ArticlePDF Available

Locative phrases and alternative concord in Tshiluba

Authors:
Studies
in
African
Linguistics
Volume
9,
Number
1,
March 1978
LOCATIVE
PHRASES
AND
ALTERNATIVE
CONCORD
IN
TSHILUBA
1.
Introduction
Susan
U.
Stucky
University
of
Illinois
107
Locative
phrases
have
been
a
subject
of
controversy
since
the
begin-
ning
of
grammatical
analysis
of
Bantu
languages.
1
Traditional
grammars
have
attributed
to
them
functions
ranging
from
that
of
subject
and
direct
object
to
that
of
adverb.
Recently,
Trithart
[1975],
Dalgish
[1976a,
1976b],
and
Dalgish
and
Scheintuch
[1976]
have
discussed
locative
noun
phrases
in
various
Bantu
languages
with
respect
to
the
grammatical
rela-
tions
these
phrases
bear
to
verbs.
This
paper
continues
the
investigation
of
Bantu
locatives
in
Tshiluba
(a
Zone L
language
spoken
primarily
in
south
central
Zaire)
within
that
cluster
of
hypotheses
known
as
relational
grammar.
In
particular,
the
paper
focuses
on two
types
of
constructions,
each
of
which
contains
a
locative
morpheme
(either
pa
'on',
ku
'at',
or
mu
'in')
followed
by
both
a noun
and
a
modifier
which
may
be
either
an
adjective
or
a
demonstrative.
Example
(1)
below
illustrates
these
constructions.
2
Note
in
(1)
that
the
locative
pa
'at'
is
followed
by
a
noun mesa
'table',
which
in
turn
is
followed
by
the
adjective
-nene
'big',
with
the
concordial
morpheme
pa-
prefixed
to
it.
In
(lb),
on
the
other
hand,
the
adjectival
concord
marker
is
ma-
in
agreement
with
the
noun
itself.
The
concord
for
locative
phrases
containing
demonstra-
lThe
research
for
this
paper
has
been
supported
by
an
NDFL
Title
VI
fellowship
through
the
University
of
Illinois
African
Studies
Center.
Special
thanks
are
due
to
Dr.
Victor
C.
Uchendu
and
the
African
Studies
Center
for
funds
made
available
to
attend
the
VIllth
African
Linguistics
Conference.
I would
like
to
acknowledge
the
assistance
of
the
language
consultant,
Mutombo
Mpanya,
with
whom
I
have
been
working
for
three
years.
I
would
also
like
to
thank
Prof.
Eyamba Bokamba,
Prof.
Charles
Kisseberth,
and
Prof.
Jerry
Morgan
for
their
helpful
comments.
Any
errors
are
solely
my
responsibility.
2Due
to
the
limitation
of
space,
the
number
of
examples
is
extremely
limited.
The
handout
at
the
conference
contained
more
examples,
which
are
available
on
request
from
the
author.
Data
in
an
earlier
version
[Stucky
1976]
are
less
complete.
lob
thres
is
contrasted
in
(lc)
and
(ld):
3
(1)
a.
pa rn-esa p-3-nene
b.
pa
m-esa
ma-nene
on
table
LC-big
on
table
NC-big
'the
big
space
on
the
table'
'on
the
big
table'
c.
pa
m-esa
a-pa
d.
pa
m-esa
a-a
on
table
this-LC
on
table
this-NC
'on
this
table
(and
not
off
'on
this
table
(and
not
on
it)
,
that
one)'
Phrases
bearing
locative
concord
are
referred
to
as
LC
phrases
while
those
exhibiting
nominal
concord
are
called
NC
phrases.
This
phenomenon,
hence-
forth
alternative
concord,
not
unique
to
Tshiluba,
has
been
noted
in
Chi-
Bemba
by
Givan
[1972]
and
in
OluTsootso
by
Dalgish
(personal
communication).
In
addition
to
these
alternative
concordial
processes,
the
data
in
this
analysis
show
that
LC
and
NC
phrases
demonstrate
different
syntactic
behavior.
Furthermore,
a
syntactic
as
well
as
a
semantic
distinction
must
be
made
between
those
locative
phrases
with
adjectives
and
those
with
de-
monstratives.
Finally,
the
acceptability
of
LC
phrases
depends
crucially
on
the
type
of
verb
in
the
sentence.
It
will
be
helpful
to
keep
in
mind
certain
facts
of
Tshiluba
grammar.
First,
noun
class
prefixes
govern
concord
of
two
types:
primary,
that
taken
by
adjectives
and
participles,
and
secondary,
taken
by
demonstratives
and
all
other
forms
requiring
agreement.
The
distinction
is
evident
in
the
concordial
markers
in
(lb)
and
(ld)
above.
Thus,
one
morphological
distinction
relevant
to
the
analysis
of
locative
phrases
exhibiting
alter-
native
concord
can
be
noted
at
the
outset.
Adjectives
take
primary
con-
cord
while
demonstratives
take
secondary
concord.
Second,
the
locative
morphemes
include
neither
the
notion
of
deixis,
handled
by
the
demonstra-
tives
nor
that
of
motion
to
and
from,
which
is
included
in
the
verb.
Fi-
nally,
unlike
other
noun
class
prefixes,
the
locative
morphemes
may
appear
3S
ym
bols
are
as
follows:
noun
class
numbers
are
not
marked,
but
a
hYPhen
separates
a noun
prefix
from
the
stem.
Ag=agreement,
OP=object
prefix,
LP=locative
pronoun,
Pass=passive,
Refl=reflexive
marker,
LC=
locative
concord,
NC=noun
concord.
Standard
orthography
has
been
employed
here
with
one
exception.
Long
vowels
that
are
not
derived
by
rule
are
written
as
sequences
of
two
short
vowels.
Tone
is
not
marked.
(See
foot-
note
4
for
tone
rules.)
before
full
nouns.
4
2.
Syntactic
Properties
of
LC
and
NC
Locative
Phrases
Besides
the
agreement
contrast
noted
in
the
previous
section,
LC
and
NC
phrases
behave
differently
with
respect
to
certain
syntactic
pro-
cesses.
A
semantic
distinction
also
emerges.
One
restriction
on
alter-
native
concord
is,
however,
noted
first.
Example
(2)
below
shows
that
alternative
concord
does
not
extend
to
the
predicate
but
is
entirely
re-
stricted
to
the
locative
phrase.
Verbal
agreement
is
always
with
the
locative.
(2)
a.
pa m-esa
pa-di
pa-nene
b.
*pa
m-esa
ma-di
pa-nene
on
table
Ag-be
Ag-big
on
table
Ag-be
Ag-big
'on
a/the
big
table'
c.
*pa
m-esa
pa-di
ma-nene
d.
*pa
m-esa
ma-di ma-nene
on
table
Ag-be
Ag-big
on
table
Ag-be
Ag-big
Syntactic
processes
illustrating
the
difference
between
LC
and
NC
phrases
include
the
non-relation
changing
rules
of
left
and
right
dislo-
cation.
The
data
in
(3)
and
(4)
below
show
that
the
LC
phrase
is
more
tightly
knit
than
the
NC
phrase.
Example
(3a)
shows
that
either
a
LC
or
a
NC
phrase
may
be
left-dislocated
by
leaving
a
pronoun
behind,
suffixed
to
the
verb.
Alternatively,
in
(3b),
the
NP
tshisalu
etshi
'this
market'
can
be
left-dislocated
from
within
a
NC
phrase,
but
not,
significantly,
from
within
a
LC
phrase.
Example (4) shows
that
right-dislocation
pro-
ceeds
analogously.
4There
is
little
phonological
or
tonological
evidence
in
Tshiluba
for
considering
the
locative
morpheme
together
with
its
noun
to
be
a
single
word.
No
noun
begins
with
a vowel
(since
there
are
no
vowel
pre-
prefixes)
and
thus
vowels
do
not
coalesce.
Locatives
stan~ng
alone
have
high
tone.
Affixed
to
a
stem
they
assimilate.in
t~ne.
~lS
does
no~
pro-
vide
any
evidence
either,
since
all
noun
pref1xes
1n
Tsh1luba
have.h1gh
tone
and
the
locatives
in
these
phrases
exhibit
high
tone.
There
1S
some
syntactic
evidence
for
considering
them
to
be
a
single
unit.
~o
other
word
such
as
a
demonstrative,
may
intervene
between
the
10cat1ve
and
the
noun:
Normally,
the
demonstrative
may
either
precede
or
follow
the
noun.
Still
traditional
orthography
has
treated
these
as
two
separate
words.
The
s~elling
in
this
paper
conforms
to
that
convention.
llO
( 3)
a.
b.
(4)
a.
b.
mu
tshi-salu
e-mu
{e-tsh
i }
mu-kaji
u-di
w-enda-
mu
in
market
{this-LC}
woman
Ag-be Ag-walk
to-
LP
this-NC
'in
this
market,
the
woman
is
walking
(to) (in)
here'
tshi-salu
market
*e-mu
{
e-tsh
i }
tthis-LC}
this-NC
mu-kaj i
u-di
w-enda-
mu
woman
Ag-be Ag-walk
to-LP
'this
market,
the
woman
is
walking
(to)
(in)
here'
mu-kaj i
u-di
w-enda-
mu,
mu
tshi-salu
e-mu
{e-tsh
i }
woman
Ag-be Ag-walk
to-LP
,
in
market
{this-LC}
this-NC
'the
woman
is
walking
(to)
(
in)
here,
in
this
market'
mu-kaji
u-di
w-enda-
mu
tshi-salu
*e-mu
{
e-tshi}
woman
Ag-be Ag-walk
to-LP
market
{*this-LC}
this-NC
'the
woman
is
walking
(to)
(
in)
here,
this
market'
Unlike
the
dislocation
rules,
relativization
is
sensitive
to
gram-
matical
relations.
This
process
also
reflects
the
structural
distinction
between
LC
and
NC
phrases.
In
(5)
and
(6) belOW,
both
LC
and
NC
phrases
are
relativized.
As
in
the
case
of
the
dislocation
rules,
the
LC
phrase
must
remain
a
syntactic
unit,
but
the
NC
phrase
may
be
relativized
out
of.
(5)
mu
tshi-Iongelu
e-mu
mu-di-bo
ba-
bala
mi-kanda
{e-tshi}
they-read
in
school
{this-LC}
Rel-be-they
books
this-NC
,
inSide/in
this
school
in
which
they
are
reading
books
...
,
(6)
mu
tshi··longe
I u *e-mu
{
"e-tsh
i }
tshi-di-bo
ba-
tok-
esha
in
school
{*this-LC}
Rel-be-they
they-clean-Caus.
this-NC
'in
this
school
which
they
are
cleaning'
Thus,
at
least
three
processes,
right
and
left
dislocation
and
rel-
ativization
indicate
that
the
LC
phrase
is
a
syntactic
unit
while
the
NC
phrase
m~
be
treated
either
as
a
single
unit
or
as
two,
the
locative
and
the
noun
plus
its
modifier.
3.
Alternative
Concord
and
Grammatical
Relations
3.1
Subject
of.
Whether
or
not
LC
and
NC
phrases
can
be
subjects
is
a
problem
that
requires
a
longer
investigation
than
is
possible
here.
Con-
structions
where
locatives
appearing
in
initial
position
govern
verbal
agreement
are
the
LOC-BE-ADJ
and
LOC-BE-NP
constructions
shown
in
examples
(T)
and
(8)
below.
Here
there
is
a
clear
difference
between
the
locative
adjective
and
the
locative
demonstrative
phrases.
That
both
LC
and
NC
phrases
are
acceptable
in
LOC-BE-ADJ
constructions
is
illustrated
in
(T).
Thus,
it
is
possible
to
predicate
a
quality
of
either
LC
or
NC
adjective
phrases.
In
(8)
however,
an
LC-adjective
phrase
is
unacceptable
in
LOC-
BE-NP
constructions.
In
contrast,
both
LC
and
NC
demonstrative
phrases
are
acceptable
in
both
BE
constructions,
as
is
evidenced
in
the
(a)
and
(b)
sentences
of
(T)
and
(8).
The
LC-adjective
(8c)
is
interpreted
as
a
locative
noun.
This
is
reflected
in
the
English
gloss
'the
clean
space
on',
rather
than
'the
space
on
the
clean
table'
.
(T)
a.
pa m-esa
a-pa
pa-di
pa-bole
b.
on
table
this-LC
Ag-be Ag-wet
'on
this
table
(and
not
off
it)
is
wet'
c.
pa m-esa
pa-tooke
pa-di
pa-bole
d.
on
table
LC-clean
Ag-be Ag-wet
'the
clean
space
on
the
table
is
wet'
(8)
a.
pa m-esa
a-pa
pa-di
mi-kanda
b.
on
table
this-LC
Ag-be
books
'on
this
table
(and
not
off
it)
are
books'
pa
m-esa
a-a
pa-di
pa-bole
on
table
this-NC
Ag-be Ag-wet
'on
this
table
(and
not
that
one)
is
wet'
pa
m-esa ma-tooke
pa-di
pa-bole
on
table
NC-clean
AG-be Ag-wet
'on
the
clean
table
is
wet'
pa
m-esa
a-a
pa-d
i
rni
kanda
on
table
this-NC
Ag-be
books
'on
this
table
(and
not
that
one)
are
books'
c.
*pa m-esa
pa-tooke
pa-di
mi-kanda d.
pa
m-esa ma-tooke
pa-di
mi-kanda
on
table
LC-clean
Ag-be
books
on
table
Ne-clean
Ag-be
books
*'the
clean
space
on
the
table
is
'on
the
clean
table
are
books'
books'
The
items
in
(7)
and
(8)
also
demonstrate
that
the
locative
morpheme
always
governs
verbal
agreement
on
the
verb.
According
to
one
early
ver-
sion
of
relational
grammar
[Postal
and
Perlmutter
19T4]
one
of
its
tenets
is
that
only
terms
(i.e.
subjects)
can
trigger
verbal
agreement.
The
fact
that
these
LC
and
NC
phrases
govern
agreement
may
constitute
an
argument
in
favor
of
termhood.
However,
there
are
at
least
two
other
explanations
for
this
pattern
of
agreement.
First,
it
may
be
that
agreement
in
Tshi-
luba
is
not
with
a
term
but
with
a noun
that
is
not
necessarily
a
subject.
Secondly,
it
could
be
the
case
that
the
simple
noun
is
the
term
but
the
agreement
is
with
the
larger
phrase
of
which
the
noun
is
a
part.
Thus,
verbal
agreement
is
not
sufficient
to
determine
termhood,
suggesting
that
more
evidence
is
needed.
Another
test
of
termhood
is
reflexivization.
If
locatives
are
sub-
jects,
they
should
trigger
reflexivization.
In
the
following
example,
both
LC
and
NC
phrases
trigger
reflexivization.
112
(9)
a.
mu
tshi-bunda
e-mu
mu-di-
shimbula
in
garden
this-LC
Ag-Refl-collapse
'inside
this
garden
(and
not
outside)
fell
in
itself'
b.
mu
tshi-bunda
e-tshi
mu-di-
sh
i
mbu
I a
in
garden
this-NC
Ag-Refl-collapse
'in
this
garden
(and
not
another
one)
fell
in
itself'
Thus,
if
it
is
true
that
only
terms
can
govern
agreement
and
trigger
reflexivization,
then
the
locatives
are
subjects
in
(6-9).
Whether
they
got
there
by
an
advancement
rule
(one
which
moves
an
NP
up
the
hierarchy
of
grammatical
relations)
or
a movement
rule
(one
which
does
not
change
grammatical
relations)
or
by
being
there
underlyingly
is
an
open
question.
3.2
Object
of.
The
relation
considered
next
is
that
of
object.
As
in
the
preceding
section
on
the
subject
relation,
the
behavior
of
locative
phrases
containing
adjectives
is
different
from
those
containing
demons-
tratives.
Consider
first
locative
phrases
with
adjectives
illustrated
by
just
examples
(a)
and
(b)
in
(10)-(13).
We
see
in
(10)
that
a
LC-adjective
phrase--but
not
a
NC-adjective
phrase--together
with
an
object,
is
unac-
ceptable.
In
the
same
type
of
construction,
though,
LC-
and
NC-ad,i
ecti
ve
phrases
in
(11)
are
both
acceptable.
(10)
a.
mu-kaj i
u-di
u-teka
mi-kanda
mu
tShi-longelu
*mu-nene
woman
Ag-be
Ag-put
books
in
school
LC-big
*'the
woman
is
putting
the
books
the
big
space
in
the
school'
b.
mu-kaj i
u-di
u-teka
mi-kanda
mu
tshi-Iongelu
tshi-nene
woman
Ag-be
Ag-put
books
in
school
NC-big
'the
woman
is
putting
the
books
in
the
big
school'
mu-kaj i
u-di
u-teka
mi-kanda
mu
tshi-Iongelu
mu
c-d.
e-{tsh
i}
woman
Ag-be
Ag-put
books
in
school
this-{LC/NC}
'the
woman
is
putting
books
in
this
school'
(
11)
a.
mu-kaj i
u-di
u-bala
mi-kanda
mu
tshi-Ionge
I u mu-nene
woman
Ag-be
Ag-read
books
in
school
LC-big
'the
woman
is
reading
books
in
the
big
space
in
the
school'
b.
mu-kaj i
u-di
u-ba
I a
mi
-kanda
mu
tshi-Iongelu
tshi-nene
woman
Ag-be
Ag-read
books
in
school
NC-big
'the
woman
is
reading
the
books
in
the
big
school'
mu
tShi-longelu
mu
c-d.
mu-kaji
u-di
u-ba
I a
mi
-kanda
e-{tshi}
woman
Ag-be
Ag-read
books
in
school
this-{LC/NC}
Again,
in
(12),
both
LC-
and
NC-adjective
phrases
may
follow
the
verb.
However,
in
(13),
only
a
NC-adjective
is
acceptable.
(
12)
a.
b.
c-d.
mu-kaj i
u-di
w-enda
woman
Ag-be Ag-walk
'the
woman
is
walking
mu-kaj i
u-di
w-enda
woman
Ag-be Ag-walk
'the
woman
is
walking
mu-kaji
u-di
w-enda
mu
tshi-Iongelu
mu-nene
to
in
school
LC-big
to
the
big
space
in
the
school'
mu
tshi-Iongelu
tshi-nene
to
in
school
NC-big
into
the
big
school'
mu
tshi-Iongelu
mu
e-{tshi}
woman
Ag-be Ag-walk
to
in
school
this-{LC/NC}
'the
woman
is
walking
into
this
school'
(13)
a.
ma-futa
ma-di ma-aya
mu
tshi-Iowu
*mu-nene
oil
Ag-be
Ag-get
rancid
in
calabash
LC-big
*'oil
is
getting
rancid
the
big
space
in
the
calabash'
b.
ma-futa
ma-di ma-aya
mu
tshi-Iowu
tshi-nene
oil
Ag-be
Ag-get
rancid
in
calabash
NC-big
'oil
is
getting
rancid
in
the
big
cala
bash'
/mu
c-d.
ma-futa
ma-di ma-aya
mu
tshi-Iowa
e-ttshi}
oil
Ag-be
Ag-get
rancid
in
calabash
this-{LC/NC}
'the
oil
is
getting
rancid
in
this
calabash'
Upon
closer
examination
one
finds
that
the
verbs
in
(10-13)
fall
into
different
classes
with
respect
to
the
types
of
NPs
required
or
permitted
to
accompany
them.
Verbs
such
as
-teka
'put',
in
(10),
rarely
surface
without
objects.
Verbs
such
as
-bala
'read',
in
(11),
often
surface
without
objects.
Verb
requiring
direct
objects
are
here
labeled
strongly
transitive
verbs.
Verbs
permitting,
but
not
requiring,
direct
objects
are
labeled
weakly
transitive.
A
verb
like
-enda
'walk
to'
in
(12)
belongs
to
another
class
of
verbs,
motion
intransitives,
which
are
distinguished
from
strictly
intransitive
verbs
such
as
-aya
'get
rancid'
in
(13).
Motion
intransitives
actually
require
locatives
while
strictly
intransitive
verbs,
although
allowing
them,
do
not
require
them.
It
appears
that
the
LC
phrase
does
not
satisfy
the
requirements
of
a
strongly
transitive
verb,
as
in
(10).
In
(11),
both
LC-
and
NC-adjective
phrases
are
accep-
table
with
weakly
transitive
verbs.
In
(12),
likewise,
both
LC-
and
NC-
adjective
phrases
can
satisfy
a
motion
intransitive.
But
only
NC-adjective
phrases
can
satisfy
a
strictly
intransitive
verb
as
in
(13).
Locative
phrases
containing
demonstratives
rather
than
adjectives
exhibit
none
of
the
idiosyncracies
of
the
adjective
phrases.
As
examples
(c-d)
in
(10-13)
show,
regardless
of
whether
the
demonstrative
locative
phrase
is
an
LC
or
an
NC
phrase,
it
'is
always
acceptable.
Thus,
it
can-
not
be
claimed
that
the
type
of
verb
determines
the
acceptability
of
LC
phrases
alone.
Rather
the
type
of
verb
and
the
type
of
locative
phrase
must
be
taken
into
account.
Returning
to
the
question
of
locative
adjective
phrases,
the
proposi-
tion
that
the
LC-
and
NC-adjective
phrases
differ
in
their
capacity
to
satisfy
the
requirements
of
the
verb
can
be
tested
as
follows.
First,
with
a
strongly
transitive
verb
having
no
object,
an
LC
phrase
should
be
acceptable,
but
an
NC
phrase
should
not.
That
this
is
indeed
the
case
is
illlustrated
by
the
examples
in
(14):
(14)
a.
mu-kaji
u-nanga
ku
n-zubU
ku-nene
woman
Ag-like
at
house
LC-big
'the
woman
likes
the
atmosphere
at
the
big
house'
b.
mu-kaji
u-nanga
ku
n-zubu
*mu-nene
woman
Ag-like
at
house
NC-big
·'the
woman
likes
at
the
big
house'
With
a
verb
such
as
-bala
'read',
a
weakly
transitive
verb,
we
might
ex-
pect
a
LC-adjective
phrase
to
have
two
readings:
one
a
patient
(where
the
LC-adjective
phrase
satisfies
the
requirements
of
transitivity)
and
the
second
a
locative
reading
(where
object
deletion
has
applied).
In
fact,
two
readings
ale
obtainable
for
such
a
verb,
as
is
shown
in
(15):
(15)
a.
mu-ntu
u-di
u-songa
mu
bu-atu
mu-nene
man
Ag-be
Ag-carve
in
boat
LC-big
'the
man
is
carving
out
the
inside
of
the
boat--making
it
bigger'
OR
'the
man
is
carving
(something
else)
in
the
big
boat'
b.
mu-ntu
u-di
u-songa
mu
bu-atu
mu-nene
man
Ag-be
Ag-carve
in
boat
NC-big
'the
man
is
carving
(something
else)
in
the
big
boat'
The
NC
phrase
in
(15b)
cannot
be
used
in
the
situation
where
buatu
'boat'
is
the
patient
or
the
receiver
of
the
action.
LC-adjective
phrases
conse-
quently
have
two
qualities
of
direct
objecthood:
they
follow
transitive
verbs
while
NC-adJective
phrases
do
not,
and
they
are
interpretable
as
patients.
Next,
consider
the
two
cases
where
a
verb
actually
requires
a
certain
kind
of
NP.
As
noted
above,
a
LC
phrase
in
(lOa),
following
a
strongly
transitive
verb
and
its
attendant
object,
was
unacceptable.
This
beha-
vior
may
be
attributed
to
the
fact
that
the
LC-adjective
phrase
behaves
like
an
object.
Since
the
verb
requires
only
one
object,
the
LC-adjective
phrase
is
then
superfluous.
Motion
intransitives,
like
strongly
transitive
verbs,
require
a NP, a
locative
in
this
case.
A
NP
not
marked
morphologi-
cally
as
a
locative
will
not
suffice,
as
example
(16)
illustrates:
(16)
*mu-kaji
u-di
w-enda
n-zubu
woman
Ag-be Ag-walk
to
house
A
LC-adjective
phrase,
as
noted
above,
could
satisfy
motion
intransitives.
Although
the
addition
of
a
second
LC-adjective
phrase
is
not
acceptable,
the
presence
of
a
second
NC-adjective
phrase
is.
As
in
the
case
of
strongly
transitive
verbs,
the
LC-adjective
phrase
is
unacceptable.
Compare
examples
(17a)
and
(17b):
(
17)
a.
}
b.
mu-kaj
i
u-di
w-enda
mu
tShi-longelu
mu-f
like
ku
woman
Ag-be Ag-walk
to
in
school
LC-dark
at
mu-soko
*ku-
{
lenga}
mu-
village
*LC-
{
NC_beautiful}
'the
woman
is
walking
to
the
dark
space
in
the
school
in
the
beautiful
village'
Thus,
the
acceptability
of
LC-adjective
phrases,
but
not
LC-demonstrative
phrases,
is
seen
to
be
directly
related
to
the
type
of
verb,
whether
strongly
or
weakly
transitive,
a
motion
intransitive
or
a
strictly
intran-
sitive
verb.
The
LC-adjective
phrase
behaves
more
like
a
direct
object
while
still
being
marked
as
a
locative
by
the
locative
prefix
and
its
objecthood
is
paralleled
by
its
nounlike
reading.
3.3
Pronominalization.
There
are
two
cases
(lla
and
llb)
where
a
LC-ad-
jective
phrase
exhibited
some
properties
normally
associated
with
direct
objecthood
that
are
not
yet
accounted
for.
LC-adjective
phrases
in
(15a
and
15b),
following
transitive
verbs,
were
interpretable
as
patients.
It
seems
prudent,
then,
to
investigate
some
syntactic
processes
that
might
reflect
their
status.
The
first
is
pronominalization.
Here
an
interesting
correlation
appears
between
the
position
of
the
locative
pronoun
and
the
appearance
of
LC
phrases
in
such
sentences
as
shown
in
(18a)
and
(19a)
below.
Normally
two
positions
are
available
for
object
pronouns,
preceding
the
verb
stem
and
following
it.
If
only
a
direct
object
is
present,
the
pronoun
appears
in
prefix
position.
Note
that
in
(18a)--corresponding
to
(15a)--and
in
(19)--corresponding
to
(14)--the
locative
pronoun
shows up
in
prefix
position
as
a
direct
object
would.
The
non-object
reading
of
the
LC
phrase--corresponding
to
(15a)--and
the
NC
phrase
pronoun
appear
in
suffix
position
as
illustrated
in
(18b).
The
examples
are
given
in
terms
of
dislocation
so
that
the
pronoun
can
be
checked
most
accurately.
(
18)
a.
mu
bu-atu
mu-nene mu-ntu
u-di
u-mu-songa
in
boat
LC-big
man
Ag-be
Ag-LP-carve
'the
big
space
in
the
boat,
the
man
is
carving
bu-atu
mu-
mu-ntu
u-di
b.
mu
{bu-nene}
u-songa-mu
in
boat
LC-
Ag-be
{NC-
big}
man
Ag-carve-LP
'in
the
big
boat,
the
man
is
carving
it'
a.}
k b k k
..
{
u-ku-nanga}
b.
u
n-zu
u
u-nene
mu-
aJI
*u-nanga-ku
at
house
LC-big
woman
{Ag-LP-like}
*Ag-like-LP
'at
this
big
house,
the
woman
likes
it
there'
it'
The
LC-adjective
and
NC-adjective
pronouns
with
-bala
'read',
a
weakly
transitive
verb,
evince
a
clear
preference
for
suffix
position
even
when
no
other
object
is
present,
as
the
following
examples
illustrate:
verbs.
These
do
not
passivize.
Compare (24-28)
below:
(24)
(25)
(26)
(27)
ku
n-zubu
e-ku
ku-nang-ibue
kudi mu-kaji
at
house
this-LC
Ag-like-Pass
by
woman
'the
atmosphere
at
this
house
is
liked
by
the
woman'
mu
bu-atu
mu-nene mu-song-ibue kudi mu-ntu
in
boat
LC-big
Ag-carve-Pass
by
man
'the
space
inside
of
the
boat
was
carved
bigger
by
the
man'
mu-
mu
tShi-longelu
{t
nene}
mU-bad
-ibue
kudi mu-kaji
5
1-
LC-
in
school
{NC_big}
Ag-read-Pass
by
woman
'in
the
big
school
was
read
by
the
woman'
mu
tshi-salu
mu-
{t
nene}
5
1-
mu-end
-ibue
kudi mu-kaji
in
school
LC-
{NC_
big
} Ag-walk
to-Pass
by
'in
the
market
was
walked
to
by
the
woman'
woman
(28)
*mu
tshi-Iowa
tshi-nene
mu-ay
-ibue
kudi
ma-futa
in
calabash
NC-big
Ag-got
rancid-Pass
by
oil
*'in
the
calabash
was
gotten
rancid
by
the
oil'
The
last
restriction
on
passivization
indicates
that
it
is
not
sufficient
to
speak
of
which
items
may
passivize.
Instead,
in
the
case
of
strictly
intransitive
verbs
it
is
necessary
to
consider
whether
or
not
the
locative
was
required
by
the
verb
in
the
first
place.
4. Summary
The
data
in
this
paper
have
shown
that
the
category
locative
is
not
a
discrete
one
in
Tshiluba.
The
morphological
distinction
evidenced
through
alternative
concord
is
reflected
in
three
syntactic
processes:
pronominalization,
passivization
and
relativization.
Semantically,
LC
phrases
focus
the
locative
while
NC
phrases
focus
the
noun.
Furthermore,
LC
and
NC
phrases
are
distinguished
along
another
parameter,
modification,
where
the
modifier
is
either
an
adjective
or
a
demonstrative.
With
respect
to
grammatical
relations,
it
was
argued
on
the
basis
of
verbal
agreement
and
reflexivization
that
locatives
could
occur
as
under-
lying
subjects,
terms.
It
was
also
noted
that
these
locatives
have
several
properties
of
objecthood;
they
may
pronominalize
and
passivize.
However,
the
ability
to
passivize,
it
was shown,
did
not
depend
on
the
locative
itself,
but
rather,
on
the
type
of
verb.
Furthermore,
LC
and
NC
phrases
appear
to
be
on a
continuum.
The
LC-adjective
phrase
is
more
like
a
direct
object,
since
only
it
may
satisfy
the
requirements
of
a
strongly
transitive
verb.
Finally,
locatives
in
Tshiluba
must
be
distinguished
from
other
ob-
jects,
since
both
may
appear
simultaneously
and
have
separate
positions
for
pronouns.
Motion
intransitives
require
locatives
in
the
same way
REFERENCES
Dalgish,
Gerard
M.
1976a.
"Locative
NPs,
locative
suffixes
and
grammat-
ical
relations."
Proceedings
of
the
Second
Meeting
of
the
Berkeley
Linguistic
Society.
pp.
139-148.
Dalgish,
Gerard
M.
1976b.
"Passivizing
locatives
in
OluTsootso."
Stu-
dies
in
the
Linguistic
Sciences.
6.1:57-68.
Dalgish,
Gerard
M.
and
Gloria
Scheintuch.
1976
"On
language-specific
sub-grammatical
relations."
Linguistic
Sciences.
6.2:89-107.
the
justification
for
Studies
in
the
Givan,
Talmy.
1972.
Studies
in
ChiBemba
and
Bantu
Grammar.
Studies
in
African
Linguistics.
Supp.
3.
Keenan,
Edward L.
1976.
"Toward a
universal
definition
of
'subject'."
In
Charles
N.
Li
(ed.),
Subject
and
Topic.
pp.
303-333.
New
York:
Academic
Press.
Keenan,
Edward L.
and
Bernard
Comrie.
1977.
"Noun
phrase
accessibility
and
universal
grammar."
Linguistic
Inquiry.
8.1:63-99.
Perlmutter,
D.
and
Paul
M.
Postal.
1974.
"Some
general
laws
of
grammar."
Handout,
Linguistics
Institute.
Stucky,
Susan
U.
sitivity."
"Locatives
as
objects
in
Tshiluba:
a
function
of
tran-
Studies
in
the
Linguistic
Sciences.
6.2:174-202.
Trithart,
Lee.
1975.
"Relational
grammar
and
ChiChewa
subjectivization
rules."
In
Proceedings
of
the
Eleventh
Regional
Meeting
of
the
Chicago
Linguistic
Society,
pp.
615-624.
Chicago:
Chicago
Linguis-
tic
Society.
... The strongest evidence that locatives in Kinyarwanda are DPs comes from the agreement properties of locative-internal modifiers. In Type 1 Bantu languages, adjectives and possessors show what is known as "alternative agreement" or "alternative concord" when they modify locative nouns (Bresnan and Mchombo 1995;Carstens 1997;Gregoire 1975;Kuperus and Mpunga wa Ilunga 1990;Marten 2012;Myers 1987;Stucky 1978). As the Chichewa examples in (27) and (28) demonstrate, modifiers can either agree with the noun class of the locative ("locative concord"; "outer agreement") or with the noun class of the base noun from which the locative is derived ("noun concord", "inner agreement"): ...
Article
Full-text available
In Bantu languages such as Chichewa or Herero, locatives can function as subjects and show noun class agreement (in class 16, 17 or 18) with predicates and modifiers. In contrast, (preverbal) locatives in Sotho-Tswana and Nguni have been analysed as prepositional adjuncts, which cannot agree. Our paper compares locatives in Kinyarwanda (JD61) with locatives in these other Bantu languages and demonstrates that the Kinyarwanda locative system is essentially of the Chichewa/Herero type. We show that Kinyarwanda locatives are nominal in nature, can act as subjects, and agree with predicates and modifiers. However, even though Kinyarwanda has four locative noun classes (16, 17, 18 and 25), there is only one locative agreement marker (class 16 ha-), which indiscriminately appears with all locatives, regardless of their noun class. We explain this fact by arguing that noun class features in Kinyarwanda do not participate in locative agreement; instead, the invariant class 16 marker expresses agreement with a generic feature [location] associated with all locatives. We offer a syntactic analysis of this peculiar aspect of Kinyarwanda locative agreement, and we propose a parameter that accounts for the relevant difference between Kinyarwanda and Chichewa/Herero-type Bantu languages.
Article
This paper addresses some semantic and syntactic aspects of Zulu locatives. Practically every noun (and pronoun) in Zulu can be locativised. The semantic effect of locativisation is to convert a thing-concept into a place-concept. In many Bantu languages, locatives are fully-fledged nominals; as such, they can function as subjects and direct objects, and control the full range of concordial agreements. Zulu locatives, however, fail to behave like regular nominals. At the same time, the locatives cannot be assimilated to any of the other syntactic categories that are standardly recognised, such as prepositional or adverbial phrases. In view of the fact that the locatives designate places, it is proposed that the locatives should be recognised as a distinct syntactic-semantic category of place-referring expressions.
Passivizing locatives in OluTsootso On language-specific sub-grammatical relations
  • Stu
  • Dalgish
  • M Gerard
  • Gloria
  • Scheintuch
1976b. Passivizing locatives in OluTsootso. Stu-Dalgish, Gerard M. and Gloria Scheintuch. 1976 On language-specific sub-grammatical relations. Linguistic Sciences. 6.2:89-107
Passivizing locatives in OluTsootso Stu-dies in the Linguistic Sciences
  • Dalgish
  • Gerard
Dalgish, Gerard M. 1976b. "Passivizing locatives in OluTsootso." Stu-dies in the Linguistic Sciences. 6.1:57-68.
Some general laws of grammar Handout, Linguistics Institute. Stucky, Susan U. sitivityLocatives as objects in Tshiluba: a function of tran-Studies in the Linguistic Sciences
  • D Perlmutter
  • M Paul
Perlmutter, D. and Paul M. Postal. 1974. "Some general laws of grammar." Handout, Linguistics Institute. Stucky, Susan U. sitivity." "Locatives as objects in Tshiluba: a function of tran-Studies in the Linguistic Sciences. 6.2:174-202.
On language-specific sub-grammatical relations Linguistic Sciences. 6.2:89-107. the justification for Studies in the Givan, Talmy
  • Gerard M Dalgish
  • Gloria Scheintuch
Dalgish, Gerard M. and Gloria Scheintuch. 1976 "On language-specific sub-grammatical relations." Linguistic Sciences. 6.2:89-107. the justification for Studies in the Givan, Talmy. 1972. Studies in ChiBemba and Bantu Grammar. Studies in African Linguistics. Supp.
Locatives as objects in Tshiluba: a function of tran-Studies in the Linguistic Sciences
  • Susan U Stucky
  • Sitivity
Stucky, Susan U. sitivity." "Locatives as objects in Tshiluba: a function of tran-Studies in the Linguistic Sciences. 6.2:174-202.