Patterns of Settlement

THE FORMATION OF THE PORTUGUESE PLANTATION CREOLES

JOHN LADHAMS

A thesis submitted in partial fulfilment of the requirements of the University of Westminster for the degree of Doctor of Philosophy

January 2003

THE FORMATION OF THE PORTUGUESE PLANTATION CREOLES

JOHN LADHAMS

2

ABSTRACT 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 16th-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.

3

CONTENTS

1.

Introduction and background

1

2.

3.

4.

1.1

Introduction

2

1.2

Geographical background

6

1.3

Linguistic background

9

1.4

Historical background: patterns of settlement

16

The Creole languages. Descriptive analysis

34

2.1

Introduction

35

2.2

Phonology

36

2.3

Lexis and semantics

63

2.4

Morphology and syntax

70

Linguistic factors I. African origins and influence

113

3.1

Introduction

114

3.2

Cape Verde

116

3.3

Gulf of Guinea

139

3.4

„Substrate‟ or „universal‟

151

Linguistic factors II. The European perspective 153 4.1

Introduction

154

4.2

„Ladinização‟

156

4.3

The „Reconnaissance Language‟ and pidgin diffusion

165

4.4

„National identity‟

167

4

4.5

5.

6.

7.

Conclusions

169

Extra-linguistic factors

172

5.1

Introduction

173

5.2

Demographic ratios

175

5.3

Family formation and „nativisation‟

179

5.4

Miscegenation

183

5.5

Manumission

188

5.6

Maroon slaves

191

5.7

The „peasant breach‟ and community formation

196

5.8

Conclusions

203

Brazil

206

6.1

Introduction and background

207

6.2

Brazilian Portuguese: descriptive analysis

214

6.3

Linguistic factors I: African origins and influence

219

6.4

Linguistic factors II: the European perspective

221

6.5

Extra-linguistic factors

223

Conclusions. The sociolinguistic perspective

229

7.1

Portuguese Plantation Creole formation: the evidence

230

7.2

Creole typology

232

7.3

Sociolinguistic theory

233

7.4

Creole formation and cultural dynamics

234

References

235

5

1 INTRODUCTION AND BACKGROUND

6

CALIBAN You taught me language, and my profit on‟t Is, I know how to curse. The red plague rid you For learning me your language! William Shakespeare - The Tempest: Act I, Scene 2.

1.1 Introduction The study of pidgin and creole languages is no longer in its infancy, having been pioneered more than a century ago by such linguists as Adolpho Coelho in Portugal, Hugo Schuchardt in Austria and Dirk Christian Hesseling in the Netherlands. On the other hand, it is only in the last 30 years or so that the modern „scientific‟ study of these languages has taken off, not without its own polemics, controversies, and problems as yet unresolved. If there has been a single constant theme in this field since the late 1960s, it is the attempt to explain the paradoxical similarity, in formal terms, between languages geographically and sociohistorically entirely disparate. One aspect that has preoccupied creolists (as these linguists can be legitimately called) is the mechanics of the formation of creole languages, and to a lesser extent, pidgins - in other words, precisely how such languages arose, and subsequently developed.

There are probably three main

perspectives on this problem: often referred to as the „superstratist‟, „substratist‟ and „universalist‟ theories. The first of these, nowadays increasingly less enthusiastically defended, suggests that in a case of language contact, creoles are merely the result of the evolution of the so-called „superstrate‟ language (normally European), brought about by not only the „superstrate‟ but also the „substrate‟ speakers.1 The substratists, on the other hand, simply reverse the perspective, and place the onus on the substrate speakers (normally African), who retain significant amounts of their own language(s) while shifting towards the superstrate. Meanwhile, the universalists, and particularly Derek

1

The terms „substrate‟, and for that matter „superstrate‟ are used here for practical convenience, and are not intended to have derogatory implications in terms of superiority or inferiority. With the term „adstrate‟, this problem does not necessarily arise. I shall continue to use inverted commas with the other two terms.

7

Bickerton, whose Language Bioprogram Hypothesis2 has revolutionised creole studies, claim that the innate universals of first- and/or second-language acquisition come in to play with the formation of creoles, and neither the substrate nor even the superstrate languages have any significant role in this process. However, where all three theories are arguably wrong is that they all refer to the superstrate as being a „target language‟, as if the non-speakers were inevitably seeking to acquire that language, and in some way failing to do so. This study wishes to show that if it can be demonstrated that those who formed a creole language in the first place were attempting to create their own means of communication, as a separate community, then there is room for a fourth, alternative, theory of creole formation, which would place the emphasis on the actual speakers and their motivation to create a language of their own. As a means to achieving this aim, I consider more the reasons why creoles have been formed, and not so much how this process took place. Indeed, even though the „how‟ and the „why‟ do go hand in hand, motivation is arguably more important for scientific study than the actual mechanics. This thesis takes the form of comparative case studies: the formation of Portuguese-based creole languages on the archipelagos of the Cape Verdes and the Gulf of Guinea Islands, off the West coast of Africa, as compared with the similar sociohistorical conditions in Brazil, where, I believe, no creole language arose. The Portuguese-based creoles are of considerable interest since they were probably the first to be formed with a European language input; on the other hand, there is one major drawback in examining these particular languages, namely that there are no early texts or descriptions available as evidence of the development of these creoles. Previous studies carried out from a similar perspective - for example, by Philip Baker on Mauritian Creole3, and by Jacques Arends on Sranan Creole in Surinam4 - have been able to draw on just such evidence in order to show that, in those cases at least, creolisation was a gradual process. Nevertheless, what can be demonstrated with the Portuguese plantation creoles is that the slaves brought to those islands soon formed a recognisably separate community, and one of the principal manifestations of their „acculturation‟ was clearly 2

See Bickerton (1981; 1984). See particularly Baker (1982) and Baker & Corne (1982). 4 See, for example, Arends (1989). 3

8

the creation of a language of their own. However, the term „acculturation‟ often implies that a community is adapting itself towards another culture: what I am suggesting here is that these communities formed, or created, their own culture, as a sign of individual identity and community solidarity. It has been customary, at least since Bickerton (1984), to classify creole languages in three groups: plantation creoles; fort creoles; and maroon creoles.5 Plantation creoles, examples of which are examined in this study, arose where African slaves from two or more ethnic groups and with mutually unintelligible languages were imported into European-created plantation complexes on islands off the coast of Africa or in the Caribbean, for example. Fort creoles were less frequent, and originated in and around European fortresses in Africa and Asia, manned by slaves from different ethnic and linguistic backgrounds, but who remained in contact with their autochthonous cultures, and the so-called adstrate languages. Examples include Guinea-Bissau Creole in West Africa,6 and the varieties of Indo-Portuguese in Asia. Maroon creoles are the result of runaway slaves, normally from a plantation context, forming a separate and distinct cultural community, deliberately severing all connection with the plantation culture and society. Examples of this kind of creole include Saramaccan and Ndyuka in Surinam, and Angolar, on the Island of São Tomé, in the Gulf of Guinea.7 Whereas the origins and development of both fort and maroon creoles are relatively transparent, the environment of plantation creoles has led many creolists to assume that they did indeed arise form being imposed from the top, as it were, conditions on the plantation being as harsh as they were. However, I hope to show that the plantation creoles were „bottom-up‟ creations, both deliberate and motivated, precisely to emphasise a separate identity. Interestingly, it is clear that this did not inevitably happen, and the situation in Brazil is a case in point.

5

However, this would seem to omit a number of creoles such as those in the Anglophone Pacific, and those in mainland Africa, such as Sango and Kituba. 6 Since Guinea-Bissau Creole is a fort creole and not a plantation one, it is not covered in detail in this study. However, although the sociohistorical setting is different from the latter, there are nevertheless certain affinities with Cape Verdean Creole, and subsequent reference will occasionally be made to Guinea-Bissau Creole in this thesis. 7 Angolar is not strictly speaking examined in this context, since it arose under different circumstances, though it will be referred to for comparative purposes.

9

Just why no sense of community solidarity - and its linguistic counterpart - did not arise will be examined below in Chapter 6. This thesis is organised as follows: the remainder of this chapter offers an overview of the geographical background to the plantation creoles, a brief introduction to the languages themselves, and the historical context, in terms of the patterns of settlement on the islands. Chapter 2 develops the comparative description of the four creoles, pointing out both similarities and differences, and establishing how far they are removed from European Portuguese. Chapter 3 examines the probable origins of the slaves, and consequently the possible linguistic sources of any „substrate‟ influence in the creoles. In Chapter 4, three theories of creole formation, based on a European perspective, are examined in the context of the Portuguese plantation creoles, and subsequently rejected as being inappropriate in these particular cases. Chapter 5 considers the demographic and sociohistorical context in the plantation complexes in question, and suggests that, though the demographic factors often used to explain creole formation are indeed significant, other factors indicative of community formation are arguably much more important. By way of contrast, Chapter 6 describes the linguistic and sociohistorical situation in Brazil, where it is clear that a similar plantation environment did not give rise to the creolisation of Portuguese. In Chapter 7, conclusions are drawn, in part using sociolinguistic models which have until now been applied synchronically: by applying them diachronically to this context, one can show that creolisation can be seen as an active, dynamic process, which far from being targetted at acquiring an alien culture, is motivated to create a language and a culture, belonging to - and instigated by - the apparently subjugated slave community.

10

1.2 Geographical Background 1.2.1 Cape Verde Islands The ten major islands of the Cape Verde archipelago, only one of which is uninhabited, are situated between 450 and 500 kilometres to the West of Cap Vert, the most westerly point of the African mainland, in present-day Senegal. The islands are normally subdivided into two groups: Barlavento (Windward), consisting of the northernmost 6 islands of Santo Antão, São Vicente, Santa Luzia (uninhabited), São Nicolau, Sal and Boavista; and Sotavento (Leeward), with Maio, Santiago, Fogo and Brava. The total land area is a little over 4,000 sq.km.; the largest island, Santiago, is 991 sq.km. in area, followed by Santo Antão (779 sq.km.), while the smallest inhabited island, Brava, has 64 sq.km. The climate is tropical, with very low rainfall and frequent prolonged periods of drought; for this reason the Cape Verde Islands are often referred to as the westward extension of the Sahel. Except on Santiago, there is very little vegetation apart from small scrubland plants, and a dearth of fresh water. Several of the islands, including Santiago and Fogo, are mountainous, reaching almost 3,000 metres at the active volcano on Fogo. The current population of the 9 inhabited islands is around 400,000; Santiago has almost half that number, while São Vicente and Santo Antão each have around 50,000. The largest towns are Praia, the capital, on Santiago island, and Mindelo, on São Vicente, each with around 45,000 inhabitants. It is significant that there are believed to be some 600,000 Cape Verdeans resident abroad - i.e. more than the population of the islands themselves -, particularly in the following communities: Lisbon (Portugal), Rotterdam (Netherlands) and S.E. Massachusetts (United States).8

11

1.2.2 Gulf of Guinea Islands

The four Gulf of Guinea Islands are, from North to South, Fernando Po (also known as Bioko), São Tomé, Príncipe and Annobon. The largest of the four, and closest to the African mainland, is Fernando Po, which is not included in this study, since it has a very different social and linguistic history. São Tomé and Príncipe form an independent republic, while Annobon is part of the Republic of Equatorial Guinea (formerly Spanish Guinea).9

1.2.2.1 São Tomé The island of São Tomé lies approximately 300 km. from the African mainland, to the West of Gabon; the Equator passes through the small uninhabited island of Rolas, at the southern tip of São Tomé. The main island has an area of 845 sq.km. São Tomé has a tropical climate, with very high rainfall, and is covered by dense rainforest; it is also extremely mountainous, reaching more than 2,000 metres at the highest peak. The current population is of around 120,000, more than one third of whom live in the capital, São Tomé.

1.2.2.2 Príncipe Príncipe lies some 160 km. North-East of São Tomé, and 100 km. West of mainland Equatorial Guinea. It has an area of 142 sq.km. The climate and vegetation are very similar to those of São Tomé; the highest mountain is almost 1,000 metres high. The current population is around 5,000, a fair proportion of whom live in the only town, Santo António.

8

For further geographical information on Cape Verde, see Amaral (1991). Further geographical information on all 4 Gulf of Guinea Islands is to be found in Tenreiro (1961:13-56); also Hodges & Newitt (1988:1-16). 9

12

1.2.2.3 Annobon The island of Annobon is considerably smaller than either São Tomé or Príncipe, with an area of only 18 sq.km. It is situated more than 200 km. South-West of São Tomé and 400 km. West of southern Gabon. The highest point is some 800 metres. The population is estimated to be around 2,000, all of whom reside in its only settlement, Pale.

13

1.3 Linguistic Background 1.3.1 Cape Verdean Creole The earliest clear evidence of the Cape Verdean Creole dates only from 1672: John Fryer, in his New Account of East India and Persia, refers to the language of the people of the Cape Verdes as follows: “their Speech is broken Portugal, as also is their Habit, imitating therein the Portugals” (Fryer 1909:45). However, earlier in the same century (1627), Father Alonso de Sandoval, writing of Cartagena, on the Caribbean coast of South America, wrote that three types of slave were transported from Cape Verde: the bozales (uneducated), who came directly from the coast of the mainland; the ladinos (educated), who spoke Portuguese and were called criollos, not because they were born in Cape Verde but because they were brought up there from an early age, having been born bozales; and the naturales, who were born and baptised in Cape Verde (Sandoval 1987:139)10. The question remains, however, as to whether the ladinos, who called themselves criollos, spoke metropolitan Portuguese or a creolised version. The issue is further complicated by the fact that Sandoval goes on to refer to the slaves coming from São Tomé as speaking a corrupt version of Portuguese (see 1.3.2.1 below). Irrespective of this, António Carreira has examined all the available documentation, and concluded that, assuming what is now Guinea-Bissau Creole was taken there from Cape Verde, and not vice versa, it is probable that Cape Verdean Creole had originated by 1550 (Carreira 1972:337-344; and, from a slightly different perspective, 1983). Also, as Dulce Fanha has pointed out (1988:293), the proportions of slaves to white Europeans would have meant that abrupt creolisation almost certainly took place at an early stage in the Cape Verdes. However, the figures she quotes - 13,700 slaves and 100 or so whites - date from 1582, more than a century after initial settlement. Nevertheless, although there are no earlier population figures for Cape Verde (see below, 2.2), she reasonably assumes that slaves outnumbered whites by a large proportion considerably earlier than the 1580s.

10

“Suelen de esta isla venir tres suertes de Negros, (...); unos bozales, al modo de los que traen de Cacheo: otros ladinos, que hablan lengua portuguesa y llaman criollos, no porque ayan nacido en Cabo Verde, sino porque se criaron desde pequeños alli, aviendo llegado bozales(...). Otros llaman naturales y son nacidos y criados en la mesma isla de Cabo Verde y bautizados niðos.”

14

By the 19th century, two varieties of Cape Verdean Creole were becoming discernable: Sotavento and Barlavento, corresponding to the southern and northern island groupings respectively.

It is relatively clear that since the former is plainly more

basilectal (i.e. more „radical‟ or differing more from European Portuguese) than the latter, Barlavento creole is a later development, based on Sotavento creole, heavily influenced by European Portuguese. This followed the rapid increase in the mixed population of the Barlavento islands from the early 19th century onwards, particularly with the arrival of relatively large numbers of (white) Portuguese convicts. Until that time, the population of the northern islands had been very small - for the first century and a half of settlement there had been virtually no permanent residents (see 1.4.1 below).11 On the other hand, there are discernable differences between the various islands within the two groupings. Bartens-Adawomu (1999) even goes as far as to imply that each inhabited island has a separate variety, but this is certainly overstating the case; Veiga (1982) makes a detailed comparative description of all the varieties of Cape Verdean Creole, and comes to the conclusion that the only valid differentiation would be between Sotavento and Barlavento.12

In this study, emphasis is laid on the basilectal Sotavento variety,

particularly that of Santiago, which is often taken as a local standard. The several varieties of Cape Verdean Creole are spoken as a first language by virtually all the 400,000 inhabitants of the islands, even though the country‟s official language is still Portuguese. Nevertheless, Creole has a relatively high prestige in the country: there is some publication in Creole, including works of literature and some journalism, while on the broadcast media, the language is used extensively. There are also some moves to introduce the language as a medium for primary education. It should also be remembered that the 600,000 or so Cape Verdean emigrants are mostly firstlanguage speakers of the Creole.

11

Whether this constitutes a case of decreolisation is a moot point. My own opinion is that Barlavento Creole is a later development from the basilectal variety in Sotavento, but the shift towards standard European Portuguese was brought about probably by the Europeans themselves, rather than the original local speakers of the Creole. 12 Lopes da Silva (1957) is largely of the same opinion, while Cardoso (1989) notes some minor differences in the São Nicolau variety of Barlavento creole; both authors are native speakers of Cape Verdean Creole, as is Veiga, while Bartens-Adawomu is not.

15

The earliest, pioneering, descriptions of Cape Verdean Creole are by Costa & Duarte (1886) and Schuchardt (1888), while in this century there are the books by Lopes da Silva (1957), Almada (1961) and Veiga (1982), the latter being written in the Creole itself. There are also descriptions of the varieties on Brava (Meintel 1975) and on São Nicolau (Cardoso 1989).

1.3.2 Gulf of Guinea Creoles On the three Gulf of Guinea islands of São Tomé, Príncipe and Annobon, there are in fact four creole languages: one each on Príncipe and Annobon, and on São Tomé, São Tomé Creole and Angolar.13 Traditionally, the origins of this creole are said to lie with a shipwreck of Angolan slaves on the coast of São Tomé Island some time in the 16th century, but it is also clear that a significant proportion of the Angolar community (in the interior and on the South coast of the island) were originally runaway slaves who escaped from the sugar plantations also in the 16th century. Ferraz has argued that Angolar would represent the language closest to a proto-Gulf of Guinea Creole, since it had little or no contact with metropolitan Portuguese from the 16th century onwards (1979:9; 1983:122), but as will be seen, the same could also be said of Annobonese. Ferraz has also shown that the African substrate influence in all the Gulf of Guinea creoles is from not only Bantu languages, but also from Edo (also called Bini), spoken in Benin, in the Niger delta of present-day Nigeria (Ferraz 1979:12-13). Since it is also known that slaves were not exported from Benin to the islands after 1553 (Ryder 1969:74), then this shows that creolisation must have at least begun before that date. 1.3.2.1 São Tomé (local language name - „forro/folo‟) The earliest clear reference to São Tomé Creole is as follows - indirect evidence from Fr. Sandoval in Cartagena (see 1.3.1 above), who wrote in 1627:

13

Although Angolar, being a maroon creole, is not covered by this study, a brief outline is included below (in 1.3.2.4).

16

Y los que llamamos criollos y naturales de S. Thome (...) las entienden casi todas con un genero de lenguage muy corrupto y revesado de la Portuguesa, que llaman lengua de S. Thome. (Sandoval 1987:140).

[And those we call criollos and naturales of São Tomé (...) almost all communicate in a kind of very corrupt language distorted from Portuguese, which they call S. Tomé language.]14

Only much later, in 1766, is there a direct reference, in a report by Gaspar Pinheiro da Câmara,15 who wrote:

Hé de saber, que a gente natural destas ilhas tem lingoa sua e completa, com prenuncia labeal, mas de que me não consta haver inscripção alguma; e hé certo que todos sabem falar a portuguesa, não sendo negros do mato, ou novamente regatados, alem dos muitos que falão a lingoa franca, ao

menos na parte que

baste para o comercio com os estrangeiros. (Neves 1989:230).

[One should know that the people native to these islands have a complete language of their own, pronounced labially, but which I understand has never been written down; and it is certain that they can all speak Portuguese, except

the

negros in the bush, or newly traded, as well as the many who speak the lingua franca, at least that much which suffices for trade with foreigners.]

São Tomé Creole is currently spoken as a first language by at least 100,000 of the approximately 120,000 inhabitants of the island. However, São Tomé Creole by no means carries great prestige, a privilege accorded to the official language, Portuguese, which is used in all public contexts, including education and journalism. Nevertheless, the Creole is certainly not an endangered language, even though it is limited to the domestic and community domains.

14

Unless otherwise indicated, all translations are mine.

17

The earliest (brief) description of São Tomé Creole is to be found in Hugo Schuchardt‟s first article on creole languages (1882), based as always on information sent to him by correspondents; this was followed by Almada Negreiros (1895:303-369), which includes a grammatical outline, and a relatively extensive Portuguese-Creole wordlist. A partial description of the language is included in Valkhoff (1966:77-115), together with some information on Príncipe and Annobon Creoles, but the only modern detailed description is Ferraz (1979).

1.3.2.2 Príncipe (local language name - „lunge iye‟) Apart from Gaspar Pinheiro da Câmara‟s reference to “these islands” in 1766, quoted above (in 1.3.2.1), there is no clear historical evidence of Príncipe Creole. However, as Ferraz has also shown, all the Gulf of Guinea Creoles are clear examples of abrupt creolisation, since not only do they retain a number of Portuguese archaisms (1983:122), but in the case of Príncipe - and also Annobon -, their relative isolation from the onset of colonisation would have meant that there was little or no influence from metropolitan Portuguese. There is some disagreement as to the current status of the language: Valkhoff (1966: 85) said that he had difficulty in finding speakers in his albeit limited field research in 1958. Günther (1973:17), on the basis of his research some 10 years later, also suggested that it was an endangered language.

However, Ferraz (1979) disagrees,

although the number of fluent speakers he encountered was small. In my own experience, during my visits to the island in 1988-1989, the language is not widely spoken, and suffers from low prestige. In any case, the island‟s population of 5,000 is often fluent in São Tomé Creole, especially with increased communication between the two islands since independence in 1975. Nevertheless, the official language, as used in primary education on the island, for example, is still Portuguese. The only descriptions of Príncipe Creole to date are Schuchardt‟s brief sketch (1889), some references in Valkhoff (1966:77-115), and the single book-length account which is Günther (1973), which includes a Creole-German wordlist. 15

AHU, S. Tomé, Caixa 10, doc. 93; Neves (1989:228-231).

18

1.3.2.3 Annobon (local name - „fa d‟ambu‟) In the case of early references to Annobon Creole, when the Spanish Capuchin friar Juan de Santiago landed on the island in 1649, he found just one white Portuguese overseer and some 500 slaves: Todos entendian y hablaban la lengua portuguesa aunque imperfectam.te pero de suerte que sin muçha dificultad se dexaban entender los que menos la sabian. (Bal 1975:121).

[They all understood and spoke the Portuguese language albeit imperfectly but in such a way that without much difficulty they could be understood by those who knew it less.]

The current sociolinguistic situation of Annobon Creole has been described in some detail by De Granda (1985:141-156). He distinguishes four different codes among the population of 2,000 or so: Annobon Creole, spoken as a first language by all; Spanish, the official language of Equatorial Guinea, with only 2 native speakers in 1985 (a priest and a doctor), though the language is used by local administrators and in (primary) education on the island, and has been acquired by a relatively large proportion of the male population as a secondary language, through education and/or emigration to elsewhere in Equatorial Guinea; Pidgin English, in the variety widely used on Fernando Po, has been acquired by those men who have returned from emigration there, and is regarded with considerable status as a medium for community solidarity; the fourth code, perhaps surprisingly, is Portuguese, which has been retained for liturgical use in the form of memorised texts probably since the eighteenth century, the last occasion when Portuguese priests were present on the island. Otherwise, Portuguese is neither used nor understood on Annobon.

19

1.3.2.4 Angolar The origins of the Angolar community on the island of São Tomé are not entirely clear. Traditionally (i.e. for example in Rosário Pinto‟s history of the island dating from 173416), they were the survivors of the wreck of a slave ship some time in the 16th century on the uninhabited southern coast of São Tomé, and that they escaped into the dense and mountainous forest. Certainly there is documentary evidence of raids on the town of São Tomé by runaway slaves, who were probably those known as Angolares. At that time, as well as later, their numbers were increased considerably by further maroon slaves, until the community eventually accomodated itself to peaceful relations with the rest of the island‟s population in the late 17th century.17 The language of the Angolares, nowadays spoken by some 5,000 people, differs somewhat from São Tomé Creole, although the two languages are normally mutually comprehensible, the main difference being in a much greater influence in Angolar from the Angolan language Kimbundu.18 The social prestige of the Angolar language is low, in relation to non-speakers on the island, though as Lorenzino has pointed out: “for Angolares (...), language still signals group solidarity.” (1998:45). The main descriptions of the language are Maurer (1995) and Lorenzino (1998). Since Angolar is clearly a maroon, and not a plantation creole, it is not examined in any depth in this study, though reference to it will occasionally be made for comparative purposes.

16

Edited by Ambrósio (1970). For further details, see Castelo-Branco (1971); Ferraz (1974:179-181); Garfield (1992:76-79); Lorenzino (1998: Chapter 2). 18 See Maurer (1992). 17

20

1.4 Historical Background: Patterns of Settlement 1.4.1 Cape Verde Two out of the ten islands of the Cape Verde archipelago - Santiago and Fogo were discovered, uninhabited, by the Portuguese in all probability in 1460 (cf. Albuquerque 1991, for a detailed examination of

the somewhat contradictory

documentary evidence).19 A grant of ownership, subsequently divided into 2 areas (captaincies), was issued by the Portuguese King Afonso V to his brother D. Fernando in 1462, with the idea that the islands could be colonised and planted in a similar way to the previously settled islands of Madeira and the Azores, where there was already a very successful sugar plantation industry.20 That settlement went ahead is clear from the Letter of Privileges issued by the King in 1466.21 However, the wideranging incentives offered to would-be settlers, “because it [Santiago] is so remote from our kingdoms” (Blake 1942:64), also show that colonisers needed considerable encouragement.

These

incentives included freedom to trade on the mainland, without having to obtain a licence, except at the trading post of Arguin, and only limited taxes were to be paid on imported goods, including slaves, which could be resold within or outside the Cape Verde Islands. Six years later, in 1472, on the pretext of safeguarding the exclusive trading rights on the Guinea coast granted in 1468 or 1469 to Fernão Gomes, another Royal Letter was issued severely restricting the privileges granted in 1466.22 The main stipulations were that only goods originating in Santiago could be traded on the Guinea coast and only residents of the island were permitted to take part in that trade. As a result, it is clear that the first Cape Verdean island to be settled, Santiago, together with the neighbouring island of Fogo, to a much smaller extent, were planted as from about 1472 with subsistence crops such as maize and sorghum, as well as cotton as a cash crop (Carreira

19

Some earlier historians suggested that at least one of the islands was in fact inhabited by Wolofs from the West African mainland, but there is no evidence to support this allegation. See Carreira (1972:294-303); Baleno (1991:125-127). 20 ANTT, Chancelaria D. Afonso V, livro 1, fl. 61; 19 September, 1462. Brásio (1958:415-416); Albuquerque & Santos (1988:17-18). See Carreira (1972:19-21). 21 ANTT, Chancelaria D. Afonso V, livro 14, fl. 104; 12 June, 1466. Brásio (1958:431-435); Albuquerque & Santos (1988:19-22); English translation in Blake (1942:64-67). See Carreira (1972:22-24); Baleno (1991:127,130).

21

1972:32-33). This latter product was extremely useful for the Portuguese, as woven cloth was used extensively for bartering on the Guinea Coast, as well as for the domestic market in Europe. Cotton is not a particularly labour-intensive crop, at least as compared with sugar; nevertheless, according to António Carreira, slaves were brought across to the Cape Verdes at an early stage from the neighbouring African coast, in considerable numbers:

Divididas as àreas das duas donatarias, e determinado o povoamento das ilhas, em especial a de Santiago, os poucos brancos tiveram de recrutar mão-de-obra no continente fronteiro, porque não possuiam condições de resistência ao clima para cultivar os campos, apascentar o gado e fazer tudo o mais que a colonização exigia. Daí o terem começado logo a trazer escravos da costa e rios de Guiné, numa primeira fase mais para o povoamento e garantia da exploração da terra. (Carreira 1972:286-7). [Having separated the two areas of the grants of ownership, and having decided upon settlement of the islands, particularly of Santiago, the small white population had to recruit manpower from the neighbouring mainland, because they were unable to resist the climate for growing crops, grazing livestock and doing everything else that colonisation required. Hence they immediately began to bring slaves from the coast and Rivers of Guinea, at this early stage more for settlement and as a guarantee for exploiting the land.]

More recently this assessment has been challenged by Ilídio Baleno, who suggests that:

Com todo o respeito que temos pela obra desse ilustre investigador da história de Cabo Verde, temos de dizer, no entanto, que as opiniões emitidas neste exerto nos parecem bastante discutíveis. Só a partir de 1472, com a necessidade de intensificar a exploração de algodão, se verificou uma considerável concentração 22

ANTT, Livro das Ilhas, fl. 2v-4; 8 February, 1472. Brásio (1958:446-450); Albuquerque & Santos (1988:25-28). See Carreira (1972:30-32); Baleno (1991:131-132).

22

de braços nas ilhas de Santiago e Fogo. Os brancos não contavam com um efectivo capaz de responder às novas necessidades do tráfico.

A solução

encontrada foi a utilização da mão-de-obra escrava, barata e praticamente acessível a todos. Antes disso (e durante muito tempo assim seria), o escravo constituía uma mercadoria para exportação muito mais do que mão-de-obra para o arquipélago. (Baleno 1991:156).

[With all due respect to the work of this illustrious historian of Cape Verde [Carreira],we must nevertheless say that the opinions quoted above are quite debatable. Only after 1472, with the need to intensify the cultivation of cotton, was there a reasonable concentration of labour on the islands of Santiago and Fogo. The whites had no effective tool to respond to the new requirements of the slave trade. The solution they found was the use of slave labour, which was cheap and accessible to practically everyone. Prior to this (and for a considerable time it was to be the case), slaves constituted merchandise for export much more than as manpower for the archipelago.]

Carreira had mentioned livestock, and from at least the beginning of the 16th century, there is documentary evidence of livestock-raising on the main islands of Santiago and Fogo, as well as goats being left to pasture on the otherwise uninhabited islands of Brava, São Nicolau and Santo Antão. One document dated July 1504 refers to livestock being taken to these islands, together with a limited number of slaves, an overseer, a few rudimentary buildings and the means to slaughter the said livestock.23 In his lengthy description of the West African coast and Atlantic islands, believed to date from around 1507, Valentim Fernandes refers to the islands of Boa Vista, Sal, Maio, Brava, São Nicolau, São Vicente and Santo Antão as being uninhabited, but did have „goats‟ (Fernandes 1940:119-121). In other words, slaves, in very small numbers, were probably sent to the outlying islands on a strictly seasonal basis; on the other hand, António Correia e Silva surmises that there were one or two slaves on these islands

23

throughout the year, to be supplemented by additional labour when required (1991:213214). In any case, even on the inhabited islands of Santiago and Fogo, livestock raising did not require much in the way of slave labour. Towards the end of the 16th century, the island colony had developed a dual but integrated economy: in the towns of Ribeira Grande and Praia on Santiago Island, it was a mercantile economy, involved in slave trading, principally for re-export to the Americas, while in the rural areas, still only on two islands, Santiago and Fogo, there was mixed agriculture, on the one hand crops for local subsistence and supplies for shipping, with such products as maize, millet, vegetables and fruit, and on the other hand cotton, and to a smaller extent livestock24, for trading on the African mainland. In the towns, there were domestic slaves, who, because of their close contact with their white masters, and because of the very nature of their work, were almost certainly Portuguese-speaking. On the other hand,

inland the slaves were involved in agriculture, not in a typical monoculture

plantation system, but nevertheless with limited contact with their European masters. That being the case, Correia e Silva (1995b:284) has written that “this diverse, or adverse, nature [of rural Cape Verde], in enforcing an effort to adapt, by bastardising inherited knowledge and institutions, has contributed in its own way to the singularisation and creolisation of Cape Verdean society”. In other words, he sees the hybrid, creole nature of Cape Verdean culture (including of course the language) as having originated in the particular circumstances of the agrarian society of the interior of the islands. Just how this acculturation took place, and why, will be examined in greater detail below. Historians of Cape Verde are agreed that as from about the 1550s, a major economic and social crisis overtook the islands, the effect of which was particularly felt in the 17th century.25 There would appear to have been two related causes - increasingly frequent attacks on Portuguese shipping and ports in Cape Verde by French, English and Dutch pirates, and the inability of the Portuguese to compete with incursions on their trade in West Africa by the same European nations. The results of this crisis were as 23

ANTT, Chancelaria D. Manuel I, livro 20, fls.6vº-7; 10 July, 1504. Brásio (1963:10-12). See also Correia e Silva (1991:211). 24 Correia e Silva (1991:186-188) has shown how horses were bred on the islands for trade particularly in the Senegambia region. 25 Carreira (1972:ch. vi); Duncan (1972:ch.9); Correia e Silva (1995a).

24

follows: a) with the severe decline in trade, especially in slaves, virtually no new slaves were imported into Cape Verde for use on the islands, the few slaves that were acquired on the mainland being re-exported to the Americas; b) since subsistence agriculture in Cape Verde had been in large part directed towards providing supplies for trade shipping, what crops were still produced were then grown simply to satisfy local needs; c) meanwhile, with the reduced trading market on the Guinea Coast, the only cash crop in Cape Verde - cotton - consequently declined, but not as much as might have been expected, since other European traders were keen to acquire cotton and cloths woven locally for use in exchange for slaves and gold on the African mainland. As far as the acculturation of the African population in the Cape Verde islands is concerned, the crisis meant that with no new slaves being imported after about 1600-1610, a wholly nativised creole society arose, and with the white population declining, there was considerably less contact between the two communities. Another significant effect of the crisis was the fact that many slaves on Santiago took advantage of the insecurity at the time of pirate attacks to flee into the hinterland; this is examined in some detail by Carreira (1972:Chapter ix). Two descriptions of the Cape Verdes by English visitors - John Fryer (1672) and George Roberts (early 18th century) - give an impression of a closed community of mostly free blacks, very few whites, and an atmosphere of economic and social decadence:

The People are of a comely Black, their Hair frizled. Tall of stature, cunning and Thievish : they staring one in the Face, and in the mean time cut a Knot from the Shoulder, or steal an Handkerchief out of the Pocket. (Fryer 1909:45).

But [in the 17th century] after the Trade to Guinea, and the East-Indies, became common to other Nations and the Portuguese Trade declined, and dwindled away, as at this Day, to a Trifle, to what it was at first; the Blacks all the Time increasing more than the Portuguese, they claimed an equal Degree of Liberty and Freedom with their Masters, who thereupon mostly returned either to St Jago, or to Portugal; and those that remained, were both poor, and few, and were then

25

necessitated to allow the Blacks, if not to be their Superiors, yet, at least, to be their Equals; and marrying with them, the succeeding Generations became by that Mixture, from Mallattoes, to Copper-coloured Negroes; so that now you shall see as great Variety of Negroes on those Islands, as is continued on the whole Coast of Guinea. (Roberts 1726:387-388).

Roberts also shows that many of the islands of the archipelago were by then inhabited, specifically Sal, Boa Vista, Maio, Santiago, S. Felipe (Fogo), St. John (Brava), St. Nicholas (São Nicolau) and St. Antonio (Santo Antão), but he only refers to white settlers on Santiago, Fogo and Santo Antão. Santa Luzia and São Vicente he states were as yet uninhabited (1726:392-452). In the 18th century and the first half of the 19th, the slave trade had declined to such an extent that exports of orchil (a lichen used as a dye) brought in more income. The trade in cotton cloth was facing severe competition from other European countries, and the economy of the islands was confined to limited exports of sugar, coffee, maize and purgueira (a medicinal plant), as well as salt and animal hides to the Americas and the African mainland. Slaves were mostly confined to Santiago, where they worked on the entailed estates, many of which were now owned by people of mixed race. Free blacks worked as sharecroppers, and were “immensely poor”(Clarence-Smith 1985:43). Because of the large numbers of people of mixed race, there were moves in Portugal to incorporate the Cape Verdes into the Metropole, like Madeira and the Azores, but ultimately the islands were “more African than European” (idem). The 19th century saw a considerable increase in the white population, however, consisting mostly of convicts, sent to settle the Barlavento islands, particularly São Vicente, until then largely neglected.26 The port of Mindelo on this latter island in fact enjoyed a brief period of prosperity as a coaling station at the end of the 19th century. However, by this time, slavery had been abolished in Cape Verde (1875) - at the time of emancipation, 6,000 slaves were registered (Newitt 1981:211). 26

This would probably account for the fact that the Barlavento dialect of Cape Verdean Creole is closer to Metropolitan Portuguese, there having been greater contact with European Portuguese settlers, and hence more likelihood of greater proximity to the European language.

26

As far as population figures are concerned, the earliest date from 1513, in a letter from Pero de Guimarães to the Portuguese King27, stating that in the only town, Ribeira Grande, on Santiago Island, there were 58 “honourable white vizinhos”, i.e. (presumably male) heads of household, 16 “black vizinhos”, who are probably freedmen, 56 “foreign estantes”, or temporary residents, and around 10 black women. He also states that there were 12 clerics and 3 friars, including 2 preachers (Albuquerque & Santos 1988:221). Thus, 50 years after the first settlement grant, it can be assumed that there were at least 16 black freedmen and 10 (free) black women, compared with just over twice as many white adult male settlers. Guimarães gave no figure for the actual slave population, which was irrelevant for his purposes. The so-called Anonymous Pilot, whose account of the voyage from Europe to São Tomé was published in Italian by Ramusio in 1550, reported that there were 500 families living in Ribeira Grande on Santiago Island (Blake 1942:149). This clearly refers to the white population, and he gave no figures for the number of slaves or freed Africans. Much more detailed population figures were given in a survey of the islands written by Francisco de Andrade in 1582.28 He states that the town of Ribeira Grande had 508 vezinhos (white heads of household) and 5,700 slaves converted to the Catholic faith (de cõfisaõ). In the interior of the island, there were 600 “white and dark” (mixed race) men and 400 freed married blacks, with a total of 5,000 slaves “working on their farms”, 3,000 of whom were de cõfisaõ, and the rest were being “taught for that purpose” (Brásio 1964:99). Meanwhile, the island of Fogo was reported to have 300 white male residents (moradores) and 2,000 slaves, 1,500 of whom were de cõfisaõ (idem:101).

Correia e Silva (1991:229-232) has calculated from the parish

records of the islands of Santiago and Fogo for 1572 that there were 8,876 “souls” on Santiago, and 765 on Fogo.29 Later population figures for the Cape Verdes, until the 19th century, are very rare - in 1606, the Jesuit Father Baltasar Barreira stated that there were 500-600 “citizens” in Ribeira Grande,30 while not many years later, in 1630, João Pereira

27

ANTT, Corpo Cronológico, Parte I, maço12, nº 120; Albuquerque & Santos (1988:219-223). AGS, Guerra Antigua, maço 122, fols. 180-185; Brásio (1964:97-107). 29 Where the number of almas (souls) was not given, Correia e Silva multiplied the number of fogos (households) by an average number of 8.5. 30 ANTT, Cartório dos Jesuítas, maço 68, doc. 119; Brásio (1968:160). 28

27

Corte Real stated there were only 51 inhabitants of the town.31 Carreira (1972:288) quotes a report by the Governor João Cardoso Pissarro in 1672 as saying that the population of Santiago island had been reduced to “little more than 20 white men”.

1.4.2 Gulf of Guinea Islands 1.4.2.1 São Tomé It is by no means certain exactly when the Gulf of Guinea Islands were discovered, but most historians have followed the early 16th-century chroniclers in calculating that they were discovered in the 1470s, probably in 1471-1472

32

, though

Hodges & Newitt (1988:17-18) prefer to date the discovery of São Tomé and Príncipe to 1478-1479 (which would mean not having to account for an apparent delay of 14 years between discovery and settlement), and of Annobon to 1486. Either way, it is clear that unlike Fernando Po, the other Gulf of Guinea islands were uninhabited when discovered by the Portuguese. The first settlement grant was issued on 24 September, 1485 to João de Paiva, who was appointed Captain.33 Less than three months later, on 16 December, this was followed by a Royal Letter of Privilege to the settlers, granting all the right to have slaves.

34

The next Royal Grant, of 29 July, 1493,35 extended the powers of the

Donatary Captain, now Álvaro de Caminha, but more importantly makes provision for the building of sugar mills, showing that the decision had been taken to plant at least the island of São Tomé with sugarcane. Thus began on the island the relatively brief but prosperous „sugar cycle‟, which was to reach its peak in the second half of the 16th century. 36 There are a number of early documentary references to slavery on São Tomé, showing that the regime was clearly not as strict as that adopted later, for example, by the Portuguese in Brazil. The last will and testament of Álvaro de Caminha, dating from

31

ANTT, Cartório dos Jesuítas, maço 68, doc. 396; Brásio (1979:236). See Cabral (1995:273). See Tenreiro (1961:57-58); Cortesão (1971:12-13); Hodges & Newitt (1988:17-18); Garfield (1992:5-6). The chroniclers are Duarte Pacheco Pereira and João de Barros. 33 ANTT, Livro das Ilhas, fl. 109r & v; Brásio (1952:50-51); Albuquerque (1989:43-44). 34 ANTT, Livro das Ilhas, fl. 109v; Albuquerque (1989:45-49); Brásio (1988:3-7). 35 ANTT, Livro das Ilhas, fl. 104; Albuquerque (1989:54-58). 32

28

1499,37 shows that slaves at that time were primarily domestic, and that they were treated almost as part of the household - the will gives detailed arrangements to be made for the numerous slaves under Caminha‟s charge. Valentim Fernandes‟ description of São Tomé, dating from around 1507, states that there were some 2,000 slaves on the island at the time (as well as 5,000-6,000 awaiting transportation to Europe), compared with 1,000 settlers (Fernandes 1940:122). In addition, Fernandes refers to yams and cocoyams (taro) being grown for consumption by “blacks and slaves” (idem:128). Since sugar plantation was not yet on a large scale, it may be assumed that slaves were used both domestically as indicated in Álvaro de Caminha‟s will - and in small-scale agriculture, including sugar, but also of maize and other food crops. By 1529, the plantation system was in full operation, as can be seen from a letter written to King João III in that year by a leading planter, João Lobato38, in which he describes the preparations being made for setting up sugar mills on two estates, on the King‟s orders, for which he had contracted 270 slaves. However, Lobato had failed to provide enough subsistence crops for the slaves, and a number of them had escaped to the bush and to plantations where there were foodstuffs (Brásio 1952:505-506). In the early decades of the sugar boom on São Tomé, the organisation of slavery must have been on the lines of what was described by the Anonymous Pilot, as published in Italian in 1550:

A master does not give his slaves anything whatever. They work for him all the week, and only on Saturdays do they work to gain a living on their own account. Nor does a master trouble to give them clothes or food, or provide them with shelter, for they attend to all these things themselves. Apart from a scrap of cotton cloth, or palm fibre, with which they cover the indecent parts of their bodies, both men and women go naked. They eat the seeds which have been described above, which are like chick pea; grinding them into flour, and from this they make a kind of bread or bun by cooking it over the embers. The yam root is their staple diet. 36

For a detailed account of sugar plantation on São Tomé, see Garfield (1992:Ch. 4). ANTT, Corpo Cronológico, Parte 3, maço 1, nº 34; Albuquerque (1989:66-91). 38 ANTT, Corpo Cronológico, Parte 1, maço 42, nº 90; Brásio (1952:505-518). 37

29

They drink water, or palm wine, of which there is abundance, and sheep and goats‟ milk. There are many big flies in this island. (...) The negroes, on account of these flies, build their houses in the following way. They set up four pieces of wood, the longest they can find, in a square, and make a roof over them of pieces of wood fastened together above and at the sides with a kind of grass like coarse straw. Then, by means of a ladder with many steps, they mount at night to sleep, carrying their children with them easily. (Blake 1942:159-160; his translation).

However, Garfield (1992:86-87) has suggested that by the end of the 16th century, the system had become much harsher, as planters made every effort to reap the financial rewards from sugar, particularly when there was increasing competition from the much more extensive plantations in Brazil. Indicative of this is the report by Carmelite priests dating from around 1587,39 which deserves to be quoted at length: Les marchands (établis) dans l‟île de S. Tomé et qui y habitent, ont beaucoup de plantations, dont ils récoltent beaucoup de sucre et d‟autres produits... Celui qui dirige ces plantations y a une maison de récréation et a proximité une clôture, et à l‟intérieur de cette clôture sont les esclaves, hommes et femmes, qui cultivent ces plantations. Ils ont leurs camps et dans l‟un vivent cent esclaves hommes et dans l‟autre cent esclaves femmes... de manière que chaque homme ait une femme esclave, ou deux, comme concubines, et ainsi ils vivent en concubinage, comme s‟ils étaient mariés, dans une cabane. Les maîtres les voient et non seulement ils ne les reprennent pas, mais ils s‟en réjouissent parce qu‟à la fin de l‟année, ils ont leurs terres travaillées et cultivées, et la moitié d‟esclaves en plus par le fait des naissances. Car chaque négresse infailliblement chaque année donne le jour à un enfant noir. Outre cela, à ces esclaves, ils n‟enseignent ni la doctrine ni les choses de la foi. Ceux-ci restent en cela aussi ignorants que s‟ils étaient demeurés dans leur pays et que s‟ils étaient païens.

30

Les maîtres n‟ont pas souci que les

dimanches et jours de fête, ils entendent la messe ni qu‟ils se confessent. Ils ne leur accordent pas même le temps nécessaire pour cela. Ils n‟ont qu‟un souci: qu‟ils travaillent leurs terres et augmentent les esclaves. Ils ne les entretiennent pas, ni leur donnent à manger de toute la semaine. Les esclaves doivent euxmêmes se procurer leur nourriture et ainsi, il est nécessaire qu‟ils travaillent les dimanches et jours de fête pour se sustenter Dans leurs maladies, ils ne les pourvoient pas du nécessaire. Finalement, ils les traitent comme autant de bufs et de vaches, ou même pis que cela, parce que des bufs on en prend soin, tandis que des esclaves on ne tient aucun compte. Et cela se trouve introduit en cette île, comme si c‟était une chose conforme la raison. (Cuvelier & Jadin 1954:154-155).

Francisco Tenreiro (1961) was probably the first to describe the 17th and 18th centuries in São Tomé - and Príncipe - as the “great fallowing” (o grande pousio), a kind of empty period following a marked decline in the sugar production, and before the coffee and cocoa „cycles‟ took off in the 19th century. Hodges & Newitt (1988:24-27) followed suit, using the very same words to describe the period, while Jaime Cortesão (1968:42), a Portuguese historian not noted for his documentary accuracy, wrote of this period that “the farmers, (...) the traders and ship-builders (...) had from the end of the 16th century abandoned the island in large numbers, most of them going to Brazil” (cited in and translated by Ferraz 1979:19). While it is true that the sugar industry had gone into some degree of decline at this time, because of competition in quantity and quality from Brazil, there is no documentary evidence whatsoever of a mass exodus from São Tomé to Brazil.40 Garfield, who tends to agree with the picture of a chronic decline in this period, misquotes a document from 1615

41

as stating that “of the 72 [sic; in fact the document

states there had been 121; JL] fazendas on the island that grew sugar in the industry´s prime, 59 had been totally abandoned by c.1615, their owners most likely gone off to

39

Biblioteca Vaticana, MS Latina 12516, fls. 103-125; French translation from original Italian in Cuvelier & Jadin (1954:108-160). 40 Indeed, a number of creolists have used this flimsy evidence - erroneously in my opinion - to surmise that socalled Popular Brazilian Portuguese derives directly from São Tomé Creole. 41 BNA, 52-VIII-58, fl. 255; Brásio (1955b:190-191). See Neves (1989:25) for a correct reading of the document. Curiously, Garfield later in the same chapter gives the correct figures (1992:183-184).

31

Brazil.” (Garfield:1992:178; my emphasis). Nevertheless, whereas sugar production in the 1580s was said to have totalled 250,000 arrobas (almost 3,700 tonnes) 42, production in 1610 is quoted by Neves (1989:25) at around 200,000 arrobas, and in 1624 reduced to 80,000-100,000 arrobas.

Neves is led to conclude that “in our opinion, the fall in

production was not so sudden and abrupt” (idem). Indeed, Fernando Castelo-Branco has shown that in the mid-17th century probably still stood at around 100,000 arrobas, and was still the mainstay of the island´s economy (1968:81). Furthermore, this was at a time when the Dutch had attacked and occupied São Tomé for 7 years, from 1641 to 1648. By the year 1712, the “Memórias da Ilha de São Tomé” by Lucas de Pereira de Araújo e Azevedo

43

was publicising the agricultural potential of the island in glowing

terms, while admitting that the population had declined considerably from 11,000 “in the glorious past” - the 1520s - to a mere 500 in the early 18th century (fls. 46-46v; 1992:180). Carlos Neves has published a detailed survey of the island in the 18th century, based on extensive documentary evidence, and concludes that the plantation society was still more or less intact, despite the fact that there had been a gradual shift in the economy away from sugar production and towards subsistence agriculture for servicing shipping, not unlike the situation in the Cape Verde Islands at the time, together with a trend for the population to move to the town of São Tomé (1992:171-176, 189-191). Life on the plantations was described by Gaspar Pinheiro da Câmara in 176644 as follows: As povoaçoins destas ilhas (...) são as roças dos particulares, nas quaes juntando os negros de ambos os sexos estabelecem familias que vivem rodeando huma suficiente praça a que chamão o terreiro da roça, e ao todo a senzala; e isto assim sucede no Brasil, e se governa pelo cazeiro da roça destinto com este mesmo nome que hé outro negro cativo do mesmo senhor que elle elege com jurisdição delegada mais ou menos ampla, segundo a sua vontade, à qual são as leis que se estabelecem tanto para o serviço, como para economia, e se tem poder hé para

42

Biblioteca Vaticana, Coll. Urbinate, Latin nº 825, Part 1, fl.134; cited by Garfield (1992:166). ANTT, Manuscritos da Livraria, codex 108; Araújo e Azevedo (1992). 44 AHU, São Tomé, Caixa 10, doc. 93; Neves (1989:228-231). 43

32

tudo o senhor da roça independente, e muitas vezes até no espiritual, pois toma satisfação publica ao Parroco se se entremete a separar de alguma forma os concubinatos dos seos escravos, por lhe embaraçar a falta da propagação, ou obrigallos ao vinculo do matrimonio que lhes deminue o valor. (Neves 1989:231).

[The settlements on these islands (...) are the private plantations, on which negros of both sexes join together to set up families which live around a broad square which they call the terreiro of the plantation, and the whole they call the senzala; and this is like in Brazil, and government is by separate foreman of the plantation with this name who is another captive negro of the same master whom he chooses delegating to him more or less ample jurisdiction, according to his wishes, whereby are the laws established both for the service, and for the economy, and if the master of the plantation has independent power it is for everything, and often even over spiritual matters, since he demands public satisfaction from the Parish Priest if he intervenes to separate in any way the concubinage of his slaves, because this creates a lack of procreation, or forces them into the bonds of marriage which diminishes their value.]

Around the beginning of the 19th century, there was another major shift in the economy, from victualling to plantation once more, firstly of coffee, and subsequently of cocoa. Hodges & Newitt (1988:26) refer to a flourishing slave trade in São Tomé especially during the Napoleonic Wars, mainly with the nearby Gabon coast. There is no doubt that some slaves were retained on the islands for use on the new plantations, but it would appear that in terms of language, they had little or no effect on the local Creoles which were by now well established and thriving. For example, there seems to be no „substrate‟ influence from the languages spoken in Gabon (Fang, Make, Nkomi). The 19th and early 20th centuries were a time of economic boom for São Tomé and Príncipe - thanks especially to the production of cocoa.45 Interestingly the plantations were almost all set up and owned by local creoles, who by this time constituted a large

45

See Hodges & Newitt (1989:27-40).

33

and separate social stratum in the islands.

Following the abolition of slavery, the

plantations were worked by forced labour imported from northern Angola (mainly Kongo ethnic group): conditions were so bad that in the first decade of the 20th century they drew the attention of the mainly Quaker chocolate manufacturers in Britain, who placed an embargo on imports of São Tomé cocoa, with catastrophic results for the local economy. Nevertheless, there was a gradual recovery, particularly during World War II, and to this day, São Tomé and Príncipe, an independent nation since 1976, survives on exports of cocoa.

1.4.2.2 Príncipe With regard to Príncipe, documentation from the 16th century is scarce. What is known is that settlement began in 1500, under the Captain António Carneiro, appointed in the same year.46 The Royal Letter of Privilege also issued in 1500, is similar to that drawn up for São Tomé in 1485 (see note 34 above). Since Príncipe was governed from São Tomé, despite having its own hereditary Captain, the pattern of plantation and slavery must have been almost identical, albeit on a much smaller scale. Indeed, towards the end of the 16th century the harsher conditions under which slaves lived in São Tomé as described by the Carmelites in around 1587 (see note 39 above) were identical on Príncipe: “Et dans l‟île du Prince, c‟est la même chose.” (Cuvelier & Jadin 1954:155). The earliest population figures date from 1607, in the anonymous Relação da Costa da Guiné,47 when there were said to be 5 or 6 sugar mills on the island, 10 Europeans, more than 18 “married creoles”, 20 “married free blacks” and up to 500 slaves. In 1620 a report in Spanish by Garcia Mendes Castelo Branco48 stated that there were 700 “vezinos” (inhabitants), including whites and local creoles (“criollos de la tierra”). By the beginning of the 18th century, Jean Barbot was writing of Príncipe:

46

ANTT, Chancelaria de D. Manuel I, livro 14, fl. 170v; Silva Marques (1971: III, 616-617). BAL, Ms. 51-VIII-25, fls. 122v-130; Brásio (1955a:382-383). 48 BAL, Ms. 51-VIII-25, fls. 73-77v; Brásio (1955b:472). 47

34

The inhabitants are nearly all black (most of them being slaves). There are only 15 or 16 families of true Portuguese, and 50-60 mulatto families, although there are reckoned to be 3,000 souls in the island. All are Roman Catholics, but of the most suspicious sort. (...) The Moorish islanders are evil, the mulattos hardly any better, and even the whites are not too reliable. The last always carry a sword and a dagger at their side or in their hand. The Moors have only a knife on their hip in the manner of the peoples of Gold Coast, and wear only a simple cloth, to cover what modesty forbids to be seen. However, all these islanders are, in general, less savage and cruel than they were at the end of the last century, when they massacred indiscriminately all those who fell into their hands. (...) The women are much more agreeable than the men. The women of the mulattos generally have a good figure, and to foreigners they are not markedly unkind. They dress in the Portuguese style and follow that way of life, being always shut up indoors and only coming out in order to go to mass, even then being veiled and followed by an old female on whom they lean. (Hair et al. 1992:724-725).

A few years later, in 1724, Daniel Defoe published the following information about Príncipe obtained from John Atkins: The manner of living among the Portuguese here, is with the utmost Frugality and Temperance, even to Penury and Starving. (...) The Negroes have yet no hard Duty with them, they are rather happy in Slavery; for as their Food is chiefly Vegetable, that could no Way else be expended, there is no Murmurs bred on that Account; and as their Business is domestick, either in the Services of the House, or in Gardening, Sowing, or Planting, they have no more than what every Man would prefer for Health and Pleasure. (...) The Negroes are, most of them, thro‟ the Care of their Patroons, Christians, at least nominal; but excepting to some few, they adhere still to many silly Pagan Customs in their Mournings and Rejoycings, and in some Measure, powerful Majority has introduced them with the Vulgar of the Mulatto and Portuguese Race.

(...)

The Portuguese, tho‟ eminently

abstemious and temperate in all other Things, are unbounded in their Lusts; (...) 35

They have most of them Venereal Taints, and with age become meagre and hectick. (Defoe 1724:189-191).

Setting aside the ethnocentric prejudices of these two European observers, one can see that Príncipe had to some extent entered a “fallow period” like its larger neighbour; indeed, Barbot made no mention of sugar being grown as a crop on the island, the “local products” are referred to as including manioc and rice, used in trading by barter (Hair et al. 1992:725). In 1753, following an attack by the French on the harbour of Santo António, the Portuguese resolved to transfer the administration of the Gulf of Guinea Islands to Príncipe, partly to secure a stonger defence against attack, and partly because of the social unrest that had beset São Tomé. The administrative capital was re-established in São Tomé only in 1852. It is perhaps significant that it was on Príncipe that both coffee and cocoa were introduced experimentally, respectively in 1800 and 1822, and plantations of both crops flourished, albeit on a reduced scale, in the 19th and early 20th centuries. Nevertheless, Príncipe has remained a virtual backwater until the present day, closely following the fortunes of São Tomé Island, with a social structure almost identical to its larger neighbour. The current population is around 5,000.

1.4.2.3 Annobon The island of Annobon (18 sq. km.) is minute by comparison with São Tomé (857 sq. km.), and even Príncipe (139 sq. km.), and the sugar plantation system never developed there as it did on the other two islands. According to Valentim Fernandes, settlement began in 1503, as ordered by the Captain of São Tomé, and by 1507 (when Fernandes was probably writing), there were just 9 (white) inhabitants. Fernandes also reports that there was a stock of pigs, goats and chicken taken there by the Portuguese (Fernandes 1940:129-130).

In 1543, the Captain of Annobon, Álvaro da Cunha,

requested rather optimistically that the island be granted the same charter (foral) as São Tomé.49 The Anonymous Pilot, whose account was first published in 1550, refers to

49

ANTT, Gavetas, X-11-6; Brásio (1988:139-142).

36

Annobon as being “uninhabited (...) and there is much fishing; and the people of San Thomé often go there to fish” (Blake 1942:156). A number of references, mainly by foreign visitors, dating from the 17th and early 18th centuries show that the island was virtually abandoned, with at most some smallscale cotton plantation. The population consisted almost entirely of a few hundred Africans, with occasionally perhaps one white overseer.50 For example, the Dutchman Willem Bosman wrote of Annobon in 1704: The Inhabitants of this Island are black, and but a sort of half Christians, tho‟ they bear the name of Christians; for if they can but read a Pater noster and Ave Maria, confess to the Priest, and bring some Offerings with them, they pass for good Christians. The White Portuguese of this Island esteem them Slaves, by reason that they are the Descendants of those Slaves which they set on this Island. They are all of them, without any Exception, Thieves and Rogues, that will injure or defraud no Man more than they can.

The Women are all common publick

Whores, which allure and mislead the Sailors, and except a very few, are monstrous, ugly and ill-favour‟d, as Monsieur Fokkenbrog has well observ‟d it. (Bosman 1967:416-417; 18th-century English translation). However, in 1754 the Governor of São Tomé suggested that “visto que nao haver aly mais, que pretos” [since there were only blacks there], 25 white couples should be sent to civilise the local population.51 In 1770, an official visit was paid to Annobon by the Portuguese authorities - they reported that there 2,000 inhabitants, “bem formadas, e de estatura mais que ordinaria” [well formed, and of more than ordinary stature].52 In another document dating from the same year, 53 it is stated that:

50

See François Pyrard de Laval in 1601 (English translation in Gray [1887:13-17]); the report by the Spaniard Garcia Mendes Castelo Branco in 1620 (BAL Ms 51-VIII-25 fls. 73-77v; Brásio [1955b:472]); Jean Barbot in 1678 (English translation in Hair et al. [1992:742-743]). See also Neves (1989:63-65) for an examination of the situation on Annobon in the 18th century. 51 AHU, São Tomé, Caixa 9, doc. 20; Neves (1989:210). 52 AHU, São Tomé, Caixa 12, doc. 23; Neves (1989:271). 53 AHU, Codex 561, fols. 25v-32v; Neves (1989:257-260).

37

porque havendo na mesma ilha de Anno Bom grande quantidade de algudão; e constando que ainda no anno de 1693, se fabricavão ali pannos, que os habitantes levavam a vender a ilha de Santo Thomé; não deixaria de ser muito conveniente, que restabelecendo-se as precedentes fabricas, se fornecessem dellas as outras ilhas. (Neves 1989:257).

[because there is on the same island of Annobon a great quantity of cotton; and it is reported that still in the year 1693 cloths were being made there, which the inhabitants would take to sell in the island of São Tomé; it would be most convenient if, by re-establishing their former manufacture, they could supply the other islands with them.]

In 1778, under the terms of the Treaty of Pardo, the islands of Annobon and Fernando Po were ceded by Portugal to Spain. However, by all accounts Annobon remained totally neglected by Spain until well into the 20th century.54 The island now forms part of Equatorial Guinea, independent since 1968; the current estimate of the population is around 2,000.

54

See particularly Unzueta y Yuste (1945:66-68).

38

2 THE CREOLE LANGUAGES DESCRIPTIVE ANALYSIS

39

Até os mesmos brancos são pouco civilizados, de sorte que são bem raros os que sabem falar a lingua portuguesa com perfeição, e só vão seguindo o estilo de falar da terra, que é uma corruptela tão rústica que não se pode escrever; e a sua mesma expressão está indicando a pirguiça e desmazelo desta gente. (Anon. [1784] 1985:27). [Even these same white people are barely civilised, such that rare are those that can speak the Portuguese language with perfection, and they merely follow the style of speaking of this land, which is such a rustic corruption that cannot be written down; and this same means of expression goes to indicate the sloth and dereliction of these people.]

2.1 Introduction This chapter seeks to identify the linguistic features in the Cape Verde and Gulf of Guinea Creoles which differ from those in the Portuguese lexifier, and which may be typical of creole languages in general. As a consequence, it may also establish to what extent any of the Creoles is more „radical‟ (i.e. further from Portuguese) than any of the others. As indicated in Chapter 1, the languages examined are the Cape Verde, São Tomé, Príncipe and Annobon Creoles; reference will also be made to Angolar, for comparative purposes, but since this is not a plantation creole, i.e. it was formed under different sociohistorical circumstances, it is not a major focus of attention in this particular study. Brazilian Portuguese will be considered separately in Chapter 6, as a probable example of non-creole development of the European language. While descriptions of all four languages do exist (as indicated in 1.3 above), none of them are entirely satisfactory, since they are often incomplete, either because their approach is to some extent outdated, or else the author did not have the opportunity to undertake the necessary depth of research. That being the case, it is in some cases possible to supplement the general descriptions with a number of useful papers and articles, but where this is not the case, the results of my own research have been included here. The linguistic features are examined and compared in the following order of linguistic levels: phonology; lexis and semantics; morphology, and syntax.

Only

specifically „creole‟ features are studied here, together with those that are recognisably 40

different from the features of European Portuguese: this chapter is not intended to be an exhaustive comparative description of the four languages in question.

2.2 Phonology As Holm (1988:105) has pointed out, it is difficult to establish whether phonological differences between the European lexifier and a particular creole, or between these creoles, are the result of language-internal or -external changes. In other words, one cannot necessarily point to palatalization, for example, as being caused by African „substrate‟ influence or by „universal‟ factors. What can be established here, however, is the comparative extent to which these changes occurred, which in turn can be used to assess the „radical‟ nature, or otherwise, of the specific creole. Phonological evidence alone, for the reasons indicated here, cannot be sufficient for establishing a meaningful comparison between separate creoles, or with their „superstrate‟ or „substrate‟ languages; nevertheless, such an attempt has been made, for example, by Norval Smith (1987), who relied almost exclusively on phonological factors to establish what amount to genetic relationships in terms of the Surinam Creoles and their genesis. Specific phonological features in the four Portuguese-based creoles studied here will be examined and compared in the following sequence: oral vowels; oral diphthongs; nasal vowels; nasal diphthongs; vowel harmony; consonants; phonotactic rules; parasegmentals.

2.2.1 Vowels 2.2.1.1 Oral Vowels a) 16th-century European Portuguese As Teyssier (1982:42) has shown, in around 1500 European Portuguese had evolved a system of 8 oral vowels, as follows: front close close mid

central

back

i

u e

o 41



open mid open





a

By the beginning of the 19th century, this had expanded to include a close mid central vowel //, making the modern European Portuguese system of 9 oral vowels. However, Holm (1988:114), in comparing the vowel systems of creoles and their European lexifiers, fails to note this subsequent development in Portuguese, and therefore assumes that the Portuguese plantation creoles had reduced their vowel systems from an original 9 rather than 8 oral vowels.

b) Cape Verde

In common with many West African languages (according to Holm 1988:114115), the oral vowel system in Cape Verde Creole consists of the following 7 vowels (Veiga 1982:28; Lopes da Silva 1957:53-93) :

front close close mid open mid

central

back

i

u e

o 

open

 a

In other words, the /e/ // distinction in European Portuguese has been retained, as has that of /o/ and //, while /i/ and /u/ correspond to the same close vowels in Portuguese. The only difference is that // and /a/ have been reduced to /a/, in common with the African substrate languages, according to Holm (1988:117). Veiga (1982:40-44) has also examined in detail the changes in the vowel system from Portuguese to Cape Verde Creole, and includes a complete comparative table of vowel distribution (1982:44).

c) Gulf of Guinea 42

São Tomé and Príncipe both have the same system of 7 oral vowels as in Cape Verde (cf Ferraz 1979:20; Günther 1983:36). In Annobon, according to Post (1995a:193-194) the oral vowel system has been reduced even further to a basic five-vowel system, possibly as a later development through contact over the last 200 years with Spanish, which has the same basic 5-vowel system:

front close

central

back

i

mid

u e

o

open

a

By contrast, Post states that vowel length appears to be a feature of Annobon Creole, and she gives the following minimal pairs as examples:

(1)

„plate‟

pá:tu

< prato

pátu

< pato

„bird‟

dé:nci

< frente (?)

„in front of‟

dénci

< dente

ke:sé

< crescer

„to grow‟

kesé

< esquecer

„to forget‟

„tooth‟

However, Post notes that “only in a small number of cases is vowel length distinctive” (1995a:194).

43

On the other hand, Barrena (1957:17-19) distinguishes /a, e, , i, o, , u/ - i.e. the same 7-vowel system as the other Portuguese creoles -, with the addition of //. For example:

(2)

gués

< igreja

„church‟

Barrena also refers to varying vowel length in Annobon (1957:18).

2.2.1.2 Oral diphthongs Teyssier (1982:43) indicates that in the modern Portuguese system of oral diphthongs, /i, i, u/ are relatively recent developments, though he does not suggest any date for these innovations. One can therefore assume that the 16th-century Portuguese system of oral diphthongs would be as follows:

-i

-u ui oi

ei

iu eu

ai

ou au

In Cape Verde, according to Lopes da Silva (1957:61-62), /ai/ is sometimes maintained (e.g. in mono- or final syllables), and sometimes reduced to /a/:

(3)

pai

< pai

(father)

bai

< vai

(to go)

káa

< caixa

báu

< baixo

but (box) (low)

/ei / is normally reduced to /e/ (1957:73-74):

(4)

mantéga

< manteiga

„butter‟

44

/oi/ is retained (ibid.:87), as is /ui/ (1957:91-92).

(5)

koitóde

< coitado

„wretched‟

kuidá

< cuidar

„to take care‟

/au/ is retained (1957:62), as are /eu/ (1957:74) and /iu/ (1957:81). But /ou/ is reduced to/o/ (1957:88). For example:

(6)

mau

< mau

„bad‟

deuz

< deus

„god‟

tiu

< tio

„uncle‟

oru

< ouro

„gold‟

See also the table of changes from Portuguese diphthongs to Cape Verde Creole in Veiga (1982:46).

In São Tomé, the Portuguese diphthong /ai/ is reduced to /a/ (Ferraz 1979:33). /ei/ is reduced to /e/ when the final vowel is /u/, and to // when the final vowel is /a/, though this “is not a general rule, but operates as a mechanism for incorporating” Portuguese /ei/ (1979:31-32); for example:

(7)

pétu

< peito

„chest‟

tmá

< teimar

„to be stubborn‟

/oi/ is reduced to /o/; /ui/ tends to be retained (1979:32-33):

(8)

dósu

< dois

„two‟

/au/ tends to be retained, while /eu/ and /iu/ are reduced to /e/ and /i/ respectively (1979:33); /ou/ is reduced to /o/: 45

(9)

désu

< deus

„god‟

ci

< tio

„uncle‟

There are, however, a few exceptions to these rules, e.g.:

(10)

kúnda

< cuidar

„to take care‟.

Príncipe for the most part has followed São Tomé in the development of Portuguese diphthongs:

/ai/ is reduced to /a/; /ei/ to /e/ or //, according to the

environment; /oi/ to /o/. However, /ui/ tends to be reduced to /u/.

(11)

basá

< baixar

„to lower‟

péi

< peixe

„fish‟

tmá

< teimar

„to be stubborn‟

depói kudá

„after‟

< depois < cuidar

„to take care‟

Unlike in São Tomé, /au/ is normally reduced to /a/ in Príncipe; /eu/ is reduced to /e/; /iu/ is retained (unlike São Tomé); /ou/ is reduced to /o/.

(12)

bakayá

< bacalhau

„codfish‟

désu

< deus

„god‟

ciu

< tio

„uncle‟

óru

< ouro

„gold‟.

In Annobon, the development of Portuguese diphthongs is as follows: /ai/ is retained; /ei/ is reduced to /i/, and /oi/ to /o/.

(13)

pai

< pai

„father‟ 46

píi

< peixe

„fish‟

ótu

< oito

„eight‟

/eu/ is reduced to /e/; /iu/ is retained; /ou/ is reduced to /u/; /ui/ is retained:

(14)

zudé

< judeu

„Jewish‟

fiu

< frio

„cold‟

útulu

< outro

„other‟

No examples of the development of the Portuguese diphthong /au/ have been found for Annobon, but since, according to Post (1995a:193), the only diphthongs in Annobon are /ai, ei, ui/, /au/ must have been reduced.

2.2.1.3 Nasal vowels

In both 16th-century and modern European Portuguese, there are nasalised vowels corresponding to each of the five oral vowels /a, e, i, o, u/ (Teyssier 1982:28; Cunha & Cintra 1984:37-38). In Cape Verde, according to Veiga (1982:29), all 7 oral vowels - /a, e, , i, o, , u/ - have nasalized equivalents. In São Tomé, the same 5 vowels as in Portuguese have nasal counterparts (Ferraz 1979:20). However, Ferraz (ibid.:37) also notes that nasalization in São Tomé can be progressive in cases of a vowel following a nasal consonant /m,n/. For example:

(15)

kamínza

> camisa

„shirt‟

imint´li

> cemitério

„cemetery‟

mínda

> medida

„measure‟

méndu

„fear‟

> medo

nãsé

> nascer

„to be born‟

mãsádu

> machado

„axe‟ 47

Príncipe has the same 5 nasal vowels as São Tomé (Günther 1973:36-38); however, it does not appear to be as progressive a feature as in São Tomé:

(16)

m´ndu

> medo

kamíza

> camisa

nasé

> nascer

masádu

> machado

but

In Annobon, according to Post (1995a:194), there are no nasalized vowels. However, Barrena (1957:18) refers to nasalized /a/ - /ã/ - as in

(17)

ampã

> pão

„bread‟.

2.2.1.4 Nasal diphthongs The three main nasal diphthongs in modern European Portuguese are /ãu, ãi, õi/ (in modern orthography: ão, ãe, õe). Of the three, /ãu/ is the most frequent, while the other two are confined to the plural forms of /ãu/, except for mãe /mãi/ („mother‟) and põe /põi/ (3rd sing. pres. of pôr „to put‟).

According to Teyssier (1982:45-46), the

development of early Portuguese -an, -on to modern -ão had already taken place by 1500, and therefore 16th-century European Portuguese can be said to have had all three nasal diphthongs. However, without exception, the Portuguese-based Creoles do not have the morphology to produce either the plural forms /ãi(), õi()/, or the 3rd. sing. present form /põi/. Therefore, the search for the development of oral diphthongs in the Creoles will be confined to /ãu/, and /ãi/ in mãe.

48

In Cape Verde, /ãu/ is retained, but /õ/ is an older, more frequent form, according to Lopes da Silva (1957:63). Lopes da Silva also notes that in Fogo, /ã/ corresponds to /õ/ in the other islands (ibid.). The word mãi („mother‟) shows that /ãi/ is also retained.

In São Tomé, /ãu/ is normally reduced to /õ/, which is seen by Ferraz (1979:31) as the “medieval variant, [which] is an indication of the early formation of S[ão] T[omé Creole]”, though if Teyssier (cited above) is correct, this statement is not valid. Portuguese mãe is reduced to me.

According to Ferraz (1979:31), in Príncipe, Portuguese /ãu/ is reduced to /ã/ (cf. Günther 1973). Portuguese mãe is mwem in Príncipe.

In Annobon, /ãu/ would appear to have been reduced to /ã/: e.g. lazã < razão („news‟). The word mãi for Portuguese mãe shows that the original nasal diphthong has been retained.

2.2.1.5 Vowel harmony Holm (1988:124-125) has shown that in many Atlantic Creoles - and indeed in a number of West African languages - vowel harmony is a notable feature, particularly in the case of paragogic vowels added to maintain the „classic‟ CV structure. It would seem that vowel harmony is especially evident in the more „radical‟ Creoles: Holm (ibid.) cites Saramaccan, the Gulf of Guinea Creoles, and Haitian Creole French. In Cape Verde, vowel harmony does not appear to be a feature, even in the most basilectal varieties of the Sotavento. No reference is made to vowel harmony in the extensive section on Cape Verde phonology in Veiga (1982), and examination of the word list in Lopes da Silva (1957) reveals no examples of this feature. On the other hand, in the Gulf of Guinea Creoles, vowel harmony is a clearly recognisable characteristic.

For São Tomé, Ferraz (1979:25-26) states that “vowel

harmony is a consistent feature of S[ão] T[omé] phonology”. However, he goes on to say 49

that “vowel harmony in ST occurs virtually only with non-nasal vowels”; among the examples he gives are the following:

(18)

bíli

< abrir

„to open‟

mesé

< mister

„to want‟

Ferraz also includes a detailed discussion on the “application of vowel harmony” in São Tomé Creole (1979:43-46), in which he shows that the paragogic vowel is normally the stressed vowel of the etymon. Ferraz also includes a section on this feature as a reflection of African phonology (1979:49-51), in which he shows that vowel harmony rules were also applied to words of African origin; he uses this as evidence for his belief that vowel harmony is a totally „substrate‟, rather than a universal, feature. In the case of Príncipe, however, it would appear that vowel harmony is not such an extensive feature. Günther (1973) makes no mention of it in the section on phonology of his description of Príncipe Creole, but examination of his word list reveals the following examples which include occasional differences from São Tomé Creole:

(19)

doutor

mister

> dotólo

(São Tomé)

> dotó

(Príncipe)

> mesé

(São Tomé)

> mesé

(Príncipe)

> ml

(São Tomé)

> mli

(Príncipe)

> bilí

(São Tomé)

> bi

(Príncipe).

„doctor‟

„to need‟

but

mel

abrir

50

„honey‟

„to open‟

With reference to Annobon, Post (1995a:194) states that “in conversation the phonemic structure of (...) words and sentences undergoes a number of modifications such as (...) vowel harmony” (my emphasis). The one example she gives of this and other modifications does not clarify this statement, however. Examination of Granda‟s 200word Swadesh list of Annobon Creole (1985:159-165), on the other hand, reveals the following examples, amongst others, of vowel harmony:

(20)

álba

< erva

„grass‟

dedé

< arder

„to burn‟

gódo

< gordura

„fat‟

fugú

< -pùk- (Proto-Bantu)

„to dig‟.

2.2.2 Consonants a) 16th-century European Portuguese By examining Fernão de Oliveira‟s Portuguese Grammar of 1536 (Silveira 1954), as well as Teyssier‟s History of the Portuguese Language (1982), one can draw up the following table for the consonant system for 16th-century European Portuguese:

51

Bilabial Plosive Nasal

p b m

Labiodental

f v l

Lateral Semi-vowel

Palatal

Velar

t d

c j 

k g

n

Trill Fricative

Apical

r s z 

R  y

w

However, it should be pointed out that a number of changes were taking place at that time which are reflected in the consonant systems of the African Creoles. The shift from /c/ to //, already discernable in the Portuguese Grammar by Fernão de Oliveira, dated 1536 (Silveira 1954), and in that of João de Barros of 1540 (1957), was not noted by Duarte Nunes de Leão in his „Orthographia‟ of 1576 (Teyssier 1982:53); therefore both phonemes appear in the table above. Syllable-final /s/ only shifted to //, or // before a voiced consonant, in European Portuguese in the 18th century - Teyssier states that the earliest reference to this change is dated 1746, and he hypothesises that the shift had begun to take place in the 16th century (1982:54-55). Teyssier also speculates about the change from the palatal plosive /j/ to the fricative //: “it is difficult to tell if this development occurred during the period under discussion [14th century] or later” (1982:27-28). I would suspect that the change did in fact take place later, judging by the reflexes in the Creoles. b) Cape Verde The consonant system in Cape Verde has been reduced to the following, taking into account recent additions from Portuguese with the consonant /v/ (adapted from the table in Veiga 1982:30):

Bilabial Plosive Nasal

p b m

Labiodental

f (v) l

Lateral Semi-vowel

Palatal

Velar

t d

c j 

k g

n

Trill Fricative

Apical

r s  

w

y

52

Veiga (1982:35-40) notes the following changes that have taken place in basilectal Cape Verde Creole:

1.

In Santiago basilect, in contrast with the Barlavento islands, Portuguese /v/ is

generally rendered as /b/, except in recent additions from Portuguese (21)

báka

„cow‟

< vaca

kabálu

„horse‟

< cavalo

but

vérbu

< verbo

„verb‟

ravalusón

< revolução

„revolution‟

Veiga (1982:36) also notes that in Sotavento basilectal varieties, /b/ in the combination /Vbu/ is frequently dropped -

(22)

2.

kau [kabu]

< cabo

„place‟

pou [pobu]

< povo

„people‟.

Portuguese intervocalic /z/ is retained as /s/ in the Sotavento islands, but in Fogo

(Sotavento), there is a shift from Portuguese /VzV/ in proximity with a nasal consonant to / VnzV/55. For example:

(23)

but

„house‟

kasa

< casa (/káza/)

kamísa (Santiago)

< camisa (/kamíza/) „shirt‟

kamínza (Fogo)

55

This feature may be compared with the progressive nasalization in São Tomé, referred to above. For example, Portuguese camisa has become kamínsa in both Cape Verde (Fogo) and São Tomé.

53

3.

The Portuguese consonant // when represented orthographically as ch is changed

in Cape Verde to /c/, but when the orthography of this phoneme is x, it is retained as //:

(24)

4.

cúba

< chuva

„rain‟

cábi

< chave

„key‟

púa

< puxar

„to pull‟

mée

< mexer

„to stir‟.

It was noted above that syllable-final /s/ only shifted to //, or // before a voiced

consonant, in European Portuguese in the 18th century: in all but recent borrowings from modern Portuguese, /s/ is retained in Cape Verde:

(25)

mas

< mas

„but‟

spértu

< esperto

„clever‟.

However, there are cases of a shift from /s/ to // in Cape Verde at the beginning of a syllable with /i/:

(26)

5.

intídu

< sentido

„sense‟

inína

< ensinar

„to teach‟.

In the more acrolectic Barlavento islands, there are cases of // replacing earlier

Portuguese /s/ where in modern Portuguese there would be //:

„but‟.

(27)

ma ~ ma

< mas

6.

Where /j/ has shifted to // in modern Portuguese, the original consonant is

retained in the Sotavento islands: 54

(28)

janta

< jantar

„to dine‟

justísa

< justiça

„justice‟.

Exceptions occur in recent borrowings from Portuguese:

(29)

7.

élu

< gelo

„ice‟

ésu

< gesso

„plaster‟.

Portuguese // has been replaced, in all but the most recent borrowings from

Portuguese, by /j/ in the basilectic Sotavento islands; in the Barlavento, it is replaced by /j/ or /i/, particularly in Santo Antão island, where it can even be dropped altogether:

(30)

8.

míju (Santiago) mij

(São Vicente)

mii

(Santo Antão)

pája

(Santiago)

páia

(São Vicente)

paa

(Santo Antão).

< milho

„maize‟

< palha

„straw‟

There are some cases, particularly in basilect of the interior of Santiago and in

Fogo (Sotavento), of a shift from /l/ to /r/:

(31)

argén

< alguém

„someone‟

ártu

< alto

„high‟

kúrpa

< culpa

„blame‟.

55

9.

Portuguese /R/, as an initial r or as rr orthographically, which in any case is a

relatively late development in European Portuguese, remains as /r/ in Cape Verde.

c) São Tomé The consonant system in São Tomé has developed from 16th-century Portuguese as follows (adapted from Ferraz 1979:21): Bilabial Plosive

p b m

Nasal Fricative Lateral Prenasalized stop Semi-vowel

Labiodental

n f v l

Apical

Palatal

Velar

t d

c j ()

k g

s z 

mb

nd

w

() nj y

São Tomé Creole has three pre-nasalized stops, or as Holm (1988:128) describes them, “word-initial nasals that are homorganic with the following stops”, e.g.

(32)

nda

< andar

„to walk‟

njánja

< já! já!

„immediately!‟

As Holm (ibid.) points out, Ferraz (1979:26-27) “analyzes these as sequences of two consonants, although the language has a predominant CV syllabic structure”. Ferraz (1979:33-37) goes on to outline the development of consonants from Portuguese to São Tomé Creole as follows: 1.

The following phonemes are retained - /p, b, k, g, f, v, m, n/. // is retained only in

certain cases, e.g.

(34)

klakañón

< calcanhar

„heel‟

sóñu

< sonho

„dream‟

sónu

< sono

„sleep‟

cf.

56

2. Palatalization of /t, d, s, z/ to /c, j, , / has occurred before oral and nasal /i/ and the semi-vowel /y/:

(33)

sóci

< sorte

„luck‟

táji

< tarde

„afternoon‟

ínku

< cinco

múika

< música

„five‟ „music‟.

Similar palatalization of /t, d/ has taken place in several areas of Brazil, and some linguists have seen this as a result of diffusion from the Gulf of Guinea Islands, but I am of the opinion that this was a parallel development, possibly with similar African „substrate‟ influence.

3.

The 16th-century Portuguese /b/ for /v/ has been retained in São Tomé, but what

must have been later borrowings from Portuguese incorporate /v/:

(35)

labá

< lavar

„to wash‟

bií

< vestir

„to dress‟

vyõ

< avião

„aeroplane‟

véde

< verde

„green‟.

but

4.

Portuguese // has normally been replaced by /y/, but occasionally by //;

infrequently // is retained, or else the original syllable containing this consonant is deleted. For example:

(36)

fya

< folha

„leaf‟

koyé

< escolher

„to choose‟

57

moñá

< molhar

„to wet‟

miõ

< melhor

„better‟

v

< velho

„old‟.

5.

There has been a complete shift from Portuguese /r, R/ to /l/ in São Tomé:

(37)

áli

< ar

„air‟

látu

< rato (/Rátu/)

„rat‟.

/r,R/ are very occasionally retained by some speakers bilingual in Portuguese and São Tomé Creole, “but it is extremely rare” (Ferraz 1979:37). d) Príncipe The development of the consonant system in Príncipe has resulted as follows (adapted from the table in Günther 1973:45): Bilabial Plosive Nasal

p b m

Labiodental

f v l

Lateral Coarticulated Stop Prenasalized Stop Semi-vowel

Palatal

Velar

t d

c 

k g

n

Trill Fricative

Apical

r s z   g nd

w

ng y

In addition to the basic consonants, Príncipe Creole has a co-articulated stop, described by Günther (1973:41, 45) as the labiovelar plosive consonant /gb/, and attributed by him to being of Igbo origin. However, Ferraz transcribes the phoneme as /g/ and classifies it as a voiced velar stop, adding that it “is not a productive unit (...), and is only found in archaic [sic] borrowings from African languages” (1975:155-156). 58

Príncipe also has pre-nasalized stops similar to those in São Tomé (see above), as in: (38)

nda

< andar

„to walk‟

ngomá

< engomar

„to iron‟.

The following consonants are retained from 16th-century Portuguese - /p, b, r, k, g, f, v, m, n/. // is retained only in certain cases of recent borrowings from Portuguese (Günther 1973:42-43). // has frequently been retained from Portuguese, much more so than in São Tomé:

(39)

máña

< manha

„craftiness‟

kakañá

< calcanhar

„heel‟

As elsewhere in the Gulf of Guinea, palatization is a distinctive feature in Príncipe Creole, but here it occurs only with /t/ before /i/, and there exceptions to this rule (1973: 44):

(40)

móci

< morte

„death‟

fitiyó

< feiticeiro

„witchdoctor‟.

but

Other features noted for São Tomé do not appear in Príncipe: the /b, v/ distinction is retained as in Portuguese, and the shift from /r/ to /l/ has not taken place; e.g.:

(41)

ve

< ver

„to see‟

ráña

< rainha

„queen‟

e) Annobon The consonant system that has developed in Annobon is as follows (adapted from the table in Post 1995a:192): 59

Bilabial Plosive

p b m

Nasal Fricative Lateral Prenasalized Stop Semi-vowel

Labiodental

n f v l

mb

Apical

Palatal

Velar

t d

c j 

k g

s z 

x

nt nd

ng

w

y

As in the other Gulf of Guinea Creoles, Annobon has a number of pre-nasalized stops, indeed more than are to be found in São Tomé or Príncipe. Another consonant, /x/, occurs in Annobon, unlike elsewhere in the Gulf of Guinea Creoles; it would appear to replace Portuguese /k/, but only selectively (?):

(42)

xansá

< cansar

„to tire‟

xabá

< acabar

„to finish‟

xomesá

< começar

„to begin‟

xamá

< queimar

„to burn‟

ke

< (o) quê?

„what?‟

ku

< com

„with‟

tabáku

< tabaco

„tobacco‟.

but

Palatalization of /d, t, s/ before/i/ - to /j, c, / respectively - has taken place in Annobon: (43)

jía

< dia

„day‟

cilá

< atirar

„to throw‟

iólo

< senhor

„gentleman‟.

The shift from /r/ to /l/ has taken place in Annobon, though occasionally /r/ is dropped altogether:

60

(44)

Malí

< Maria

[personal name]

pá:tu

< prato

„plate‟.

The Portuguese palatal // does not occur in Annobon, and has been replaced by the semi-vowel /y/ or by /l/, or omitted altogether: (45)

ozóyo

< joelho

„knee‟

myéle

< mulher

„woman, wife‟

mádu

< molhado

„wet‟.

2.2.3 Phonotactic rules Holm (1988:108-109) claims that “almost all West African languages” have a basic CV syllable structure; this also applies to many Atlantic Creoles, “particularly those whose structure is least influenced by that of their European lexical-source language such as the Portuguese-based creoles of the Gulf of Guinea (...) (once called the most “conservative” creoles in reference to their African features, but rechristened the „radical‟ creoles by Bickerton 1981)” (1988:109). Holm adds that “as Ivens Ferraz (1987[:342343]) notes, the Upper Guinean varieties of creole Portuguese (i.e. those of the Cape Verde Islands and Guinea-Bissau) do not follow the same phonotactic rules as those spoken on the Gulf of Guinea islands”, and notes that in the Upper Guinean varieties, there is “no constraint against words ending in a consonant”, attributing this to different African „substrate‟ influence in the two regions (1988:113). In other words, in Cape Verde, the language is not based on a CV syllable structure, and therefore such features as syncope, epenthesis, etc. are not to be found in the development from Portuguese. On the other hand, final consonants and consonant clusters are permitted:

(46)

rapáz

< rapaz

„boy‟

búska

< buscar

„to look for‟.

61

However, there is a strong tendency to omit an initial vowel, by aphaeresis, - though not /o-/ ou /u-/, and /-i/ only rarely -, which to some extent does preserve a CV structure. :

(47)

kába

< acabar

„to finish‟

spéra

< esperar

„to wait for‟

gréja

< igreja

„church‟.

Gulf of Guinea Islands As Holm indicated (quoted above), the Gulf of Guinea Creoles have followed a CV structure, and therefore a set of phonotactic rules have come into play to avoid initial vowels, final consonants and consonant clusters. Aphaeresis occurs, with the omission of (normally) an unaccented vowel:

(48)

< andar

„to walk‟

< acabar

„to finish‟

(São Tomé)

< esquecer

„to forget‟

kéda

(Príncipe)

< esquerda

„left‟

skyi

(Annobon).

nda

(São Tomé)

nda

(Príncipe)

nda

(Annobon)

kabá

(São Tomé)

kabá

(Príncipe)

xabá

(Annobon)

kesé

Syncope, or the omission of one or more sounds from the middle of a word, occurs in the Gulf of Guinea Creoles, usually to avoid consonant clusters:

(49)

kõsé

(São Tomé)

< conhecer 62

„to know‟

buká

(Príncipe)

< buscar

„to look for‟

gáñi

(Annobon)

< grande

„big‟.

Apocope - omitting one or more sounds from the end of a word - is “an unproductive means of deletion, [and] occurs only in a few words” in São Tomé Creole (Ferraz 1979:39); nevertheless, it does also occur in Annobon:

(50)

v

(São Tomé)

< velho

al

(Annobon)

< areia

„old‟ „sand‟.

Epenthesis, the insertion of a sound in the middle of a word, “is quite widespread in the Atlantic creoles” (Holm 1988:110), normally with the addition of a vowel to break up a consonant cluster:

(51)

álima

(São Tomé)

álima

(Príncipe)

sógolo

(Annobon)

< alma

„soul‟

< sogro

„father-in-law‟.

Paragogue, or the insertion of a sound at the end of a word, occurs in the Gulf of Guinea Creoles, usually with a vowel to avoid a final consonant:

(52)

dósu

(São Tomé)

dósu

(Príncipe)

téi

(Annobon)

< dois

„two‟

< três

„three‟.

Metathesis, or changing the order of two sounds in a word, is relatively common in the Gulf of Guinea Creoles, often to break up a consonant cluster:

63

(53)

bsupla56

(São Tomé)

< véspera

„eve‟

kiryá

(Príncipe)

< criar

„to bring up‟

namína

(Annobon)

< menina

„child‟.

Sandhi rules Another notable phonological feature of the Gulf of Guinea Creoles - but not of Cape Verde Creole - is the use of sandhi rules, or the reduction of syllables in connected speech. For example, elision takes place between the final vowel of one word and the initial vowel of the following word:

(54)

(São Tomé)57 pe

„for him‟

dúdu di áwa >

dúdu dáwa

„jug of water‟

ku alé

>

kwalé

„with the king‟

da 

>

d

„to him‟

kabá 

>

kab

„finished it‟

>

déli

„for him‟

namenénsy

„those brothers‟.

pa e

>

(Príncipe) da éli (Annobon)58 namén nensyi >

56

Though one consonant cluster is broken up, another is created, but /Cl/ is a permissable combination in São Tomé Creole. 57 cf Ferraz (1979:28-29). 58 According to Post (1995:194), these sandhi rules are particularly productive in Annobon.

64

2.2.4 Parasegmentals This is another area of phonology which has been little studied in relation to creoles. As far as Portuguese-lexifier Creoles are concerned, the only serious attempts to analyse prosodic features are Günther‟s (1973:48-51) and Ferraz & Traill‟s (1981) discussion of Príncipe Creole (see below), as well as Couto‟s section on the subject in his description of Guinea-Bissau Creole (1994:76-80) and Maurer (1995:10-23) on Angolar. The principal distinction to be made here is between intonational, tonal and pitchaccent languages: Portuguese, in common with most European languages is an intonational language, while the majority of Niger-Congo languages of Africa - with the exception of Wolof and a few others - are tonal languages. It has been claimed that a number of creole languages are wholly or partially tonal, and significantly these are often classified as being „radical‟ also: Saramaccan and Ndyuka, maroon creoles in Surinam (but not the principal Surinam Creole, Sranan, which is not a maroon language), 65

Papiamentu, in the Netherlands Antilles, and Príncipe Creole. Therefore, it is important to assess to what extent the Creoles studied here have remained intonational or have shifted towards being tonal. The distinction between intonational and tonal languages has been defined as follows:

Basically tone is a feature of the lexicon, being described in terms of prescribed pitches for syllables or sequences of pitches for morphemes or words; whereas intonation is a feature of phrases or sentences. (Cruttenden 1986:8).

However, the distinction is not always clear-cut, and there are for example intonational languages with tonal features, such as Norwegian.

Pitch-accent languages (such as

Japanese, for example) belong to a category which is roughly intermediate between the other two - reference will be made to these below, in relation to Príncipe Creole. Characteristic of intonational languages such as Portuguese or English is the “occurrence of recurring pitch patterns, each of which is used with a set of relatively consistent meanings, either on single words or on groups of words of varying length” (Cruttenden 1986:9), as well as regular stress patterns on polysyllabic words. This is indeed the case in Cape Verde Creole, where intonational and stress patterns exist, similar to those in Portuguese: Veiga, in his analysis of the prosody of Cape Verde Creole, states (albeit somewhat ambiguously) that distinctive tones are not a feature of the language, attributing this characteristic to “some languages in Africa” (1982:64). What is clear is that one cannot find minimal pairs making tonal distinctions between words with identical phonology. According to Ferraz (1979:25), “there is no phonologically significant tone” in São Tomé Creole, but he adds that “as a stylistic or emphatic device, a high tone may be placed on the nucleus of a monosyllabic word or on the last two nuclei of a word of more than one syllable”. However, this “device” is more suitably classified as intonational, rather than tonal. It is concerning Príncipe that there has been a certain amount of controversy as to whether the Creole is a tonal language or not. Günther has claimed that the language has 66

high tones which correspond to stressed syllables in Portuguese, and low tones corresponding to unstressed syllables, and is thus a tonal language. (1973:48). He goes on to give examples of minimal pairs which illustrate this distinction (ibid.:50):

(55)

„beach‟

pá

(high tone)

< praia

pa

(low tone)

< para

„for‟

fá

(high tone)

< falar

„to speak‟

fa

(low tone)