Complex sentences
Полное изложение текстаПлан
Верх страницыПолное изложение текста
13.1. Postverbal embedded clauses
1This chapter describes various types of complex sentences attested in Beng. Two types of subordinate clauses, complement and goal clauses, can be properly embedded inside another clause in a postverbal position inside the main clause, while temporal and conditional clauses occupy a position before the main clause.
13.1.1. Complement clauses
2Complement clauses are marked by comlementizers kē, kēsá̰, sá̰ and sâ̰, which appear to be mutually interchangeable. Often, the complement clause is coreferent with a 3sg pronoun in an argument position in the same clause:
(232) |
Ŋ́ |
à |
pè |
kē |
mḭ́ |
nṵ̂. |
1sg:Pst+ |
3sg |
say:L |
that |
2sg:Pst+ |
come:L |
|
‘I said that you had come’ (literally ‘I said it that you came’). |
(233) |
Ò-ó |
ŋ̄ |
nḭ̀ |
kē |
mḭ̀ |
nṵ̄. |
3sg-St+ |
1sg |
Benef |
that |
2sg:Hab+ |
come |
|
‘I want you to come’ (literally ‘It’s for me that you come’). |
3Such Beng sentences with complement clauses are reminiscent of the English sentences with embedded clause extraposition, and the 3sg pronoun seems analogous to the dummy it in English. The pronoun is not always present in sentences with complement clauses, but only when required by the argument structure of the main predicate. For instance, the verb ‘to say’ as in (232) is transitive, and requires the direct object position to be filled in, and the subject position is obligatory in the volitive construction in (233). In contrast, the intransitive verb wé ‘to reply’ does not select for a direct or indirect object with the semantic role of content of response, and it combines with a complement clause without any ‘dummy’ pronoun present:
(234) |
Ó |
wé |
kē |
à |
nṵ́ |
ɛ́. |
3sg:Pst+ |
reply |
that |
3sg:Pst- |
come:L |
Neg |
|
‘He replies that he wasn’t coming’. |
4Furthermore, complement clauses can combine with arbitrary verbs, adding the speech component to the sentence meaning regardless of whether the main verb has anything to do with speech:
(235) |
Ó |
bé |
kē |
ó |
mlɛ̰̄ |
yē-nā. |
3sg:Pst+ |
run:L |
that |
3sg:Pst+ |
snake |
see-Prf |
|
‘He was running saying that he had just seen a snake’ (literally: ‘He ran that he has seen a snake’). |
13.1.2. Goal clauses
5The main strategy of marking goal clauses uses complementizer nà̰ followed by a clause in optative mood:
(236) |
Zrṵ̄á̰ |
nà̰ |
mḭ̀ |
gbō |
pí. |
hunker |
for.to |
2sg:Hab+ |
feces |
excrete |
|
‘Hunker down in order to defecate’. |
(237) |
Ŋ-ó |
sɔ̰̀ŋ̀ |
cícá-ló |
nà̰ |
wà |
wó |
cá |
ŋ̄ |
nḭ̀. |
1sg-St+ |
person |
search-Prog |
for.to |
3sg:Hab+3 |
IN |
watch |
1sg |
BENEF |
|
‘I am looking for someone to read me this letter’. |
(238) |
Ŋ-ó |
ŋṵ̀ɛ̰̀ |
dǎ |
nà̰ |
sɔ̰̀ŋ̀ |
bì-lɛ̀ |
ò |
gā. |
1sg-St+ |
fetish |
fall |
for.to |
person |
this-Def |
3sg:Hab+ |
die |
|
‘I will pray the fetish for this person to die’. |
6Complementizer nà̰ can be omitted if the subject or the direct object of the matrix clause is coreferent to one of the participants of the goal situation:
(239) |
Ŋ́i |
kpɛ̰̀ -pɔ́ |
lù |
ŋ̄ |
lɛ́ŋ́ |
gɔ̄ŋ̄ |
nḭ̀ |
(nà̰) |
1sg:Pst+ |
play-Mns |
buy:L |
1sg |
child |
man |
BENEF |
for.to |
ŋ̀i |
drɛ̰̄ |
wō |
mḭ̄dàlɛ́ |
lō. |
1sg:Hab+ |
work |
do |
calmness |
with |
‘I bought my son toys in order to work undisturbed’. |
(240) |
Ŋ́ |
kɔ́pɛ́i |
lù |
ŋ̄ |
lɛ́ŋ́ |
gɔ̄ŋ̄ |
nḭ̀ |
1sg:Pst+ |
hoe |
buy:L |
1sg |
child |
man |
Benef |
(nà̰) |
ò |
drɛ̰̄ |
wō |
ài |
lō. |
for.to |
3sg:Hab+ |
work |
do |
3sg |
with |
‘I bought my son a hoe to work with’. |
7When such coreference does not hold, the complementizer is obligatory, compare (241) where the subject of the goal clause is not coreferent to the subject or the direct object of the matrix clause but to the indirect object:
(241) |
Ŋ́ |
kɔ́pɛ́ |
lù |
ŋ̄ |
lɛ́ŋ́ |
gɔ̄ŋ̄ |
nḭ̀ |
nà̰ / *ø |
ò |
drɛ̰̄ |
wō |
1sg:Pst+ |
hoe |
buy:L |
1sg |
child |
man |
Benef |
for.to |
3sg:Hab+ |
work |
do |
|
‘I bought my son a hoe in order for him to work’. |
8Sometimes, the goal semantics is expressed with the complementizer kē followed by a sentence in optative mood, but it has an additional semantic component in addition to the goal. kē is only compatible with a verbalized goal:
(242) |
Ŋ́ |
kɔ́pɛ́ |
lù |
ŋ̄ |
lɛ́ŋ́ |
gɔ̄ŋ̄ |
1sg:Pst+ |
hoe |
buy:L |
1sg |
child |
man |
nḭ̀ |
kē |
ò |
drɛ̰̄ |
wō |
à |
lō. |
BENEF |
that |
3sg:Hab+ |
work |
do |
3sg |
with |
‘I bought my son a hoe (saying) that he should work with it’. |
9Essentially, this goal usage of kē is a special case of its function, described in 13.1.1, of adding the speech component to the meaning of the sentence. The goal semantics in this case is just a pragmatic consequence of the optative.
13.2. Ways of encoding clausal arguments
10Strategies of encoding clausal arguments include:
-
complement clause with kē / sá̰ / kēsá̰, with various matrix predicates:
(243) |
Ò-ó |
pɔ̀ |
kē |
mḭ̀ |
nṵ̄. |
3sg-St+ |
necessary |
that |
2sg:Hab+ |
come[Bsq] |
|
‘You have to come’. (literally ‘It is necessary that you come’.) |
(244) |
Má̰ |
à |
là |
kē |
ó |
nṵ̂? |
1sg:Pst+ |
3sg |
ask:L |
that |
3sg:Pst+ |
come:L |
|
‘I asked if he had come’. |
-
nominalization is used in NP positions:
(245) |
Mɛ̰̄ |
[mḭ̄ |
vɔ̰̄ŋ̄ |
dɔ̄-lɛ̀] |
mà̰ |
begin |
2sg |
hole |
build-Nmlz |
CONT |
|
‘Start to dig your yam field!’ |
(246) |
Ó |
[sɛ̀wɛ́ |
pē-lɛ]̀ |
tùà |
3sg:Pst+ |
paper |
say-Nmlz |
leave:L |
|
‘He stopped reading’. |
-
subordinate clauses with the goal complementizer nà̰ or often asyndetical are used with predicates of causation:
(247) |
À |
túà |
[ŋò |
drɛ̰̄ |
wō]. |
3sg |
leave |
3pl:Hab+ |
work |
do |
|
‘Make them work!’ |
(248) |
Ó |
mḭ̄ |
gblè |
[nà̰ |
ŋ̀ |
tá |
mḭ̄ |
lō]. |
3sg:Pst+ |
2sg |
force:L |
for.to |
1sg:Hab+ |
go |
2sg |
with |
|
‘He forced you to go with me’. |
-
verb phrases with verbs in the base form are used with the verb zḭ̄ ‘can’:
(249) |
Mà̰ |
zḭ̀ |
yātrɔ́ |
flɔ̰́ɔ̰̄ |
ɛ̰́. |
1sg:Hab- |
can:L |
sit[Bsq] |
today |
Neg |
|
‘I cannot sit today’. |
13.3. Serial construction
11Beng has a very limited instantiation of the serial construction in the form of “nṵ̄ or tá + verb phrase”. The verb nṵ̄ ‘to come’ or tá ‘to go’ has the same morphological form as the second verb:
(250) |
Ó |
tá |
à |
wò. |
3sg :Pst+ |
go:L |
3sg |
do:L |
|
‘He went and did it’. |
12Another constraint on the serial construction is that it is used only in those TAM values where the verbs are not marked by suffixes: in the preterite, the habitual, the optative, and the conditional; so the two verbs not only have identical form but the form is suffixless. In other TAM constructions (perfect, progressive, stative, and future) a goal converb (see 13.4) of the second verb is used instead of the serial construction with identical verb forms:
(251) |
Ó |
tá-nā̰ |
drɛ̰̄ |
*wō-nā̰ |
/ *wō / |
OKwō-yà. |
3sg:Pst+ |
go-Prf |
work |
do-Prf |
do[Bsq] |
do-GL |
|
‘He went and worked / He went to work’. |
13.4. Converb constructions
13A goal converb can depend on three verbs: with tá ‘to go’ and nṵ̄ ‘to come’ it describes the goal of movement; with bɔ̄ ‘to come from’ it depicts the subject’s activity at the point of departure:
(252a) |
Ŋ́ |
nṵ́ |
drù-yâ. |
1sg:Pst+ |
come:L |
walk-GL |
|
‘I came to walk’. |
(252b) |
(*Ŋ́ |
drɛ̰̄ |
wò |
drù-yâ). |
1sg:Pst+ |
work |
do:L |
walk-GL |
|
(‘I worked to walk’.) |
(253) |
Ŋ́ |
bɔ́ |
drù-yâ. |
1sg:Pst+ |
come.from:L |
walk-GL |
|
‘I came from a walk’. |
14The goal converb cannot be separated from the motion verb by any constituent, behaving as a typical argument, rather than a modifier (see 12.2; compare (Gusev 2004)).
15As mentioned in 6.2, the locative nominalization in –ya can be used to express action simultaneity, although this usage is rare:
(254) |
Ŋ-ó |
jàtèlí |
kɛ́-ló |
drù-yá. |
1sg-St+ |
thought |
V-Prog |
walk-Plc |
|
‘I am thinking while walking’. |
13.5. Temporal and conditional clauses
16This section describes subordinate clauses that precede the main clause and are structurally outside of it. Goal and complement clauses that are embedded inside the main clause have aleady been characterized; interestingly, the preposed vs. embedded subordinate clauses are marked with two distinct positional classes of complementizers. Embedded clauses (such as goal clauses) have a complementizer on the left edge; preposed subordinate clauses have a complementizer on the right edge. The distinction follows the predictions of J. Hawkins’ theory of word order whereby the head of a subordinate constituent should gravitate towards linear proximity to the head of the superordinate phrase (Hawkins 1990).
13.5.1. Temporal clause: the topic construction
17It is noteworthy that two common constructions (not counting juxtaposition) that express temporal relations between clauses are marked exacly like information structure relations of topic and focus. The main temporal complementizer is ná̰, identical to the topic marker:
(255) |
Gbɔ̌ŋ̀ |
ó |
tá |
pɔ̄ú |
ná̰ |
ó |
zrá |
klɛ́ŋ́ |
nḭ̀ |
wó. |
Gbong |
3sg:Pst+ |
go |
field |
Top |
3sg:Pst+ |
get.lost |
forest |
Def |
IN |
18‘Gbong went to the field and got lost in the forest’ (literally: ‘When Gbong went to the field, he got lost in the forest’. Beng’s fields are often located quite far from their villages, and there are even special temporary settlements for people working in those remote fields.)
(256) |
Ŋ́ |
nṵ́ |
ná̰ |
ŋ́ |
zrô. |
1sg:Pst+ |
come:L |
Top |
1sg:Pst+ |
wash:L |
|
‘I came and washed’ (literally: ‘When I came, I washed’). |
19The selection of TAM values in the main clause (after ná̰ ) follows general TAM semantics. TAM marking in the embedded clause adheres to special rules. If the situation of the subordinate clause precedes that of the main clause, as in (256), the preterite construction is used in the subordinate clause. Simultaneity of the two situations is marked in the subordinate clause by the future (sic!) construction, which has in this case progressive interpretation, or with a semantically appropriate construction with default present time reference (stative, adverbial clause, etc.). Clearly, this usage of the future construction reflects the fact that the future construction historically had a progressive meaning, even though it was replaced in the core progressive contexts by a newly grammaticized form in -lɛló, and was only retained in subordinate contexts and as a future form. Compare (255) and (257a):
(257a) |
Gbɔ̌ŋ̀ |
ò-ó |
tá |
pɔ̄ú |
ná̰ |
ó |
zrá |
klɛ́ŋ́ |
nḭ̀ |
wó. |
Gbong |
3sg-St+ |
go |
field |
Top |
3sg:Pst+ |
get.lost |
forest |
Def |
IN |
|
‘Gbong got lost when he was going to the field’. |
(257b) |
Ŋ-ó |
klóó |
ná̰, |
ŋ̄ |
dā |
ò |
1sg-St+ |
little |
Top |
1sg |
mother |
3sg:Hab+ |
gbéné |
lɛ̀ |
zɔ̰̀ |
fɛ̰́ |
dōdō. |
manioc |
Def |
pound |
day |
some |
‘When I was little my mother would pound manioc sometimes’. |
20Finally, conditional mood is used in the sense of habitual aspect:
(258) |
Gbɔ̌ŋ̀ |
ô |
tá |
pɔ̄ú |
ná̰ |
ò |
zrà |
klɛ́ŋ́ |
nḭ̀ |
wó. |
Gbong |
3sg:Cnd+ |
go |
field |
Top |
3sg:Hab+ |
get.lost:L |
forest |
Def |
IN |
|
‘When Gbong goes to the field he usually gets lost in the forest’. |
21Particle fɛ̰̄ when added to a temporal clause gives it a conditional flavor, which can be expressed in English with the complementizer since:
(259) |
[Fɛ̰̄ |
ā |
pè |
wà-ā |
tá |
ɛ́ |
ná̰], |
dé |
Rel |
3sg:Pst+3 |
say:L |
3sg-St- |
go |
Neg |
Top |
who |
|
ó |
blɔ̄ |
nà̰ |
ò |
à |
bɛ̄ |
srá |
ɛ̄? |
St+ |
here |
for.to |
3sg:Hab+ |
3sg |
trace |
take |
Foc |
‘Since he said he’s not going, who’s here to replace him?’ |
13.5.2. Temporal subordinate clause: the focus construction
22Temporal sequence of two clauses can also be marked by a special construction marked in the same way as the focus contruction: clause A + ɲɛ̰̄ + clause C + ɛ̄, meaning ‘A, then C’. Example:
(260) |
[Ŋó |
kpà̰ŋ́ |
nṵ̀ŋ̀ |
ŋò |
srà] |
ɲɛ̰̄ |
3pl:Pst |
whip |
PL |
3pl |
take:L |
Foc |
[ŋó |
nṵ́ |
à̰ŋmô |
mɛ̰̀ |
kpà̰] |
ɛ̄. |
3pl:Pst+ |
come:L |
hyena |
beat:L |
much |
Foc |
‘They took whips and then came and beat the hyena hard’. |
13.5.3. Conditional clause
23The structure of conditional clauses is protasis + conditional complementizer + apodosis. There are two conditional complementizers, dɛ́ɛ̄, used with protasis in affirmative polarity and nḭ̄, used with negated protasis. Before the protasis one can also find an optional marker fɛ̰̄ or ò dɔ̄ kē, literally ‘let it be set that…’
24TAM marking in the apodosis follows the general semantics of TAM. Protasis exhibits some special rules of TAM marking:
-
in case of a condition in the past or present that the speaker believes can be true (‘real conditional’), the same TAM constructions are used as in independent clauses:
(261) |
Fɛ̰̄ |
wǎ |
mlɔ̰́ |
à |
lō |
ɛ́ |
nḭ̄ |
Rel |
3sg:Pst- |
meet:L |
3sg |
with |
Neg |
if.Neg |
wà-ā |
à |
jrɛ́ŋ́ |
dɔ̰̄-lɛ̀ |
ɛ́. |
3sg-St- |
3sg |
enough |
know-Res |
Neg |
25‘If he didn’t meet her he doesn’t know much’ (the protasis exhibits regular preterite construction);
-
in case of a condition in the future or a habitual condition without concrete time reference (‘potential conditional’), protasis is marked with conditional mood (or more rarely with the appropriate indicative TAM constructions, future or habitual):
(262a) |
Fɛ̰̄ |
ô |
srǒ |
dɛ́ɛ̄ |
ŋó |
nṵ̄ |
gbɔ̀. |
Rel |
3sg:Cnd |
arrive |
if |
1sg-St+ |
come |
also |
|
‘If he comes I am coming too’. |
(262b) |
Mḭ̂ |
gō |
sɔ̰̀ŋ̀ |
mà̰ |
dɛ́ɛ̄ |
Ècí |
ɛ̀ |
ó |
mḭ̄ |
yē-lɛ̀. |
2sg:Cnd |
hide |
person |
CONT |
if |
sky |
Def |
St+ |
2sg |
see-Res |
|
‘If you hide from people, God still sees you;’ |
-
in case of a condition that the speaker believes to be false (counterfactual condition), protasis is marked with optative mood when referring to past events, or appropriate indicative forms when referring to the present. Besides TAM, counterfactual conditionals are obligatorily marked with particle ŋ́gǒ after the conditional complementizer (this particle can also optionally appear with potential future conditions that only possibly can be false). Example:
(263) |
Ŋ̀ |
yí |
ɲḭ́-lɛ́ |
lú |
dɛ́ɛ̄ |
1sg:Hab+ |
water |
cool-Nmlz |
buy[Bsq] |
if |
ŋ́gǒ |
wálí |
wā |
ŋ̄ |
wɔ̄lì |
drɛ̄ |
ɛ́. |
NGO |
money |
St- |
1sg |
POSS |
anymore |
Neg |
‘If I had bought cold water, I wouldn’t have money anymore’. |
(protasis is marked with optative mood expressed by a combination of a habitual subject pronoun with the base form of the verb).
13.6. Relative clause
26Discussion in this section follows (Paperno 2008b), omitting the relativizing function of nominalizations that have been briefly characterized in sections 6.2, 6.3, 6.6, and 6.7.
13.6.1. Head-external relative construction
27Relative clauses are marked with the combination of a preposed particle fɛ̰̄ (which can also be thought of as a relative determiner, see 13.6.2 for arguments to this effect) and a postposed marker ná̰ that equals the topic marker. The ná̰ element can be omitted before a pause, and is always omitted before another ná̰ marker in the topic-marking function. The relativized position in the relative clause is filled by a resumptive pronoun that agrees in person and number with the head NP if a pronoun is possible in the given position or left empty otherwise. Rarely, when the noun phrase is topicalized and separated by a pause, the complementizer can, but does not have to, be omitted (264b).
28As mentioned above, the most common complementizer in relative clauses is ná̰. Relative clauses can also employ a conditional complementizer dɛ́ɛ̄ /nḭ̄ (dɛ́ɛ̄ occurs after affirmative conditional clauses and nḭ̄ after negative conditional clauses). Relative clauses with the conditional (dɛ́ɛ̄) differ in meaning from the main type of relative clauses (with ná̰) and include a conditional element in their semantics (246d). Unlike in relative clauses with ná̰, the statement expressed by a relative clause with a conditional is not presupposed to be true. Relative clauses with dɛ́ɛ̄ (examples 264d,e) can be roughly rendered in English using words whenever, whichever, etc. All of the following examples come from real texts; all of them feature topicalization of the whole relative construction.
29The conditional construction can be used to modify a noun with reference to a future event (264e), even when there is no sense of uncertainty as to whether this event will happen (uncertainty as to whether condition would hold seems to be a common meaning element of English conditionals). Compare (264e) to an analogous example but with a past event in the relative clause (264f), and no conditionality involved (the relative clause is presupposed true); ná̰ is used in this case. Examples:
(264a) |
Ó |
blànâ |
[fɛ̰̄ |
ŋó |
à |
klà |
Kùàsí |
dḭ́ |
ná̰] |
klṵ̀à̰. |
3sg:Pst+ |
banana |
Rel |
3sg:Pst+ |
3sg |
put |
Kouassi |
APUD |
Top |
steal:L |
|
‘He stole the banana that had been put next to Kouassi’. |
(264b) |
Zrɛ̈ |
[fɛ̰̄ |
mḭ̄-ó |
à |
yé]i |
mḭ̀ |
mḭ̄ |
[zrɛ̈ |
bì-ɛ̀]i |
yā. |
way |
Rel |
2sg-St+ |
3sg |
on |
2sg:Hab+ |
2sg |
way |
this-Def |
walk |
|
‘Walk on the way you’re standing on’. |
(264c) |
Ŋ̄ |
bábá |
nṵ̀ŋ̀ |
[fɛ̰̄ |
mḭ́ |
ŋò |
dɛ̀ |
ná̰], |
ŋmā |
ŋò |
yɔ̀. |
1sg |
sheep |
pl |
Rel |
2sg:Pst+ |
3pl |
kill:L |
Top |
1sg:give |
3pl |
other |
|
‘Give me the replacement for the sheep you killed’. |
(264d) |
Pɔ̄bɛ̄ |
[fɛ̰̄ |
ǒ |
yōnó |
dɛ́ɛ̄] |
wà |
zìn |
gō |
ɛ́. |
scar |
Rel |
3sg:St+ |
forehead |
if |
3sg:Hab- |
can:L |
hide |
Neg |
|
‘You can’t hide a scar on your forehead’ (literally: ‘Whichever scar is on a forehead, it can’t be hidden’). |
(264e) |
Yrámà̰ |
[fɛ̰̄ |
bàāŋ̀á̰nḭ́ŋ́yé |
lɛ̀ |
ô |
pɔ̰́ |
dɛ́ɛ̄], |
wà |
mḭ̀ |
wálɛ́ |
lɛ̀ |
gá. |
time |
Rel |
end.of.rain.season |
Def |
3sg:Cnd |
come |
if |
3sg:Hab- |
2sg:Hab+ |
yam |
Def |
pick |
30‘When the rain season ends, gather yams’ (literally: ‘whenever there’s end of the rain season, gather yams’).
(264f) |
Gblē |
[fɛ̰̄ |
ŋ́ |
zīē |
yä̰ |
ló |
wē |
cà |
yesterday |
Rel |
1sg:Pst+ |
kapok |
this |
on |
there |
look |
sḭ́ŋ́ |
ná̰] |
ŋ́ |
dóbà |
dō |
yè |
wē. |
|
closely |
Top |
1sg:Pst+ |
monkey |
one |
see:L |
there |
|
‘Yesterday while watching this kapok tree closely, I saw a monkey there’. |
13.6.2. Head-internal fɛ̰̄-construction with dɛ́ɛ̄ and other arguments for treating fɛ̰̄-constructions as originally head-internal
31Some properties of fɛ̰̄ suggest that it is not simply a relative clause marker but a relative determiner. For instance, before fɛ̰̄, the definite article lɛ̀ is blocked. After a demonstrative, lɛ̀ is generally required, but lɛ̀ is absent in the presence of a relative clause, cf. (265a) vs. (265b).
(265a) |
[Pɔ̄ |
bì |
fɛ̰̄ |
mḭ́ |
à |
lù |
ná̰]NP, |
wà-ā |
gɛ̄ŋ̄ |
ɛ́. |
thing |
this |
Rel |
2sg:Pst+ |
3sg |
buy:L |
Top |
3sg-St- |
good |
Neg |
|
‘This thing that you bought, it is not pretty’. |
(265b) |
Pɔ̄ |
bì |
*(lɛ̀) |
wà-ā |
gɛ̄ŋ̄ |
ɛ́. |
thing |
this |
Def |
3sg-St- |
good |
Neg |
|
‘This thing is not pretty’. |
32The interaction of fɛ̰̄-relativization with the expression of the definite article can be explained if fɛ̰̄ is, at least historically, a determiner occupying the same position as the article lɛ̀. And fɛ̰̄ does occupy the position of a determiner (after the head noun) in a rare variant of the relativization construction with dɛ́ɛ̄, in which a head NP with fɛ̰̄ is found within the relative clause (such constructions are head internal):
(265c) |
[Ô |
[pɔ̄ |
fɛ̰̄]NP |
srá |
dɛ́ɛ̄] |
wà |
klà |
à |
bɛ̄ló |
ɛ́. |
3sg:Cnd |
thing |
Rel |
take |
if |
3sg:Hab- |
put:L |
3sg |
place |
Neg |
|
‘He doesn’t put things where they belong’ (literally: ‘whatever thing he takes, he doesn’t put it in its place’). |
(265d) |
[Tɔ̰̄ |
wā |
[sɔ̰̀ŋ̀ |
fɛ̰̄]NP |
mà̰ |
ɛ̰́ |
nḭ̄] |
yròbítà |
wà |
sɔ̀bì |
ɛ́. |
curse |
St- |
person |
Rel |
on |
Neg |
if |
water.snake |
3sg:Hab-3 |
bite:L |
Neg |
|
‘If there’s no curse on a person, a water snake won’t bite him’ (literally: ‘If a curse is not on whichever person…’). |
(265e) |
[Mḭ̂ |
wlō |
[gbě |
fɛ̰̄]NP |
wó |
dɛ́ɛ̄] |
2sg:Cnd |
move |
village |
Rel |
in |
if |
ŋà |
gbōpì |
à |
yé |
ɛ́. |
3pl:Hab- |
defecate:L |
3sg |
in |
Neg |
33‘One doesn’t defecate in the village one is moving from’ (literally: ‘whatever village you are moving from …’) .
34These examples suggest that even when the head noun is outside the relative clause, fɛ̰̄ might still be a determiner of the head NP.
35Another instance of head internal relativization can be seen in relative clauses with presentative markers like ɛ̀ ‘this is’:
(265f) |
[[pɔ̄ |
fɛ̰̄]NP |
ɛ̀ |
ná̰]Rel, |
wà-ā |
pɔ̄ |
bì-lɛ̀ |
dɔ̰̄-lɛ̀ |
ɛ́ |
thing |
Rel |
this.is |
Top |
3sg-St- |
thing |
this-Def |
know-Res |
Neg |
|
‘This thing (lit. the thing that this is), he doesn’t know this thing’. |
36Since ɛ̀ cannot function as a full sentence on its own, we have to assign this relative clause a structure where the head noun is its subject. This construction is essentially idiomatic, functioning as a complex demonstrative. Such demonstrative relativizations form a closed class, so they could be treated as lexicalized relics of head-internal fɛ̰̄-relativization.
37An additional piece of evidence comes from topicalization of phrases modified by relative clauses. Such topicalized phrases are never accompanied by an additional topic marker (ná̰ ), to which the complementizer ná̰ at the end of relative clauses is phonologically identical. The topic marker ná̰ can often be omitted, but it is usually present in topicalization of adverbial elements (265h). However, such a topic marker is not introduced if the topicalized adverbial is modified by a relative clause, compare (265g) vs. (265h). Topic markers are not used after relative clauses with the complementizer dɛ́ɛ̄, either, compare (265i) vs. (265j).
(265g) |
[Flɔ̰́ɔ̰̄ |
[fɛ̰̄ |
à |
lɛ́ŋ́ |
nḭ̀ |
ó |
ŋ̄ |
wɔ̄lì |
ná̰]Rel]NP, |
today |
Rel |
3sg |
child |
Def |
St+ |
1sg |
at |
Top |
ŋ̄ |
vḭ̄‑lí |
ɛ̀. |
1sg |
love‑Ag |
this.is |
‘Today when I have her daughter she is my friend’. |
(265h) |
[Flɔ̰́ɔ̰̄]NP |
ná̰, |
ŋ̄ |
vḭ̄‑lí |
ɛ̀. |
today |
Top |
1sg |
love‑Ag |
this.is |
|
‘Today she is my friend’. |
(265i) |
[Fɛ̰́ |
dō]NP |
ná̰ |
mḭ̄ |
nà̰ |
à |
lō |
kà |
blè |
mṵ̂ŋ́. |
day |
one |
Top |
2sg |
and |
3sg |
with |
2sg:Hab+ |
agree:L |
again |
|
‘Some day you and she agree again’. |
(265j) |
[Fɛ̰́ |
[fɛ̰́ |
à |
tètè |
ó |
à |
bō |
à |
yúó |
day |
Rel |
3sg |
self |
3sg:Cnd+ |
3sg |
take |
3sg |
mouth |
kē |
ma̰ |
sí |
nā̰ |
dɛ́ɛ̄]Rel]NP |
that |
1sg:Pst+3 |
take |
Perf |
if |
mḭ̄ |
nà̰ |
à |
lō |
kà |
blè |
mṵ̂ŋ́? |
2sg |
and |
3sg |
with |
2sg:Hab+ |
agree:L |
again |
‘The day she (mother-in-law) says “I took her (your wife) away”, will you and she agree again?’ |
38The facts outlined above can be given a straightforward interpretation: the relativization marker fɛ̰̄ is, at least historically, a determiner; the most widespread relativization strategy features an extraction of the NP with fɛ̰̄ from the relative clause.
39The original syntax of relative clauses could have been correlative which is still found in the cases discussed above, examples (265c-265f):
[…[N fɛ̰̄]NP … Comp]S S2
with the option of topicalizing the fɛ̰̄ noun phrase:
[[N fɛ̰̄]NP … Comp]S S2,cf. example (265j).
These types of sentences could have been reanalyzed as involving a topicalized noun phrase with a relative clause, extracted from the main clause S2:
[N [fɛ̰̄ … Comp]S]NP S2
Undoing such topicalization gives the basic relativization pattern:
[…[N [fɛ̰̄ … Comp]S]NP…]S2, see example (264a).
The development proposed here is reminiscent of the scenario proposed by Nikitina (2012) for what she labels as ‘the rise of clause-internal correlatives’ in Southeastern Mande languages.
40To summarize the argument of this section, relative clauses in Beng are originally head-internal, at least in the historical sense. The argument can also be interpreted in favor of analyzing relative clauses as originally head-internal in syntactic derivation (Kayne 1994). Historical and derivational interpretations of head-internal syntax of relative clauses are compatible but not isomorphic, reminiscent of the relation between historical and derivational processes in phonology.
41There is one argument in favor of the historical rather than synchronic interpretation of the head-internal status of Beng relative clauses. One important piece of evidence used in the argument above was that the determiner lɛ̀ is not used in the presence of a relative clause marker fɛ̰̄. However, the definite article nḭ̀ (variant of lɛ̀ after /ŋ/, see 8.3), is attested before fɛ̰̄:
(265k) |
Ŋ̄ |
baba |
lɛ̀ |
à |
sá̰nḭ̄ŋ̄ |
lɛ́ |
[à |
1sg |
sheep |
Def |
3sg |
mark |
3sg:Pst+:Cop:L |
3sg |
blɛ̀ŋ̀ |
trōŋ̄ |
nḭ̀ |
[fɛ̰̄ |
ò |
cḭ́-lɛ̀ |
ná̰]]NP |
ɛ̄. |
left |
ear |
Def |
Rel |
3sg:St+ |
cut‑Res |
Top |
Foc |
‘The mark of my sheep is its left ear which is cut’. |
42Examples like this suggest that the incompatibility of fɛ̰̄ and determiners is not strict, and the determiner status of fɛ̰̄ might not be synchronically valid but could rather be a historical relic.
Чтобы цитировать эту статью
Ссылка в печатном виде
Denis Paperno, «Complex sentences», Mandenkan, 51 | 2014, 89-101.
Электронная ссылка
Denis Paperno, «Complex sentences», Mandenkan [Онлайн], 51 | 2014, Выложить онлайн 01 juin 2014, Наводить справки в 29 mars 2024. URL: http://journals.openedition.org/mandenkan/569; DOI: https://doi.org/10.4000/mandenkan.569
Верх страницыАвторские права
The text only may be used under licence CC BY-NC-SA 4.0. All other elements (illustrations, imported files) are “All rights reserved”, unless otherwise stated.
Верх страницы