Atlas of Pidgin and Creole Language Structures Online

1. Introduction ⇫ ¶

Juba Arabic is an expanded pidgin spoken in South Sudan, a country which split off from Sudan in 2011 (following the signing of the Comprehensive Peace Agreement in 2005). As it is spoken either as first language (L1) or as second (L2) or third language (L3) by the majority of South Sudanese people, it is quite difficult to estimate the exact number of Juba Arabic speakers. On the one hand, Juba Arabic represents the L1 of the greater part of the urban population of Juba, the capital city of South Sudan.¹ On the other side, Juba Arabic is widely spoken as L2/L3 by mother-tongue speakers of Nilotic languages such as Bari or Dinka. In addition, another factor that makes it hard to evaluate the boundaries of Juba Arabic as a linguistic entity is that this pidgin is involved in a reconstructing continuum with its major lexifier, which is Northern Sudanese Arabic (or Sudanese Standard Arabic). For decades the coexistence with Sudanese Standard Arabic has been affecting Juba Arabic in southern Sudan, although the influence of Arabic is even stronger among the large diaspora communities living in Khartoum (Sudan) and Cairo (Egypt). In this overall situation, it is possible to distinguish between acrolectal/mesolectal and basilectal varieties of Juba Arabic according to the different degrees of structural interference of the Arabic lexifier. Other Arabic-based contact languages are Kinubi (see Luffin 2012 in this volume) and Turku Arabic.

2. Historical background ⇫ ¶

The Arab conquest of Africa began soon after the death of the prophet Muhammad. As far as the Sudan is concerned, the penetration of Arab groups from the Upper Nile started in the 12th century and reached its climax by the end of the 15th century.

     Powerful states, such as the Funj Sultanate and the Muslim Sultanate of Dar Fur, gradually took the place of the ancient Christian Nubian kingdoms, and Arabic became the lingua franca of the area. In 1820 the Khedive of Egypt Muhammad Ali sent a military expedition against the Mamluks who had sheltered in Nubia and eventually conquered a great part of what is now modern Sudan. The way to southern Sudan was opened for long-distance trade: ivory and slave traders started to set up permanent trading camps, called zaribas (literally ‘cattle enclosure’), which spread all over the White Nile Basin, the Bahr al-Ghazal and Equatorial Province. These settlements were inhabited by a heterogeneous population composed of a dominant minority of native speakers of Arabic and a vast majority of autochthonous people engaged as dragomen, slave soldiers and domestic slaves, who were forcibly drawn from various Nilotic tribes (Mundu, Lugbara, Dinka, Shilluk, Bari; see Owens 1985). Soon afterwards the traders’ camps were progressively turned into military camps and those who had worked in the zaribas as traders’ private troops were gradually incorporated into the government army, while local inhabitants started to join the camps deliberately. In this environment a pidginized Arabic arose. After 1870 slave-trade became so large that the Egyptian government had to take anti-slavery measures and began to restrain the zaribas which were effectively dominating the economy and politics of the south. The provinces were assigned to several governors (Baker, Gordon, Gessi, Emin Pasha) in order to bring them under control of the Egyptian government in Khartoum. Despite this, at the end of the 19th century, southern Sudan faced a complete upheaval: the Mahdi revolt of 1882 resulted in the capture of Khartoum and the occupation of several provinces. The loss of power of the Egyptian government led to numerous rebellions among local tribes and desertions in the army, until the Egyptian forces had to withdraw from the south. These events hampered the linguistic influence of Arabic in southern Sudan immediately after the Mahdi revolt. Ultimately, the declaration of independence in 1956 marked the beginning of a new, overt Arabization policy that greatly affected the development of Juba Arabic.   

3. Sociolinguistic situation ⇫ ¶

As already stated, Juba Arabic represents the main lingua franca of South Sudan. Thus, even if it is only in the last thirty years that Juba Arabic has gradually begun to be acquired as a first language, it has long been a vehicular medium widely used in inter-ethnic exchanges. Urban speakers are generally more affected by the interference of Sudanese Arabic; thus, they are to be considered the repositories of the acrolectal/mesolectal varieties. In this regard, it is important to remark that Juba Arabic speakers living in Khartoum are much more exposed to the prestigious variety of Sudanese Arabic. By contrast, the basilectal varieties of Juba Arabic can generally be associated with rural speakers who are more influenced by the Nilotic substratum. In truth, there are no discrete boundaries between acrolectal and basilectal varieties since Juba Arabic speakers are involved in a socio-linguistic continuum which implies a high degree of phono-morphological variation.

            The signing of the Addis Abeba peace agreement in 1972 granted a certain degree of administrative autonomy to South Sudan, and consequently Juba Arabic became an important linguistic means for expressing a new super-tribal South Sudanese identity (see Miller 1991: 159). As a result Juba Arabic went through a significant functional expansion that led it to cover formal communicative settings such as religious functions in missionary churches, judgments in local courts (Miller 2007) and, more recently, radio broadcasting. Finally, it should be observed that even though Juba Arabic does not possess an established orthography, it is widely written in Latin script. This is particularly true for prayer booklets produced by Christian missions that have significantly contributed to the spreading of Juba Arabic in South Sudan. 

4. Phonology ⇫ ¶

The consonant inventory consists of twenty phonemic segments. Juba Arabic presents a five vowel set (a, e, i, o, u). The predominant syllable structure is open, although syllable-internal consonant clusters are also allowed. Stress is phonetically realized as a high pitch and can be distinctive, both lexically and grammatically. In Table 1, a broad allophonic inventory of consonants is given so as to show the secondary segments (in brackets)².

Phonemic variation occurs between s ~ š, p ~ b and z ~ j:

šókol ~ sókol              ‘work’

ašán ~ asán                ‘because’

tálabu ~ tálapu                       ‘to ask’          

zol ~ jol                       ‘individual’   

ázu ~ áju                    ‘to want’       

     [x] is an allophone of /k/ in acrolectal varieties, whereas [h] may stand for the Arabic [x] in basilectal varieties: kámsa (< Arabic xamsa)[ˈxamsa] ~ [ˈhamsa] ‘five’.

     The appearance of [ɣ] as an allophone of /g/, together with other recurrent phenomena of palatalization and aspirations of plosive segments, are to be considered as instances of hypercorrection applied by the speakers under the influence of Arabic (e.g. tagalíd [taɣaˈlid] ‘traditions’).

     The nasals ŋ, ɲ and the implosive ɓ only appear in loanwords from the Nilotic substrate, especially in Bari borrowings (Miller 1989: 3; Bureng Vincent 1986: 72–74):

ɓéko                ‘to find’         

ɲeré                ‘marginalization’

toŋoɲú                        ‘to mix’

     Juba Arabic has five vowel phonemes. Vowel quantity is not distinctive, either lexically or grammatically, although long allophonic variants are found in acrolectal registers; vowel length is also predictable in stressed syllables, and it can be prompted by pragmatic factors such as emphasis.

Furthermore, a phonological process which may affect vowel quantity is vowel assimilation, a tendency which is quite extensive and has an impact especially on verbal morphology. Vowel assimilation processes take place across morpheme boundaries. Accordingly, two consecutive vowels may merge into one long vowel, e.g. le íta [ˈle:ta] ‘to you’, ya ána [ˈja:na] ‘I’, úo ligó ána [ˈuo#liˈga:na] ‘he found me’. Assimilation gives rise to allomorph forms for some of the clitic particles, such as the TAM markers (see §6); the irrealis marker ge, for instance, can be realized as [ˈga] or [ˈgi] depending on the vowel context: úo ge=ágra [ˈuo#ga#ˈagra] ‘he is studying’; ána ge=šílu [ˈana#gi#ˈʃilu] ‘I am carrying’. Vowels optionally merge between prepositions and personal pronouns, e.g. le úo [ˈlɔ] ‘to him’; ma úmon [ˈmɔmɔn] ‘with them’. Verbal morphology is likewise influenced by assimilation processes in so far as passive derivation is concerned. When the 3rd singular pronominal form -u is suffixed to bisyllabic verbs ending with a vowel in order to form the passive voice (see §7.3), it undergoes phonetic assimilation, as in dáfa ‘pay’ > dáfau [ˈdafaɔ] ‘be paid’; dába ‘slaughter’ > dábau [ˈdabaɔ] ‘be slaughtered’; ába ‘refuse’ > ábau [ˈabaɔ] ‘be rejected’.

     Stress is confined to the first syllable if there are no heavy syllables, otherwise the first heavy syllable in the word will be stressed. In a few cases stress is lexically distinctive, e.g. sába ‘seven’ vs. sabá ‘morning’, záman ‘when’ vs. zamán ‘ancient times’. In verbal derivation (see §7.1), stress dislocation distinguishes the active (transitive) voice from the passive (intransitive) for verbs ending in -u#: kátulu ‘kill’ vs. katulú ‘be killed’, jíbu ‘bring’ vs. jibú ‘be brought’. Stress shifted to the penultimate syllable in verbs ending in -u# is associated with deverbal nominal forms (see Miller 1993: 157): kúruju ‘cultivate’ vs. kurúju ‘cultivation’.

5. Noun phrase ⇫ ¶

Juba Arabic noun phrases are essentially head-initial: postmodification and agreement of possessive pronouns, adjectives, determiners and other modifiers will be shown in the following paragraphs.

5.1 Nouns     ⇫ ¶

Juba Arabic distinguishes number, but not gender. This applies to the category of nouns and their modifiers. However, for a few animate nouns which have biological gender a suppletive feminine form exists:

rájil        ‘man’      vs.          mára        ‘woman’

wéled     ‘boy’       vs.          binía        ‘girl’

akú        ‘brother’               vs.            úkut   ‘sister’

Plural noun derivation is mainly obtained by suffixation and stress shift onto the last syllable: the suffix -át is usually affixed both to Arabic and Nilotic nouns, e.g. hayawan-át ‘animals’, nisab-át ‘bows’. Some lexemes code plurality through change of stem vowels (ablaut, or “internal plurals”), others by suppletion. Internal plurals are productive in Arabic and its dialects, but not in Juba Arabic, where they have been retained as frozen forms with a plural meaning:

múškila      ‘problem’       vs.       mašákil         ‘problems’

béled           ‘country’                   vs.                 bilád            ‘countries’

mára          ‘woman’                    vs.                 nuswán       ‘women’

wéled          ‘boy’               vs.       iyál                ‘boys’

The word nas ‘people’ is preposed to nouns to express collective plurals: nas gazál ‘(the group of) the gazelles’, nas ajús ‘(the group of) the elders’.

5.2 Pronouns ⇫ ¶

The pronominal system is composed of two numbers and three persons:

Unlike dialectal Arabic, Juba Arabic does not have pronominal suffixes, and personal pronouns can have the function of subject, object and indirect object:

Possessive pronouns, except for the first and second singular³, result from the lexicalization of analytical constructions formed by the genitival particle ta (see §5.5 for further pronominal forms) and the independent personal pronoun:

5.3 Determiners ⇫ ¶

Juba Arabic has neither a definite nor an indefinite article. Definiteness in a noun phrase is morphologically determined by elements such as personal or relative pronouns; otherwise it can be inferred from the context. Indefiniteness can sometimes be emphasized by preposing the numeral wáhid ‘one’ to a noun.

     Demonstratives distinguish two categories of spatial and temporal deixis: proximity to the speaker and remoteness from the speaker (cf. Table 4). Demonstratives are sometimes used in the sense of the English definite article.

hayawán da    ‘this animal’

iyál dol            ‘these boys’

béled dak         ‘that country’

5.4 Adjectives ⇫ ¶

Juba Arabic adjectives are marked for number in two different ways: suffixation, and in some cases stem ablaut. The plural suffix -ín shifts the word stress toward the last syllable, and it is sporadically used in combination with internal plurals, giving rise to mixed plural forms:

ketir-ín ‘many’ (pl.)

murta-ín ‘happy’ (pl.)

ja(h)iz-ín ‘ready’ (pl.)

kebír ‘big’ (sg.) - kubár ‘big’ (pl.) - kubar-ín ‘big’ (pl.)

The position of both attributive and predicative adjectives is after the noun that they modify:

     Comparative and superlative adjectives are formed by the postposition of šedíd ‘strong’ or kalís ‘much’ to the adjective at issue:

In a few cases, superlative adjectives preserve the Arabic-derived form which Semitists refer to as elative⁴. In Arabic and its dialects this invariable pattern is applied to adjectives to form the superlative and the comparative degree. Juba Arabic has few cases of elative adjectives which are not used systematically: ‘you are the biggest’ can be expressed both by the elative íta ákbar and by the simple form íta kabír.

     A further strategy employed to convey a superlative meaning is iteration:

     Comparison is expressed in two different ways. In the prevailing construction, which derives from Arabic, the standard NP (the noun phrase with which the subject of the clause is compared) is preceded by the locational preposition min ‘from’.

The other option is represented by the ‘exceed’ comparative. This construction, contrary to its recurrent use in Kinubi and Turku, is marginal in Juba Arabic. The standard noun phrase is marked by the grammaticalized item fútu, which derives from the Arabic verb fāta ‘pass, surpass’:

5.5 Possessive phrases ⇫ ¶

Possession is expressed through an analytic construction: The possessor follows the head noun and is preceded by the genitive marker ta (or bitá):

The particle ta derives from the Sudanese Arabic genitival marker bitá (originally meaning ‘merchandise’). The two markers in Juba Arabic are interchangeable, the second one occurring more under Arabic interference. For this reason bitá can occasionally appear in place of ta in possessive pronominal forms: bitái ‘my’, biták ‘your’, bitá(u) ‘his’, bitánna ‘our’.

5.6 Numerals ⇫ ¶

Numerals are based on a decimal system:

Cardinals precede the head they modify: arba yom ‘four days’, sítta bagar-át ‘six cows’. Ordinals are formed by the head noun and the cardinal number bound together by the genitive particle ta (Head ta Numeral): yom ta kámsa ‘the fifth day’, síka ta síta ‘the sixth street’. The first two ordinals (áwal ‘first’, táni ‘second’) are borrowed from Arabic and behave like the cardinals; they are thus preposed to the head noun, e.g. áwal hája ‘the first thing, first of all’, táni hája ‘the second thing’.

6. Verb phrase ⇫ ¶

6.1 Tense, aspect and mood

Juba Arabic shows three overt pre-verbal markers (bi=, ge= and kan) together with zero marking. Bi introduces future and conditional meanings in most cases; hence it is generally considered an irrealis marker (irr). On the other hand, ge typically marks progressive aspect (prog). Kan is an anterior marker (ant) which adds past tense reference to nominal and verbal predicates and can co-occur with bi and ge. The tense values expressed by unmarked verbal forms vary according to the lexical aspect of the verb. Table 5 summarizes simple and complex TAM expressions in Juba Arabic.       

If unmarked action/motion verbs typically express a perfect (cf. 6), stative unmarked verbs express a generic present (cf. 7). 

Unmarked verbs also express imperative mood. In this case, the plural is formed by postposing the second plural pronoun tákum (see §5.2) to the verb:

rówa ‘go!’ (IMP)       rówa tákum ‘go!’ (IMP.2PL)

gáta  ‘cut!’ (IMP)      gáta tákum ‘cut!’ (IMP.2PL)

If ge⁵ has progressive and present habitual meanings, then it is generally considered a non-punctual marker. The occurrence of ge is mostly independent of the tense reference of the VP. Bi⁶ is basically an irrealis marker, although as a consequence of the growing influence of Sudanese Arabic it is gradually acquiring the habitual values which were formerly expressed by ge. In particular, bi started to cover potential habitual actions, while ge reduced its habitual value to actual actions (Tosco 1995: 434). Another important point to remark is that, in contrast to Kinubi (see Luffin 2012 in this volume), bi and ge cannot occur together to mark progressive future. This is because in Sudanese Arabic the imperfective bi= and the progressive gā‘id (see notes 5 and 6) are mutually exclusive. The following examples show some common occurrences of bi and ge.

- ge as progressive (motion verb):

- ge as durative present (action verb):

- ge as factual habitual (action verb):

- bi as generic present (stative verb):

- bi as generic future (motion verb):

- bi as potential habitual (epistemic possibility, see §6.3):

- bi as factual conditional (action verb)

The anterior marker kan⁷ introduces the past perfect when occurring with unmarked action/motion verbs, while it marks an imperfect aspect before unmarked stative verbs. The imperfect of motion and action verbs is expressed by the sequence kan ge. By contrast, the sequence kan bi typically marks a habitual past or counterfactual conditional. Independent subject pronouns can be interposed between the two preverbal sequences. In NPs, kan occurs between the subject and the nominal predicate. Some examples of anterior marking are shown below:      

- kan + motion verb (past perfect):

- kan + ge (imperfect):

- kan + bi (counterfactual):

6.2 Tense-aspect auxiliary verbs

In addition to TAM particles, also auxiliary verbs that express tense and aspect are found in Juba Arabic.

Simple auxiliary verbs such as bíga ‘become’ and gum ‘get up’ generally express an inchoative meaning. Whereas bíga occurs with bi- marked verbs, gum usually occurs before unmarked verbs. The complex auxiliary verbs bíji and birówa, which derive from the Arabic imperfective forms b=i-ji ‘he is coming, he will come’ and b=i-rawwaḥ ‘he is leaving, he will leave’ respectively, introduce a resultative and a near future. The auxiliaries bíji and birówa always precede unmarked verbs. 

- gum (inchoative):

- bíji (resultative):

- birówa (near future):

6.3 Modal verbs

In Juba Arabic epistemic, dynamic and deontic modalities are expressed by the modal verbs:

In Juba Arabic, as well as in Sudanese Arabic, the pseudo-verbs yúmkin, múmkin introduce both deontic and epistemic possibility. Bukún is generally related to future and present presumptive expressions. If bukún is joined with an action/motion unmarked verb, it conveys a deontic possibility in the past. Bágder ‘can’ denotes dynamic possibility, always preceding unmarked verbs. Lázim, for its part, requires the irrealis bi to express deontic necessity, while it precedes unmarked verbs in epistemic contexts:

– bukún ‘could’: deontic possibility in the past:

– bágder ‘can’: dynamic possibility:

– lázim ‘it is necessary that’: epistemic necessity:

6.4 Reduplication

Reduplication of a verb emphasizes the action described by the verb phrase. In certain cases, reduplication indicates specific semantic notions such as repetitivity or habituality of the action. There are no semantic restrictions to reduplication in Juba Arabic and in theory any verb can be reduplicated.

It is not rare to find instances of reduplication applied to sequences of verbs:

6.5 Negation

The basic negative operator in Juba Arabic is ma, which always precedes the verb and the preverbal markers (if any). It follows the anterior marker kan (e.g. 17) and the auxiliary verbs.

In existential NPs the negative operator is máfi. It generally precedes the subject, although in topicalized sentences it might occur in final position:

The prohibitive mood in Juba Arabic is formed by the negative operator ma prefixed to -ta for the singular and to  -tákum for the plural. 

7. Simple sentences ⇫ ¶

7.1 Word order

Juba Arabic is predominantly SVO. The basic word order may be changed by passive verbs (§7.3) or in presence of contrastive topicalization (Y-movement). In the last case, the object might be placed before the verb:

The indirect object follows the direct object, as illustrated in the following example:

The recipient, or beneficiary, of a ditransitive verb is signalled through adpositional marking (the dative preposition le) and it often precedes the direct object, or patient (indirect object construction; but see example (5d)):

7.2 Causative, reflexive and reciprocal

Since Juba Arabic lacks valence-increasing and valence-decreasing devices, it resorts to periphrastic constructions in order to add causative, reflexive and reciprocal meanings to verbs.

     The analytic causative is formed by the verb ámulu ‘make, do’ preceding the sentential complement which refers to the caused event:

Reflexives and reciprocals are not isomorphic. The analytic reflexive náfsa, lit. ‘soul’, followed by a possessive pronoun, expresses objects that are coreferential with the subject:

The reciprocal is expressed by means of the lexicalized form badúm (from Arabic baʕd-hum ‘some of them’) which acts as a reciprocal pronoun:

7.3 Passive constructions

Juba Arabic commonly displays prototypical passive constructions, as in (41b):

The example is in line with all the prototypical hallmarks of passives because: (a) the passive construction represents a transitive clause containing an optional agent, (b) the patient (i.e. the object of the active clause) occupies the slot of the subject and represents the only topical argument of the construction, (c) the agent (i.e. the subject of the active clause) is demoted to an oblique case introduced by the comitative adposition ma ‘with’,⁸ (d) the verb has a specific phono-morphological marking for the passive voice: the displacing of the demarcative stress of the active form from the first to the last syllable of the morphological domain. The occurrence of stress displacement for passive marking depends on the morphology of the active verbs. Consequently, prototypical passive constructions are subjected to some restrictions.

            Stress displacement typically correlates with disyllabic and trisyllabic transitive verbs presenting a final vowel -u:

Other -u verbs:

∫ílu ‘carry’      ∫ilú ‘be carried’

ʤíbu ‘bring’  ʤibú ‘be brought’

kútu ‘put’       kutú ‘be put’

kátulu ‘kill’    katulú ‘be killed’

kátibu ‘write’ katibú ‘be written’

góbudu ‘catch’           gobudú ‘be caught’

Besides, there are a number of transitive verbs with a final -a. These verbal items derive from Arabic perfective and imperfective stems presenting a final pharyngeal consonant that has been elided in Juba Arabic (e.g. dabaħ > dába ‘slaughter’, asmaʕ > ásma ‘listen’). Contrary to -u marked verbs, this class of transitive verbs shows a considerable degree of variation in the formation of the passive voice. At the present time we can isolate the following three morpho-phonological variants of passive marking for -a verbs.

These variants represent three stages of the grammaticalization of passive marking in Juba Arabic. Stage (a) shows a reflexive 333vvvdrd singular independent pronoun in object position. In (b) the independent pronoun is cliticized and triggers assimilation but does not induce stress displacement. Lastly in (c), the cliticization of the earlier 3rd person independent pronoun results in a final stressed syllable. Thus, we can argue that stress displacement for the derivation of passive voice in Juba Arabic finds its origin in reflexive (non-causative) constructions in which the patient is expressed by a third singular independent pronoun in object position. Actually, in many languages, reflexive (non-causative) constructions constitute the basic source for the grammaticalization of passive morphology (Haspelmath 1990).

     Lastly, Juba Arabic also possesses a few ambitransitive verbs, which always show a final stressed syllable. Ambitransitive verbs in Juba Arabic have been lexicalized on the basis of Arabic verbal stems followed by a 3rd singular bound pronoun realized as a final stress (e.g. add-í ‘give it!’ > wedí ‘give, be given’, warr-í ‘show it’ > werí ‘show, be shown’, lig-ó ‘he found it’ > ligó ‘find, be found’). Lacking overt morphological marking, the following passive construction is non-prototypical. 

8. Interrogative and focus constructions ⇫ ¶

Polar interrogative clauses can be distinguished from declarative clauses only by intonation. Q-words in content questions always occur in situ:

Le ‘why’ is always sentence-initial:

Šenú ‘what’, mitén ‘when’, yátu ‘which’ and wenú ‘where’ generally take sentence-final position:

The focused element in a clause is indicated by the focus marker yáwu following it:

9. Complex sentences ⇫ ¶

9.1. Relative clauses

In Juba Arabic the relative pronoun has the invariable form al ‘who, that, which’. Furthermore, the lexeme abú (from Arabic abū ‘father’) can optionally introduce subject and direct object relative clauses modifying an animate head noun. Subject and direct object relative clauses generally do not contain resumptive pronouns. 

By contrast, instrumental relative clauses need a resumptive pronoun which is variably introduced either by the comitative ma ‘with’, or the instrumental be ‘by’. Because of Arabic influence, in acrolectal varieties only be occurs.

Headless relative clauses are also common in Juba Arabic:

Lastly, it should be remarked that in basilectal registers the relative pronoun is often missing:

9.2. Complementizer with verbs of speaking and thinking

In Juba Arabic subordinate complement clauses usually follow the primary verb directly (uo fékir úo zol batál ‘He thinks he is bad’). However, the verb gal(e) ‘say’ may function as complementizer with verbs of speaking (in indirect speech) and thinking (cf. Miller 2001).

In the majority of cases, nothing can intervene between the verb and the complementizer. In acrolectal varieties, though, an indirect object may optionally separate the verb of speaking and the complementizer gál(e):