2.1 Portuguese-Lexified Creoles in Africa

The Portuguese-lexified creoles spoken in Upper Guinea (Cape Verdean creole from Santiago, Brava, São Vicente; Guinea-Bissau Creole; and Casamance Creole in Ziguinchor; see map 1) have nouns and adjectives unmarked for gender for the most part, but vestiges are still found. For instance, for CVC from Brava, Baptista (2013, p. 14) notes that some animate nouns are marked for gender (e.g., galinha-galu ‘hen-rooster’ [< Pt. galinha, galo], tia-tiu ‘aunt-uncle’ [< Pt. tia, tio]). However, grammatical gender marking is rarely if ever found on inanimate nouns, lost in the creolization process. The creoles exhibit variable overt plural marking with the suffixal plural marker -s, inherited from Portuguese. Overt plural marking in these creoles depends on semantic factors (e.g., animacy of the referent of the noun), syntactic factors (e.g., the presence of a numeral or a quantifier in the noun phrase), and pragmatic factors (e.g., definiteness). The Portuguese-lexified creoles spoken in the Gulf of Guinea (Angolar and Santome spoken on São Tomé, Principense spoken on Príncipe, and Fa D’Ambô spoken on Annobón; see map 1) have not retained Portuguese -s.² Instead, each has developed an overt plural-marking strategy, using the third person plural pronoun, equivalent to ‘them’ (Angolar ane/ene, Santome inen, Principense ine, and Fa d’Ambô nen), whose use is sensitive, as well, to syntactic (e.g., to co-occurrence with other elements) and pragmatic constraints (e.g., definiteness) (Hagemeijer, 2013; Maurer, 2013a, 2013b; Post, 2013).

With regard to verb-related features, two key changes have altered the way verbs are conjugated between Portuguese and the Portuguese-based creoles in Africa: In the creoles, there are no person/number-marking verbal suffixes retained from Portuguese, nor is the verb class distinction retained, which in Portuguese is marked by thematic vowels (TVs) and corresponding morphology (e.g., the preterit allomorphs -ou and -u in 3sg, as in cant-Ø-ou [sing-tv-3sg.pret] ‘s/he sang’, beb-e-u [drink-tv-3sg.pret] ‘s/he drank’, sub-i-u [go.up-tv-3sg.pret] ‘s/he went up’). This notwithstanding, these creoles did retain some suffixal morphology from Portuguese, shown in Table 1.

Open in new tab

Map 1. Africa.

Reprinted with permission from Clements (2022).

The creoles developed forms and form combinations, both pre- and postposed, to mark different tenses, aspects, and moods. Generally, bare stative verbs in these creoles have present-tense interpretation, as shown in (1), from Lang (2013, p. 8), while bare dynamic verbs receive past-tense interpretation, as with fika in (2), from Baptista (2013, p. 15).

(1)

(2)

In (2), preposed ta is used to mark imperfectivity; in Cpv.-B ta also marks the progressive, and the habitual/iterative in the present and, in combination with -ba, habitual in the past. The examples in (3), from Baptista (2013, p. 15), show the uses of the postposed anterior marker -ba, which yields a past interpretation when appearing with a stative verb ((3a)) and a pluperfect interpretation with a dynamic verb ((3b)), from Baptista (2013, p. 15).³

(3)

A construction common to all the Portuguese-based African creoles in question is the double-object construction, shown in (4), from Maurer (2013b, p. 77), characterized by the order of the indirect object immediately after the verb, followed by the direct object, and no overt marking on either the direct or indirect object.

(4)

Serial verb constructions are commonly found in the Gulf of Guinea creoles. An example is shown in (5), from Hagemeijer (2013, p. 54). In this type of construction, one verb (da ‘give’) takes on a functional interpretation. That is, in (5), da ‘give’ is functionally a benefactive marker, with the interpretation ‘for’.

(5)

Turning now to the pronominal systems, Table 2 displays the subject, object, and disjunctive pronouns in the creoles being discussed in this section. Of note is that no creole expresses the notion of ‘I’ with a form traceable back to the Portuguese 1sg eu ‘I’. Rather, ‘I’ in these creoles is expressed by forms coming from the disjunctive, stressed 1sg pronoun mim ‘me’. Also of note is that some creoles maintain the 2sg informal–formal (T/V) distinction⁴, and one language, Santome, also displays the T/V distinction in 3sg. The Santome forms sun (< Portuguese senhor ‘sir’) and san (< Portuguese senhora ‘madam’) can be used to formally address an interlocutor directly (2sg formal), as well as referentially, to formally talk about someone (3sg formal, as in ‘that gentleman/lady there’). Moreover, the Atlantic Portuguese-lexified creoles make at least one subject-object case distinction in their respective pronominal systems, and some make the distinction in most persons and numbers (e.g., Cape Verdean varieties). Finally, the 3pl forms are often used as plural markers in the Gulf of Guinea creoles.

By way of wrapping up this section, we have discussed the noun and verb forms as well as pronominal systems of the Portuguese-lexified Atlantic creoles. We have suggested that the forms and structures discussed can be accounted for, at least in part, by appealing to the processes underlying second-language acquisition according to which highly detectable, frequently used content words are learned first and content words are learned before function words. Among the function words, such as pronouns, we would expect that the forms incorporated into varieties that became creoles would initially have a sound form that would make them more detectable in discourse. The pronominal forms shown in Table 2 all contain at least one consonant and often consist of a consonant-vowel structure, although over time these have reduced in some cases. For example, the 3PL form es ‘they’ in CVC varieties has a reduced form -s. As already pointed out, it is noteworthy that the Portuguese unstressed object pronouns -o ‘him’ and -a ‘her’ are not retained in any of the creoles in question. This follows from the observation that in creole formation, the co-creators of the new varieties incorporated forms they could easily detect and parse in communicative interactions. Unstressed pronominal forms consisting of only one vowel would not be easily detectable, and thus not easily parsable, in discourse.

Returning to the question briefly noted about the possible origins of the Portuguese pidgin that could have served as a point of departure for the Portuguese-lexified creoles in Africa, it may have formed in Portugal before their colonial expansion, as argued by Kihm and Rougé (2013, 2016), in Santiago, or it may have originated in the interaction between the Portuguese and the Africans in Africa (Clements, 1992, 1993a, 1993b; Jacobs, 2010). It could also be that both hypotheses are correct. What does seem to be the case, however, is that these creoles developed in situations where naturalistic language acquisition occurred among speakers of multiple languages, as described earlier.

2.2 Portuguese-Lexified Creoles in Asia

The colonization of parts of India, Sri Lanka, Malaysia, Indonesia, and Macau by the Portuguese took place in the fourth stage of its colonial expansion (see map 2). In 1498, the Portuguese explorer Vasco da Gama, crossed the Arabian Sea to the tip of the Indian subcontinent, reaching Calicut in the summer of 1498. The Portuguese fortified trading posts were established with relative rapidity in Asia, and by 1550, the Portuguese had settled a large part of the Indian west coast (Calicut, Goa, Daman, and Diu, among others), as well as Sri Lanka (e.g., Batticaloa). The city of Daman—which put up the most resistance—fell in 1559, and by 1581 it was recognized as a Portuguese city. The Portuguese also moved quickly to the east, founding settlements in Malaysia (Malacca), throughout Indonesia, and in China (Macau) well before the end of the 16th century.

Before surveying the salient features of the Portuguese-lexified creoles in Asia, a brief discussion of the connections between these creoles is useful. Clements (2000) hypothesized that there was an Asian Portuguese pidgin that served as the basis for all the Asian Portuguese-based creoles. With the exception of Diu Indo-Portuguese (which exhibits many acrolectal features), all Asian creoles in question share a reflex of the 16th-century phrase que laia ‘what way/manner’ for the question word ‘how’: Daman kilay (also kom), Korlai kilɛ, Sri Lankan Creole kilaay, Papiá Kristang klai, Makista klai. Clements (2000) postulated that such a variety of Portuguese could have spread from Calicut in the south of India north to the northern provinces (Goa, Chaul-Korlai, Mumbai area, Daman, Diu), and at the same time east to Sri Lanka, Malaysia, Indonesia, and China. There is lexical evidence of southern Indian influence in the northern varieties of Indo-Portuguese. In Diu, Daman, and Korlai Indo-Portuguese varieties, for instance, the lexical item ap (< Malayalam appaṃ ‘(rice) flatbread’) is present (Cardoso, 2009; Clements, 1996; Clements & Koontz-Garboden, 2002), and the Korlai lexical item khədya ‘tiger’ is most likely derived from Malayalam kaʈuʋa.⁵ Baxter (2013, p. 122) also mentioned that Malayalam lexical features, as well as grammatical features shared with Indo-Portuguese creoles, are found in Papiá Kristang. Moreover, there is also evidence that Makista in Macau is at least partially derived from Papiá Kristang, spoken in Malacca (Baxter, 2009; Pinharanda, 2010, and references therein). Thus, it seems that whatever variety of Portuguese (pidgin/creole) was used or formed in the areas of South India, it traveled with speakers eastward.

In this section on the Portuguese-lexified creoles spoken in Asia, the same focus is adopted: on nominal and verbal forms and on the pronominal systems that form part of the morphosyntax of the languages surveyed, all of which are still spoken, though Makista has ever fewer native speakers (Pinharanda, 2010), some of whom live in diasporic communities.

Open in new tab

Map 2. Asia.

Reprinted with permission from Clements (2022).

As one might expect, nouns and adjectives in the Portuguese-lexified creoles in South Asia do not retain suffixal marking of grammatical gender, with the exception of some animate word pairs (e.g., Korlai irmã-irmãw ‘sister-brother’ [< Pt. irmã-irmão]). Vestiges of plural -s are found in written Sri Lanka Creole Portuguese, in a restricted number of human-referent nouns in Kristang (Baxter, 2013) and in acrolectal Makista.⁶ Otherwise, it is not found in these creoles. In Diu and Daman Indo-Portuguese, the universal quantifier tud (< Pt. tudo ‘all’), although still used in the sense ‘all’, has also undergone grammaticalization and is often used to mark the plural (Cardoso, 2013; Clements & Koontz-Garboden, 2002). Papiá Kristang uses reduplication to mark plural (e.g., kaza kaza [lit. house house] ‘houses’ if there is not a numeral or a quantifier. The same applies to non-acrolectal Makista. Regarding pronouns, DiuDaman, and Sri Lanka Indo-Portuguese distinguish gender in the 3sg: el ‘he’ vs ɛl ‘she’ in Diu and Daman IPg. (Cardoso, 2013, p. 93; Clements & Koontz-Garboden, 2002; Maurer & Bollée, 2016) and eli ‘he’ vs ɛla ‘she’ in Sri Lanka IPg. (Smith, 2013, p. 114). This last one also contrasts gender in the plural (elis 3m.pl vs ɛlas 3f.pl).⁷

With regard to verb form-related features, there is one key change at the paradigmatic level between Portuguese and the Asian Portuguese creoles: these creoles have no person/number-marking verbal suffixes retained from Portuguese. However, in Diu, Daman, and Korlai IPt., the distinction among verb classes is maintained, marked by TVs and corresponding morphology (e.g., the preterit allomorphs ‑ou and ‑u in 3sg, as in cant-Ø-ou [sing-tv-3sg.pret] ‘s/he sang’, beb-e-u [drink-tv-3sg.pret] ‘s/he drank’, sub-i-u [go.up-tv-3sg.pret] ‘s/he went up’) (Clements & Luís, 2014; Luís, 2011).

The creoles also retain vestiges of the gerund suffix -ndo and the past participle suffix -do. Table 3 shows this. Of course, all these creoles have developed forms and/or form combinations to mark different tenses, aspects, and moods. Like African creoles, some bare stative verbs in Asian creoles have a default present-tense interpretation, as illustrated by the Korlai examples in (6a–b), from Clements (2007, p. 154). Note, however, that bare verbs in Papiá Kristang can have a present-tense or past-tense interpretation, as shown in (6c), from Baxter (2013, p. 125).

(6)

However, unlike the African creoles, in the Asian creoles bare dynamic verbs do not have a default indicative past-tense interpretation but rather an imperative interpretation or, in the case of Papiá Kristang and Makista, no categorical tense interpretation at all (see discussion to follow). In (7a), the Korlai bare verb kata is imperative, and the corresponding indicative present-tense sentence is either marked, obligatorily, for ongoingness ((7b)) or for habituality ((7c)).

(7)

In Papiá Kristang and Makista, the bare verb form of a dynamic verb can have present-tense or past-tense interpretation, dependent on contextual factors. Baxter (2013, pp. 124–125) gave an illustrative example of this, shown in (8a), in which the bare verb form can have a habitual reading in the present or the past. The aspect markers (imperfective ta and perfective ja) are used to mark a progressive or iterative reading in the case of ta ((8b,c)) or a perfective reading in the case of ja ((8d)).

(8)

For Makista, Ferreira (1996, p. 242) cited examples like those in (9), for which I have provided a morpheme-by-morpheme gloss and a translation of his Portuguese renderings into English. (9a), with a bare form of the verb, has a generic (habitual) present-tense reading, whereas the sentence in (9b) has a progressive present-tense reading.

(9)

Pinharanda (2010, p. 70) noted that present- and past-tense expression in Makista are unmarked more often than overtly marked, and that ta and ja behave more like aspectual markers than tense markers, as is the case with Papiá Kristang.

In the South Asian Portuguese creoles, past tense is overtly and obligatorily marked. Thus, northern Indo-Portuguese creoles mark simple past with the suffix -o for the verb class -a and -w for the verb classes -e and -i. Examples of the -a type are given in (10), from Diu (Cardoso, 2013, p. 95), Daman (Clements’s field notes), and Korlai (constructed example). The verb class morphology is retained directly from Portuguese and is not sensitive to any morphophonological processes. That is, in the formation and development of these creoles, Portuguese verb classes (-a, -e, -i) are maintained.

(10)

Noteworthy is that in the Portuguese-based Asian creoles being discussed here, there is no double-object construction as found in the African creoles. If a verb has two objects, the indirect object is overtly marked, as illustrated by the Papiá Kristang example in (11), taken from Baxter (2013, p. 126).

(11)

Finally, in these creoles, serial verb constructions are found but are usually constructions also found in the substrate/adstrate languages (see Baxter, 2013, p. 126; Clements, 2007, pp. 165–166).

Turning now to the pronominal systems, Table 4 displays the subject, object, and disjunctive pronouns in the creoles being discussed in this section. Of note is that Daman and Korlai IPt. maintain a T/V distinction in 2SG, in contrast with Diu IPt., which has selected the formal address forms, discarding those of informal address. Diu IPt. represents an acrolectal variety, also spoken by many in Small Daman which also has only the formal address forms. In the more basilectal variety of Daman IPt., by contrast, the T/V distinction is still maintained, as it is in Korlai (Clements & Koontz-Garboden, 2002). In Sri Lanka Creole Portuguese, there is no T/V distinction in the 2SG but rather in the 3SG, as is the case in the adstrate languages of the areas. In all South Asian Portuguese creoles, subject/object case marking is found in all persons and numbers. In Papiá Kristang and Makista, by contrast, the pronouns do not show a case distinction.