SOME ASPECTS OF THE PREDICATE PHRASE IN GEGBE
by
Ed!n~ Elemawuss!
Ayaba Jandah
Submitted to the FaCUlty of the Gr~duate School
in partial
fulfillment of the requirements for
the deQree
in the
_
-
Doctor
-
-
--
-
-
of
--
-
Philosuohv
•
J
Ueoartment of LinQuistics
--
-,- -
-
-
Indiana Unlversity
October 1980
FACULTE DES LETTRES ET SCIENCES
HUMAINES
UNIVERSITE
DU
BENIN
LOME - T 0 G 0

---

· ~
Edina E. A. Jondoh, i980
All rights reserved.
- i i -=

---

This thesis. by Edina E. A. Jondoh has been accepted by
the faculty of the Graduate School. Indiana University in
partial fulfillment of the requirements for the degree of Doctor
of Philosophy.
Committee
Dr. Linda J. Schwartz
Dr. Harry L. Gradman
Dr. Carleton T. Hodge
Dr. Fred W. Householder
Dr. Ruth Stone
- i i i -

---

ACKNOWLEDGEMENTS
First of all,
I would like to extend my sincere apprecia-
tion and gratitude to the members of my committee for their
invaluable help in making this thesis a reality:
Professor
Linda J. Schwartz, chairman of the committee, for her patient
listening, her valuable discussion and suggestions throughout
these months of search for an appropriate solution to each of
the problems dealt with in this work: Profes'sor Fred W. House-
holder for his valuable comments and help in clarifying many
of the theoretical aspects involved and also for his help in
many administrative matters throughout my career as a student
at Indiana University; Professor Carleton T. Hodge for his
continual support and very helpful comments and advice; Professor
Harry L. Gradman for his help and Professor Ruth S. Stone for
her valuable and helpful comments.
Needless to say that any
error in the present thesis is my own.
I would also like to
thank Professor Charles Bird for his help and support.
I sincerely wish to thank my parents and other members of
my family without whose sustained moral support and care for
my children, it would not have been possible for me to pursue
, .
my studies toward this degree.
Special thanks to my children,
Marie-Claude and Regis for their understanding and letters of
encouragement all the time I was away.
I
am grateful to my friends and other members of the Lin-
gUistics Department who, one way or another, have helped me
- iv-

---

-v-
during these past few years.
I acknowledge the financial support of the African-American
Institute and the Internaticnal Students Services at Indiana
University which made i t possible for me to study in this country.
At last, but not the least,
I wish to express my appreciation
to my typist,
Augusta Jesus,
for her patience and understanding
in laboring through my manuscripts and dealing with a language
rather unfamiliar to her, not only for this thesis but for other
papers and reports as well.
,
, .

---

Resume'.
smH. ASPECTS er THt: P,iEOIC/\\TE F'Hfil>,':;E IN GD;BE
( CERTAINS ASPECTS DE LE PHR."SE PREDICATIVE t.I,
La presente these a pour but d'analyser differants aspects syntaxiques de la
phrase predicative en Gegbe ,
une langue de 1 'hfrique
de l'Ouest appartenant
au group e KloJa.
Nous avons essayc de det eDi1in er la Lo<;on de class er 1 es di ffe-
rentes composantes de la phrase predicative et les generalisations que l'on
peut faire au suj et de sa structure. Le cadre th~orique est ?rincipalement celui
de la grammaire
Generative transfonnationelle.
Les di fferents aspects considl~-
res sont les sui'itsnts : les auxiliaires, le groupe verbal(GV), la
serialisation
Verbale, les phrases. predicatives modi ficatrices o
En nous basant sur des donnes inten,es a la langue telles que la position
des elements dans la phrase predicative, leur mouvement ou non-rnouvement et leur
rOle syntaxique et aussi sur les universaux de la langue relati fs a la nature du
Aux et du GV nous avons tirales conclusions suivantes :
(1) i l
est psssible
d'instituer une categorie Aux pour la langue comportant des elements de temps,
d'aspect et de mode,
(2) une categorie GV aussi peut ~tre determinee comportant
le verbe principal, les Groupes Nominaux (GN) objet\\O,
d'autre verbes de modifica-
tion et des adverbes; i l a ate demontre que le Aux ne peut ~tre incorpore au GV;
(3) l'analyse de la senalisation verbale a montre que de telles constructions
verbes
n'ont pas la structure Aux + V dans laquelle certain~rsont considere Comme des
auxiliaires et les autres les verbes principaux, mdis ce sont plut6t des construc-
tions comportant des verbes principaux en Berie dirivees
de
phrases coordonnees;

---

2.
•
J
(4)
en Gegbe, certaines phrases predicatives peuven t ~tre
utilisees pour
quali fisr les substdntives:
le noyau
DC'
C8[C
phrases predic"tives peut ~tre
un verb e,
un nom ou un nGmin':::il
derive d'un .:Jojcctif ou d'un ver::8;
quelques-
etre
uns des morphemes derives des v srb es p euVent
consideres syn ta :dquem en t
comme
x
des adj ecti fs et des noms,

---

ABSTRACT
~
v
SOME ASPECTS OF THE PREDICATE PHRASE IN GEGBE
The purpose of this dissertation is to present an anal-
ysis of different syntactic aspects of the predicate phrase
,
in Gegbe, a West African language of the Kwa group.
An at-
tempt is made to determine how the components of the predica-
te phrase can be categorized and what generalizations can be
made about its structure.
The theoretical framework used is
primarily that of the Transformational Generative Grammar.
The different aspects considered are: auxiliaries, the verb
phrase, verb serialization, modifying predicate phrases.
Drawing from language internal evidence such as the position
of the constituents in the predicate phrase, their movement
or lack of movement and their syntactic role, and al~o from
language universals dealing with the constituent nature of
the Aux and the VP, the following conclusions have been
reached: 1)
a general constituent category AUX can be set up
for the language, including items used to convey tense
(time),
aspect or mood; 2)
a constituent category VP can also be set
up for the language inclUding the main verb, the object NP's
and other.~odifying verbs and some expressions of manner; i t
has been shown that the Aux cannot be part of the VP;
3)
in
the analysis of serial verb constructions,
i t has been shown
that such constructions are not of the form AUX + V where some
of the verbs are considered as auxiliaries and the others as
main verbs, but rather, they are constructions with true
-vi-

---

-vii-
verbs occurring in series, derived from underlying coordinated
,
sentences~ 4)
in Gegbe, some predicate phrases attribute a
quality to substantives: the nucleus of these modifying pre-
dicate phrases can be a verb, a noun,or a nominal derived
from an adjective or a verb~ some of the morphemes derived
from verbs can be classified both as adjectives and as nouns
syntactically.
Dr. Carleton
Dr.
~..£~--
Dr. Ruth Stone

---

Abbreviation with Alphabetic Letters
Adj, adj.
adjective
adjal.
adjectival
ADV
adverb or adverb phrase
ADV
adverb of manner
m
adverb of place
adverb of time
aspect
Art
article
AUX
category of auxiliaries
Aux
aux
auxiliary
CONJ, conj.
conjunction
camp.
comparative
Dem
: . demon strati ve
DET, det.
determiner
d-object
direct object
emp.
emphasis
fut.
future
hab.
habitual
I-object
indirect object
M, Mod.
modal
neg.
negation
nom.
nominal
N
noun
NP
noun phrase
-viii-

---

-ix-
Num
numeral
obj.
: object
part.
particle
pI.
plural
Poss, poss.
possessive
Post, post.
postposition
PP
prepositional phrase; postpositional phrase;
prepositional-postpositional phrase
Prep, prep.
preposition
pre/post
prepositional/postpositional
QUANT
quantifier
quest.
question
reI.
: relative
s
sentence
subj.
subject
sing.
singular
sov
subject object verb
(order)
svo
subject verb object
(order)
v
verb
VP
verb phrase
vso
verb subject object (order)

---

TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS • • • • • • • • • • . • • • • • • • • . . . . . . . • . • • . . • • . . . • • • ~ v
ABSTRACT •••••••••••.•••••• ~ •.••••.••••••••••••.••••.••• • vi
Abbreviations with Alphabetic Letters
.••••••...••.•.•.•• vii
INTRODUCTION • • • . • • • . • • • • . • • . . . . . . . • . . . . • • . • • . . • • . • . • • • • • 1
1.
The Prob] em ••..•.••••'. . . . . . . . . • • • . . . . . . '" . . . . . . . . . . . . 1
2.
General Outline of the Thesis ••••••••••••••••••••••• 2
,
CHAPTER I
GENERAL FACTS ABOUT GEGBE
•••••••••.•.•••...• 4
1.
History and Relationship to Other Languages
••.•.•.•• 4
2.
Relevant Typological Facts
•••.••••..•..••••••.•••••• 7
3.
Sound System and r-lorphology ••••.••••.••••••••••••.•• 9
FOOTNOTES TO CHAPTER I
•••••••.•••••••••.••..•••••••••••• J.2
CHAPTER II
WORD ,ORDER,
_
v
AUXILIARIES AND THE VERB PHRASE
IN GEGBE
• • • • . • • • • . • . . . • . • . • . • . • . . • • . • . • . • • • • 1·3
1.
Word Order ••••••••••.••••.••.•••••••••••..•••••••.•• 13
1.1. Introduction ••••.•••.••••.•••••..•.•••••••••••• 13
1.2. Determining Wor~ Order •••.••.••••••••.•••.••••• 13
1.3.. Word Order in Gegbe
•••••••...•••••••••••.••.•.• 14
2.
The Category AUX •••••.••••••••..•...•••••••••••••.•• 21
2.1. A Definition of t-he Category AUX ••••••••••••••. 21
2.2. Auxiliaries in Gegb~ ••••.•••••••••.•.••••.•••.. 26
2.2.1. General Remarks on Predicative Construct-
,
•
ions in Gegb~ •••..•••••.••••••••••••.••• 27
2.2.2. The Different Constituents of the Catego-
ries of Auxiliaries ••••••••••••••••.••••• 29
2.2.2.1.
Tense
.•.•.••••.••.•••.•••..••• 29
2.2.2.2.
Modali ty . • • • . . • . . • • • • • • . • • • . . . 32
2.2.2.3.
Aspect ••••..••.••••..•...••••• 37
-x-

---

-xi-
,
2.2.3.
A Category AUX for Gegb~ •••..••.••.••.• 51
\\
3.
The VP Category in Gegbe . . . . . . . . . • . . . • • . . . . • . . • . • • • • 54
3.1. Introduction ••....••.. (' . . • . • . • . . . . . • . . • . . . . • • ••
54
3.2. Constituency Test for G~gbe Predicate Phrase •.•
57
.
\\
3.3. AUX and VP 1n G~gbe • • . • . . • . . . . . . . . • . . . . . . . . . . • .
63
4.
Conclusion • . . . . • • . • • • . . . . . . . . . . . . . . . . . . . . . . . . . . . • • • .
70
FOOTNOTES TO CHAPTER I I
•••.•.•...•.•.•...•...••..•...•.•
73
l.
v
CHAPTER III
SERIAL VERB CONSTRUCTIONS IN GEGBE ••...•..•
80
1.
Introduction.... . . . • . .•• . . .. . .• .. . . . . . . . .. . .• .•.. .•.
80
2.
The Analysi.s of Serial Verb Constructions •••.......•
83
2.1. The Facts of Gegbe . • . . . . . . . . . . . . . . . . . . . . ~
84
2.1.1.
Sentences with the Same Subject and One
AUX • • . . . . . . . . . . . . . . . . • . . . . . . . . • . . . • . . . .
86
2.1.2.
Sentences with Different AUX's but the
Same Subject . • • . . . • • • • . . . . . . • . . . . . . • •
89
2.1.3.
Sentences with Both Surface Subject and
Object of Initial Sentence as Possible
Deep Subject of Non-initial Predicate
Phra.ses. •. • . . . .. . • • . .. . . . . .. . ... . • .•••.•
91
2.1.4.
Sentences with only the Object of the
Initial Predicate Phrase as Possible
Deep Subject of Non-initial Predicate
Phrases • • . . • . • • . . • • . . . . • . . . . • . • • . . . . . • .
94
2.1.5.
Comparatives and Superlatives . . . . . . • . • •
95
2.1.6.
Spli t
Verbs • • . . . . . . . . . . . . . . . . . . . . . . . • . •
98
2.1.7.
Modifying Verbs .•...••.•....•..••••••••
100
2.1.8.
Verbs as Prepositions .....•.•.•...•••••
102
3.
Derivation of Serial Verb Constructions in Gegbe •••..
105
4.
Conclusion .. . .. . . .... . . .. .. . . . . . . . . . . . . . . . . . . . . . . . . . 116
FOOTNOTES r6 CHAPTER III •••.•••..•..•.••••..•........... 119
CHAPTER IV
MQDIFYING PREDICATIVE CONSTRUCTIONS IN
G~GB!; ••..••.••....•...........•.••........•• 121
1.
Introduction • • . • • • . . • . • . • . . . . . . . • . . • . . . . . . . . . . . • • • • •
121

---

-xii-
2.
Different Types of Modifying Predicative Con-
structions . • . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • 123
2.1. Verbal Constructions • • . . . . • • . . . . . • . . . . • . . . . . • • . • 124
2.2. Other Types of Modifying Predicative Con-
structions • . . • . • . • • . . . . • . . . • . • . • • . . • . . . . . . . • • . • . 128
3.
Conclusion • • • . . . . . . . . . . . • . • . . . . . . • . . . • . . • . • . . . • . • . . . . 139
FOOTNOTES TO CHAPTER IV ••.•••.•...•.••..•••.•••...•...••• 146
EPILOGUE • • . . . • • • • • . • . • . • • • . • . . . . . • . . . • . . . . . • . . . . . • . . . . . . • 148
BIBLIOGRAPHY . . . . . . . . . ....... . . . . . . . . . .. . . . ... .. . . .. . . . . . . l52
,
~
.

---

INTRODUCTION
1.
The Problem
One of the main problems the linguist has to deal with
when describing the grammar of a language is that of determin-
ing what the different grammatical categories of the language
are.
This problem is especially important when examining a
language about which very little has been written.
One such
,
language is Gegbe, also known as Mina, a language belonging
to the Kwa group of languages spoken in West-Africa.
The main goal of the present thesis is to present an ana-
lysis of different syntactic aspects of the predicate phrase
,
in G~gb~.
The different questions I will be addressing are
as follows:
- Is there a category "Auxiliary" in
"
G~gb~ dlstinct from
a category of main verbs?
- Can a category Verb Phrase be set up for the language?
If so, what are the different constituents of such a category?
- What is the nature of serial verb constructions in G~gb~?
Are such constructions distinct from AUX + Verb constructions?
Can they be derived from other more basic structures such as
..I
coordinated' verb phrases?
- What is the syntactic nature of the different grammatic-
al elements that can be the nuclei of the predicate phrase in
modifying predicative constructions?
Are predicate adjectives
,
and predicate nominals distinct categories in ~gb~?
-1-

---

-2-
The theoretical framework adopted in this study is to a
great extent that of the Transformational Generative Grammar.
The data used is based on my knowledge of the language as a na-
tive speaker.
2.
General Outline of the Thesis
Chapter one presents general information about the histo-
rical background
of the language and its structure.
Section 1. in Chapter II gives the different word order
types found in the language.
The criteria used are based pri-
marilyon Greenberg's universals of language.
Section 2. pre-
sents an analysis of auxiliaries as distinct from main verbs.
The general conclusion is that a constituent category AUX can
be set up for the language.
Support for such a position is
drawn from Susan Steele's studies on the AUX as a universal ca-
tegory of grammar and from language internal evidence.
Section
3. presents an anlysis of the Verb Phrase.
Support drawn from
language internal evidence and Arthur Schwartz'
studies on the
category VP in different languages is used to set up a category
VP for the language, excluding the AUX.
Chapter III presents a brief analysis of serial verb con-
structions.
It is shown that such constructions are not of the
Aux + Verb type but are rather derived from underlying coordinat-
ed sentences
Chapter IV is a study of modifying predicative constructions.

---

-3-
The main conclusion is that there are three different types of
nuclei for the predicate phrases: a verb, a noun, or a nominal
derived from a verb or an adjective.
Finally, the epilogue gives a summary of the different find-
ings of the thesis.

---

CHAPTER I
\\
GENERAL FACTS ABOUT
- -
GEGBE
1.
History and Relationship to Other Languages
G~g~, also referred to as Ge or Mina, belongs to the Kwa
group of languages in the Niger Congo family according to
!
Greenberg's classification of African languages
(1955).
It
!I
is primarily spoken by some 62000 people in the southern re-
I~
I
gions of Togo along the coast and in southwestern Dahomey,
I
l
or Republic of Benin.
The historical linguists who have
!
analyzed the language to any extent do not seem to agree on
I
its origin.
According to L. Homburger (1949),
the present
!.I
day Ge (people)
originated from the Ga in Ghana.
They fled
i
i
their hometown of Elmina near Accra in the sixteenth century
to escape from the warfaring Fanti and Akwapims and settled in
the town of Little Popo (present day Agbodrafo)
and the Aneho
region in Togo, among Ewe speaking people.
Later on, the ad-
vent of the slave trade caused more of the Ga people to mi-
grate eastward and settle as far away as the area south of
the Republic of Benin.
Schroeder
(1936),
Seidel
(1902)
Westermann
(1935)
and Dalby (1977)
all agree in consider-
\\
.
ing
Gegbe a dialect of Ewe, a language widely spoken in
Southern Togo and South-eastern Ghana.
In Voegelin's Anthro-
2
pological Linguistics
(1964), G~gbe is classified as a
-4-

---

\\ '
-5-
dialect of Ga and is called Ge
(Gain,
Amina)
and also as a
dialect of Ewe and is again refered to as Ge
(Guingbe, Gaingbe
or Mina).
His reason for this double classification is the
,
fact that even though Gegbe is historically related to Ga,
its grammatical structure is closer to that of Ewe than to
the structure of Ga.
Gabriel K.
Johnson,
a native speaker
,
of Gegbe does not consider the language a dialect of Ewe.
As he states in an article in Etudes Togolaises
(1965, Tl,
,
Vol I), Ge{gb~) along with its neighbors or "cousins" Ewe,
Fe and Gu are languages with the same status, descending from
the same "mother-language".
Finally, Remy Bole-Richard in an
article in Afrigue et language
(N.
5, 1976)
says the following
about the origin of the Ge people and their language:
The Ge
p~ople are descendants of two groups of migrants from Ghana,
the Ga from Accra
(1663 and 1733)
and the Fanti from Elmina
(1720).
The Ga founded the kingdom of Glidji and the Fanti
(or Ane)
the kingdom of Aneho.
These two groups intermarried
and became one ethnic group.
They also mingled with the
natives, most likely the WatSi and adopted their language,
a dialect of Ewe.
,
In spite of the fact that Gegbe as i t is spoken today
is much closer to Ewe than i t is to Ga or Fanti
(Ewe and
G~gbe are mutually intelligible while Ga or Fanti and G~gbe
are not), I believe along with Johnson and Bole-Richard that

---

-6-
G~gbe is not a dialect of Ewe but a language made up of Ga,
Fanti and Ewe, with a predominating Ewe influence, and I make
this assertion on historical as well as grammatical grounds.
According to oral tradition and the few recorded stories
about the history of the people, the Ge people today consider
Elmina and Accra as their place of origin.
Many of their
traditional customs are similar to those of the Ga people and
unlike the customs of the Ewe-speaking people.
This is further
supported by the fact that the term Mina used in refering to
the language derives from the name of a fort, Fort Elmina,
which had given its name to the town in which it was: situated.
That town. as mentioned earlier, was the place of origin of
i
most of the Ge people in Togo.
This fort played an important
Ir
role during the period of slave trade.
The people call them-
selves the Ge and refer to their language as G~gbe or the lan-
guage of the Ge (gbe = language).
f
,
Beside Ewe, Gegbe had also been influenced by the lan-
I
guage of the Adangbe people, also from southeastern Ghana,
I
who had fled their homeland under the pressure of slave trade.
I
,
I
The only remnants of Ga and Fanti to be found in Gegbe today
I
are limited to a few lexical items.
Following is a broad
informal description of some of the syntactic characteristics
of the language.
I

---

-7-
2.
Relevant Typological Facts
According to the principles laid down in Greenberg's
(1963)
Language Universals, G~gbe can be said to have an SVO
word order since subject, verb and object always occur in that
order in a simple non-emphatic sentence.
However there are
also characteristics pertaining to an SOV word order to be
found in the language.
A more detailed discussion of word
.l.
y
•
order in Gegbe w~ll be found in section 1 of Chapter II.
,
Suffice i t to say here that, as example (1.1)
shows,
in Gegbe
simple sentences the object follows the verb: the verb
pre-
cedes the adverbs and prepositional phrases:
intentional
3
verbs and modal auxiliaries . precede the infinitive: tense
always precedes the verb and aspect markers, except for the
4
habi tual marker na,
also precede the verb.
(1.1)
aft
I~
t ~~
.,
n;)
pie
921 I f
n~
Afi
future can
progressive buy
gali
for
subject tense modal aspect
verb d-object prep.
,
,
m
le
,
,
aSI
d?o~ti
9 be
me
at
market
Monday on
I-object prep. noun
noun
postposition
'Afi will be able to buy gali for me at the market on
Monday. '
,
GegbE; uses mostly pos·tposi tions and very few prepositions.
Noun modifiers generally follow the noun.
The order of cons-
tituents in the Noun Phrase is as follows:
Noun - Adjective - Possessive - Numeral - Demonstrative
(or Definite Article)
- Plural marker.
Cf.
(1.2) below

---

-8-
,
...
.e
-,
(1. 2)
awu
d3£
ne
ve
wa
w6
dress 'red
my
two
those plural
N.
Adj. Poss. Num.
Dem.
PI.,
'Those two red dresses of mine.
Wi th the first and second person pronouns the possessive may
precede and/or follow the noun.
With the other persons and
with nouns i t always precedes the noun.
Cf.
(1.3).
, .
(1. 3)
a.
ape awu
a
or
ape aw u
ne a
,my
dress the
my
dress my the
or
awu
ne -
a
I
dress my the
'My dress.'
f
b.
marie-b~
awu
a
Marie-Poss dress-the
!
'Marie's dress.
I
I
c. *awu marie
I
In some of the syntactic processes or structures which occur
rI
in the language, the following word order of constituents prevails.
,
The negative is formed by placing mY before the item being
negated and adding the particle Q at the end of the sentence.
Cf.
(1. 4) .
,
(1.
...
4)
a.
m&
aff y~
yl
0
Neg. Afi emphasis go part.
'It is not Afi who went'
b.
af r
,
, ,
mu
yl
0
Afi neg. go part,
'Afi did not go.
aff m~
pie ga I r ,
c.
o.
Afi neg. buy gali post
'Afi did not by gali.'
,
d.
9 i:li r (y ~)
aH pie ,
mu
0
Neg. gali
(emp)
Afi buy post.
'It is not gali that Afi bought. '

---

-9-
The interrogative is formed by adding ~ at the end of the
,5
sentence for yes / no questions and Q
for wh-questions.
In
the latter case also,
the wh-word occurs at the beginning of
the clause.
Cf.
(1.5).
,
(1. 5)
a.
o
y'i
a?
you went quest.
'Did you go?'
,
b.
f i kt
o
y'i
o?
where you went quest.
'Where did you go?'
In relative formations,
the antecedent NP is immediately
followed by ke and the particle a is added at the end of the
clause.
Cf.
(1.6).
(1. 6)
aw~ ..d 3i ke
mpie
a
h~
t If
dress red that I bought the
took money
'The red dress that I bought was expensive. '
In comparative sentences, the comparative always precedes the
standard:
(1.7)
k~d?6 1010
w~
paul
Kodj6 be-big than Paul
Compar.
Standard
'Kodj6 is bigger than Paul.'
3.
Sound System and Morphology
The following phonemic systems are found in
>-
"
Gegbe for vowels,
consonailt.s and tones:
Vowel system
oral
Nasal
i
u
i
u
e
0
E
:>
e
5
a
a

---

-10-
Tonal system
,
..
high
high-low
...
mid
low-high
low
Consonant system
Stops
Fricatives
p
voiceless bilabial
f
voiceless labio-dental
b
voiced bilabial
v voiced labio-dental
t
voiceless alveolar
5
voiceless alveolar
d
voiced alveolar
z voiced alveolar
d
retroflex voiced alveolar
h voiceless laryngeal
if voiceless palatal
d,
x voiceless velar
voiced palatal
y voiced velar
k
voiceless velar
9
voiced velar
Nasals
kp voiceless labio-velar
gb voiced labio-velar
m bilabial
n alveolar
n palatal
I)
velar
Glides
Liquids
y
palatal
lateral
w
bilabial
r palatal t r i l l
1.
...
In Gegbe there are two types of consonant clusters:
..,(~Ttr-'-\\'·
·....\\v\\,
- " ' , ...•
1) stops, fricatives or nasals fo
p~~a:tiqUid;
2) Il)l followed by Itl, Idl, Ik/, f~}:g/\\j~b/. In
'c
\\
, - "
I " :
the clusters of consonant + liquid, Ir -~,~~~~.~1 with the
palatals and III with
the other consoncfu€s·.OCl skiff/is never
-,... _~I'·
used in clusters.
As far as morphology is concerned, there is a great amount
of vowel assimilation in G~gbe.
Sometimes certain consonants
are deleted in the process.
Cf.
(1.8) below.
!
r
[

---

-11-
(1.8)
a.
mf sa
e
-+
m f see
we sold i t
b.
fikt
y~
0;
-+
fft;
where emph. quest.
'Where? '
Most nouns have an initial vowel ~ or ~ which deletes when
preceded by another word.
The ~ deletes more readily than the
\\
g.
\\
(1.9)
a.
eta
; afi
head
mouse
\\
b.
~
gba
ta
he broke head
'He knocked his head. '
c.
~
\\
wu
afi
he killed mouse
'He killed a mouse'
.
but d.
ete +
"
aSI
-+
etesT
I
,
yam
market
yam D1arket
f!
Sometimes when the vowel is deleted, the tone remains and be-
!
comes attached to the preceding word.
This occurs most fre-
I
r
I
~
"
•
,
(1.10)
a.
pie
es:>
he bought horse
r A '
b.
~
kp6 ed
-+
ekp:>s:>
I
he saw horse
'He saw a horse.'
!
\\
In this thesis, the orthography used to transcribe the Gegbe
I
examples uses the same symbols as those listed above for the
~r
sound system of the language.
However, for practical reasons,
II,
the language itself will be transcribed G~gbe instead of G~gbe.
fi,
It!!

---

FOOTNOTES TO CHAPTER I
IVoegelin, C.F .• and F.M. Voege1in, 1964.
Anthropological
Linguistics, Vol.
6, No.
15, p.
28.
,
2Ibid •
I
3whether these are auxiliaries or main verbs will be dis-
II
cussed in section 2 of Chapter 11.
t
4
!
Na always follows the main verb in the sentence.
More
I
,
will be said about this. aspect marker later.
I
5This Q is probably the same word as the one used above
,
}
as a negative marker.
This process of using the negation
!
marker both for negatives and interrogatives is found in other
r
languages too; for example, Chinese for some yes/no questions.
[
I[
III
i
!
-12-

---

CHAPTER II
WORD ORDER. AUXILIARIES AND
THE VERB PHRASE IN G~GBE
1.
Word Order*
1.1
Introduction
According to studies conducted in the classification of
African languages, Kwa languages are said to have descended
from Proto-Niger-Congo.
Talmy Givon in his article "Serial
Verbs and Syntactic Change:
Niger-Congo"
(1975)
and Larry
Hyroan in his "On the Change from SOV to SVO:
Evidence from
Niger-Congo"
(1975)
have both tried to prove that Proto-Niger-
Congo had.an SOV word order and that many of the daughter lan-
guages are in a transitional state,l moving from an SOV word
order to a SVO word order, some having effected this change
more completely than others.
The main purpose of the
present section is to try to determine what the word order
,
2
patterns in Gegbe are.
1.2
Determining Word Order
Studies conducted to determine the word order type pre-
valent in a language have generally been based on the prin-
ciples laid down by Greenberg in his Universals of Language
(1966).
In my analysis I will adopt the abridged version
-13-

---

-14-
commented upon in Bartsch and Vennemann
(1972)3 and Vennemann
,
!
!
(1973).4
In their paper, Bartsch and Vennemann define 'The
i
principle of Natural Serialization' which shows the kind of
,l:I
consistency in the order of sentential elements found in lan-
I
II
guages depending on the kind of word order these languages
I
have adopted.
This principle states:
r
~
••. elements belonging together in the hierarchy of
semantic representation tend to be lexicalized and
reserialized in the surface representation in such
If
a way that hierarchical dependencies ,are directly
reflected in categorial operator-operand relation-
ships and the closeness of constituents to each
other in the surface string •.•
Furthermore the
I
operator-operand relationship tends to be expressed
i
by unidirectional serialization:
{operator {operand}}
r
tends to be serialized either as [operator [operand]]
I
throughout, or as [[operand] operator] throughout.
Less mathematical terms for operator and operand would be
I
"modifier" and "modified element" or "head" respectively.
t
The following chart taken from Foundations of Syntactic
I
Theory by R. P. Stockwell gives a summary of some of the
IiI
characteristics of SOV and SVO word orders.
This chart is
based on Bartsch and Vennemann's
'Principle of Natural Seria-
I[i
lization I.
(See p.15,
Table 1.)
fI,f
\\
f
1.3
Word Order in Gegbe
!
Now let us take a look at each of the characteristics
,
shown in the chart overleaf and with the help of examples
I
,
li
from Gegbe try to determine into which word order group the
I
I
language fi ts •
f
Ij
f
t
r

---

-15-
Table 1
~',
VO Languages
OV
I
Languages
\\I
(Head)
(Modifier)
(Modifier)
(Head)
I
!
I
f
1
2
1
2
t
~
1)
Verb
Object
Object
Verb
If
2)
Verb
Adverb
Adverb
Verb
I
I
,
3)
Intensional
Infinitive
Infinitive
Intensional
I
Verb.
Verb
!t
i
4)
Modal
Infinitive
Infinitive
Modal
(
!
Auxiliary
Auxiliary
,
5)
Aspect,
tense
Infinitive
Infinitive
Aspect, tense
I
If
,
6)
Auxiliary
Participle
Participle
Auxiliary
.!
7)
Noun
Adjective
Adjective
Noun
,
It
8)
Noun
Relative
Relative
Noun
I
clause
clause
I
9)
Noun
Genitive
Genitive
Noun
I
i
(Possessive)
I
10)
Comparative
Standard
Standard
Comparative
!
1
!
ll) Preposition
NP
NP
Postposition
f

---

-16-
I.
In an SOV language the object precedes the verb and
in an SVO language the opposite prevails.
Examples
_.l
_
(2.1)
nonua
pie
d,e
woman-the bought salt
'The woman bought salt. '
,
(2.2)
dev( 2~
s~
e va
na rn
child some took i t came to me
'A child brought i t to me.'
I
,
v
As these sentences show, Gegbe follows the SVO pattern in verb-
!
!
object order.
Object may precede the verb only in case of
!i
emphasis.
(2.1) I
ed,e y~
n5nua
pie
salt emp. woman-the bought
'It is salt that the woman bought.'
I
11.
In an SOV language the verb follows the adverb whereas
!
-
~
in SVO languages the verb precedes the adverb.
t
Examples
!~
(2.3)
e le
s(du sisi2~
he aspect run
hard
!;.
,
'He is running fast.'
,
(2.4)
rnl
debl~ k~b~
you (pI.)
hurry quickly
'Hurry up (pl.)! '
In the above examples G~gbe shows characteristics of an SVO
word order.
Ill.
In SVO languages the intentional verb precedes the
infinitive and in SOV languages the opposite prevails.

---

-17-
Examples
I
(2.5)
aff le
d3 f
la
yl 5
Afi asp. want fut. go
'Afi wants to go. I
I
,
(2.6)
~
d3 r6
b~
~
~
ye
yl
I1
he wanted that he inf. go
,
I
'He wanted to go.
,
I
Here again Gegbe shows characteristics of SVo languages.
i
[
IV.
In SVO languages, the modal auxiliaries 6 precede the
I
infinitive and in SOV languages they follow it.
I
Examples
,
(2.7)
ne mG
t~u
yl o
I
neg. could go particle
i
'I could not go.'
I
(2.8)
~y~k~ ~61~ QunJ
i
Ayoko must eat
'Ayoko must eat.'
I
The characteristics displayed here are those of anSVO word
I
order.
I,
v.
Aspect and tense precede the infinitive in SVO lan-
guages and follow i t in SOV languages.
Examples
(2.9)
mf I~
yl as~
we will go tomorrow
(2.10)
k~sf I~
las)
Kosi progr. take a bath
'Kosi is taking a bath.'
(2.11)
mf du
,
"
na
nu
ga
w6eve
me
we eat habitual thing hour twelve in
'We eat at twelve o'clock.'
In (2.9)
and (2.10)
tense and aspect precede the verb as is
characteristic of an SVO language but in (2.11)
the habitual

---

-18-
!
aspect marker follows i t , a characteristic of an SOY word
order.
1
,
I
VI.
Auxiliary and participle order.
Gegbe does not have
~
participial constructions because the form of the verb used
in the simple past tense,
for example, is the same as that
I
used after the progressive aspect markers le or n3.
Cf.
(2.12)
and (2.13)
below.
I
,
"
,
(2.12 )
ad yl
aSI-me
I
Afi went market-in
!
'Afi went to the market.'
, "
,
(2.13)
aft le
yl
aSI-me
Afi prog. go market-in
I
'Afi is going to the market.'
!I
VII.
For nouns and adjectives, the order of constituents
is noun-adjective for SVO languages and adjective-noun for SOY
languages:
I!
" "
- ,
I
'
(2.14)
en:::>
ne
na
m
8WU
nti
Q~
mother my gave me dress beautiful some
r
'My mother gave me a beautiful dress.'
I
(2.15)
at!
g~g~-w6
le
agble--a
me
I
tree big-plural are field-the in
I
'The big trees are in the field.'
I
,
Again Gegbe has an SVO word order for adjectives and nouns.
i
VIII.
In SVO languages, relative clauses follow the nouns
they modify while in SOV languages the former precedes the
I
latter.
I
Examples
!1r
(2.16)
dev r
ke
m (
kp:5~
child relative we saw-the
I
'The child we saw left.'
•

---

-19-
(2.17)
ah6~
ke
be tu-a
n5
1t~
house rel. he built-the be beautiful very
'The house that he built is very beautiful. I
,
Gegbe has an SVO word order pattern for relative constructions.
IX.
The genitive (possessive) and noun order in SVO lan-
guages is noun-genitive and in SOV languages it is genitive-noun.
Examples
(2.18)
d~d~
ne le lorne
elder sister my is Lome
'My elder sister is in Lome.'
(2.19)
a.
~d
wo
v~
f
father your came
'Your father came. '
It_
b.
devi~
b~
t~
v~
child-the poss. father came
I,
'The child's father came.'
!t
,
t
(2.20)
~
b6
awu-a
ze
~
she poss. dress-the be torn
I
i
'Her dress is torn.'
I!
(2.21)
a.
~
h:>t
"
ru-a
b 'e
"l)~
al:>
.-
door-the poss. hand be broken
'The door knob is broken.'
b.
k~d'36 b~
aI:' l)i
Kodjo poss. hand broken
'Kodj 0' s hand is broken.'
Except for the possessive adjectives in the first and second
person singular,? the genitive always precedes the noun in
If
G~gb~ with a particle be inserted between the two.
This order
I
,f
is that of an SOV language.
I
X.
In comparative constructions, the comparative precedes
I
the standard in SVO languages while the opposite order prevails
in SOV languages.

---

-20-
Examples
(2.22)
kod,6 /010
w6
paul
Kodjo be big
than Paul
'Kodjo is bigger than Paul.'
,
, ,
(2.23)
h6m~
\\
ya
ve
wu
ewu8
book this be expensive than that one
'This book is more expensive than that one.'
In this case G~gb~ shows an SVO word order.
XI.
SVO languages generally have preposition while SOV
languages usually have postpositions.
As the following examples
.
.
show Gegbe has postpositionsprimarily and therefore displays
SOV characteristics.
(2.24)
atav(
I e ~h6-me
uncle
is house-in
'Uncle is (at) home.'
(2.25)
kr(sm~s
dz( be v~
Christmas on
he came
'It is at Christmas that he came.'
(2.26)
so
e d6 ad -~
gb6
take it put tree-the near
'Put it near the tree.'
There are however few instances of what can be called pre-
.
: \\~" "'-;~~""<-'~
positions.
For example naB (fori ,';j >~ -k6--~ibl:)'
(2.27)
\\
~.L'
-:F'/
ho
h E n Cl
m
r' i C /\\ ~Ji ["
,,_
recei ve hold for me 'i'.r~ ;:--- -,
'Hold (this) for me.'
/
~.~--~~-~. ,.<..1 ,
(2.28)
abl~v( YI
k6
ay~b~
, :>~.. ~.,_:: >
Ablavi went with Ayaba
.-'
I
As the above discussion shows, G~gbe is of neither a straight-
I
forward SOV nor SVO word order type.
This is not surprising if,
I
as mentioned before, G~gbe and many of the Kwa languages have
evolved from Niger-Congo, a primarily SOV word order language.
\\' , .

---

-21-
Besides, it is a common fact that languages displaying primarily
one kind of word order characteristics in the order of their
constituents will also have characteristics of another type of
word order (see fn. 1).
English is a good example of this (e.g.
genitive constructions).
One thing that is clear in the above examples is the fact
that, for its major categories constituents that is, subject,
verb, object, adjective, adverb, etc., G~gbe displays an SVO
type of word order.
Furthermore, of the eleven items considered
in the above chart, only two are really characteristic of an SOV
word order.
Therefore we can conclude that, in light of the
facts examined in this section, G~gbe displays primarily SVO
characteristics with a few characteristics of an SOV word order.
2.
The Category AUX
2.1
A Definition of the Category AUX
In Syntactic Structures (1957), Chomsky sets up a category
AUX as part of the phrase structure rules for English.
Included
in this category were elements which are used to confer the idea
of tense, modality and aspect and the PS rule
AUX ~ Tense (Modal)
(have+en)
(be+ing)
constrains the relative distribution of the constituents in
this category.
Such a proposal has had supporters in Jackendoff

---

-22-
(1972), Emonds (1976), Culicover (1976) and others, as well as
challengers such as Chafe (1970), Ross (1969), Pullum and Wilson
(1977) .
Susan M. Steele in a paper entitled "The category AUX as a
language tn'liversa1" (1978) and in another paper "The category
AUX in Universal Grammar" (1979)) she co-authored with A. Akmajian and
T. Wasow) supports the idea of a category AUX for English and
goes further to propose such a category for universal granunar.
This category must be made available to all languages of the
world since,
as she tries to prove, quite a few languages seem
to exhibit elements in their grammar which can
readily fit
into such a category.
Thus, including the category AUX in uni-
versal grammar will avoid having to add new categories every time
a category is fotn'ld in a language which includes elements which
in some respects can be considered auxiliaries.
Of course, not
all languages will choose to include this category in their
grammar.
Before going further into the analysis of the cate-
gory AUX. let us take a look at the different constituents that
make up this category.
Grammarians have not always been consistent in their defi-
nitions of the term 'auxiliary'.
For example Jespersen (1940), among
other~ only considers be, have, shall, will as auxiliaries and
the modals are considered verbs "that take another verb as their
object."
Akmajian, Steele and Wasow include the moda1s among
the auxiliaries.
As a general rule, auxiliaries are recognized
1
by grammarians as performing a role different from that of any

---

-23-
other element to which they relate in a sentence.
In particular,
they differ in some respect from main verbs; even the modals
which are often cited in support of a position against setting
up a category AUX for Indo-European languages (Ross (1967a))9
are still different from main verbs in many respects.
One of
the main characteristics of auxiliaries is that they are used
to modify the meaning of a sentence in regard to tense (time),
aspect or modality depending on the meaning of the individual
auxiliaries.
Let us try now to provide a more formal definition
for the category AUX.
Stee1e (1978), in her analysis, adopts the definition of
the category.AUX advanced in Chomsky 1957, that is, "The AUX
contains a certain notional set, a set containing tense, aspect
and modality elements ...
This notional set involves elements
which are sententia1 in scope, i.e. they place the situation
described in the sentence in a certain time (tense), ascribe a
temporal contour to it (aspect), and assess its reality (moda1-
. t
)" 10
1.
Y
.
Stee1e (1978) also focuses on three claims made in
Syntactic Structures, that is:
1)
There is a deep structure unit AUX.
2)
The AUX and the main verb occur in the same clause;
that is, the main verb is not (at any point in the
derivatiori) subordinated to the AUX.
3)
The unit AUX involves certain notional categories
(see quote above).

---

-24-
In the same paper, Steele gives different characteristics
appropriate to the different grammatical units that can be con-
sidered as auxiliaries.
She also points out some characteris-
tics of languages which have auxiliaries as opposed to those
which do not.
There are two different types of category AUXi
a category AUX, labelled AUX ' which includes elements used to
V
modify verbs, traditionally referred to as auxiliary verbs, ex-
pressing aspect, and a category AUX, labelled AUX_ ' which does
V
not have such elements.
There seems to be a correlation between
the type of AUX a language has and the kind of word order it
displays and the relative freedom of this word order.
For exam-
ple, a language with a rigid SVO word.Qrder tends to have a
category AUXv.
Furthermore, languages
with subject agreement
or a subj~ct marking of some sort are generally those which have
the AUXv t!pe category, and very often the agreement is expres-
sed in the AUX.
Another important characteristic of AUX is the position it
occupies in a sentence.
In general, there are three main posi-
tions favored by the category AUX:
initial position, second
position or final position.
Again, the position adopted depends
on the word order and the rigidity of this word order in the lan-
guage.
For example, languages with a rigid SVO or relatively
free SOV word order tend to prefer a second position AUX.
How-
ever there are languages with rigid VSO/VOS word order which
have the AUX in second position even though most languages in
this group seem to prefer a sentence initial AUX.
Rigid SOV

---

-25-
languages tend to have a sentence final AUX.
Another common
assumption in the description of AUX is that it is generally
contiguous to the main verb.
Stee1e (1978) shows that this is
not always the case and that "the contiguity and non-contiguity
of the AUX to the main verb is a characteristic which correlates
. with which of the two types of AUX the language has.
AUXV are
generally contiguous to the verb. AU~V are not.
However. there
are exceptions.
~or example. in German. a language with AUXV'
the auxiliary is not always contiguous to' the main verb.
Another characteristic of the AUX discussed in both Stee1e
(1978) and Akmajian. Stee1e and Wasow (1979) is the form some of
the constituents of AUX may have.
In most cases they are either
c1itics (AU!v languages) or are c1iticizab1e (AUXv languages).
For example in Luiseno tense and aspect are expressed with
c1itics. in English the auxiliary will is c1iticizab1e.
Among
the three groups of the catego!y mentioned earlier. tense is
the most likely to be c1iticized. aspect the next most likely
and modality the least likely.
Finally. because in many languages with AU!v there seems
to be no clear-cut differentiation between tense and modality.
Stee1e (1978) suggests the expansion of AUX as follows:
AUX
~
Super Tense (Aspect)
in which Super Tense includes both tense
and modal elements in
cases where they are differentiated.
The following excerpts
from Stee1e (1978) summarize the different characteristics of
AUX dealt with in the above discussion.

---

-26-
... [T]he AUX will have certain characteristics.
It
will occur in one of three sentential positions,
primarily sentential second position; it will almost
always include subject agreement; it will contain
either clitics or elements which are cliticizable ...
[T]here are two types of AUX, one of which is based
on aspectual auxiliary verbs and the other of which
is based on modal elements ... [W]hich type a lan-
guage exhibits is dependent on certain other aspects
of the language, primarily basic word order type and
the relative freedom or rigidity of word order.
(p.
41)
In the same paper, the author gives restrictlons regarding
the,AUX in what is a possible language:
a.
No language of AUXV type which has subject
agreement will not include it in the AUX.
b.
No language of AUXV type will exist in which
the AUX does not take what is usually verbal
inflection.
c.
No language with free word order will have a
clause final AUX.
d.
No language with SVO or VSO ~asic word order
will have a clause final AUX.
e.
No language will exist where AUX does not po-
tentially include tense; that is, in a sen-
tence with tense and an AUX, tense will always
be part of the AUX.
.
(p.
42)
2.2
Auxiliaries in G~gbe
The main purpose of the following discussion is to examine
in the light of the assumptions brought forth in section 2.1,
the constituents in G~gbe which may be described as auxiliaries
and to determine whether it is necessary to conclude that a des-
cription of G~gb~ includes a category AUX, and finally whether

---

-27-
it is necessary or even desirable to have a category AUX in the
phrase structure rules in the base component.
2.2.1
General Remarks on Predicative Constructions in
Before going into any further details about· auxiliaries,
some general remarks about verbal constructions and the relation
between tense and aspect are required at this point.
In G~gbe predicative constructions, the verb always has the
same form no matter whether it is used with an auxiliary or with-
out (to express the simple past tense for example).
Cf.(2.29)
below.
,
, ,
,
(2.29)
a.
ad I~
yl
aSI-me
Afi will go market-in
,
'Afi will go to the market.
,
, ,
,
b.
ad y I
aSJ-me
Afi went market-in
,
'Afi went to the market.
As will be shown later, however, tense (except for the
future perhaps) is not disjunctive with modality or aspect.
As
the following examples show, only the context and the kind of
adverbs used differentiate the past from the present at times.
(2 30)
,
t I.
"
"
" ( '
L
-a) 11
.
e
teu yl
as/-me
es~
ke
Vd
yl
he can go market-in day (that come go the)
'He could go to the market yesterday.'
(2.31)
e t~~ yl asl-me
es~ (ke
gbana
a)
he can go market-in day (that be coming the)
'He can go to the market tomorrow.'
,
(2.32)
~
t~u YI aSI-me
'He can go to the market. '
(No indication of tense)

---

-28-
In (2.33) the particle la (future) can be used to precede teG to
indicate the future tense:,
(2.33)
~
IJ
t~u y1 as1me
~s~
he future can go market-in tomorrow
'He will be able to go to the market tomorrow. I
The same situation generally obtains for the other modals and
some of the aspect markers.
So at this point we may assume that
,
in Gegb~ only two simple tenses can be differentiated, the future
tense and the non-future tense.
Only the former can be overtly
expressed with the use of la.
However there are a few excep-
tions to this rule.
For example with the modal g6la expressing
an obligation la is never used.
Only the context will indicate
whether reference is being made to the past, present or future.
This is exemplified in the sentences in (2.34) below.
(2.34)
a.
aff 261~ y1 es~ ke v~ yT a
Afi must go yesterday
'Afi had to go yesterday.'
b.
aff d61~ y1 as1me
gbesf~gbe
Afi must go market-in everyday
'Afi must go to the market everyday.'
c.
aff 261~ y1 ket~ krismas
d?f
Afi must go Keta Christmas on
'Afi {:~~f have to} go to Keta at Christmas.'
i
,
1
This behavior of d61a may prompt us to say that the la in the
i!
word is really the future marker.
However there is no empirical
~,
!
reason for saying so since the different meanings of the word
2£ occurring alone do not come anywhere near the idea of obliga-
tion. 12
42 may have one of the following meanings:
'have (pos-
session)', 'to gather in small bunches',
'to line up',
'to
II

---

-29-
prepare (rain)', 'to have a bump'.
Finally, since in the following section we will try to de-
termine whether the constituents under consideration here can be
considered as main verbs, it is appropriate at this time to
point out the different characteristics of a verb in G~gbe. In
G~gbe sentences, a verb:
1)
is the nucleus of a predicate phrase.
2)
can be modified for tense, aspect and modali ty ,13
3)
can be modified by an expression of time, place or
manner.
This is exemplified in the following sentences:
(2.35)
k;,d36 ~u
gal(
Kodjo ate gali
(2.36)
~
,~
t~n no
pie a~T
he fu t. can prog. buy soap
'He can be buying soap.'
(2.37)
k;,d36 QU
galr-~
k~b~k~b~
Kodjo ate gali-the quickly
'Kodjo ate the ga1i quickly.'
In (2.35) QU is the nucleus of the predicate phrase; in(2.36),
,
~ ,
y
~ is modified by lA tell n~ (tense, modal and aspect) and in
(2.37) ~ is modified by the adverb of manner kabikaba.
2.2.2
The Different Constituents of the Categories of
Auxiliaries.
2.2.2.1
Tense
As discussed above in section 2.2.1, the only two distinc-
tions to be made in this category in G~gbe are the future and

---

-30-
the non-future.
The future is generally expressed by the use of
la and the non-future by 0 as shown in the following sentences.
,
(2.38)
a.
'\\
..
'\\
af r I ~
yl aSlme
Afi will go market-in
I Afi
will go to the market. '
,
, ,
b.
ad
yl aSlme
Afi o go market-in
'Afi went to the market.'
c.
af(
lti
,!.
'\\
'\\
"
'\\
teu yl aSlme
Afi will can go market-in
'Afi will be able to go to the marke t. '
, ,
,
d.
ad
I '
,
teii yl
aSlme
Afi o can go market-in
'Afi was able to go to the market. '
e.
ad
I~
no
yl
aslme
Afi will prog. go market-in
'Afi will be going to the market.'
f.
•
..
'\\
'\\
'\\
I e yl
aSlme
prog. 14
'Afi is going to the market. '
La cannot be considered as a main verb because it does not meet
any of the criteria for verbness in G~gbe listed above in section
2.2.1:
- It cannot be the sole nucleus of a predicate; it must
always be followed by a verb or at least a modal ele-
ment, preferably t~u.
(2.39)
a.
ad
I~
yl
aSlme
'Afi will go to the market.'
,
, ,
*b.
aSlme
Or, in answer to a question such as (2.40) one can use the
sentence (2.41) but not (2.41').

---

-31-
,
(2.40)
af f I ~
t~jj k~k:S adi3ka a
Afi will can lift trunk question
'Will Afi be able to lift the trunk?'
, ,
(2.41)
ee,
e
I~
te6
yes, she will can
'Yes, she will be able to. I
L~ cannot be modified by an adverb:
La cannot be modified for aspect or modality:
(2.43) a. *af( t~6 I~ v~
,
Another criterion for verbness in Gegbe is the ability for
.
15
verbs to be followed by the habitual aspect marker na.
Though
this will be discussed later in the appropriate section, it can
be pointed out here that, whereas all other aspect markers always
precede the main verb in a sentence, ~ always immediately follows it.
(2.44)
a.
aft
"
"
" ' \\ , ,
yl
na
aSlme
Afi go hab. market-in
'Afi (generally) goes to the market.'
b.
af (
,
,
, I (
<;tu
na
ga
Afi eat hab. gali
'Afi (generally) eats gali. '
c. *atf
I~ n6 9u gal(
Finally, l~ can be cliticized when it is preceded by the
negative marker mU or the first person singular pronoun mu
(or ~).

---

-32-
,
(2.45)
a.
ma r i e mu
I~
\\
VI
0
+
Marie neg. will go negative particle
b.
marie m~
\\
\\
VI
0
'Marie Won't go. I
(2.46)
mu I~
v1
V
\\
ma V I
I
will go
, I '11 go.'
2.2.2.2
Modality
Three main morphemes used in G~gbe to express modality are:
,
teii, Q6U and ke.
,
~ii (or tem) express the idea of possibility, physical abi-
lity and permission.
,
(2.47)
mf t~ii k~k~ adakaa
we can lift trunk the ,
'We can lift the trunk.
,
.
(2.48)
\\
m b~
\\
ml
teu VI
I
said you can go
, I
give you permission to go. I
,
(2.49)
pap~
teii gbo
wetd
va
father may come back afternoon this ,
'Father may come back this afternoon.
(possibility)
,
As a general rule, t~u is always followed by a verb but is not
necessarily adjacent to one.
However, there are cases such as
short answers where this is not so; in such cases the main verb
is still understood even though it is not expressed:
,
(2.50)
~evf ~
I~
t~u k~k~ E a
child-the will can lift it quest.
'Can the child lift it?'
,
\\
\\
(2.51)
EE,
~
I~ t~u
'Yes, he can.'
The same process of elision is found in predicative sentences
given in answer to a question, where the object of the verb can
be omitted:

---

-33-
(2.52)
rn)
la
du
. ga I ( a
you will eat gali quest
'Will you eat gali? '
, ,
,
(2.53)
'\\
~
\\
00, rnl
ma
2u
o
no, we not-will eat part.
'No, we will not. I
J.
T~u can also be preceded by the future tense marker la but it
cannot be preceded by the aspect markers le or n3 which as we
will see later are always adjacent to the main verb.
,
(2.54)
af( I~
t6jj YI
'Afi will be able to go. I
,
(2.55)
a.
afi Un no
yl
fff1
Afi can prog. go now
'Afi may be going now. I
,
,
*b.
ad n3 t~u yl fft)
,
Furthermore, t~u cannot be followed by the habitual marker na,
which, as mentioned earlier in section 2.2r2.l, is a characte-
ristic of main verbs:
,
(2.56) *afr t~u n~ yl
,
but
a f ( t ~u y1 na
Afi can go hab.
'Afi is able to go (usually).'
,
. T~u can occur in constructions with g6la but not ke.
It
can precede or follow ~:
,
,
(2.57)
a.
ay~b~ d61~ t~u yl
.
Ayaba must can go
,
or b.
ay~b~ t~& d61~ y I
'Ayaba shouid be able to go. I
,
,
,
(2.58) *a.
e t~u ke yl
*b.
~ ke t~a
,
yl
,
T~u alone cannot be modified by an adverb, it must be followed

---

-34-
by a verb:
,
(2.59)
a.
~ I~ t~u y) fff)
*b .
~ I ~ t ~ ~ f f f I
,
Finally, t~u is not c1iticizab1e.
~61~ is used to express an obligation, moral or otherwise,
or a necessity.
The different meanings of this modal are exem-
p1ified in the following sentences:
(2.60)
devf-~
d 6 L
ICI
"
dunu
'f'
fl
I
.
.
, ~hi1d-the must eat
now
'The child must eat now (it is time for him to).'
(2.61)
afr 961~ YI Q~klta
eSO
Afi must go hospital tomorrow
'Afi will have to go to the hospital tomorrow.'
(2.62)
o m6
d61~ yo
slga
le flye 0
you neg. must smoke cigarette' in here part
'You must not smoke cigarettes here (it is prohibited).'
,
,
(2.63)
devr sr~ devf
d61~ b~
t6~~
k~
n5-~
every chiid ~st respect father-his and mother-his
'Every child should respect his father and mother.'
~ never occurs as the main verb of a sentence.
It must always
be followed by another verb.
The verbal complement of d61A can-
,
not be omi ttedas in the case of t~u.
(2.64) *afr Q61~ galf
,
(2.65)
a.
0
d61~ k~k6 E
kpo
y~
a
you must lift him necessarily emphasis quest
'Must you necessarily pick him up? '
, ,
,
b.
EE, m Q61~ k~k~ E
yes I must lift him
'Yes, I must.'
, ,
*c.
EE. m
~61a is never preceded by the future tense marker la but it can

---

-35-
be followed by the aspect marker n3.
It is not used in sentences
along with ~.
(2.66) *f3
I~ d61~ YI kpo Y~ a
to mean 'Will he necessarily have to go?'
This idea is generally expressed by (2.67):
(2.67)
~
d61~ y) kpo
ye
a
he must go necessarily emphasis question
'Will he necessarily have to go?'
(2.68)
f3
~61~ n3
gun6 f(fl
he must prog. eat
now
'He must be eating now.'
(2 • 6 9) *'
61 '
"
• "
,
('
e
,16
~
a Ylna aSlme gbe S agbe
to mean 'He must go to the market everyday.'
,
As mentioned before, g6la and teu can occur together in a sen-
tence.
See example (2.57) above.
Finally, ~ cannot be modified by an adverb and it is not
cliticizable.
Ke is used to express a wish not likely to be realized,or
a regret.
(2.70)
nf3 mse
yf3
~
mke
yl
if I heard emp. part. I mod. go
'Had I heard, I would have gone.'
(I'm sorry I didn't go)
(2. 71)
ed?T I~.
ke
d?6
se
up
will mod. drip emphasis
'It should rain (it would be nice).'
Ke does not meet any of the criteria for verbness discussed
above ~d it is not cliticizable.
Of the other two modals, d6la
~
is the only one with which ke can occur in a sentence.
In such
cases, d6la must always precede ke.
- - .
-

---

-36-
~
(2.72)
a.
.£
kef( <)61 ~ ke
V~ kp~ n:>-a
Kofi must mod. go see mother-his
'Kofi should have gone to see his mother (too bad
he didn l t go). I
, ,
*b.
n513
(2.73) *ilf( t~6 ke k~k~ kpl5a
As is the case with d6la, the verbal complement cannot be omitted
--
after ke:
,
(2.74)
a.
o
b~
~
ke
VI aneh:S a
you said he mod. go Aneho quest.
'Did you say he should h~ve gone to Aneho?'
, ,
b.
EE.
~
ke
V'
'Yes. he should have gone.'
, ,
*c.
Ee, e ke
Ke can cooccur with 1<£, n:S or na in a sentence:
(2.75)
m(
I~
ke
V'
iln~h~ ~gbe
we fut. mod. go Aneho today
'We should have planned to go to Aneho today.'
'-
(2.76)
, ( "
at
ke
'"
no
'
VI " ' ' ' '
aSlame
d:;e l)k~
na mr
Afi mod. prog. go market-the-in ahead
of us
'Afi should have been on her way to the market ahead
of us.'
(2.77)
e ke
tutu na
m n~ ma
pe
e
he mod. push hab. me so I-will beat him
I If
only he'd pushed me so I'd have had an excuse
to beat him up. '
Sentence (2.77) expresses more the idea of threatening than a
wish or a regret.
Na here does not really convey the idea of
habi tuali ty.
As the discussion in this section shows, the modal elements
cannot play the role of main verbs in a sentence in G~gbe.
They
are not cliticizable.
They always occupy a pbsition after the
negation marker mU. the tense marker la and before the aspect

---

-37-
marker n~.
Finally only two of them can occur at the same time
,
in a sentence, that is, ~ and tell or ~ and ke.
2.2.2.3
Aspect
,
To mark aspect in Gegbe, the following morphemes are used:
le and no for the progressive or continuous, d3~ for the incho-
ative aspect, na for the habitual.
To express intention, the
morpheme gb~ is used.
The continuous can also be expressed by
a eonstruction using w~ with le and n~, two stative verbs.
In
the following discussion, we will also take a look at some aspec-
tua1 constructions using particles such as d3{, k~ and the verb
v~.
However, as the discussion will show, not all of these
morphemes can be considered to be auxiliaries.
,
As ~tated above, the continuous aspect is expressed in Gegbe
by two main types of constructions: . a) a construction with le
and no preceding the verb and b) a construction with le and n~
and the verb w~ (to do or make).
In an unpublished manuscript
written in 1979, I tried to show that le and n~ are two stative
verbs, translatable by 'be', with complementary functions:
le is
used
primarily in the present tense and in some idiomatic
constructions; n~ is used in tenses other than the present, in
all aspectua1 constructions and with all the moda1s.
For a more
detailed analysis of le and n~ in constructions with the progres-
sive aspect, let us consider the following examples:
. ,
.
(2.78)
d~vr~
le 5 151
W~
child-the is fleeing m~e
'The child is running away. '

---

-38-
,
,
(2. 79)
9~vf~
n:>
na
Slsf
w:,
child-the is hab. fleeing make
'The child used to n.m away.'
(2.80)
rnf
le
n6
du
,
. w:>
we are thing eat make
,
'We are eating.
,
(2.81)
~
le 10(0
w:>
he is get fat make ,
'He is getting fat.
(2.82)
aH le V1Vl
w:,
Afi is going make
'Afi is going (in the proces s of).'
(2.83)
aff ' , '\\ ,
"
,
le aSlrne
VI
w:>
Afi is market-in going make
'Afi is going to the market.'
(2.84)
atf-~
le 1)~l)l
W~
tree-the is breaking make
'The tree is breaking.' ,
(2.85)
devf~
le atf-~
1)£
w:,
child-the is tree-the break make
'The child is breaking the tree. I
,
In Gegbe, nominals derived from verbs are formed by a re-
duplication of the verb stem for verbs with one syllable.
For
verbs with more than one syllable the nominalization can be
expressed by a change in tone.
Thus sIs! 'flight' in (2.78)
and (2.79), for example, is derived from the verb si 'to flee'
or 'run away'; yiyi and l1Yl (going) are derived from Yi 'to go'
when the verb has a high tone, the first syllable of the nominal
bears a rising tone and when the verb has a low tone the second
syllable bears
the rising tone.
In verbs with two syllables,
the same tonal alternation takes place though there is no redu-
plication.

---

-39-
(2.86)
)010 'fatness' from 1010 'to be or grow fat'
,
,
(2.87)
tTtr£ 'persistence' from tftr£ 'to be persistent'
According to Schroeder (1936) and Johnson (1967) construc-
tions with le and n~ + wo involve an auxiliary verb le or no, a
verb (with a reduplicated form for one-syllable verbs (~, va,
etc ... » and a particle w~ following the verb.
Sentences
(2.78)-(2.85) above exemplify this.
This type of analysis agrees
with analyses done for Ewe progressive sentences (cf. Westerman,
Diedrich (1907), Warburton, I.
, Kpotufe P. and Glover, R.
(1968».
As examples (2.88) and (2.89) show, such sentences in Ewe have
the auxiliary verbs le or n~, the verb (reduplicated form or not)
and the particle m suffixed to the verb.
,
(2.88)
keff le
y)yl
m
Kofi Aux go (redup lica ted from Y!.) part
,
'Kofi is going.
,
(2.89)
keff n~
xevf I~
rfi
yesf~yT
Kofi Aux bird catch part. all the time
,
'Kofi is always catching birds.
The fact that in both languages the object precedes the "main"
verb may be considered a proof that G~gbe in particular does
exhibit an SOV order in the surface order of the main components,
that is, subject, object and verb.
However, as will be explained
presently,I am not quite convinced that the above analysis of this type
,
of progressive sentences also holds true for such sentences in Gegbe.
,
In spite of the structural similarity between Gegbe and
Ewe, I still consider constructions with le or n~ + w~ not as
constructions with auxiliary + verb + particle (aspect marker)
as they have been analyzed by Johnson and Schroeder (see above)

---

-40-
but rather as constructions with two verbs le or no (stative verb
'be') and w~ 'to do or make' and a nominal.
Thus in an example
like (2.78) above, the child is considered to be in the state of doing
the running away. My reasons for holding such a position are based on
,
two factors to which previous analyses of Gegbe do not seem to at-
tribute much importance.
Such factors are 1) the use of the
habitual marker ~ in such constructions and 2) the alternation
of tonal patterns on the reduplicated forms ;.of the verbs.
As mentioned earlier in the discussion in this se~tion
(see p.3l ) the habitual marker ~ generally follows the main
verb in the sentence (cf.
(2.90) and (2.91) below) .
(2.90)
rnf du
,
na
ga If
we eat hab. ?ali
'We eat galL
,
,
v
(2.91)
a.
rnf tt$ii du
,
na
hala
we can eat hab. pork
,
'We can eat pork.
,
,
,
v
*b.
rnf tt$ii na du hala
As the ungrammatica1ity of (2.91b) shows, the na cannot follow
the auxiliary or modal element.
In progressive constructions
with n~, na always fol.lows n:> when it is used, rather than the verbal
constituent which Johnson and Schroeder have analyzed as the main verb.
,
,
,
(2.92)
a.
llf f no
na
ga I ( pie wo
Afi be hab. gali buy make
,
. 'Afi is generally buying gali .
,
,
*b.
af r ,
no ga I r pie na wo
From this we can conclude that p1e (and any other form of the
verb occupying this position in a sentence) does not play the
role of the main verb.
Therefore the order of elements in a

---

-41-
sentence like (2.92) above is not SOV with Aux between Sand 0
but SVO as (2.92a') shows:
(2. 92a')
at(
non~
'
ga r
I
'
pie
I
[Subj] [V+Asp]
In other words, as (2.92a) describes it, Afi is in a particular
state and that state is the performance of the action of buying
gali.
The second factor, that of tonal alternation on the verb
form, supports my assumption that the form of the verb used in
progressive sentences with le and no is the nominal form.
For
example in (2.78), (2.79) and (2.84) above, the nominal forms
of the verbs si and QE are used.
However in (2.81) and (2.82)
the nomina1s 1010 and Y1Y1 do not have a rising tone on the last
syllable as should be the case according to (2.86) above.
My
explanation for this is that there are two forms of nomina1s
with different tone patterns derived from verbs:
for high tone
verbs, the two syllables have either a high-high or a rising-
high pattern and for low tone verbs. they have either a low-rising
or a low-low pattern.
Cf.
(2.93) and (2.94) below.
,
"
(2.93)
QE
{Q£QE}
break Q~ Q~
'breaking'
(2.94)
y";
{Y~Yl}
go
YlyT
'going'
Constructions with le or no + wo use the nomina1s with rising-
high tone pattern for high tone verbs and the nomina1s with low-
low tone pattern for low tone verbs.
Cf.
(2.82) and (2.84) above.

---

-42-
The other forms of nominals, that is nominals with a high-high
tone pattern for high tone verbs and a low-rising tone pattern
for low tone verbs are used in modifying constructions of the
type of (2.95) and (2.96) below:
(2.95)
at f
.l
.el7
,
fa
\\
l)El)E
mu
0
wood breaking neg. be easy part.
'Wood breaking is not easy. 1
(2.96)
sUk61u viv1
ve
n~ dev(
va
l)t:5
school going be bitter to child this very
'This child hates going to school.'
In G~gbe, any construction with verb + noun phrase or pre-I
post-positional phrase or adverbial phrase should be considered
a transitive construction because the PP's and the AdvP behave
like any NP, as the following examples with le and no show.
In these examples, the object (m:5lu, asime or blEwU) comes
-
immediately after thestative verbs and is followed by a short
form of the nominal expressing the action being performed in
,
,
the sentence, that is du, ri, z5 in (2.92), (2.93) and (2.94)
respectively.
'"
(2.97)
mf
le m:5 J 6 t!a
\\
W;)
we be rice picking do
'We are picking rice. '
(2.98 )
m(
le • , •
,
\\
aSlme
VI
W;)
we be market-in going do
,
'We are going to the market.
,
,
(2.99)
~
le b I EWU
z5
wo
he be slowly walking do
'He is walking slowly. I
My reason for considering the form of the verb following the
object as a short form of the nominal rather than a verb is
that, though it is not common, we do get the reduplicated form

---

-43-
in that position.
(2.100) below is a perfectly grammatical
sentence.
,
(2.100)
~
le tI(tla
sf s (
W:J
he be teacher be afraid (or flee) make
'He is (being) afraid of the teacher.'
However, there is a problem with this analysis of transi-
tive sentences.
I have shown earlier that the form of the
. nominals used in modifying constructions is the one with a
high-high tone pattern for high tone verbs and low-rising for
low tone verbs.
Example (2.100) agrees with this hypothesis.
But as (2.101) shows, only the low-low form of the nominals
is used for low tone verbs in transitive constructions.
(2.101)
aft le kpl5a
t ut u
,
a.
W:J
Afi is table
pushing do ,
'Afi is pushing the table.
,
at(
t ut'~
,
*b.
le kpl5a
W:J
It could be suggested, as a solution to this problem, that
these kinds of transitive constructions be considered as some
kind of serial verb construction where le or n~, the morpheme
following the object (what I have so far considered a nominal)
and w~ are the verbs.
The form of the verbal nominal used in
such constructions can
then be considered as a participial
form of the verb.
Another solution would be to include a clause
in the grammar stating that in this type of aspectual construc-
tion with le or n~ + w~ only the nominal form with a low-low
tone pattern is used for verbs with low tones.
To sum up the above discussion on le and n~, I have tried
to show that in these types of aspectual constructions, le and

---

.-44-
no are stative verbs followed by a nominal derived from a verb
plus wo, a verb meaning 'do' or 'make'.
In other words, in
such sentences, the actor or subject is in the state of per-
forming the action expressed by the nondnal.
The nominal is
then the complement of the stative verb.
Another way of expressing the progressive or continuous
aspect is by using a form of le and no with rising tone fol-
lowed by a verb as shown in examples (2.102) through (2.104):
(2.102)
m( le
, , 1 '
yl
as me
we prog. go market-in
,
'We are going to the market.
. , , ,
(2.103)
m(
lti
no
yl
aSlme
we will prog .. go market-in
,
'We will be going to the market.
(2.104)
m(
le
(jU
m~J6
we prog. eat rice
. 'We are eating. rice. '
The syntactic distribution of l~ and n~ is the same as that of
the verbs le and no, that is, le is used primarily for the pre-
sent tense and n~ is used for the past and with other auxilia-
ries.
Schroeder (1936) gives different possible explanations for
the rising tone or low-high tone on the two morphemes.
In one
of these explanations, he provides examples of the construction
illustrated by (2.105) below with a morpheme ! be~ore the verb:
(2.105)
m le ~ y1 wo
I be it go particle
'I'm going.'
This ~ is a third person pronoun referring to the goal of the
action expressed by the verb in the case of intransitive verbs.

---

-45-
For transitive verbs, this e is a trace left behind by moving
the object into a position after the verb as a result of the
change from a SOV to a SVO word'order.
The segment of this
trace is later deleted, leaving the tone behind and the result
is the rising tone on le and n~.
This kind of explanation
assumes that le and n~ are some kind of auxiliaries and not
main verbs as is advocated in this thesis.
Though this exp1ana-
tion may be plausible
on historical grounds, it is too abstract
because synchronically, all objects appear after the verb~
Another explanation for the rising tone is provided by
Tohouegnon (1968).
In his analysis, Tohouegnon refers to an
infinitive form of the verb used in constructions of the form
of (2.106) :
, ,
(2.106) , ~
vfJ
is fifi
he came to steal
The a is the infinitive marker of the second verb.
The
rising tone on le and n~ could be derived from the combination
of the low tone on le and n~ and the high tone on a, the seg-
ment of the latter having deleted.
Cf.
(2.107).
(2.107)
~ le a du n~
~
e le du n~
Le and n3 are no longer verbs in their own right as are
the two verbs from which they are derived.
They cannot be the
nucleus of a predicate phrase;
they must always be followed by
a verb.
Furthermore they are always adjacent to the main verb
in a sentence.
They can be preceded by the moda1s and the
future marker.
They are not c1iticizab1e.
For illustrations

---

-46-
of the uses of le and n3 with the other constituents of the
auxiliary group, cf. examples (2.55),
(2.68) and (2.76) above.
There are two other constructions used for expressing the
progressive:
those with the morphemes d3{ 'on' and k~ 'at'.
These morphemes, as the following examples show, are postposi-
tions preceded by nouns or nominals rather than auxiliaries.
(2.108)
~
Ie n ~
dud U
k~
"he be thing e~ting at
'He is eating. I
(2.109)
'\\
"
,
r '
rnl
le
d:>a
d~
a
you(pl.) be work-the on quest.
'How is the work coming?'
,
,
(2.110)
0
gba
le
sTsr-~
d~r a
you still are running away-the on quest.
'Are you still running away? '
(2.111)
~
le '
.t. -
-
aple:go a
k~
it be nail-the
at
, It
,
is hanging on the nail.
,
(2.112)
e le hpl5a
d~r
i t is table-the on
, It
,
is on the table.
In (2.108) and (2.110) dud~ and sIs{a occupy the same syntactic posi-
~
,
,
tion as aplegoa and ekp15a in (2.111) and (2.112) respectively.
Therefore in the first two examples d3! and k~ are just post-
positions used after nominals, comparable to English preposi-
tion 'at the table' or 'on the table' for example.
They are not
auxiliaries.
Now let us turn to other aspect markers.
The morpheme gb~ is used in sentences to express intention.
Cf. (2.113) and (2.114):
(2.113)
at r v~
ga I r sa
gb~
Afi came gali sell for
'Afi came to sell galL'

---

-47-
(2.114)
ad,6 YI
an~ke
gb~
Adjo went firewood for
'AdjQ went to fetch firewood.'
In (2.113) Afi came with the intention of selling gali and in
(2.114) Adjo went away with the intention of fetching firewood.
The form of the verb used in (2.113) is the nominal form because
here sa plays the same syntactic role as anake.
Sometimes the
complete form of the nominal may be used.
Thus (2.115)
is a
perfectly acceptable sentence with the same meaning as (2.113).
(2.115)
at( v~ ga I r .sasa gb~
When gbe is used in a sentence with le, it only expresses an
intention in the future, never in the past as may be the case
for other verbs.
Cf.
(2.116)-(2.119):
,
(2.116 )
-,
,
I"
d'
b~
ne mu
e
nu
.u
9 e
o
I
neg am thing eat intention part.
'I'~ not going to eat (I don't intend to).'
*-'
,
...
,...
(2 117)
~
...
.
ne mu n~ nu ~u gbe 0
(to mean' I was not going to eat. ')
(2.118)
~
v~
ga I r sa
gb~ gb~€
m d,6
she came gali ·selling for before I left
'She came to sell gali before I left. I
In this case, gbe in combination with the nominal phrase pre-
ceding it can be considered as an adverbial phrase used to
modify the main verb, attributing to the meaning of the sen-
tence a sense of intentionality.
Thus in (2.118) gal! sa gb~
is the adverbial phrase modifying va.
The morpheme d3a is used to express the idea that an action
is about to take place; it is the inchoative aspect marker:

---

-48-
(2.120)
e be
ye d,~
d,6
he said he about to leave
'He{said} he {~as} leaving.'
says
1S
d3a cannot be preceded by an NP object and therefore cannot be
considered a postposition.
It always occurs in a position be-
tween the subject NP and the verb, that is in a second posi-
tion, the position usually occupied by auxiliaries in G~gbe.
(2.121)
a.
a d3~
~~
m~l~
he about to eat rice
'He is going to eat rice. '
d3 a can be followed by n3 but never by le:
.
(2.122)
a.
e d3&
no
50
sab~l€
he about to prog. cut onion
,
'He is going to be cutting onion.
. ,
*b.
e d3~ le •
[.
50
sabul
Sometimes d3a can be preceded by le with no change in meaning.
For example (2.123) has the same meaning as (2.120):
(2.123)
e be
•
ye
le d3~
d3 6
he says he be about to leave
'He says he is leaving. '
,
Of the moda1s, only teu and ke can be used with d3 a .
te~
can precede or follow it.
ke can only precede it.
(2.124)
e tau d3&
y)
he can about to go
or e d 3 & t eu y)
'He can go (for sure).'
,
(2.125)
e ke d3~
y)
se
he mod about to go
part.
or a d3~ ke y) se
'If only he were going. I

---

-49-
(2.126) *a.
~ ~61~ d?~ v1
*b .
6 d ? ~ <j 6 I ~ V'j
d3a cannot be used with the future tense marker la or the ha-
bitual aspect marker na:
(2.127) *a.
mr I ~ d?~ d?6
*b .
mr d? ~ d? 6 n~
Finally d3a cannot be the nucleus of a predicate phrase
and it is not cliticizable.
As mentioned earlier, the habitual aspect marker ~ always
follows the main verb in a sentence.
It does not have a lexical
tone of its own because the tone it bears is always the same
as that of the preceding verb.
Na cannot be the nucleus of
a predicate phrase because it must always be preceded by a verb.
Na is not cliticizable.
The only constituents of the auxiliary
group na can be used with are teu and ke.
,
(2.128)
6
t~ii k:5 k6 n~ .
kploa
he can lift hab. table-the
'He can usually lift the table. t
,
,
,
(2.129)
6
ke
VI
na
se
he mod. go hab. part.
, If only he had gone!'
,
,
(2.130) *a.
e
I~ k:5 k:5 n~ kploa
,
1.,
*b.
6 d61~ k6k6 na kpl;)a
0
,
*c.
~ le k6 k:5 n~ kp loa
The fact that na always comes after the verb while all the
constituents so far considered as auxiliaries precede the verb
,
is another instance of Gegbe having characteristics of both SOV
and SVO word order.
As far as the nature of na is concerned,

---

-50-
there is no empirical reason for considering it an auxiliary.
It can very well be classified as an adverb since, as will be
discussed in section 3 of this chapter, in G~gbe most adverbs
occupy a position after the verb and they do not move to any
other position in a sentence.
Na too always occupies a posi-
tion after the verb.
However I still consider na an auxiliary
because it is more closely related to the verb than other ad-
verbs which can be separated from the verb by an object or pre-
positional phrases.
One last item to be considered in the group of aspect
markers is the use of the verb v~ to express the perfective
aspect as expressed in (2.131) below:
(2.131)
d~v(-w6
d6
n6
v~
child-pI. eat thing finished _
'The children have eaten.'
This use of v~ does not however make it an auxiliary.
Rather,
in sentences such as (2.131). what we have is a case of a verb
being used to modify another verb.
(2.131) can also be trans-
lated as 'The children have finished eating'.
Besides, v~ can
be used as a main verb in a sentence.
(2.132)
ga I (-~
v::,
gali-the finished
'There is no more gali.'
Furthermore, this use of two verbs in a sentence to express
one action is very common in G~gb~ (and other Kwa languages for
that matter).
It is as if each detail of the action is spelled
out.
Cf.
(2.133) and (2.134):

---

-51-
,
(2.133)
a.
e 56
k6po-a gba
he took glass
break,
'He broke the glass.
~
b.
~
gba
k6po-a
,
'He broke the glass.
,
(2.134)
a.
~
I~
gbo~
56
wu
he caught goat-the take kill
'He killed the goat. I
b.
~ wu gb3~
'He killed the goat. I
In summarY,'of all the morphemes discussed in this section
on aspect markers, only l~ and no, d3a~and na can be considered
as auxiliaries.
Except for na,. they all occupy the second posi-
tion in a sentence, one of the characteristics of AUX.
Ko and
d3! are postpositions used after noun phrases or nominals.
W~
and vo are true verbs.
Gbe is a particle used in adverbial
phrases.
2.2.3
A Category AUX for G~gbe
,
Now that we have determined which items in Gegbe can be
considered as auxiliaries, let us turn to the question of de-
termining whether there is any justification for including
a category AUX in the Phrase Structure rules of G~gbe.
In
order to do this, let us turn once more to the characteristics
of AUX discussed in Steele (1978) which were examined earlier
,
in section 2.1 of this chapter and see how the Gegbe auxiliaries
relate to them.

---

-52-
I.
In languages with AUX, AUX tends to prefer a certain
position in the sentence.
As has been obvious all through the discussion in section
2.2.2, all the constituents in G~gbe except na tend to occupy
the second position in the sentence, that is, a position be-
e
tween the subject NP and the main verb.
ef.
(2.135).
(2.135)
~
J ~
t , l
y
eu
no
du
gal(
•
he fut can prog. eat ga1i
NPS[T
M
A
] V
NPObj
,
'He will be able to eat ga1i.
11.
The order of constituents in AUX is generally, Tense,
Modality, Aspect.
As the ungrammatica1ity of (2.136) shows, the auxiliaries
,
of Gegbe generally follow that order.
(2.136) *lJ
n:>
NP SeA
The only exception to this order of occurrence are cases where
d3~ can occupy a position before the modal as in (2.137) below.
Notice, however, that in such cases, 1~ cannot be used and d3a
then becomes both an aspect and a tense marker.
(2.137)
marie
d?~
ud n:>
y)
as)me
Marie about to can prog. go market-in
'Marie is going to be able to go to the market.'
111. Languages with AUXv generally have subject agreement.
As the discussion in section 2.2.2 showed, none of the
auxiliaries of G~gbe can be considered main verbs in their own
right.
le and n~ which at first look may appear to be auxi-
1iary verbs, are not; they are verbs like any other verb in

---

-53-
,
Gegbe even though they are complementary in their syntactic func-
,
tion.
Cf. section 2.2.2.3 (PP.38-44).
Therefore Gegbe can
be considered a language with AUZv and, as should be expected,
it has no subject agreement of any sort.
This is exemplified
in sentences (2.l38a,b and c). . , , , ,
(2.138)
a.
marie
I~ t~u n::> yl aSlme
,
'Marie will be able to go to the market.
,
, , ,
b.
kof i
I~ t~u n3 yl aSlme
'Kofi will be able to go to the market. I
c.
devf-:S I~ t~u n5 y1 as1me
'The children will be able to go to the market. I
Even though in (a) we have a feminine singular subject, in
(b) a masculine singular subject and a plural subject in (c),
the rest of the sentence is the same in all three cases.
There
18
is no agreement in gender or number expressed.
,
According to this analysis of auxiliaries in Gegbe in the
light of characteristics I, 11 and III above, we have a pretty
,
good case for setting up a category AUX for Gegbe in the phrase
structure rules.
However, a problem arises with the placement
of na, the habitual aspect marker, since this auxiliary always
occupies a position after the main verb in the sentence.
If
we have a category AUX in second position in the PS rules, na
will then have to be generated in that position.
We will need
an obligatory rule moving na into the position after the verb.
The next problem to be solved is, what node will na be attached
to after it moves out of the AUX?
Due to the fact that having
the AUX in the second position is one of the main reasons for

---

-54-
,
.
setting up a category AUX for Gegbe, we cannot have a second AUX
position after the verb in the base component.
The only solu-
tion I can think of at this point is to sister-adjoin na to
the main verb at the surface level under a node "aux".
The na-
movement rule can be expressed as follows:
(2.139)
na-movement rule (obligatory)
.. 1 o 3+2 4
19
condition:
2 = na
(2.139')
S
..
S
N~VP
NP
I ------VP
I
I
/ ' - - .
~
Afi
na
V
pp
Afi
V
aux
pp
I
~
I
I
A
yi
N
P
yi
na
N
P
I
I
I
I
asi
me
asi
me
Finally, of the two types of AUX discussed in Steele (1978),
,
Gegbe shows primarily characteristics of an AU~V:
1) none of
the elements included in this category can be considered a verb;
2) there is no strong cases of subject agreement in the lan-
guage; 3) G~gbe does not display characteristics of a rigid SVO
word order type, a necessary condition for an AUXV.
,
3.
The VP Category in Gegbe
3.1 Introduction
Most syntactic analyses since the advent of N. Chomsky's
'Standard Theory' as defined in Syntactic Theory and Aspects
have stated the phrase structure rules for different languages

---

·-55-
in terms of categories such as Noun Phrases
(NP), Verb Phrases
(Vp), Prepositional!Postpositional Phrases
(PP), Adverbial
Phrases (AdvP), etc ••• However, in the last few years many in-
vestigative studies in the field have raised the question regard-
ing the validity of Such a universal view of the grammar of dif-
ferent languages, especially the tendency to include a VP cate-
gory in the anlysis of languages for which there is no reason
to do so.
One such study is John V. Hinds'
(1974)
"On the status
of the VP node in Japanese" in which he concludes that there is
no emprical reason for a VP category in Japanese, a language
with a primarily SOV word order, because the so-called Japanese
VP constituent does not stand,any of the traditional tests for
VP constituency.
Even the status of the VP constituent for
English, the language on· which Chomsky's theory was based, has
often been questioned; one of the studies dealing with th~s sUb-
ject is
MCCawley's
(iL970)
"English as a VSO language" in which
the author suggests that, because of some of the surface structu-
res found in the language, there is really no strong reason
for:prefering an SVO word order over any other and having a VP
node for English.
Arthur Schwartz, in his paper entitled "The VP constituent
of SVO languages"
(1972), deals with this problem of a universal
VP category in great detail.
His arguments are based on four
main hypotheses with respect to the distribution of the VP consti-
tuent in the three main systems of word order, namely,
VSO,
SOV, and SVO.
The hypotheses are stated as follows:

---

-56-
(1)
All systems have VP.
(2)
Only SOV and SVO systems have VP.
(3)
No' systems have VP.
(4)
Only an SVO system has VP.
(p.
221)
Along with the~e hypotheses he also has some general constraints
that they presuppose with respect to notions such as agent, pa-
tient, subject and predicate.
The constraints are stated thus:
I.
Agent to precede patient in the unmarked transitive
construction.
n.
"Subject" to be a nominal immediate1?, dominated by
S node.
Ill.
"Subject" to have a uniform time orientation with
respect to predicate.
IV.
Only one nominal to be immediately dominated by S
node.
(p.
221)
In his analysis which we will not cover in great detail here,
Schwartz rejects hypotheses (1) and (2) on the basis of their
inability to describe his data accurately, data including
examples from languages with SVO, SOV and VSO orders, transi-
tive, intransitive and ergative systems.
Even though it seems
more accurate in describing the different systems, hypothesis
(3) which presupposes constraints I, 11 and III fails to explain
the fact that, as the data in the paper shows, "while VSO and SOV
systems manifest an openness with respect to movement toward
the predicate, SVO systems show a resistance to interrupting

---

-57-
the predicate-patient sequence" (p. 225).
In the case of hypothesis (4) the crucial constraint is
constraint 11 which requires that the "subject" nominal be im-
mediately dominated by S node.
In this case, the patient no-
minal in an SVO system can never be the subject of an unmarked
transitive sentence.
This explains the fact that as a rule of
universal grammar, SVO-ergative systems are ruled out.
There-
fore hypothesis (4), even though it is more costly because it is
language specific, must be chosen over hypothesis (3) which seems
more universal in nature.
In light of these hypotheses and conclusions from Schwartz's
paper, let us try to answer the question which is the subject
of this section, namely,
'is it necessary or even desirable
,
to set up a category VP for Gegbe?'
3.2
Constituency Test for G~gb~ Predicate Phrase
In order to answer the question of whether G~gbe should
have a VP category, we will be primarily looking at simple
affirmative sentences in the language, that is, sentences which
do not have an emphatic meaning, or do not include embedded
clauses of any sort since such sentences will most likely be
derived by transformations applying to simple underlying sen-
tences.
Furthermore, the predicate phrases in the sentences
in this section will not include auxiliaries.
These latter
will be dealt with in the following section.
Consider the
examples below:

---

-58-
, ,
(2.140)
atf yl
aSlme
le n5a
9b6
es:,
Afi went market-in at mother-her near yesterday
Subj V
Locative
Locative
Advtime
'Afi went to the market near (to see) her mother
yesterday. '
(2.141)
ad pie
ga If
n<!i
rnf
Afi 'bought gaH
for
us
Subj. V
Objdirect Prep Obj. d'
~n ~rect
'Ali bought ga1i for us. '
(2.142)
devf~
n<!i
mf
oH
child-the gave us
oranges
Subj
V
Objindirect Objdirect
'The child gave us oranges.'
(2.143)
kofi
du
. n6
k~b<!
le t!ft!a
gb6
Kofi ate thing quickly
at teacher near
Subj
V
Obj
Advmanner Locative (PP)
le suk61u es:,
at school yesterday
PP
Advtime
'Kofi ate quickly in the presence of the teacher
at school yesterday. '
Each of the sentences above contains a subject and a verb f01-
lowed by a set of constituents.
In this section, we will try
to determine whether the verb and the items which follow it
should be considered as one constituent group generated under
the VP node or whether they should be considered as separate
constituents.
In other words. do G~gbe simple sentences have
phrase structure (2.144a). (2.l44b),
(2.l44c,or (2.144d) at the
underlying level?

---

-59-
(2.144)
a.
s
c.
s
NP~P
~
V
NP
pp
AnV
NP
v
NP
pp
AnV
b.
S
d.
S
NP~V
NP
VP
V~PP
~
V
NP
At this point some remarks are in order concerning G~gbe
transitive and intransitive constructions.
In intransitive sen-
. tences of the form of
(2 •. 140) above, time adverbials 20 can be
moved to the beginning of the sentence without making the sent en-
ce emphatic.
Thus
(2.145) has the same meaning as
(2.140):
(2.145)
as5a 2l
atf yl
aslme
le n5~
gbo
yesterday-the Afi went market-in at mother-her near
None of the two pre-/post-psitional phrases can be moved to the
front.
, ,
(2.146)
*a. asTmea atf yl
le
n5a gb:S as:'
, ,
, ,
*b.
la nSa gb6 ~ at( y1
eso
However, such preposing is perfectly acceptable in emphatic sen-
tences with ~ (emphasis marker) :
, ,
(2.147)
~n5a
gb6
y~
at( yl
le aslme
as:'
mother-her near emph. Afi went at market-in yesterday
' I t is to see her mother that Afi went to the market
yesterday. '
In transitive sentences like (2.143)
above place adverbials
and time adverbials can be preposed individually or together
without making the sentence emphatic~ manner adverbials cannot.

---

-60-
(2.148)
I~ tJftJ~
gb~
I~ s~k61~ is5i
kbf( ~~
at teacher near at school yesterday-the Kofi ate
nu
k~b~
thing quickly
'Near the teacher, at school, yesterday, Kofi ate
quickly. '
(2.149) *k~b~~ koff 2~ n6 le tJftJa gb~ le suk61~ ~s~
With the emphatic ~ (2.149) is perfectly grammatical.
Direct object NP's and indirect object NP's can never occur
before the verb in simple non-emphatic sentences.
Even in cases
where the indirect object is emphasized only the noun moves to
the front of the sentence, leaving the preposition behind.
For
example in (2.141) only m! can be preposed becoming the emphatic
miaw6; nA remains in the same position.
Cf.
(2.150):
(2.150)
mr~w6
yl!,
aff pIe
gal f n~
us-emph-emph Afi bought ga1i for
'It is for us that Afi bought &a1i. '
The only type of non-emphatic movement that can occur with NP
objects is with the objects of the verb na 'to give'.
For
example in (2.142) mi and Qt! can switch places with no changes
in meaning.
The same is not true of gal! and m! in (2.141).
As is obvious from the discussion so far, none of the cons-
tituents following the verb in a simple non-emphatic sentence
can move to a position between the verb
and the subject NP.
Expressions of time and place can move to initial position but
never between the verb and the subject or the verb and its
objects in transitive sentences.
Expressions of manner never
move to initial position.
However, in case of emphasis any
of these constituents can move to an initial position.
Therefore,

---

-61-
we can say at this point that as a general Tule, object NP and
adverbs of manner have a closer relationship to the main verb
than the other constituents in the sentence.
So, in case a VP
constituent is deemed necessary for G~gbe, it will most
likely
be of the form in (2.151).
(2.151)
VP
+
V(NP)
(NP) (ADV )
m
where thefirst NP is the direct object and the second NP the in-
direct obj ect.
ADV
stands for adverbs of manner.
m
However, an exception must be made in the case of intransi-
tive constructions.
As said earlier, in such constructions only
the time adverbia1s can be preposed in non-emphatic sentences.
All other constituents which come after the verb cannot.
Cf.
(2.145a and b) above.
Nevertheless, in transitive sentences,
expressions of place can be moved to initial position.
Faced
with such a situation we may wish to include a special clause
in the grammar differentiating between a transitive VP and an
intransitive one.
Such a solution is not very economical and
is, to say the least, undesirable.
Since in both types of cons-
tructions adverbia1s of manner cannot move to initial position
one possible solution,at this point would be to adopt
PS rule
(2.151) above as a rule of the grammar of G~gbe since it best
describes both kinds of constructions.
Pre-/post-positiona1
phrases and adverbs of time and place
will be outside the VP
22
and directly dominated by S. More will be said about this in 3.3.
One final point that needs clarification at this point is
the preposing of constituents as expressed in (2.148) above

---

-62-
(repeated here for convenience).
(2.148)
I~ tSftSa
gb~
le suk~lu ~s~a
kof( 2u
at tea~her near at school yesterday-the Kofi ate
n6
k~b~
thing quick
'Near the teacher at school yesterday, Kofi ate
quickly. '
As this example seems to suggest, all three constituents, le
!f:!tIa gb~ 'near the teacher', le sukulu I at school', and eso.
'yesterday', might seem to form a whole since they can move
together and furthermore, the last constituent ends with the
definite marker ~ generally used as a definite article which
always comes at the end of an NP.
(Placing the! at the end of
each individual constituent would render the sentence unaccep-
table).
However, this is a misleading fact because as the
examples in (2.152) shows these constituents do not always
have to move together ..
(2.152)
a.
ti.....1ll.:t.U 9 b~ ~ ka t f <1 U n~
,
c.
es5 a kaf f du
nu· ••
•
This behavior of these constituents support the idea that
pre-/post-positional phrases and adverbials of time and place
should be directly dominated by S; this way they can easily
move to initial position since in both positions they will be
,
under the same node.
Thus we can state for Gegbe at this point
PS Rules I and 11 of the following form:

---

-63-
(2.153)
PSR I.
S ... NP VP (PP) (ADVt' )
,p
PSR n. VP... V (NP) (NP) (ADV )
m
However these are very sketchy rules which are bound to change
when other facts such as constructions with auxiliaries and
serial verb constructions are taken into account.
,
Going back to Schwartz's analysis, it is clear that in Gegbe,
the object NP in an unmarked sentence can never come to be do-
minated by S or occupy a subject position since the subject NP
always precedes the verb and hhe object NP always follows. This
fact then is partly in agreement with Schwartz's Hypothesis (4).
,
3.3
AUX and VP in Gegbe
In section 2 of this chapter, we have presented arguments
,
justifying the need for a category AUX in Gegbe phrase struc-
ture rules.
In the present section, we will go one step further
and examine arguments for or against including the AUX in the VP
node.
Consider the following examples:
,
,
(2.154)
kef f
I~
t~u no
yl
tff'j
Kofi fut can prog
go now
[ T
M
A ]
,
"Kofi will be able to go now.
(2.155)
'atf dol~ pie galf le ak6wa gbo
le aslme
as;)
Afi must buy ga1i at Akuwa near at market-in tomorrow
[M
]
'Afi must buy ga1i from Akuwa at the market tomorrow.'

---

-64-
(2.156)
kof( t~D k~k~ na
kpl5a
Kofi can lift hab. table-the
EM]
E A ]
'Kofi can usually lift the table.'
As discussed in section 2, the auxiliaries which occur
between the main verb and the subject have a rather rigid posi-
tion.
Most of them always occupy the same position in the sen-
tence.
Furthermore, the order of occurrence of constituents
within the AUX is also consistent.
They always occur in the
order Tense, Modal, Aspect.
The only exception in this case ~s
d3A 'inchoative' which as we recall may have tense overtones
in some circumstances.
The question in the present discussion
is:
are there any justifiable reasons for making AUX a part of
the VP in G~gbe or should the two categories be kept separate
and individually attached to S directly?23
In G~gbe simple sentences very few items have freedom of
movement.
As said in section 3.2, PP's other than indirect
object NP's,and adverbials of time and place are the only cons-
tituents that can move and they generally can only move to ini-
tial position.
Therefore they can never come between the AUX
and the main verb.
Cf.
(2.157) and (2.158):
(2.157) *kof( I a
t~d n~
f (fl
y)
Kofi fut can prog now
go
(2.158) *~f( ~61a le ak6wa gbo pie gal( le as)me
To form negative sentences in G~gbe, the morphemes mu and 0
are inserted, the first in front of the item being negated and
the second at the end of the sentence.
In the following

---

-65-
sentences, kef! (2.159) and the predicate phrase (2.160) are
negated.
The mu can never occur between~he AUX and the verb
(2.161).
, ,
(2.159)
mo
kof( V~24
I~
t~6 no
y I
0
Neg. Kofi ernp,
will can prog. go part.
'It
is
not Kofi that can go.'
,.t
, ,
(2.160)
koff m~
t eu VI
0
Kofi neg-fut can go part.
'Kofi will not be able to go.'
(2.161) *kof ( I ~ tJ~ n5 mJ vl <>
It would have been easier to prove constituency if mu and
e always delimited the item being negated, that is, if we could
have a sentence of the form of (2.162) as a grammatical sen-
tence showing the subject as the scope of negation:
(2.162) *mu kef' V~·o I~ t~ij no V)
As it appears, however, the 0 is sentential in scope, not
phrasal.
The emphasis marker ~ mentioned at various times throughout
the discussion can also be used as a test of interruptability
for constituency.
y~ generally occurs at the end of the item
being emphasized.
For example~ in the following sentences of
(2.163), Af! (b, e), la tea pIe gal! le asime (c), and asime
(d) are emphasized.
, , ,
(2.163)
a.
atf I~
tJ6 pie ga If le aSlme
Afi will can buy gali in market-in
,
, ,
b.
at( V~
I~ t~~ pie ga I f le aSlme
(emph. )
c.
af f la tJ& pie ga If le , , ,
aSlme
yJ
(emph. )

---

-66-
(2.163)
d.
as)me y~ id r la t~t pie ga I r le
(emp. )
t~6
, , ,
,
e.
af (
y~
ma
pie gad le aSlme
0
(emph) (Neg+l8.)
(part)
Placing the Y.l innnediately after the AUX or even after the
object gal! will make the sentence ungrammatical:
,
,
ga I r
,
, ,
*f.
aff IcS Uli ye pIe
le aSlme
, , ,
*g.
i:lf (
la t~6 pie ga I ( y~ le aSlme
A final test of constituency that can be used at this
point is that of nominalization.
In G~gbe. the predicate, phrase
can be nominalized; in such cases the constituents (with a re-
duplicated form for one-syllable verbs) move to initial position
,
followed by an optional y! (emphasis marker) and the verb W:J
('do', 'make') is used to fill in the vacant position:~5
Consider the following examples from G~gbe:
(2.164)
kof( le
da
kp~
<;16
SI-me
le funu
Kofi prog. throw stone into water-in at there
'Kofi is throwing stones into the water over there.'
(2.165)
af( I~
ua dk:5
kpf5a
yi na
tas(
Afi will can lift (carry) table-the go give aunt
. I e as I -me
at market-in
'Afi will be able to carry the table to aunt
at
the marke t. '
(2.166)
ad'36 d61a ·Ie SI na dev(:5
v2>
Adzo must bathe to child-plo finish
'Adzo should have finished bathing the children.'
The nominalization of the predicate phrases in the above exam-
ples yields the following:

---

-67-
(2.164)"
a.
~pk~
dada
~6
si-me
le f~n~
(y~)
stone throwing into water-in at there emph.
keff le
wo
Kofi prog. do
'Throwing stones into the water is what Kofi
is doing over there. I
,
(2.165)
a.
~kpI5&
k:5d
YI n&
t&s(
le asrme
table-the carrying go give aunt at market-in
,
(Y~)
aft I ~
t~u wo
emph. afi will can do
'Carrying the table to aunt at the market is
what Afi will be able to do.'
(2.166)
a.
es'i
l€lle
n& d€lvf6
vo
(Y~)
ad?6
water bathing to child-pI finish emph. Adjo
d61& wo
must do
'Finish giving a bath to the children is what
Adjo should have finished doing.', .
As these sentences show. the auxiliaries le. la teu and d6la in
-
" " ' - - -
(2.164). (2.165). (2.166) respectively cannot be moved into sen-
tence initial position along with the other constituents of the
predicate phrase.
An attempt to do so will produce the fol-
lowing ungrammatical sentences:
(2.164) *b.
(2.165) *b.
,
,
,
,
(
L)
le
as Ime
ye
(2.166) *b.
eSl
d61~
•
Furthermore. when a negated predicate phrase is nominalized.
the negation particle mU optionally moves along with the phrase
into initial position even though as mentioned before it never
comes between the auxiliary and the main verb in a simple de-
clarative sentence.
Cf.
(2.167 a-d) below:

---

-68-
,
(2.167)
a.
at r m6
le
da
n6
o
.
Afi neg. prof. cook thing part.
'Afi is not cooking.'
, ,
Nomina1iza-
b.
en6
~a9a
(y~)
at r m6
le
W:J
0
tion
thing cooking (emph) Afi neg. prog. do part.
'What Afi is not doing is cooking.'
c.
~ n6
ma
da
(y~)
atr le
wo
thing neg. cooking (emph) Afi prog. do
'What Afi is doing is not cooking. I
An appeal to the "unit movement constraint,,26 makes this behavior
of the auxiliaries a strong indication that the AUX should not
be included in the VP since the two do not move as a single
constituent.
According to the above tests of interruptabi1ity (first
with mu - £, then with rl.), there is no reason not to generate
the AUX and the rest of the predicate phrase under the same
node, the node VP.
However, I believe that the cases of nomi-
na1ization with movement illustrated in (2.164)-(2.167) re-
present a stronger argument for not including the AUX in the VP.
It should also be pointed out here that the locative pre-
positional phrase le funu in (2.64) and the verb vo in (2.166)
do not always have to move along with the other constituents
while le asime in (2.165) must.
ef.
(2. 164c) - (2 . 166 c) below:
(2.164)
c.
hp~
dada
96
s'lme
(y~)
katf.le
stone throwing into water-in (emph) Kofi prog.
wo le tun6
do at there
'Throwing stones into the water is what Kofi
is doing over there.'

---

-69-
(2.165) *c.
hpl5~
dkj
y~ n6
t&sr
(ye)
at r
~ab1e-the carrying go give aunt (emph) Afi
I~
t~& w::, le as~me
£ut. can do at market-in
Tp mean 'Carrying the table to aunt at the
market is what Afi will be able to do.'
2 166)
,
,
I' I Y
L
d'
"
( L)'
L
"
,
,
(
•
c.
e 5 l e e
n a
e.v I;)
ye
ad ,0 <i 0 I a w;) V;)
'Finish giving a bath to the children is what
Adzo should have finished doing. I
Sentence (2.165c> can be grammatical if it is used to express
the idea that the action of carrying the table will be taking
place at the market and not from somewhere else to the market.
This is a question of outer locatives versus inner locatives.
According to a discussion of these types of locatives presented
in Tai (1977),27 inner locatives locate the recipient of an
action while outer locatives locate an event.
A prepositional
phrase of the form le + N28 can be either type of locative de-
pending on its semantic function.
In (2.165), le astme has an
ambiguous semantic function disambiguated in (2.165a) and (2.165c).
In (2.165a) le as1me locates tas!, the recipient of the action
of k~k~ y1 (inner locative) while in (2.165c) it is the action
itself that is located (outer locative). 29
From this observa-
tion, on the le + N type of locatives, we can conclude that
while inner locatives must be generated in the deep structure
as part of the VP, outer locatives must be generated under a
sister node outside the VP. The node under which the locative will
be generated in the deep structure will correlate with whether it is
an inner or outer locative.
In any case, the auxiliaries do not move

---

-70-
into initial position along with the rest of the VP.
In conclusion to this section, .it is obvious that we
have some reasons both for making AUX part of the VP and for
not doing so.
However I believe at this point that with the be-
havior of the AUX in VP norninali·zation~ we have a ·stronger case for
not making the AUX part of the VP.
Therefore, PSR I in section
3.2 should be changed to PSR If.
PSR If.
S
~
NP AUX VP (PP) (ADV)
and PSR 11 to PSR 11'.
PSR 11'. VP ~
V (NP) (NP)
(PP)
(ADV)
(V)
4.
Conclusion
In this chapter, I have tried first to show in section 1
,
that Gegbe is a language with a predominan~ly SVO word order.
This is shown by the order in which the major constituents
occur, that is, subject + verb + object.
However, it also
has some characteristics generally found in languages with an
,
SOV word order, that is, Gegbe has mostly postpositions instead
of prepositions; in genitive constructions, the genitive pre-
cedes the noun etc ... The presence of these SOV characteristics
may be due to the fact that Proto-Niger Congo from which Kwa
languages have originated has been hypothesized to have an SOV
word order.
In section 2, I have tried to show that the verb modifiers
used for tense, modality and aspect cannot be considered as
main verbs in their own right.
They are auxiliaries belonging

---

-71-
to a category AU~, to use Steele's terminology.
I have also
shown that it is desirable, for practical and economical reasons
and also based on some empirical reasons, to set up a category
AUX in the base component for generating these auxiliaries,
based on the position they occupy in a sentence and the fact
that the order of occurrence of the constituents in the AUX is
a rather consistent one.
In section 3, arguments have been presented for the need
to set up a category VP for G~gbe including AUX, the main verb,
the object NP's and the adverbs of manner.
To justify the,in-
elusion of AUX in the VP, the main argument used was that of
the non-interruptability of the order AUX + V by any other
constituent especially the PP's, the adverbials of time, the
emphasis marker ~ and the negative particle mu.
However,
cases of nominalization of the VP with movement to initial posi-
tion in the sentence have been presented against including the
AUX in the VP.
It has then been decided for a good reason that
AUX should not be made part of the VP.
The argument for set-
ting up the category VP in the first place was based on the
fact that in G~gbe simple non-emphatic sentences some of the
constituents, especially the NP objects and the adverbs of
manner do not move free.ly, they always follow the. verb in a
certain order, that is, NP then ADV of manner; thus they form
a kind of unit with the main verb.
Two tentative PS rules, repeated below, have been proposed
for expanding Sand VP.

---

-72-
PSR I'.
S
~
NP AUX VP (PP)
(ADV)
PSR 11'. VP ~
V (NP) (NP) (PP) (ADV)
(V)
In light of the discussion in the following chapter on
serial verb constructions, some modifications will have to be
made in these rules, especially PSR I',

---

FOOTNOTES TO CHAPTER 11
*The main reason for considering word order in this study
is the fact that my analysis of auxiliaries is based on S.
Stee1e's paper "The category AUX as a language universal"
(1978) and she uses word order as one of the basis for her
discussion.
1There is disagreement among linguists concerning the idea
of languages being in a transitional state from one word order
type to another.
The fact is that, even though each language
is said to have one type of word order or another, there are
always exceptions to the rule.
Furthermore, there are langua-
ges which have been qualified as being in a transitional state
and yet no subs tantia1 change had· been noticed over a long
period of time.
Therefore some reservations have to be made
concerning Givon and Hyman's contention that some Kwa languages
are in a transitional stage from an SOV word order to an SVO
word order.
There is no way to tell whether they are ever
going to reach a perfect SVO word order.
We should rather say
that these languages have reached a stage where they have a new
word order primarily with remnants from the word order of the
mother language.
This is the stand"John Hawkins adopted in his
analysis of word order in his paper entitled "lmp1icationa1 uni-
versa1s as predictors of word order change" (1979).
-73-

---

-74-
2According to J. Greenberg,
the SOV,SVO and VSO word
orders are the most common in languages of the world.
However,
only the SOV and SVO are found in Kwa languages.
Therefore
the discussion in this section will only deal with those two
types of word order.
3Bartsch, Renate and Theo Vennemann,
1972,
"Semantic
structures, a study in the relation between semantics and
syntax" .
4vennemann, Theo, 1973,
"Explanation in syntax".
SIntentional verbs are not always followed by an infini-
tive verb in Gegbe.
In fact the infinitive is not commonly
used in the language.
6see footnote 2 in Chapter I.
7sometimes, as mentioned in section 2 of Chapter I
(p. 5),
the possessive may precede the noun in the first and second
persons singular.
Bna is derived from the verb na 'to give'.
This na has
often been considered as a verb used in a serial construction
with the other verb or verbs of the sentence.
(Cf.
(2.27).)
I disagree with that assumption.
More will be said about this
in Chapter Ill.

---

-75-
9 Ross, J., 1967a, "Auxiliaries as main verbs".
lOsteele, M. 5.,
1978, "The category AUX as a langauge
universal".
In J. Greenberg,ed. Universals of Human Language,
VoL Ill, p.ll.
lles~ can be used to mean 'yesterday or 'tomorrow'. Ad-
ding a relative clause after it
(in this case 'that went by')
helps di£ferentiate between the two meanings if necessary.
12However, as Dr. Carleton Hodge pointed out to me, this modal
could be derived from d6 with the meaning of
'have' (possession)
or-
plus l~ (future:
'about to').
Compare the English expression
'to have to'
used to express the idea of obligation.
13Except for the modals, a member of a category of auxi-
liaries (tense or aspect)
cannot generally be modified by other
members of the same category
141~ and n3 are two morphemes used to express the progres-
sive aspect.
They are complementary in their syntactic role.
More will be said about these two morphemes in section 2.2.2.3
of this chapter.
15~ does not have a lexical tone of its own.
It always
bears the same tone as the preceding verb: so if the verb has
high tone ~ gets a high tone and if the verb has a low tone
the tone on ~ is also low.
Cf.
(i) and
(ii)
below:

---

-76-
,
(i)
~
pie na
ga I r
he buy hab. gali
,
'He buys gali.
(ii)
~
I~
n~
hev(
he catch hab. bird,
'He catches birds.
16 The idea of "every" is expressed in G~gbe by repeating
the noun and inserting sia between the two words.
(i)
ame
5r~
(a)me
person part. person
I Everybody. I
(ii)
ad
s(~
(a)tf
tree part. tree
'Every tree.'
17,
:!.
.!.
I.
at
.I
!. .r.
l)cl)c
_is derived from .at~ be l)cI)C
(tree, wood-
particle-breaking) and the same way sUkulu yiy1 is derived
from sukulu be yiyl (school-particle-going).
18 However, there is a case of correference in the language
where number agreement is expressed in reference to the subject
of the sentence when this subject is a coordinated NP.
In such
cases, as shown in (i)-(iii), a plural pronoun is used corres-
ponding to the person of the NP's in the subject:
(i)
aft k6
marie w6
YI
Afi and Marie they go
I Afi
and Marie went. I
(ii)
ne kG
at( mr yi
me and Afi we go
'Afi and I went.'

---

-77-
,
,
'\\
'\\
(iii)
wo
ku
ma r i e mI
y I
you(sg.) and Marie you(p1.) go
'You and Marie went. '
When the subject is emphasized the pronouns are not used.
Cf.
(i' ) - (iii ' ) :
(i' )
af r ku
marie y~ y)
emp.
'It was Afi and Marie who went.'
(ii')
ne k6 aff yIJ y)
'It was Afi and I who went. '
(iii')
wo kJ marie y~ y)
'It was you and Marie who went.'
As it is obvious from the above examples, this process has no
bearing on the form of the verb and cannot be considered a true
case of agreement.
19In case AUX contains more than one auxiliary, th~ condi-
tion on rule (2.139') should state that only the auxiliary na
can be moved out of AuX.
~OIn G~gb~, expressions of manner, time and place can be
rendered
with adverbs or with pre-fpost-positiona1 phrases.
No matter which form is used, the different expressions in each
group behave the same way with regard to their movement in
the sentence.
21 When adverbia1s of time or place are preposed (non-
emphatic) the definite article a is always placed at the end
of the expression.
22 See p. 57f.

---

-78-
23 In other words should the grammar of Gegbe include the
surface structure in
(i)
or the one in
(ii)?
(i)
S
(ii)
S
~
~
NP
VP
NP
AUX
VP
~
AUX
V
•••
24
,
Ye .1S used
h
w en any cons t .1 t uen t
0 ther th an t h e
VP'1S
negated because such a constituent must move to sentence initial
position, a position of emphasis.
25schwartz (1972)
tells of a similar case in Indonesian
where an enclitic -kah is used for emphatic questions.
This
morpheme is attached to the emphasized element which is then
moved into initial position.
According to the
'unit movement
constraint', any sequence of elements followed by -kah that
move into initial position must form a constituent.
The fol-
lowing sentences from Schwartz I
paper exemplify this:
Base
The patient went to-the-hospital yesterday?
yesterday-kah the patient went-to the hospital?
To-the-hospital-kah the patient went yesterday?
Went to-the-hospital-kah the patient yesterday?
26 See Scwartz (1972) p. 219.
27 For a brief discussion of inner and outer locatives, see
Tai (1977),
"Semantic and Syntax of inner and outer Locatives."

---

-79-
28,
N '
d h
f
.
Le +
1S use
ere
or conven1ence.
This type of pre-
positional phrase can be of the following forms:
(i)
le + Noun
(le + asl
'at the market
(in another town) ')
(ii)
le + Noun + Postposition
(le + asi + me
'at the
market (in the same town) ')
(iii)
le + Adverb of place
(le + funu
'over there')
29This behavior of the.locative is dependent upon the type
of verb used.
The outer versus inner distinction is valid only
with verbs of motion.
In
(2.164)
there is no such distinction
because da 'throw'
is not a verb of motion.
In
(2.165)
the
compound verb k~k~ yi
'take to'
is a verb of motion.

---

CHAPTER III
,
¥
SERIAL VERB CONSTRUCTIONS IN G~GBE
1.
Introduction
In most West African languages of the Kwa group, there
are sentences which contain two or more predicate phrases
with one subject NP and one or more auxiliaries.
This will
be referred to as a
'Serial verb construction'.
This same
phenomenon has been observed in many languages of East Asia,
namely Chinese, Vietnamese, Thai and others.
Various attempts have been made by linguists to charac-
terize precisely what a
'serial verb construction' is.
This
construction is also referred to as verb serialization, verb
clusters, verbal complexes, double-base and multi-base com-
pression or simply serial constructions.
Further attempts
have been made to differentiate true serial verb constructions
from other types of constructions such as causative construct-
ions { Stahlke (l972»,
splitting verb constructions ( Awobu-
luyi
(1967»,
close-knit constructions
{ Bendor-Samuel
(l968»,
etc •••
The main reason for wanting to differentiate between
true serial verb constructions and these other types of con-
structions is the fact that in the case of the former there
is generally just one subject at the beginning of the complex
sentence for the different verb phrases; this subject is the
-80-

---

-81-
surface as well as the deep sUbject.
If there is only one
auxiliary,
it-too is generally located at the beginning of
the sentence.
In the case of the other types of construct-
ions, there is more of a dependent relatio~ship between the
first sentence and the other predicate phrases,
in other
words, the latter seem to modify the former in certain ways
or to result from i t .
Furthermore, there may be other auxi-
liaries different from the one(s)
in the initial predicate
phrase overtly expressed in the other predicate phrases.
Sometimes the subject of the initial predicate phrase cannot
logically be interpreted as the sUbject of some of the other
predicate phrases in the sentence, instead the object of that
initial sentence is the only possible logical sUbject of those
predicate phrases. l
The following examples illustrate the
different types of sentences mentioned above.
(3.1)
marie ple gal! yi n~
Qev!~
Marie bUy gali go give children
'Marie bought some gali and took i t to the children'
(3.2)
marie Qa
te
l~
to
fufu!
l~
du
Marie cook yam prog pound pounded yam fut eat
'Marie cooked yam and is pounding i t to eat.'
(3.3)
marie tutu af! d?e
an!
Marie push Afi fall ground
'Marie pushed Afi and she fell.'
In (3.1) Marie is the surface as well as the deep subject of the
different verbs in the sentence and all the verbs are in the
simple past tense.
In
(3.2) Marie is also the sUbject of all

---

-82-
the verbs but in this case the verbs are in different tenses
and aspects: ~ is in the simple past, t6 is in the present
progressive and du is in the future tense.
In
(3.3)
Marie
,....-
is the subject of tutu but only Afi can be the logical sUbject
Whatever the case may be, all those constructions
have one thing in common: they all express one event or action
consisting of different subparts expressed sometimes in the
same sometimes in a different tense, aspect or mood with a series
of predicate phrases following each other and no overt coordi-
nating element.
Furthermore, there is only one overt surface
sUbject for the whole sentence.
Westermann
(1965)
expresses
this idea for Ewe in the following terms:
"The explanation for this [serial verb con-
structions] is that the Ewe people descri-
be every detail of an action or happening
from beginning to end and each detail has
to be expressed by a special verb: they
dissect every happening and present i t in
its several parts, whereas in English we
seize only the leading event and express i t
by a verb, while subordinated events are either
not considered or are rendered by means of a
preposition, adverb, conjunction or a prefix
to the verb." 2
Example (3.4) below illustrates this difference between English
and Ewe.
While Ewe uses two verbs,
English uses only one to
express the idea of
'visit'.
(3.4)
marie dltsa
yi aslme
ets~
Marie take a walk go market-in yesterday
'Marie visited the market yesterday'

---

-83-
2.
The Analysis of Serial Verb Constructions
The traditional method used by most Africanists to deri-
ve serial verb constructions has been from coordinated or em-
bedded type of sentences, with the application of deletion
rules to
eliminate the identical subjects, objects and auxi-
liaries in the non-initial predicate phrases as well as the
complementizers and the coordinators.
For such analyses see
Awobuluyi
(1967)
and Bamgbose (1974)
for Yoruba, Stewart (1963)
for Twi, Williamson
(1963)
for Ijo,etc •.•
More recent
proposals have been made to restrain the power of deletion
rules in Transformational Generative Grammar.
For example, in
Katz and Postal
(1964), i t is the opinion of the authors that
transformation operations must be reconstructible; arbitrary
elements cannot be introduced and then deleted in order to
arrive at the desirable surface structure (p.93).
Chomsky (1964)
advocates a close relationship between the deep structure and
the surface structure so as to facilitate recoverability of
items deleted with the use of transformational rules.
Finally
Jackendoff (1972) expresses his opposition to deletion rules,
even deletion under identity where the deleted items can easi-
ly be recovered; he states: "it would be a highly significant
restriction on linguistic theory i f all deletions under iden-
tity could be eliminated"
(p.382).
The main reason for this
recent position among linguists is the fact that very often,

---

-84-
in order to be able to apply deletion rules in their analyses,
authors have made up very far-fetched underlying structures with
items not easily recoverable from the surface structure because
they were too arbitrary.
In line with the above objection to deletion rules in
general, other attempts have been made to derive the surface
sentences from underlying sentences very similar to the surface
string, thus doing away with the need for deletion rules of any
sort.
Schachter (1974)
and Lord
(1974)
advocate an analysis
for s.erial verb const"ructions which does not require deletion
rules.
More will be said about these different types of pro-
posals later.
The purpose of this chapter is to examine the different
-
typ~s of serial verb constructions in G~gbe and in the light of
the different proposals already offered for deriving serial
verb constructions in other languages, try to find a suitable
L.
..
solution for the problem in Gegbe.
2.1
The Facts of G~gbe
In the analysis of these constructions,
I will deal with
the following cases:
1)
Sentences with one subject and one AUX for the different
predicate phrases; this subject is the surface as well as the
deep subject.
2)
Sentences with different auxiliaries but the same subject

---

-85-
for the different predicate phrases.
3)
Sentences with both subject and object of the initial
sentence as possible logical subject of non-initial predicate
phrases .•
4)
Sentences with the object of the initial sentence as the
only possible logical subject of non-initial predicate phrases.
5)
Comparative and superlative constructions.
6)
Sentences with 'split verbs'.
7)
Sentences with modifying verbs.
8)
Sentences with verbs used as prepositions.
In my analysis I will consider sentences of the type of
1) through 4)
above as serial verb constructions main~y because
of the fact that at the surface level only one sUbject is overt-
ly expressed and the other predicate phrases follow the ini-
tialone in serie~ and also because the actions expressed in
the non-initial predicate phrases follow from the action ex-
pressed in the initial predicate phrase.
For reasons that will
be stated later I
do not consider sentences of the type of
5) through 8)
as serial verb constructions.
I will first give
examples of the different types of sentences with an explana-
tion of the syntactic processes taking place in these senten-
ces.
Next,
I will try to make a proposal concerning the deriva-
tion of the different types of sentences.

---

-86-.
2.1.1
Sentences with the Same SUbject and One AUX
(3.5)
af! yl pIe gal! s~
va
~h6me
Afi go buy gali take come home-in
[NB ][vp] [ VP
] [vp
][ vp] [
pp
]
'Afi went and bought gali and brought i t home.'
(3.6)
afi l~
te~ va
\\
y~
pIe adI
s~
yl na
tas{
Afi fut. can
come
go
buy soap take go give aunt
'Afi will be able to go buy some soap to take to aunt. I
(3.7)
afi nu
b{ya
mu
Afi
drink
beer
get drunk
'Afi got drunk on beer.'
(3.8)
af! g.a
d~sl
ku
n~
m
Afi
cook
stew
take some give me
'Afi cooked stew and gave me some. '
(3.9)
ay~ba
la
d~6
yl
sukulu
Ayaba
fut.
leave
go
school
'Ayaba will leave for school. I
In each of the above sentences, there is only one subject
expressed at the beginning of the· sentence and, as sentences
{3.6) and (3.9)
show, there is only one AUX position in the
surface string occupied by l~ and te~.
These sentences can be
paraphrased with the repetition of the subject NP before each
of the non-initial predicate phrases, and in general that NP is
pronominalized.
In case any of the non-initial predicate
phrases contains a transitive verb with the object of the ini-
tial predicate phrase as its object, this object can also be
repea.ted either as a noun or as a pronoun.
Thus
(3. 5)
and (3.8)
can be ~araphrased as
(3.5') and (3.8').

---

-87-
,
(3.5')
af:! yJ., ~
pIe gal1,
~
s6
'E, 3~
va
ah6me
Afi go she buy gali she take it she come home-in
'Afi went, she bought gali,
took ,she came home. I
,
,
(3.8')
aft
da
desl
, ~
ku
J.,
~
na
m
Afi cook stew
she take i t
she give
me
'Afi cooked stew, she took some and gave me.'
These sentences can also be paraphrased using the conjunction
~y~ 'and' .
Cf.
(3.5")
and (3.8") :
,
,
(3.5")
at! yJ. ~y~ ~
pIe
gal:! ~y~ ~
s6
E
~y~
Afi go and she buy
gali
and she take i t
and
~
va
ah6me
she come
home-in
,
(3.8")
af! da
desl ~y~
~
ku
J.
~y~
~
na
m
Afi cook stew and
she take i t
and
she give
me
Though both types of paraphrase are acceptable in the langua-
ge, the one with ~y~ sounds more natural.
One thing that
,
should be mentioned here is the fact that in Gegbe, i t is
common to omit the object of a verb without the sentence being
ungrammatical,when the context permits i t .
Thus sentences
such as
(3.10)
and (3.11) below are perfectly grammatical sen-
tences in G~gbe.4
(3.10)
aft
s~
va
Afi
take come
'Afi brought (it) .'
(3.11)
~
na
m
he
give
me
'He gave me
(some). I
Sentences
(3.5)
through
(3.9)
can each be negated as a
whole, in which case the negation marker mU is placed imme-
diately after the subject NP and 2 is placed at the end of
the sentence.
In this case also,
the sentences are ambiguous.

---

-88-
For example, a sentence like
(3.7a),
the negative version of
(3.7)
can be interpreted as meaning: 1) Afi did not get drunk
on beer but on something else; 2) Afi did not drink beer and
she did not get drunk.
(3.7)
a.ad mJ
nu
bfVa m6
0
Afi neg. drink beer get drunk part.
'Afi did not get drunk on beer.'
In case just the non-initial predicate phrases are to be negated,
the coordinator ~
'but' must be introduced.
Though anyone
of these VP's can be negated,
the negation generally encompas-
,
ses any VP ( 's) between mu and Q.
For example,
let us take
sentence (3.6)
above:
,
ad I ~
t~u v~
v1 pIe adT
s~
v1 n~
t~sf
.
Afi fut.
can come go buy soap take go give aunt
'Afi will be able to go buy some soap to take to aunt.'
The negation can be introduced in the following manner:
,
(3.6)
a. aft m6
t~u v6
v1 pIe agT
55
v1 n~
Afi neg.+la can come go buy soap take go give
,
t6s f 0
aunt part.
'Afi will not be able to go buy some soap to take
,
to aunt.
,
,
•
,
b.
aff If!
t t$cr v~
v~~ ma
VI
pie adT
5:>
Afi fut.
can come but neg.+la go bUy soap take
v1 n6
t~sf 0
go give aunt part.
'Afi will be able to come but will not go buy
soap to take to aunt.'
,
c.
aft If!
t~cr v6
y1 v:>a m6
pie adT
so
Afi fut.
can come go but neg.+la buy soap take
V1 n~
tasL <>
go give
aunt
part.
'Afiwill be able to go but she will not buy soap
to take to aunt.'

---

-89-
,
,
(3.6)
*d.
aff I~
Uti v~
YI m6
pie adT
5:5
.
yl
Afi fut.
can come go neg.+la bUy soap take go
,
n6
t65f 0
give aunt part.
Whenever the part of the negated sentence after v5a contains
more than one VP, i t is ambiguous.
For example in
(3.6b)
the
ambiguity of that part of the sentence is expressed as follows:
1) Afi will not buy soap to take to aunt but she will probably
buy something else; 2) Afi will not buy soap or take i t to aunt.
Sentence (3.6d)
is unacceptable because of the absence of v~a
before m~.
There are few cases where the non-initial predicate phrases
can be negated without the introduction of v3a being necessary.
Thus
(3.8a)
below in wh.1.ch the VP ku n;;{m is negated is perfect-
ly acceptable with or withoutv~a.
(3.8)
a.
aff ga
d~s) {v5a) m6
ku
n6
m "
Afi cook stew (but)
neg. take some give me part.
'Afi cooked stew but she did not give me any. '
In my investigation so far, I
cannot find any particular pattern
setting
a·part cases where v~a can be omitted from cases where
i t cannot.
2.1.2
Sentences with Different AUX's but the Same SUbject
(3.12 )
kotf 55
bll
d,.! .
,
po
' r
dev
'a
Kofi take stick is going to beat child-the
'Kofi took a stick and is going to beat the child. '
(3.13)
a.
ak6 da
te
to
fuftf
I~
Aku cook yam pound
pounded yam fut.
eat
'Aku cooked yam pounded i t and will eat i t . '

---

-90-
(3.13)
b.
aku da
te
I~
.
t6
fuf~1 I~ QU
Aku cook yam prog. pound p.yam fut.
eat
'Mu cooked yam, is pounding i t to eat. '
, ,
*c.
ak6 I~
Qa
te
le
t6
fuf'Ui 16
cju
Mu fut. cook yam prog. pound p.yam fut.
eat
,
As the above examples show,
i t is possible in Gegbe to
have different auxiliaries for the different predicate phrases
of a serial construction.
This is also found in Yoruba but not
in Akan.
The above sentences can also be paraphrased with the
repetition of the subject NP and with eye.
Cf.
(3.12') 'and
(3.13a')
below.
(3.12' )
a.
kof( 56
atf , ~
d,6
po
devf~
Kofi take stick he be going to beat child-the
'Kofi took a stick and is going to beat the child.'
b.
koff 55,
atf
~y~ ~
d,~
po
d~vr~
Kofi take stick and he be going to beat child-the
'Kofi took a stick and he is going to beat
the child.'
, ,
(3.13a')
a.
ak6 ~a
te, ~
t6
fufui, ~
16
du
Aku cook yam
she pound p.yam
she will eat
"Aku cooked yam ,
she made pounded yam, she
will eat i t . '
, ,
b.
,ak6 ~~
t~
~y~ ~
t6
fuf'Ui 'y~ ~
I~
du
Aku cook yam and she pound p.yam and she will eat
'Aku cooked yam and she made pounded yam and
'she will eat i t . '
As the unacceptability of
(3.13c)
shows,
i t is not possible
to have the future tense marker
la preceding the present progres-
sive aspect marker
le.
There must be a logical sequence in the
occurrence of the tenses.
The negation of the sentences is the
same as the negation of the sentences in section 2.1.1 above.
The negative markers mu and 0 are inserted immediately after the

---

-91-
NP subject and at the end of the sentence respectively, making
the sentences ambiguous.
In case a non-initial VP is negated
independently from the initial one, v~a must be inserted before
md.
2.1.3
Sentences with Both Surface Subject and Object of Initial
Sentence as Possible Deep Subject of Non-initial Predi-
cate Phrases
(3.14)
aff 5:"
amTa
k:,
qe5 5
slme
Afi take oil-the pour arrive water-in
'Afi poured the oil into the water. '
,\\,
(3.15)
kos r
5&
k5poa
gba
Kossi take glass-the break
'Kossi broke the glass. I ,
(3.16 )
kosf
tutU k~poa
gba
Kossi push glass-the break
, Kossi pushed the glass and i t broke. '
(3.17)
aff yl yfJ
ak~ v~
Afi go call Aku come
'Afi. went to get Aku. and she./they came.
].
I
J
J
(3.18)
af r y'i kp 10
ak~ vti
Afi go accompany Aku come
'Afi went and brought Aku.'
(3.19)
affble
ak~
yl
Afi trick Aku
go
'Afi tricked Aku and she/they went. '
(3.20)
aft
't'
tu
u "
,
aku
"
yl
Afi push Aku go
'Afi pushed Aku along and they wen t. '
In sentences
(3.14)
through
(3.16)
the subject NP's Afi
and Kossi and the object NP's amia and k5p6a can be the deep
,
subject of k3 and gba in the respective sentences.
Thus
(3.14'),

---

-92-
(3.14")
and
(3.16 ' ), (3.16") are possible paraphrases of (3.14)
and (3.16)
respectively.
(3.14')
afT 56
amTa
~V~ afr k~
amTa
96
51me
Afi take oil-the and Afi pour oil-the arrive water-in
'Afi took the oil and poured i t into the water. I
, ,
(3.14")
aff 56
amia
~V~ amia
k~
q6
5 I me
Afi take oil-the and oil-the pour arrive water-in
'Afi took the oil and the oil poured into the water.'
,
(3.16' )
k~5 r
tutu k~p03
~V~ k~5r
gba
k5poa
Kossi push glass-the and Kossi break glass-the
'Kossi pushed the glass and broke it. 1
,
(3.16")
k~5 r
tutu R5poa
~V~ k5poa
gba
Kossi push galss-the and glass-the break
'Kossi pushed the glass and i t broke. '
In sentences
(3.17)
and
(3.19) either Aku or both Aku and
Afi can be the subject of the verb in the second predia:ate phrase.
But Afi alone cannot be the subject.
Thus
(3.17) can be para-
phrased as
( 3 • 17 I)
or ( 3 • 17 ") ~
(3.17')
at( VI V!J· ak6
~Vf3 ak6 v~
Afi go call Aku and Aku came
'Afi went to call Aku and Aku came.'
(3.17")
at! vl V5
ak6 ev~ w6
v~
Afi go call Aku and they come
'Afi went to call Aku and they came. I
In (3.18)
and (3.20),however, only Afi and Aku together can
be the subject of the second predicate phrase.
For example, if
in(3.18) Afi goes to bring Aku back, the only logical thing to
expect is for both of them to come back~ not just one of them.
In (3.20)
the action of pushing is continuous,
that is, Afi
pushed Aku t i l l they both got there.

---

-93-
Sentence
(3.14)
through
(3.20)
can be paraphrased with
, ,
eye+ NP sUbject .or NP.object if only one of the NP's is in-
tended as sUbject of the non-initial predicate phrases~ they
are paraphrased with ey~+ 3rd person plural pronoun wo when
both subject and object NP's are intended as subject of the
non-initial VP's.
In case the object alone is intended as
the subject,
i t is preferable to use the noun rather than the
.
.
G)., b V
h
pronoun s~nce ~n
eg e t ere is no gender specification for pro-
noun.
Thus
(3.19')
is preferable to
(3.19"):
,
(3.19')
ad ble
1:1 k~
6y~ akJ yl
Afi trick
Aku and Aku go
'Afi tricked Aku and Aku went. '
(3.19")
afr ble
ak~ ~y~ 6
y1
Afi trick
Aku and she go
'Afi tricked Aku and she went.'
In sentences (3.17)
and (3.18)
there are two verbs pre-
ceding the object NP
(Yl ~ in (3.17) and Yi kplo in (3.18».
In such cases, even though the second predicate phrase is non-
initial, only the subject of the initial sentence can be un-
derstood as the sUbject. of that predicate phrase.
The object
INP can only be the subject of VP's coming after it.
The negation of sentences (3.14)
through
(3.20) works
the same way as the negation of the sentences in the preceding
sections.
When the negation marker mU is placed at the be-
ginning of the first predicate phrase, the sentence is ambi-
guous~ when only the non-initial predicate phrases are negated,
v~a must be placed before
•
!ffi!•

---

-94-
2.1.4
Sentences with only the Object of the Initial Predi
cate Phrase as Possible Deep Subject of Non-initial
Predicate Phrases
(3.21)
at( tutu h6 d~e
anr
Afi push Aku fall ground
'Afi pushed Aku and Aku fell.'
,
(3.22)
marie do
ad y1
Marie send Afi go
'Marie sent AfL '.
.. ,
,
, --
avua
I~
kek~a
gi
dog-the
knock
bicycle-the fall
'The dog knocked the bicycle down. '
In the above examples, only the object NP's Aku, Afi and kekea
can be understood as the sUbjects of the non-initial predicate
phrases in their normal semantic interpretation.
This is further
supported by syntactic facts in the l~guage.
In case the
subject of the initial sentence is intended as the subject
of the non-initial predicate phrase,: i t has to be overtly
expressed.
Thus in
(3.21) , i f Afi falls after pushing Aku,
the sentence must be expressed as in
(3.21a).
(3.21)
a.
atr t ut u aku {~~~ } ~ 6H
d~e
Et n(
voa
Afi push Aku {and} she herself fall ground
but
'Afi pushed Aku but she herself fell down. '
Sentences
(3.21),
(3.22)
and (3.23)
can only be para-
phrased as follows:
(3.21')
atf t 't' , , , , 'k' dze'
'a-nf
u u aKu eye a u
)
Afi push Aku and Aku fall ground
'AH pushed Aku and Aku fell.'

---

-95-
,
(3.221)
marie do
aff ~y~ af f •
yl
Marie send Afi and Afi go
,
'Marie sent Afi and Afi went.
,
,,~
(3.23')
, --
avua
16
kek~~
~y~ kek~~
gg
dog-the knock bicycle-the and bicycle-the fall
'The dog knocked the bicycle and i t fell.'
While paraphrases of these sentences with a third person pro-
noun is awkward, i t is possible to paraphrase them without
~yJ.
Cf.
(3.23").
(3.23")
af f t 'utu' 'k
a
'
u,
'
a k
u' d '
?e
'-f
an
Afi push Aku, Aku fall
ground
'Afi pushed Aku and Aku fell. '
The negation of these sentences works the same way as the
negation of the sentences in the previous sections.
2.1.5
comparatives and Superlatives
(3.24)
at[ 1010
wu
rnf
Afi be big
surpass
us
'Afi is bigger than us. I
(3.25)
desiEi
I~
vrv(
wu
nJ
stew-the fut. be delicious surpass thing
'The stew will be too delicious.'
(3.26)
kof r QU
wu
rnol6; ~
QU
eta n~
Kofi eat surpass rice; he eat yam also
'Kofi ate more than rice; he also ate yam.'
(3.27)
, f
'
i
kof
wu
rnl
ga
Kofi surpass us strength
'Kofi is stronger than us.'
Most analyses dealing with comparatives and superlatives
of this form have considered sentences (3.24)
and (3.25)
above
as serial verb constructions because wU
I
surpass'
can also
occur as a main verb in sentences which do not involve seria-

---

-96-
lization.
(Cf.
(3.27)
above).
In that case,
(3.24)
can be
considered
as being derived from a sentence such as
(3.24').
(3.24')
af ( 1010
6y~ aft wJ
mf
le 10)0
me
Afi be big and Afi surpass us in bigness-in
'Afi is big and she surpasses us in bigness.'
However the first part of the paraphrase implies that Afi is
big~ this is not necessarily so in the case of comparatives.
It. may simply
be the case that compared to us Afi is big,
not that she is big in the absolute sens~ of the word.
Further-
more, in sentences
(3.25)
and (3.26), the surface subject can-
not be understood as the underlying subject of w6 because the
sentences resulting from such a use of that subject either do
not make sense
(cf. (3.26'»
or do not have the same meaning as
the surfacre sentence
(cf.
(3.25'».
(3.26')
*kof r du
(n~)
~y~ kof f wG
Kofi eat (thing)
and Kofi surpass
'Kofi ate and Kofi surpass rice.'
(3.25')
desTZi
I ~
v(vf
~y6 desTa
I~
stew-the fut.
.
be delicious and stew-the fut.
~
wu
nu
surpass thing
'The stew will be delicious and will be too much.'
(3.25") and (3.25'")
which are closer in meaning can also be
offered as underlying forms for
(3.25):
(3.25")
desTa
I~
v(v(
~y~ desTa
b~
stew-the
fut. be delicious and stew-the pass.
,
v(vr
la
wu
nu
deliciousness fut.
surpass thing.
'The stew will be delicious and its deliciousness
will be too much.'

---

-97-
(3.25 ''')
des i a
I ~
v r v r
~ y~ d ~ 5 i a
I ~
stew-the fut.
be delicious and stew-the fut.
wu
n6
le vYv(
me
surpass thing in deliciousness in
'The stew will be delicious and i t will surpass
(any)
thing in deliciousness.
The same way I
(3.24")
and
0.26'1)
can be underlying forms
for
(3.24)
and (3.26)
respectively.
(3.24")
af( 1010
eye af( be
1010
wu
mr~
Afi be big and Afi poss bigness surpass us
be
10 I 0
poss bigness
I Afi
is big and she surpasses us in bigness. I
(3.26")
kof f d u
n6
eye enJ
ke
kof r du5 w6
Kofi eat thing and thing that Kofi eat surpass
molu; e d 'u 'et'e --'-
na
rice; he eat yarn also
'Kofi ate and what he ate is. more than rice:he
also ate yarn. I
Though sentences
(3.25") I
(3.25''') I
(3.24'),
(3.26-') I (3.24")
and (3.26")
are acceptable sentences,
i t is clear that using
them as underlying sentences for
(3.25),
(3.24)
and (3.26)
repectively
will require rather ad hoc and arbitrary dele-
tion rules involving items which are not in any way syntactic-
ally or lexically identidal to any other item in the sentence.
This is the type of deletion rules
Jackendoff
(1972)
and
and Postal
(1964) have objected to (see p.83 above).
For
this reason I consider 'NU used in comparative and superlative
sentences as a modifier to the main verb.
Another syntactic
fact that supports my claim is the use of the habitual aspect
marker na (see section 2.2.2.3 of chapter II).
As a general
rule, this aspect marker is placed immediately after the main

---

-98-
verb in a sentence.
However in constructions with compara-
tives and superlatives
na cannot be placed after w.
CL (3.28)
,
and (3.29).
wu cannot be the
nuleus of the predicate phrase
,
,
wu m~·
(3.28)
a.
ad du
,
,
#
na
nu
wu
rn(
Afi eat hab.
thing surpass us
,
'Afi eats more than us.
*b.
aft du
,
,
na
nu
w6
n~
rn(
Afi eat hab. thing surpass hab. us
,
,
(3.29)
~
aha
I.
I.
..
,
a.
nu
na
se: se: wu
nu
he drink hab. drink hard surpass thing
'He drinks too much. I
, ,
,
,
,
,
,
*b.
~
nu
na
aha
SESE wu
na
nu
he drink hab. drink hard surpass hab. thing
2.1. 6
Split Verbs
,
,
(3.30 )
e ho
•
rnawu d3( se
he receive God
on . hear
'He believes in God. I
,
(3.31 )
rn ho
,
se
b~
~
t6
I
receive hear that he go out
,
'I believe he went out.
(3.32 )
- --
nusua
d3ra
kek~~
q6
man-the prepare bicycle-the arrive
'The man repaired the bicycle.'
,
(3.33)
rn( d3 ra
Q6
vo
we pr.epare arrive finish
'We have finished getting ready. '
(3.34 )
ad d~
desTa
kp:5
Afi taste stew-the see
'Afi tasted the stew. I
In sentences
(3.30)
and (3.31)
above, ho
'receiva,-get!
and
se' hear'
can be used as main verbs in other sentences.

---

-99--
Cf.
(3.35)
and (3.36) below:
(3.35)
marie
h ~'"
gaa
Marie receive money-the
'Marie received the money.'
(3.36)
m se
ehaa
I hear song-the
'I heard the song.'
d3ra.in sentences (3.32)-(3.33)
and d:5 in sentence (3.34)
..-
do not occur alone as main
verbs in a sentence; at least I
cannot recall any instances where they do.
However d6 'arrive'
r -
and kp5 'see' do.
Cf.
(3.37)
and (3.38).
(3.37)
mf d&
'
ah 6 '
me
we arrive home-in
'we arrive home. '
,
(3.38)
m kp~ hevf~
I
see bird-the
' I saw the bird. I
The constructions in (3.30)
through (3.34)
are differ~nt
from serial verb constructions in two main respects.
First,
they have no corresponding
paraphrases with ~y~; second, it
is not possible to negate the second verb alone.
This is
illustrated in
(3.30')
and (3.32') :
,
(3.30 ')
*a. ~
h~
A '
mawu d,f
~ye ~
se
he receive God
on
and he hear
,
,
. b. m.5
h~
A
mawu
d,f
s~
0
(he) neg. receive God
on
hear
part.
'He did not believe in God. '
. ,
,
,
,
*c. ~
h~
mawu
d,f
v;)a mu
se
0
he
receive God
on
but
(he)
neg. hear part •.
,
(3.32 ')
*a.
-
nusua
d,ra
kElk·6~
~y6 e
96
man-the
prepare bicycle and he arrive

---

-100-
(3.32')
b.
nDsiHi
m~
~,ra
kek6a
d6
•o
man-the neg. prepare bicycle-the arrive part.
'The man did not repair the bicycle.'
*c.
nusua
d,ra
kek~a
v~a m6
d6
•o
man-the prepare bicycle-the but
(he)neg. arr. part.
In the case of the morphemes being discussed here,
I
will agree with earlier analyses proposed for them
(see Schach-
ter (1974»
and enter them in the lexicon as verbs with two
parts, specifying which verb comes first and the position of
the object NP in relation to the two parts in case there is one.
Using a system Schachter used for the same verbs for Akan,
~h~~~.~.~.~s~e can be entered as follows:
(3.39)
h~ •.• se
[
+V ••• v ]
[
+NP AVX _NP
]
x
Y
[
+NP AUX
S ]
X
y
BELIEVE ( x,y )
Morphologically related to h~ 'receive' and
,
-
~ 'hear'.
2.1.7
Modifying Verbs
(3.40)
dev(a
..Q.3..Q.
l e d u n 11
kpb
:z:::=
child-the leave prog. eat thing
continually
I The
child does not stop eating.'
,
(3.41 )
m .Yl :t2
I
go
finish
'I have already gone. I
•
(3.42)
~ev (a
56
gee
mi
child-the take money swallow
'The child swallowed the coin. I
,
(3.43 )
e
ill
b6
ge
E
he
leave lose money part.
'He lost the moneyl'
(expressing disapproval)

---

-101-
,
(3.44)
e
du
nu
sugb2)
~
he eat thing be many
'He ate a lot.'
,
(3.45)
afr te
ak:'du~
kp5
Afi poke banana-the
'Afi poked the banana
(to see if i t was good)
,
(3.46)
~
Qebl~ s6
mi
he
hurry take swallow
'He swallowed i t quickly.'
The verbs singly underlined
in the above sentences can
all occur as main verbs in sentences.
What makes these con-
structions different from serial verb constructions is the
fact that none of them can be paraphrased as coordinated sen-
tences and no subpart of the sentences can be negated.
These
constructions also differ from those containing';' spli t
verbs I
because the verbs do not form a unit; the main verb (doubly
underlined) does not have to be the same all the time. For
example compare (3.40)
and (3.43): in these two sentences,
even though the
modifying verb is the same,
(d30),
the main
verbs are different
(du and bU).
One final reason why I do
not consider the modifying verbs
(singly underlined)
as main
verbs is the fact that they are not followed by the aspectual ~
when it is used in the sentence; rather i t may follow only the main
verb.
Cf.
(3.45a and b).
,
,
(3.45 )
a.
at( te
na
'k'd"
a o.ua
kp:S
Afi poke hab. banana-the see
"Afi usually pokes the banana. I
,
,
,
*b.
at( te
na
ak:, du~
kp:S na
Afi poke hab. banana-the see hab.
These modifying verbs generally convey an aspectual

---

-102-
meaning to the verbs they modify.
Sometimes their meaning
is somewhat similar to the meaning they convey when used as
main verbs
(cf. kp~, gebl~, ~) and sometimes there does not
seem to be any relationship between the two meanings
(cf s5,
d30).
sugb~ in (3.44) can also be used to modify a noun;
cf.
(3.47)
below:
, ,
(3.47)
awu
sugb~
le marie sf
dress be many be Marie hand
'Marie has many dresses.'
As is the case in Yoruba, (see Bamgbose
(1974»,
there
,
are two types of modifying verbs in Gegbe: pre-verbal modify-
ing verbs and post-verbal modifying verbs.
In the examples
above, d3b , s~ and qebla (or its equivalent s{du 'run race')
are pre-verbal while kp5 and v; are post-verbal.
2.1.8
Verbs as Prepositions
(3.48)
mf pia
g21/ f
la as)me
we buy
gali
be market-in
'We bought gali at the market.'
,
, ,
(3.49)
~
ml5 anf
d6
kpl5a
he lie ground arrive table
'He lay down on the table. I
,
,
(3.50)
marie kpl~
anfgba
n~
m
Marie sweep ground
gJ.ve
me
'Marie swept the yard for me. I
,
(3.51)
at(
1135)
!@.
dav r~
Afi give bath give children-the
(pI.)
'Afi gave a bath to the children.'
.
(3.52)
56
~gbe
d,(a
mf liS
n3
dU
t;mela
come from today on-the we fut.
prog. eat fish
'Starting today, we will be eating fish.'

---

-103-
There has been much controversy regarding the nature of
the items underlined in the sentences above.
Some linguists
consider them as verbs used in a serial verb construction; see
Stahlke (1970); Boadi
(1968); Awobuluyi
(1967)
etc . . . ).
Others
such as Gilbert Ansre consider them as verbids or prepositions
derived from verbs
(see Ansre (1966)6,.
Even though all these
items can occur as main verbs in other contexts,
I believe
like Ansre that they are not verbs, but prepositions with a
meaning somewhat similar to that of the verbs they are derived
from.
My position in this respect is based on three main ar-
guments.which are given
in the following paragraphs.
First of all, let us suppose
the items under considera-
tion here are verbs used in serial verb constructions; in such
a case, i t phould be possible to conjoin the predicate phrase
containing them to the initial sentence in paraphrases using
, ,
eye.
However this is not the case because, as the unaccepta-
bility of sentences
(3.49')
and (3.50')
below show,
such para-
phrases do not make any sense.
,
*{ .
(3.49')
,J:·'-r
m ::> an
~y~ ~
Q6
kpl5~
~ri f
he lie ground and he arrive table-the
,
(3.50')
* ma r j e kplo
'-f b.t.
an 9 a ~y~ ~
n~
m
Marie sweep ground and she give me
In the case of sentence (3.48), one may argue that ~ can be
used as the deep subject of le and give a perfectly valid
paraphrase such as
(3.48'):

---

-104-
. , .
(3.48')
mr pie g~1I r 6y(; g~1I r le
aSlme
we buy gali and gali be
market-in
'We bought gali and there is gali at the market.'
This may be true but if le were a true verb in that position,
i t should be possible to modify i t with different tenses,
aspect or mood.
But as i t turns out,
le in constructions
like the above never changes its tense, . mood or aspect.
(See
next argument.)
The next argument is a syntactic one dealing with the use
of the habitual aspect marker ns or any of the auxiliaries dis-
cussed in Chapter II.
These auxiliaries cannot be used with any of
the items under consideration here.
Cf.
(3.48a)
through
(3.51a):
(3.48 )
*a.
mf pie
•
ga I r no'na
• , •
na
aSlme
we buy hab. gali be hab. market-in
,
, ,
(3.49f
*a.
~
I~
nil5 an r
I~
d6
kpl5a
d;r
he flit.
He ground fut_ arrive table-the on
,
,
(3.50)
*a.
marie kplb
na
a;;rgb~ na
na
m
Marie sweep hab. ground give hab. me
. ,,
(3,51)
*a.
aff t(;a les1
tea na
devl::>
Afi can give bath can give
children
My third argument is a semantic one dealing with the meaning
these items convey in the different sentences.
For example in
(3.49), do (preposition)
expresses more an idea of position than
..--
that of an action generally conveyed by the verb do.
In
(3.50),
..--
na (preposition)
expresses the idea of 'on behalf of' rather than
that of benefit conveyed by the verb.
In
(3.50), however, the
meaning of the preposition na is closer to that of the verb since
in a way the children benefit from the bath or they are 'given'

---

-105-
the bath: nevertheless this is not a problem because, in
spite
of the syntactic differences between the prepositions and the
verbs, i t is normal that there should be at times some kind of
semantic relationship between them since one
(the preposition)
is derived from the other (the verb) • 8
Finall~ example (3.52)
is a particular case because, even
though sd 'come from'
may occur as the main verb of a sentence
in other,contexts (see
(3.53)
below), i t is obvious that here
i t is not used as a verb in a predicate phrase: i t is a preposition
in a pre-post-positional phrase, a structure commonly found in
"
Gegb~: so (preposition), egbe (NP), d~f (postposition).
(3.53)
marie 56
an~h~
Marie come from Aneho
'Marie came/comes from Aneho.'
,
3.
Derivation of Serial Verb Constructions in Gegbe
.,
As mentioned earlier, ,.at the beginning of section 2 of this
chapter, I consider only the sentences in sections 2.1.1 through
2.1.4 as serial verb constructions in
"
"
Gegbe because only in those
sentences are
all the verbs involved true verbs.
One test that
has been used for determining which verbs in the constructions
in the other sections are true verbs and which are not is their
behavior with regard to auxiliaries in case they are present in
a sentence.
The same test can be used for the verbs in sections
2.1.1 through 2.1.4.
For example in all the sentences in 2.1.1
all the verbs can be preceded by the same auxiliary.
Cf.
(3.54)

---

-106-
below as a paraphrase of (3.5):
,
(3.54)
aft I ~
y)
I~
pie
ga If
I~
s~
E
I~
v~
Afi fut. go fut. bUy gali fut.
take it fut.
come
, 6 '
ah
me
home-in
'Afi will go bUy gali and bring it home. I
In this sentence, the future tense marker la can be used to
precede each of the verbs and the sentence is still acceptable.
fAs
is illustrated. in the examples of 2.1.2, i t is possible
to use different auxiliaries with each verb. in those sentences.
In sections 2.1.3 and 2.1.4 the same auxiliary can be used
with each of the verbs in the different sentences except with
gQ in example (3.14) for reasons discussed earlier (see section
2.1.8).
In the case of the
'split verbs' in section 2.1.6,
the auxiliary can only be used with the first of the two mor-
phemes involved: however, as the discussion of the examples
in this section has shown we are dealing here with a kind of
compound verb; therefore
only one auxiliary is necessary.
In sections 2.1.5,
2.1.7 and 2.1.8, the use of the auxiliary
~ has made it clear that only the main verbs can be modified
by an auxiliary.
ef (3.28b),
(~.45b) and (3.50a) repeated here
for convenience.
,
,
(3.28)
*b. af r du
na
n6
wu
n~
rnf
Afi eat hab.
thing surpass hab. us
,
(3.45)
*b. a f i
te
na
ak~du~
kp~ ncS
Afi poke hab. banana-the see hab.
,
,
,
(3.50 )
'*a. marie kpl~
na
aii(gba
n~
n~
m
Marie sweep hab. ground. give hab. me

---

-107-
These sentences are grammatical when only the first verb is
followed by na.
This then shows that in these cases we are
not dealing with true verbs in series but with one true verb
followed by a modifying verbal element.
Earlier in this chapter, reference was made to different
proposals regarding the derivation of sentences like those in
sections 2.1.1 through 2.1.4 in other languages.
Two such
proposals, Awobuluyi
(1967)
and Bamgbose (1972)9 favor a solu-
tion with underlying mUltiple sentences and deletion rules to
eliminate the coordinating elements and the subject and object
NP's in the non-initial sentences when the latter are identic-
al "to those in the initial sentence.
The main reasons for
favoring such a solution are based on the fact that 1)
someti-
mes serial constructions can be paraphrased with multiple sen-
tences with or without coordinators: 2)
very often, the seman-
tics of these constructions is such that, in order to account
for ambiguities, the syntactic component must be made to show
the different meanings and this can be easily done by positing
multiple underlying sentences where the subject and object
intended for the verbs are overtly expressed:
3)
finally,
there are cases where the auxiliaries in the different non-
initial predicate phrases differ from the one(s) "in the initial
predicate phrase and having different sentences underlyingly
is really the only way to handle this problem
in a satisfac-
tory
way.
For more details on these two proposals see the

---

-108-
individual documents.
Two other proposals have been made to account for the same
type of sentences,
favoring an underlying string with one sUb-
ject and one auxiliary node foll~wed by a series of concatena-
ted verb phrases.
Schachter (1974) who favors this approach
bases his proposal on the fact that in general, only one sUb-
ject NP and often one auxiliary node
are present in serial
constructions at the surface level.
Furthermore., some of the
new. tendencies in linguistics have been to restrain the power
of the transformational component of Generative Grammar and
one of the ways of doing so is to avoid deletion rules when
possible (see Katz and Postal
(1964); Jackendoff
(1972».
Therefore, in Schachter's opinion, the underlying structure
of serial verb constructions should be very similar to the
surface structure so as not to require any deletion rules of
any sort.
He offers the following base rule for serial con-
structions based on data from Akan,
a Kwa language:
(3.55)
S ~
NP
AUX
VP
VP*
In this rule AUX dominates tense and aspect morphemes
and VP* represents zero or more occurrences of VP.
Though he
does not offer any semantic rules,Schachter suggests that
cases of ambigUity can be taken care of by rules of semantic
interpretation.
Lord (1974)
also favors the use of con catenated VP's to ac-
count for serial verb constructions in Yoruba,
for the same

---

-109-
reasons as those given by Schachter. However in addition to
a base rule A, she offers a rule of semantic interpretation B
to take care of ambiguities which may arise in such construct-
ions:
0
(A)
S.... NP
VP
Vp 1
~
MP
VP
VP
(B)
Rule of Semantic Interprettion for Serial Verb
Con structions:
(i)
The VP's all refer to sub-parts or aspects of
a single event.
(ii) The second VP is always in some sense a further
development,result or goal of the first. l l
Rule B is more specific to the African languages type of serial-
ization because, as the author points out,
i t may not apply
to members of other language families.
For example, in Manda-
rin Chinese, the predicate phrases involved do not necessari-
ly describe a single overall
event~ they may describe events
that are just consecutive.
For more details see Lord
(1974),
·Causative
Constructions in Yoruba~'.
Let us now turn to the derivation of the sentences in
sections 2.1.1 -2.1.4 above.
In view of the facts presented in
the discussion of the data, I do not think that Schachter I s
solution with concatenated VP's can adequately account for the
,
different cases we are dealing with in Gegbe because of the
great semantic variation involved and also because of the pos-
sibility of having more than one kind of tense auxiliary in
I

---

-llO-
one sentence.
Of course if one chooses to consider as serial
verb constructions only those constructions which have one
subject which is the surface as well as the deep subject,
and
one auxiliary for all the predicate phrases, Schachter's solu-
tion can be considered a desirable one because of its simpli-
city and because i t does not require any complicated deletion
rules.
However,! do consider sentences with different auxi-
liaries as serial constructions because, as Lord
points
out,in her Rule B,the predicate phrases with the different
auxiliaries all form parts of a whole event.
Another point to be taken into consideration is the nega-
tion of the serial verb constructions in ~gbe.
As the negated
versions of the s.entences in the different sections show, the
auxiliaries tend to surface in non-initial predicate phras~s.
CL (3.6) and (3. 6c) repeated here for convenience:
,
,
,
,
,
(3.6)
afT I~
t~u va
yl
pie aQT
50
y)
na
t~sr
Afi fut. can come
go buy soap take go give aunt
'Afi will be able to go buy some soap to take to
aunt. '
,
,
(3.6)
c.aff
I~
t~u va
y)
vSa m~
pie as! i
Afi fut.
can come go but neg.+la
buy soap
,
5~
n~
t ~5 r
,
yl
0
take go give aunt
part.
'Afi will be able to go but she will not bUy
,
soap to take to aunt.
Though the future marker la is not overtly expressed in (3.6),
i t surfaces in
(3.6c)
after vS:!
(ma= m15+la).
For this reason,
there must be a provision for a node for these auxiliaries at
the underlying level.
This cannot be accomplished if the base

---

- l l l -
string has only one AUX node.
The solution proposed by Lord is very similar
to Schach-
ter's.
The only advantage i t has is the inclusion of Rule B,
rule of semantic inter~retation which allows the recognition
of SUbjects
of the non-initial predicate phrase different
from the subject of the initial predicate phrase.
Even though
this solution may be preferable to Schachter's because i t pro-
vides a solution to the problem of the semantics of the dif-
ferent constructions, i t still does not solve all the problems
,
of the Gegb~ data.
In her study, Lord does not deal at all
with sentences containing different auxiliaries, unless the
VP's in her base rule are of the form
Aux+V.
In my discus-
sion on auxiliaries in Chapter II, I consider the AUX as ex-
ternal to the VP because in cases of VP nominalization, the
AUX does not move along with the rest of the VP as a unit. In
such a case then, Lord's solution does not solve my problem
of deriving sentences with different auxiliaries, that is, if
she considers the AUX as part of the VP.
However,
if,
like
Schachter, she considers the AUX as separate from the VP, i t
should have been stated in the base rule and the problem
still remains unresolved.
The solution I propose for the derivation of serial verb
,
constructions in Gegbe is similar to that offered by Awobuluyi
and Bamgbose for Yoruba, that is, an underlying form for the
sentences consisting of coordinated sentences and deletion

---

-112-
rules of the Equi-NP deletion type to take care of the iden-
tical NP I S and auxiliaries in the different sentences of the
serial construction; a rule is also offered to delete the co-
ordinating conjunctions.
My main reason for adopting this
type of solution rather than the one offered by Schachter or
Lord is the fact that sometimes the coordinating conjunction
does surface in front of one of the non-initial predicate
phrases along with the appropr~ate sUbject, especially after
an object NP had been expressed.
Cf.
(3.56 ) below:
,
,
(3.56)
a. aft
•
yl
pie ga If
le , "
a51me
56
yl
na
ak~
Afi go bUy gali in market-in take go give Aku
,
,
or
b.
af (
"
yl
pie ga I ( le , , ,
a51me
~y~ ~.
5:>
yl
na akJ
Afi go bUy gali in market-in and she take go give A.....u
'Afi went to bUy gali at the market and took
i t to Aku. I
Therefore, positing a base string with coordinated sentences
.
.is not an improbable solution since such strings do occur in
the language as a paraphrase for serial verb constructions~
besides, such a solution allows me to solve more of the pro-
blems than I could with a solution using conca~ena~ed vp'&.
Following are the different deletion rules necessary to
arrive at the surface string of serial verb constructions in
~gbe from underlying coordinated sentences.

---

-113-
I
Conjunction Deletion Rule
(optional)
[S[ S X NP AUX[Vp V NP Y]Vp Z]S CONJ[S X NP AUX[vp V NP Y]vp Z]S W]S
S:D. 1 2
3
4 5
6
7
8
9 10 11
12 13 14
15
16 9
s.c. 1 2 3
4 4
6
7
ft1
9 19 11
1'213 14
15
16
Where 10=2 or 10=5 or
Iterative when more
10=2 conj.5
than two sentences
involved.
I l SUbject NP Deletion Rule
(optional)
[S[S X NP AUX[vp V NP yl
.~ X NP AUX[vp V NP Y]vp
vp Z ]S
S.D. 1 2
3
4 5
6
7
8 9
10
11 12 13
S.C. 1 2
3
4 5
6
7
8 tp
10
11 12 13
14
15
Where :
9 =2
or 9 =5
Iterative when more
than two
sentences
or 9 = 2 conj. 5
involved.
The clause '9 = 2 conj. 5· refers to cases where the· subject and
object of the initial sentence can be understood as the sub-
j ect of the second sentence.
In this case both NP I S generally
refer
to humans.
III Object NP Deletion Rule
(optional)
[ S[S X NP AUX fvp V NP Y]vp z]S CONJ ~ X NP AUX fvp V NP Y]vp Z]S W]S
S.D. 1 2
3
4 5
6
7
8
9 10 11
1213 14
15
16
)
S.C. 1 2
3
4 5
6
7
8
9 10 11
12
ft1 14
15
16
Where:
13=5
Iterative when more
than two sentences
involved.

---

-114-
IV
AUX Deletion Rule
(optional)
[S[S X NF AUX fvp V NF Y]vp Z]S
'fs X
AUX fvp V NP Y]vp Z]S W]s
S~D. 1 2
3
4 5
6
7
B
9
10 11 12
13
14~
S.C. 1 2
3
4 5
6
7
B
J1
10 11 12
13
14
Where:
9 =3
Iterative when more than
two sentences involved.
Few conunents are in order regarding the
application of
these rules.
The Conjunction Deletion Rule can be applied
first
to the base string since,as' I have mentioned i t in
section 2.1.1,
serial verb constructions can be paraphrased
with concatenated sentences even though they do not sound as
natural as coordinated sentences.
In case the iqentical sub-
ject NF's are deleted before the conjunctions, the application
of that rule,
the SUbject NF Deletion Rule, will produce_un-
acceptable strings as illustrated in
(3. 57a)
below;
(3.57)
is
the underlying form for
(3.5).
(,3.57)
~ff yl ~y~ ~fr pie galf ~y~ atr s~
~
~y~'aff
Afi go and Afi bUy gali and Afi take i t and Afi
v~
ah6me
come home-in
The application of Rule II gives:
(3.57)
a.afr y) eye pie gal r ~y~ s~
e ~y~ v~
ah6me
Afi go and buy gali and take i t and come home-in
an unacceptable string.
It is therefore desirable to order
the Conjunction Deletion Rule before the Subject NP Deletion
Rule.
The structural description of RUle IV makes i t necessary
to apply i t only after -Rules I and II have applied. - Rule III

---

-115-
need not be ordered in any particular way.
NP's in non-initial positions can also be pronomina1izedJ
and the pronoun used depends on the nature of the NP being
pronominalized.
For example, i f both the subject and the
object in the initial sentence are intended as the subject
of the non initial sentence(s), then the third person plural
pronounw6 is used: otherwise-the singular form of the pronoun
' i s used.
Cf.
(3.17")
and (3.58).
(3.17")
atr yl ,ytJ' ak~ ~y~we5
vIJ
Afi go call Aku and bhey come
'Afi went to call Aku and they
(both)
came.'
,
(3.58)
at' yl ~y' ~
tr~
v:)
Afi go and,she return come finish
'Afi went and returned already. I
This pronominalization is optional •
Let us now turn to the derivation of sentences involving
modifying verbs and split verbs.
In these cases I believe
the expansion of the VP in the Phrase structure rules should
make i t possible for these verbs to be generated directly in
the base string.
For this purpose I propose the following
rule as a modification of PS Rule II in Chapter II:
Original rule:
VP'" V (NP)
(NP)
(Pp)
(ADV)
(V)
New rule
:
VP'" (V)
(V) V (NP)
(NP)
(pp)
(ADV)
(V)
This rule says that a VP is composed of a verb preceded
by two optional modifying verbs and followed by a direct ob-
ject NP, an indirect object NP, aprepositional/postpositiona1
phrase, an adverbial phrase and a final modifying verb.
The

---

-116-
cooccurence restriction of the different verbs will determine
which verbs can be used as pre-verb modifiers, as post-verb
modifiers or as main verbs.
Split verbs will be directly ge-
nerated under the node of the main verb in the verb phrase. 12
4.
Conclusion
In this chapter, I have presented an analysis of serial
verb constructions in Gegb~ based on the different semantic
interpretations of the sentences involved and the syntactic
behavior of the verbal elements used in those sentences.
I
have also given a very brief account of other studies done
on the same sUbject for other West African languages of the
Kwa group.
It is my personal opinion that what others, Bam-
gbose in particular, have considered moqifying types of serial
constructions are not true serial verb constructions in Gegb~,
since in this language,verDs used as modifying verbs do not
have the same syntactic characteristics as main verbs.
I
consider as serial constructions only sentences where the
different verbs involved have the same status, that is, they
behave syntactically the same way with respect to modification
by auxiliaries in particular.
In order to make clear the
difference between modifying verbs and main verbs,
I have pro-
posed an expansion of the VP in the phrase structure rules
allowing modifying verbs to be generated directly in the base
string with no use of coordination.
Serial verb constructions

---

-117-
in Gegb~ are derived from coordinated sentences in the base
to which deletion rules apply to eliminate the NP's in the
non-initial sentences identical to NP's in the initial sentence.
Another deletion rule is given to eliminate auxiliaries in
subsequent sentences identical to auxiliaries in preceding
sentences.
Finally ,a conjunction deletion rule
which must
be ordered before the subject NP deletion rule eliminates all
the coordinating conjunctions.
These rules are optional since
paraphrases of serial constructions with coordinated sentences
are found as surface strings in the language.
In light of
this analysis of serial constructions, there should be a rule
in the phrase structure rules of the language expressing this
idea of coordinated sentences in the base.
Therefore,
I pro-
pqse the following two rules as PS rules for the expansion of
the sentence in G~gbe:
A)
So ~ S (CONJ S)*
B)
S
~ NP AUX VP (pp)
(ADV)
The '*' in rule A makes i t possible to generate more than two
coordinated sentences in a serial construction.
Notice that
this rule will also allow the generation of all coordinated
sentences in the base; however, in order for the conjunction
to delete giving a serial construction at the surface level,
the subsequent NP subjects must be identical to the subject
or object in the initial sentence.
Rule B is the same as
PSR I I in Chapter II.

---

-118-
Finally, the material presented here has shown that serial
verb constructions in ~tl!gbe are not derived from base strings
of the form AUX + verb where some of the verbs are the main
verbs and the others constitute the auxiliaries.
Rather, in
the constructions with serial verbs, we are dealing with true
verbs occurring in series which are main verbs in underlying
coordinated sentences.
Furthermore, we cannot derive serial
verb constructions from concatenated VP's because doing so
does not make clear what the deep subjects of the non-initial
predicate phrases are, especially in cases where either the
subject or the object of the initial sentence or both can be
the sUbject(s)of the non-initial sentences.
Therefore in
order to arrive at the correct interpretation of the surface
string tner-e must be sentences at the underlying level expres-
sing what the logical subject and object of the different
surface predicate phrases are.

---

FOOTNOTES TO CHAPTER III
lIt should be pointed out here that, even though i t has
been the opinion of many linguists that semantics should be
kept separate fom syntax in grammatical analyses,
since they
belong to different levels, i t is very difficult to do so
in the study of serial verb constructions.
2see Westermann (1965), fifth edition of
Westermann
(1930),
p. 126.
3As expressed earlier in Chapter I, the third person
pronoun object generally assimilates to the preceding vowel.
Thus, e 'him,her, it' becomes 1 after lal, IEI and I~I and
1 after lul and li/.
4These sentences can only be used in cases where the
speaker and the hearer are aware of the object of the verb;
i t must have been mentioned in a previous discourse.
5d6 will not be considered here as a verb participating
in the serial construction.
More will be said about this·
morpheme later in section 2.1.8.
6Ansre , Gilbert (1966).
liThe verbid-- a caveat to I serial
verbs'
• Journal of West African Languages 3,1.
7As
i t may be recalled from the discussion in section
2.2.2.3 of Chapter II, only the stative verbnb can be used with
auxiliaries.
-119-

---

-120-
8Furthermore, there is a case of a true preposition not
derived from a verb.
Such a preposition is the morpheme kuor
kudo 'and' mentioned earlier in section 1.3 of Chapter II
(p.20).
Therefore there is empirical ground for setting up
a category 'Preposition'
for the language.
9Awobuluyi, A. 0.,1967.
Studies in the Syntax of the
Standard Yoruba Verb; and Bamgbose, A.,
1974.
"On Serial Verbs
and Verbal Status."
lOsee Lord
(1974) p. 195.
She only has the surface structu-
re representation in her article.
The rewrite rule is my own
interpretation of it.
Ilsee Lord l1974) p. 197.
12Movement rules will adjust the position of these verbs in
the sentence, depending on whether or not there is an object
present.

---

CHAPTER IV
,
-
¥
MODIFYING PREDICATIVE CONSTRUCTIONS IN GEGBE
1.
Introduction
In linguistic analysis,
the classification of modifying
predicative constructions has often been the subject of much
controversy.
One of th~ main problems has always been that
of determining the syntactic nature of the nucleus or focus
of these predicate phrases; in other words, which syntactic
categories can such morphemes be assigned to?
Ross
(1969b) and Lakoff (1970)
have dealt with this pro-
blem of classification to a certain extent.
Ross has pre-
sented arguments in favor of classifying as noun phrases,.
adjectives which are one of the possible nuclei of modify-
ing predicate phrases.
Lakoff argues in favor of classify-
ing them as verbs.
Schachter (1973)
discusses both Ross
(1969b) and Lakoff (1970)
and concludes that there is really
no empirical or formal grounds presented in both analyses
for considering adjectives as either noun phrases or verbs.
When analyzing modifying predicate phrases in different
languages of the world, especially non-Indo-European langua-
ges, the main problem the linguist faces is that of being
influenced by the traditional classification of parts of
speech based on Indo-European languages.
The result is that,
-121-

---

-122-
very often, morphemes have been labelled as adjectives, nouns
or verbs simply because they seem to correspond to the trans-
lated equivalent of the Indo-European counterparts, while a
more careful
consideration of their syntactic characteris-
tics would have shown that they should really be labelled
1
differently.
The main purpose of the present chapter is to examine
different types of modifying predicative constructions in
,
~gbe and try to determine, on the basis of their syntactic be-
havior, the syntactic categories which the different nuclei of
such constructions can be assigned to.
Before getting into
~
y
the details of the Gegbe data,
i t is appropriate at this point
to try to give a definition of what is meant by 'modifying
predicative constru~tions'.
As a general rule, the main purpose of a predicate phrase
is to communicate some kind of information about the subject
of the predication.
The predicate phrase can describe an
action performed (active sentences)
or endured
(passive sent en-
ces) by the sUbject of the sentence; it can also describe a
state or a position the subject is in or the nature of the sub-
ject.
The term 'modifying predicate phrase',.as used in this
study, refers to the latter type of description, that is,
i t
may describe the physical nature of the subject or a state the
subject is in or is getting into.

---

-123-
2.
Different Types of Modifying Predicative Constructions
in G~gb~
In modifying predicative constructions in general,
the
nucleus can be either a verb,
a noun or nominal,
an adjective
or adjectival.
In the following discussion, we will try to
determine whether,
syntactically speaking, these three cate-
,
gories can clearly be distinguished in G~gb~.
Following is a
list of the syntactic characteristics which can be used to dis-
tinguish a verb, a noun or an adjective in the language.
Verb:
2
1) The verb is the nucleus of the predicate phrase;
2)
the verb can be modified for tense,
aspect and modality;
3)
the verb can be modified by an expression of time,
place or manner.
1) The noun can be the syntactic subject of a sentence;
2)
i t can be the direct or indirect object of a verb1
3)
i t can be modified by a determiner, an adjective or an-
other noun.
Adjective:
1) An adjective can be in the same NP with the noun it
3
modifies, i t always follows the noun;
2)
i t can be the nucleus of a modifying predicate phrase;
3)
in predicative constructions i t is always preceded
4
by a stative verb or an inceptive verb.

---

-124-
2.1
Verbal constructions
- .-
(4.1)
eSI8
bla
water-the be red
'The water is red.'
(4.2)
, --
nusua
1010
man-the be fat,big
'The man is fat.'
(4.3)
hev ( de
k6
bird some die
lA bird is dead. '
(4.4)
" "
" 5
timati5
d 1
iiu'Ed6
tomato-the-pl. be ripe weli
'The tomatoes are nicely ripe.'
In the above examples, the underlined words attribute some
kind of quality to the items the predication is about or they
describe the state of these items.
As the following examples
snow, bla, 1010, kl! and di are true verbs because they have
~
all the characteristics of true verbs listed above.
.-
(4.1' )
eSla
I~
bla
~d
water-the fut.
be red very
'The water will be very red. '
,
(4.2')
n~s[j~
t~u 1010
naw6
man-the may be fat truly
'The man may truly be fat.'
(4.3')
hey! d~
I~
ku
5:'
bird some fut.
die tomorrow
'A bird will be dead tomorrow.'
, , "
'\\
\\ " ' ,
" , '
,
\\
(4.4')
timati5
eJl
na
iiu'E~e le
ad?afla
me
tomato-the-pl. be ripe hab. well in basket-the in
'The tomatoes get ripe very well in the basket.'
In sentences
(4.1')
through
(4.4'), each of the verbal
elements are modified by an expression of time, place or
,
manner and by an auxiliary
(la or t~u).
Though,
like predi-

---

-125-
cative adjectives, they are the nuclei of the individual
predicate phrases, they cannot be considered adjectives becau-
se, as the following ungramrnatical sentences show, they can-
not form an NP with the noun they describe and they cannot
be preceded by a stative verb or an inceptive verb.
, ,
(4.1)
*a.
eSI
bia
a
le t6n6
water-be red-the be there
. ...-
*b.
eSla
td
bH
water-the become be red
, ,
(4.2)
*a.
nusu
1010
a
vIJ
man-be fat-the come
*b.
' ~-
nusua
le
1010
man-the be be fat
(4.3)
*a.
hevf k6
d~
le h5'a
tIJ
bird-die-some be room-the top
*b. hev( g~
le ku
bird some be die
"
"-,,
(4.4)
*a. timiHi d i
nuede:>
I e ad,~lf (IJ
me
:tomato-be ripe.:....,ell-the-pl. be basket-the· in
, "
"
\\.
*b. timati5
le ~)
nUEd~
tomato-the-pl. be beripe well
In spite of the ungramrnaticality of the
(a)
sentences
,
y
above, they are cases in Gegbe where a verb phrase can be in
the same NP with the noun i t is describing.
Cf.
(4.5)
and
(4.6)
below:
,
,
(4.5)
des)
v~ d,e
d~
me
[NP [vp
vp]·
• NP]
stew be bitter salt some
be pot-the in
'There is some salty stew in the pot.'

---

-126-
(4.6)
awu
p~dT
~
le f~nJ
[ NP
[ VP •
VP ]
NP ]
clothes be dirty the-pl
be there
'The dirty clothes are over there.'
v~ d~e and pod! may be derived from relative clauses of the
fbrm ke v~ d3~a 'which is salty' and ke podI~ 'which is dirty'
respectively.
However no more will be said about these cases
here because relative clauses are not the concern of this
thesis.
Nevertheless i t should be mentioned that in cases
like this (attibutive vp's)
the verb phrase must always be
composed of a verb, a noun and maybe some other modifier; that
position is never occupied by a single verb.
The words underlined in
(4.1)
through
(4.4) cannot be
considered as noun either, because they cannot be the subject
or the direct or indirect object of a sentence.
They cannot be
modified by a determiner of any sort.
ef (4.lc)-(4.1e)
below:
,
,
,
(4.1 )
*c.
bl' aa
mu
ii5
?5
be red-the neg. be good part.
,
,
(4.1 )
*d.
~
10
na
b il'Hl
he like hab. be red_the
,
*e.
biaa
N
wa
le f~nJ
be red that be there
The same test applied to the other verbs in examples
(4.3)-
(4.4) will produce ungrarnrnatical sentences.
AS the above dis-
cussion shows, biA, 1010, kd and di are true verbs.
All other
lexical items having the same syntactic characteristics and
occupying the same position in modifying predicate phrases can
also be considered as verbs.

---

-127-
,
One process that is very productive in Gegb~ and in other
Kwa languages is the derivation from verbs of morphemes that can be
referred to as nominals or adjectivals.
This is done by redupli-
cation of the verb stern for verbs with one syllable; thus biA
and di will give biabi~'red/redness and did! respectively; ku
--
~
will give either kuku used as an attributive adjective
(cf.hev!
kUkd 'dead bird') or kukd used in predicative constructions
(cf. 4.7):
(4.7)
heY r~
kuku
le f6nu
bird-the
dead
at there
'The bird
lying dead over there. I
In case the verb has two or more syllables, there is no redupli-
cation but the tone· pattern of the verb changes to low-rising
for verbs with a low-low tone pattern and to rising-high for
verbs with a high-high tone pattern.
When the morphemes derived
from verbs with a high-high tone pattern are used as attributive
adjectives,
they retain the high-high tone pattern; forms derived
from verbs with a low-low tone pattern have the same pattern in
attributive as well~predicative positions.
Thus 1010 'be fat'
will give lol~ and tdtd 'wipe' will give t~td and tutu.
The re-
duplicated forms behave the same way as these latter forms with
regard to tone pattern when used attributively or in predicate
phrases.
In the following section we will examine more closely these
derived forms in their syntactic use and determine whether they
should really be considered as adjectivals or as nominals 6 or

---

-128-
both and what makes them different , if at all,
from the verbs
examined earlier in this section.
2.2
Other Types of Modifying Predicative Constructions
.,
(4.8)
awiia
le d~£
dress-the be red
'The dress is red. I
1.. .. =
(4.9)
n::>nua
le trlll
woman-the be thin
~The woman is thin. I
(4.10)
awiia
td
ylb5
dress-the become black
'The dress became black. '
(4.11)
m61 u~
le dada
rice-the be cooked
'The rice is cooked.'
(4.12)
kok I ca
le kuku Ie f~nu
chicken-the be dead at there
'The chicken is lying dead over there.'
.,
In the above examples, d3e, tr{ll, yib5, gad~ and kUku
can be considered adjectives because they are linked to the
substantives theymodify by the stative verb le
'be'
and the
inceptive verb
tr~ 'become'.
However, how is i t possible to
differentiate them syntactically as adjectives from nouns or
nominals?
In order to answer this question, let us consider
the following sentences:
. ,
(4.8)
*a.
- --
·awua
la
d 3~ wunu
dress-the fut.
red too much
,
_:w"==:r.
(4.9)
*a.
n::>nua
Ha tdlt naw6
woman-the can thin
truly

---

-129-
,
(4.10 )
*a.
awua
261~ y1bo
le toka
me
dress-the must black in bucket in
,
(4.11)
*a.
m61~a
d61~
dad a
le ze~
me
rice-the must
COO'ked in pot-the in
,
,
(4.12)
*a.
k~loa
k~ ku eso
chicken-the dead yesterday
In the above sentences, the underlined morphemes are modified
by auxiliaries and expressions of manner or place as could be
any verb.
However, as the ungrammaticality of the sentences
show,these items cannot be considered as verbs because they
,
do not agree with the definition of a true verb in G~gb~ ex-
pressed in the characteristics listed on P. 123.
Sentences
(4.8a) through
(4.12a)
would be grammatical if a stative verb
were inserted between the auxiliary and the underlined word.
Cf.
(4.11b).
,
(4.11)
b.
, "
mo I·ua
~61t1 no ga~a
le ze~
me
rice-the must be cooked
in pot-the in
'The rice must be cooked in the pot.'
Now take sentences
(4.13) through (4.19).
"
,
(4.13)
i l l li1 k~la
liITede
red be color
nice
'Red is a nice color. '
,
,
(4.l4)
ne m~
15
na
y1bo
o
I neg. like hab. black part.
'I do not like black.'
,
(4.15)
kUk~
m6
f~
o
dying
neg. be easy part.
'Dying is not easy.'
(4.16)
trllE~
d,od,1
n~ m
thin-the please
to me
'I like the thin one. I

---

-130-
v'
(4.17)
d3ea
y~
fif marie
t~
red-the emph. be Marie
poss.
'The red one is Marie's.'
V
"
(4 .18)
dada
sugb5
wu
mumua
cooked-the be many surpass raw-the
'The cooked one is more than the raw on. I
(4.19)
k~k6tj
d6
v5vS
dead-the instill fright
'The dead one is frightful. I
The words underlined in the above examples have the characteris-
tics of nouns because they are either subject'NP's or object
.,
NP's in the individual·sentences.
Furthermore
trtl{, doe, dad~,
., ,
mUmtt and kuku in
(4.16)-{4.19)
are modified by the definite ar-
v
ticle~.
However, there is a difference between d,e in (4.13)
.,
and the same word in
(4.17).
While in the former case dOE
refers to the color 'red~, in the latter it refers to something
red the speaker and the hearer have been talking about.
The
same thing is true of all the words underlined in
(4.16)-{4.19),
that is, they all refer to some substantive which is the topic
of the discourse even if i t is not mentioned overtly. This is
a semantic rather than a syntactic difference.
Syntactically
all those items are nominals with many of the attributes of
true nouns.
Therefore we can say at this point that morphe-
mes derived from verbs which are used as nuclei in modifying
predicative constructions can be classified both as adjectivals
and as nominals.
Further support can be found in the language for the cross-
\\
classification of these morphemes.
In Gegbe,
the verb fi! 'be'

---

-131-
is generally followed by a noun or a noun phrase.
Cf.
(4.20)
and (4.21).
(4.20)
Afr fir tIrtIa
Afi be teacher
'Afi is a teacher.'
,,.
(4.21)
11s~e t6k5e
y~
nfmra b~
suk~lu
Lycee Tokoin emph. be our poss. school
'Lycee of Tokoin is our school.'
In the above sentences, ~[tt[a and mfA be sUk~lu , a noun and
a noun phrase respectively, follow fit. The items under consi-
deration here can be used in the same position after fir. This
is exemplified in
(4.22) and
(4.23)
, ,
,
¥
,
(4.22)
-,.
t5ne
ye
nl
d?ea
mine
emph. be
red-the
'Mine is the red one. '
, ,
,
..
..
,
(4.23)
enu
ke
d6
.ll
na
vovo
ne
sugb3 a
thing that instill hab. fright to-him a lot part. -
y~
fir k~ ku
emph.
be dying
'What frightens him most is dying. '
Furthermore there are predicative constructions with the verb
d3~ 'be by nature' which have true nouns and noun phrases with
true nouns as their nuclei.
Cf.
(4.24)
•
"
I.l. t '
(4.24)
marle
d?o kUVIO 0
Marie be laziness-owner
'Marie is lazy' or 'Marie is a lazy person.'
In the same line of reasoning, many col or terms in Gegbe
are based on substances generally
having the same color:
thus 'grey' is af!", the term used for 'ash':
'green' is gbe-
mffm6 or 'green grass': brown is k~fe from kofl 'coffee'.
These words, though they are nouns or noun phrases are used

---

-132-
both attributively and predicatively:
(4 25)
1.
~
w
' y
,
I
•
a.
dWU
nell"
le gbemumu
dress my-the be green
'My dress is green.'
, ,
b.
~wu
gbemumu
d~
le hf
ne
dress green
some be hand my
I I have a green dress. 1
The fact that these words can be used both as nouns and as
adjectives and the fact that we have modifying predicative
constructions with nouns (or noun phrases) as their nuclei
provide additional support for the possibility of classify-
ing the morphemes derived from verbs both as adjectives and
as nouns because as i t will become clear in the following
discussion these morphemes can also be used both as nouns
and as adjectives, in the same positions as true nouns and true ad-
7
jectives. A characteristic mentioned earlier for adjectives is the
fact that they can be in the .same NP With the noun they modi-
fy.
This is particularly true of attributive adjectives.
. ,
Thus in (4.26) below, tr£l£ can be considered an attibutive
, ,
adjective since it is in the same NP with fi5nu.
, ,
(4.26)
i'i5iiu
[NP
woman
thin some
come
'A thin woman came. I
All other morphemes derived from verbs can be used attributively
to modify a noun or any substantive.
Cf.
(4.27)-(4.31)
,
,
(4.27)
aWl!
d?e ~
le
"
dr5
me
dress red the be closet in
'The red dress is in the closet. I

---

-133-
(~.28)
nJsu
lol~
d~
v6
es~
man
big
some come yesterday
I A big
man came
yesterday. I
. ,
(4,29)
ame
trtll d~
gbana
person thin
some
oe coming
'Someone thin is comimg.'
(4.30 )
, f 'k
Y
a.;)
pa
ylb~
w~·
mJ
d ?e
marie 6
shoe
black
that neg. suit Marie part.
'Those black shoe.s do not suit Marie. '
,
.
(4.31)
m61~
mumJ ~
mu
sa
o
rice
raw
the neg. be enough part.
'The raw rice is not enough. '
As shown before, trB.eandylb5in
(4.29)
and
(4.30)
respectively can also be used as predicative adjectives/nouns
in modifying;predicative constructions (see pp 128-129 above).
Lol~
in
(4.28) can also be used predicatively:
(4.32)
nusua
le IoU>
d~
man-the be big
some
'The man is somewhat big.'
Therefore in these cases. the same words are used as attribu-
tive adjectives and as predicative adjectivals/nominals. However,
..,.
in the case of d3e in
(4.27)
and mumu in
(4.31), there is a
difference of tone between the words used attributively and
predicatively.
In attributive constructions they have a high
tone pattern while in predicative constructions they bear a
rising or a rising-high tone pattern.
ef
(4.27),
(4.31) above
and ·(4.33) and (4.8)
repeated here for convenience.
(4.33)
m~IH
le mum~
le ze~
me
rice-the be raw
in
pot-the in
'The rice is raw in the pot.'

---

-134-
..
(4.8)
1iwu~
le d?e
dress-the
be red
'The dress is red.'
The same tonal alternation is true of all adjectivals/nominals
derived from verbs with a high tone pattern:
(4.34)
Verb
Attributive
Predicative
,
, ,
ye
awD'~
le
y
ye
awu
ye
'be white'
dress white
dress-the be white
'white dress'
'the dress is white. '
kiS
hevf k~ku
hey f~
le kUku
'be dead'
bird dead
bird-the
be dead •••
'dead bird'
'the bird is dead ••• '
,
,
f~
eSI
f~f~
- ...-
eSla
le fad
'be cold'
water cold
water-the be cold
'cold water'
'the water is cold. I
, ,
tSf
ezo
tSrtSf
ezoa
le tIltSf
'put out'
fire put, lout
fire-the be put out ,
·'dead fire'
'the fire is put out .
...
t Jt u
~gba t ~t ~
,
b::.1.
ag aa
le t~H
'wipe'
dish wiped
dish-the be wiped
'a dish that
'the dish is wiped. '
. was wiped. '
Following are some examples of words derived from verbs
with low tone patterns.
They have the same tonal alternations
as the words discussed earlier which are derived the same way.
(4.35)
Verb
Attributive
Predicative
da
m:5l~ dada
m:5 I is ~
led ada
'cook'
rice cooked
rice-the be cooked
'cooked rice'
'the rice is cooked.'
b' ,
la
es1
blabT~
esTi3
le b1abH
'be red'
water red
water-the be red
I
red water'
'the water is red.'
,
, ,
dl
timati d1dT
:l.
.I.
_=
tl/;lat,a
le ~19T
'ce ripe'
tomato ripe
tomato-the be ripe
'ripe tomato'
'the tomato is ripe.'

---

-135-
The same tonal alternations can be found in other words
used attributively and predicatively and which are not derived
,
from verbs. Thus vi
'small'
and ga 'big' behave the same way
as the morphemes derived from verbs.
,
(4.36)
a.
aw~
vf~
ii5
dress small-the be pretty
'The small dress is pretty.'
b.
awua
le vY
dress-the be small ,
'The dress is small.
,
(4.37)
a.
ahu~
ga
d~
le marie sf
.
house big some be Marie hand
'Marie has a big house. '
..
b.
ah6~a
le ga
house-the be big
'The house is big. '
,
,
Like vi and ~ , d,e is not derived from a verb either.
The same thing is true of yib~ which behaves like words derived
from verbs with low tones.
It may be suggested that the tonal variation found in the
morphemes under consideration here is due to the tone pattern
of the preceding words.
However, that is not the case.
In
(4.38)
the modifying morphemes have the tone pattern ap-
propriate for their syntactic use regardless of the tone of the
preceding word.
(4.38)
a.
heY f
r 'white' 'white bird'
'bird'
yl~5
'black'
'black bird'
d,e
'red'
'red bird'
b·
"awu'dress'ry 'white 'white dress'
Yl~~
'black'
'black dress'
d,e
'red'
'red dress'

---

-136-
(4.38)
c.
atr
G:,b:,
'long'
'long tree'
'tree'
t r~ I {
'thin'
'thin tree'
d.
aw~
~ ~b:'
'long'
'long dress'
'dress'
t r~ I ~
'thin'
'thin dress'
e. ~
' i t '
r
' i t is white'
'W
}
t~5
'become'
ye ' it became white'
wo
'make'
' i t looks white'
We can neither say that the tonal variation is a result of
these words being derived from verbs since as
(4.36)
and
(4.37) above show, there are words which are not derived and
yet they have the same behavior with regard to tone.
vI. and g[ like yibo and s13i'can also be used as substanti-
ves in sul:;)ject as well as object position in a sentence.
But
we have the same semantic interpretation expressed earlier for
,
d'3e: (see p. 130), that is, they generally modify a sub~tantive
that is not overtly expressed.
Cf.
(4.39)and
(4.40):
¥.,
(4.39)
a.
so
gaa
na m
take big-the give me
·Give me the big one.'
,
Of"
b.
gaa
n5
w6
~ya
big-the
be nice supass this one
'The big one is better than this one.'
(4.40)
a.
~
55
vU
nc!i aff
he take small-the giveAfi
'He gave Afi the small one.'
b.
vU
y~
fif
ilff to
small-the. emph. be
Afi pOSSe
'The small one is Afi' s. '
Finally, a group of words with a high-high or a low-low
tone pattern, not derived from verbs keep the same tone pattern

---

-137-
no matter whether they are used attributively or predicative-
ly. A semantic characteristic which these words have in common
is the fact that they all refer to
'dimension'.
ef
(4.41):
(4.41) :
Attributive
Predicative
tdl~
eka trll€'
ekaa
le tr{l{
'thin'
string thin
string-the be thin
'thin string'
'the string is thin.'
,
t ra Ia
emS trala
,1.1.
" I '
em~a
le tra a
'wide and long'
street wide and long
street-the be w.and 1.
'wide and long street'
'the street is w. and 1.
\\
, L
'
,
ahue gbad,a
ahJe~
le gbad,a
house broad
house-the be broad
'broad house'
'the house is broad.'
I~b~
aw~ I~b~
awua
le I~b~
"'long'
dress long
dress-the
be long
'long dress'
'the dress is long'
l~b5{
af~ l~bM
af:>a
le I:$b~~
I long
and thin'
foot long and thin
I foot-the
be "1. and th.
'long and thin foot'
'the foot is 1. and th.
gb{d,{
agb~ gb€d,~
"
~.l
agbaa
le gb{d:;{
'small and flat'
plate small and flat
plate-the
be s.and f.
'small and flat plate'
'the plate is s. and f.'
kl~boto
eze kloboto
ezea
le kloboto
'round'
pot round
pot-the be round
'round pot'
'the pot is round.'
k16b6t6~
al~gavr k16b6t6~
~I~gav(a
le k I6b6t6~
'small and round'
ring
small and round
ring-the
be s. and r.
'small and round ring'
'the ring is s. and r.'
As mentioned above for tr{l{, these words can also be used as
substantives and have the same characteristics as nouns.
The
semantic interpretation is the same as that of tr{l~ (see p.
129-131).
This is illustrated in
(4.42)
below:

---

-138-
(4.42)
l;b;i
d3~nad31
"
,
na m
wu
v~a
"
ma
long-the please+hab. to me surpass but
1+fut.
h~
gb':d3':~
take
flat-the
'I like the long one better but I will take the
flat one. I
These morphemes can never be used as verbs; they do not have
any of the characteristics of verbs listed above.
This has
already been shown for tr':I':(cf.(4.9a)
repeated here for con-
venience. )
,
(4.9)
*a.
==
J
n~nua
trEl£ .niwo
woman-the can
thin truly
(4.43) and '(4.44) below are ungrammatical because l~b~ -and
gb£d3£aremodified by an auxiliary and an expression of manner,
characteristics only verbs have.
_.!.:
,
- (4.43)
* awiIa
1';
l~b:' nU£g e
dress-the fut.
long
nicely
" ,
, , ,
(4.44)
* agbaa
d~l'; gb£d3 i
' -
wu
nenea
plate-the
must flat
surpass that-the
The introduction of a stative verb will make these sentences
grammatical:
,
, ,
(4.43')
awua
la
n~
l~b~
Iiu£de
dress-the fut. be long
nicely
'The dress will be nicely long.'
" ,
(4.44')
,
--
agbaa
~o'la'
,
b'd'"
" ,
w~
9 £ 3£
wu
nenea
plate-the must
make flat
supass that-the
I The
plate must be flatter than that.'
Since the words under consideration at-·this point can only
modify a substantive, -overtly expressed
or not,
in their
semantic interpretation, we can consider them as adjectives
which can sometimes be used syntactically as
nominals.
Words

---

-139-
,
,
like ~, vi, d3£ and ylbo can also be included in the same
category for the same reasons; however an exception must be
made for yibo since,unlik~ the other three, its adjectival
form is not tonally different from the nominal form (cf. (4.14)
and (4.30)
above~.
3.
Conclusion
In tqe discussion in this chapter, we have investigated
,
different types of modifying predicative constructions in G~-
gbe based on the syntactic nature of the nuclei of the differ-
ent predicate phrases.
So far the main findings are as
follows: there are three main types of modifying predicative
.
,
constructions in ~gb~: 1) predicate phrases with verbs as
their nuclei
(cf.
~a blA 'the dress is red' where bla is the
nucleus)i 2) predicate phrases with pouns as their nuclei
(cf.
,
,
Marie d3~ kuvi3to 'Marie is lazy', where kuv{3to is the nucleus);
3) predicate phrases with adjectivals/nominals as their nuclei
v
v
(cf. awUa le d3E
'the dress 1s red', where d3E is the nucleus).
The nuclei in the third group are classified both as adjectives
and as nounsi these include morphemes drived from verbs through
reduplication and/or tone change and other morphemes which are
not derived but behave the same way as the former with regard
to their tone pattern •. These morphemes can be used attributive-
ly or predicatively.
The different tone patterns are as follows:
in attributive position, the morphemes derived from high tone

---

-140-
,
,
verbs and other morphemes such as ye 'white', £ll 'red', ~
'big', y! 'small', have a high tone pattern
(cf. m;lu mumu
,
'raw rice'
or awU d3£ 'red dress').
In predicative position,
these words have a rising tone on one-syllable words and a
rising-high tone pattern on two-syllable words
(cf m::l1 tia le
m~mti 'the rice is raw' or awua le d3l 'the dress is red'.
As for morphemes derived from verbs with a low tone pattern,
and other words such as ylb~ 'black', they have a low-rising
tone pattern in both predicative and attributive position
(cf.
" ~ \\
' v , ,
: s i ' ' ' '
af~kpd le y~b~ 'the shoe is black' or ~a~f~~~k~p~d~y~~~b~~ 'black shoe';
~Inea le lol~ 'the person is fat' or ame lol~ 'a fat person').
Both types of morphemes can be used as nominals in contexts
other than that of the modifying predicative construction,
that is, they can be used as subject or object NP's in a sen-
tense.
In such cases they retain the same tone pattern as the
one they have in predicative constructions.
Finally, there
are morphemes which do:"ncit change their tonal pattern whether
us.ed predicatively or attributively; these words generally re-
fer to a dimension of some sort (cf.
(4.40)
above).
In the light· of the above remarks on tonal alternations,
especially the fact that some of the words have the same forms
when used attributively as well as predicatively, and also
when they are used as nouns in subject or object positions, is
i t really necessary to set up a class of adjectives or adjecti-
vals separate from a class of nouns or nominals?
My answer is

---

-141-
yes for two main reasons: first,
some of the items,those with
a high tone patternJdo have a distinction of form between a
word used attributively and one used predicatively
(a high
tone pattern in attributive position and a rising or rising-
high pattern in predicative position).
Second, in case we do
not want to make a distinction between adjectivals and nominals,
especially in items which have a low-rising tone pattern pri-
marily because they have the same form in all their syntactic
uses, we will then have in attibutive position a noun being
modified by another noun; however, as i t has been mentioned
earlier (see footnote 7 above),when a noun is modified by an-
other noun, the modifier generally precedes the modified.
This shows then that there is a differenee between the words
under consideration here and true nouns when they are both
used as modifiers; for this reason we should make a syntactic
difference between the attributive use and other uses of the
words with low-rising tones we have investigated here: in the
former use they should be considered as adjectives or adjec-
tivals and in the latter use as nouns or nominals.
,
I therefore believe that though Gegb~ has true verbs and
true nouns and few words that can be considered as true adjec-
tives, there are words that should be cross-classified as
nominals
(or nouns)
and as adjectivals
(or adjectives)
depenc-
ing on their syntactic use.
Thus in attributive position the
words used can only be considered as adjectives.
However, the

---

-142-
distinction is more difficult in predicative position since
in some cases the morphemes have only one representation used
in all
constructions.
Nevertheless, taking into consideration
the fact that true nouns do occur in that position and also
the fact that in the case of words which have a high tone con-
figuration in attributive position only the nominal form is
,
used in predicative position, we can say that in Gegbe, only
nominals and nouns are used in predicative position after sta-
tive verbs and inceptive verbs.
Finally,. a look at a semantic classification of "adjec-·
,
tives" in Gegbe based on Dixon's
(1970)
classification shows
that the semantic function of these terms does not, in any
way, determine their syntactic function as the remark on P.137
about words·like t r t l t , denoting dimension, may lead us to
believe.
,
(4.45)
Semantic Groups of Adjectives in Gegbe
a)
Dimension
ga
'big'
(adj.)
vt
'small'
(adj.)
huehue 'small'
(from hue
I be
small')
(adj al./nom • ).
11
11
dldi
'long'
(from dIdI
'be long')
" )
,
"
kpoppoe
'short'
(from kpokpoe'be short')
(adjal./nom. )
trala
'long, tall'
(adj.)
trtl t
'thin'
( "
)
gbad 3a
'wide, flat'
( "
)
kloboto
'round, in a circle I
( "
)
godo
'round'
( "
)
.

---

-143-
a) Dimension (continued)
nb~
'long'
(adj.)
1 :>b:>£
'thin and long'
(
"
)
,
,
gb£d3£
'small and flat'
l
"
)
b) Position
,
Y1 d3 1
'be high'
(V)
b~b~
'be low'
( " )
,
,
,
va g5me
'be low'
( " )
c)
Physical property
b~b~£
'50ft'
(from b~b~ 'be 50ft' )
(adjal./nom.~ )
A
,
A
,
5£0.
'hard'
(from s€U
'be hard' )
(
"
"
" )
,
..
,
kp€kp'£
'heavy'
(from kp'£
'be heavy' )
(
11
"
11
)
fofCS
'light'
(from fo
'be light' )
(
11
11
" )
, ,
W:>s1
'liquid'
(literally: make water)
(
"
"
" )
d)
Color
...
,
ye
'white'
( (N)
ye
(adj.)
....
,
d3'£
'red'
en)
d3£
(
11
)
'b"
y1
:>
'black'
(N:
adj.)
ebl~
'blue'
(N)
ijt:tsrdidr
'yellow'
(literally : small ripe. orange)
(N)
. .
f!
'
,
gb ·mumu
'green'
(literally
green grass )
( " )
k:>fe
'brown'
(from the word for coffee)
( " )
afi.'
'grey'
(from
' f ' 'ash ')
( " )
S-..!.
e)
Human propensity
kp;d31d3~£ 'be happy'
(literally: see happiness)
lV)
fianu
, intelligent,
smart'
( literally: know thing) ( ")
d31makpla
'impolite'
(literally: born but not raised
(adj~
properly)
, ,
fir3.ku
, (be) wicked)
(V)
bada
'(be)
bad)
(
11 )
',;
,
f'i3'£
'good'
(from n3 'be good')
(adj.)

---

-144-
f)
Age
, ,
y~ye
'new'
(adjal./nom. )
yeye
(adj. )
, ,
Mho
'old'
(
11
11
11
)
hoho
(
If
)
, .,
sI sI
'old'
(from sI
'be old' )
(adjaL/nom. )
,
ku amega 'become an old man'
(V)
, ,
ku naga 'become an old woman'
(
11 )
g)
Value
, ,
nQl
'good'
(adj.)
gbegble
'bad, spoiled'
(from gble t be bad, sJ?oil t)
sogbe
'be right'
(V)
h)
Speed
(V)
debl~
'go qUickly'
Verb+ bltwU
'Verb + slowly'
(V + adv.)
s{du
' run fast,go fast' (literally
'run race')
(V)
A comment is in order regarding the examples in h).
There is
no word that can be derived from these expressions and used
attributively to modify nouns or substantives.
Only these
verbs can be used in predicate phrases to convey the idea in-
tended.
Thus 'the car is fast'
can only be rendered by the
Gegbe expression meaning 'the car goes fast'
or 'the car runs
race'.
The words yeye and h~h6 in f)
can also be used syntac-
tically as nouns and as adjectives but they have the same se-
mantic interpretation as the words classified as adjectives
(see p.138).
Judging from the words and expressions in
(4.45), the
idea expressed
with adjectives in other languages are con-
,
veyed in ~gb~ with verbs, nouns, and nominals used predi-

---

-145-
catively and adjectives and adjectivals used attributively.
So the semantic category the different words belong to does
not determine what syntactic category they are part of since
most of the semantic categories contain verbs as well as nouns
and adjectives.
,
To summarize then, G~gbe modifying predicative construc-
tions are of three kinds: 1)
those which have a true verb as
their nuclei1 2)
those which have true nouns as their nuclei;
3) those which have as their nuclei nominals derived from verbs
or adjectives.
In this last category, some of the words have
only one lexical representation used both as adjectives and as
nouns, (cf yib5, tr£l£,
lol~, etc ••• ) 1 others have two forms,
one used as an adjective in attributive position and one used
,
v
as a noun or nominal in other positions
(cf. d3£ vs d3£, tutu
vs t~tu etc .•. )

---

FOOTNOTES TO CHAPTER IV
1See Welmers (1973) pp. 249-250 and Hagege (1974) p. 125.
2 This is a+so true of predicative nominals and predicative
adjectives. Therefore this characteristic alone is not enough
to classify an item as a verb; the other characteristics must
also be taken into consideration.
3
' v
In Gegbe, the determiner always comes at the end of the
NP and .attributive adjectives occupy a position between the
noun being modified and the determiner.
4These characteristics of adjectiveR are based primaily on
the fact that ~here are morphemes in the language such as trill,
~11b2 etc ••• which are used both in attributive position and in
predicative position to modify a substantive; they have the
same tonal pattern in both cases.
There are also morphemes
which exhibit different tonal patterns depending on whether
they are used attributively or predicatively.
Therefore there
is empirical motivation for having a category of adjectives in
the lan~lage with the characteristics listed here.
More argu-
ment will be presented later to justify my position.
5
"
,
oQ£ is derived from the verb 03 'be good/right'.
6The term 'adjectival' is used to differentiate adjectives
derived from verbs from those which are not;
'nominal' refers
to terms used syntactically as nouns but which are derived from
-146-

---

-147-
verbs and, as we will see later, from adjectives.
7
'
In G~gb~, there are also cases like
(i)-(iii)
below
where a noun is used to modify another noun; in such cases
the modifier always precedes the modified,
resulting in what
is generally referred to as associative constructions:
,
(i)
r5ba
tt.kpa
rubber bottle
lA plastic bottle.'
,
,
,
(H)
h5tru 1::>
door
hand
I door
knob. I
(Hi)
\\
,
, ,., /
s1ka tomegav1
gold earing
I golden
earing. I
In the above expressions, r5ba, h5tru and slka are the modifiers
,
and athkpa, alb ~d tdmegav! are the~odified.
These modifiers
are not derived from verbs and unlike the items classified as
adjectives earlier (see p. 138-139), they cannot be interpreted
semantically as modifying other sUbstantives when they are used
alone in a sentence.
Furthermore, while a noun
(modifier) + -
noun construction can be paraphrased with the possessive parti-,
cle be, a noun + nominal
(modifier)
cannot.
(Cf. rpba b'atukpa
as opposed to *awh be d3i.}

---

EPILOGUE
This thesis has investigated some of the syntactic cate-
,
gories that can be found in the predicate phrase in G~gb~.
The results aDe as follows:
In Chapter 11, i t has been shown
that a constituent category AUX can be set up which includes
terms expressing tense, aspect and mood.
The tense morphemes
are mainly the future tense marker la and the non-future ren-
dered by a zero morpheme~.
Aspectual morphemes include the
habitual marker ~, the progressive le/n~, the inchoative d3a,
,
the intentional gbe and the perfective expressed with the mo-
difying verb v~ at the end of the sentence; the progressive
aspect can also be rendered by an expression made up of the
stative verb le7n~ plus a nominal (derived from a verb) and
,
, ,
the verb ~.
Modal morphemes include ten I~an/mayl, d61a Imust',
and ke (counterfactual).
The main reason
for setting up this
category, based on studies carried out by Susan Steele and
others on the category AUX as a language universal, is the fact
that the different auxiliaries (except the habitual~) always
occupy a position between the subject NP and the main verb and
they always occur in the same order, that is, tense,mood and
aspect.
They form a constituent because their order can never
be interrupted by other constituents.
A constituent category VP is also set up for the language,
excluding the AUX and including a main verb, object NP's, some
prepositional phrases and adverbs of manner; place
and time
-148-

---

-149-
adverbials generally occur outside the VP.
The main reason for
setting up this category is the fact that in case of VP nominali-
zation all the contituents of the VP move as a unit to
the be-
ginning of the sentence.
The investigation of serial verb con-
structions in Chapter III has shown that some of the verbs in such
constructions can only be considered as modifying verbs because
they do not display all the characteristics of main verbs; there-
fore, the VP has been expanded to include pre-verb modifying
verbs as well as post-verb modifying verbs.
In this Chapter, i t
has also been shown that serial constructions should be derived
from underlying coordinated sentences; so the Phrase Structure
Rules generating sentences are formulated in a way making i t pos-
sible to generate these coordinated sentences in the base.
Sur-
face $erial constructions are then related to these coordinated
sentences by means of deletion rules which remove identical con-
stituents from the non-initial sentences and also remove the
coordinators.
Chapter IV investigates "modifying predicative contructions"
which describe qualitatively the object of the predication.
The
conclusion reached is that such predicate phrases can general-
ly have as their nuclei, a verb, a noun, or a nominal derived from
a verb or an adjective.
Taking into account the findings of the investigation in
this thesis, the following Phrase structure Rules can be pro-
,
posed
for generating simple, non-embedded sentences in G~gb~:

---

-150-
PSR I
So
.....
S ( CONJ S )*
PSR II
:
S
.....
NP AUX VP
(PP)
(ADV)
PSR III
VP
.....
(V)
( V)
V (NP)
(NP)
(pP)
(ADV)
(V)
PSR IV
:
NP
.....
~
(QUANT)
(DET)
]
(ADJ)
Pro
PSR V
PP
.....
1 NP Post
Prep
Prep
NP
NP
Post
PSR VI
AUX -+
Tense
(Mod)
(Asp)
PSR VII
DET -+
(Num)
(Poss)
({:~ J)
Turning now to the theory of grammar in general,
the in-
vestigations in this thesis bring support to the idea of a
universal constituent category AUX for all human languages~
the nature of the constituents which may occur in this cate-
gory and their order of occurence will mainly depend on the
word order type (s), found in the individual languages.
Even-
though i t has not been shown that there should also be a uni-
,
versal category VP; the facts of G~gbe show that such a cate-
gory is plausible in order to make certain generalizations for
some languages.
The investigation of modifying predicative constructions
has pointed out once .again the· fact that i t is very difficult
to set up clear cut grammatical categories at times; one proof
in this case is the fact that one of the items that are pos-

---

-151-
sible nuclei for these predicate phrases, the nominals derived
from verbs and adjectives, must be cross-classified both as
nominals and as adjectivals.
Finally,
though there are still many aspects to be in-
vestigated, i t is ~y hope that the results of the investiga-
tions in this thesis have in some ways contributed to the
elucidation of some of the syntactic aspects of the grammar
,
of ~gb!.

---

BIBLIOGRAPHY
Abbott, B.
1976.
Right Node Raising as a Test for Constituent-
hood.
Linguistic Inquiry 7.39-142.
Ab1orh-Odjidja. J. R.
1968.
Ca for Beginners.
Accra:
Water-
vi11e Publishing House.
Akmajian, A. and F. Heny.
1975.
An Introduction to the Prin-
ciples of Transformational Syntax.
Cambridge:
MIT Press.
_____ and T. Wasow.
1975.
The ,Constituent Structure of VP and
AUX and the Position of the Verb Be.
Linguistic Analysis
1. 3.205-245.
, S.Stee1e and T. Wasow.
1979.
The Category AUX in Uni-
-----versa1 Grammar.
Linguistic Inquiry 10.1-64.
Ansre, G.
1963.
Reduplication in Ewe.
Journal of African Lan-
guages 2.2.128-132.
· : 1966.
The Verbid.
A Caveat to 'Serial Verbs'. Journal
-----of West African Languages 3.1.29~32.
Awobu1uyi. A. O.
1967.
Studies in the Syntax of the Standard
Yoruba Verb.
Columbia University Ph.D. dissertation.
_____•
1971.
On the Status of Prepositions.
Journal'of West
African Languages 8.2.101-111.
•
1973.
The Modifying Serial Construction.
A Critique.
-----Studies in African Linguistics 4.1.87-112.
Bach, E.
1965.
On Some Current Types of Transformations.
Monograph Series on Languages and Linguistics 18.3-18.
·
1968.
Nouns and Noun Phrases.
In E. Bach and R. Harms
----(eds.), Universals in Linguistic Theory. New York: Ho1t.
Rinehart and Winston.
•
1974.
Syntactic Theory.
New York:
Ho1t, Rinehart and
---;Wins ton.
Bamgbose, A.
1974.
On Serial Verbs and Verbal Status.
Journal
of West African Languages
9.1.17-48.
_____ .
1973.
The Modifying Serial Construction.
A Reply.
Studies in African Linguistics 4.2.207-217.
-152-

---

-153-
Bartch, R. and T. Vennemann.
1972.
Semantic Structure.
A
Study in
the Relation Between Semantics and Syntax.
Frank-
furt am Main:
Athen~um Verlag.
Bedou-Jondoh, E. A.
1977.
Some Syntactic Aspects of Ewe, Mina
and Fon.
Unpublished manuscript.
,
v
1977.
Vowel Assimilation and Rule Application in G~gbe.
Unpublished manuscript.
,
,
'
v
1978.
Word Order in Gegb~, Evegbe and F~gbe.
Unpu-
blished manuscript.
,
,
,
1979.
The Status of ~ and ~1 in G~gb~.
Paper read
at the 10th Annual Conference on African Linguistics at
Champaign-Urbana. ,
Le and no in Gegbe.
Unpublished manuscript.
Bendor~Samuel, J. T.
1968.
Verb Clusters in Izi.
Journal of
West African Languages 5.2.119-128.
Berry, J. and N. A. Kotei.
1969.
An Introductory Course in
Ga.
U.S. Department of Health, Education and Welfare.
Bloomfield, L.
1933.
Language.
New York:
Holt, Rinehart
and Wins ton.
Boadi. L. A.
1968.
Some Aspects of Akan Deep Syntax.
Journal
of·West African Languages 5.2.83-90.
Bole-Richard, R.
1976.
Le Phoneme JpJ en Language Gen.
Afrique et Langage 5.29-32.
1978.
Probl~matique des Series Verbales avec Applica-
tion au Gen.
Afrique et Langage 10.24-47.
Bowers, J.
1971.
Adjectives and Adverbs in English.
Blooming-
ton:
Indiana University Linguistics Club.
Brame, M. K.
1975.
On the Abstractness of Syntactic Structure:
The VP Controversy.
Linguistic Analysi s1.l9l-203.
Bresnan, J.
1974.
The Position of Certain Clause Particles
in Phrase Structure.
Linguistic Inquiry 5.614-619.
Bull. W. E.
1971.
Tim~ Tense and the Verb.
Berkeley:
Uni-
versity of California Press.

---

-154-
Burt, M. K.
1971. From Deep to Surface Structure:
An intro-
duction to Transformational Syntax.
New York:
Harper
and Row.
Callow, K.
1966.
Preliminary Notes on Serial Constructions
in Kasem.
In Kenneth L. Pike (ed.), Tagmemic and Matrix
Linguistics Applied to Selected African Languages.
Ann
Arbor, Mich.:
University of Michigan Center for Research
on Language and Language Behavior.
Chomsky, N.
1957.
Syntactic Structures.
The Hague:
Mouton.
1964.
Current Issues in Linguistic Theory.
In J.
Fodor and J. J. Katz (eds.), The Structure of Language:
Readings in the Philosophy of Language.
Englewood Cliff,
N.J.:
Prentice Hall.Inc.
1965.
Aspects of the Theory of Syntax.
Cambridge:
MIT Press.
_ _~.
1970.
Remarks on Nominalization.
In R. Jacobs and
P. Rosenbaum (eds.), Readings in English Transformational
Grammar.
Waltham, Mass.:
Ginn.
•
1972.
Some Empirical Issues in the Theory of Trans-
----formational Grammar.
In S. Peters_(ed.), Goals of Lin-
guistics Theory.
Englewood Cliff, N.J.:
Prentice Hall,
Inc., pp. 63-129.
.
1973. Conditions on Transformations.
In S. Anderson
---and P. Kiparsky (eds.), Festschrift for Morris Halle.
New York:
Holt, Rinehart and Winston, pp. 232-286.
and H. Lasnik.
1977.
Filters and Control.
Linguistic
------Inquiry 8.425-504.
Comrie, B.
1976.
Aspect.
Cambridge:
Cambridge University
Press.
Culicover, P.
1973.
On the Coherence of Syntactic Descrip-
tions.
Journal of Linguistics 9.35-51.
1976.
Syntax.
New York:
Academic Press.
Dalby, D.
1977.
Language Map of Africa.
Londun:
Interna-
tional African Institute.
Dixon, R. M. W.
1970.
Where Have all the Adjectives Gone?
An Essay in Universal Semantics.
Duplicated ms. Camberra:
Australian National University.

---

-155-
Emmonds, J. E.
1976.
A Transformational Approach to English
Syntax:
Root,
Structure-Preserving and Local Transforma-
tions.
New York:
Academic Press.
Eulenberg, J. B.
1970.
Conjunction Reduction and Reduplica-
tion in African Languages.
In Chin-Wu Kim and Herbert
Stahlke (eds.), Papers in African Linguistics, pp. 71-80.
Fillmore, C. J.
1968.
The Case for Case.
In Bach and Harms
(eds.), Universals in Linguistic Theory.
New York: Holt,
Rinehart and Wins ton.
George, I.
1976.
Verb Serialization and Lexical Decomposi-
tion.
Studies in African Linguistics, .Supp1. 6.63-72.
GivOn, T.
1969.
Studies in Chi Bemba and Bantu Grammar.
Studies in African Linguistics,
Suppl. 3.
1975.
Serial Verbs and Syntactic Change:
Niger-Congo.
In Charles N. Li (ed.), Word Order and Word Order Change.
Austin:
University of Texas Press, pp. 49-112.
Greenberg, J. H.
1955.
Studies in African Linguistics Clas-
sification.
Branford, Conn:
Compass Publishing Company.
.
1961.
Some Universals of Grammar with Particular Re-
---ference to the Order of Meaninfu1 Elements.
In J. H.
Greenberg (ed.), Universals of Language.
Cambridge, Mass.:
MIT Press, pp. 58-82.
.
1966.
The Languages of Africa.
B1oomington:
Indiana
University Press.
Hagege, C1aude.
1974.
The "Adjective" in Some African Lan-
guages.
Studies in African Linguistics,
Suppl. 5.125-133.
.
1975.
Le Prob1eme Linguistique des Prepositions et
--~la Solution Chinoise.
Linguistique 71.
Paris:
Societe
de Linguistique de Paris.
Hale, R. and C. Lejewski.
Parts of Speech.
Aristotelian
Society,
Supp1. 39.173-204.
Halliday, M-A. K.
1961.
Categories of the Theory of Grammar.
Word 17.241-292.
Hankamer, J.
1971.
Constraints on Deletion in Syntax.
Un-
published Ph.D. dissertation, Yale University.

---

-156-
Hawkins, J. A.
1979.
Implicational Universals as Predictors
of Word Order Change.
Language 55.618-648.
Hinds, J. V.
1974.
On the Status of the VP Node in Japanese.
Bloomington:
Indiana University Linguistics Club.
Homburger, L.
1949.
The Negro-African Languages.
London:
Routledge and Kegan Paul Ltd.
.
'
Houis, M.
1968.
Pour une Typolog~e des Sequences de Propo-
sitions.
The Journal of West African Languages 5.1.13-24.
Householder, F. W.
(ed.) 1972.
Syntactic Theory 1: Structura-
list.
Baltimore, Md.:
Penguin Books Inc.
Huddles ton , R.
1978.
On the Constituent Structure of VP and
AUX.
Linguistic Analysis 4.31-60.
Huvanandana, S. V.
1980.
The Syntax and Semantics of Thai
Verbs.
Ph.D. dissertation, Indiana University.
Hyman, Larry.
1971.
Consecutivization in Fe?Fe1.
Journal
of African Languages 10.2.29-43.
1975.
On the Change from SOV to SVO:
Evidence from
Niger-Congg.
In Charles N. Li (ed.), Word Order and Word
Order Change.
Austin:
University of Texas Press, pp.
113-147.
Iwakura, K.
1977.
The Auxiliary System in English.
Lin-
guistic Analysis 3.101-136.
Jackendoff, R. S.
1972.
Semantic Interpretation in Generative
Grammar.
Cambridge:
MIT Press.
-
____~.
1975.
Introduction to X-bar Syntax.
Bloomington:
Indiana University Linguistics Club.
1977.
X-Bar Syntax:
A Study in Phrase Structure.
Cambridge:
MIT Press.
Jacobs, G.
1966.
The Structure of the Verbal Clause
in Bimoba.
Journal of West African. Languages 3.1.47-53.
Jakobson, R.
1971.
Selected Writings.
2nd Edition.
The
Hague:
MouLon.
Jespersen, O.
1933.
Essentials of English Grammar.
Uni-
versity of Alabama Press.

---

-157-
Jespersen, O.
1969.
Analytic Syntax.
New York:
Ho1t, Rine-
hart and Wins ton.
1974.
The Philosophy of Grammar.
London:
A11en and
Unwin.
Johnson, G. K.
1965.
Un Prob1eme de Termino1ogies:
Le Mot
Ewe lEVEl du Bas Togo dans la Classification des Langues
Negro-Africaines.
Etudes Togolaises T1, Vol. 1, pp.38-50.
"
1965.
Morpho1ogie des Nominaux dans la Langue Ge ou
----~Gegbe (Mina-Popo) du Bas Togo. Etudes Togolaises T1, Vol.
1, pp. 51- 61.
,
1967.
Mina or Gegbe Basic Course.
B1oomington:
Indiana University Linguistics Club.
Kamp, H.
1975.
Two Theories about Adjectives.
In Edward L.
Keenan (ed.), Formal Semantics of Natural Language.
Cambridge:
Cambridge University Press.
Katz, J. and P. Postal.
1964.
An Integrated Theory of Lin-
guistic Descriptions.
Cambridge, Mass.:
MIT Press.
Keyser, S. J.
1968.
Adverbial Position in English.
Review
of Jacobson.
Language 44.357-374.
Kock, J. de.
1975.
Pour une Nouve11e Definition de la Notion
d' Auxiliarite.
La Linguistique 11.81-92.
Kokora, P. D.
1976.
Studies in the Grammar of Koyo.
Ph.D.
dissertation, Indiana University.
Koze1ka, P.
1980.
Verbs as Preposition in Ewe.
An SOV Lan-
guage "Over the Hill"?
Paper presented at the 11th
Annual Conference on African Linguistics.
Boston.
Kpotufe, P. and N. Scheeberg.
1966.
Spoken Ewe.
B1ooming-
ton:
Indiana University.
Kropp~Dakubu, M. E.
1970.
The Categories of the Ga Verbal
Group.
Journal of African Languages 9, Part 2.70-76.
Lakoff, G.
1965.
Irregularity in Syntax.
New York:
Ho1t,
Rinehart and Wins ton.
_____ .
1966.
Stative Adjectives and Verbs in English.
The
Computation Laboratory of Harvard, Report No. NSF17.
_____•
1968.
Is Deep Structure Necessary?
Mimeo.
B1oomington:
Indiana University Linguistics Club.

---

-158-
Lakoff. G.
1970.
Adjectives and Verbs.
Appendix A of IT'-
regulaT'ity in Syntax.
New York:
Ho1t. Rinehart and
Wins ton.
_____ .
1971.
On Generative Semantics.
In D. Steinberg and L.
Jakobovits (eds.). Semantics.
London:
Cambridge Uni-
versity Press. pp. 232-296.
_____ and J. R. Ross.
1966.
A Criterion for Verb Phrase
Constituency.
In A. G. Oettinger et al.
(eds.). Mathe-
matical Linguistics and Automatic TT'anslation, pp.
(ii-1)-(ii-ll) .
and S. Peters.
1969.
Phrasal Conjunction and Symmetric
Predicates.
In Reibe1. D. A. and S. Schane (eds.).
Modern Studies in English.
Eng1ewood Cliff. N. J.:
Prentice Hall. Inc.
Laughren, M.
1976.
Serial Verbs.
Bulletin de
l'IFAN.
Tome
38. Serie B. No. 4.
Dakar. pp. 872-889.
de Lavergne de Tressan M.
1953.
Inventaire Linguistique de
l'Afrique Ocidenta1 Fran~aise et du Togo.
MJmoiT'es de
l'Institut Franqais d'Afrique Noire 30.
Levi. J. N.
1973.
Where do All Those Other Adjectives Come
.
From?
CLS 9.332-345 . .
1976.
The Syntax and Semantics of Non-PT'edicating
Adjectives in English.
Ph.D. dissertation. Northwestern
University.
(Previously distributed by the Indiana
University Linguistics Club.)
Li. C. N. and S. A. Thompson.
1973.
Serial Verb Construc-
tions in Mandarin Chinese.
Chicago Linguistic Society:
You Take the High Node and I'Zl Take the Low Node,
pp.
96-103.
Lord. C.
1973.
Serial Verbs in Transition.
Studies in
African Linguistics 4.4.269-296.
•
1974.
Causative Constructions in Yoruba.
Studies in
-----African Linguistics. Supp1.
5.195-204.
_____ .
1977.
How Igbo got from SOV Serialization to SVO
Compounding.
Studies in African Linguistics.
Supp1. 7.145-156.·
Lyons. J.
1966.
Towards a Notional Theory of the 'Parts of
Speech'.
Journal of Linguistics 2.209-235.

---

-159-
Lyons, J.
1968.
Introduction to TheoreticaZ Linguistics.
Cambridge:
Cambridge University Press.
1977.
Semantics II.
Cambridge:
Cambridge University
Press.
McCaw1ey, J.
1968.
The Role of Semantics in a Grammar.
In
Bach, E. and R. T. Harms (eds.), UniversaZs in Linguis-
tic Theory.
New York:
Ho1t, Rinehart and Winston, Inc.
1969 .. English as a VSO Language.
Language 45.286-299.
1971.
Tense and Time Reference in English.
In C.
Fillmore and D. T. Langendoen (eds.), Studies in Lin-
guistics Semantics, pp. 97-113.
1975.
The Category Status of English Moda1s.
Founda-
tions of Language 12.597-601.
Maxwe11, D.
1977.
Coordinate Deletion and the Position of AUX.
Proceedings of the 1976 Mid-America Linguistics Conference.
Minneapo1is:
University of Minnesota, pp. 223-236.
Mei11et, A. and M. Cohen.
1952.
Les Langues du Monde.
Paris:
C.N.R.S.
Mushira, E.
1977.
On the Nature of Coordinate Structure Cons-
traint.
Proceedings of the 1976 Mid-America Linguistics
Conference.
Minneapo1is:
University of Minnesota, pp.
81-94.
Newmeyer, F. J.
1971.
The Source of Derived Nomina1s in
English.
Language 47.786-796.
Pai11et, J-P and A. Dugas.
1977.
Principes d'AnaZyse Synta-
xique.
Montreal:
Les Presses de l'Universite de Quebec.
Per1mutter, D. and P. Postal.
1974.
Deep and Surface Struc-
ture Constraints in Syntax.
New York:
Ho1t, Rinehart
and Winston, Inc.
Postal, P.
1964.
Constituent Structure:
A Study of Con-
temporary ModeZs of Syntactic Description.
The Hague:
Mouton.
Pu11um, G. and D. Wi1son.
1977.
Autonomous Syntax and the
Analysis of Auxiliaries.
Language 53.741-788.

---

-160-
Rankin, R. L.
1977.
From Verb to Auxiliary to Noun Classi-
fier and Definite Article:
Grammatica1ization of the
Siouan Verbs 'sit', 'stand',
'lie'.
Proceedings of
the 1976 Mid-America Linguistics Conference.
Minnea-
po1is:
University of Minnesota, pp. 273-284.
Ross, J.
1967.
Constraint on Variables in Syntax.
Ph.D.
dissertation.
Cambridge, Mass.:
MIT Press .
. 1969a.
Auxiliaries as Main Verbs.
Journal of Philo-
---sophical Linguistics 1.77-102 ..
. 1969b.
Adjectives as Noun Phrases. In Reibe1 D. and S.
----Schane (eds.), Modern Studies in English.
Eng1ewood Cliffs,
N.J.:
Prentice Hall.
Sanders, G.
1975., On the Explanation of Constituent Order
Universa1s.
In C. N. Li (ed.), Word Order and Word
Change.
Austin:
University of Texas Press, pp. 197-268.
Sapir, E.
1921.
Language.
An Introduction to the Study of
Speech.
New York:
Harcourt Brace Jovanovich.
Schachter, p.
1974.
Constraints on Coordination.
Mimeo.
B1oomington:
Indiana University Linguistics Club.
•
1974.
A Non-Transformational Account of Serial Verbs.
----Studies in African Linguistics, Supp1. 5.253-270.
1974.
Serial Verbs as Verbs.
A Reply to a Reply.
Studies in African Linguistics, Supp1. 5.279-281.
Schroeder, J.
1936.
Formenlehre des Ge-Dialektes der Ewe-
sprache~ Dur1ach (Baden):
Christ1iche Ver1agsdruckerei
Gebr. Tron.
Schwartz, A.
1972.
Constraint on Movement Transformations.
Journal of Linguistics 8.35-85.
~
.
1972.
VP-Constituent of SVO Languages.
In J. Kimba11
---(ed.), Syntax and Semantics 1.213-235.
Seide1, A.
1902.
Grammatische E1emente des Anecho-Dia1ektes
der Evhe-Sprache in Togo.
Zeitschrift fur afrikanische,
ozeanische und ostasiatische Sprachen.
Berlin, pp. 174-
185, 189-193.
.
Shopen, T.
1971.
Caught in the Act.
Unpublished ms., Indiana
University.

---

-161-
Shopen, T.
1972.
A Generative Theory of Ellipsis:
A Consi-
deration of the Linguistic Use of Silence.
Ph.D. dis-
sertation, University of California, Los Angeles.
Ann
Arbor:
University Microfilms.
Sprigge, R. G. S.
1967.
Tone in the Adangbe Dialect of Ewe.
University of Ghana:
The Institute of African Studies.
Stah1ke, H.
1970.
Serial Verbs.
Studies in African Linguis-
tics 1.1. 60-99.
1974.
Serial Verbs as Adverbs:
A Reply to Paul
Schachter.
Studies in African Linguistics, Supp1.
5.271-277.
Steel, S.
1975.
On Some Factors that Affect and Effect Word
Order.
In C. Li (ed.), Word Order and Word Order Change.
Austin:
University of Texas Press, pp. 199-268.
1978.
The Category AUX as a Language Universal.
In J.
Greenberg (ed.), Universals of Human Languages, Vol. Ill.
Stanford, CA:
Stanford University Press.
Stewart, J. M.
1963.
Some Restrictions on Objects in Twi.
Journal of African Languages 2.2.146-149.
St~ckwe11, R. P.
1977.
Foundations of Syntactic Theory.
Eng1ewood.C1iffs, N.J.:
Prentice Hall, In~.
Tai, J.
1977.
Semantics and Syntax of Inner and Outer Loca-
tives.
Proceedings of the 1976 Mid-America Linguistics
Conference.
Minneapo1is:
University of Minnesota, pp.
393-402.
Takiza1a, A.
1974.
On the Similarity Between Nominal Adjec-
tives and Possessive Forms in Kijungan.
Studies in
African Linguistics,
Supp1. 291-305.
Tohouegnon,
1968.
Manuel Progressive de Conversation en
Langue Guen.
Ouidah, Repub1ique du Benin.
Vennemann, T. G.N.
1973.
Explanation in Syntax.
In J. P.
Kituball (ed.), Syntax and Semantics 2.
New York:
Seminar Press, pp. 1-50.
Voege1in, C. F. and F. M. Voege1in.
1964.
Anthropological
Linguistics 6.5. B100mington:
Indiana University Press.
Warburton, I., P. Kpotufe and R. Glover.
1968.
Ewe Basic
Course.
B100mington:
Indiana University Linguistics Club.

---

-162-
We1mers, W. E.
1973.
African Language Structures.
Los
Angeles:
University of California Press.
Westermann, D.
1907.
Grammatik der Ewe ~prache.
Dietrich
Reiner.
1930.
A Study of the Ewe Language.
Translated by
A: A. Bickford-Smith.
London:
Oxford University
Press.
•
1935.
Die Glidji-Ewe in Togo;
Zuge aus ihren Gesell-
-----schaftsleben.
Berlin:
W. de Gruyter & Co.
1943.
Der Wortbau des Ewe.
Akademie der Wissen-
schaften~Vienna.
Phi losophi sch-hi s torische Klasse 9.
Berlin:
WaIter de Gruyter & Co.
1961.
Die Ewe-sprache in Togo.
Ein praktische Einfuhrung.
Berlin:
WaIter de Gruyter & Co.
and M. A. Bryan.
1952.
The Langua.g~.z:/~.rN'~k~
...:"A.frica.
Oxford Univers i ty Pres s . : , \\ ; -
~" '\\\\
Wi11iam, W. R.
1971.
Serial Verb Const(~ ttiqd~C~~)~\\·o.
Studies in African Linguistics,
Sup 'lSt:47-~d,i~
Wi11~amson, K.
1963.
The Syntax of Veros afMat~~V in Ijo.
Journal of African Languages 2.2. 150-1;..;.,','..·~
1965.
A Grammar of the Kolokuma Dialect of Ijo.
Cambridge:
Cambridge University Press.
Witte, A.
1917-1918.
Sprichworter der Ewhe-Neger Ge-Dia1ekt
(Togo i Westafrika).
Anthropos 11.58-83.
Zwicky, A.
1970.
Auxiliary Reduction in English.
Linguistic
Inquiry 1.323-j36.
Additional Reference
Tai, J.
1970.
Coordinate Reduction.
Ph.D. dissertation,
Indiana University, B1oornington.

---