Sango is one of the official languages of the Central African Republic (CAR), the other being French. The name derives from one of the dialects of its major lexifier, Ngbandi. Various means are used to distinguish it from this dialect (sango riverain ‘river Sango’). Some educated Central Africans refer to it as sango standard in French although there is no standardized form. It has also been called a sabir (jargon), sango commercial (‘commercial Sango’), and vehicular Sango (sango véhiculaire). In popular speech until independence it was referred to as Sango ti turugu ‘soldier’s Sango’. It is spoken by most people in the Central African Republic, with a population of about 3,000,000, and by many inhabitants of the adjoining Cameroon and Chad republics. In the diaspora it is spoken in Brazzaville and in France. It is the dominant language of Bangui, the capital of the Central African Republic, with a population of half a million, where it is the first language of an increasing number of young people.
Sango is in origin a pidgin, created in the last two decades of the 19th century, when representatives of the Congo Free State (Etat Indépendant du Congo) and the French government took possession of the two sides of the Ubangi River. The first expedition of Europeans to arrive at the headwaters of the Ubangi River, 1,500 km from Kinshasa, was in 1887, led by Alphonse Van Gele, representing King Leopold II of Belgium. He spent a year among the riverine Ngbandi people, who were quickly named Sangos, Yakomas, Dendis, and Abiras, buying ivory and establishing the claims of the Congo Free State in the region. The French established themselves in what came to be known as Bangui in 1889 and immediately made forays northward into the hinterland and eastward on the river, establishing what came to be known as the Haut-Oubangui (Upper Ubangi) and then Oubangui-Chari (Ubangi-Shari).
Since the culture of the Yakomas and their Ngbandi co-ethnics was based on fishing, iron work, and trading and they possessed canoes twenty metres in length, they figured importantly in the occupation of the territory by the colonists in providing canoes and canoers. The African employees and militiamen brought by the Europeans–men who came from East and West Africa, plus speakers of Bantu languages from the equatorial regions of the Congo River–would have jargonized Ngbandi at a time when Kituba (see Mufwene, this volume) and Bangala-Lingala (see Meeuwis, this volume) were also coming into existence. The new language that emerged was in this process reduced in grammar and lexicon by comparison with the source language and for that reason is called a pidgin. The earliest attestation of a jargon is 1893. Others claim that a lingua franca based on Ngbandi existed before Europeans arrived and that it underwent changes after their arrival without becoming a pidgin.
The indigenous languages that could have figured in the origin of Sango belong to the Ubangian family of Niger-Congo, primarily Gbanziri and Ngbandi, both riverine, and possibly Nzakara; any significant influence from Banda dialects must have come as it spread northward. Foreign Africans serving the whites spoke several Bantu languages and numerous languages from the Guinea Coast and the West Coast, as well as the Western Sudan.
Sango spread first among the people living along and close to the banks of the upper Ubangi River, when it accompanied the military and administrative occupation of the French territory and the rapacious exploitation by the colonial government and the capitalist commercial societies. Since speakers of Ngbandi in the French militia and the foreign Africans in the employ of traders were numerous, and since they had direct contact with the inhabitants, they must be considered the agents of spread. Catholic missions, the first two established in 1894, played a lesser role. However, Baptist Mid-Missions, established in the 1920s, adopted a Sango-only linguistic policy, quickly introducing a written form of the language and translations of scripture that were used by other Protestants. The Catholic standard, not fully established until after the Second World War, was different.
Education has always been in French, provided by Catholics and the government but not Protestants in most of the country. Despite its use in religious education by the latter, Sango has never achieved the status of a written language except for the adoption of an orthography that recognizes three levels of tone. It has yet to be used by any institution. Sango is hardly ever seen in posters and advertisements. It began to appear in radio broadcasts at the end of the 1950s and since independence it has been used for news reports and broadcasts of an educational and religious nature. Since the 1970s, music in the “Kinshasa style” with Sango lyrics has been popular.
Responding to international initiatives in popular education, the government adopted projects to introduce Sango in the first years of schooling and the Institut de Linguistique Appliquée was created to further this aim, but Sango has never been in the curriculum of any government-approved school. French continues to be the language of education, the means of social and economic advancement, the spoken language in formal settings, and, most of all, the written language, as in newspapers, both those published by the government and those published privately. French is also the dominant language in radio and television broadcasts. It has since independence in 1960 had an increasing influence on Sango in syntax and phonology as well as lexicon (see Taber 1979).
Immediately after independence Sango was considered constitutionally the vehicular language of the country. In 1964 it was declared the national language (langue nationale). In January 1965 a decree established the Commission Nationale pour l’Etude du Sango (‘National Committee for the Study of Sango’). In 1974 there was established the Institut Pédagogique National (IPN, the ‘National Pedagogical Institute’). Its purpose was to initiate research on Sango and its use in schools. It was replaced in October 1983 by the Institut National d’Education et de Formation (INEF). In January 1984 a decree established an official orthography. In the final years of the presidency of André Kolingba (of Ngbandi ethnicity) Sango was made co-official with French (April 1991).
The prognosis for Sango, given continual political and civil strife and the lack of funds, is not encouraging. It remains the Central African Republic’s spoken language, French its written language.
There are seven oral vowels, only five have counterparts with nasalization (see Table 1), but nasalized vowels are infrequent. Vowels appear long in contractions, as in tee < tene ‘say, speak.’ The central vowel [ə] occurs as a variant in the speech of young people, possibly derived from French or from some Banda dialect. French rounded vowels also appear in French words. The letter <n’> after a vowel is used to indicate nasalization in the orthography (e.g. nyen’ [ɲɛ̃] ‘what’). The close-mid vowels are not, and never have been, represented in the orthography, and there are no lexemes that differ only by them (in minimal pairs).
Table 1. Vowels |
|||
front |
central |
back |
|
close |
i ĩ |
u ũ |
|
close-mid |
e |
o |
|
open-mid |
ɛ ɛ̃ |
ɔ ɔ̃ |
|
open |
a ã |
The consonants (Table 2) are almost identical to the ones in the source language, except that /mv/ seems to be disappearing, and /l/ and /r/ are distinct, whereas in Ngbandi they are phonologically conditioned variants of a single phoneme. The variation between /h/ and /ʔ/, which occur only word initially, may also be a vestige of Ngbandi. Some variation is also due to the influence of substrate languages: for example, the macrolectal /nz/ occurs as [nz] and [ndʒ]. The greatest amount of variation occurs in the contracted speech of young people, a topic that cannot be covered in a few words.
Table 2. Consonants |
||||||||
bilabial |
labio-dental |
labio-velar |
alveolar |
palatal |
velar |
laryngeal |
||
plosive |
voiceless |
p |
kp |
t |
k |
ʔ |
||
voiced |
b |
gb |
d |
g |
||||
prenasalized |
mb |
mv |
mŋgb <ngb> |
nd, nz |
ŋg <ng> |
|||
nasal |
m |
n |
ɲ <ny> |
(ŋ <ng>) |
||||
trill |
r |
|||||||
fricative |
voiceless |
f |
s |
h |
||||
voiced |
v |
z |
||||||
approximant |
w |
l |
j <y> |
Sango has three levels of pitch, as does Ngbandi; the few instances of falling and rising contours can be represented as sequences of two vowels, as they are here, or with <â> and <ǎ>. There is great variation in what occurs phonetically, and no study of this feature of Sango’s phonology has yet been undertaken. One gets the impression, anyway, that contours are being replaced by level pitch, as in kumase from commencer, which frequently ends with mid-tone. Pitch features in only a few words to differentiate lexemes and has very little role in grammar (except for a rare vestige of Ngbandi) and only as an alternate to a periphrastic construction. In this introduction, the definite article with high tone is distinguished from the logophoric pronoun by marking the latter as nï. Tone was indicated in religious publications only for a few words, a practice continued in the latest version of the Bible. The introduction of tone marking in the 1970s was primarily to help readers to correctly pronounce neologisms like the word for ‘environmentalist’. The Central African government nonetheless adopted in 1984 the recommendation of the Institute of Applied Linguistics (IAL): as in <â> for high tone, <ä> for mid-tone, and <a> for low. The IAL, with the collaboration of members of SIL International, beginning in the 1990s, published several booklets with tone diacritics to be used in literacy programs.
Questions are not ordinarily marked by a special feature of pitch, but when a question is repeated or when expressing emotion, a rising-falling contour with greater emphasis occurs on the final syllable represented by <?!>.
Nouns can be morphologically simple, compounded (like koli-kondo [man-fowl] ‘rooster’), and morphologically derived (e.g. verb + -ngo). Only nouns and the demonstrative can be pluralized with the prefix a-, e.g. a-wali ‘women’, a-so ‘these’.1
Natural gender is conveyed with ‘man’ and ‘woman’ in phrases with or without ti, except for animals: melenge (ti) koli [child (of) man] means ‘male child, son, young man.’
The numerals are the following. The decades are made with bale ‘unit of ten’, hundreds with ngbangbu, and thousands with saki.
(1) oko one omana six
use two mbasambala seven
ota three miombe eight
osyo four ngombaya nine
uku five bale oko ten
These can occur (a) independently as an utterance (as in ‘I want five’), (b) in ordinal function with the definite article, e.g. use ni [two DEF] ‘the second one’, and (c) as an ordinal preceding a noun, as in use wali ni [two woman DEF] ‘the second wife’, or (d) as a cardinal numeral following a noun, as in lango use [night two] ‘two nights (or days)’.
The connective ti ‘of, to’ relates modifiers to the head noun in noun and verb phrases. As with the genitive in other languages, ti is used in noun phrases to signal possession and qualification or attribution. It is the most frequently used word in the language. Examples of noun phrases with a ti modifier are the following.
(2) koli ti mbi [man of 1SG] ‘my husband’
fen’ ti zo [scent of person] ‘smell of human being’
nyama ti nyen’ [meat of what] ‘what kind of meat?’
nzara ti zo [hunger of person] ‘desire for human flesh’
The semantically general oblique preposition na introduces noun phrases in clauses to mark dative, locative, benefactive, and other roles. Frequent are relational phrases with body parts in a spatial sense (a usage common in many African languages): na ya (ti) [OBL belly [of]] ‘in, inside’. These phrases with the definite article ni can be translated adverbially.
(3) Lo gwo ndo ape.2
3sg go place neg
‘She didn’t go anywhere.’ (gwo < gwe)
(4) Lo leke a-kungba (ti) lo na a ti da ni kwe.3
3sg prepare pl-belongings (of) 3sg obl interior of house def all
‘She gathered up all her belongings in the house.’ (a < ya)
(5) Na peko ni lo ken’.
obl back def 3sg refuse
‘He later refused.’
Personal pronouns are of two types: pronouns of direct discourse, and logophoric pronouns (i.e. pronouns of indirect discourse, that is, reported speech). In addition, the definite article ni and sometimes lo are used in reference to inanimate objects with the meaning ‘it’. The indefinite marker mbeni can be used pronominally.
(6) Mbeni zo a-ga awe.
indf person pm-come already
‘Someone has come.’
(7) A-mbeni a-ga awe.
pl-indf pm-come already
‘Some (people) have come.’
Pronouns of direct discourse include three pronouns in the singular number but only two distinct forms in the plural (see Table 3).
Table 3. Personal pronouns (of direct discourse) |
|
subject/object |
|
1SG |
mbi |
2SG |
mo |
3SG |
lo |
1PL |
e/i |
2PL |
ala |
3PL |
ala |
indf |
mbeni |
Some of these occur in variant forms resulting from contraction or assimilation: e.g. mi < mbi, me < mo, aa < ala. The ambiguity created by homophony for second and third persons plural in ala is increased by its use for both second and third persons singular (‘you’ and ‘he’ or ‘she’) in a respectful or deferential manner, similar to that of French. The first Protestant missionaries appear to have introduced ani for the first person plural. It is still used for ‘we’ by some Central Africans and marks them as Protestants.
Plural pronouns are used in a resumptive manner when singular pronouns are coordinated:
(8) Laso mbi na lo, i ke gwe biani.
today 1sg and 3sg 1pl cop go truly
‘Today he and I will go for sure.’
The forms for logophoric pronouns are nï (with mid-tone) for singular and anï for plural. (In the latter a- is the plural prefix.) Although the grammatical distinction between direct and indirect pronouns is categorical, their use is variable. These are matters yet to be studied. There is also a usage one might call semi-direct that is signalled by the use of the particle o and occurs most frequently in reported speech with mbi and lo.
(9) Mama na a-tene, a-ninga nï baa lo ape.
mother def pm-say, pm-be.long.time 1sg.log see 3sg neg
‘The mother said that she (the mother) hadn’t seen her (the daughter) in a long time.’
(10) Lo o, lo zo na nï yi ti nï so nï te.
3sg clt, 3sg roast obl 1sg.log thing of 1sg.log rel 1sg.log eat
‘(He said) that she (his wife) should roast for him this stuff of his to eat.’
Adjectives are the only words that can co-occur with nouns, some preceding them, others following them (some perhaps only in frozen phrases), but they can also occur alone with the definite article as nominals. Examples of those that precede nouns and noun phrases are kete ‘little, small’, nzoni ‘good’ and vuru ‘white’; those that follow nouns and noun phrases are mingi ‘much, many’ and kwe ‘all’. In some phrases with postposed adjectives, which might now be considered compound nouns, the meaning is different: mama-kete [mother-small] ‘aunt (mother’s younger sister)’, as well as some other kinship terms, and the phrase zo voko [person black] ‘an African’. In a pluralized noun phrase the adjective carries the plural prefix.
(11) A-kete a-melenge ti a-kondo
pl-small pl-child of pl-chicken
‘small chicks’
Comparison of adjectives and adjectival verbs is expressed with the predication a-hon’ ‘(it) surpass’ (see 12), and a superlative is expressed with a-lingbi ape [PM-be.possible NEG] ‘it has no equal’ (see 13).
(13) A-melenge so ka a-ke na yoro a-lingbi ape.
pl-child dem there pm-cop obl charm pm-be.possible neg
‘Those kids over there have an awful lot of charms.’
Verbs in commands are usually preceded by a pronoun: ala ga ma! [2PL come PCL] ‘Come on now!’ All verbs and the copula can occur with the predicate marker (PM) a-, which obligatorily follows a noun subject, where it is comparable to a dummy predicate marker, and never follows a pronoun subject. In the speech of young people in Bangui the prefix is frequently deleted in verb chains: lo ga lo hunda ‘she/he came and asked’ can be rendered lo ga a-hunda or lo ga hunda. Verbs, except the copula yeke (in speech usually ke), can also be nominalized with -ngo. Some verbs –such as hinga ‘know’, ken’ ‘refuse’, and yi ‘want’, and including French verbs such as peut (pe in Sango) ‘be able to’ and commencer (with many variants) ‘begin’– can occur with a dependent verb that is linked with ti, much like infinitival to in English.4 (But there is no reason to consider these as auxiliary verbs at the present time.) Sango’s “filler verb” is sara ‘do’. With a nominal complement it replaces verbs that were lost in pidginization, e.g. sara inon’ [make urine] ‘urinate’.
The copula yeke is an innovation in Sango; the source language has none. Common variants of yeke are eke, ke, e, i. It is used with predicated adjectives, locative phrases, and noun phrases.
(14) Mo eke?!
2sg cop
‘Are you (there)?’ (a common greeting)
(15) Da ti ala na a-ke na ya ti gbako.
house of 3pl def pm-cop obl interior of forest
‘Their house was in the forest.’
(16) Koli ti mbi na a-ke nzoni zo ape so.
man of 1sg def pm-cop good person neg thus
‘My husband is not a good person, as you can see.’
Predications are qualified for time (hence, tense) by temporal adverbs, as in (17). There is no contrastive system for tense-mood-aspect marking, but the copula with a verb lacking the PM functions as an auxiliary for continuative, habitual, imperfective, unrealized (the future), repetitive, etc., with no constraint for time (e.g. (8), (17)). (In such instances the copula is no more reduced phonologically than in other uses.) One is justified in denying awe (from Ngbandi a-we [PM-be.finished]) the function of perfective; it can be rendered ‘already’. In casual speech, and especially among young people, awe occurs in reduced forms, one of them being w following a vowel. This adverb is based on the predication a-we [PM-be.finished], which still occurs authentically in Sango (as in ‘it [the gasoline] is all gone’). It marks realization, completion, and so forth, but should not be considered part of the verbal system. The lexically independent use of awe is seen in (42). Although I once believed (as did others) that it was a grammatical marker of tense or aspect, recent analysis of a large number of examples demonstrates that this is not the case. The gloss of ‘already’, of course, is not always necessary as some of my translations reveal.5 For a different view on the verbal system see Diki-Kidiri (1988).
The most frequent device to express conditional mood is the use of an equivalent construction with tongana (or to) ‘if, when’.
(17) Fade gbanda Nzapa a-ke mu na mo [...] ande.
quickly distant.future God pm-cop give obl 2sg [...] some.day
‘Some day God will give it to you.’
(18) To mo te a-hon’ ndo ti mo aw [...]
if 2sg eat pm-pass top of 2sg already […]
‘But if you eat too much [...]’ (to < tongana, aw < awe)
A redundant marking with habitual ka after the verb has appeared in recent years, probably adopted from a regional Bantu language. In the following examples both the copula and the habitual marker occur, as they often do. They are also used by themselves.
(19) Nï e te ka ti nï gi nyama ti a-zo.
1sg.log cop eat hab of 1sg.log only meat of pl-person
‘[He said that] he made it a practice of eating only the flesh of human beings.’
(20) Ala ke te ka ti ala nyama pepe.
3sg cop eat hab of 3pl meat neg
‘As for her, she doesn’t make it a practice of eating meat.’
Predications are negated with the marker pepe, or more frequently ape, even in verb chains, followed only by particles. Negative markers and awe never co-occur in a clause:
(21) Ala ga (na) ndo so ape ma!
2pl come (obl) place dem neg pcl
‘Don’t come here!’
(22) A-mbeni a-ga ape.
pl-indf pm-come neg
‘Some (people) have not come.’
The rendition of something in English like (a) ‘I didn’t know that (she/he died)’ and (b) ‘I knew that she didn’t die’ would have at one time been something like the following: mbi hinga titene lo kwi ape [1SG know that 3SG die NEG]. Persons who are literate in French, on the other hand, in written and spoken tests moved the negative marker according to the meaning: e.g. (a) mbi hinga ape titene lo kwi for the first and (b) mbi hinga titene lo kwi ape for the second.
Reflexive and reciprocal voice are expressed in the same way with tere ‘body’. In (23) one could say tere ti mbi ‘my body.’ The logophoric pronoun can also be used, as in (24).
(23) Za, mbi sukula tere si.
leave 1sg wash body then
‘Let me have a bath first.’
(24) Lo sukula tere ti nï.
3sg wash body of 3sg.log
‘He/she washed himself/herself (and not someone else).’
(25) Ala hinga tere depii.
3pl know body since.long.ago
‘They’ve known each other for a long time.’
(26) Ala pika tere sioni.
3pl hit body badly
‘They fought terribly.’
Causative voice is expressed with sara ‘make.’
(27) Ala sara si a-ga nzoni mingi.
3pl make then pm-become good much
‘They made it very good.’
The canonical word order in a clause is Subject-Verb-Object amply illustrated in preceding examples. With a ditransitive verb the recipient is introduced with oblique na as indirect object. The positions of the direct and indirect objects can be reversed.
(28) Ala hunda na mbi nginza ti nze.
3pl ask obl 1sg money of month
‘They asked me for the month’s wages.’
(29) Ala hunda nginza ti nze na mbi.
3pl ask money of month obl 1sg
‘They asked the month’s wages of me.’
A predicated verb can occur without a specific noun subject: a-ga ti si na bi [PM-come to arrive OBL night] ‘just about nightfall’. Some such constructions evoke the passive, a device used in the recent translation of the Bible with a separated PM: a pe ti kangbi ape? [PM can to divide NEG] ‘Couldn’t it have been divided?’ This might be considered a device by subject suppression.
Polar questions are marked with a change in intonation, or with the question marker eski/esi from French est-ce que. The use of eski eliminates a change in pitch, as in (30).
(30) Eski mo hinga ala?
pcl 2sg know 3pl
‘Do you know them?’
In content questions (‘who?’ / ‘what?’ / ‘when?’ / ‘why?’) the interrogative element occurs in initial or non-initial position, but ‘where?’ is always in final position. With these there is no change in intonation. The interrogative nyen’ ‘what?’ occurs as an independent word in a sentence and in interrogative phrases, such as nda-li-ti-nyen’ [end(of)-head-of-what] ‘why?’, ten ti-nyen’, and tongana-nyen’ ‘how?’, as in (34), (36) and (39). These phrases occur in several phonologically variant forms. The nominal function of nyen’ is revealed in a tape-recorded question: a-nyen’ ni [PL-what DEF] ‘what things (are you talking about)?’6 Wa is an interrogative that means ‘where?’ as in Ngbandi, but in Sango it occurs with this meaning in isolation only with the copula: a-k wa [PM-COP where?] ‘Where is he/she/it?’ It occurs most frequently in phrases where its function is that of an interrogative determiner ‘which’: na ndo wa [obl place which?] ‘where?’ and zo wa [person which] ‘who?’ One could consider ndwa and zwa, two of their frequently used reduced forms, as words rather than phrases because wa no longer means ‘where?’ in them. Some interrogatives in content questions can be fronted with focusing laa (FOC) or with a repeated interrogative as in (45) and (46).
(31) Zowa a-ga laso?
who pm-come today
‘Who came today?’
(32) Melenge so a-ke zowa?
child dem pm-cop who
‘Who is this child?’
(33) Mo ga ti sara nyen’ la?
2sg come to do what foc
‘What is it that you have come to do?’
(34) Nyen’ a-ke so mo?
what pm-cop hurt 2sg
‘What’s hurting you?’
(35) Mo si lawa?
2sg arrive when
‘When did you arrive?’
(36) Ten ti nyen’ (laa) lo tene so?!
because of what (foc) 3sg speak thus
‘Why in the world did she speak this way?’
(37) Kogara (ti) mbi a-k (na ndo) wa?
mother.in.law (of) 1sg pm-cop (obl place) what
‘Where is my mother-in-law?’
(38) Nyen’ la a-saa mbi?
what foc pm-do 1sg
‘What would happen to me?’ (la < laa, -saa < -sara)
(39) Mama ti wal na a-tene, nda l ti nyen’.
mother of woman def pm-say end(.of) head of what
‘The wife’s mother said, “Why?”‘ (wal < wali, l < li)
(40) Mo fa nyama ni toneen’?
2sg kill animal def how
‘How did you kill the animal?’ (toneen’ < tongana-nyen’)
(41) Koli ti mbi a-gwe n ndwa
husband of 1sg pm-go obl where
‘Where did my husband go?’ (n = na)
(42) Ala ga awe?
3pl come already
‘Have they come?’
(43) Ala ga lawa?
3pl come when
‘When did they come?’
(44) Mo e te nyama ti nyen’?
2sg cop eat meat of what
‘What kind of meat do you usually eat?’
(45) Samba ti nyen’ laa ala ke nyon’?
alcoholic.drink of what foc 2sg cop drink
‘What kind of alcoholic beverage do you drink?’
(46) Zwa a-ga zwa?
person.what pm-come person.what
‘Who came?’ (zwa < zo wa)
(47) Lawa si mo ke ga?7
when then 2sg cop come
‘When are you coming?’
A noun or verb phrase can be focused with the particle laa, frequently shortened to la, which follows the focused element. It has also been called a topicalizer (“présentatif-indicateur”), a predicating morpheme (“morphème prédicatif”), and thematic marker (“marqueur thématique”) (Diki-Kidiri 1977, 1998).
(48) Nda ti tere ti mbi ni la.
end of story of 1sg def foc
‘That’s the moral of my story.’
(49) Lo oko la a-ke lango na a ti da.
3sg one foc pm-cop sleep obl interior of house
‘She alone slept in the house.’ (a < ya)
(50) Mo goo na ndo ti melenge ti mbi ti sara so laa?!
2sg go obl place of child of 1sg to do dem foc
‘So you’re going to my daughter’s place to do that, are you?’ (goo < gwe)
(51) Mo zonga la mbi.
2sg swear foc 1sg
‘You swore at me there.’
(52) A-ninga, mbi baa ala ape laa, le ti mbi a-gi ala,
pm-be.long.time, 1sg see 2sg neg foc, eye of 1sg PM-seek 2sg
‘It’s because it has been a long time since I saw you that I missed you
si mbi ga ti baa ala.
conj 1sg come to see 2sg
and came to see you.’8
(53) Lo duti na poko da si lo toto so o.9
3sg sit obl back(of) house and 3sg cry thus clt
‘She’s sitting behind the house and is crying, I’m sorry to say.’ (poko da < poko ti da)
(54) A-ke melenge ti lo so a-duti na ya ti da ti
pm-cop child of 3sg rel pm-sit obl interior of house of
‘It was her child who had spent a long time living with her
kogara ti lo a-ninga laa a-sigi na nda ni.
in.law of 3sg PM-be.long.time foc PM-appear obl end def
son-in-law who explained the matter.’
(55) Ni laa mbi kpe mbi ga so.
def foc 1sg run 1sg come thus
‘That’s the reason why I came running like this.’
The coordinating conjunction na is rarely used, but si is versatile with several functions. Even though there are several examples of na in Samarin (1967: 88–91), some of them sound like “white man’s Sango.” The verbs ngba ‘remain’, ga ‘come’, and gwe ‘go’ occur in narratives in the first clause of a verb chain with a discourse or rhetorical function. They occur so frequently that they might be translated ‘and then.’
Although there are several lexical means to relate clauses, juxtaposition also characterizes subordination and dependency, occurring at the beginning of a sentence or following the first clause. Some of these are the following: ka oppositive, ‘but, on the other hand, so’; tenti (teti in early Protestant usage) ‘because’ consists of the noun tene (with mid-tones) ‘word, affair’ and ti; titene (and variant tite) ‘so that’ consists of ti and the verb tene (with low tone) ‘say’; tongana ‘if, when’. French words with temporal meaning like moment followed by so mean ‘when’, probably a calque on one or more Ubangian languages. The recent use of chance is quite original.
(56) <Chance> melenge na a-gwe ti to ngu, [...]
perhaps child DEF PM-go to draw water […]
‘Perhaps the child went to draw water [...]’
(57) Ifo mo zaa gi mbi ge.10
it.is.necessary 2SG leave only 1SG here
‘You have to put me here.’ (zaa < zia)
(58) To lo ga ti mu na mo kasa [...]
if 3SG come to give OBL 2SG stew […]
‘If he should offer you stew [...]’ (to < tongana)
(59) Mbi ga, tite mo sar na mbi.
1SG come so 2SG do for 1SG
‘I’ve come so that you might do the same for me.’ (tite < titene, sar < sara)
So can have the same function as comme from French, usually followed immediately by the subject. The co-occurrence of so [...] so can be interpreted, as I have in the following example, as ‘since [...] thus’ or as the bracketing of a clause analogous to what occurs with relative clauses.
(60) So mo zia ere ti Nzapa da awe so, [...]
since 2SG put name of God there already thus […]
‘Since you invoked God’s name just then, [...]’
So, as the marker of a relative clause (rel), is an innovation in Sango –that is, not derived from Ngbandi, which has its own way of marking a relative clause. Relative clauses are enclosed here in curly brackets {…}. The construction probably arose from a construction with the adverb so ‘thus’, which quickly became used as the demonstrative (Samarin 2007). The verb following the relative-clause marker in restrictive clauses always occurs with the predicate marker, as does every noun phrase functioning as subject of a verb. The postposition of the adverb so in a relative clause, as in (62), may be evidence of the emergence of a bracketing or parenthesized construction for relative clauses, similar to the one reported for Tok Pisin. Presently it may be arbitrary to gloss it one way or another. Subjectival, objectival, and instrumental relative clauses are illustrated here.
(61) mbeni {kasa so a-ke mafuta} 11
some stew REL PM-COP fat
‘some stew that is fatty’
(62) {Koli ni so nï mu so} a-ke nzoni koli ape.
man DEF REL 1SG.LOG marry REL PM-COP good man NEG
‘She said that the man she married was not a good husband.’
(63) {Zeme so mbi doroko na nyama ni} a-za ape.
knife REL 1SG butcher OBL animal DEF PM-be.sharp NEG
‘The knife with which I butchered the animal is not sharp.’
(64) Ala ma {wango so a-<médecin> a-ke mu na ala so}.
2PL hear warning REL PL-doctor PM-COP give OBL 2PL REL
‘Listen to the advice that the doctors will give you.’
(65) {Bongo so lo yu na teer lo so} a-e ti lo?!
clothes REL 3SG wear OBL body(of) 3SG thus PM-COP of 3SG
‘Are the clothes she wears really hers?’ (teer < tere)
Adverbs occur in the verb phrase –some following the verb, some preceding it, and some in both positions or alone as a predication– adding meanings of location, time, quantity, etc. Some of them, in addition to ideophones (descriptive adverbs), are the following: biani ‘certainly, for sure’, da ‘here, there, in place’, fadeso ‘now’, ge ‘here’, gba ‘in vain’, hio ‘quickly’, and mingi ‘much, very’ (from some Bantu language), and so ‘here, thus’. In narratives fadeso is used frequently to carry the story along as ‘and then’, and it can be used as a noun: a-wali ti fadeso [PL-woman of now] ‘women of this age’. There are only very few ideophones (such as ngbii ‘for a long time’), evidence of the drastic reduction of Ngbandi during pidginization, for which see Samarin (1979).
(66) Mo te kasa ni mingi ape.
2SG eat stew DEF much neg
‘Don’t eat much of the stew.’
(67) Mo te ti mo nde o.
2SG eat of 2SG apart CLT
‘Please eat yours separately.’
(68) A-saa kodro ge.
PM-make village here
‘(They) made their home there.’ (saa < sara)
(69) Lo te kasa ni ngbii, a-hunzi kwe faaso.
3SG eat stew DEF long.time PM-finish.off all now
‘She kept on eating the food until it was completely gone.’ (faaso < fadeso)
A nominalized verb that repeats the main verb is used adverbially to show intensity, as a-gi melenge-wali ni gingo [PM-look.for child(of)-woman DEF looking.for] ‘(the girl’s mothers) looked and looked for the girl.’
Many words in Sango, apart from some that are arguably compounds, consist of one and two syllables; words of three and four syllables are generally Bantu in origin. Most of Sango’s very limited vocabulary is Ngbandi, and only Banda of other indigenous ethnic languages has contributed a substantial number of words. There are also words from Bantu languages, even Swahili, which were adopted early in the language’s history and a few more in recent years. French, of course, has contributed the greatest number of foreign words of all kinds, and it has more of an influence on the language now than ever before. Spelling of words has not yet been standardized except in religious publications. Each author has his own conventions, and there are always inconsistencies.