verb has no
OM, the NP object must immediately follow the verb, whereas
the subject may be reordered. Simply stated, when there is
no OM, word order is restricted as shown in (744) examples,

repeated here as (23):

(23) "a. SuVO: Njfichi zi-na -lum -a alenje
bees SM-PAST-bite-INDIC hunters

’The bees bit the hunters.’

VOSu:

OVsu:

VSuO:

SuOV:

OSuV:

248

Zinaldma alenje njfichi.

. . I .A .
X Alenje Zinalfima njuchi.

. I ' . . .
X Zinaluma njachi alenje.

.~ . . . z
X Njuchi alenje Zinaluma

X Alenje njachi zinélfima

But when the OM is present, word order is free; that is, all

the above orders are possible as shown in their examples,

repeated here as (24):

(24) "a.

b.

C.

d.

SuVO:

VOSu:

OVsu:

VSuO:

Njfichi zi-na —wa—lum -a alenje
bees SM-PAST-OM-bite-INDIC hunters
’The bees bit them, the hunters.’
Zinawéluma alenje njfichi.

. .II .A .
Alenje z1nawaluma njuchi.

. z I .A . .
Zinawaluma njuchi alenje.

249

e. SuOV: Njfichi alenje zinawéluma

f. OSuV: Alenje anchi zinawaluma" (7451.

The relevant question at this point is: how may these
data be accounted for? B & M (745) give five explanations.

(a) SM plays two roles: those of grammatical and
anaphoric agreement. ‘The pronominal interpretation arises
when the subject NP is omitted. As a matter of fact, when
the subject NP is omitted in the above examples in (23) and
(24) the pronominal interpretation holds (that is,
agreement).

(b) The OM is used as an incorporated object pronoun
and not an object marker.

(0) All object NPs in Chichewa appear in post verbal
position. Therefore the appearance of a post verbal object
in a VP is only possible when there is no OM on the verb.

(d) Sentences in Chichewa consist of an optional
subject NP, a VP, and an optional topic NP, which are all
unordered. Such combinations allows six different orders as

schematized below:

(25) a. NPsu VP NPt

b. NPsu NPt VP

0. VP NPsu NPt

250

d. VP NPt NPsu
e. NPt VP NPsu
f. NPt VP NPsu

where NPsu means noun phrase subject, NPt means noun phrase
topic.'
(e). Finally, B & M argue that the cooccurrence of OM

with an object NP is explained as the 'anaphoric binding of
an object pronoun, incorporated in the verb, to a topic NP
in S. (746)."

The five explanations above help to explain the
differences between examples (23) and (24). The SM is both
a grammatical agreement marker and an incorporated pronoun,
whereas the OM is only an incorporated pronoun. The SM is a
grammatical agreement marker when it cooccurs with the NP
which controls its 'agreement features; in the absence of
this NP it is an incorporated pronoun. The OM is an
incorporated pronoun whether or not its controlling NP is
present or absent. This controlling NP of the OM is not an
object but a topic. The data on reordering constitutes the
evidence for this difference between the SM and the OM. If
the SM were an incorporated pronoun when cooccuring with its

controlling NP, then this NP would be a topic and could be

reordered within the VP, even in the absence of an OM.

251

However, it can only be reordered to a position outside of
the VP, as indicated in the examples in (23). Therefore, in
this case, the SM is purely a grammatical agreement marker.
On the other hand, when the OM cooccurs with its controlling
NP, this NP can be reordered anywhere, within the VP or
outside of it, as in the examples in (24). The OM in this
case is an incorporated pronoun.

B & M replace the examples (24) with examples (26) below:

(26) a. Su[vp v1 TOP: Njfichi zI-né-wélium-a alenje
bees SM-PAST—OM-bite-INDIC hunters

’The bees bit them, the hunters.’

b. [vp V] TOP Su: Zinawaluma alenje njfichi.
’they bite them, the hunters, the

bees.’

c. TOP[Vp V] Su: Alenje z1nawaluma anchi.
’the hunters, they bite them the

bees’

d. [vp V] Su TOP: Zinawaluma njfichi alenje.
’they bite them the bees, the

hunters’

252

e. Su TOP[vp V]: Njfichi alenje zinawaluma
’the bees, the hunters, they bite

them’

f. TOP Su[vp V]: Alenje njuchi zinawaluma" (745).

’the hunters, the bees bite them’

B & M’s analysis raises two questions: (747).

(1) "Why is gender class agreement required between the
topic NP and the incorporated object pronoun?"

(2) "Why must the topic NP be anaphorically linked to
an INCORPORATED pronoun, which suspiciously' resembles an
agreement marker, rather than to an independent pronoun in
the object NP position (as in English example I love him
dggrlxw_m1_fggh§r1?" (italics omitted).

The answer to question 1 above is that person, number,
and gender are pronominal categories which universally show
agreement in anaphoric relations.u Several examples are
given to support this point. We show two of those examples

here (10) and (11) repeated here as:

(27) (Pointing to a lion lying on the ground)
UWU

this (class 3)

253

(28) (Pointing to a lion on the ground1X
X Ichi

this (class 7).

In Chichewa the word lion, mkgwgg, belongs to gender class
3. Indicating deixis to a lion, as shown in (27) above,
requires the class 3 deictic pronominal form; the class 7
form (28) shown would be used for other class 7 things.
These examples indicate that the choice of agreement
features of person, number and gender in the anaphoric use
of pronominals is independently motivated, and not motivated
exclusively by the sentence internal mechanism of agreement.

As to the explanation of the use of incorporated
pronouns to anaphorically link the topic NP’s to the
predicate argument structure, the answer lies in the
internal. structure of the language. In Chichewa,
"independent object pronouns are used only for introducing
new topics or for contrast" (748) (see B & M’s examples
(12)-(15) 748-749).

As the examples in (29) show, the pattern in Haya is
identical to that in Chichewa. Simple transitive sentences
in Haya, in which there is no OM on the verb, show fixed

word order (the object must immediately follow the verb):

(29) a. SuVO: E-njoki zi-ka -lum -a A -ba-hiigi
IV-bee IO-RMP-bite-IND IV-Z -hunter

’The bees bit the hunters’

254

b. VOSu: Zikaluma Enjoki Abahiigi

c. OVSu: X Abahiigi zikaluma Enjoki

0"

d. VSuO: Zikaluma Enjoki Abahiigi

e. SuOV: X Enjoki Abahiigi zikaluma

96

f. OSuV: Abahiigi Enjoki zikaluma

But when the OM is present, all the above orders are

possible in Haya as well:
(30) a. SuVO: E-njoki zi-ka- b; -lum -a A -ba-hiigi
IV-bee IO-RMP OM(2) bite-IND IV-2 -hunter
’The bees bit them, the hunters’
b. VOSu: Zikabgluma Enjoki Abahiigi
c. OVSu: Abahiigi zikabgluma Enjoki
d. VSuO: Zikabgluma Enjoki Abahiigi

e. SuOV: Enjoki Abahiigi zikabaluma ‘

f. OSuV: Abahiigi Enjoki zikahaluma

255

In Haya too, as was the case in Chichewa, independent Object
pronouns are also used only in introducing new topics or as

also is the case in Chichewa, for contrast.

(31) a. Omwana akalya ekiliibwa. Olwo yamazile
child ate cassava(7). When he finished
Ku -ki -ly -a yagya omushule
INF-it(7)-eat-IND he went to school
’The child ate the cassava. Having eaten it (the

cassava) he went to school.’

? b. Omwana akalya ekiliibwa. Olwo yamazile
child ate cassava(7). When he finished
Kulya ebyo yagya omushule
eating them(7) he went to school.
’The child ate the cassavas. Having eaten them

(something other than cassava) he went to school.’

While example (31.a) is natural in Haya, (31.b) is odd. .The
independent pronouns as in Chichewa are interpreted as
referring to topics not mentioned in the previous sentence
even though they agree with objects of the previous sentence
in person, number and gender class. That is, example (31.b)
is interpreted as meaning that the child ate several things

including cassava.

256

We turn now to the second criterion, that of locality.
Locality is defined with respect to the proximity of
agreeing elements within the clause structure; local
agreement holds between elements of the same simple clause,
while non-local agreement may hold between elements of
different clauses. In accordance with the assumptions
developed in the discussion of word order, locality
constitutes another difference between arguments of the verb
and topics and focuses linked to the verb 'by :anaphoric
agreement. Only non-arguments may be separated from the
clauses to which they are linked.

B & M illustrate locality in Chichewa with the

following examples:

I
(32) Chigawénga ichi asilikali 5 ganyu

terrorist (7) this soldiers of temporary work

a -na -Gz -é mtsogoleri wathu kuti

SM-REC.PAST-tell-INDIC leader our that
I f ' ’

s -a -ngath-e ku -chi -gwir -a

not-SM-can -SUBJN INF-OM(71-catch-INDIC
’This terrorist, the mercenaries told our leader they

cannot catch him.’

257

. / I . . . . I .’ A
(33) X Chigawenga ichi a31likali a ganyu

terrorist (7) this mercenaries

’ I . A .’
a -na -uz -a mtsogoleri wathu kuti
SM-REC.PAST-tell-INDIC leader our that
I e
s -a -ngath-e ku -gw1r -a

not-SM-can -SUBJN INF-catch-INDIC
’This terrorist the mercenaries told our leader they

cannot catch.’

These examples indicate that an object marker in
Chichewa may be used to bind a topic which is outside of the
-clause in which the object marker appears, and that such
binding is obligatory in Chichewa.

Additional evidence that this NP is a topic is seen in
the following examples, which indicate that it may be
reordered within the topmost clause in the same manner as in
the examples above N34.a-c) given earlier to illustrate the

reordarability of topics.

(341 a. Asilikali é ganyu a -na -fiz -§ mtsogoleri
mercenaries SM-REC.PAST-tell-INDIC leader
/
wathu kuti s -a -ngath-e ku -chi -gwir -a

our that not-SM-can -SUBJN INF-OM(7)-catch-INDIC
chigawéngé ichi

terrorist(7) this

’The mercenaries told our leader that they

cannot catch him, this terrorist.’

258

I
b. Asilikali é ganyu chigéwengé ichi

mercenaries terrorist(7) this

a -na -52 -a mtsogoleri wathu
SM-REC.PAST-tell-INDIC leader our

3 -a -ngath-é ku -chi -gwir -a

not-SM-can -SUBJN INF-OM(71-catch-INDIC

/

c.X ? Asilikali 5' ganyu a -na -Gz -é
mercenaries SM-REC.PAST-tell-INDIC
chigawénga ichi mtsogoleri wathu
terrorist(7) this leader our

. I .’ .’
kuti s -a -ngath-e ku -chi -gwir -a

that not-SM-can -SUBJN INF-OM(7)-catch-INDIC

As seen in examples "(34.a) and (34.b1 the topic may be
placed at the end of the sentence, or may follow the highest
subject. Example (34.c) shows that it may not occur within
the VP, following the main verb. This is because only
arguments, not non—arguments, may reappear within the VP.
With respect to the locality criterion, again the
pattern in Haya is identical to that in Chichewa. As in
Chichewa, the floating topic NP can be non-local from the OM
to which it is linked and the non-local topic shows the same
ordering possibilities within its higher clause, as examined

in examples (35) and (36), respectively.

259

(35) Omwisi ogu abai shelukali bo okulipwa

terrorist (1) this soldiers of temporary work

(mercenaries)

ba-a -gambil—a omubembezi waitu bati
SM-TNs-tell INDIC leader our that
ti -ya-shoboil-e ku -mu -kwat -a
not-SM—can SUBJN INF-OM(l) catch-INDIC

’This terrorist, the mercenaries told our leader they

cannot catch him.’

In (35), the class noun (1) omwisi ’terrorist’ is a floating
topic NP. There are three levels of level embedding above
class (11 OM which agrees with it. If it is also removed,
as was the case in (12), the sentence becomes ungrammatical

as shown in (36).

(36) X Omwisi ogu abaishelukali bokulipwa

terrorist (1) this mercenaries

ba-a -gambil-a omubembezi waitu bati
SM-TNs-tell INDIC leader our that
ti -ya-shoboil-e ku°-mu ~kwat -a
not-SM-can SUBJN INF-OM(l) catch-INDIC

The ungrammaticality of (36) as was the case in Chichewa,
follows from the fact that in Haya too, topicalizations are

constructed by anaphoric binding.

260

Concerning the ordering possibilities within the higher
clause, in Haya too, the topic NP can appear sentence-
finally as in (37.a1, and after the highest subject, as in
(37.b) but NOT after the main verb inside the VP, as in

(37.0):

(371 a. Abaishelukale bo okulipwa ba-a -gambil-a
mercenaries SM-TNs-tell -INDIC
omubembezi waitu bati ti -ya-shoboil-e
leader our that not-SM-can -SUBJN
ku -mu . -kwat -a omwisi ogu
INF-OM(7) catch-INDIC terrorist (1) this
’This mercenaries told our leader that they

cannot catch him, this terrorist.’

b. Abaishelukale bo okulipwa omwisi ogu
mercenaries terrorist (11 this
ba-a‘ ~gambil-a omubembezi waitu
SM-TNs-tell -INDIC leader our
ti -ya-shoboil-e ku -mu -kwat —a

not—SM-can -SUBJN INF-OM(7) catch-INDIC

261

c.X ? Abaishulukale bo okulipwa ba-a -gambil-a

mercenaries SM-TNs-tell -INDIC
omwisi ogu omubembezi waitu °
terrorist this leader our

bati ti -ya-shoboil-e ku -mu -kwat -a
that not-SM-can -SUBJ INF-OM(l) catch-INDIC

The application of the Haya data to the criteria established
by B 81. M, shows identical results for Chichewa and Haya.
The word order criterion shows that both Chichewa and Haya
follow a fixed word order: the object immediately following
the verb. The locality criterion shows that the floating NP
can be non-local from the OM to which it is linked and that
non-local topics show the same ordering possibilities within
their higher clauses. ‘These results indicates that the (wt
in Haya (as well as Chichewa) is not an agreement marker but
something else.

Let us now turn to the SM. It has been suggested that
the OM is always an incorporated pronoun and never a non-
referential marker of grammatical agreement (B & M:
ibid:755). The relevant question at the moment is what
about the SM? It was pointed out earlier that SM is indeed
such a marker and it also has a referential use, under
appropriate conditions, as an incorporated pronoun. That
is, when the SM is used as a grammatical agreement marker,
it also agrees with the subject NP, and when the SM is used

for anaphoric binding, it functions as a subject topic as

262
shown in B & M’s work as example (321 repeated here as (38).
The topic mkango uwu ’this lion’ appears three levels of
embedding above its verb- initial sentence position in
(38.a), following the highest subject. in (38.b), and in

sentence final position in (38.c).

(38) a. Mkéngd uwu, alenje a -ku -géniza kutf’
lion(31 this hunters SM-PRES-think that
‘6 -ma -ffina ku -gumdla nyumba yé'mffimu
SM(3)-HAB-want INF pull.down house of chief
’This lion, the hunters think that it wants to pull
down the chief’s house.’
b. Alenje mkangd uwu a —ka ~ganiza kuti
hunters lion(3) this SM-PRES-think that
5 -ma -funa ku —gdmdla nyumba ya mfumu

SM(3)-HAB-want INF pull.down house of chief

. . .’I I/

c. Alenje a —ka -ganiza kuti u -ma -funa
hunters SM-PRES-think that SM(3)-HAB-want

ku -gfimula nyumba ya mffimu mkang6 uwu

INF pull.down house of chief lion(3) this

These examples, show that SM(3) is used as an incorporated
pronominal for non-local anaphora to the topic and the other
SM is used as a true grammatical agreement marker, a pattern

which is also evidenced in Haya.

263

(39) a. Entale egi, abahiigi ni -ba -tekeleza bati
lion(91 this, hunters FOC-SM(21—think that
NP SUBJ(2)
e -yenda ku -teneka enju yo omukama

SM(91-want INF pull down house of chief‘
’This lion, the hunters think that it wants to pull

down the chief’s house.’

b. abahiigi entale egi ni -bg -tekeleza bati
hunters lion(9) this FOC-SM(21-think that
NP SUBJ(21
e -yenda ku -teneka enju yo omukama

SM(3)-want INF pull down house of chief

c. Abahiigi ni -b§ -tekeleza bati
hunters FOC-SM(21—think that
NP SUBJ(2)
e -yenda ku -teneka enju yo omukama
SM(3)-want INF pull down house of chief
entale egi

lion(9) this

As was shown in the Chichewa examples, in Haya too, SM(9) is.
used as an incorporated pronominal for non-local anaphora to

the topic while the other SM (SM(2) underlined) is used as a

264

true grammatical agreement. marker agreeing in (class and

number with the NP subject.

The SM is also used referentially when it does not co-

occur with a nominal to which it has to agree in class and

number as exemplified in both Chichewa (40) and Haya (41).

(40) Chichewa

(41)

a. gi-na -lum -a alenje

SM-PAST-bite-INDIC hunters

b. gi-na -wa-lum -a

SM-PAST-OM-bite-INDIC

c. gi-na -wa-lum -a ' alenje

SM-TNs-OM-bite-INDIC hunters

Haya
a. Zi-ka -lum -a abahiigi

SM-TNs-bite-IND hunters

b. zi-ka -ba-lum -a

SM-TNs-OM-bite-IND

c. Zi-ka -ba-1um -a abahiigi

SM-TNs-OM-bite-IND hunters

265

In such cases the SM is itself the subject in the absence of
the overt NP subject. As nicely summarized by Moshi
(1985:139) the ”subject has neither something to control nor
something by which it is controlled." This adds evidence to
B & M’s treatment of SM as both used for grammatical
agreement and anaphoric binding.

Bresnan and Mchombo present evidence that in Chichewa
the SM is both a grammatical agreement marker and an
incorporated pronoun, whereas the OM is exclusively an
incorporated pronoun. This evidence also exists in Haya.
(Similar conclusions have been reached by Byarushengo, et
al., (1976) in their study of tone in Haya; cf. B & M (1987:
751, fn.13).) The difference between the two suggests an
account of their different positions with respect to the
root within B’s theory. The SM functions partially as an
agreement role in determining its position. ij the SM is
regarded primarily as an agreement marker, then it may be
seen that it occupies the position predicted by B’s theory.
What about the OM? The evidence suggests that it is
exclusively an incorporated pronoun, a category which is not
included in B’s list of inflectional categories whose linear
order is predicted by Relevance and Generality. However, a
pronoun represents an argument of the verb and this
observation may indicate a connection between a pronominal
inflectional category and valence, even though valence, in
B’s account, refers to morphemes which alter argument

structure, not just represent it. (29) Insofar as this

266

connection may be made, the position of the OM with respect
to the root fits B’s prediction concerning the position of a
valence morpheme. This explanation is problematic, since
the SM may also function as an incorporated pronoun under
Bresnan & Mchombo’s analysis, but B’s theory may be
considered to offer a tentative account of the connection
(if any) between the different functions of the SM and the
OM and their different positions in the verb.

Thus far, a distinction has been made between the OM
and SM. It has been established that the SM functions as a
grammatical marker as well as an anaphoric binder. Such an
analysis based on the Haya data supports B’s prediction and
B 8: M’s analysis. We have examined the SM and OM in Haya
and concluded that they show different syntactic behavior.
What about the reflexive marker, RFM? As seen in chapter 3,
the RFM resembles the OM in that it immediately precedes the
verb root and refers to an object of the verb. It differs
in that it does not encode the noun-class of the object it
refers to. The question now is: does the RFM behave like an
incorporated pronominal (in B & M’s terms), like the OM?

The RFM, unlike the OM, cannot occur with the NP that

it refers to:

(42) X A -ba-hiigi ba-ka -e -tem-a a -ba-hiigi
IV 2 hunter 2 RMP-RFM-cut-IND IV 2 hunter

’hunters cut themselves hunters’

267

Therefore, neither the criterion of word order nor the
criterion of locality used earlier to argue that an OM which
cooccurs with the object it refers to is an incorporated
pronominal can be used with the RFM, since both of these
criteria. rely crucially’ on the» OM cooccurring with the
object it refers to.

In the next section we raise an issue which we have not

discussed so far, concerning our earlier analysis of stems

and non-stem morphemes? How may we further justify this
analysis? HOW’ about the traditional approach of
derivational and inflectional morphemes. Does this approach
concur with B’s approach. These are some of the questions

we try to answer next.

5.3.3.5. Split between sets of morphemes

In chapter 4 we split the Haya verb morphemes into two
groups: the stem mOrphemes and non-stem morphemes, both
occurring before and after the verb root. The non-stem
morphemes. may" be further divided. into prestem morphemes

(43.a) and poststem morphemes (43.b):

(43).non-stem morphemes

IV

a. COND - SM - NEG2 - TNS - (NEG3 - (INF (NEGZ)) - X-
NEGl
FOC

b. x - IND - LOC - MNR - Q - INT
SUBJ

268

These two schemata may be shown according to their relation
to the Root, (marked x in 43) as shown in (44). Since the
Root is obligatory, we will show its cooccurrence with the

prestem and proststem morphemes.

W

(44) COND - SH - NEGZ - TNS - NEG} - (INF (NE62)) - ROOT -{IND }- LDC - HNR - 0 ‘ INT
NEGI SUBJ
FM

The stem morphemes are arranged as follows:

(451 RFM - ROOT -{CAUS}-{PASS}- APL - RPT
(DOl)(1031 APL RCP
Each set of morphemes really constitutes a separate set as

exemplified in (46):

(461 a. a -ba-laa-shal-a -ha
IV-Z -IF -cut-IND-LOC

’who will cut here’

b. ku -gul-a -ki-kwo
INF-buy-IND-Q -INT

’to indeed buy what’

269

0. ti -mu—chumb-ile
NEGl-Zp-cook -RP

’you haven’t cooked’

Examples (46.a-c) indicate that non-stem morphemes may
cooccur on their own without stem morphemes. This supports
our discussion in chapter 2 of the non-stem morphemes in
terms of their relationship to the stem, without reference
to the internal structure of the stem. In addition, the
stem morphemes were described without reference to non-stem
morphemes.

In chapter 4, the claim was made that, with very few
exceptions, the cooccurrence restrictions on morphemes hold
between morphemes belonging exclusively to one group or the
other, the non-stem morphemes or the stem morphemes. We
again show some examples to clarify this point. The begin
with examples which show the cooccurrence restrictions of

the non-stem morphemes:

(47) a. X ka -ti -ba-li-bon-a

COND-NEGl-Z -NF-see-IND

b. X ba-ta -li-ku -lum -a

2 -NEG2-NF-INF-bite-IND

Examples (47.a-b) are ill-formed because in (47.a) COND and

NEGl are mutually exclusive and therefore cannot cooccur.

270
In (47.b) it is shown that INF cannot cooccur with the SM,

NF or NEGZ. The examples above are repeated here with

necessary modified requirements and they are well—formed.

(48) 8. ka -ba-li-bon-a
COND-2 -NF-see-IND

’when they see’

b. ti -ba-li-bon-a
NEG1-2 -NF-see-IND

’they wont see’

(49) a. ba-ta -li-lum -a
2 -NEG2-NF-bite—IND

’they should not bit (in the near future)’

b. ku -lum -a
INF-bite-IND

’biting/to bite’

Furthermore, the stem morphemes by themselves may have
noun formation morphology attached rather than the verbal
morphology discussed in chapter 3. Noun formation

morphology is exemplified as follows:

(50)

(51)

(52)

Note

phonological changes as described in chapter 2.

a.

8..

a.

that examples (50.b) and

271
ku -kol -a

INF-work-IND

’to work’

0 -mu-koz -i
IV-l -work-DA

’worker’

ku -eelez-a
INF-serve—IND

’to serve’

a-ba-eelez-i
N-2 -serve-DA

’servants/helpers”

ku -lwal -a
INF-be ill-IND

’to be ill’

o-mu-lwail-e
IV-I-sick-PAT

’sick person (patient)’

(52.b)

are

subject to

Note also

that the vowels following the verb stem in these examples

272
imply broadly the doer of the action (DA) in example (50.b),

(51.b), and patient or sufferer of action (PAT) in example
(52.b). The following examples show various stem morphemes

which may appear with this noun-formation morphology:

(53) a. o -mu-e -yeg -es -elez-o —-> (omwegeshelezo)
IV-3 -RFM-learn-CAUS-RPT -RST

’teachings (i.e. religious teachings)’

b. o -mu-sig -il -w -a
IV-3 -trust-APL-PASS-IND

’steward/trustee’

c. a -ba-shab-iliz-i
IV-2 -beg -RPT -DA

’beggars’

The above discussion of the order of the Haya verb
morphemes fits B’s prediction of the "mirror-image’ order.
In Haya where the ROOT is lbcated in the middle of the Haya
verb morphemes, we suggest that the most relevant morphemes
occur closest to the root and the least relevant occur at
the greatest distance from the root. Therefore the stem
morphemes are more relevant to the root than the prestem or
poststem morphemes (cf. the schemata of the verbal morphemes
of Haya in chapter 3 (53). These two groups of morphemes

are reminiscent of the more familiar division made between

273

inflectional and derivational morphology.“ We may
tentatively propose that the non-stem morphemes are
inflectional morphemes, whereas the stem morphemes are
derivational.

As we pointed out in.(5.2.1.1), B’s idea is that the
inflectional/derivational distinction is a continuum rather
than a sharp division. Note that support for this idea
comes from the cooccurrence restrictions which hold between
the groups. They appear to be restrictions which hold
between morphemes which are at the border of the continuum
between inflectional and derivational. This is where you
expect to find morphemes which behave somewhat like
inflectional. morphemes and somewhat. like derivational
morphemes. We will here discuss one such morpheme, the
subjunctive.

The SUBJ morpheme must cooccur with SM or OM. (cf.
examples (56.a,b)). (Note that when it cooccurs with OM,
there is always an option of SM). This, therefore, is an
instance of a cooccurrence restriction holding between a
non-stem morpheme (SUBJ) and a stem morpheme (OM). Note

that the SUBJ morpheme occurs immediately after the stem

morphemes of the root.

(54) a. mu -ba -twal-e o -mu-shule
SM(2p)-OM(2)-take-SUBJ IV-9 -school

’you (pl) should take them to school’

274

b. ba -twal-e omushule
OM(2)-take-SUBJ school

’take them to school’

5.4. Carstairs theory

We would like to conclude this chapter by briefly
discussing Carstairs’ (1983) theory of the Paradigm Economy
Principle (PEP) which poses an interesting observation that
is relevant to this study. Carstairs investigated the
distribution of the inflexional resources available in some
word-classes or part of speech among the languages of world
class (16); and found out that the ’inflexional resources’
as he calls them, do not approach the upper limit, but
rather the minimum. In other words, languages which have a
large number' of inflectional morphemes, and therefore a
large number of potential combinations of these morphemes,
do not exhibit the maximum number of combinations but rather
the minimum. These inflexional resources includes for
example nominal and verbal morphemes, among others. With
evidence from five languages-Latin, Hungarian, Zulu (a Bantu
language), Dyirbal and. German, and by applying what he
called the "traditional notion of paradigm" he found out
that those languages do not approach the "logical maximum"
but are at or close to the minimum possible inflexional
resources. As an example, suffix combinations of classical
Latin were worked out, totalling 1,371,686,400 possibilities

but only 5 actually occur (119). With these results,

275

Carstairs hypothesizes that "there exists a real tendency
towards ’paradigm economy’ - towards keeping the total
paradigms for any word class close to the logical minimum"
(127). The question is, is this prediction borne out in
Haya?

With 23 morphemes in the Haya verb (see chapter 3 (7)),
it is logically possible to have a maximum of 1,334,832
morpheme combinations, reached by multiplying the 22
morphemes with the two or more possibilities of appearance
or non-appearance of each category, for those which are
optional. These logical possibilities extend from two

obligatory morphemes within a single stem - . . .

(55) STEM + IND}
SUB

- to the maximum. The maximum possible number of morphemes
in a verb stem is approximately 12,960. (see rule 71
chapter 4). This number is approximately 9.7% of the total
logical maximum combination. Haya therefore confirms

Carstairs theory, as does B’s work (see chapter 3 of her
book), that there exists a real tendency of keeping the

total paradigms for any word class to the minimum.

5.6. Summary and conclusion
Testing B’s hypothesis with Haya data tends to confirm
her work, with minor modifications. Although it is not very

clear whether B groups the'voice and valence categories with

276

derivational morphemes, we have done so on the basis of
evidence internal to Haya. Given this conclusion, this
study supports B’s prediction concerning hierarchy and order
and more generally RELEVANCE and GENERALITY in so far as the
hierarchy and order. predictions follow from the two
principles. We have also discussed the observation that the
OM in Haya fits the prediction in B’s model only when it is
regarded as an incorporated object pronoun, considered as a
valence marker, following the evidence derived from B & M’s
(1987) study. These conclusions support our earlier general
conclusions reached in chapter 4, that semantic, syntactic
and pragmatic factors determine the linear- order of
morphemes. Syntactic factors have not only been discussed
in chapter 4 but also in this chapter when discussing among,
other things, B & M’s proposals of the analysis and order of
the OM and SM. Finally, we mentioned in chapter 4, that the
linear order of morphemes can be barred by'the meaning
derived through combinations of morphemes well formed
syntactically but not accepted culturally (pragmatic

factors).

277
CHAPTER 5 ENDNOTES

1 B’s theory has been reviewed by Aronoff (19871 and
Walker (1986), but has not (as far we know) been applied to
any Bantu language.

3 These ways are not discrete, but fall along a
continuum, whose intermediate points are not discussed here.
Cf. B: Chap.2.1, in particular 12.)

3 This example illustrates the problematic character of
Generality, since a particular category can be made more or
less general depending on how the "appropriate semantic and
syntactic category" is defined. She the Latin motion verb
example, if the "appropriate .. . . category" is defined as
all verbs, as she defines it, then motion prefixes are not
general; however, if the category is defined as all motion
verbs, then it is. It seems to us that Generality is most
useful as a relative notion: one morpheme is more or less
general than another with respect to a particular syntactic
category.)

‘ B mentions three other independent predictions:
language acquisition, language change tuul word order. As
for language acquisition, B notes that children acquire
categories that are closer to the stem before those further
away from the stem. For example, according to the child
language literature, aspectual distinctions on 'verbs are
made long before person/number and agreement distinctions.
(Antinucci and Miller 1976, Stephany 1981, Simoes and Stoel

Gammon 1979, Bloom et al 1980).

278

Evidence cited for language change indicates that
changes occur in the exterior positions vs the interior
positions with respect to the stem (Bybee and Brewer 1980).
Finally, a comment is also made that the order of morphemes
within a word reflect an earlier ordering of words within a
sentence (Givon 1971, Vennemann 1973). We did not get into
details of these predictions for Haya.

5 See Aronoff’s (1987:117) reservations on these
principles.

6 We will return to this point during the discussion of
the application of B’s theory to Haya data (Section 5.3.3).

7 The thirteen least Relevant categories are IV, COND,
NEG, FOC, SM, TNS, INF, IND, SUBJ, LOC, MNR, Q, and INT.
The ten highly Relevant categories are: RFM, DO, IO, CAUS,
BEN, PASS, RCP, ST, RPT and RV.

8 A fourth type not discussed by B is what Comrie
(1981:39) has called polysynthetic (or incorporating). In a
polysynthetic language it is possible to combine a large
number of morphemes, lexical or grammatical, into a single
word, often corresponding to a whole sentence of English
e.g. Eskimo etc. (Ibid:42) (see details).

9 Garo is one of Perkins sample languages (see B
ibid:215). It is a Sine—Tibetan type of languages
(Burlington 1961).

1° However, note that B acknowledges that an exhaustive

survey of derivational morphology was not undertaken .

279
that "there are many more possible derivational categories
than the ones mentioned . . ." (29).

11 Chichewa is a Bantu language spoken in East Central
Africa - particularly in Malawi and its neighboring
countries of Mozambique, Zambia, and Zimbabwe, where it is
also known as Chinyanja. See B & M (1987:7431 fn3.

13 B & M cite one other piece of evidence for the
pronominal use of object markers, that they may be used as
resumptive pronouns in the repair of island violations (752-
764); we will not discuss this use of the object marker
here.

13 Givon (1976) proposed that, "in the historical
development of Bantu, subject and object pronouns used for
reference to topics (’The man, he came’) became cliticized,
and then morphologically bound to the verbs (’The man he-
came’); B & M (19871741).

1‘ There is a current debate going on as to whether
derivation and inflection must be separated or not. Scalise
(1986:1021 lists all those who regard inflection in grammar
as operating in the syntactic component (the so-called ’Weak
Lexalist Hypothesis’ (WLH)) as morphologists who assume that
Derivation and inflection are different in nature. These
morphologists include Chomsky and Halle ((19681, Chomsky
(1970), Siegel (1974) Aronoff (1976), Allen (1978.a),
AnderSOn (1982) and others. A list of those who maintain

that_ they. are not different and therefore must. not be

280

separated is long. the reader is referred to Scalise
(ibid:102).
Scalise (ibid:103-115) lists 8 reasons in favor of the

division of the two blocks of rules.

CHAPTER 6

CONCLUSION

6.1. Summary

In this chapter we summarize the major findings of this
study and state some suggestions for future research. We
also state the contribution this study has made in the field
of linguistics.

In chapter 1, we provided a brief overview of the Haya
language, including its .classification and dialects. We
stated the major goals of this study, which are:

(1) to identify the Haya verb morphemes and their
linear order by answering the question: is the linear
ordering of the Haya verb morphemes predictable?

(2) to discover the. factors which determine the
ordering of the Haya verb morphemes, and

(3) to determine whether the factors in (2) above,
follow B’s theory.

In the last section of chapter 1 we discussed the
sources of the data and the organization of this
dissertation.

In chapter 2 we provided a phonological sketch of Haya,
which includes the phonological inventory, a word on tones,

the distribution of consonant and vowels, the phonological

281

282
rules and finally a word on syllable structure. Only those
phonological aspects which are relevant to the Haya verb
morphology are discussed. .
In chapter 3, we identified 23 Haya verb morphemes and
we discussed the functions of these morphemes.

In chapter 4 we discussed four factors which determine

the cooccurrence restrictions of the verb morphemes. These
factors are (1) phonological, (2) syntactic, (31 semantic,
and (4) pragmatic. For example, certain morphemes cannot

cooccur because they are phonologically identical also, some
morphemes have cooccurrence restrictions because they clash
semantically. Such extensions includes the passive (-w-_)
and stative (-ik-).

In identifying the four factors above, we have taken a
broader perspective on the interaction of the various
language components. Our work as we have mentioned earlier
supports other linguists (for example, Aronoff 1981,
Anderson 1982, 1986, Scalise 1986, Selkirk 1982) in arguing
that morphology interacts with other parts of grammar,
namely, phonology, syntax, semantics and pragmatics.

In chapter 5 we tested B’s theory with Haya data. We
found' out that Haya data tends to support B’s work, with
minor modificationsQ For example, by grouping voice and
valence categories with the stemt morphemes in Haya B’s
prediction concerning hierarchy and order (of her 50
language sample) and. more generally, RELEVANCE and

GENERALITY, is supported.

283

6.2. Suggestions for further research.

As we mentioned in chapter 1 and chapter 5, our work is
a single language approach, but this study may be—expanded
and modified to include ten or more Bantu languages. Such
expansion of the data base would ensure that all the
morphemes that are most common in Bantu languages are
covered and the linear order and or co-occurrence
restrictions is determined. Such a large data base could
then be used to test B’s theory such as her theory about the
frequency of occurrence of inflectional categories as verbal
morphemes. Such a prediction, as we pointed out could not
be evaluated in a study of one language.

In addition to that, a huge data base gathered from ten
or more Bantu languages could be compared to another ten or
more non—Bantu languages and the results could be related to

B’s theory.

6.3. Contribution
Despite the limited nature of this work, it has made
several contributions.

1. This study as far as we know is the first to pub B’s

theory to the test. In so doing we have fulfilled B’s
invitation which she made to readers, to test her
predictions: ”no systematic test of these predictions has

been carried out, but the reader is invited to apply his or

284
her knowledge of the languages of the world to test these

predictions" (page 20).

2. Although this study used B’s (ibid) general
principles of: RELEVANCE and GENERALITY, it differs in the
method applied. Whereas B examined the inflectional
morphology of verbs in 50 languages, our methodology was to
investigate one Bantu language (Haya) in detail, a language
which was not represented in her sample (a modified version
of Parkins (1980)). It was shown that Haya has all of B’s
categories along with additional categories that did not fit
into B’s categorization. Moreover, because Haya is a highly
agglutinating language, it was easier to see the order of
morphemes relative to the stem. This study has therefore
added to 3’3 data.

3. In addition, because this work is embodied in a
recent theory of the linear order of morphemes, it will
(hopefully) be of interest to morphologists in general and
African morphologists in particular. That is, this work has
added on to the growing literature on morphology,

particularly on African languages.

APPENDICES

. APPENDIX A

A FEW EXAMPLES OF
PRESTEM MORPHEMES IN USE

A FEW EXAMPLES OF
PRESTEM MORPHEMES IN USE

EE.L.A.T.I.Y.E§.

Eibala kalikwenda lwa Bunyana

Her name is Beloved-one

Enkuku eyonakateka omumutwe!

What knuckles which I tapped her head with!

Akati k’engoma ilikwekonya

Litle scourge of fearful rulers

Aina enkoni y’Omukoma-mpango

His quarter-staff wraps up the matter

Manti leba nibo balikumweteza

Certainly they bring it on themselves

Mwenda-ngabo atumilwemu Nkwele

A shield-breaker sent by King Nkwele

285

286

Omwibobo elya wakakya

At the iron bars that chime a warning

Ekyalabile mbele ekya Kalemela

First to bear fruit of Kalemera

Eyayoky’ engoma yomukama ali mulenzi

He who burned a royal drum while still a youth

Olunyana lwawe olwo wagisize

Your favorite calf which you gave the special pasture to

Waitu olwo wagisize Kijongo

My lord the one you pastured specially nat Kijongo

Ndiwa Bushalita na Muchwampaka#Ekyakukwata

I am of bushalita and the Decider-of-cases#The -matter-that-
holds-you-in-its-grasp

Alikubandama kegy’ eigulu
One who knealt to give his arrow flight
CONDITIONAL

Kabatamu zeleba zaikiliza

When they consulted there the oracle saw and favorably spoke

287
Nkahulila kayagobile nyalulembo nyalulembo nyalulembo

I heard the king as he reached his palace forecourt
forecourt forecourt

Kayagobile omwilembo ly’Abakulwa bonene Kayanga Kayanga

When he entered the Bakulwa forecourt at Kayanga (?)

Kayachwile kuleba ati Omumuti omumushasha

Suddenly he glanced up in a mushasha tree

Kabashabukile Kyenyabasa

When they had crossed the ford at Kyenyabasa

Kayaleba ati Kamachumu yaija yabona omuli Lwanda n’omuli
Busingo

He looked thus at Kamachumu he came and saw into Lwanda and
Busingo

Mh Kayatukula yayema ahali Lutenge yakokoloma

When it ran red he stood up at Lutenge and crowed like a
rooster

Emihambi kelikulaba omubishansha

When the arrows began to fly among the leaves of the banana
plants ‘

Kayahikile omukashekela mwonene

When he arived at laugh-at village

288
Kabalengile bwankya

When they judged it to be daybreak

Komubembe Tibagiteile

Since they did not beat the Leper

Kabagaba omumugamba

When they piled them in the back pantry

Kabaganywa gabakwata

When they drank the beer it took hold of them

Kayayetengya bati Atem’ ente niwe

When he’d draw attention to himself with kingly gestures
they’d say The slasher of cattle herds is you

Mwakutengya olubingo tulutema bwoli bwoli

Draw our attention to a bamboo grove and we’ll slash it
truly truly

Kayayetengya bati Atem’ente niwe

When he’d draw attention to himself with kingly gestures
they’d say The slasher of cattle gerds is you

NEGAIIV S

Nka naba ntakujumile

Never have I said an angry word to you

289

Mbali obwile bwakyela

Nor since the dawn last lightened have we fought

Ati Omuloto naba ntamanyile

I did not see a dream she said

Waitu tinabaliz’ ekindi

My lord I did not ask another word

Tiyabaliz’ ekindi

He did not ask other word on that account

Kolaba otamanya okwo agolala

If you really don’t know how he sets things straight

Tiyabaliz’ ekindi

He did not need to ask another word

Endulu tiyalek’ endulu

The cry did not leave off the cry

muses

Present Universal Habitual.

Ninye omu Kishenye ekya minazi

I was there at Sandy-spot of the munazi trees

290

Ati Obuso inye nitongana ekilamyo

I beg to tender my respect in his presence

Omushaija ntukuganyila

Fellow we pity you (?1

Nagolol’ata omukama?

How does the king set things straight?

Abantu natema nemiholo angili nasiga omubazi?

Does he out people up with machetes or does he administer
medicines?

Angili leba nanyuka omubili

Or does he massage the body?

Lwamanumi nagoloza ebintu

The owner-of-bulls straightens things with wealth

Twasiga nayesha obulaza

We left him playing stockpile strategy in makala

Omukama nayesh’ ekigumba

The king plays the strategy of cattle herd

Emboni nazihanga Lutwiga

But his pupils point to the slavegirl long-necked Lutwiga

291

Ati Waitu nintongana

My lord I beg a favor

Waitu obuso nintongana ekilamyo

My lord I beg to tender my respect in his presence

Obuso ninyija kuitayo omwiyo

I’m coming soon to order that the court be cleared

Ninkulamya ekyobutini

I tender my respect in words of fearfulness

Ninshuba ninkutaho ekyobwegula moyo

As well I set before you words to buy my life
Immediate Past

Nakutawe ekitebe oli w’enketo

I place before you a throne to sit

Kisa-nshone nayesig’ eiganzi

She-who-can-make-a-man-feel-shame is certain that she’s
loved

292

Akakazi kangutuz’ omuloto gwange
Well this little one stopped my dream for me
Kati Baba wantulaki?

She says Why chastize me love?

Kuba wangutuz’ omuloto gwange

Because you broke my dream for me

Buti Walotaki mbandwa?

What was your dream spirit-medium?

Waitu walotaki embandwa?

What was your dream our lord spirit medium?

Omunaga bwateka ahakyoto

The placed an earthenware pot among the flames

Akashumu omungemu bwandagaza

With the pith of-a banana plant they wipe my hands

Eichumu bwateka ahakilindi

My spear they placed on the sharpening block

Ekilele bwatamu amabale

My calabash they cleaned inside with stones

293

Otwalw’ otwo kulangaiza

Banana beer to sip relax and thind with

Amajuta bwanteka ahamulundo

With butterfat they rubbed my shins

Near Past

Obukazi tibwakeile kilekyo

The ladies did not delay that day

Tibabaliz’ ekindi

They did not ask another word

Nilwo nasokile omuly’entale

And then I descended into the wilderness of lions

Omu Kamezile-nfuka

At Hoe—swallower field

Ah Muhimbile-ngundu

Ah Protector Bull

Bati Omushaija anyoile amalwa obukyakya

The man’s drunk beer at dawn they said

294

Omushaija anyoile amalwa olukyakya

The man’s drunk beer at dawn

Nkome kaisho kaigaile

The One-eyed Binding -rope

Ati Wenene asingile njuki?

He said In which house is he now?

Asingile njuki?

In which house is he now?

Ayegekeile Njwanga

He’s leaned against his wife Njwanga

Lwa Mugimbile-ngundu

Oh Protector Bull

Tuhulil’ endulu Kalabagaine

We heard a cry from Kalabagaine village

Bahi Ikilakulya kihigile bwankya

They said the What-will-chew-you-up troop went to hunt at
dawn

Olwo yakasingile akyaija

When he had come

295

Olwo yakamutaileho amaisho

When he set eyes on him
Remote Past

Nkabanduka omwibanga ly’empumi

I climbed the mountain of hyenas

Nkataho omutana gw’etaba

And then a roll of tobacco

Nkahulila ati Telikunda ya Byabato

I heard him call Never-gives-in child of children’s-matters

Nkakuba omuti

I climbed a tree

Nkagya Buganda

I went to Uganda

Nkatabala Buzinja Nagya natabala Bunyoro

I raided in Buzinza and then raided in Bunyolo

Nkaitamu omuntu

I killed a man there

Akabikila Makuba Ngimbwa

He summoned Causer-of—retreats Obliged-one

Nkahulila Omuhamba

I heard a Muhamba

Nkabona ati Yoi yoi yoi yoi

I heard someone say Yo yo yo yo He’s speared you

Omuli Ikunyu mwonene nkabona

In Ikunyu village I saw

Obulwano bweila bukankwasa enkoju ahalisho

A longstanding feud has fixed a scar upon my eye

Nkahulila wenene Omunyihangilo

I heard the Munyaihangilo

Akataishula abashaija tukalwana
And free his men to fight
Infinitive/ "Immaginary" Marker

Kusisimuka kitwe yampinda

Awake I started with a big head roaring

297

Waitu bwatamu otwo kulangaiza

My lord they poured in just enough to sip relax and think
with

Nikyo kwija kundagila omutambi

Which is why I came to be directed to a healer

Nikwo kwija

Which is why I came

Kutola mugambwage yange

To catch my Well-spoken-of-thing

Ahi tata kugya kushwela mawe

He said When my father went to court my mother

Okugyeta mashayaye

Calling it a broken-mouthed vessel

Nakuba nimpulila Nyamunyalila Ishanga

I would be hearing the Urinataor-in-the-marsh

Nakuba nimpulila Nyamwitwa ababi

I would have heard Ruined-by—the-ugly

Ati Bakwitayo enjobe

They must have killed a bushbuck (?) there he said

298

Bakwitayo eikombo

They must have killed a mongoose there

Nakuba nimpulila lwa magabe

I would be hearing the Apportioner-of-trails

Nakuba nimpulila bilimu lw’ enanga

I would be hearing the Sound—of—an-enanga

Bakwitayo omunyima

They must have killed an ’omunyima’ there

Nakuba nimpulila ebijumilo nkenzaile

I would be hearing groaning like a calving cow

APPENDIX B

SOME STEM MORPHEMES: RESTRICTION OF
OCCURRENCE BASED ON VERB TYPES

2

10

11

12

13

14

15

16

17

18

19

20

Some

STEM
jum-
shagal—
yongel-
igan-
kul-
ikiliz-

laga-

gendelel-

bon-

lek-

hindul—

hulul-

nigaal—

bik-
olool-
yetab-
hakan—
hik-
baz-

shab-

stem morphemes:

based on verb types.

restriction of occurrence

RV CAUS BEN ST PASS RCP RPT

’abuse’ X +
’accompany’ X +
’add to’ X XX
’be equal’ X +
’grow’ ~' X +
’agree’ X X
’promise’ X +
’progress’ X +
’see’ + +
’leave’ X +
’alter’ X +
’prune (coffee1’. X +
’get anger’ ' X +
’announce death’ + +
’answer’ X +
’answer (call)’ X +
’argue + +
’arrive’ X +
’ask’ 3 X +
’ask for’ X +

299

¥

X

+

+

+

+

+

+

21

22

23

24

25

27

28

29

3O

31

32

33

34

36

37

38

39

40

41

42

43

44

45

STEM
sisiimuk-
8&8-
shob-
gamb-
jumbik-
shaabul-
boigol-
gumb-
yog—

b-

zaal-
banz-
jugut-
hend-
shombf
inam-
log-
hangu—
lum-
shaalil-
ilagul—
hum-
huut—
fuu—

yatan-

300

’wake up’

’go bad’

’go bad

’say’

’bake in ashes’
’bail out of
water’

’bark’

’barren’
’bathe’

:ben

’bear children’
’begin’

’blow’

’break’

’carry things’
’bend’

’witch’

’big, become’

’bite’

’become bitter’
’become black’
’become blind’
’blow’

’blow noise’

’border on

RV CAUS BEN ST PASS RCP RPT

46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69

70

STEM
tiil—
yat-
gutul-
iky-
leet-
lel-
monyok-
ey-
yombek-
yoky-
yak-
hy-
balul-
tond-
baluk-
baluul-
ziik-
tem-
gul-
yet-
zing-
heek-
shutul-
tweek-

twaal-

’borrow’
’break’

’break’

301

’breathe’

’bring’

’bring up (child1’

’break off’

’sweep’
’build’

’burn’

’take away’

’become

cooked’

’burst open’

’boast of’

’burst’
’retell’
’bury’
’butch,
'buy’
’call’

’fold’

cut down’

’carry on back

(child)’

’carry in arms

(child)’

’carry on head

(child)’
’take’

RV CAUS BEN ST PASS RCP RPT

71
72
73
74

75

77
78
79
80
81
82
83
84
85
86
87
88
89
9o
91
92
93
94

95

STEM
shumul-
lekel-
ky-
bing-
kanj-
nokel-
lond-
jwal-
juul-
chuchul-
ij-
lug-
chumb-
kolol—
bal-
shweek-
yatik-
yajul-
shaabuk-
bandam-
lil-
lime
lumik-
tamb-

log-

302

’castrate’
’cease’
’Xcease of rain’
’chase’
’chew’
’choke’
’choose’
’clothe’
’take off
(clothes1’
’comb hair’
’come’

’come from’
’cook’
’caugh’
’count’
’cover’
’crack’
’craw (of child)’
’cross over’
’crouch’
icrye
’cultivate’
’cupp’
’cure’

’curse’

RV CAUS BEN ST PASS RCP RPT

I

X

+

+

+

+

+

t

4.

+

+

+

XX

XX

XX

XX

+

4.

XX

96

97

98

99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119

120

STEM
shal-
zin-
ilagul-
8W-

f-
beiy-
sing-
nYY'
yetong—
kelelelw-
imuky-
shuuk-
shush-
end-
songok-
lem-
iy-
bul-
ihul-
fundul-
tamw-
gab-
kol-
hanantuk-

t...

303

’cut up’

’sing’

’get black/dark’
’fall’

’die’

’deceive’
’defeat’
’defecate’
’deny’

’delay’

’depart’
’descend’

’look like’
’desire’
’descend’

’be difficult for’
’dig (farms)
uproot’

’get lost,
disappear’
’dish up food’
’open’

’dislike’

’give, sacrifice,
divide’

’do’

’get down’

’go home/draw
wather’

RV CAUS BEN ST PASS RCP RPT

*

X

+

+

+

+

X

X

+

XX

4.

XX

XX

XX

+

+

304

STEM ' RV CAUS BEN ST PASS RCP RPT
121 loot— ’dream’ + + + X X + +
122 nyw- ’drink’ X + + + X X X
123 toony— ’drip’ X + + + X X +
124 tamiil- ’get drunk’ + + + X X X +
125 ikal- ’dwell’ X + + + + X +
126 ly- ’eat’ X + + + + XX X
127 laal- ’go without food’ X + XX XX X X XX
128 gool- ’embarass’ ' X + + X X + X
129 hing- ’exchange’ X + + X + X X
130 shoboolol- ’explain’ X + + + + + +
131 laaz- ’extinguish’ X + + + + X +
132 lemw- ’get fired/fail’ + + X X X + X
133 kom- ’wrat, fasten’ + + + + + + 2+
134 gomok- ’become fat’ X + + X X X +
135 mal- ’finish’ X + + X + + +
136 honder- ’follow’ X + + X + + +
137 kuntil— ’foot fall’ X + + + X X +
138 lagul- ’foretell’ + + + + + + +
139 tang- ’forbid’ X + + + + + +
140 nywan-i ’become friends X + + + + + +
141 yeb- ’forget’ X + + + + X +
142 tiin- ’frighten’ X + + X + + +
143 kalang- ’fry’ X + + + + + +
144 ijul- ’become full’ X + + X X X +
145 igut- ’become full’(with X + + + + X +

food)

STEM RV CAUS BEN ST PASS RCP RPT
146 sheeny- ’gather firewood’ X + + + + X +
147 h- ’give’ ‘X + X X X + X
148 lagil- ’give a message’ X + + + + + +
149 gend- ’go’ X + + + + X +
150 ebembel- ’go ahead’ X ‘+ + X + + +

 

Ker:
X ill-formed
XX ill-formed but becomes a different lexical item

+ well formed (blank space)

B IBLI OGRAPHY

BIBLIOGRAPHY

ALLEN, M. 1978a. Morphological investigations. Unpublished
doctoral dissertation. University (MS Connecticut,
Storrs, Connecticut.

1978b. The morphology of negative prefixes in
English. North Eastern Linguistic Society, Toronto,
Canada. 8:1-11.

 

1980. Semantic and phonological consequences of
boundaries: A morphological analysis of compounds, M.
Arnoff and M. L. Kean (eds.) Juncture, Anna Libri, 9-
27.

 

ANDERSON , HENNING. 1980 . Morphological change: toward
typology. Historical Morphology, ed. by J. Fisiak, The
Hague: Mounton.

ANDERSON, STEPHEN. R. 1982. Where’s Morphology?, Linguistic
inquiry 13, 571-612.

1986. Morphological theory. To appear in
Linguistics: The Cambridge Survey.

 

ANTINUCCI, FRANCESCO and RUTH MILLER. 1976. How children
learn to talk about what happened. Journal of Child-
Language 3. 169-189.

ARONOFF, MARK. 1976. Word formation in generative grammar.
Cambridge, Massachusetts: MIT Press.

1980. The relevance of productivity in a
synchronic description of word formation. Historical
Morphology, edited by J. Fisiak, The Hague: 1980.

 

and ANSHEN, FRANK. 1981. Morphological
productivity and phonological transparency. Canadian
Journal of Linguistics 26.63-72.

 

1982. Potential words, Actual words, Productivity,
and Frequency. Preprints of the plenary session papers.
The XIIIH1 International Congress of Linguistics,
Tokyo, 141-148.

 

306

307

1987. Review: Morphology: A study of relationship
between meaning and form, by J. L. Bybee. 1985.
Language. 63.1:115-129.

 

ARNOTT, D. W. 1970 The nominal and verbal system of Fula.
Oxford Clarendon Press.

ASHTON, EDITH O. - 1944. Swahili Grammar (including
intonation). Longman Group Ltd.

et al. 1954. A Luganda Grammar. London: Longmans,
Green.

 

AYERS, M. DONALD. 1986. English words from Latin and Greek
elements. The University of Arizona Press. TUCSON.

BAMMERSBERG, A. 1984. A sketch of diachronic English
morphology. Regensburg: Press.

BARBER, E. J. W. 1975. Voice--beyond the passive. Bulletin
of Linguistic Science 1.16-24.

BARLOW, RUFFEL A. 1951. Studies in Kikuyu grammar and idiom,
Edingurgh. Blackwood.

BATIBO, HERMAN M. 1980. Review: Nurse, Description of Sample
Bantu Languages of Tanzania In Journal of Africa
Languages and LlngUISthS. 2.2:172-176.

BAUER, LAURIE. 1979. Against word based morphology.
Linguistic Inquiry 10.3: 508-9.

1983. English word formation. London. Cambridge
University Press.

 

BEARD, ROBERT E. 1977 On the extent and nature of
irregularity in the lexicon. Lingua 42:305-341.

BEAUCHAT, P. D. 1963/64. A restatement of the Zulu Verb
Conjugation. African Studies. 22.4:137-169, 23.1:36-49
ptz, 2302:67-87 pt3.

1966. The verb in Zulu. Johannesburg,
Witwatersrand University Press.

 

BENDER,‘ LIONEL M. and Hailu Fulas. 1978. Amharic Verb
Morphology. Language and Linguistics, 1. Monograph 7.
East Lansing. MI. African Studies Center. Michigan
State University.

BENDIX, EDWARD HERMAN. 1966. Componential analysis of
general vocabulary: The Hague: Mouton.

308

BENNETT PATRIC R. 1969. A comparative study of four Thagiau
Verbal systems: The inflectional systems of Kikuyu
Kamba Embu, and Mwimbi. Ph.D. Thesis. School of
Oriental and. African Studies. University' of Lonson.
1982. The status of predicator in Dhaagicw and Bantu.
Etudes Linguistiques Vol. 3

1982. The status of the predicator in Dhaagicw and
Bantu. Etudes Linguistiques Vol. 3.

 

BERGVAL VICTORIA. 1987. A typology of empty categories for
Kikuyu and Swahili. Current approaches to African
linguistics, ed. by Gerrit J. Dimmendaal, 3 Dordrecht:
Foris, 198x.

BETBEDER, PAUL W. F. 1943. Empondera Yezu ((Haya): Followers
of Jesus). Vic. Apost. De Bukoba. Kashozi.

and JOHN JONES. 1949. A handbook of the Haya
Language. White Fathers Printing Press Bukoba.
Tanganyika territory (Tanzania).

 

BLOOM, et al. 1980. Semantics of verbs and the development
of verb inflection in child language. Language 56.386-
412.

BLOOMFIELD, LEONARD.1933. The structure of learned words.
The structure of language, ed. by Charles F. Hockett
1970:252-256.

BOAS, HANS U. 1974. On Halle’s "Prolegomena to a theory of
word formation" or what is a linguistic generalization?
Linguistics 134:5-8. ‘

BOKAMBA, EYAMBA. 1985. Verbal agreement as a noncyclic rule.
In Bantu African Linguistics. Studies in the sciences
of language series 6:9-54.

BOLINGER, DWIGHT L. 1950. Rime, assonance and morpheme
analysis. In Bolinger 1965:203-226.

1965a Forms‘ of English, Accent, Morpheme, Order.
Cambridge Massachusetts. Harvard University Press.

 

1965b. On defining the morpheme. Reprinted in
Bolinger I). 1“ Forms of English; Accent, Morpheme,
Order. Cambridge, Harvard University Press, 183-89.

 

BONA BAISI, JOSEPH (1957). IkanVI—Ngambo (Haya dictionary)‘
Rwamishenye, Bukoba Pallotine Fathers Hamburg Lahn
(Germany) Printing Press.

309

BOOIJ, GEERT. 1977. Dutch Morphology. A study of word
formation in Generative Grammar. The Peter de Ridder
Presse, Lisse.

1979. Semantic regularities 111 word. formation.
Linguistics 17, 985-1001.

 

BOTNE, ROBERT DALE OLSON. 1980. Phonological and
morphological aspects of verbal ‘ structures in
Kinyarwanda. M. A. Thesis, Northwestern University.

1981. On the nature of tense and aspect. Studies
in the semantics of temporal reference in English and
Kinyarwanda. Unpublished Ph.D. dissertation,
Northwestern University. Studies in African Linguistics
December 1986, 17:3:317.

 

1983. The semantics of tense in Kinyarwanda.
Studies in African Linguistics 14:3:235-265.

 

BRESNAN, JOAN and SAM A. MCHOMBO. 1987. Topic, Pronoun, and
Agreement in Chichewa. Language 63.4:741-782.

1982. Control and Complementation. LI 13.343-434.

 

BROWN, GILIAN and G. YULE. 1983. Discourse analysis.
Cambridge University Press. Cambridge.

BRYAN, MARGARET A. 1959. The Bantu languages of Africa.
London. Oxford University Press.

BUKOBA DIOCESE. 1987. Rumuli ("The torch"). A Haya Newspaper
V010 10-150

BUREAU OF STATISTICS. 1982. 1978 Population CenSus Volume
IV: A summary of selected statsitics. Ministry of
Planning and Economic Affairs. Dar-es-Salaam.

BYARUSHENGO, ERNEST RUGWA. 1975. An Examination of the
Segmental Phonology of Haya. Unpublished M. A. Thesis,
University of Dar-es-Salaam.

and SARAH TENENBAUM. 1976a. Agreement and word
order: A case for pragmatics in Haya. Berkely
Linguistic Society 2.89-99.

 

et al. 1976b. Tone, Accent, and Assertion. Haya
Studies in Bantu tonology, ed. by L. Hyman. Southern
California Occasional Papers in Linguistics 3: 183-205.

 

et al. 1977. Haya Grammatical Structure:
Phonology, Grammar, Discourse. Southern California
Occasional Papers in Linguistics No. 6. University of
California Los Angeles, California.

 

310

BYBEE, JOAN 'L. 1985. Morphology: a study of the relation
between meaning and form. Amsterdam, Philadelphia. John
Benjamin Publishing Company.

and C. In MODER. 1983. Morphological Classes as
natural categories. Language. 59.251-270.

 

CANNON, GERALD. 1986. "Blends in English word formation."
Linguistics: 24.4:725-754.

CARLSON, GREG and T . ROEPER. 1979 . Morphology and
subcategorization case auui the unmarked complex verb,
Lexical grammar ed. by T. Hoekstra and M. Mootgat Glot
3/4:123-64.

CARSTAIRS, ANDREW D. 1983. "Paradigm economy." Journal of
Linguistics 19:115-125.

1984» .Constraints (n1 allomorphy 1J1 inflection:
Bloomington: Indiana University Linguistic Club.

 

CHAFE, WALLACE L. 1970. Meaning and the structure or
language. The University of Chicago Press. Chicago.

CHAGAS, JEREMY E. 1977. The preprefix. Byarushengo, E. (ed.1
et al. Haya Grammatical Structure. SCOPIL 6.35-44.
University' of Southern California” Los Angeles,
California.

1977. The preprefix. Haya grammatical structure,
ed. by Byarushengo E. et al. SCOPIL 6:35-44.

 

CHOMSKY, NOAM. 1957. Syntactic structures. The Hague:
Mouton.

1965. Aspects of the Theory of Syntax. MA: MIT
Press.

 

and MORRIS HALLE. 1968. The Sound Pattern of
English. New York: Harper and Row.

 

1970. Remarks on Nominalization. A. Jacobs and P.
S. Rosembaun (eds).

 

1972 Studies on Semantics in Generative Grammar.
The Hague: Mouton.

 

1984. Lectures on Government and Binding. The Pisa
Lectures. Foris Publications. Holland/Cinnaminson
UOSOAO

 

CLARK, EVE V. and HERBERT H. CLARK. 1979. When nouns surface
as verbs. Language 55:767-811.

311

CLARK, M. 1978. A dynamic treatment of tone with special
attention to the tonal system of Igbo. Unpublished
doctoral dissertation. University of Massachusetts.

CLEMENS, GEORGE N. 1984. "Principles of Tone assignment in
Kikuyu." Clemens, G. N. and J. Goldsmith (eds).
Autosegmental Studies in Bantu. Tone. Foris
Publications. Dordrecht-Holland/Cinnaminson. U.S.A.

and JOHN, GOLDSMITH. 1984. Autosegmental studies
in Bantu Tone. Publications in African languages and
linguistics. Foris Publications. Dordrecht-
Holland/Cinnaminson. U.S.A.

 

COLE DESMOND T. 1955. An introduction to Tswana Grammar.
London. Longmans, Green and Co.

COMRIE, BERNARD. 1976. Aspect. Cambridge. London University
Press. Cambridge.

1981. Language universals and linguistic typology:
syntax and morphology. Oxford: Blackwell. 3.

 

1985. Tense. Cambridge University Press.
Cambridge: London.

 

CONTINI, ELLEN MORAVA. 1983. Tense and non tense in Swahili
Grammar. Semantic asymmetry between affirmative and
negative (Bantu, Africa). Unpublished Ph.D.
dissertation. Columbia University.

COWIE, A. P. 1981. The treatment of collocations and idioms
in learners dictionary. Applied Linguistics 2:223-35.

1982. Polysemy and the structure of lexical
fields. Nottingham Linguistic Circular. 11.2:51-64.

 

CRUSE, D. A. 1982. On lexical ambiguity. Nottingham
Linguistics Circular. 11.2:65-80.

1986. Lexical Semantics. London, ’Cambridge
University Press. Cambridge.

 

CUTLER, ANNE. 1980. Productivity in word formation. Papers
from the 16‘! Regional meeting of the Chicago
Linguistic Society, Chicago. 45-51.

1981. Degrees of transparency in word formation.
The Canadian journal of Linguistics 26:73-77.

 

DAHL, OSTEN. 1984. Temporal distance: remoteness
distinctions in tense-aspect systems, Explanations for
language universals, ed by B. Butterworth, B. Comrie
and O. Dahl. The Hague: Mouton.

312

DALGISH, GERALD M. 1977. "Past tense formation in OruHaya"
African Language/Language Africaines 3:78-92.

1978. The syntax and semantics of morpheme NI in
Kivunjo (Chaga). Kiswahili, Journal of the Institute of
Kiswahili Research. ed. by A. G. Gibbe, Pp. 42-46.
University of Dar-es-Sallam.

 

DE BLOIS, KEES. 1970. The augment in the Bantu languages.
Africana Linguistics IV, 85-165 Tervuren.

DELL, F., and E. O. SELKIRK. 1978. On a Morphologically

Governed Vowel Alternation in French, Recent
Transformation Studies in European Linguistics, ed. by
S. J. Keyser, Linguistic Inquiry' Monograph 3,

Cambridge, MA: MIT Press.

DI SEIULLO, ANNA-MARIA. 1987. On the definition of Word.
Cambridge, MA: MIT Press.

DIK, SIMON C. 1967. Some critical remarks on the treatment
of morphological structure in transformational
generative grammar. Lingua 18: 352-83.

DOKE, CLEMENT M. 1927. Textbook of Zulu Grammar. Longmans
Southern Africa (PTY) LTD. Cape Town.

1943. Outline grammar of Bantu. Mimeo.

 

1945. Bantu; modern grammatical, phonetical and
lexicographical studies since 1860. London. Dawnsons of
Pall Mall.

 

1954. The Southern Bantu Languages. London.
Longmans, Green and Co.

 

and MOFOKENG, S. M. 1957. Textbook of Southern
Sotho Grammar. London. Longmans, Green and Co.

 

DRESSLER, WOLFGANG, U. 1985. Morphophonology. The dynamics
of derivation. Ann Arbor. Karoma Press.

DRIEVER, DOROTHEA. 1975. Aspects of the case of grammar of
Mombasa Swahiii. Hamberger Philogische Studien 43.

DUBNOVE, ELLEN ZINOV’EVNA. 1984. The Rwanda Language. Nauka
Publishers. Central Department of Oriental literature.
Moscow.

DURANTI, A. 1979. Object clitic pronouns in Bantu and the
topicality hierarchy. Studies in African Linguistics
10.1:31-46.

313
EASTMAN, CAROL M. 1969a. A descriptive statement of

operative verbal extension in Swahili: African Language
Studies, X, SOAS.

1969b. The status of the reversive extension in
modern Kenya Coastal Swahili. Journal of African
Linguistics 8.29-39.

 

FARNSWHORTH, E. M. 1954. A Kamba Grammar. Cyclostyled, AIM.

FLEICHMAN, SUZANNE. 1977. Culture and Linguistic factors in
word formation. University of California Publications
in Linguistics. Berkley and Los Angeles. University of
California Press.

FORTUNE, G. S. J. 1955. An analytical Grammar of Shona.
London. Longmans, Green and Co.

FREIDEN, ROBERT. 1975. The analysis of passives; Language.

FROMKIN, VICTORIA A. 1978. Tone: a linguistic survey. New
York: Academic Press Inc.

GARDNER, JEAN LOUIS. 1985. Relational Grammar. The Catholic
University of America Press. Washington D.C.

GARY, JUDITH 0. AND EDWARD L. KEENAN. 1977. "On Collapsing
Grammatical Relations in Universal Grammar." Syntax and
Semantics: Relational Grammar, ed. by Cole, P. and J.
M. Sadock. New York. Academic Press. 85-120.

GAZDAR, GERALD. 6H: al. 1985. Generalized Phrase Structure
Grammar. Harvard University Press. Cambridge,
Massachusetts.

GIVON,.TALMY. 1972. Studies in ChiBemba and Bantu grammar.
Studies in African Linguistics. Supplement 3:174-179
and 203-216.

1976. Topic, pronoun, and grammatical agreement.
LI, CHARLES N (ed.1 Subject and topic. New York:
Academic Press.

 

GOLDSMITH, JOHN. 1976. Auto segmental phonology. Unpublished
MIT Ph.D. dissertation, reproduced by Indiana
University Linguistics Club.

GREENBERG, JOSEPH. H. 1954. A quantitative approach to
morphological typology of language. IJAL 26.178-194.

314

1963a. Some universals of grammar with particular
reference to the order of meaningful elements .
Universals of Human Language. Greenberg J. H. 73-113.
Cambridge, MA. MIT Press.

 

1963b. Universals of Language. 2nd ed. Cambridge,
MazMIT Press.

 

GRICE, PAUL H. 1975. Logic and Conversation. Speech acts
(Syntax and Semantics,3), eds. by Peter Cole and Jerry
L. Morgan, 41-58. New York: Academic Press.

GUNTER, RICHARD. 1972. English derivation. Journal of
Linguistics 8.1-19.

GUTHRIE, MALCOLM 1948. The classification of Bantu
languages. London. Oxford University Press.

1967. Comparative Bantu I-IV Farnborough: Gregg

 

Press.

1971-74. Comparative Bantu. Farnborough: Gregg

 

Press.

GUTT, ERNST AUGUST. 1985. A multi-tiered to silt’i verb
morphology. Studies 1J1 African Linguistics. 16.2.183-
222.

HALLE, MORRIS. 1973. Prolegomena to a theory of word
formation“ Linguistic inquiry, 4.1:3-16.

HARTMANN, RENE. (ed.1. 1983. Lexicography: principles and
practice. London and New York Academic Press.

HAWKINSON, ANNE and LARRY M. HYMAN. 1974. Hierarchies of
natural toch in Shona Studies in African Linguistics
5: 147-170.

HEDINGERN ROBERT. 1985. The Verb in .Akoose. Studies in
African Linguistics. 16.1.1-56.

HEINE, BERND. and ULRIKE CLAUDI. 1986. On the rise of
grammatical categories. Berlin Dietrich Reimer Verlag.

HENY, FRANK and B. RICHARDS. (eds.) 1983 Linguistic
categories: Auxiliaries and related puzzles. Vol. 1:
Categories, Vol. 2: The scope, order and distribution
of English auxiliary verbs. Syntheses language library
19 and 20 Dordrecht: Reidel.

HOBLEY, JUNE. 1965. Bassa verbal formations. Journal of West
African languages. 11:2.39-50.

315

HOCKETT, CHARLES. 1947. Problems of morphemic analysis.
Language 23:321-43.

HOOPER, JOAN B. 1976A. Word frequency in lexical diffusion
and source (fl? morphophonological change. Current
progress in historical linguistics, ed. by W. M.
Christie. Amsterdam North Holland 95-105.

1976B. An introduction to natural generative
phonology. New York: Academic Press.

 

HUDSON, GROVER. 1979. Is deep structure linear? Linear order
and Generative theory. ed. by Meisel, J. M. and M. D.
Pam. Philadelphia John Benjamins B. V.

1982. Review of Amharic Verb Morphology. M. L.
Bender and Hailu Fulas. Afroasiatic Linguistics 8, 4:3-
8.

 

1984. The Arabic double verb conspiracy and
morpheme invariance. Paper presented at the 15th
Conference on African Linguistics. Los Angeles.

 

1986. Arabic root and pattern morphology without
tiers. Journal of Linguistics 22:85-122 printed in
Great Britain University Press.

 

HUDSON, RICHARD. 1984. Word Grammar. New York. Basil
Blackwell Ltd.

HUTCHISON, JOHN P. 1981 The Kanuri: a reference grammar
University of Wisconsin.

HYMAN, LARY. 1976. Studies in Bantu tonology. SCOPIL 3.
University of Southern California. Los Angeles.

and A. DURANTI. 1982. On the object Relationship

 

in Bantu. Syntax and Semantics: Studies in
Transitivity. Ed. by Hopper, P. J. and S. A. Thomson.
15.217-259.

and E . R . BYARUSHENGO . 1 984 . A model of Haya
tonology. Autosegmental studies in Bantu tone, Ed. by
G. M. Clemens and J. Goldsmith 3.53-104.

 

ISHUMI, ABEL G. M. 1980. Kiziba: The cultural heritage of an
Old African Kingdom. Syracuse University.

JACKENDOFF, RAY. 1975. Morphological and semantic
regularities in the lexicon, Language. 51:639-71.

JANSEN, THEO J. and M. STRONG-JENSEN. 1984. Morphology is in
the Lexicon! Linguistic Inquiry 15:474-98.

316

JOHNSON, MARK. 1980. A semantic description of temporal
reference in Kikuyu verb. Studies in African
Linguistics 11:269-320.

JOOS, MARTIN. 1957. Readings in linguistics I. Chicago. The
University of Chicago Press.

JUILLAND, ALPHONSE and P. M. EDWARDS. 1971. The Rumanian
Verb system. The Hague: Mouton.

1978. Transformational and structural morphology.
About two rival approaches to the Rumanian Verb System.
Anma Libri and Co.

 

KASTOVSKY, D. 1977. Word formation, or: at the crossroads of
morphology, syntax, semantics and Lexicon. Folia
Linguistica 10:1-33.

KHAMISI, ABDU M. 1983. The stative derivational affix in
Swahili. Working Papers in Lingu1stics, Dept of
Linguistics University of Hawaii Vol. 15, 2:93-113.

KIDIMA, LUKOWA. 1987. Object agreement and topicality
hierarchies. Studies in African linguistics 18.2:175-
208.

KIEFER, FERENC. 1972. A propos derivational morphology.
Kiefer F. (ed.) Derivational processes. Stockholm,
KVAL, 42-59.

KIMENYI , ALEXANDRE . 1979 . Studies in Kinyarwanda . and Bantu
phonology. Carbondale (U.S.A.) and Edmonton (Canada).
Linguistic Research, Inc.

KIPARSHY, PAUL. 1982. Lexical morphology and phonology. In
Linguistics in the morning calm, ed. by I. S. Yang.
Seoul: Hanslin 3-91.

KISSEBERTH, CHARLES W. and M. I. ABASHEIKH. 1977. Object
Relationship in Chi-mwizni, a Bantu Language. Syntax
and Semantics: Grammatical Relations, ed. by Cole, P.
and J. M. Sadock. Academic Press. New York. 179-218.

KOOPMAN, HILDA. 1984. The syntax of verbs: From -Verb
Movement Rules in the Kru Languages to Universal
Grammar. Dordrecht-Holland/Cinnaminson- U.S.A. Foris
Publications.

KOUTSOUDAS, ,ANDREAS. 1972. ’The strict order Fallacy.’
Language 48:88-96.

1981. Constraints on linearization rules. Language
57.4:818-840.

 

317

KURYLOWICZ, JERZY. 1964. The inflectional categories of Indo
European. Heidelberg, C. Winter.

LACTOT, A. 1959. Evangeli entakatifu nebikorwa byentumwa
((Haya) ("The holly bible and the work of the
angels")). Marianum Press. Kisubi, Uganda. Marianum
Press.

LADEFOGED, PETER et al. 1972. Language in Uganda. London.
Oxford University Press.

LAPOINTE, STEBENG. 1983. A comparison of two recent theories
of agreement. Papers from the parasession on the
interplay of phonology, morphology, and syntax. Chicago
Linguistics Society. Chicago. 122-134.

LI, CHARLES N. (ed.). 1975. Word order and. word order
change. Austin: University of Texas Press.

LIEBER, ROCHELLE. 1980. On the organization of the lexicon.
Doctoral dissertation, Cambridge, MA: MIT Press.

LIGHTER, THEODORE M. 1975. The role of derivational
morphology in generative grammar. Language 51.617-38.

LWAMGIRA, F. X. 1949. Amakuru ga Kiziba na Abakma bamu
(History of Kiziba and its chiefs). Rwamishenye, Bukoba
Printers.

LYONS, JOHN. 1968. Introduction to theoretical linguistics:
New York: Cambridge University Press.

1977. Semantics. Vol 1 and 2. New York: Cambridge
University Press.

 

MACHOBANE, MALILO. 19XX Tense and Aspect in Sesotho.
Bloomington. Indiana University Linguistics Club.

MAGIER, DAVID. 1985. Temporal reference points and the
semantics of framing. Evidence from Marwari. Presented
at the 60‘“ meeting of the linguistic society of
America. Seattle.

MALKIEL, YAKOV. 1978. Derivational categories. Universals of
human language, ed. by J. H. Greenberg. Vol. 3 Word-
structure. Stanford. Stanford University Press 125-49.

MANCZAK, WITOLD. 1980. Laws of analogy. Historical
morphology. ed. by J. Fisiak. 283-6. The Hague: MoUton.

MARLE, J. V. 1985. On the paradigmatic dimension of
morphological creativity. Dordrecht-Holland/Cinnaminson
U.S.A. Foris publications.

318
MASAMBA, DAVID. P. B. 1976. Kiambisho-Awali (Pre-prefix)

Katika Kiruri (Pre-prefix in Kiruri). KISWAHILI.
Journal of the Institute of Kiswahili Research). 46/2:
23-34 0

MATHEWS, PETER H. 1970. Recent development in morphology.
New horizons in Linguistics ed. by Lyons J. 96-114.
Harmondsworth, Baltimore, Penguin Books.

1972. Inflectional morphology. A theoretical study
based on aspects of Latin Verb Conjugation. Cambridge:
Cambridge University Press.

 

1974. Morphology: an introduction to the theory of
word-structure. Cambridge: Cambridge University Press.

 

MAW, JOAN. 1985. Twende! A practical Swahili Course. London.
Oxford University Press.

MCCARTHY, JAMES D. 1979. Formal problems in semantic,
phonology and morphology. Doctoral dissertation,
Cambridge MAzMIT Press.

MEEUSSEN, ACHIEL EMIEL. 1976. Bantu grammatical
reconstructions, Africana Linguistics III:79-128.

MORRIS, HENRY FRANCIS and B. E. R. KIRWAN. 1957. A
Runyankore grammar. The Eagle Press. East African
Literature Bureau. Kampala. Uganda.

MOODY, MARVIN D. 1978. Some preliminaries to a theory of
morphology, Glossa, 12:16-38.

MORSE, MARY LYNN ALICE. 1976. A sketch of the phonology and
morphology' of Bobo (Upper 'Volta). Unpublished. Ph.D.
dissertation. Columbia University.

MOSHI, LIOBA. 1986. "Functional typology of "hi" in
Kivunjo." A paper presented at the 17th African
Linguistic Conference at Indiana University, April 2.5.

1988. Tense, aspect and other time markers "in
Kivunjo (Chaga). A peper presented at the 19th African
Linguistics Conference. Boston University.

 

MOULD, MARTIN. 1976. Comparative Grammar Reconstruction and
Language Subclassification: the North Victoria
Languages. Ph.D. Dissertation, U.C.L.A.

MUJWAHUZI, BERNARD. 1950. Obugorozi bwa Buhaya ((Haya).
("The leadership of Buhaya"). Bukoba. Tanzania.

NEWMAN, STANLEY S. 1948. English suffixation: a descriptive
approach. WORD 4:24-36.

319

NIDA, EUGENE A. 1946. Morphology. The descriptive analysis
of words. Ann Arbor. University of Michigan Press, 2“‘1
ed.

1948 . The identification of morphemes . Reprinted
in Readings in Linguistics 1. ed. by Joos. The
University of Chicago. 1957:255-271.

 

NURSE, DEREK. 1979. Description of sample Bantu languages of
Tanzania. African languages/Langues Africaines 5.1.
London: International African Institute.

NYMAN, M. 1987. Is the paradigm economy relevant? Journal of
Linguistics 23:251-262.

ODENDAL, F. F. 1963. Limitations of morphological processes.
A note. Lingua 12:220-225.

OKUNOR, V. 1967. Tone in the GA verb. Institute of African
Studies. Accra. University of Ghana.

PENNANEN, E. V. 1972. Current views of word-formation.
Newphilologische Mitteilungen 73:292-308.

PEREZ, CAROLYN H. 1986. Aspects of Complementation in three
Bantu Languages. Indiana University Linguistics Club.

1987. Successive cyclic movement in Kitharaka. A
paper presented at the 18th African Linguistic
conference in Montreal Canada.

 

PERKINS, REVERE D. 1980. The evolution of culture and
grammar. SUNY/Buffalo. Ph.D. Dissertation.

PESETSKY, DAVID. 1985 . Morphology 'and logical form.
Linguistic inquiry Vol. 16.2:193-246.

PHILLIPS, JUDITH W. 1983. A partial grammar of the Haitian
creole verb system: forms, functions and syntax.
SUNY/Buffalo Ph.D. Dissertation.

PLATT , D . 198 1 . The balance between regularity and
irregularity in verbal inflection. Inflectional
morphology: Introduction to the extended word-and-
paradigm theory , ed . by T . Thomas Flinders . U . C . L . A
Occasional papers, 4:41-71.

POLOME, EDGAR C. 1976. Swahili language Handbook. Washington
D.C. Center for applied Linguistics.

PORT, ROBERT F. 1982. Morphophonemics of Swahili Verb
suffixes. Studies in African Linguistics. 13:3:248-271.

320

RARDIN, R. B. 1975. Studies in derivational morphology.
Ph.D. dissertation MIT.

RASCHER, ANNA. 1958. Guide for learning the Ruhaya Language.
Bielefeld, Bethel Mission.

RAY, SIDNEY H. 1933. A grammar of Kiwai language, Fly Delta,
with a Kiwai vocabulary by E. B. Riley. Port Moresby,
Australia: Goverment Printer.

REICHENBACH, HANS. 1947. The tenses of verbs. Section 51.
Elements of symbolic logic. McMillan.

REYNODS, KARL. 1985. Swahili Morphology. Manuscript.

ROEPER‘, THOMAS and MUFFY E. A. SIEGEL. 1978. A lexical
transformation for Verbal Compounds. Linguistics
inquiry 9.199-260.

RUBANZA , YUNUS I . 1979 . The relationship between Ki swahili
and other African languages: the case of Kihaya.
Unpublished M.A. thesis. University of Dar-es-Salaam.

1988. Restrictions in the linear ordering of the
Haya verb morphemes. A paper presented at the 19th
African Linguistic Conference. Boston University.

 

SABIMANA, FIRMARD. 1986. The relational structure of the
Kirundi Verb. (Burundi) Unpublished Ph.D dissertation.
Indiana University Club.

SALONE, SUKARI. 1977. Conditionals. Haya grammatical
structure ed. E. Byarushengo et al. 149-159.
California, University of Southern Carolina.

1979. Typology of conditionals and conditionals in
Haya. Studies in African Linguistics. 10.1:65-80.

 

SAMPSON, GEOFFREY. 1979. The indivisibility of words.
Journal of Linguistics 15.39-47.

SANDERS, GEORGE. 1970. Constraints on constituent. ordering.
Papers in Linguistics 2.460-502.

1975a. Invariant ordering. The Hague: Mouton.

 

1975b. "On the explanation of constituent order
universals." C. N. LI 389-436.

 

SCALISE, SERGIO. 1986. Generative. Morphology. Foris
Publications. Dordrecht-Holland/Cinnaminson. U.S.A.
(Second Edition)

321
SCHADEBERG. THILD C. 1984. A sketch of Swahili. Morphology.

Dordrecht. Holland/Cinnaminson. U.S.A. Foris
Publications. '

SCHRAMM, GENE M. 1962. An outline of classical Arabic verb
structure. Language 38:360-75.

SCOTTON , CAROL MYERS . 1 96 7a . Semantic and syntact ic
subcategorization in Swahili causative verb shapes.
Journal of African Linguistics 6:249-267.

1967b. Aspects of the Swahili extended verb system
with special reference to some deep structure syntactic
and semantic restrictions. Unpublished Ph.D.
dissertation. University of Wisconsin. Madison.

 

SEITEL, PETER. 1970’s-1980’s. A collection of Haya Epic.
Mimeo.

 

1987. Haya epic and heroic poetry. Mimeo.

SELKIRK, ELISABETH O. 1982. The syntax of words. Cambridge,
MA: MIT Press. .

SIEGEL, DOROTHY. 1974. Topics in English morphology.
Unpublished doctoral dissertation, Cambridge, MA: MIT
Press.

1977. The adjacency condition and the theory of
morphology. Proceedings of NELS VIII, ed. by Mark, J.
Stein. 189-197.

 

SIMOES, MARIA C. P. and CAROL STOEL-GAMMON. 1979. The
acquisition of inflections in Portuguese: a study of

the development of person markers on verbs. Journal of
child language 6.53-67.

SINGLER, JOHN VICTOR. 1985. The status of lexical
association and the obligatory contour principle in the
analysis of tone languages. African Linguistics. SSLS
6:491-507.

SKOUSEN, ROYAL. 1974. An explanation theory of morphology.
Chicago Linguistic Society. 318-27.

STEPHANY, URSULA. 1981. Verbal Grammar in Modern Greek early
child language. P. Dale and D. Ingram (eds) Child
language: an international perspective. Baltimore:
University Park Press. '

STRAUSS, STEVEN L. 1980. How abstract is English morphology?
Glossa 14:89-112.

322

STUMP, GREGORY T. 1984. Agreement vs. inflection in Breton.
Natural language and Linguistic Theory 2.289-348.

TALMY, LEONARD. 1985. Lexicalization patterns: semantic
structure in lexical forms. Language typology and
syntactic description, III: Grammatical categories and
lexicon, ed. by Timothy Shopen, 57-149 Cambridge:
University Press.

TAYLOR, CHARLES. 1985. Croom Helm Descriptive Grammars:
Nkore-Kiga. Croom Helm. London.

THOMAS-FLINDERS, TRACY. 1981. Evidence for inflection as a
phonological process. Inflectional morphology:
Introduction to the extended word-and-paradigm theory
ed. by T. Thomas-Flinders. U.C.L.A. Occasional papers.
4:72-83.

TROMMELEN, MIEKE. 1986. Dutch Morphology: Evidence for the
right hand head rule. Linguistic inquiry 17.1:147-170.

UBELS, EDWARD H. 1983. Mood and aspect in Karang. Studies in
African Linguistic 14.1:47-70.

ULTAN, R. 1978. The nature of future tenses. Universals of
human language, ed. by J. Greenberg. California.
Stanford University Press. 3:81083.

VAIL, HAZEN LEROY. 1972. Aspects of the Tumbaku Verg.
Unpublished Ph.D. dissertation University of Wisconsin.
Madison.

VENDLER, ZENO. 1967. Verbs and times. Chapter 4 of Z.
Vendler. Linguistics in Philosophy. Ithaca, New York:
Cornell University Press. 97-121.

VENNEMANN , THEO . 1 97 3 . Explanation in syntax . J . Kimball
(ed.). Syntax and. Semantics II. New ‘York: Academic
Press. ' ,

VINCENT , NIGEL . 1 979 . Word order and grammatical theory .
Linear Order and Generative Theory, ed. by Meisel, J.
M. and M. D. Pam. Philadelphia: John Benjamins.

VITALE, ANTHONY J. 1981. Swahili Syntax. Foris Publications
Cinnaminson. U.S.A.

VOEGELIN, C. F. and F. M. VOEGELIN. 1966. Index to the
languaes of the world . Anthropological linguistics ,‘ 6
and 7.

WALKER, D. 1986. Review. Morphology: A study of the relation
between. meaning and form, by’ Bybee, J. Journal of
linguistics 22:493-496.

323
WELMERS, WILLIAM E. 1970. The derivation of Igbo verb bases.
Studies in African Linguistics 1:49-59.

1973. African Language Structures. Berkley and Los
Angeles: University of California Press.

 

WHEELER, CATHY J. & SCHUMSKY, DONALD A. 1980. The morpheme
boundaries of some English derivational suffixes.
Glossa. 14:3-340

WHORF, BENJAMIN LEE. 1936. The punctual and segmentative
aspects of verbs in Hopi. Language 12:127-31.

1950. An American Indian model of the universe.
International Journal of American Linguistics, 16.

 

WILLIAMS, EDWINS. 1981a. On the notions "lexically related"
and ’head of a word.’ Linguistic Inquiry 12:245-274.

1981b. Argument structure and morphology.
Linguistic review 1:81-114.

 

WILT, TIMOTHY L. 1988. Bukavu Swahili. A sociolinguistic
study of language change. Ph.D. Dissertation Department
of Linguistics: Michigan State University.

YULE, GEORGE. 1985. The Study of Language. An introduction:
Cambridge University Press. Cambridge.

ZAWAWI, SHABRIFA M. 1979. Loan words and their effect on the
classification of Swahili nominals. E. J. Brill,
Leiden: The Netherlands.

ZIMMER, KARL E. 1964. Affixical negation in English and
other languages . Supplement to Word 20 . 2 Monograph 5 .
N.Y. Linguistic circle of New York.

ZWANEMBURG, W. 1980. Form and meaning in morphology.
Quaderni di Semantica. 2:327-38.

ZWICKY, ARNOLD M. and R. E. WALLACE (eds.) 1984. Working
papers in linguistics. 29. The Ohio State University.
Papers on Morphology.