ChapterPDF Available

Article agglutination and the African contribution to the Portuguese-based Creoles



Content may be subject to copyright.
Article agglutination and the African contribution
to the Portuguese-based Creoles
John Ladhams
1. Introduction: article agglutination1
Article agglutination is an etymological feature of various creole languages whereby an
element which occurs very frequently immediately before a particular noun in the lexical
source language – most often a definite or partitive article – is incorporated into, or
agglutinated to, the noun, and this combination is re-analysed as a single morpheme (a
noun) in the emergent creole. For example: French le rat ‘the rat’, les rats ‘the rats’ >
Mauritian Creole lera, ‘rat(s)’ (en lera ’a rat’, kat lera ‘four rats’); French du thé (of) tea’ >
Mauritian Creole dite ‘(brand/measure of) tea’ (dite-la ‘the tea’, en dite ‘one (brand, cup, etc.
of) tea’, trwa dite ‘three teas’) etc.
Although article agglutination occurs in at least 11 non-French creoles (Parkvall 2000:
81), it is a notable feature of all French lexicon creoles except Réunionnais, and is
particularly common in Mauritian Creole and its derivatives.. The first major study of this
phenomenon, Baker (1984), concentrated on the Indian Ocean Creoles and, for
comparison, Haitian French Creole, and he noted well over 400 examples from Mauritius
and the Seychelles, more than 300 from Rodrigues, just 12 from Réunion, and over 100
from Haiti. The explicit aim of his paper was to try to explain the very unequal
distribution of the feature amongst the French-based Creoles (1984: 89), which he
ultimately attributed to the influence of Bantu languages in the formation of these creoles,
in that the definite and partitive articles in French were perceived to be noun-class
markers by speakers of Bantu languages, and were thus agglutinated into the noun. The
exceptions were clearly Réunion, and to some extent Haiti, but this could be explained by
the fact that in Réunion, very few Bantu-speaking slaves were imported, at least initially,
while in Haiti, the Bantu influence was attenuated by speakers of other African languages.
More than a decade later, Grant (1995) published a study of article agglutination in the
French-based Creoles, largely based on Baker’s data at that time (including additional
material collected for Baker 1988). But Grant also added data from pidgins with a lexical
input from French: Bislama in the Pacific, Sango in Africa, and Chinook Jargon and Michif
in North America.2 Including examples from the latter languages, he was able to report
that his database of article agglutination in the French-based Creoles numbered “some
2500 separate agglutinated nominals” (1995: 155). By and large, Grant accepted Baker’s
1 I am very grateful to Philippe Maurer for sharing his data on Príncipe Creole (since published in Maurer
2009), without which this article would be much the poorer. My thanks also go the the editors of this
volume for their helpful comments. Any remaining errors or shortcomings are solely my responsibility.
2 He could also have added Turku Creole Arabic, spoken in Chad, where Muraz (1931) gives a dozen or so
lexical items with agglutinated articles derived from French.
suggestion of the influence of Bantu languages, but failed to explain why such languages
as Tayo or the North American Pidgins, where there was almost certainly no African
influence, let alone Bantu, should have article agglutination at all.3
A few years later, Parkvall (2002: 81-83) addressed the question of why there were so
many more examples of article agglutination in French-based Creoles than in other
pidgins and creoles, for example English-based ones: “This might be explained by the
tendency of English, Portuguese, Dutch and Spanish nouns to occur without a definite
article more often than their counterparts in French, due to the frequent marking of
partitive articles in French. In addition to this, French word-stress being considerably
weaker (…), this may have made it more difficult to adequately perceive the boundary
between noun and article” (2002: 81). Like Grant, Parkvall agreed with Baker about Bantu
influence in the Indian Ocean Creoles, but he pointed out “that does not hold for the
Atlantic languages” (2002: 82).
McWhorter & Parkvall (2002: 193-94) also address the question of article agglutination
in the French-based Creoles, in an article which amounts to a strong critique of
Chaudenson’s ‘superstratist’ approach to creolisation. They do not refer to the possible
influence of Bantu languages in this feature, but do come to the “inevitable conclusion”
that these and other features are the result of “interrupted transmission” (2003: 194).
In response to all four of the above-mentioned studies of article agglutination in
French-based Creoles, Chaudenson (2003: 262-69) launches a fierce attack on the whole
idea of Bantu influence, particularly as outlined by Baker (1984), but also includes Grant
(1995), since “practically three-quarters of the cases [of article agglutination] derive from
[Baker’s works]” (Chaudenson 2003: 266). Chaudenson sums up his case by saying that
“L’hypothèse du substrat bantou, si évidente pour P Baker il y a quelques années encore,
est totalement indéfendable en raison du mode de selection des données, à la lumière de la
comparaison avec le rodriguais et du fait de l’étude de la bantouisation des emprunts au
français. Il y a plusieurs arguments dont chacun est suffisant à soi seul” (2003: 269).4
This is not the place to discuss his arguments in detail, but suffice it to say that this writer,
at least, finds them unconvincing.5 Indeed, he himself could be accused of selecting some
data and ignoring others, particularly in relation to the work of Grant and Parkvall.
Furthermore, Chaudenson seizes upon the fact that McWhorter & Parkvall (2002) do not
refer to Bantu influence: “on se contente de parler désormais de ‘transmission
interrompue’ (2002: 194); je préférerai personnellement ‘appropriation approximative’
(…)” (Chaudenson 2003: 269). However, this is not necessarily the same thing.
Chaudenson aside, this brief overview of article agglutination in the French-based
Creoles does appear to show a link between the frequency of this feature in a particular
Creole and African, probably Bantu, influence. The remainder of this article now turns to
article agglutination in the Portuguese-based Creoles, and once again the possible link
with African influence.
3 However, see Speedy (2007) for an account of possible influence from Réunionnais on Tayo Creole.
4 It is not clear whether Chaudenson’s “plusieurs” is referring to two or three arguments. I am assuming
that there are three.
5 For example, of Grant’s computerised data base, Chaudenson states: “À l’ère de la Play Station 3 et des
ordinateurs intégrés au téléphone mobile, l’informatique n’impressionne plus autant qu’autrefois. Si
l’on met des données douteuses dans un ordinateur, elles n’en sont pas validées pour autant.” (One
nevertheless assumes that he would not have wanted Grant to return to the use of handwritten index
2. Article agglutination in Portuguese-based Creoles
Though not nearly as widespread as in the French-based Creoles, article agglutination is a
feature of a number of Portuguese-based Creoles, albeit confined to one particular group,
those in the Gulf of Guinea. Elsewhere, the feature is marginal: for Capeverdean Creole,
Lopes da Silva (1957:129) gives the following 7 examples, 5 with common nouns, and 2
with toponyms:
azágua < as águas ‘rainy season’
azilya < as ilhas ‘Cape Verde Islands’
pazdreita < para as direitas ‘to the right (-hand side)’
spiʃine < os peixinhos ‘Hell’ [lit. ‘the little fishes’]
stoʃa < as tochas ‘Hell’ [lit. ‘the torches’]
skorde < os cardos (toponym) [lit. ‘the thistles’]
spedra < as pedras (toponym) [lit. ‘the stones’]
[Note that all seven examples are derived from the plural definite article in Portuguese.]
There are no documented occurrences of article agglutination in the related Guinea-Bissau
Creole. In all the Asian Portuguese-based Creoles,6 one word might seem to be an instance
of article agglutination: anote, ‘night’, but it is in fact derived from Portuguese à noite, ‘at
night’, i.e. a preposition.
However, it is in the four Gulf of Guinea Creoles – São Tomé (ST), Príncipe (Pr),
Annobon (Ann) Creoles, and the Maroon Creole, Angolar (Ang), on São Tomé Island
that a relatively high number of examples of article agglutination occurs (120), as shown in
Appendix 1. The following points may be made about this list:
The number of examples of agglutinated articles in Pr is considerably greater than
in the other three (88, as opposed to 15 in ST, 14 in Ann and only 3 in Ang). Why
that should be the case will be discussed in Section 3 below.
While most of the examples are derived from the singular definite article in
Portuguese – o [u], masculine, and a [ɐ], feminine -, there are two examples of
words derived from the feminine plural definite article (as): zálima and zonda, in
both ST and Pr (cf the examples from Capeverdean Creole cited above).
There are also two examples of words derived from the Portuguese indefinite
article (um, uma): umpan and umuña, both in Pr; there is no partitive article in
Portuguese, unlike in French.
There are 9 examples of derivation from Portuguese definite articles of the ‘wrong’
gender, all but one in Pr, replacing the feminine article a with the initial vowel u, as
indicated in Appendix 1 by an asterisk. Indeed, the vast majority of the examples in
Pr have initial u, and only 4 with initial a.
Again in Pr, there are 6 examples of words derived from noun phrases in
Portuguese taking on different grammatical categories: verb, adjective, preposition
(udêntu, udôdô, ugagu, ukuru, ukurú, unú).
As for the semantic domains for all these Gulf of Guinea examples, they are mostly
such basic vocabulary as parts of the body, or domestic items such as tools, food,
plants, animals, etc. Also, they largely observe what Baker (1984: 111) calls the
6 Norteiro, Korlai and Sri Lanka varieties of Indo-Portuguese, Papia Kristang (Malacca), Tugu Creole (Java),
and Macau Creole. See, respectively, Dalgado (1906), Clements (1995), Dalgado (1900), Baxter & De
Silva (2004), Schuchardt (1890), and Batalha (1988).
‘frequency of collocation’ principle, whereby article agglutination takes place
generally when there is only one of the particular noun (sun, moon, nose, mouth,
In Pr, there is one example of agglutination in a strictly 20th-century word: ukaru
< o carro, ‘car’. This shows that article agglutination, in Pr at least, is an ongoing,
gradual process.
3. The African contribution to the Gulf of Guinea Creoles.
Having shown that article agglutination in the Portuguese-based Creoles is largely
confined to the four Gulf of Guinea Creoles, and the examples in Pr greatly outnumber
those in the other three, it remains to establish why this is the case, and, particularly, to
what extent it is the result of African (“substrate”) influence. For this it is necessary to give
a detailed account of the patterns of early settlement in the islands, with emphasis on the
precise origins and timing of slave imports from the African mainland.
São Tomé was the first of the three islands to be discovered, probably in 1471. In 1485,
the first settlement grant was issued, closely followed by a Royal Letter of Privilege for the
settlers, granting all the right to have slaves. By around 1507, Valentim Fernandes’
description of São Tomé states that there were some 2,000 slaves on the island, compared
with 1,000 European settlers (1940: 122).
Príncipe was probably discovered around 1472, but it is known that settlement of the
island only began in 1500, under the Captain, António Carneiro, appointed in the same
year. The Royal Letter of Privilege, also issued in 1500, is similar to that drawn up for São
Tomé in 1485. The earliest population figures for Príncipe date only from 1607, when
there were 10 Europeans and 500 slaves, as well as 18 “married creoles” and 20 “married
free blacks”.
Annobon, a much smaller and more remote island than the other two, was discovered
possibly around 1500. According to Valentim Fernandes, settlement began in 1503, and by
1507 there were just 9 white inhabitants (1940: 129-30). Because of the island’s size and
difficult access, there was probably only one shipment of slaves around the time of the
initial settlement, almost certainly distributed from the main island of São Tomé.
However, there is no clear documentary evidence of this.
The Angolar maroon community, on the island of São Tomé, according to oral
tradition, is understood to consist partly of the descendants of survivors from the
shipwreck of a slave ship from Angola on the southern coast of the island, around the
middle of the 16th century, and partly of runaway slaves from the plantations on the
northern part of the island. Certainly throughout the 16th century there are many reports
of slaves escaping into the dense forest in the interior of the island.7
The pattern of slave trading by the Portuguese for the Gulf of Guinea Islands is
reasonably well documented, and it is clear that the areas of supply for the island of São
Tomé, from where slaves were redistributed to the other two islands of Príncipe and
Annobon8, were in fact only two in number: the Niger Delta, and the coastal region
Southwards from the estuary of the Congo River. The actual trading places were as
7 For a more detailed account of the settlement of the 3 islands and the formation of the Angolar community,
see Ladhams (2003).
8 There was one notable exception, in the early 16th century, as described below.
The so-called Slave Rivers, on the Western side of the Niger Delta, were known to the
Portuguese from the 1470s. According to Duarte Pacheco Pereira (Esmeraldo de Situ Orbis,
Book 2, Chapter 8; Pereira 1988: 148-53), they consisted of the ‘Rio Primeiro’, ‘Rio Fermoso’
(nowadays the Benin River), ‘Rio dos Escravos’, Rio dos Forcados’ and ‘Rio dos Ramos’.
On the second of these rivers was the port of Ughoton, which gave access to the
‘Kingdom’ of Benin:
E indo pelo segundo braço acima, espaço de doze léguas, é achada ũa vila que se chama
Hugató, que será lugar de dous mil vizinhos: e este é o porto da grande cidade do Beni
que está no sertão nove léguas de bom caminho. (...) O reino do Beni será de oitenta
léguas de comprido e quarenta de largo. E o mais do tempo faz guerra aos vizinhos, onde
toma muitos cativos que nós compramos a doze e quinze manilhas de latão ou de cobre,
que eles mais estimam. (Esmeraldo de Situ Orbis, Book II, Chapter 7; Pereira 1988:149).
[And going up the second branch of the river, for a distance of twelve leagues, there is a
town called Ughoton, which would have around two thousand inhabitants: and this is the
port of the great city of Benin which is nine leagues inland by a good road. (...) The
kingdom of Benin must be eighty leagues long and forty wide. And most of the time they
make war on their neighbours, where they take many captives which we buy at the rate of
twelve and fifteen manilhas of brass or copper, which they appreciate most.]
Pereira adds that these slaves were then (early 16th century) transported to El Mina, where
they were traded for gold. It is known that the Portuguese set up a trading factory
(feitoria) at Ughoton in 1487, which survived, somewhat precariously it seems, for around
twenty years (see Ryder 1969: 32-33). Meanwhile, the first settlement grant for the island
of São Tomé, in 1485, had given the colonisers the right to acquire slaves in the five “Slave
Rivers”. However, subsequent grants for São Tomé in 1493, and for both São Tomé and
Príncipe in 1500, no longer referred to the Slave Rivers as being within the permitted
trading area.
Following the closure of the factory at Ughoton, in 1506 or 1507, the trading rights
with Benin were leased out by the Portuguese Crown, and in 1514 a four-year exclusive
lease was granted to António Carneiro, the donatário (lord) of Príncipe, which meant that
between 1514 and 1518, slaves were being imported from Benin directly to the island of
Príncipe (Ryder 1969 :44). This was clearly to have linguistic consequences in terms of the
proportion of African influence on Príncipe Creole, as indicated below.
However, in 1516 the Oba - or ruler - of Benin laid down that male and female slaves
were to be traded separately, and within a few years, only female slaves could be sold to
the Portuguese, or to any other traders (Ryder 1969: 45). The surviving ships’ books, from
the 1520s, show how trade continued at Ughoton, but on a smaller scale than before, and
for female slaves only.9 Therefore, the Portuguese were at this time seeking alternative
markets, which they found elsewhere in the Niger Delta, as well as in the Congo region.
By the middle of the century, trading by the Portuguese at Ughoton had ceased, as
formalised by an official ban issued from São Tomé in 1553 (Ryder 1969: 74).
Forcados River
In the fourth of the five Slave Rivers, slave trading was carried out by the Portuguese at
least from the beginning of the 16th century, at a point five leagues from the river mouth,
according to Duarte Pacheco Pereira (1988: 152). Following the limits placed on trade at
Ughoton after 1516, trade at the Forcados River would appear to have increased, as
9 See for example, the ship’s book for the ‘São Miguel’ (1522), translated into English in Ryder (1969: 295-
attested for example by the ship’s book for the 1522 voyage of the “Santa Maria da
Conceição” to the Forcados River, and to Ughoton, where the slaves acquired were all
female (Ryder 1969). It is unclear, however, how long trading continued at the Forcados
River: it is unlikely to have gone on beyond the 1553 ban on trading with Benin.
Rio Real
This river, nowadays called the River Bonny, at the Eastern end of the Niger Delta, is
mentioned by Duarte Pacheco Pereira as being a location for slave trading in the first
decade of the 16th century (1988: 156). In the settlement grants for São Tomé in 1493 and
1500, and for Príncipe in 1500, mention is made of the Rio Real, as being within the limits
of the area (as far as the Congo) where trading was permitted. Once again, just how long
trading here continued is uncertain.
This port on the South side of the mouth of the Congo River was first reached by the
Portuguese with Diogo Cão, probably in 1483 (see e.g. Birmingham 1966: 21-26). Trade
with the Kongo Kingdom rapidly took off, particularly for the supply of slaves to the Gulf
of Guinea Islands. In the 1493 and 1500 settlement grants for São Tomé and Príncipe, it
was made quite clear that residents could freely trade in slaves at the Congo River
(Mpinda). This was reiterated in a Grant from King Manuel in 1504, which limited trading
to the Congo River, and banned commerce further South along the coast (Birmingham
1966:29). That Mpinda was by far the major source of slaves for the Gulf of Guinea Islands
throughout the century can be seen for example in the 1548 Enquiry into the São Tomé
trade, which stated amongst other things that 4,000 to 5,000 slaves were exported from
Mpinda every year in the 1530s (Brásio 1953: 197); however, by no means all of these were
retained for use on the Islands. According to Birmingham (1966: 78), trade at Mpinda
went into decline at the beginning of the 17th century, in favour of Luanda, further South.
Kwanza River; Luanda
Despite the 1504 ban on trading South of Mpinda, it is clear that for most of the century
slaves were acquired along the coast, particularly at the mouth of the Kwanza River, South
of Luanda, as indicated by Birmingham (1966: 32-33), quoting documents from 1532, 1548
and 1551. However, by 1575, when the Portuguese under Paulo Dias de Novais were
attempting to take over the site of Luanda as their first colonial settlement in Africa, there
was already a great deal of slave trading there, as much as 10,000 slaves a year, according
to one Jesuit priest quoted by Birmingham, and trade increased considerably at Luanda up
to the end of the century (1966: 41, 50).
Having shown where and when the slaves for the Gulf of Guinea islands were obtained, a
pattern for the import and distribution of those slaves can be established. From the final
decades of the 15th century until the ban in 1553, the Portuguese were obtaining slaves at
Ughoton and the Forcados River for transportation to São Tomé, and thence a partial
redistribution to Príncipe and on a very small scale to Annobon. However, for a four-year
period only (1514-1518), slaves were also being imported from this region directly to
Príncipe. Meanwhile, trading at Mpinda had begun before 1500, and as indicated above,
slaves were imported to São Tomé in large numbers up to the end of the 15th century, and,
no doubt, redistributed to some extent to Príncipe and Annobon. From the early 16th
century onwards, slaves were also being imported, in increasing numbers, from the
Kwanza River and Luanda. However, because of the steady decline in the sugar
plantation production in the latter half of the 16th century, together with foreign
competition in slave trading, it is very unlikely that slaves were imported to São Tomé
after around 1700, and redistribution to the other islands would have ceased long before
that. This being the case, one can now consider the ethnolinguistic background at the
trading ports in question, in order to estimate the language input in the slave population
on the Gulf of Guinea Islands. The language communities are as follows:
The Edo ethnolinguistic group corresponds to the ‘kingdom’ of Benin, which in the 16th
century also ruled over the area occupied by the Itsekiri, including the port of Ughoton,
and the Forcados River. The people and the language have in the past also been referred
to as ‘Bini’ (e.g. by Ferraz), but the recognised ethnonym is nowadays Edo. The language,
formerly classified as a member of the Kwa family, is now considered to be part of the
Benue-Congo group (cf Elugbe 1989b).
The Itsekiri occupy the coastal area roughly between the first three of the five ‘Slave
Rivers’, including the ‘Fermoso’, or Benin River; they were referred to by Duarte Pacheco
Pereira as “Huela” (Pereira 1988: 152). Although direct trading with them by the
Portuguese is not documented, it is known that the Edo sold slaves acquired from that
ethnolinguistic group. Their language is closely related to Yoruba, one of the principal
languages of modern Nigeria - indeed, “in a list of over 100 common Itsekiri words over
90% are almost identical with Yoruba words” (Lloyd, in Bradbury & Lloyd 1957: 174).
The Ijo ethnolinguistic group occupies the coastal area of the Niger Delta between the
Forcados River and the Bonny River (‘Rio Real’). They are referred to by Duarte Pacheco
Pereira as the “Jós”, with whom trading was undertaken both at the Forcados River and
the Rio Real; despite this, no less than three times, Pereira (1988: 152-53, 155-56) states that
the Ijo were cannibals. The Ijo language is part of the Benue-Congo family, but it is not
closely related to either Edo or Yoruba/Itsikeri.
The Kikongo ethnolinguistic group occupies a large area on both sides of the Congo River:
Portuguese contact and trade in the 16th century was concentrated on the South side of the
River at Mpindi, where the main variety of the Kikongo language was Kishikongo. In any
case, the varieties of Kikongo are closely related and are normally mutually
The division between the Kikongo and Kimbundu ethnolinguistic groups was normally
considered to be the Dande River, North of Luanda; as indicated above, direct trading
with this area, albeit clandestine, was flourishing throughout the 16th century, but it must
also be borne in mind that the Kongo often obtained slaves through raids into
neighbouring ethnic groups, including the Mbundu (Birmingham 1966: 26). The Kim-
bundu language is spoken over a large area inland from Luanda and the Kwanza River; it
10 In common with modern practice for Bantu names, I use Kikongo (which includes the class prefix ki-) to
refer to the language but Kongo to refer to the people and the territory they occupy.
11 This is not to be confused with the neighbouring language, Umbundu, from whose community some slaves
were traded to the Portuguese at Luanda..
is reasonably close to Kikongo, as a loosely related Bantu language, and to a small extent
the two languages are mutually comprehensible.
One can now estimate that on São Tomé there would have been a mixture of Edo,
Kikongo and Kimbundu among the slave population, with Edo, the language of the early
input on the island, being added to, and later replaced, by the two Bantu languages in the
course of the 17th century. On Príncipe, however, one can expect a higher proportion of
Edo-speakers, because of the direct import to the island of slaves from the Niger Delta
between 1514 and 1518, and also because redistribution from São Tomé had probably
ceased by about 1550, before the boom in imports from the Bantu-speaking areas.
Meanwhile, there was probably one small shipment of slaves to Annobon from São Tomé,
and depending on the timing, this would include proportionate numbers of Edo- and
Bantu-speaking slaves. As for the Angolar community, the majority would consist of
Kimbundu-, or possibly Kikongo-speaking slaves, from the shipwreck of the vessel from
‘Angola’, though there could well be some imput from Edo-speakers as well from the
runaway plantation slaves who joined them, probably around the 1550s.
Evidence from this conjectured pattern of ethnolinguistic input into the Gulf of Guinea
Creoles can be obtained from an examination of the distribution of non-Portuguese (i.e.
African-derived) lexical items in the four creoles. In Appendix 2, I have listed all those
lexical items I could find which have African origin. However, the list is necessarily
incomplete, as a result of the limited resources currently available, not only for the creoles
themselves, but also for the African languages.12 The latter would appear to be largely
limited to three (groups of) languages: Edo (including a few items from closely related
Edoid languages), Bantu languages (mainly Kikongo and Kimbundu), and Yoruba.13 Of
the identifiable etyma, I have calculated the proportions as follows:
Table 1. Percentages of words identified as African by source languages
São Tomé: Bantu 56%, Edo 35%, Yoruba 6%, others 3%;
Príncipe: Bantu 23%,14 Edo 65%, Yoruba 10%, others 2%;
Annobon: Bantu 52%, Edo 38%, Yoruba 6%, others 4%;
Angolar Bantu 92%, Edo 6% , Yoruba and others 2%.
The large proportion of Edo-derived words in Príncipe Creole is clearly the result of the
exclusive contract between Benin and this island as indicated above. In contrast, Angolar
has only 6% of Edo-derived words compared with 92% Bantu-derived, while both the São
Tomé and Annobon Creoles have more than 50% Bantu-derive items and a little less than
40% from Edo. Thus the proportions of African ethnolinguistic input projected above are
borne out: São Tomé has a greater proportion of Bantu items than Edo, reflecting the
rapid growth of plantation production in the mid-16th century, and the increasing import
of slaves from Mpinda and Luanda; on Annobon, because the proportions are roughly
equal, the redistribution of slaves to the island from São Tomé probably took place in the
first quarter of the 16th century; meanwhile, on Príncipe, the projection that Edo-speakers
would outnumber Bantu-speakers is borne out by the figures quoted above. As for the
12 There are 35 Angolar items which have cognates in one or more of the other Gulf of Guinea Creoles. A
further 343 items of African origin, identified in Maurer (1995) and/or Rougé (2004), are almost
exclusively of Kimbundu origin.
13 The Yoruba items could be accounted for by the fact that slaves were traded from the Itsikiri, whose
language, asmentioned above, is very close to Yoruba. Given Yoruba’simportance as a regional
language, it may also have been known as a second language to some Edo speakers.
14 I had earlier (Ladhams 2003) calculated the proportions of lexical items in Príncipe as being Edo 63% and Bantu 37%
but this has now been revised thanks to data obtained from Maurer (p c, 2007).
figures for Angolar, unsurprisingly the vast majority of non-Portuguese lexical items is
Kimbundu in origin. Further support for this interpretation of the data can be found by
comparing the origins of items currently known to exist in only one of these languages: in
Príncipe Creole, 76% of the latter are of Edo origin whereas the majority is of Bantu origin
in the other three: 81% in São Tomé Creole, 67% in Annobon Creole, and ca. 96% in
Angolar (P Baker, p c).
To go back to the possible source for article agglutination, and particularly the
question as to why there is such a large number of examples in Pr, one should now seek an
explanation in examining the Edo language, to see if there is any feature that could give
rise to the article agglutination in this creole. Indeed there is such a feature. While the
agglutinated articles in the Indian Ocean French Creoles could be explained by the
reanalysis of Bantu noun class markers, it should be remembered that French articles, and
Bantu noun class markers are structured CV. This is not the case in the Gulf of Guinea
Creoles, where the Portuguese articles consist of V only, so article agglutination cannot be
attibuted to Bantu influence. On the other hand, in Edo, “nouns invariably have a prefix
attached to the stem” (Elugbe 1989b:199), and those prefixes always consist of a vowel
only, and that vowel is very often [u] (1989b:200).15 This of course corresponds very
closely with the Pr data for article agglutination.17 Therefore, one can safely say that there
is indeed a case of African influence in a particular feature of the Gulf of Guinea Creoles,
and the origin of that influence can be ascribed to the Edo input into those creoles.
4. Analyses by word class and origins
Angolar is excluded from the analyses which follow because I do not have complete data
on the word classes of the many items which are attested only in this language.
Words are sorted into three groups; 1. nouns; 2. adjectives, verbs, and adverbs (abbre-
viated here as ”adjverbs”); and 3. others, comprising what are traditionally known as
closed classes but also including ideophones.
Table 2. Analysis of all items by word class and Creole
nouns adjverbs others total
ST 63 (60%) 35 (33%) 8 ( 7%) 106 (100%)
Pr 51 (55%) 36 (39%) 5 ( 6%) 92 (100%)
Ann 52 (79%) 14 (21%) 0 66 (100%)
The proportion of nouns indicated above is undoubtedly much lower than is found in
many Caribbean Creoles. However, a proper assessment of the significance of this
requires comparisons to be made with a wider geographical range of languages. The
editors attempt to do this in the final article in this volume.
Table 3. Analysis of all items by Creole and source language
Bantu Edo Yoruba Others
ST 56% 36% 6% 2%
Pr 24% 67% 9% 0%
Ann 52% 38% 6% 4%
15 See also Elugbe (1989a) for further details, including his reconstruction of noun class markers consisting of
an initial vowel in Proto-Edoid.
17 Hagemeijer (2003: 167; 2009) suggests that the addition of an initial vowel to nouns in Pr is not the result of
article agglutination, but rather a question of vowel harmony. The data presented here would seem to
prove this not to be the case. For details on vowel harmony in Edo and Proto-Edoid, see Elugbe (1989a:
47-48, 118-19).
Table 3 is the same as Table 1 except that figures for Angolar are excluded for reasons
stated above. The information is repeated here to avoid the reader having to turn back the
page in order to make comparisons with the following tables. Note that all percentages
are give to the nearest whole number with the consequences here and below that 0% may
in some cases mean less than 0.5% rather than zero, and that, if percentages are totalled,
they may in one or two cases add up to 99% or 101%.
Table 4. Analysis of nouns by Creole and source language
Bantu Edo Yoruba Others
ST 65% [56%] 29% [36%] 3% [6%] 3% [2%]
Pr 30% [24%] 58% [67%] 12% [9%] 0% [0%]
Ann 53% [52%] 37% [38%] 6% [6%] 4% [4%]
The figures in square brackets here indicate the percentage contribution of each source
language to the vocabulary as a whole. Thus it can be seen that the Bantu contribution to
nouns is appreciably greater than its overall input while, for Edo, the reverse it true. There
will be further comment on this following Table 6. Note that the figures for Annobon are
remarkably similar to its overall averages.
Table 5. Analysis of adjverbs by Creole and source language
Bantu Edo Yoruba Others
ST 51% [56%] 40% [36%] 8% [6%] 1% [2%]
Pr 15% [24%] 78% [67%] 7% [9%] 0% [0%]
Ann 49% [52%] 39% [38%] 9% [6%] 2% [4%]
Rather predictably, Table 5 shows the opposite of the Table 6 insofar as the Bantu
contribution to “adjverbs” is less than, and the Edo contribution greater than, their
respective overall averages. There will be further comment on this below. The figure for
Annobon are again strikingly similar to its overall averages.
Table 6. Analysis of other items by Creole and source language
Bantu Edo Yoruba Others
ST 19% [56%] 81% [36%] 0% [6%] 0% [2%]
Pr 20% [24%] 80% [67%] 0% [9%] 0% [4%]
Ann 0% [52%] 0% [38%] 0% [6%] 0% [4%]
What is immediately striking about Table 6 is that the dominant contributor to São Tomé’s
closed class items is Edo rather than Bantu languages, and that the percentage figure is
almost identical to the Príncipe figure. Also notable is that fact that no items in this class
have yet been identified for Annobon.
Taking Tables 4, 5, and 6 together, there is a fall in the Bantu contribution to São Tomé
from 65% for nouns, to 51% for adjverbs, and to 19% for closed class items whereas there is
a rise in the Edo contribution to Príncipe from 29% for nouns, to 78% for adjverbs, and to
80% for closed class items. It is widely accepted that nouns are the class of words most
easily borrowed from one language into another. If that is true, having an increasing
proportion of Edo-derived items for both adjverbs and close class items –which is true of
both São Tomé and Príncipe – might be interpreted as an indication that Edo played a
more important or at least more specialized role than Bantu languages in the formation of
these two creoles. While the information about the peopling of Príncipe presented earlier
can readily account for this on that island, it is not obvious that there can be simple of this
in São Tomé. Before attempting to answer this question, data from other contact
languages needs to be examined. This will be untertaken by the editors in the final article
in this book.
5. Conclusions: the African contribution to Portuguese Creoles
Studies of the African influence in the Gulf of Guinea Creoles were made by Ferraz in the
1970s and 80s,18 showing that apart from the lexical items of African origin, there was a
number of grammatical features in the languages in which parallels could be drawn from
appropriate ‘substrates’. It should be pointed out that Ferraz, as an avowed ‘substratist’,
was anxious to disprove any ‘universalist’ influence in these creoles, particularly in
relation to Bickerton’s Language Bioprogram Hypothesis (see especially Ferraz 1984).
Therefore, one should be cautious about accepting all the many examples he gives of
African influence. However, in fact he has very little to say about article agglutination in
the Gulf of Guinea Creoles, and that is confined to Ferraz (1979). Other descriptions of the
four creoles do refer to article agglutination: Günther (1973) and Maurer (2009) on Pr, De
Granda (1994: 424-39) and Maurer (1995) on Ang, and Schang (2003) on ST, but there is
little or no discussion of possible African influence. Only in Parkvall (2000) and Ladhams
(2003) is there any detailed and reliable account of this influence in the Gulf of Guinea
Creoles. Parkvall includes some phonological features, with differences noted between Pr
and the other three creoles, pronoun forms, and especially the double negative
construction, analysed in detail in Hagemeijer (2003). Ladhams (2003) includes all these
features, plus the use of ideophones of African origin in the creoles, as well as
reduplication, also analysed in more detail in Ladhams et al. (2003).
It is clear, therefore, that the Gulf of Guinea Creoles do exhibit quite a number of
features which can be attributed to African influence, but what of the other African
Portuguese-based Creoles? Ladhams (2003) also examines African influence in
Capeverdean Creole but concludes: ‘substrate’ influence in Capeverdean Creole is not
particularly evident: indicative of this is the fact that in percentage terms (of the whole
vocabulary) the number of lexical items in the creole identifiable as being of African origin
is quite small, while at the other linguistic levels, the influence is at best slight.” (2003:
138).19 There has little research done on the adstrate African influence in Guinea-Bissau
Creole: one might expect that being located on the African mainland, adjacent to at least 20
different African languages, there would be considerable African influence on the creole,
but my impression is that there is as little influence as in the related Capeverdean Creole.
Possibly the key factor is that no one or two African languages are dominant over the
others in Guinea-Bissau, and therefore the influence on the creole is very limited.20 But
another question arises: why should there be so much more African influence in the Gulf
of Guinea Creoles than in the other two? The answer lies, I feel sure, in differences in the
sociohistorical background: the plantation system in the Gulf of Guinea was so intense
that it led to the formation of a separate slave community, with its own identity, and
therefore separate languages (creoles) of their own.21 Consequently, the creoles in turn
would be more subject to African influence.
18 Papers on the African influence on the creoles respectively in Príncipe (1975) and Annobon (1984), as well
as a section on this subject in his book on São Tomé Creole (1979: 110-15).
19 On the question of the number of lexical items of African origin, note that Brazilian Portuguese contains at
least 2,500 items of African origin (see Schneider 1991), but there is no evidence at all that Portuguese
was ever creolized in Brazil.
20 However, Rougé (2004) gives 333 lexical items of non-Portuguese origin in Guinea-Bissau Creole,
compared with only 145 in Capeverdean Creole.
21 For a detailed statement of this position, see Ladhams (2003).
Since the mid-1980s, there has been a great debate on the role of ‘substrate’ influence
in creole formation: see especially Muysken & Smith (1986) and Mufwene (1993).
However, in both these volumes, there is no mention of the Portuguese-based Creoles, and
until now the dicussion has been largely confined to the English- and French-based
Creoles. This also the case with the recent publication by self-affirmed ‘superstratists’:
Chaudenson (2001, 2003) and Mufwene (2001, 2005). There is no mention of Portuguese-
based Creoles, and though Mufwene does acknowledge some ‘substrate’ influence in the
English- and French-based Creoles, Chaudenson, as shown above in section 1, seems
extremely reluctant to accept any such influence.22 More than 20 years ago,23 Baker was
somewhat pessimistic about assessing the African influence, at least in the French-based
Creoles: “Identifying the African contribution to French-based Creoles in areas other than
the lexicon is a very difficult matter. The basic problem is that we do not know enough
about how creole languages originated and evolved” (1993: 140). By 2000, this problem
had been overcome to a great extent, as shown by the detailed overall survey of African
influence in creole languages in Parkvall (2000). Nevertheless, Parkvall also sounds a note
of pessimism when he states, in his concluding discussion: “The first thing that I myself
learnt from this study is that there are far fewer clearly substrate-induced structures in the
Atlantic Creoles than I had expected to find” (2000: 154). It is hoped that this article can
make a significant contribution to the debate on African influence, by showing that in the
Portuguese-based Creoles at least, that influence is evident.
Baker, Philip 1984 Agglutinated articles in Creole French: their evolutionary significance.
Te Reo 27: 89-129.
–––– 1988 Combien y a-t-il eu de genèses créoles à base lexicale française? Etudes Créoles
10(2): 60-76.
–––– 1993 Assessing the African contribution to French-based Creoles. Mufwene, S (ed.),
Batalha, Graciete 1988 Glossário do Dialecto Macaense. Macao: Instituto Cultural de Macao.
Baxter, Alan N & De Silva, Patrick 2004 A dictionary of Kristang. Canberra: Pacific
Bentley, W Holman 1887 Dictionary and Grammar of the Kongo language. London: Baptist
Missionary Society / Trübner & Co.
Birmingham, David 1966 Trade and Conflict in Angola. Oxford: Clarendon Press.
Bradbury, R E & Lloyd, P C 1957 The Benin Kingdom and the Edo-speaking Peoples of South-
Western Nigeria, [& The Itsekiri]. London: International African Institute.
Brásio, António (ed.) 1953 Monumenta Missionaria Africana. Vol. II. Lisbon: Agência Geral
do Ultramar.
Chaudenson, Robert 2001 Creolization of Language and Culture. London & New York:
–––– 2003 La créolisation: théorie, applications, implications. Paris: L’Harmattan.
22 Neumann-Holzschuh (2007: 384) states that: “Chaudenson, who, although often considered a “Euro-
centrist”, does not deny that African languages may have played a considerable role in the development
of creoles”. However this is clearly not the case in Chaudenson (2003: 262-69).
23 The paper published as Baker (1993) was first presented at a conference in Athens, Georgia in 1987.
Clements, J Clancy 1995 The Genesis of a Language. Amsterdam: John Benjamins.
Dalgado, Sebastião Rodolfo 1900 Dialecto Indo-Português de Ceilão. Lisbon: Imprensa
–––– 1906 Dialecto Indo-Português do Norte. Revista Lusitana 9: 142-66, 193-228.
De Granda, Germán 1985 Estudios de Lingüística Afro-Románica. Valladolid: Universidad
de Valladolid.
–––– 1994 Español de América, español de África y hablas criollas hispánicas: cambios, contactos
y contextos. Madrid: Gredos.
Elugbe, Ben Ohi 1989a Comparative Edoid: Phonology and Lexicon. Port Harcourt:
University of Port Harcourt Press.
–––– 1989b Edoid. Bendor-Samuel, John (ed.), The Niger-Congo Languages. Lanham et al.:
University Press of America, 291-304.
Fernandes, Valentim 1940 O manuscrito de Valentim Fernandes. Lisbon: Academia
Portuguesa da História.
Ferraz, Luiz Ivens 1975 African influences on Principense Creole. Valkhoff, Marius (ed.)
Miscelânea Luso-Africana. Lisbon: Junta de Investigações Científicas do Ultramar, 153-
–––– 1979 The Creole of São Tomé Johannesburg: Witwatersrand University Press.
–––– 1984 The Substrate of Annobonese. African Studies, 43(2): 119-36.
Grant, Anthony P 1995 Article agglutination in Creole French: a wider perspective.
Baker, Philip (ed.) From Contact to Creole and Beyond. London: University of
Westminster Press, 149-76.
Günther, Wilfried 1973 Das portugiesische Kreolisch der Ilha do Príncipe. Marburg:
Marburger Studien zur Afrika- und Asienkunde.
Hagemeijer, Tjerk 2003 A negação nos crioulos do Golfo da Guiné: aspectos sincrónicos e
diacrónicos. Revista Internacional de Lingüística Iberoamericana 2: 151-78.
–––– 2009 Initial vowel agglutination in the Gulf of Guinea creoles. Aboh, Enoch O &
Smith, Norval (eds.) Complex Processes in New Languages. Amsterdam: Benjamins,
Ladhams, John 2003 The Formation of the Portuguese Plantation Creoles. Unpublished PhD
dissertation, University of Westminster.
Ladhams, J, Hagemeijer, T, Maurer, P & Post, M 2003 Reduplication in the Gulf of Guinea
Creoles. Kouwenberg, Silvia (ed.) Twice as Meaningful. London: Battlebridge, 165-74.
Laman, Karl Edvard 1936 Dictionnaire kikongo-français. Brussels: Georges van
Le Guennec, Grégoire & Valente, José Francisco 1972 Dicionário português-umbundu.
Luanda: Instituto de Investigação Cientifica de Angola.
Lopes da Silva, Baltasar 1957 O Dialecto Crioulo de Cabo Verde. Lisbon: Imprensa Nacional.
Maia, António da Silva 1964 Dicionário complementar português-kimbundu-kikongo.
Cucujães: Editorial Missões.
Maurer, Philippe 1992 L’apport lexical bantou en angolar. Afrikanische Arbeitspapiere 29:
–––– 1995 L’angolar. Un créole afro-portugais parlé à São Tomé. Hamburg: Buske.
–––– 2009 Principense. London & Colombo: Battlebridge.
McWhorter, John & Parkvall, Mikael 2002 Pas tout à fait du français: une étude créole.
Études créoles 25(1): 179-231.
Melzian, Hans 1937 A concise dictionary of the Bini language of Southern Nigeria. London:
Kegan Paul, Trench, Trubner & Co.
Mufwene, Salikoko (ed.) 1993 Africanisms in Afro-American language varieties. Athens GA:
University of Georgia Press.
–––– 2001 The Ecology of Language Evolution Cambridge: Cambridge University Press.
–––– 2005 Créoles, écologie sociale, évolution linguistique. Paris: L’Harmattan
Muraz, Gaston [ca. 1931] Vocabulaire du patois arabe tchadien ou “tourkou” et des dialectes
sara-madjinngaye et sara-m’baye. Paris: Charles Lavauzelle.
Muysken, Pieter & Smith, Norval (eds) 1986 Universals versus substrata in creole genesis.
Amsterdam: Benjamins.
Negreiros, António de Almada 1895 Historia ethnographica da Ilha de S. Thomé. Lisbon: José
Neumann-Holzschuh, Ingrid 2007 Review of Mufwene (2005). Journal of Pidgin and Creole
Languages 22: 382-86.
Parkvall, Mikael 2000 Out of Africa. London: Battlebridge.
Pereira, Duarte Pacheco 1988 Esmeraldo de Situ Orbis. Lisbon: Academia Portuguesa da
Rougé, Jean-Louis 2004 Dictionnaire étymologique des créoles portugais d’Afrique. Paris:
Ryder, Alan 1969 Benin and the Europeans. London: Longmans.
Schang, Emmanuel 2003 L’émergence des créoles portugais du Golfe de Guinée. Villeneuve
d’Ascq: Septentrion.
Schneider, John T 1991 Dictionary of African Borrowings in Brazilian Portuguese. Hamburg:
Schuchardt, Hugo 1890 Ueber das Malaioportugiesische von Batavia und Tugu.
Sitzungsberichte Wien 122: 1-256.
Speedy, Karin 2007 Réunion Creole in New Caledonia – what influence on Tayo? Journal
of Pidgin and Creole Languages 22: 193-230.
Swartenbroeckx, Pierre 1973 Dictionnaire kikongo et kituba-français. Bandundu.
Valkhoff, Marius F 1966 Studies in Portuguese and Creole. Johannesburg: Witwatersrand
University Press.
São Tomé Príncipe Annobon Angolar Portuguese English
abyá a beira (river) bank
afé afé a fé faith
aglasa a graça grace, name
alê arê alê el-rei king
aleduña el-rei doninha weasel
apá a pá spade, shovel
idintxi o dente tooth
ifi o fio thread
ikyabu ukyabu o quiabo okra
imin o milho maize
injezu o jejum fasting
irizi o nariz nose
isengi o sangue blood
ite a terra land
ivin o vinho wine
ixize a cinza ash
oali uari o ar air
omali umwe omal o mar sea
opé opé opé o pé foot
opo opó opó ompo o pó dust
opó o polvo octopus
oryó oryô o rio river, diarrhoea
osé osé osé onɵe o céu heaven, sky
otés *o testa forehead
ubaaku o buraco hole
ubaasu o baixo underneath
ubaasu o braço arm
ubanku o banco bench
ubasu o baço spleen
ubaw o barro clay
ubên os bens goods
ubuka *o boca mouth
ubwê o boi ox
udedu o dedo finger
udêntu o dentro inside [prep.]
udôdô o doido stupid [adj.]
udyabu o diabo devil
ufaka *o faca knife
ufatu o fato suit
ufeu ofelu o ferro iron, metal
ufôgô o fogo fire
ufundu o fundo bottom, base
ufya *o folha leaf
ugagu o gago stutter [v.]
ugalu o galo rooster
ugatu o gato cat
ukabu o cabo end
São Tomé Príncipe Annobon Angolar Portuguese English
ukaku o casco turtle shell
ukalu o caldo broth
ukampu o campo field
ukaru o carro car
ukantu o canto eyebrow
uku uku o cu arse
ukuru o escuro dark [adj.], darkness
ukurú o cru raw [adj.]
ukwatu o quarto room
uladu o lado side
ulasu o laço knot, ribbon
ulatu o lado [?] way, path
ulensu o lenço headscarf
uman oman *o mão hand
umatu o mato forest
umpan um pão bread, loaf
umundu o mundo world
umuña uma unha fingernail
unfenu o inferno hell
unôtxi *o noite night
unu unú ono o nu naked [adj.]
unwan onuja *o lua moon
upá opá o pau tree, stick
upan ãpã o pão bread
upanu o pano cloth
upasu o passo step
upetu o espeto spit, fork
upêtu o peito chest
urabu o rabo tail
uramu o ramo branch (tree)
uratu o rato mouse, rat
uremu o remo oar, paddle
urôsô o arroz rice
usaku o saco sack, bag
usalu o sal salt
usan o chão ground
usolu o sol sun
usuva *o chuva rain
utabu *o tábua board, plank
utasu o tacho pan
uventu oventu o vento wind
uwê o olho eye
uzên osojo o joelho knee
zálima zálima as almas ghost
zonda zonda as ondas wave
Gulf of Guinea Valkhoff (1966: 92, 133); Schang (2003: 103-13); Rougé (2004)
São Tomé Negreiros (1895: 355-69); Ferraz (1979: 27)
Príncipe Günther (1973); Maurer (p c 2007, 2009)
Annobon Ferraz (1984: 125-26); De Granda (1985: 159-65, 171-79; 1994: 433)
Angolar Maurer (1995).
ST Pr Ann Ang English etymon African language(s)
a a a [ind. pr.] a Edo
ába abá trunk, branch aba Twi/Igbo
akê clay pot àkhé Edo
akelé akaré frog ekire/akere/akiri Edo/Yoruba/Igbo
andjí adí andji palm-nut edin/adin Edo/Yoruba
anká kara áxala anka crab nkala Kikongo
baa burn [vb.] bàá Edo
bambí bambí pneumonia m-bambi Kikongo/Kimbundu
bangá palm-oil sauce bannga Kikongo
bánsa bansa hook m-basa/lu-basa Kikongo/Kimbundu
bánsa mbasá mbasá rib lu-banji/m-banji Kikongo/Kimbundu
baya bayubya baya bewitch, spell gbye Yoruba
be cut, pierce bele Edo
benku benku tortoise ? Edo?
bi shove (v.) bi Edo
ba where is? vbòó a Edo
bobó bobó bobó bobó ‘piggy-back’ bobo Edo
bôbô bôbô bôbô bôbô ripe, red booba Kikongo
bomon hurry off bomon Edo
búdu ubúdu du stone bodo Edoid
búnga tree sp. m-bunga Kikongo
byê byê byê cook (v.) biê Edo
byogo slip mioghon Edo
dedé embrace dede Edo
dindin elderly person ódèdè Edo
dúmu udumu pound, grind dunvu Edo
éli moonshine - xeli ‘Proto-Bantu
fefé suck fefeña Kimbundu
fingí fingí mouse m-fingi Kikongo
go away Ewe
fofó fofó fo fo winnow fofo/hoho Yoruba/Edo
fúba fúba maize flour m-fuba/fuba Kikongo/Kimbundu
fufú pounded yam fufu Yoruba
fu fugú turn (soil) hùgú Edo
fumá fumá fumá swell fuma Kikongo
fúnda fun funda bundle funda Kikongo
gbê crush gbé Edo
gegé tree sp. mu-ngyengye Kikongo
complain go Edo
gogó gogó ngongó congratulate gogo Edo
gólo go goló search, dig gwalo Edo
gomon germinate mon Edo
gyen leak gin Edo
ibi charcoal ibi Edo
ídu ídu ídu louse iru Edo
ig ogé body egbe Edo
igbegbé snail igbin Yoruba
ST Pr Ann Ang English etymon African language(s)
iglígu igugu ígogo idigô smoke igogo Edo
iki hill oke Edo
ikilí ikiri pad (on head) kili Bantu(?)
iku iku rubbish iku Edo
ingléva bweva twins ivin eva Edo
inen/ine ine ane they ian Edo
íze izé íse/áse idhe crayfish, shrimp ize Edo
jambá sling lukaámba Kikongo
jojó horn oko Edo
jojoló cut (v.) -kohol-a Kimbundu
jongo banana sp. koongo Kikongo
jongul bracelet lu-kongolo Kikongo
tiny xerhé Edo
ketéketé very old kya-kete Kikongo
kili coil (v.) kii ri Edo
kisamá scorpion ki-sama Kikongo
kiséngle hatchet n-sengele Kikongo
kitxibá kintxiba kitxíba banana sp. tiba/tʃiba Kikongo
kite a little/small kere/kete Yoruba/Kikongo
klón-klón koko neck, throat gogongo Yoruba
klókoto krokotó kokotó cockroach kokoto Kikongo
resemble xo Edo
kôkô kôkô crawl ku-dikoka Kikongo
kóni blow with fist n-kome Kikongo
kumbé large drum n-kuúmbi Kikongo
kumbili tree sp. n-kuúmbi Kikongo
kundu ukundu pubis n-kundu Kikongo
kunkunú bedbug ki-ngunu-ngunu Kimbundu
kusá scratch, scabies -kús-a Kikongo
kútu kutú thicken (food) -kukuta Kikongo
kwa be heavy xua Edo
kwali ugberi basket ? Edo?
larya be bitter riara Edo
lelé lelé lelé follow lele/le Edo/Yoruba
lemá overcast lema Kikongo
levá twins ivi eva Edo
lóla turn off, detour -kondoloka/bolola Kikongo/Kimbundu
loló completely ulolo Kikongo
loló loló lolo lick lalo Edo
lululú lululu black (ideoph.) duduudu/nunuu Edo
makoyá scurf kooya Kikongo
makúku makuku hearthstones ma-kuku Kikongo
maku palm-leaf bowl nkutu Kikongo
mankwe smart (ma)wete Kikongo
masánga necklace mu-sánga Kikongo
mbá hang, hook ba Edo
mene be sweet myemye Edo
mikondó baobab n-kondo Kikongo
mukumb tree sp. mu-kumbi Kikongo
muléle fishing net mu-lele Kimbundu
munjá mund myilá stand, stop mu dia Edo
munkén múnkele dove sp. mu-nkele Kikongo
musámbi dried fish mu-nsambu Kikongo
ST Pr Ann Ang English etymon African language(s)
muténde mutende young palm tree mu-tende Kikongo
mutete palm-leaf basket mu-tete Kimbundu
mwíndu bush sp. mw-indu Kikongo
ngándu ngandú ngandu shark n-gandu Kikongo/Kimbundu
ngembú bat n-gembo Kikongo
ngené-ngené twinkling ngedi-ngedi Kikongo
ngumbá jegunba ngúmba nguba groundnut n-gumba Kikongo/Kimbundu
ngunú-ngunú grumble ki-ngunu-ngunu Kimbundu
njánja nyányi quickly l-onjanja Umbundu
nyonyón snail nyonyo Kikongo
-o [respect tag] -o Edo
obó ovyó ogó obo forest ogo Edo
odó idó udum mortar odo Edo/Yoruba
ofí/ofía witchcraft ubio Efik
ógo yard ugbo Edo
ógogo pot ikoko Yoruba
óka uká oxá banyan tree oka Edo
óke ôkyê okyé íki hill oke Edo/Yoruba
okó oxó calabash uko Edo
okori woodworm oxoye Edo
ôk ogó clearing ogo Edo
otó neck urhu Edo
owó you (pl.) ouo Edo/Kimbundu
oxál basket kwalu Kikongo
óxo navel uvu ihue Edo
pénu-pénu eyelash, eyelid ipenpenju Yoruba
pete flat perhe Edo
te-péte peté-peté soft, gentle pete-pete Kikongo
pimbí pímpi penis pimbi Kimbundu
to-to potopotó drenched potopoto Ijo, Kikongo, Yoruba24
púnda pidi punda because rhun-da Edo
safú sáfu safu fruit sp. n-safu Kikongo
gwa medicine man sangwa Kikongo
samangungú tarantula esangangungu Kikongo
selelé termite n-selele Kikongo
sénkwa thekwa bug ki-nsekwa Kikongo
súxu xuxu dagger, spike susu Kikongo
tanga tanga loincloth tanga Kimbundu
tokólo fruit sp. -tokol-a Umbundu
twadu biting/acid tuisa Kimbundu
tumbó large calabash n-túmbu Kikongo
túmbu dust u-tumbu Kimbundu
ubága baga saucepan ukpabo Edo
úbua ugba fence ugba Edo
ubwé ibé ogé body egbe Edo
ubya spell (n.) gbeye Yoruba
ufu ufu lather (v.) ihu Edo
uka tree sp. okan Edo
úku íku íku rubbish iku Edo
ukugbá belt ukugba Edo
24 Potopoto also occurs in other West African languages, including Mandinka and Twi, and is found in a wide range of
Caribbean Creoles as well as in Turku Pidgin Arabic.
ST Pr Ann Ang English etymon African language(s)
úkwe igbé ikwe seed ikpe Edo
ulalu skin rash alalu Edo
uruba boil (n.) uruvba Edo
ururu umbilical cord eru Edo
utú utú mushroom utun, otu Edo
uu thread oru Edo
va split va Edo
vanganá stagger vungana Kikongo
vúgu vugu struggle out vuku Kikongo
vunvú vunvú buzz m-vofom Fang?
vunvún bunbun bee mvuvu Kikongo
vuvá howl boba Kikongo
wangá wangá wangá drizzle, spill - mwanga Kikongo
xambá climbing ring lukaámba Kikongo
xóngo banana sp. kóongo Kikongo
xongúl bracelet lu-kongolò Kikongo
xoxó horn oko Edo
xololo cut (v.) -kohol-a Umbundu
xoxoso elbow kokói [Bantu]
yangá tear (v.) yanza Kikongo
please (v.) ye Edo
much, many yo Edo
zámba dhanba elephant n-zamba Kikongo
zowó/zó then zevo Kikongo
zúmbi dhunbi ghost zumbi/n-zumbi Kikongo/Kimbundu
Gulf of Guinea Rougé (2004)
São Tomé Ferraz (1979: 90-100)
Príncipe Günther (1973); Ferraz (1975); Maurer (p c 2007, 2009)
Annobon Ferraz (1984); De Granda (1985: 159-65)
Angolar Maurer (1992; 1995).
Edo Melzian (1937)
Kikongo Bentley (1887); Laman (1936); Maia (1964); Swartenbroeckx (1973)
Kimbundu Maia (1964)
Umbundu Le Guennec & Valente (1972)
John Ladhams has spent most of his working life in Portuguese-speaking
countries, particularly as a Lecturer in English Language at universities in
Portugal and Brazil. In the past ten years he has taught linguistics, translation,
and English language at three of London’s Universities.
His research has concentrated on Portuguese as a contact language
in Africa, Asia and South America. <>
... Príncipe received a small number of prisoners from Bantu regions in contrast to São Tomé that continued to receive a large Bantu input (Hagemeijer 2011). 5 Contrary to Ferraz (1979), who considered both Edoid 6 and Bantu languages to be the main substrates of the Gulf of Guinea creoles, the current view on their history is that languages from the Niger Delta, particularly Edoid languages, were the main substrate, while Bantu languages, such as Kikongo and Kimbundu, played a secondary role (Hagemeijer 2011; see also Agostinho et al. 2019;Ladhams 2012) (see Map 2). Ladhams (2012) shows that Lung'Ie has the lowest percentage of Bantu words and the highest percentage of Edo words compared to the other Gulf of Guinea 5 Genetic studies show a predominance of Benin (52.3%) and Bantu (36.4%) haplotypes in the city of São Tomé (Tomás et al. 2002: 403), which also confirms that "Central-West Africa and the western areas of Bantu-speaking Africa were the major sources of the slave settlers in São Tomé" (Tomás et al. 2002: 407). ...
... 5 Contrary to Ferraz (1979), who considered both Edoid 6 and Bantu languages to be the main substrates of the Gulf of Guinea creoles, the current view on their history is that languages from the Niger Delta, particularly Edoid languages, were the main substrate, while Bantu languages, such as Kikongo and Kimbundu, played a secondary role (Hagemeijer 2011; see also Agostinho et al. 2019;Ladhams 2012) (see Map 2). Ladhams (2012) shows that Lung'Ie has the lowest percentage of Bantu words and the highest percentage of Edo words compared to the other Gulf of Guinea 5 Genetic studies show a predominance of Benin (52.3%) and Bantu (36.4%) haplotypes in the city of São Tomé (Tomás et al. 2002: 403), which also confirms that "Central-West Africa and the western areas of Bantu-speaking Africa were the major sources of the slave settlers in São Tomé" (Tomás et al. 2002: 407). 6 Since Edoid languages were earlier classified as Kwa (vs. ...
... Percentages of words identified as African by source languages from Ladhams (2012) in the Gulf of Guinea creoles: São Tomé: Bantu 56%, Edo 35%, Yoruba 6%, others 3%; Príncipe: ...
Creole languages have generally not figured prominently in cross-linguistic studies of word-prosodic typology. In this paper, we present a phonological analysis of the prosodic system of Lung’Ie or Principense (ISO 639-3 code: pre), a Portuguese-lexifier creole language spoken in São Tomé and Príncipe. Lung’Ie has produced a unique result of the contact between the two different prosodic systems common in creolization: a stress-accent lexifier and tone language substrates. The language has a restrictive privative H/Ø tone system, in which the /H/ is culminative, but non-obligatory. Since rising and falling tones are contrastive on long vowels, the tone must be marked underlyingly. While it is clear that tonal indications are needed, Lung’Ie reveals two properties more expected of an accentual system: (i) there can only be one heavy syllable per word; (ii) this syllable must bear a H tone. This raises the question of whether syllables with a culminative H also have metrical prominence, i.e. stress. However, the problem with equating stress with H tone is that Lung’Ie has two kinds of nouns: those with a culminative H and those which are toneless. The nouns with culminative H are 87% of Portuguese origin, incorporated through stress-to-tone alignment, while the toneless ones are 92% of African origin. Although other creole languages have been reported with split systems of “accented” vs. fully specified tonal lexemes, and others with mixed systems of tone and stress, Lung’Ie differs from these cases in treating African origin words as toneless, a quite surprising result. We consider different analyses and conclude that Lung’Ie has a privative /H/ tone system with the single unusual stress-like property of weight-to-tone.
... Príncipe received a small number of prisoners from Bantu regions in contrast to São Tomé that continued to receive a large Bantu input (Hagemeijer 2011). 5 Contrary to Ferraz (1979), who considered both Edoid 6 and Bantu languages to be the main substrates of the Gulf of Guinea creoles, the current view on their history is that languages from the Niger Delta, particularly Edoid languages, were the main substrate, while Bantu languages, such as Kikongo and Kimbundu, played a secondary role (Hagemeijer 2011; see also Agostinho et al. 2019;Ladhams 2012) (see Map 2). Ladhams (2012) shows that Lung'Ie has the lowest percentage of Bantu words and the highest percentage of Edo words compared to the other Gulf of Guinea 5 Genetic studies show a predominance of Benin (52.3%) and Bantu (36.4%) haplotypes in the city of São Tomé (Tomás et al. 2002: 403), which also confirms that "Central-West Africa and the western areas of Bantu-speaking Africa were the major sources of the slave settlers in São Tomé" (Tomás et al. 2002: 407). ...
... 5 Contrary to Ferraz (1979), who considered both Edoid 6 and Bantu languages to be the main substrates of the Gulf of Guinea creoles, the current view on their history is that languages from the Niger Delta, particularly Edoid languages, were the main substrate, while Bantu languages, such as Kikongo and Kimbundu, played a secondary role (Hagemeijer 2011; see also Agostinho et al. 2019;Ladhams 2012) (see Map 2). Ladhams (2012) shows that Lung'Ie has the lowest percentage of Bantu words and the highest percentage of Edo words compared to the other Gulf of Guinea 5 Genetic studies show a predominance of Benin (52.3%) and Bantu (36.4%) haplotypes in the city of São Tomé (Tomás et al. 2002: 403), which also confirms that "Central-West Africa and the western areas of Bantu-speaking Africa were the major sources of the slave settlers in São Tomé" (Tomás et al. 2002: 407). 6 Since Edoid languages were earlier classified as Kwa (vs. ...
... Percentages of words identified as African by source languages from Ladhams (2012) in the Gulf of Guinea creoles: São Tomé: Bantu 56%, Edo 35%, Yoruba 6%, others 3%; Príncipe: ...
Full-text available
Creole languages have generally not figured prominently in cross-linguistic studies of word-prosodic typology. In this paper, we present a phonological analysis of the prosodic system of Lung’Ie or Principense (ISO 639-3 code: pre), a Portuguese-lexifier creole language spoken in São Tomé and Príncipe. Lung’Ie has produced a unique result of the contact between the two different prosodic systems common in creolization: a stress-accent lexifier and tone language substrates. The language has a restrictive privative H/Ø tone system, in which the /H/ is culminative, but non-obligatory. Since rising and falling tones are contrastive on long vowels, the tone must be marked underlyingly. While it is clear that tonal indications are needed, Lung’Ie reveals two properties more expected of an accentual system: (i) there can only be one heavy syllable per word; (ii) this syllable must bear a H tone. This raises the question of whether syllables with a culminative H also have metrical prominence, i.e. stress. However, the problem with equating stress with H tone is that Lung’Ie has two kinds of nouns: those with a culminative H and those which are toneless. The nouns with culminative H are 87% of Portuguese origin, incorporated through stress-to-tone alignment, while the toneless ones are 92% of African origin. Although other creole languages have been reported with split systems of “accented” vs. fully specified tonal lexemes, and others with mixed systems of tone and stress, Lung’Ie differs from these cases in treating African origin words as toneless, a quite surprising result. We consider different analyses and conclude that Lung’Ie has a privative /H/ tone system with the single unusual stress-like property of weight-to-tone.
... Um dos processos tratados pela autora é a aglutinação de vogal, em que artigos definidos de étimo português teriam sido interpretados por aquela comunidade de falantes como parte da raiz do item lexical que o seguia, resultando numa nova palavra prosódica constituída por duas ou três sílabas. Alguns autores já se debruçaram sobre o assunto analisando esse processo fonológico nas línguas-filhas, levantando hipóteses de influência das línguas bantas (SCHANG, 2000), e edoides (LADHAMS, 2007) e decalque do português (HAGEMEIJER, 2009b). Bandeira (2017) levanta a hipótese de que a criação dessas palavras poderia ter alguma motivação prosódica. ...
Full-text available
O contato entre lusofalantes e populações africanas multilíngues escravizadas na ilha de São Tomé, África Ocidental, a partir do final do século XV, implicou a emergência de uma nova língua: o protocrioulo do Golfo da Guiné (PGG). Posteriormente, o PGG deu origem ao santome, lung’Ie, angolar e fa d’Ambô. A partir da hipótese de que a formação do léxico do PGG teria motivação prosódica, investigamos se fatores como eurritmia associado ao peso silábico (GHINI, 1993) estariam envolvidos na formação de palavra prosódica mínima (VIGÁRIO, 2003). Levando em consideração a produtividade de alguns processos fonológicos que contribuíram para a configuração (em número de sílabas) do léxico do PGG, damos atenção especial ao processo de aglutinação de vogal (O pé (português) → *ɔpɛ “pé”). Tal processo consiste na interpretação de artigos definidos de étimo português como parte da raiz do item lexical seguinte, resultando numa nova palavra prosódica, constituída por duas ou três sílabas. Entretanto, nossos resultados mostram que o sistema fonológico e prosódico do PGG possibilitou a constituição de palavras tanto com mais quanto com menos sílabas do que no étimo português participante de sua formação. Assim, as hipóteses levantadas não se sustentam: a aglutinação de vogal inicial, por exemplo, realizou-se em poucos itens do PGG, caracterizando-se como processo marginal na língua. Contudo, é possível que exista uma restrição que evite verbos iniciados por vogais, o que parece dialogar com a restrição encontrada nas línguas edoides (que participaram da formação do PGG), de que palavras nominais sejam iniciadas por segmentos vocálicos.
... Tal afirmação demanda ponderação, posto que, sem um estudo fonológico sobre as línguas edóides, não é possível ter certeza sobre essas supostas peculiaridades linguísticas do grupo edo. Ademais,Ladhams (2007) sugere uma diversidade étnica maior nos primeiros anos formativos do PGG. ...
Full-text available
Este estudo propõe uma análise comparativa das fonologias do santome, lung’Ie, angolar e o fa d’Ambô. A formação dessas línguas autóctones da região insular do golfo da Guiné, na África Ocidental, está diretamente relacionada a uma língua ancestral, o protocrioulo do golfo da Guiné (PGG). Com base no método histórico-comparativo, o presente estudo comparou os sistemas fonológicos das quatro línguas-filhas do PGG. Na comparação dos inventários consonantais, todas as línguas compartilham treze fonemas (/p b t d k g f v l m n w j/) e divergem quanto a treze fonemas (/kp gb s z θ ð ∫ ʒ ʎ r t∫ dʒ ɲ/). Os sistemas de vogais orais simples, por seu turno, convergem em todos os pontos, ao passo que os sistemas de vogais longas – idênticos em fa d’ambô, em lung’ie e em angolar – diferem apenas do santome que não apresenta os referidos segmentos. As divergências entre os quadros de fonemas refletiram os caminhos tomados pelas quatro línguas a partir da especiação do PGG após a separação geográfica de seus falantes e a consequente formação de cada língua.
Full-text available
Resumo: Este trabalho objetiva apresentar uma breve descrição comparativa dos aspectos da não concordância de gênero do sintagma nominal, atestados de forma semelhante em três variedades africanas da língua portuguesa, a saber: o português falado em Angola, especificamente do Município do Libolo; o português falado em São Tomé e Príncipe; e o português falado em Helvécia, na Bahia, Brasil. Trabalhos sobre a concordância nominal para as variedades brasileira e europeia do português apontaram, por exemplo, para uma regra variável para a primeira e categórica para esta última (BRANDÃO; VIEIRA, 2012; BRANDÃO, 2018). Alguns estudos sobre as variedades africanas e afro-brasileira do português já destacam as descrições e análises acerca da concordância nominal em entre outros). A partir da descrição e análise de dados coletados em trabalho de campo, proporemos aproximações entre as três referidas variedades e os possíveis fatores sociais que possam influenciar a não realização dessa concordância nominal, principalmente no que concerne ao contato linguístico, contexto sociohistórico e aprendizado do português como L2 ou como L1 a partir de um modelo de L2. Palavras-chave: Concordância Nominal. Concordância de Gênero. Português de Angola. Português de São Tomé e Príncipe. Português de Helvécia. Contato Linguístico. 1 Os autores estão dispostos em ordem alfabética uma vez que contribuíram igualmente para o desenvolvimento e escrita do texto. Assim, a ordem apresentada não reflete ordem de autoria.
Congruence and convergence are frequently mentioned to explain crosslinguistic phenomena, such as code alternations, borrowings etc., which arise when languages are in contact (Matras 2009). As a rule, these linguistic processes are generated by the coexistence of languages in the same area over a long period of time (hence the creation of sprachbünde), and by situations of language acquisition and bilingualism. These processes are fueled by the acquisition strategies developed by plurilingual speakers (Haspelmath 1998). Congruence and convergence are often used as synonyms to name the linguistic factors that shape the material (linguistic matter) and the modes of organization (patterns) of languages in contact (Matras 2009: 148), but sometimes these terms may refer to distinct linguistic processes. In this contribution, convergence and congruence are discussed in the light of the socio-historical and linguistic development of Creole languages. The meaning and use of these notions are examined through a discussion of the case of nominal agglutination (i.e. the reanalysis of the French Determiner + Noun sequences into new lexical units), presented by Baker (1984) as an example of transfer of a Bantu substratum in the development of Isle de France Creole (i.e. French Creoles from Mauritius, Rodrigues and Seychelles). Since Weinreich (1953) and Haugen (1956), language contacts are known to foster the emergence of diaphonemes and diamorphemes, that is linguistic units which span languages in contact and are perceived by bilingual speakers as identical. These interlingual phenomena bear both on linguistic materials and on modes of organization. The notion of conceptual change of Jarvis and Pavlenko (2008) describes the conditions that trigger these crosslinguistic effects. To pose that the appropriation of the dominant language of a colony by plurilingual subalterns is at the origin of creolization is to accept that the emergence of Creole languages follows a course where transfer, convergence and congruence between languages in contact, and other general learning processes are at play. The crystallization of these learner varieties is made possible because of the development of socialization among the slaves, and of the related sociolinguistic phenomenon of focussing (Lepage & Tabouret-Keller 1985). Gumperz and Wilson (1971) and Myers-Scotton (2002) define the socio-linguistic and cognitive dimensions of convergence while Baptista (2020) uses the notion of congruence to describe the same linguistic phenomena. For these researchers, convergence and congruence, are unitary processes that engender new combinations of units and functions in languages in contact. For other researchers (for instance, Olko, Borges and Sullivan 2018), convergence or congruence is the result of a sum of mechanisms of linguistic change, including grammaticalization, lexical and grammatical borrowing and reanalysis. In the present paper, convergence and congruence are defined as processes based on targeted linguistic transfer which, according to the dynamics of the languages in contact, give rise to reanalysis of the model languages, leading to the creation of new grammatical meanings in the replica languages. If the link is direct between congruence, convergence and reanalysis, i.e. the reinterpretation of the units and categories of the model language, it is indirect vis a vis grammaticalization (Hopper & Closs-Traugott 1993: 220). Convergence and congruence, like other processes, contribute to the development of grammaticalization in emerging languages. Nominal agglutination in French-related Creoles is of two types: a) a consonant is added to a French etymon beginning with an initial vowel; b) a syllable derived from the French determiner (mau. di, li etc.) agglutinates with the French lexical etymon to form new Creole lexemes, examples lera (rat) (from) < (fr. le rat), dizef (eggs) < (fr. de l’œuf), lizje (eyes) < (fr. les yeux) etc. According to Baker (1984), the massive presence of agglutinated nominal forms of the second type in the Creoles of Mauritius, Rodrigues and the Seychelles is due to the presence of Bantuphone slaves in Mauritius during the period of formation of the so-called Isle de France Creole (IdFC), i.e. between 1773 and 1810. Baker (1984: 111) sketches an acquisition scenario whereby the Bantuphone slaves, the majority of the inhabitants of Ile de France at that time, reanalyze the Determiner + Noun sequences of French into unique lexical items because of the transfer of their knowledge of Bantu class markers, leading to a case of isomorphic convergence (Baptista 2020). The paper attempts to show that factors other than convergence and congruence could be at play in the inception of nominal agglutination and more widely of creolization in French Creoles. In the light of the definitions of convergence and congruence proposed, the paper investigates the phenomenon of lexical reanalysis called nominal agglutination, keeping in mind that transfer presupposes the identification of similar sites in the target language and that any isomorphic convergence or any congruence implies that identities have been perceived between the languages in contact. The paper claims that nominal agglutination in French Creoles cannot be considered to constitute a strict case of convergence or congruence in spite of Baker (1984), Myers-Scotton (2002) and Baptista (2020)’s analyses. A cursory examination of the system of class markers in Bantu languages shows that the nominal morphology of these languages is different from that of French. The nominal classes of the Bantu languages and the French NP and its agreement specifiers are organized differently on the morphological and semantic levels. Of course, speakers of Bantu languages tend to be sensitive to prefixes in the model NP and tend to add augments to the substantival base of the model language in certain contexts. However, these arguments are not sufficient to prove that nominal agglutination in French-related Creoles a case of congruence or convergence. Nominal agglutination in French Creoles has been and is produced by multiple factors, from the conservation and the intra-systemic generalization of regional uses of French to possible partial transfers between the languages of the slave populations and colonial French during creolization. Understanding creolization as a case of acquisition of the dominant language of the colony by multilingual adults presupposes an understanding of the conditions that govern the mobilization of convergence or congruence and other linguistic and cognitive mechanisms of transfer, analogy, reanalysis, convergence and congruence. Sociohistorical factors are also to be taken into account in creolization because they shape the very nature of inter-linguistic contacts.
Full-text available
The Proto-Creole of the Gulf of Guinea is the common ancestor of the Santome, Angolar, Lung’Ie and Fa d’Ambô languages. Based on a supposed occurrence of vowel harmony (vh) in these modern languages and the proven influence of the languages of the Niger Delta (HAGEMEIJER, 2009), in which vh is attested, this study aims to discuss the existence of vh processes with [atr] mid-vowels in the Proto-Creole of the Gulf of Guinea, considering as a starting point its phonological and lexical reconstruction (BANDEIRA, 2017). Therefore, we intend to verify what phonological parameters related to vh in the literature (trigger, target, domain, application direction and blocking element) indicate about vh processes in Proto-Creole, based on the lexical reflexes of a set of cognates in daughter languages and proto-forms. We will show that many of the items interpreted as harmonic derive from the maintenance of mid-vowels in Portuguese etyma and from the insertion of a vowel copy in order to avoid roots ending in consonants in the Proto-Creole of the Gulf of Guinea. In this analysis, we will argue for the hypothesis according to which parasitic vh processes (COLE; TRIGO, 1988) with [atr] mid-vowels in contiguous syllables occurred in Proto-Creole within the prosodic word domain. However, the vh was limited because, based on the observation of proto-forms and harmonisation patterns with [atr] mid-vowels in contiguous syllables, there were many cases in which there was neutralisation of final non-stressed mid-vowels and, therefore, no application of the harmony process in favouring contexts.
This chapter describes the history of French Guiana during the time when French Guianese Creole emerged (1660–1700). After providing background information about the Native American, French and African peoples who lived in French Guiana, this chapter explores the French Caribbean colonial system founded in the Antilles. It then focuses on the linguistic experience of African slaves in French Guiana’s capital Cayenne, drawing on a plantation archive to describe daily life in the colony and providing evidence that, unusually for a slave colony, almost all the first slaves came from a single region of Africa. This chapter also provides a likely pathway for the invention of French Guianese Creole as it evolved from basic pidgin to the dominant language of the community.
Full-text available
The general aim of the thesis is to assess the circumstances of the origin of Portuguese plantation creole languages, from both linguistic and extra-linguistic perspectives, by studying the Creoles in the Cape Verde Islands and the Gulf of Guinea Islands (São Tomé, Príncipe and Annobon), as compared with the situation in Brazil, where it would appear that no creole language arose. This is despite the fact that similar sociohistorical conditions existed in all three territories, from the point of view of plantation settlement and slavery, and at roughly the same period. The thesis attempts to answer not so much how, but more particularly why plantation Creoles are formed and develop, in particular sociolinguistic circumstances. A linguistic analysis is made of the extent to which the plantation creoles developed away from 16 th-century Portuguese, and a comparison made between the individual languages as to how "radical" they are. An examination is also made of the origin of the slaves involved in the formation of the Creole languages, to assess the proportions of slaves from different ethnic groups, and also to see what linguistic influences can be traced in the actual Creole languages. A number of theories have been proposed which regard the development of creole languages as being primarily European-based. However, none of these theories would appear to be appropriate for the case of the Portuguese plantation creoles. On the other hand, evidence is produced showing the connection between community formation and language formation, as an indication of solidarity within that community. A number of aspects of community formation are examined in relation to the Portuguese plantation islands and the relationship established between these aspects and the fact that creole languages were formed there. By contrast, it is shown that in Brazil, where creolisation almost certainly did not take place, there is very little evidence of similar community formation, thereby reinforcing the connection between linguistic and extra-linguistic factors in the formation of creole languages.
Full-text available
According to Ehrhart and Corne, Tayo is an endogenous creole that crystallized under the peculiarly plantation-like circumstances present at the St-Louis mission in the late 19th century. Noting some linguistic similarities with Reunion Creole, Chaudenson (1994) raises the question of whether Reunion Creole had had any influence on the development of Tayo. This notion is refuted both by Ehrhart (1994) and Corne (1994, 1995, 1999, 2000a, 2000b), although Corne (2000a) concedes that due to some linguistic and socio-demographic evidence, Reunion Creole influence on Tayo cannot be excluded. This paper revisits this debate and reopens questions that earlier researchers appear to have closed by discussing the implications of two texts written in Reunion Creole and published in New Caledonia. The first is a Georges Baudoux text containing the ‘Reunion Creole’ of Socrates, a black Reunion Creole taken to New Caledonia in 1870 to work as a coolie. The second is a political text attacking a ‘Creole’ candidate running for election on the Conseil Supérieur des Colonies published in 1884 by journalist Julien Bernier, an immigrant from Reunion. Accepting the authenticity of these texts raises questions pertinent to the debate on Tayo genesis. Given that réunionnais was being spoken in New Caledonia when Tayo was developing, were any speakers in contact with the Kanaks of St-Louis? What, if any, influence did their language have on the developing St-Louis patois? I discuss these questions by re-examining socio-historical evidence and by making some brief comparisons between the New Caledonian Reunion Creole texts and Tayo.
Texte intégral accessible uniquement aux membres de l'Université de Lorraine