John Holm-Languages in Contact The Partial Restructuring of Vernaculars (2004) PDF

This page intentionally left blank
Languages in Contact
The Partial Restructuring of Vernaculars
There is widespread agreement that certain non-creole vernaculars are

structurally quite different from the languages out of which they grew:
African American English, Afrikaans, Brazilian Vernacular Portuguese,
Nonstandard Caribbean Spanish, and the Vernacular Lects of Réunion
French. Until now, however, these languages have proved remarkably
resistant to the attempts of linguists to provide a plausible theory to
account for either their genesis or their synchronic structure. Informed
by the first systematic comparison of the social and linguistic facts in
the development of these languages, this book argues that the transmis-
sion of their source languages from native to non-native speakers led to
partial restructuring, resulting in the retention of a substantial amount
of the source languages’ morphosyntax, but also the introduction of a
significant number of substrate and interlanguage features. This study
concludes with the proposal of a formal theoretical model identifying
the linguistic processes that lead to partial restructuring, throwing into
focus a key span on the continuum of contact-induced language change
which has not been coherently analyzed up to now. It demonstrates how
the insights gained from the comparative study of such vernaculars cast
much-needed light on the relationship between the diachronic develop-
ment and synchronic structure of this important group of languages,
with some 200 million speakers.
John Holm is Chair, English Linguistics and Director, Graduate Pro-

gram in Descriptive Linguistics at the University of Coimbra, Portugal.
He has written widely on the issues surrounding contact languages,
and his previous publications include Introduction to Pidgins and Creoles
(Cambridge University Press, 2000) and Pidgins and Creoles (Cambridge
University Press, Volume 1, 1988; Volume 2, 1989).
Languages in Contact
The Partial Restructuring of Vernaculars
John Holm
  
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo
Cambridge University Press

The Edinburgh Building, Cambridge  , United Kingdom
Published in the United States by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521430517
© John Holm 2004
This book is in copyright. Subject to statutory exception and to the provision of

relevant collective licensing agreements, no reproduction of any part may take place
without the written permission of Cambridge University Press.
First published in print format 2003
ISBN-13 978-0-511-06755-6 eBook (EBL)

ISBN-10 0-511-06755-0 eBook (EBL)
ISBN-13 978-0-521-43051-7 hardback

ISBN-10 0-521-43051-8 hardback
Cambridge University Press has no responsibility for the persistence or accuracy of

s for external or third-party internet websites referred to in this book, and does not
guarantee that any content on such websites is, or will remain, accurate or appropriate.
For Mary, with love
Contents
List of maps page ix

List of tables x
Preface xi
Acknowledgments xvii
List of abbreviations xix
1. The study of partially restructured vernaculars 1

Introduction 1
1.1 Partial restructuring versus decreolization 4
1.2 The study of African American English (AAE) 10
1.3 The study of Afrikaans 13
1.4 The study of Brazilian Vernacular Portuguese (BVP) 15
1.5 The study of Nonstandard Caribbean Spanish (NSCS) 17
1.6 The study of Vernacular Lects of Réunionnais French
(VLRF) 19
1.7 The comparison of partially restructured vernaculars 21
2. Social factors in partial restructuring 24

Introduction 24
2.1 AAE: the sociolinguistic setting of its development 29
2.2 Afrikaans: the sociolinguistic setting of its development 41
2.3 BVP: the sociolinguistic setting of its development 47
2.4 NSCS: the sociolinguistic setting of its development 60
2.5 VLRF: the sociolinguistic setting of its development 65
2.6 Common sociolinguistic factors 70
3. The verb phrase 72

Introduction 72
3.1 The AAE verb phrase 73
3.2 The Afrikaans verb phrase 77
3.3 The BVP verb phrase 80
3.4 The NSCS verb phrase 83
3.5 The VLRF verb phrase 85
3.6 A comparison of the verb phrase 90
vii
viii Languages in Contact
4. The noun phrase 92

Introduction 92
4.1 The AAE noun phrase 93
4.2 The Afrikaans noun phrase 96
4.3 The BVP noun phrase 101
4.4 The NSCS noun phrase 105
4.5 The VLRF noun phrase 108
4.6 A comparison of the noun phrase 112
5. The structure of clauses 116

Introduction 116
5.1 The structure of AAE clauses 116
5.2 The structure of Afrikaans clauses 120
5.3 The structure of BVP clauses 123
5.4 The structure of NSCS clauses 127
5.5 The structure of VLRF clauses 129
5.6 A comparison of clause structure 133
6. Conclusions 135
Introduction 135
6.1 Social factors in partial restructuring 135
6.2 Linguistic factors in partial restructuring 137
6.3 Linguistic processes in partial restructuring 142
6.4 Comparative studies in restructuring 144
References 147
Index 166
Maps
1 United States: ‘Coloured Population, 1900.’ Originally

published in Darby and Fuller (1978), The New Cambridge
Modern History Atlas (Cambridge University Press). page 31
2 Southern Africa. Originally published in Holm
(1989), Pidgins and Creoles Volume 2 (Cambridge
University Press). 43
3 ‘South America, 1830–1956.’ Originally published in Darby
and Fuller (1978), The New Cambridge Modern History Atlas
(Cambridge University Press). 49
4 ‘Localización aproximada de áreas afro-hispanas en América.’
Originally published in Perl and Schwegler (eds.) (1998),
América Negra: panorámica actual de los estudios lingüı́sticos sobre
variedades hispanas, portuguesas y criollas (Iberoamericana).
Reproduced by kind permission of Iberoamericana. 61
5 (a) ‘Carte de l’ı̂le de la Réunion.’ Réunion Ethnic groups.
Originally appeared in the Chaudenson (1998), Créolisation du
français et francisation du créole: Le cas de Saint-Barthélemy et de
la Réunion (previously unpublished paper presented at the
Conference on Degrees of Restructuring in Creole Languages,
University of Regensburg, June 24–27, 1998). Reproduced by
kind permission of Robert Chaudenson.
(b) ‘Réseau ALRé.’ Réunion Lects. Originally appeared in the
Chaudenson (1998), Créolisation du français et francisation du
créole: Le cas de Saint-Barthélemy et de la Réunion (previously
unpublished paper presented at the Conference on Degrees of
Restructuring in Creole Languages, University of Regensburg,
June 24–27, 1998). Reproduced by kind permission of Robert
Chaudenson. 66
ix
Tables
1 Degree of restructuring and social factors page 26

2 Estimated population of (East) Virginia, 1685–1790 32
3 Estimated proportion of blacks in American colonies
in 1750 33
4 Estimated population of South Carolina, 1685–1775 35
5 Estimated population of the Cape Colony, 1658–1798 45
6 Estimated population of Brazil, 1538–1890 50
7 Estimated proportion of whites in the Greater Antilles in
the late eighteenth century 64
8 Estimated proportion of whites in various societies in the
late eighteenth century 71
9 Key morphosyntactic features in partial restructuring 138
x
Preface
The naming of books is a tricky business, and it has consequences. In the

case of this book – a comparison of five languages with a surprising com-
monality in structure and social history, and an account of the linguistic
processes that formed them – the first thing that may strike anyone who
knows my work is not so much the title, as what the title is not.
Since the late 1980s, I have used the term “semi-creolization” for
the process I discuss here. This book, however, is not called “semi-
creolization”. The change is in part strategic: I wish to make an argu-
ment which is, I think, original about the nature of these language vari-
eties, which all raise current and important issues of politics and culture.
I am not interested in exercising an imperial right to label, especially if
it obscures discussion about the issues raised by this book. Creole lan-
guage studies was one of the first post-colonial disciplines: Reinecke was
surely far ahead of his times in seeing creoles from the perspective of
their speakers rather than those who sought to be their speakers’ imperial
masters.
I am very aware that the social sciences’ long demand to label at will
is always problematical. I also know that the use of a new or unfamiliar
term in an established field is an irritant to others who, perhaps, have a
word they like better or even a certain resistance to thinking again about
issues of taxonomy.
“Semi-creolization” has been the traditional term for what I am about
to discuss since Reinecke first referred to Afrikaans as “semi-creolized”
(1937:559). As used among linguists, this term has the advantage of a cer-
tain transparency: its first morpheme clearly means “half ” or “partly” in
reference to the process of creolization. The problems that have emerged
in connection with this term have entirely to do with the second mor-
pheme.
Creole linguistics is part of sociolinguistics, which is the most politically
charged area of language study. At the very least, creoles involve the con-
tact of different cultures and thus cultural frontiers and cultural politics.
Usually the social setting of creolization has been that of enslavement,
xi
xii Languages in Contact
with all the racism and brute violation of human dignity on which slavery
depended. Nobody who studies creole languages can be indifferent to all
the resonances of their work, particularly among the people who speak
the languages we study. Mufwene (2000c:67) has written that
linguists’ self-licence to go around the world baptizing some vernaculars “cre-

oles”, when in some cases their speakers do not even know the word creole, let
alone how it is used in linguistics, is . . . the disenfranchising act by which some
vernaculars are marginalized . . .
The problem with this honorable concern is that all social scientists, to
some degree, work by setting the reality of other people’s lives within
a framework and a vocabulary that is particular to a discipline; it is in-
evitably a question of power, but it is also how analysis becomes possible.
To refuse that task is to refuse the science. Not every Englishman is glad
to know he speaks a “Germanic” language; but the classification has its
uses, even if it reflects the German origins of historical philology. Not
everyone knows what a “clitic” and a “copula” are, but these are terms
of art which linguists need to analyze language. And the disenfranchising
of a language, and of its speakers, may have more to do with economic
and political issues than the choices of labels made by social scientists far
away.
Still, Kautsch and Schneider are certainly correct in their assertion that
“Terms such as ‘semi-creoles’ . . . suggest certain undesired associations
and are found inappropriate by many” (2000:247). Linguists do indeed
have a duty to be sensitive to such questions. But these questions may
be exceedingly complex, and relative. Many creole-speaking areas are
now independent countries where creole-speakers themselves control the
media, and to that extent the local use of words; most of them seem to find
the term creole to have positive connotations. Some do not, particularly
in some areas where partially restructured languages are spoken.
Few words are truly neutral. The term Geechee (sometimes used in ref-
erence to Gullah) can have undesirable connotations in African American
English, i.e. “a Gullah or any Black person whose speech is peculiar or un-
intelligible to the hearer” (Folb 1980); “She was a big, burly-looking, dark
type sort of girl, a real geechy-looking girl” (quoted by Labov 1972a:390).
In fact, the American Dialect Dictionary defines Geechee as “a negro from
the islands, as from the Bahamas” (Wentworth 1944), while for Bahami-
ans, “A Geechee is what you could call a Merican who’s work field”
(Holm with Shilling 1982). And therein lies the analogy: a descriptive
word will carry the burden of attitudes to those it describes, and come in
time to be identified with those attitudes. Whatever terms linguists may
Preface xiii
use, many African Americans and Brazilians find semi-creole or semi-crioulo

unacceptable in reference to their own speech – while many Réunionnais
find semi-créole unacceptable because it is not creole enough.
For all these reasons, I have decided to refer to semi-creoles as “par-
tially restructured languages,” although I will continue to refer to creoles
as such, rather than as “fully restructured languages.” My logic is that
linguists still need to distinguish creoles from other fully restructured
languages, such as those that are intertwined (partially intertwined lan-
guages having never, to my knowledge, been reported). There may be
some loss of precision here, since I am not convinced that the processes
which produced the languages discussed in this volume are in all points
the same as those which produced partially restructured language vari-
eties such as Irish English (Hickey forthcoming, pace Winford 2000:216).
It is also true that the languages examined here share at least part of their
substrate with the Atlantic creoles, which means that they also share to
varying degrees a broader cultural heritage. The more neutral term does
not convey this cultural difference. But it may be that this cultural and
historical difference is why some find the term semi-creole unacceptable.
I believe that my choice of terms is not euphemistic; that is, it is not
an attempt to deny historical realities, like the social position of slaves,
in order to make the world seem more tasteful or untroubling than it
actually is. Thus I will continue to use one term with which linguists have
developed the most surprising problems: the world “creole” itself.
“Creole” is a term of art for linguists, as is “creolization”; we try to give
it a precise meaning in order to analyze and understand a particular kind
of language contact and change. It must clearly be something beyond
the obviously circular definition: “the process which produces a creole
language.” We think we mean something like McWhorter’s pithy defini-
tion: “radical reduction by non-natives followed by reconstitution into a
natural language” (2000:115), but the problems begin as soon as we try
to explain how and why this happens. It does sometimes seem that we
are more interested in debating the general processes even before we have
provided a proper account of the specifics involved, but this failing may
be understandable if not forgivable: there are so many creole languages
and they appear to have been produced by such a wide range of social
and linguistic circumstances that it often seems unclear precisely what a
creole language is, much less what the processes are that produce such a
language.
This has led to a singularly unhelpful confusion. Parkvall (2002:362)
has written that “for most creolists, there is nothing at all that sets creoles
apart from non-creoles.” If this is so, it has to do with the belief that
xiv Languages in Contact
the typological similarity that creole languages may or may not bear to
one another is analogous to the typological similarity of languages whose
structural affinities can be attributed to a family resemblance – that is, the
similarities found between genetically related families of languages such
as the Germanic or the Bantu languages.
The confusion undermines the fundamental and useful meaning his-
torical linguists give to the genetic relationship between languages when it
comes to discussion of the relationship of creoles to their superstrates.
Thomason and Kaufman (1988) maintain that because creoles (and the
pidgins from which they developed) were not passed from one genera-
tion of speakers to the next by normal language transmission, they are
not genetically related to the languages spoken by their creators – in
the technical sense used in historical linguistics, which requires system-
atic correspondences between languages not only in lexicon but also in
structure. This position has been attacked by Mufwene (2000a:9–11)
as “disenfranchising” the creoles, and by DeGraff (2001) for implying
that creoles are “abnormal.” Thomason (2002) responds that it would
nonetheless be unwise to abandon the traditional definition of genetic
relatedness: it is the central concept of the comparative method, which
has, for well over a century, enabled historical linguistics to accumulate
a vast and reliable body of knowledge about language change. Perhaps,
before attempting to dismantle a science, we should have reasons more
substantial than the possibility that others will misconstrue its technical
terms.
In any case, the category to which creole languages belong is sociolin-
guistic rather than typological in the genetic sense. To indulge again in the
tautology, creoles are creoles because they have undergone a sociolinguis-
tic process called creolization, not because they “attest the features that
define creoles as creoles” (Markey 1982:170). Because the study of At-
lantic creoles has dominated the field, and because these varieties happen
to share genetically related Indo-European superstrate and Niger-Congo
substrate languages as a historical result of the Atlantic slave trade, it has
perhaps come to seem that certain syntactic features found in most of
the basilectal Atlantic creoles are “creole features” in a sense compatible
with Markey’s. But this is not the case.
Creolization is a sociolinguistic process: there is widespread consensus
that its defining characteristics must include social as well as linguistic
phenomena (Holm 2000a:68–71). Because it is a sociolinguistic process,
like language attrition or code switching, there is no logical reason to ex-
pect that languages which have undergone creolization necessarily bear
any more typological similarity to one another than languages that have
not. The nub of the problem seems to be confusion over what “typology”
Preface xv
means in linguistics. Genetically dissimilar languages can, by coincidence,

be typologically similar (i.e. isolating, synthetic, etc.). In respect to its
paucity of inflectional endings, English is typologically closer to Chinese
than it is to Latin (Crystal 1986:319) despite the fact that genetically it is
closer to Latin. On the other hand, certain linguistic processes can also
result in closer typological similarity: two genetically unrelated languages
undergoing drift can eventually emerge with a similar degree of loss of
inflectional morphology. There is, indeed, abundant evidence that genet-
ically unrelated languages that are creolized undergo an extensive if not
complete loss of inflectional morphology.
Building on Chaudenson’s (1992) view of creoles as simply varieties
of their superstrate languages – approximations of approximations of the
earlier regional dialects spoken by European colonists – Mufwene (1994,
1996, 1997) concluded that creole is not a valid term for classifying lan-
guages, and therefore one language cannot be said to be more or less
“creole” than another. McWhorter (1998) countered that creoles are in-
deed synchronically distinguishable from non-creole languages in that
they combine all three of the following traits: they have little or no in-
flectional affixation; they make little or no use of tone lexically to con-
trast monosyllables or encode syntax; and they have semantically regular
derivational affixation. This creole prototype, graded like most phenom-
ena, is “the direct result of severely interrupted transmission of a lexifier,
at too recent a date for the traits to have been undone by diachronic
change” (1998:812). Later, McWhorter (2000) specified that the gradi-
ence in the proximity of creoles to the synchronic prototype that he had
proposed could be accounted for by sociohistorical factors such as the
continuing presence of the lexifier language.
This reaffirmation of the validity of the concept of “creole” has
nonetheless been challenged by the claim that the differences between
creoles and other languages, such as their lexifiers, are in fact the result
of differences in parameter settings, which determine various interrelated
morphosyntactic configurations in current generative theory. For exam-
ple, historical syntacticians such as Roberts (1999) have argued that in
languages with highly inflected verbs, the verb tends to move to I, the
locus of tense and agreement, whereas in languages with verbs having
few or no inflections (such as creoles) verbs tend to remain in place (but
note the strict verb-second rule of Afrikaans, a language with scarcely any
verbal inflections, section 3.2.1).
McWhorter (2002) further argues for the validity of creoles as a delin-
eable syntactic class with evidence that “the syntax-internal model . . . is,
by itself, inadequate to explain creole genesis” (2002:31). Distinguishing
between features that are universal to all languages for cognitive reasons
xvi Languages in Contact
(e.g. definite/indefinite opposition) and those which are “ornamental”

(e.g. grammatical gender), he points out a number of differences between
creoles and their source languages that cannot be explained by the syntax-
internal hypothesis. This is “because much of what distinguishes creoles
from older grammars is that, as recent descendants of pidgins, their gram-
mars are less accreted with such ‘grammaticalizational overkill’ ” of highly
particular, non-universal features (2002:9).
Whether or not McWhorter’s characterization of creoles as a delineable
synchronic class of languages is supported by future research, I must point
out here it will still be the case that creoles differ from languages that did
not result from contact: creoles develop much more rapidly. One of the
earliest creolists, Van Name, recognized this:
The changes which [creoles] have passed through are not essentially different in
kind, and hardly greater in extent than those, for instance, which separate the
French from the Latin, but from the greater violence of the forces at work they
have been far more rapid . . . here two or three generations have sufficed for a
complete transformation. (Van Name 1869–70:123)
For example, there could have been no Caribbean Creole French before
the French brought their first slaves to St. Kitts in 1626 and then captured
Martinique and Guadeloupe in 1635. Yet we find the existence of the
language documented by a text dated 1671 containing linguistic features
unequivocally identifying it with the modern creole of the latter islands
(Carden et al. 1991). The structural gap between French and Creole
French is at least as great as that between French and its source language,
vulgar Latin: it is the difference in the speed of formation – a lifetime
instead of a millennium – that is justification enough for distinguishing
creoles from languages whose genesis was not induced by contact.
All this confirms a simple point which has direct relevance here: the
concept of a “creole” remains an essential point of reference in our under-
standing of language contact and its relation to language change. Without
it, we have a less accurate, less truthful, view of language.
Acknowledgments
My greatest debt is to my former doctoral students at the City University

of New York (CUNY) for the insights on language restructuring that I
gained from their research. A number of talented young linguists partici-
pated in the seminars I organized at CUNY on partially restructured lan-
guages, some of whom were native speakers of the varieties they described
or the source languages involved. These seminars were extraordinarily
productive, leading to publications on American Indian English (Craig
1991), Dominican Spanish (Lorenzino 1993, 1998), Surinamese Dutch
(Healy 1993), and Afrikaans (Slomanson 1993; de Kleine 1997). Oth-
ers wrote relevant dissertations on Nagamese (Bhattacharjya 2000), Do-
minican Spanish (Green 1997), Brazilian Vernacular Portuguese (Mello
1997), and the Vernacular Lects of Réunion French (Chapuis forthcom-
ing). I am happy to have this opportunity to acknowledge my gratitude
to them.
I would also like to thank the following linguists for help and ad-
vice while I was working on this volume: Alan Baxter, Jeanette Béraha,
Hans den Besten, Ernesto d’Andrade, Raymond Hickey, Liliana In-
verno, Alexander Kautzsch, Alain Kihm, Ernst Kotzé, Dante Lucchesi,
John McWhorter, John Rickford, Paul Roberge, Edgar Schneider, Armin
Schwegler, Dominika Swolkien, Sarah Grey Thomason, Jean-Philippe
Watbled, and Walt Wolfram. I am also grateful to my editors at Cam-
bridge University Press, particularly Andrew Winnard and Helen Barton,
for their excellent support. Of course the present volume builds on my
earlier work, and I would like to thank once again those who helped me
with that. I am particularly grateful to Michael Pye for unearthing the
quilombo song (section 2.3.2), possibly one of the earliest examples of
restructured Brazilian Portuguese. However, responsibility for the errors,
omissions, and other shortcomings in this book is mine alone.
Finally, it is a pleasure to thank the following institutions for travel
grants that allowed me not only to exchange ideas at international meet-
ings, but also to do fieldwork and library research relating to language re-
structuring: from the Comité International des Études Créoles to present
xvii
xviii Languages in Contact
a paper at a conference on Réunion (2002); from the University of Puerto

Rico to give a mini-course in San Juan on restructuring (2001); from the
Linguistic Society of Southern Africa to present a paper at a conference
at the University of Cape Town (2000); from the University of Coim-
bra to present a paper at a round table discussion of the sociohistorical
origins of Brazilian Vernacular Portuguese in Recife (2000); from the As-
sociação Brasileira de Lingüı́stica to teach a course on language restruc-
turing at their XIV Instituto Lingüı́stico at the Universidade Federal de
Santa Catarina in Florianópolis (1999); from the Deutsche Forschungs-
gemeinschaft to present a paper at the International Symposium on De-
grees of Restructuring in Creole Languages at the University of Regens-
burg (1998); from the Colombian government to present a paper at a
colloquium in Cartagena (1996); and from Hunter College of the City
University of New York to present a paper with Gerardo Lorenzino at a
conference at the Universidad de Oriente, Santiago de Cuba (1992).
Abbreviations
A, AFK (standard) Afrikaans

AAE African American English
AAVE African American Vernacular English (term used in some
quotations)
ANT anterior tense
ART article
ASM agentive subject marker
AUX auxiliary
BVP Brazilian Vernacular Portuguese
CD Creole Dutch
CE Creole English
CF Creole French
CLP copula-like particle
CP Creole Portuguese
CS Creole Spanish
CV consonant-vowel
CVC consonant-vowel-consonant
D (standard) Dutch
DARE Dictionary of American Regional English (see Cassidy and
Hall, [eds.])
DEM demonstrative
DET determiner
DO direct object
E (standard) English
EP European Portuguese
F (standard) French
HP Helvécia Portuguese
IO indirect object
INFL inflection
LPV Lı́ngua dos Pretos Velhos (see section 1.4)
KA Kaaps Afrikaans
NEG negator
xix
xx Languages in Contact
NP noun phrase
NSCS Nonstandard Caribbean Spanish
OBJ object (marker)
ORA Orange River Afrikaans
P (standard) Portuguese
PAST past tense
PERF perfective aspect
POSS possessive
PRES present tense
PS Palenquero Spanish (section 1.5)
QW question word
REL relativizer
S (standard) Spanish; subject
SB Substrate
SBP Standard Brazilian Portuguese
SOV subject-object-verb word order
SVO subject-verb-object word order
SWVE Southern White Vernacular English
V verb
VLRF vernacular lects of Réunion French
VOS verb-object-subject word order
VP verb phrase
1s first person singular, etc.
3p third person plural, etc.
{} boundaries of an embedded clause
Note: the varying orthography used in this volume for VLRF and some
other language varieties is in each case that of the author cited.
1 The study of partially restructured
vernaculars
Introduction
Language is a kind of social behavior, one of the many ways in which
individuals interact with those around them. Thus linguistics is a social
science, and linguists take pride in thinking of themselves as scientists,
with all the objectivity that word denotes. Unfortunately, objectivity is
very hard to achieve, especially in the social sciences, and linguistics is no
exception. It is hard to imagine any study of language which manages to
put away all ideology, but in the case of the languages discussed in this
book, the task is unimaginable.
African American English – also called AAE, Ebonics, or just Black
English – is a good case in point. Until at least the middle of the twentieth
century, the overwhelming majority of white Americans saw their country
and its culture as the product of their European roots flourishing in a
new land. This ideology allowed very little room for the contribution
of other cultures, so that even the distinctiveness of the folk ways and
speech of African Americans was attributed to their frequent lack of access
to education and general ignorance – if not to their very intelligence.
Thus well into the 1950s Negro Nonstandard English (as AAE was
then called) was usually considered bad English in need of eradication
rather than study. In so far as its origins were considered at all, it was
assumed to have descended solely from British dialects that had been left
untended in America.
In the 1960s the civil rights movement sharply changed this ideology:
equal citizens could not logically be unequal human beings, and there
was a new willingness to reconsider African Americans, as well as the
development of their language and culture in the United States. By the
1970s there was widespread agreement – at least among linguists – that
the distinctive features of AAE identified it as a post-creole: the descen-
dant of a variety of English that had first been creolized or restructured
when it was learned by adult African slaves on plantations (as English
had been creolized in Jamaica, for example). Subsequently this speech
1
2 Languages in Contact
underwent decreolization, or the loss of many of its distinctive creole fea-

tures through contact with standard English. Dillard’s influential book,
Black English (1972) popularized this view, convincing many that AAE,
like its speakers, was much more African than anyone had realized. This
was part of another growing ideology, supported by many blacks, that
affirmed a very separate cultural identity for African Americans.
But there were problems in explaining AAE as a post-creole. Most im-
portantly, no one could find reliable historical evidence of the widespread,
stable creole from which AAE had supposedly decreolized. The known
passages purporting to represent the speech of blacks in North America
during the eighteenth and nineteenth centuries remain ambiguous. Quite
aside from the inherent problem of the authenticity of such fragments,
almost all of which were actually written by speakers of standard English,
there is an even greater problem in accurately identifying the kind of
speech represented. Unless the purported speaker’s background is docu-
mented, it is impossible to determine whether it represents the foreigner’s
English of Africans, the Caribbean Creole English of slaves imported from
the West Indies, a pidginized variety of English from West Africa, or an
indigenous creole such as Gullah, the fully restructured variety spoken
along the coast of South Carolina and Georgia.
Up to this point linguists had generally assumed that decreolization
could account for the varying distance between the grammatical structure
of different creoles and that of the European language they were based
on: AAE and Caribbean creoles based on English were viewed as post-
creoles at different stages of decreolization away from some very early
fully creolized variety. By the mid-1980s there were growing objections
to this all-or-nothing model of creolization and skepticism that it could
account for what was becoming known about the earlier structure of
AAE (Hancock 1987:264–265; Schneider 1989; Holm 1991:247). Much
of the most recent debate focuses on the nature the language of blacks
born in North America (outside of the creole-speaking Gullah area):
whether it was from its very beginning a fully restructured creole or rather
a compromise between the pidgin or creole brought in by slaves from
the West Indies and Africa and the regional speech of British settlers
(Winford 1997; Rickford 1997, 1999), and whether partial restructuring
can account for the known sociohistorical and linguistic facts concerning
AAE and some other languages that apparently had a similar genesis, such
as nonstandard Brazilian Portuguese and Caribbean Spanish, Afrikaans
(the South African language descended from Dutch), and the vernacular
French spoken on the island of Réunion in the Indian Ocean (Holm
1992, 2000).
These language varieties, which appear to have grown out of the partial
restructuring of older varieties spoken in Europe that came into contact
The study of partially restructured vernaculars 3
with non-European languages, today have some 200 million speakers,

placing them among the major languages of the world. They present
formidable challenges not only to linguistic theory but also in practical
matters like the language-related problems encountered in education by
speakers of nonstandardized varieties, which include all of the language
varieties discussed here, except for standard Afrikaans. These problems
have shown no signs of going away. And each of these languages has been
studied through the prism of particular, often local, ideologies, as Heliana
Mello has shown for her own language, Brazilian Vernacular Portuguese
(Mello 2001).
Of course the concept of a partially restructured language as opposed to
a post-creole (which was fully restructured but then decreolized through
contact with its lexical source language) has its own ideological impli-
cations. If the restructuring of the English spoken by blacks in most of
North America was only partial, this implies that the transmission of
the English language (and, indeed, other aspects of English culture) to
African Americans was much more complete than it has been fashion-
able to assume. The cultural separatism of the 1960s and 1970s may have
distorted the issue by insisting on the Africanness of African Americans
to the virtual exclusion of their Europeanness.
These languages, then, would require new study if only because our
sense of identity and ideology shift with time. But there is a more press-
ing scientific reason for reassessing them. The genesis and development
of such partially restructured languages have become one of the most
important leading edges of contact linguistics as a whole. The languages
discussed here have a number of the structural features of creoles but
appear, nonetheless, never to have undergone full creolization. Their re-
duced inflectional morphology – particularly in the verb phrase and noun
phrase – seems to have been transmitted from one generation to another
largely like that of unrestructured overseas varieties, rather than having
been reacquired by more basilectal varieties during decreolization, which
distinguishes them from post-creoles. Some of the most interesting re-
search in this area has been the effort to correlate the synchronic struc-
ture of these languages to the sociolinguistic history of their speakers: the
demographic balance of native versus non-native speakers of the target
language at the beginning of the speech community’s settlement, their
relative power, their migrations, and the nature of their contact.
There has also been a shift in theoretical perspective that is facilitating
progress in this area of inquiry. More of us working in pidgin and creole
linguistics are coming to see our field as only one part of a broader area of
research: contact linguistics, as defined by Thomason (1997). The scope
of this wider field includes language varieties that have resulted not only
from pidginization and creolization (to whatever degree) but also from
such processes as intertwining (Bakker and Muysken 1994), koineization,

or indigenization (Siegel 1997). Such studies promise to increase our
understanding of the range of possible outcomes of language contact by
encompassing varieties that fail to fit neatly into the definitional boxes in
which we have often tried to restrict pidgin and creole linguistics.
In addition to the five partially restructured varieties mentioned above,
which have received considerable scholarly attention, there are a number
of less well-studied varieties that seem likely to have undergone a similar
process, such as the nonstandard English of American Indians, Australian
Aborigines, and others. There are also partially restructured varieties
which appear to have evolved solely through community-wide language
shift, such as Irish English. Whether these are indeed the same kind of
language, which is the position of Winford (2000:216), has yet to be
demonstrated. Specialists in Irish English such as Hickey (forthcoming)
are not convinced (see section 2.1.1).
This chapter examines how scholarship on each of these five varieties –
based on five different European languages – has taken its own course, the
literature on each being largely in the corresponding standard language.
Although language barriers are still surprisingly effective in limiting the
horizons of linguists, there has been a certain amount of communication
across these barriers so that research on one variety has sometimes cast
light on theoretical problems connected with another. After surveying
general views on full and partial restructuring from the earliest creolists
until the 1980s (section 1.1), this chapter examines scholarship on each
variety, beginning with AAE (1.2). To a limited extent (especially in more
recent years) AAE studies have provided models for interpreting the his-
torical development of the other varieties, from (a) the model of a purely
European dialect reflecting general Western European tendencies such
as the loss of inflections; to (b) the model of a post-creole retaining sub-
stratal features; to (c) the model of differing degrees of restructuring,
varying according to social factors. This review of the theoretical under-
pinnings of research on AAE will then be compared with that of work on
the other four varieties: Afrikaans (1.3); Brazilian Vernacular Portuguese
(1.4); Nonstandard Caribbean Spanish (1.5); and the Vernacular Lects
of Réunionnais French (1.6). The final section (1.7) describes recent
comparative research in partial restructuring.
1.1 Partial restructuring versus decreolization

The theoretical foundations for the study of fully creolized languages have
been developing since the eighteenth century – particularly since the mid-
dle of the twentieth century (Holm 1988–89:13–70). However, linguists
have had more difficulty developing an adequate theoretical model for

dealing with partially restructured languages – one that would allow reli-
able predictions about the interrelationship between the social history of
their speakers and the linguistic structure likely to emerge from a partic-
ular context.
We have long known that fully creolized languages exist – languages
whose linguistic structure differs radically from that of the older lan-
guages from which they drew most of their lexicon. For example, the
generally synthetic structure of the Western European languages used by
colonists (Portuguese, Spanish, French, Dutch, and English – which still
use a number of inflections to convey grammatical information) was re-
placed by an analytical structure in the Atlantic creoles derived both from
these European languages and from the isolating Niger-Congo languages
spoken by Africans brought to the New World. There is fairly general
agreement that the isolating structure of the creoles – using free rather
than bound morphemes to convey grammatical information – was deter-
mined by several factors: (1) the tendency towards isolating structures
that was already widespread in the European superstrate languages; (2)
the almost categorical use of isolating structures in the African substrate
languages; (3) the universal tendency of adults to use isolating structures
when learning a second language (e.g. the pidgins that developed into
creoles); (4) the internal systematicity that would have spread the use
of isolating structures as the creoles developed; and (5) the converging
influence of two or more of these tendencies.
Of all the structural similarities of the Atlantic creoles, the common trait
that indicates most clearly the completeness of their restructuring is the
completeness of their analyticity. If we leave aside the non-Atlantic creoles
(which have not been compared as systematically), we find that basilectal
creoles – those closest to their earliest form – seem to have very few true
inflections, and that varieties that do have true inflections seem not to
be the same kind of language as basilectal Atlantic creoles (Holm 1989).
The existence of fully restructured creoles (whatever they may have
been called) has been acknowledged since the early eighteenth century,
and references to what can only be interpreted as more and less fully
restructured Caribbean varieties date from the latter part of that century:
die creolische, oder Negersprache . . . wird aber von den blanken Creolen feiner
gesprochen, als von den Negern. [. . . the creole, or language of the blacks . . . is
spoken better by the white Creoles than the blacks.] (Oldendorp 1777:263,
quoted by Stein 1984:92)
(Of course feiner, translated as ‘better,’ here means more like the
European source language.)
However, it was not until the late nineteenth century that a linguist ob-
served that there were language varieties that combined features of creoles
with those of non-creoles. Schuchardt (1889:480) coined the term Halb-
kreolisch (literally ‘half-creole’) for certain varieties of Indo-Portuguese
and Caribbean Creole French that had taken on superstrate features:
Ueberall wo eine kreolische Mundart gesprochen wird, liegt Mischung mit

der europäischen Muttersprache sehr nahe, mit anderen Worten: es stellt sich
leicht ein individuelles oder gelegentliches Halbkreolisch ein. Indem wir die
Bedingungen für dasselbe näher suchen, bemerken wir einerseits dass Europäer
die des Kreolischen nicht wirklich mächtig sind, sondern nur dunkle Vorstellung-
en davon haben, such bemühen von den Einheimischen verstanden zu werden –
kreolisiertes Europäisch; anderseits dass Europäer die des Kreolischen mehr oder
weniger mächtig sind, irgend eine Form der Darstellung wählen, für welche das
Kreolische nicht ausreicht, oder dass Kreolen die des Europäischen nicht mächtig
sind, ihren sprachlichen Ausdruck zu verfeinern sich bemühen – europäisiertes
Kreolisch.
(Wherever a creole dialect is spoken, mixture with the European mother tongue
lies very close at hand; in other words, an individual or occasional semi-creole
easily appears. When we look more closely into the underlying conditions, we see
on the one hand that [1] Europeans who do not really know the creole, having
only a confused notion of it, may strive to make themselves understood by the
natives, producing a creole-influenced variety of the European language. On the
other hand, [2] Europeans who are more or less at home in the creole may use
constructions not found in it, or [3] Creoles who have not mastered the European
language may attempt to refine their creole, producing a European-influenced
creole).
The first situation produces a variety similar to what Mühlhäusler

(1982:456–457) calls Tok Masta; the second situation produces a variety
like the lects of Negerhollands and Papiamentu spoken by Europeans;
the third produces what are now called decreolized varieties.
Schuchardt’s idea of Halbkreolisch was interpreted by Tagliavini
(1931:834) as a language that was half-way in the process of being cre-
olized, and so he translated the term into Italian as “lingue creolizzanti.”
Unfortunately the present-participial ending might suggest that such lan-
guages are “creolizing” in the sense of still undergoing restructuring;
Reinecke (1937:22) translated the term as “those tending toward the
creole, the creolisant dialects.”
Schuchardt also noted that African American English seemed to be
losing its creole features:
The Negro English that is most widely known is spoken in the southern United
States . . . those variants which still show a creole-like character are increasingly
falling into disuse by being accommodated to the English of the whites by means
of an intermediate speech variety. (Schuchardt c. 1893, in Gilbert 1985:42)
In this view, African American English originated as a full creole that later
acquired non-creole features from contact with regional English. Later
Bloomfield (1933:474) reasoned that a restructured variety of English
had become nativized
among Negro slaves in many parts of America. When the jargon has become
the only language of the subject group, it is a creolized language. The creolized
language has the status of an inferior dialect of the masters’ speech. It is subject to
constant leveling-out and improvement in the direction of the latter. The various
types of “Negro dialect” which we observe in the United States show us some
of the last stages of this leveling. With the improvement of social conditions, this
leveling is accelerated; the result is a caste-dialect . . . It is a question whether
during this period the dialect that is being de-creolized may not influence the
speech of the community – whether the creolized English of the southern slaves,
for instance, may not have influenced local types of sub-standard or even of
standard English.
This view was not elaborated into a full-blown theory of decreolization

until interest in AAE and the English-based Caribbean creoles became
widespread in the 1960s. Stewart asserted that
the non-standard speech of present-day American Negroes still seems to exhibit

structural traces of a creole predecessor. . . . One of the more important changes
which have occurred in American Negro dialects during the past century has
been the almost complete de-creolization of both their functional and lexical
vocabulary. (1968:51–52)
DeCamp (1961, 1971) developed the idea of a continuum of lects for

Jamaican, ranging from the most creole-like to the most English-like.
Stewart (1965) applied this idea to African American English, introduc-
ing the terms acrolect for the variety closest to the standard and basilect
for the variety furthest from it, with mesolect for those between. Later the
continuum model was further refined by others (e.g. Bickerton 1973,
Rickford 1987).
By the end of the 1970s there was a general assumption that decreoliza-
tion explained the varying structural distance between different creoles
and their lexical source language: Caribbean creoles based on English, for
example, were actually post-creoles at different stages of decreolization
away from a very early fully creolized variety that may have resembled
the modern Surinamese creoles, which were cut off from contact with
English in the seventeenth century.
The idea behind the modern meaning of partial restructuring origi-
nated in Hesseling (1897), who pointed out that “the Dutch on the Cape
was on the way to becoming a sort of creole . . . [but] this process was not
completed” (1979 translation, p. 12). Shortly afterwards, Vasconcellos
noted that
les Portugais ont été obligés d’apprendre quelquefois les langues indigènes, et
les indigènes d’apprendre la langue du Portugal. Le second fait est le seul qui
m’intéresse pour le moment, parce qu’il en est résulté la formation des dialectes
créoles, et d’autres variétés du portugais; entre les uns et les autres, on peut admettre
des degrés (1901:157–158; my emphasis)
[The Portuguese were sometimes obliged to learn the indigenous languages and
the indigenous people Portuguese. The second fact is my only interest for the time
being because it resulted in the formation of creole dialects and other Portuguese
varieties. Between the two groups, one could say there is a question of degree.]
The first recognition of a whole category of such languages can be found

in Reinecke (1937:61):
In several instances the slaves were so situated among a majority or a large minority
of whites (and there were other reasons as well for the result), that they, or rather
their creole children, learned the common language, not a creole dialect; or the
plantation creole dialects that had begun to form never crystallized, never got
beyond the makeshift stage. This happened in . . . Brazil, Cuba and the Spanish-
speaking Caribbean countries in general, and in the southern United States in
general.
Reinecke was also the first to put this meaning together with the term
semi-creolized, which he used in reference to Afrikaans (1937:559). He
also pointed out that the English-based creoles of the Caribbean did not
seem to have been completely restructured:
The Surinam dialects, like West African Pidgin English, are unmistakably creole
dialects in the sense of being simplified to a purely analytic structure. The other
West Indian dialects are not, however, so completely pruned down [. . . and] may
be regarded as what Schuchardt called creolizing languages – dialects on the way
to complete analytic simplification, but which for various reasons stopped a little
short of it. (1937:274–275)
As recently as l962, Stewart considered Suriname to have the only real

creoles based on English in the Caribbean area: “Jamaican and other
regional varieties of English are best treated as dialects of English”
(1962:50–51). In a personal communication, Stewart explained that at
the time it seemed more prudent to exclude these varieties from the dis-
cussion of creoles since it was unclear whether they were creoles that had
acquired non-creole features or vice versa. By 1967, however, he felt con-
fident that additional historical sociolinguistic information had made it
clear that the West Indian varieties were in fact post-creoles.
However, the fact that this view came to be widely accepted among cre-
olists does not in itself prove that Reinecke had not been right – that these
varieties had never been as fully creolized as the Surinamese varieties. An
additional possibility that could explain the considerable structural gap
between the Surinamese and West Indian varieties of creolized English is

that Sranan may have been repidginized in the late seventeenth century,
leaving it even further from English than it had been prior to 1667. Bloom-
field (1933) had indirectly implied that a non-creole language might take
on creole features (a process that could lead to partial restructuring) when
he asked, “whether the creolized English of the southern slaves, for in-
stance, may not have influenced local types of sub-standard or even of
standard English” (1933:474).
Later Silva Neto (1950a:12) followed Schuchardt (1889) in referring
to re-lusitanized Indo-Portuguese as a semi-crioulo. That same year he
extended the use of the term to the Portuguese spoken by non-whites
during the early settlement of Brazil:
constituiu-se, no primeiro século da colonização (1532–1632), na bôca de ı́ndios,
negros e mestiços, um falar crioulo ou semi-crioulo. [. . . there arose during the
first century of colonization (1532–1632) a creole or semi-creole language used by
Indians, blacks and people of mixed race] (1950b:166)
Although Silva Neto never spelled out the sequence of social and lin-
guistic events that may have led to the partial restructuring of a language
variety from the very beginning of its existence, this possibility struck me
as worth exploring when I was working on the same problem of the de-
velopment of Brazilian Vernacular Portuguese (Holm 1984). Thomason,
who was then working on a comparative study of a number of differ-
ent kinds of languages resulting from contact (Thomason and Kaufman
1988), agreed that it would be useful to reserve the term semi-creole for
those varieties that appeared never to have been fully creolized. Therefore
I contrasted the term with
post-creole varieties such as (according to some) American Black English . . . or
vernacular Brazilian Portuguese. . . . Others would call these varieties semi-creoles,
which also means that they have both creole and non-creole features but does
not necessarily imply that they were ever basilectal creoles, since both creoles and
non-creoles (e.g. Caymanian English . . .) can become semi-creoles by borrowing
features. Thus some believe that Afrikaans . . . particularly the variety spoken by
some people of mixed race . . . could safely be called a semi-creole but not a
post-creole (Holm 1988–89:9–10)
The term is also used in this sense by Thomason and Kaufman in

reference to Afrikaans (1988:148). Around the same time, Mufwene
(1987:99) noted that
the results of half-creolization and decreolization may look alike, but the processes
responsible for the structural likeness of their outcomes are certainly not the same.
Whichever is the case for B[lack] E[nglish] still needs to be demonstrated.
Bickerton (1984:176–178) proposed what he called a pidginization in-

dex to explain why the structure of some creoles is quite close to that
of their lexical source language (e.g. Réunionnais) while that of others
is quite far from it (e.g. Saramaccan). Although the mathematical for-
mula which he proposed to indicate the degree of restructuring proved
“unworkable” (Singler 1990:645), Bickerton did recognize that creoles
stand at different distances from their source languages in terms of the
degree of restructuring that they have undergone, and that this differen-
tiation could occur at the beginning rather than the end of the process of
restructuring (see the introduction to chapter 2).
It was during this period that linguists began to question whether de-
creolization alone could adequately account for the varying distance of
the structure of different creoles from that of their lexical source language.
Hancock (1987) put it thus:
I do not, then, believe that, for example, Black English was once like Gullah,
or that Gullah was once like Jamaican, or that Jamaican was once like Sranan,
each a more decreolized version of the other along some kind of mystical contin-
uum. . . . My feeling is that most of the principal characteristics that each creole
is now associated with were established during the first twenty-five years or so
of the settlement of the region in which it came to be spoken: Black English has
always looked much the way it looks now . . . (1987:264–265)
The theoretical importance of gradience in creolization was signaled

by a conference on “Degrees of Restructuring in Creole Languages” at
the University of Regensburg in Germany in 1998, resulting in an entire
volume on this topic (Neumann-Holzschuh and Schneider 2000).
1.2 The study of African American English (AAE)

The decreolization theory for the origin of Black English – the “creolist”
theory that finally received the imprimatur of Labov (1982) – was a much
more satisfactory explanation for that variety’s creole features than ear-
lier hypotheses that traced its origins solely to British dialects. However,
my own work on the lexicon of two much more restructured varieties –
Nicaragua’s Miskito Coast Creole English (Holm 1978) and Bahamian
Creole English (Holm with Shilling 1982) – made it clear to me that
archaic and regional British English must have played a primary role in
the genesis of all three African American varieties. Research on possible
British origins of specific creole grammatical features had been unfash-
ionable in the 1970s, but in the 1980s two such studies – Schneider 1981
(translated in 1989) and Rickford 1986 – had an important impact on
the field, reopening the question of the degree to which British syntactic
patterns had been preserved in African American varieties.
I was further led to question some of the basic assumptions of decre-

olization theory through work with several non-creoles that appeared to
have acquired creole features: White Bahamian English (Holm 1980),
and Caymanian and Bay Island English (Washabaugh and Warantz in
Holm 1983). I concluded that
Although long contact with creolized varieties of English has influenced the
English spoken by white Caymanians and their kin on the Bay Islands of
Honduras, this influence seems to be confined largely to areal contact phenom-
ena such as word-borrowing and phonological shifts. Considering the English
system of verbal inflections in the speech of Utila . . . as opposed to the system
of preverbal tense and aspect markers that characterizes C[entral] A[merican]
E[nglish] creoles . . . the former would seem to be not a creole but rather a re-
gional variety of English influenced by contact with creolized English, much like
the folk-speech of the southern United States. (Holm 1983:15)
In 1986 there began a debate as to whether AAE and white varieties of

American English were historically converging (through the decreoliza-
tion of AAE) or diverging (through AAE’s increasing isolation) as argued
by Labov and Harris (1986) (see section 2.1.7). The latter interpretation
seemed to support the implausible view of Poplack and Sankoff (1987)
that early nineteenth-century AAE had been more similar to white va-
rieties than current AAE is. However, what convinced me that decre-
olization alone could not account for the present structure of AAE was
listening to tape recordings of the speech of former slaves (Bailey et al.
1991, see section 2.1.5). Even taking into account that their speech may
have shifted considerably between their childhood in the mid-nineteenth
century and the time they were recorded in the 1930s and later, it was
clear that what I was hearing was a variety of English with some creole
features rather than a variety of creole with some English features. The
only honest conclusion that I could reach was that
The present study supports the view that the language of the ex-slaves, like earlier
attestations of the speech of blacks in the American South, indicates in the light
of the relevant sociohistorical and demographic data discussed above that the
language of blacks born in North America (outside of the Gullah area) was from
its very beginning a semi-creole representing a compromise between the creole of
slaves imported from the West Indies and the regional speech of British settlers.
While American Black English has certainly undergone decreolization over the
past 300 years in the sense that it has replaced many of its original creole features
with those of English, this is not actually evidence that American Black English
itself ever constituted an autonomous creole system. (Holm 1991:247)
A more radical view (which seems inherently unlikely, given what is

known about language contact phenomena) is that the very concept of
decreolization is misguided, and that it played no role in the development
of varieties such as Gullah and AAE, which stand at differing distances

from English structurally solely due to their having undergone differing
degrees of restructuring. Mufwene (1991:382–383) seems to support
such a view.
Schneider (1990) re-examined the idea of “creoleness” as a graded
phenomenon in reference to varieties of English and English-based cre-
oles in the Caribbean area with a view to casting light on the debate over
the creole origin of AAE. He concluded that the question as to whether
or not a particular variety is a creole can be very difficult to answer:
There is a variety of constitutive factors that contribute independently to the
notion, and the label applies to some language varieties better than to others,
without implying that the latter are necessarily “non-creoles.” We may distinguish
prototypical, or full, creoles that combine all or almost all of these features from
varieties that are less typical of the category. Even the notion of semi-creoles
does not seem to be very helpful in this dilemma, because its applicability, if not
defined too loosely, seems limited, and should not be taken to include the non-
prototypical – but nevertheless true – creoles . . . In linguistic matters, more and
less are frequently more appropriate responses than yes and no. (1990:105–106)
While Schneider considered the term semi-creole unhelpful because of the

limited number of languages it could be applied to (despite the numer-
ical importance of the speakers of partially restructured languages, as
discussed above), Kaye dismissed the validity of the very notion with an
analogy beyond the reach of logic:
There can be no such thing, of course, as partial pidginization or partial creoliza-
tion (this is why the terms post-creole, semi-creole, and creoloid are imprecise),
just as there is no such thing as partial pregnancy. (Kaye 1990:301)
More recent work on AAE has focused increasingly on those sociolin-

guistic factors which have long been considered relevant to the study of
full creoles (e.g. demographic figures suggesting the proportion of native
versus non-native speakers during the early period of language contact)
but which have not been systematically explored for AAE until now.
Winford (1997) traces the social histories of Virginia and the Carolinas,
citing early demographic figures from Wood (1989), and compares the
key structures in Gullah, AAE, and Southern White Vernacular English,
concluding that “AAVE was never itself a creole, but it was created by
Africans, and bears the distinctive mark of that creation.” Rickford (1997,
1999) has followed a similar methodology and reached a similar conclu-
sion; Mufwene also suggests that AAE “may simply have resulted from
a restructuring which was not as extensive as what produced Gullah”
(2001). However, Mufwene (2000b) lends his credibility as a creolist to
support the position of Poplack (2000:1) that “. . . the many grammatical
distinctions between contemporary varieties of AAVE and American and

British English are relatively recent developments,” i.e. not the result
of earlier contact with restructured varieties of English. Such a position
allows for less external influence on the development of AAE than the
apartheid-era linguists in South Africa allowed on the development of
Afrikaans (section 1.3).
Hackert and Holm (1997) have shown that the only hard evidence
ever offered for the full creolization of AAE resulted from a historical
misinterpretation:
the creole nature of the folk speech on the southern Bahamian islands should not
be interpreted as evidence that AAVE had been fully creolized on the mainland be-
fore 1780 (and later decreolized) since the language that was brought there was in
all likelihood eighteenth-century Gullah rather than eighteenth-century AAVE.
1.3 The study of Afrikaans

Afrikaans, derived from Dutch, is spoken by some 6 million South
Africans; about half are white and the rest are of mixed ancestry. Afrikaans
is unique among the language varieties examined here in that it was stan-
dardized and made an official state language. Its exhaustive documen-
tation makes it much easier to contrast its structure to that of its lexical
source language, which was actually seventeenth-century regional and
nautical varieties of Dutch. Also of particular relevance to tracing the
development of Afrikaans are its nonstandard varieties spoken by various
groups, particularly those of mixed race with little education.
The history of the study of Afrikaans and its origins has been sum-
marized by Reinecke et al. (1975:323ff.) and updated and expanded by
Roberge (1994), the sources of much of the following. Hahn (1882)
claimed that although Afrikaans is “phonetically Teutonic, it is psycho-
logically an essentially Hottentot idiom. For we learn this patois first from
our nurses and ayahs. The young Africander on his solitary farm has no
other playmates than the children of the Bastard Hottentot servants of his
father, and even the grown-up farmer cannot easily escape the deterio-
rating effect of his servants’ patois.” Viljoen’s 1896 dissertation, focusing
mainly on the phonetic system of Afrikaans, claimed it was derived from
the dialects of North Holland. Hesseling (1897) provided the first ex-
tended discussion of the origins of Afrikaans. Although he recognized
the influence of Hottentot (now called Khoi), he emphasized the in-
fluence of the Malayo-Portuguese creole of early Indonesian slaves and
claimed that “the Dutch on the Cape was on the way to becoming a sort
of creole . . . [but] this process was not completed” because of the con-
tinuing influence of metropolitan Dutch (1979 translation, p. 12). This
characterization set off a debate that continued for a century, at times

with considerable heat.
Hesseling’s Malayo-Portuguese theory was adopted by the Afrikaner
Du Toit (1905) and later developed further by Valkhoff (1966, 1972).
It was opposed by the “spontaneous development” theory initially pro-
posed by the Dutch linguist Kruisinga (1906), who saw Afrikaans evolv-
ing early on out of seventeenth-century Dutch dialects through what was
essentially normal language transmission. This model was taken up by
Afrikaans-speaking linguists such as Boshoff (1921, 1959) and Smith
(1927, 1952), who agreed that Afrikaans had developed according to
trends already present in earlier Dutch dialects under minimal influence
from other languages. The Dutch linguist Kloeke (1950), usually in-
cluded in this camp, attributed a strong “founder effect” to the South
Holland speech of the first Dutch colonists. The spontaneist model was
later revived in a more drastic form by Van der Merwe (1963, 1968), who
went so far as to claim that Afrikaans emerged within a half dozen years
after the colonists’ arrival (1968:66) due to accelerated drift, and ruled
out the possibility that people of color had influenced it in any significant
way (1968:29).
The approach of Bosman (1923, 1947) is considered eclectic by
Reinecke et al. (1975:323), who note that “this view admitted foreign
influence, chiefly from Low German colonists and Hottentots, but did
not admit a situation favorable to outright creolization (unless of the
Dutch spoken by Hottentots).” What den Besten (1987) calls the “South
African philological school” came to prevail in that country from the
1960s until majority rule in 1994. Its leading writers were Scholtz (1963,
1980) and Raidt (1974, 1983, 1991), who concerned themselves less
with the origins of Afrikaans as such than the history of specific linguis-
tic phenomena. However, their underlying theoretical model was that of
ordinary language change within varieties of Dutch accelerated by the
influence of non-native speakers in a multilingual setting, whose speech
was influenced by their first languages and had interlanguage features,
but never underwent outright pidginization and creolization. One of the
most complete histories of Afrikaans is Ponelis (1993), which stresses the
restructuring resulting from imperfect second language acquisition.
During this period, both black and white South African linguists be-
gan examining nonstandard varieties of Afrikaans more closely for the
light they might cast on the issue of the language’s origins, including
“Coloured” Afrikaans and Flytaal (Makhudu 1984), Malay Afrikaans
(Kotzé 1989), and Orange River Afrikaans, including Griqua Afrikaans
(van Rensburg 1984, 1989) – the last variety having been studied in a
book-length work by Rademeyer (1938).
What might be called the “Amsterdam school” of Afrikaans scholars

has evolved around den Besten (1985, 1986, 1993) and his colleagues,
who have focused on the effects of contact with Khoi and other lan-
guages. They see the South African school as antiquatedly Eurocentric
in its approach: “If a feature can possibly be European, then it must be
European,” provoking the opposite caveat regarding Valkhoff ’s approach:
“If a feature can be a creolism, it must be a creolism” (Roberge 1994:40).
Now that South Africans are reassessing their cultural identity with the
advent of majority rule, the composite identity that “creolism” suggests
has become increasingly attractive, and the ideological pendulum in lin-
guistics may now be swinging wide of the mark in that direction. At a con-
ference workshop on Afrikaans sociohistorical linguistics at the University
of Cape Town (Mesthrie and Roberge, 2001–02) a reference was made
to what distinguishes “. . . Afrikaans from other creole languages”
(Holm 2001:353).
1.4 The study of Brazilian Vernacular Portuguese (BVP)

The history of the study of BVP from a language-contact perspective has
been outlined by Holm (1987) and updated by Mello (1997). The simi-
larity of some of BVP’s structural features to those of Portuguese-based
creoles was pointed out over a century ago by Coelho (1880–86 [1967]),
who concluded that “it shows a tendency towards creolization” (p. 170),
but many Brazilian linguists still resist the view that the development of
BVP involved significant restructuring. They have done studies of how
the Brazilian lexicon has been influenced by indigenous languages like
Tupi (Sampaio 1928; Marroquim 1934) or African languages (Raimundo
1933; Mendonça 1933 [1973]), but in general they have followed the ad-
vice of Melo (1946), who cautioned against exaggerating the importance
of such external influences when parallels could be found in archaic or
regional usages in Portugal. However, Silva Neto (1950b:131) asserted
that creole and what he called “semi-creole” (semi-crioulo) varieties of Por-
tuguese had existed in Brazil, defining the latter as closer to the European
variety but not speculating as to how they had evolved. Révah (1963)
discounted substratal influence on BVP in favor of a general tendency
towards simplification of morphology in Western European languages, a
line of thought taken up later by Naro and Lemle (1976). They assumed
that BVP was in the process of losing number agreement rules, which
were being obscured by certain phonological rules.
Among non-Brazilian linguists, Valkhoff (1966) identified BVP fea-
tures shared by Portuguese-based creoles as evidence of the latter’s influ-
ence on it. Jeroslow did a detailed study of a rural dialect (1974) that led
her to suspect prior creolization (McKinney 1975). Guy (1981) exam-

ined the same BVP phenomenon as Naro and Lemle (1976) but reached
the opposite conclusion, i.e. that number agreement in BVP was spread-
ing as a final stage in decreolization, comparable to that of AAE in the
United States. In 1981 the Brazilian linguist Celso Cunha called for the
study of BVP from the perspective of modern creole studies, the goal
of Holm (1984, 1987, 1992b), who concluded that partial restructur-
ing was clearly evident in the BVP varieties of Helvécia (Silveira Ferreira
1985) and Ceará (Jeroslow 1974), and began attempting to work out the
development of BVP as the product of this process.
This approach has been taken up by Baxter (1992, 1997), who evalu-
ates the importance of creole-like features through quantitative methods,
and in the recent work of some Brazilian linguists such as Couto (1997)
and Careno (1997). The most comprehensive of these is Mello (1997),
who concludes that
creolization and partial restructuring did not occur throughout colonial Brazil,
but mainly in isolated areas which favored these processes . . . later decreoliza-
tion through contact with B[razilian] P[ortuguese] occurred. In most of settled
Brazil, the likeliest scenario was a process of imperfect language shift to Por-
tuguese by the African and Amerindian populations and their descendants. This
shift led to the establishment of BVP as the predominant dialect of Portuguese.
However, as the shift was taking place, substratum structural features and inter-
language patterns were transferred to the target language, becoming fossilized.
(Mello 1997:270)
Lucchesi (2000) basically takes the same position, but calls BVP the
product of “irregular language transmission” that was mais leve (‘lighter’)
than full creolization. Bonvini (2000) describes a creole-like lect of BVP
called Lı́ngua dos Pretos Velhos (LPV) traditionally used by Brazilian prac-
titioners of candomblé religious ceremonies for the light it could shed on
earlier language contact.
Studies approaching BVP from a language-contact perspective have
contributed to the growth of Afro-Iberian linguistics as a distinct field.
For obvious historical reasons, scholars working in Afro-Portuguese and
Afro-Hispanic studies are natural allies, and there has been a movement to
join the two camps since the first conference on Portuguese-based creoles
was held in Lisbon in 1991 (d’Andrade and Kihm 1992). Since then the
journal Papia: Revista de Crioulos de Base Ibérica has been published in
Brazil in both Portuguese and Spanish, encouraging further research in
the coalescing field. An international colloquium on creoles based on
Portuguese and Spanish in Berlin (Zimmermann 1999) was an important
forum for debating the role of partial restructuring in the emergence of
both BVP and NSCS, as is the recently founded Associação: Crioulos de

Base Lexical Portuguesa e Espanhola.
1.5 The study of Nonstandard Caribbean Spanish (NSCS)

NSCS is spoken by a substantial portion of Puerto Ricans, Dominicans,
Cubans, and coastal Venezuelans and Colombians, as well as many of
the Spanish speakers of New York City and Miami. Research since the
1960s indicates that a number of features in these varieties have parallels
in Spanish-based Caribbean creoles. An overview of these studies can be
found in de Granda (1975, 1987, 1998) and Green (1997), the sources
of part of the following summary.
One of the earliest references to external influence on a variety of
Caribbean Spanish is that of Sandoval (1627), who describes the lan-
guage spoken by Africans on the coast of what is today Colombia as
“corrupt Spanish . . . influenced by the Portuguese they call the language
of São Tomé.” This and the emergence of Palenquero Creole Spanish
nearby provide sufficient evidence that a Spanish-based pidgin built on
Afro-Portuguese did in fact exist in the Caribbean, but it does not confirm
the speculation of Bickerton and Escalante (1970:262) that there existed
“a Spanish-based creole spoken in many parts of the Caribbean during
the sixteenth and seventeenth centuries,” since there is no evidence that
pidgins or jargons spoken elsewhere in the the region ever developed into
stable creoles.
One such pidgin or jargon was the habla bozal spoken by the large influx
of Africans brought to Cuba (and elsewhere) in the first half of the nine-
teenth century to work on sugar plantations (section 2.4.2). Pichardo
(1862: vii, iii) described their “mutilated Castilian, without concord,
number declension or conjugation,” but noted that “Negroes born in
Cuba talk like the local whites.” Van Name (1869–70:125) referred to it
as only “the beginning of proper Creole” – an assessment later confirmed
by Reinecke (1937:271). Van Name was also among the first to recognize
that Curaçao’s Papiamentu was a creole language rather than a dialect of
Spanish.
Although Cuba’s habla bozal was never nativized as a creole, it did leave
its traces in the local vernacular; Ortiz (1924) documented its lexicon of
African origin. Still, most linguists in the Spanish-speaking Caribbean re-
mained reluctant to admit any significant influence of African languages
on local Spanish. Henriquez Ureña (1940:130, 169), for example, de-
scribed Dominican Spanish as having no more words of African ori-
gin than did general Spanish; he argued for an Andalusian origin for
the alternation of /r/ and /l/, seeing African influence only in the loss of
syllable-final -s. Most linguists of this period who recognized the possibil-
ity of external influence on local Spanish, such as Wagner (1949), were not
from the Caribbean. Cabrera (1954) was an exception; a Cuban anthro-
pologist who interviewed older people of African descent around 1930,
her portrayal of their Spanish is considered accurate, revealing a number
of African-like features both in their normal speech and in the special
language thought to resemble that of their ancestors, used in religious
contexts.
An equally exceptional linguist was Alvarez Nazario (1961), who ex-
amined early texts of bozal Spanish in Puerto Rico, concluding it was a
“criollo afroespañol” linking local Spanish (especially that of black com-
munities such as Loiza Aldea outside San Juan) to African languages via
an Afro-Portuguese pidgin. His work served to encourage others studying
Caribbean Spanish from the approach of contact linguistics, and he is now
recognized as one of the principle founders of Afro-Hispanic linguistics
(Ortiz 1999). Another founder, de Granda (1968), identified the speech
of Colombia’s Palenqueros as the New World’s other Spanish-based cre-
ole, and went on to identify features from African and restructured lan-
guages in NSCS, focusing on the theory of an early pan-Caribbean creole
that gradually decreolized (1970, 1976, 1978).
Meanwhile, Otheguy (1973), working from Cabrera’s Cuban data,
identified certain phonological and morphosyntactic traits in the vernac-
ular that had survived from the habla bozal as being specifically creole
features supporting the pan-Caribbean creole hypothesis (although he
has since retreated from this position). Ziegler (1976, 1977) linked the
bozal Spanish of Puerto Rico to that of Cuba, also arguing for decreoliza-
tion. Megenney has focused on non-Peninsular features in the vernacu-
lars of coastal Colombia (1976), Venezuela (1985), and the Dominican
Republic (1990), as well as African-derived vocabulary used in religious
rites in Cuba and Brazil (1999). Lipski, coming from within the creolist
camp, has offered counter arguments to the pan-Caribbean creole theory
(1993, 1994), seeing substratal influence as more likely to have come into
Caribbean Spanish through imperfect second language acquisition, and
bozal Spanish as never having undergone complete creolization since it
was not nativized (2000).
Schwegler, another creolist but one working primarily on Palenquero
(1993, 1996a), has also studied the effects of restructuring on Caribbean
Spanish (1996b). Schwegler and Morton (2002) document the features
of the NSCS of bilingual speakers of Palenquero CS (PS), casting crucial
new light on the link between restructured varieties like bozal Spanish and
modern NSCS. Schwegler helped to organize one of the first international
conferences on Palenquero and NSCS (Moñino et al. 2002) as well as the

first book-length survey of Afro-Hispanic linguistics (Perl and Schwegler
1998). Perl, coming from Afro-Portuguese studies, has worked on the
Cuban vernacular from a creolist perspective (Perl 1985, 1988, 1989),
as have Ortiz (1998) and Figueroa (1998). Other younger scholars who
have dealt with varieties of NSCS as a the product of partial restructur-
ing include Alvarez (1990), Lorenzino (1993, 1998), and Green (1997),
the last describing a hitherto unknown basilectal variety of Dominican
Spanish.
1.6 The study of Vernacular Lects of Réunionnais

French (VLRF)
The vernacular French of Réunion, a small island in the Indian Ocean, is
spoken or understood by most of the 500,000 inhabitants. Although lo-
cally called créole, its structure seems to be descended mainly from that of
seventeenth-century French dialects, including maritime varieties; how-
ever, it has a number of features also found in creoles. The creole or
non-creole identity of Réunionnais has long been the subject of consid-
erable debate; this identity is the focus of the following brief review of
the literature, based largely on Chapuis (forthcoming), rather than the
other main point of contention, which is the historical relationship of
Réunionnais to the fully restructured Ile de France creoles of Mauritius
and the Seychelles.
Adam (1883) grouped Réunionnais with the French-based creoles:
“Creole is the adaptation of French . . . by and for the slaves of Africa . . . in
the Antilles, in Réunion and in Ile de France.” Schuchardt (1885, trans-
lated 1979:15–17) analyzed a text of Réunionnais spoken by whites as
“totally French . . . foreign elements merely float on the surface [making
it] . . . only an apparent creole.” Reinecke (1937:526) noted that in
comparison to the creole of Mauritius, “the dialect of Reunion has not
departed quite so widely from the original French.”
Valkhoff (1964:724) suspected authors of Réunionnais texts “of using
an artificial and gallicized language which can be called ‘semi-creole’ ”
and decided that the language was a form of “Creole-influenced French
rather than French-influenced Creole” [“plutôt du français créole que du
créole français”] and that “there are two varieties of this Creole (without
counting many intermediary nuances), namely an urban speech form and a
popular speech form, and the former is more gallicized.” Vintilă-Radulescŭ
(1976:129) was the first to mention the the possible influence of the Indo-
Portuguese spoken by the wives of the earliest settlers in Réunion. She
also realized that “the mountainous relief of the islands . . . explains the
dialectal mosaic” (1967:126). Deltel (1969) proposed that there was a

continuum among the different lects of Réunionnais.
The most important work on Réunionnais to date is that of
Chaudenson (1974ff.), who was the first to distinguish among what are
now understood to be the three principal lects: (1) the Créole des Bas, spo-
ken by the coastal Réunionnais of African, Malagasy, and Indian origin;
(2) the Créole des Hauts, used by the highland whites; and (3) the urban
Creole, which is strongly gallicized. He also proposed that Mauritian and
Seychellois Creole were derived from Bourbonnais (an earlier form of
Réunionnais), which since decreolized due to the continuing presence of
French on Réunion. He argued that there had been no substratal influ-
ence on Réunionnais, which had simply evolved out of the français avancée
or colloquial French that had developed beyond the reach of those who
would have kept it more in line with the standard.
Bollée (1977:116) argued for Mauritian and Seychellois having re-
sulted from a “higher degree of reduction . . . than Réunionnais.” Papen
(1978) provided a comparative study of the grammar and social history
of all three varieties, concluding that Réunionnais represented a post-
creole continuum. Valdman (1978) was the first to draw parallels between
Réunionnais and the patois of St. Barts, a comparison later furthered by
Calvet and Chaudenson (1998). Hull (1979) pointed out the difference
in structure between Réunionnais and the Ile de France creoles, conclud-
ing that “Maur[itian] Cr[eole] evolved on Mauritius out of a nucleus of
Pidg[in] Fr[ench], with only secondary borrowings from Réu[nionnais]
Cr[eole]. Seych[ellois] Cr[eole] derives from early Maur[itian] Cr[eole],
not Réu[nionnais] Cr[eole].” He noted that “Where black influence was
subsequently removed, a somewhat decreolized form of Cr[eole] could
remain on the island, as on Réunion, or on St. Barts. . . . But on the whole
Cr[eole] and French remain psychologically distinct . . . [and] No ‘post-
Creole continuum’ has formed, as in English Cr[eole]-speaking areas”
(Hull 1979:211–213).
Baker and Corne (1982) also rejected Chaudenson’s Bourbonnais the-
ory, arguing that Réunionnais settlers were not present in sufficient num-
bers on Mauritius during the crucial period, while “West Africans formed
a majority of the slave population of Mauritius in the period 1730–35”
(1982:241), explaining the striking parallels between the French Creoles
of the Atlantic and Indian Oceans. In the same work, Corne pointed out
that the “verbal system of R[éunion] C[reole] is fundamentally ‘French’
in its make-up . . . [while its] Creole features . . . are rather marginal”
(1982:102).
In more recent work, Chaudenson (1992, 1995, 2000) has described
creolization as a restructuring process that is not so different from the
kind of restructuring found in normal language transmission. This

would certainly account for the partial restructuring of varieties such
as Réunionnais, from the perspective of the present study.
1.7 The comparison of partially restructured vernaculars

In 1991 I organized a seminar on partial restructuring at the City
University of New York (CUNY), followed by another in 1996. A number
of talented doctoral students participated, several of whom were them-
selves native speakers of partially restructured vernaculars or their source
languages. These seminars led to a number of publications, ranging from
conference papers to journal articles (e.g. Craig 1991 on American Indian
English) and dissertations (e.g. Mello 1997 on Brazilian Vernacular
Portuguese, and Green 1997 on nonstandard Dominican Spanish), one
of which is still in progress (Chapuis forthcoming on Réunionnais). The
goal of these seminars was to identify some of the problems that needed
to be solved in developing a workable theoretical model for this linguistic
process by tracing the genesis and synchronic morphosyntax of a number
of partially restructured languages, comparing the results, and looking for
the possible relationship between the social history of the speakers and
the linguistic outcome. Some of the initial results were described in Holm
(1992a) and are briefly outlined below along with some further develop-
ments that grew out of later work on these varieties.
The social factors that we considered potentially relevant to the lin-
guistic outcome included the following:
1. the precise origins of superstrate and substrate speakers;
2. the (changing) ratio of superstrate to substrate speakers; if the latter
came to outnumber the former, the length of time this took;
3. the degree of intimacy of early social relations between superstrate
and substrate speakers (i.e. the likelihood of pidginization as opposed
to normal second language acquistion);
4. the likelihood of either group’s contact with a pidgin or creole spoken
elsewhere;
5. demographic changes (e.g. immigration, emigration, wars, plagues)
and the effect on intergroup relations;
6. social, economic, and political changes and the effect on intergroup
relations;
7. the degree of rigidity of any racial caste system;
8. education: accessibility, actual language of instruction;
9. communications: degree of geographical isolation;
10. any changes in the variety’s status (e.g. new domains of use, stan-
dardization).
Regarding the linguistic make-up of each variety, the following factors

were considered:
1. the sources of lexicon: archaic, regional, or sociolectal usages in su-
perstrate; substrate languages; adstrate languages; pidgins or creoles
spoken elsewhere;
2. phonology: contrasts with superstrate; similarities to any varieties in
(1) above, i.e. in phonotactic rules or actual phonemes and their allo-
phones;
3. morphosyntax: contrasts with superstrate; similarities to any varieties
in (1) above, e.g. the loss or retention of inflections in the NP (e.g.
number/gender marking on articles, adjectives; possessive construc-
tions) and VP (bound vs. free tense/aspect morphemes; uses of tense
and aspect), as well as any other constructions not found in the lexical
source language (e.g. use of prepositions and conjunctions; word order
in main clauses; structure of dependent clauses).
4. the typological distance between the superstrate and substrate.
The study of each variety concluded with a summary of the scholarship
relating to its status as a creole or non-creole, and an assessment of its
status as a partially restructured language. The group’s ultimate task was
to compare the results of each study to determine whether the similarities
and differences among these varieties would justify their inclusion in a
group of partially restructured languages, and then to extrapolate the
defining sociolinguistic and structural characteristics of that group.
Since partial restructuring is a graded phenomenon, any specification
of what proportion of features a variety so designated must share with
creoles but not the lexifier language has to be intrinsically arbitrary. As a
common-sense guideline, it seemed unhelpful to designate any language
that has borrowed any creole feature as being partially restructured. For
example, standard American English has borrowed some lexical items
and even set phrases from various creoles such as go for broke, but this
hardly seems to be grounds for classifying it as being partially restruc-
tured. On the other hand, the number of features shared with creoles
but not with British varieties of English in AAE does seem significant: it
has not only lexical items but also frequently occurring phonological and
morphosyntactic features.
It is difficult to measure objectively the degree of restructuring that a
language variety may have undergone. Nonstandard Caribbean Spanish
appears to have undergone less restructuring than either Afrikaans (in
which one can say “Ons is bly,” literally ‘Us is happy’ in terms of Dutch)
or Brazilian Vernacular Portuguese (e.g. “Eu chamei ela,” corresponding
to ‘I called she’ in European Portuguese). Of course the literal transla-
tions into English do not provide an objective indication of the degree
of restructuring, since English speakers associate violations of stan-

dard pronominal case marking with basilectal Caribbean Creole English,
which is not the impact these structures have in modern Brazil and South
Africa. (Parkvall 2000 uses a more objective method of measuring degree
of restructuring, discussed in the introduction to chapter 2, and another
method is discussed in section 6.2). We considered ways of indicating the
degree of restructuring, e.g. by designating varieties as having undergone
“weak” or “very weak” restructuring if they have few features shared with
creoles, and as “strong” or “very strong” if they have many. The idea was
to distinguish between varieties like AAE and the nonstandard varieties
used by some Southern whites, containing fewer such features (albeit sig-
nificant ones such as copula deletion). Such distinctions would seem to be
helpful in describing the status of vernacular lects of Réunionnais French
as opposed to that of the nonstandard French of Louisiana’s Cajuns or
the patois of St. Barts.
The criterion of distinguishing partially restructured varieties from
post-creoles through the existence of basilects in the latter’s history
(Holm 1988–89:9–10) proved problematical, quite aside from the very
real difficulty of finding written evidence of extinct basilects. For exam-
ple, while the ancestors of the Afrikaners apparently never spoke a fully
creolized variety and their seventeenth-century Dutch simply underwent
creole influence, the African ancestors of the so-called Coloureds appar-
ently did speak a full creole (or at least a variety that underwent a very
strong degree of restructuring, to extrapolate from modern Orange River
Afrikaans). However, the modern lects of Afrikaans – both standard and
nonstandard – have speakers from both groups, and there is a basilect or
near-basilect in the history of these lects, although we can say that the
nonstandard lects seem to be descended from this (near) basilect more
directly than the standard lects.
We encountered other problems as well with our own hypotheses and
those of others, but recognized that these difficulties could not be ade-
quately dealt with until we found and organized the relevant linguistic
and sociolinguistic facts and then tested our theories against them. This
volume, which builds on the work of the students participating in those
seminars as well as my own research and that of others, is an attempt to
do that.
2 Social factors in partial restructuring
Introduction
The kind of partial restructuring of languages examined here is clearly
related to the more complete kind of restructuring called creolization,
and there is a widespread consensus that the defining characteristics of
creolization include social as well as linguistic factors (Holm 2000a:68–
71). In fact, Thomason and Kaufman (1988:35) demonstrated that “it
is the sociolinguistic history of the speakers, and not the structure of
their language, that is the primary determinant of the linguistic outcome
of language contact.” It will be argued here that partial restructuring
is also a sociolinguistic process, and to understand it we must consider
phenomena that are both social and linguistic.
One of the crucial social factors in creolization has to do with the
proportion of native versus non-native speakers of the lexical source lan-
guage present at the time the creole language emerged. Although Parkvall
(2000) has questioned the centrality of this demographic ratio in deter-
mining the degree to which creoles are restructured, his comparative work
actually reaffirms the validity of this correlation, which is the topic of the
following discussion.
Bickerton (1981:4) claimed that the maximum percentage of native
speakers of the superstrate language associated with full creolization was
20 percent. He was not the first to ponder this part of the equation; over
a century earlier, Van Name surmised that
Of the causes which have contributed to the formation of these dialects the chief
are: first, the mature age of the slaves . . . secondly, the fact that they constituted
the great body of the population, the whites being in a minority seldom as large as
one-fourth. (Van Name 1869–70:124, my emphasis)
While there is general agreement that non-native speakers need to con-

stitute a substantial majority of the population for full creolization to
occur, there is debate as to whether this ratio is the overriding fac-
tor in the process, and what the minimal majority might be. Rickford
24
Social factors in partial restructuring 25
(1999:235) points out that there are known instances where Africans
made up less than 80 percent of the population in places that produced a
creole:
By 1690, 35 years after the settlement of Jamaica, Blacks constituted 75 percent

of the population (Williams 1985:31) . . . In Haiti the Black population was only
34.8 percent 25 years after the colony’s founding, and in Martinique, only 51.5
percent. (Singler 1995)
The discrepancy between 75 percent and 80 percent in Jamaica seems

too small to worry about; that in Haiti seems much more considerable,
but this may have to do with the date simply being too early – raising the
issue of the relevance of the amount of time that was necessary for the
slave population to reach the required majority. Haiti’s population was
over 80 percent non-European by 1713 (Singler 1995:210). The figures
from Martinique seem like better counter evidence to Bickerton’s claim
because of the existence of a creole text from 1671, when the population
was 62 percent non-European (ibid., Carden et al. 1991).
To test the correlation of such demographic ratios to the degree of re-
structuring, Parkvall (2000) designed a restructuring index to quantify
the typological distance between some Atlantic creoles and vernaculars,
on the one hand, and their lexifiers on the other. The linguistic features
he used to measure restructuring included 36 grammatical features often
found in Atlantic creoles but not their lexifiers (e.g. no gender distinction
in the third person singular pronoun) and 9 phonological ones (e.g. the
presence of tonal distinctions). By quantifying the presence or absence of
these features, he was able to quantify the degree of restructuring of each
language, resulting in scores (see table 1 below) varying from 90 for the
completely “radical” creole to −45 for the ideal lexifier. He then com-
pared these to social factors: the first documented slave imports (event 0);
the year the non-white population exceeded 20% (20%); the interval
in years; the year the non-white population exceeded 50% (50%); the
interval between this and event 0; the year the non-white population ex-
ceeded 80% – purportedly triggering the bioprogram (Bickerton 1981) –
(80%); and the interval between this and event 0.
Table 1 was originally published with a number of qualifying footnotes;
here we need note only that the demographic data for St. Thomas∗ ac-
tually refer to St. Barthélemy, the source of the former’s French-based
patois. Since this table raises a good number of questions, the reader is
referred to the original article, which is quite transparent considering the
constraints of space: the results of the linguistic comparison are given for
each feature in each variety, and the obvious problems of comparability,
reliability, etc. are openly discussed.
Table 1. Degree of restructuring and social factors (Parkvall

2000:205–206)
Score Variety Event 0 20% Interval 50% Interval 80% Interval
68 Sranan 1651 1652 1 1662 11 1672 21

63 São Tomé 1485 1493? 8 1499? 14 1546? 61
51 St. Kitts 1626 1650? 24 1680 54 1729 103
48 Jamaica 1655 1657 2 1670 15 1690? 35
48 Haiti 1634 1635 1 1698 64 1710 76
42 Gullah 1670 1672 2 1708 38 never n/a
41 Negerholl. 1665 1668 3 1675 10 1710 45
41 Fr. Guiana 1664 1665 1 1668 4 1697 33
41 Guadeloupe 1640 1652 12 1667 27 1725 85
39 Louisiana 1708 1711 3 1726 18 never n/a
38 Martinique 1635 1636 1 1655 20 1733 98
35 Mauritius 1721 1722 1 1730 9 c. 1750 29
35 Papiamentu 1634 1652 18 1657 23 1685 51
23 Cape Verde 1462 1515? 53 1525? 63 1540 78
20 Barbados 1627 1643 16 1657 30 1705 78
16 Virginia 1619 1712 93 never n/a never n/a
−5 Dom. Rep. 1502 ? ? ? n/a 1520s c. 20
−5 Brazil 1549 ? ? ? n/a never n/a
−23 Bermuda 1612 1618 6 never n/a never n/a
−24 St. Thomas∗ 1660 1671 11 1787 127 never n/a
n/a Fr. St. Kitts 1626 1626 0 1668 42 1710 84
30 Average 13 33 60
Parkvall concludes that “. . . there is indeed a correlation – whether

causal or not – between the degree of restructuring and the demographic
ratios in the formative period” (2000:197). Assuming the soundness of
his method and execution of this inquiry, however, it is also the case that
demographic ratios are not everything, as he himself points out.
To deal first with two apparent exceptions: Table 1 indicates that two
areas which developed full creoles never had a non-white population ex-
ceeding 80% (Gullah, scoring 42; Louisiana, scoring 39). “Gullah” ap-
pears to refer to the Gullah country – certain coastal counties of South
Carolina and Georgia – rather than to either of these states as a whole.
But, if that is the case, we must note that by the nineteenth century blacks
made up some 80% of the the population of the tidewater area of South
Carolina – and over 95% in its rural districts (Reinecke 1937:491). As for
Louisiana, a footnote clarifies that “the non-white population reached an
all-high low [sic: ‘all-time high’?] of 77,9% in 1741” (Parkvall 2000:5).
In other words, both areas seem to have had a slave population very close
to 80%.
Having dealt with the apparent exceptions to the 80% rule, we now
need to consider the relevance, if any, of the amount of time that it took for
the slave population to reach 80% of the total. This may account for the
discrepancy noted above, namely the observation of Rickford (1999:235)
that in Haiti the slave population was only 34.8% twenty-five years after
the colony’s founding, and in Martinique, only 51.5%. According to
Table 1, it took 76 years for Haiti’s slave population to reach 80%, and
98 years for Martinique’s. But if we look at the number of years in the
final column of Table 1, there seems to be no pattern: the most restruc-
tured varieties (Sranan, São Tomé, St. Kitts) took 21, 61, and 103 years
respectively to reach 80%, while the least restructured creoles (Parkvall
draws the bottom line at a score of 20) are Papiamentu, Cape Verde,
and Barbados (to follow his nomenclature), taking 51, 78, and 78 years
respectively. Thus, if there were a correlation, these results seem counter-
intuitive: the places developing the most restructured creoles took both
the shortest time (Sranan: 21 years) and the longest time (St. Kitts: 103
years) to reach 80%.
A concern with the amount of time that was required for the slave
population to reach a majority and the possible correlation this might
have with the degree of restructuring can be seen in Bickerton’s (1984)
discussion of the amount of time between the beginning of immigration
of non-native speakers and what Baker (1982:852) had called
“Event 1” – the point at which slave and master populations achieved numerical
parity. . . . The longer this period, the greater the exposure of early arrivals to the
dominant language and hence the richer the second-language version that would
be transmitted to the first influx. (Bickerton 1984:177)
This led to Bickerton’s pidginization index (PI):

P
“Y × = PI
R
Where Y represents the number of years between colonization and Event
1, P the total substratum-speaking population at Event 1, and R the
yearly average of post-Event 1 immigrants. . . . A higher PI indicates a
‘richer’ pidgin, one that retains more features of the dominant language”
(1984:178). However, this mathematical formula to indicate the degree
of restructuring proved “unworkable” (Singler 1990:645).
Later Mufwene (1996) dealt with a similar idea in his article on the
founder principle:
In some colonies, such as South Carolina, Virginia and Réunion, the Africans
remained minorities for the first 30–50 years, whereas in some others such
as Suriname, Mauritius (Baker, to appear [= 1996]) and apparently Guyana
(Winford, p.c. 1994), the plantation phase came about rather rapidly bringing
along an early slave majority. (Mufwene 1996:98)
While the first three colonies gave rise to less restructured vernaculars,
the latter three have fully creolized languages. However, the length of
time it took for the slaves to become a majority may not be the relevant
social factor here. All three of the latter colonies could have imported
restructured languages along with their slave force. Many, if not most,
of the slaves brought to Suriname between 1651 and 1667 came from
Barbados, and many, if not most, brought to Guyana after 1760 came
from the British West Indies and West Africa. Even in Mauritius, where
the slave population grew from 30 to 1,000 between 1727 and 1730,
some 600 came from French entrepots in West Africa and might well
have spoken some pidginized French.
Thus the relationship between the degree of restructuring and social
factors like demographic ratios is not always a straightforward one, a
factor Parkvall recognizes:
One possible reason why there appears to be a lack of correspondence between
the demographic and the linguistic data is that the creoles under investigation did
not arise in the areas where they are presently spoken. (2000:193)
This complexity of the interrelationship of social factors has long been

recognized in the acknowledgment of the effect of factors such as maroon-
age or the early withdrawal of the superstrate on a creole’s proximity to
the “bioprogram” (Bickerton 1984:177–178) or its proximity to a proto-
type (McWhorter 2000).
As to the existence of a straightforward relationship between the
amount of time that it takes for a slave population to become a sub-
stantial majority and the degree of restructuring in the ensuing language,
this would appear to be a question that has not yet been answered clearly,
although Parkvall (2000) has cast some light on the matter. The most
we can say with certainty is that it appears to have taken an average of
sixty years for the non-native-speaking population to reach its greatest
proportion of the total, at least in the history of the language varieties he
examined – and possibly the same amount of time for these varieties to
reach their present degree of restructuring. While Bickerton (1984) was
certainly right to assert that time is a relevant factor, its relevance de-
pends on still other variable, complicating factors. For example, the slave
population of Réunion reached 80 percent within 100 years, but this re-
sulted in partial rather than full restructuring because of the particular
sociolinguistic circumstances of the colony’s early history (see section 2.5

below).
It will be argued below that, like creolization, partial restructuring is
a sociolinguistic process, that a relationship does indeed exist between
sociolinguistic factors in the development of these languages and their
synchronic typology, and that this relationship is parallel to the one be-
tween these factors in fully creolized languages.
In order to examine this hypothesis, we will first look at the sociolin-
guistic setting of the formation of each language variety, focusing on social
factors related to its genesis and development – the topic of the following
sections of this chapter (the disparate lengths of which reflect in part the
disparate amounts of published research to draw on). Chapters 3 to 5 will
then compare corresponding areas of their morphosyntax, and the final
chapter will attempt to relate the social and linguistic facts in a theoretical
model that will account for what we know about these languages.
2.1 AAE: the sociolinguistic setting of its development
2.1.1 AAE: the seventeenth century

The proportion of Africans in most of Britain’s North American colonies
was so much lower than the ratio of blacks to whites in the Caribbean
that it seems doubtful that full creolization ever took place outside of a
few demographically exceptional areas such as coastal South Carolina
and Georgia, where Gullah developed. While settlers in both the British
Caribbean and North America were largely European-born indentured
servants in the early seventeenth century, by 1680 the black population
equalled or surpassed the white population in many Caribbean colonies,
whereas in North America blacks made up less than 5 percent of the
total population (Rawley 1981:329). By the 1730s blacks still made up
only 14 percent of the North American population, but 80 percent of the
population of the British and French Caribbean (ibid.).
The sociolinguistic history of African slaves in the Americas was largely
determined by the economic forces shaping the societies to which they
were brought, which in turn depended on particular crops. The tobacco
plantations in Virginia and Barbados to which the first slaves were brought
were relatively small farms worked mostly by indentured servants from
Britain. The first Africans, arriving in Virginia in 1619 and in Barbados
in 1627, were outnumbered by the British laborers among whom they
worked and from whom they most likely learned English largely through
normal adult second language acquisition; their contact was much closer
at this early period than in the plantation societies that developed later.
The social conditions necessary for creolization – lack of access to native-

speaker models of the target language – did not occur on Barbados until
the cultivation of sugar was introduced in the middle of the seventeenth
century. As the supply of tobacco in Europe rose, its price fell and farmers
began looking for other products to grow in the West Indies. The pro-
duction of sugar was more profitable, but this required large plantations
and a large workforce. Great quantities of sugar cane had to be produced
to offset the initial investment of capital in the mill and other machinery,
as well as land and labor. The work was much more grueling than raising
tobacco, and indentured servants were unwilling to do it. Their labor was
gradually replaced by that of the African slaves: while there were only a few
hundred of the latter in Barbados in 1640, by 1645 there were 6,000 –
and some 40,000 whites. By 1685 the number of blacks had risen to
46,000 and the number of whites had actually dropped to 20,000.
The effect of the sugar revolution on Barbados was to force the small-
holders (mostly former indentured servants) off the land as large sugar
plantations took over the island. The displaced peasantry chose emigra-
tion over starvation, and populated new English colonies in Suriname,
Jamaica, the Carolinas, and the Leeward Islands. Thus in the second
half of the seventeenth century Barbados played a central role in the
dispersal of English in the New World (Holm 1994). The question is
to what degree that English reflected contact with the languages spo-
ken by Africans, especially the speech of slaves exported from Barbados
and other Caribbean islands such as St. Kitts. Winford (2000:216, 242)
considers modern Barbadian an “intermediate” variety, “. . . the result
of targeted shift . . . quite comparable with other outcomes of commu-
nal shift such as Hiberno English . . .” Yet Hickey (forthcoming) finds
that Winford has misinterpreted the facts about Irish English, such as
“its preference for invariant does” (Winford 2000:232). Hickey (forth-
coming) says “There is no evidence of this in Irish English. Perhaps the
claim derives from a misunderstanding . . .” At any rate it seems clear
that more basilectal varieties were used in Barbados in the seventeenth
century; one such lect was documented in the testimony of a Barbadian
slave at the 1692 Salem, Massachusetts, witch trial, e.g. “He tell me he
God” (Dillard 1972:79).
The cradle of the development of African American English was
Virginia, the first and most populous colony in the southern part of British
North America. In Virginia the status of Africans was comparable to that
of indentured servants during the first part of the seventeenth century;
most worked as domestics or on small farms and seem likely to have had
sufficient access to local English. Mufwene (2000b:240) reasons that this
would have led to their normal acquisition of the British settlers’ English
Lake
Winnipeg
Seattle
bia
m
lu
Co
ce
en
wr
Portland M
La
iss
o
St
Super ior
ur
ke
i
La
Lake
Lake Ontario
Sn Huron Boston
a ke
Minneapolis Lake
Erie
Detroit New York
Lake Michigan
Great Salt Lake
Chicago Cleveland Philadelphia
Salt Lake
P A C I F I C
San Francisco City
Denver Cincinnati Washington

o
io
ad Kansas
lo r Norfolk
Oh
O C
Co Ar City
k St Louis
an
sas
Los Angeles
E A
Chattanooga
A T L A N T I C
Oklahoma Memphis
N
Albuquerque
i
Little
Rock
Red Atlanta O C E A N
Charleston
COLOURED
Mississipp
POPULATION Dallas Montgomery
Over 50%
Ri
Jacksonville
oG
ran
30–50%
Houston
de
New Orleans
San
10–30% Antonio
Under 10%
0 250 500 750 km
Areas without coloured population or unsettled 0 250 500 miles
Map 1 United States: ‘Coloured Population, 1900. Originally published in Darby and Fuller (1978) The New Cambridge Modern
History Atlas (Cambridge University Press).
Table 2. Estimated population of (East) Virginia, 1685–1790
1685 1700 1715 1730 1745 1760 1790
Total 43,600 63,500 96,300 153,900 234,200 327,600 747,800

Red 7% 3% 1% 0.6% 0.1% 0.1% 0.03%
White 87% 88% 77% 67% 63% 60% 59%
Black 6% 9% 22% 32% 36% 40% 41%
(Based on Wood 1989:38, cited by Winford 1997.)
as a second language. Africans in Virginia did not become slaves for

life until the latter part of the century (2000b:238), when their num-
bers became more considerable. Although slaves then made up a larger
proportion of the local population, Mufwene (2000b:240) believes that
these later arrivals tended to adapt to the local language norms rather
than imposing their own norms due to the founder principle (Mufwene
1996). However, it was established above that the founder principle can
be overridden by the importation of a variety that was restructured else-
where, and Rickford (1999:241–242) objects to the scenario proposed
by Mufwene, since during this period most slaves were imported from
England’s Caribbean colonies and were already likely to know English in
a pidginized or creolized form. It seems doubtful that they would have
abandoned such linguistic habits on arriving in North America since these
would clearly have been useful to them and their local interlocutors, not
only in facilitating communication but also in establishing sociolinguistic
boundaries.
2.1.2 AAE: the eighteenth century

As tobacco farming spread during the eighteenth century, the importation
of slaves – increasingly from Africa rather than from the Caribbean – grew
as indentured labor became more expensive. The proportion of blacks
in Virginia increased from 22 percent in 1715 to 40 percent in 1760.
Thus later newcomers were less likely to have as much access to native-
speaker models of the target language as did those who arrived earlier,
but whites still made up a substantial majority of the population: In fact,
whites made up some 60 percent of the population of all the southern
colonies in 1750, although it must be borne in mind that this was a large
and demographically heterogeneous area.
Rickford (1999:244) points out how these statistics are misleading for
Georgia, which in 1750 had been founded only recently: by 1776 blacks
Table 3. Estimated proportion of blacks in American

colonies in 1750
Colony/region Total blacks % of blacks % in 13 colonies
New England 10,982 3.1% 4.6%

Middle Colonies 20,736 7.0% 8.8%
Southern Colonies 204,702 39.8% 86.6%
Maryland 30.8%
Virginia 43.9%
N. Carolina 27.1%
S. Carolina 60.9%
Georgia 19.2%
(Based on Rickford 1999:238.)
made up 48 percent of Georgia’s population and were concentrated on

the coast, where Gullah was spreading.
In other words, the most likely scenario after 1715 in most parts of the
American South is that English was not fully creolized but did undergo
some degree of restructuring as it was acquired by speakers of African lan-
guages and of more fully restructured varieties of English brought in from
the Caribbean, and was increasingly learned from other African Ameri-
cans speaking an emerging sociolinguistically distinct lect of local English
that became the forerunner of modern AAE. While Mufwene is surely
right in surmising that “children, regardless of whether they were locally
born or imported . . . slowed down the divergence of Africans’ English
vernaculars from those of Europeans, because they were more likely to
acquire colonial English with minimal restructuring” (2000b:240), the
important issue here is the very existence of the divergence, for which the
only credible explanation is restructuring. The roots of that divergence
are likely to go back to at least the early eighteenth century.
The recognition of a distinctively black way of speaking seems implicit
in a passage in a journal of 1705, in which one white criticizes another for
speaking “Negro” to an Indian (quoted by Dillard 1972:80). However,
there is a problem in what the designation of “Negro” might mean. Since
colonial times African Americans have spoken many varieties of English,
ranging from the foreigner’s English of Africans; the creole or partially re-
structured English of slaves imported from the West Indies; the pidginized
English brought from West Africa; the indigenous creoles of North
America (such as Gullah and its offshoot, Afro-Seminole); the partially
restructured varieties that had emerged by the eighteenth century; un-
restructured regional English (which may have borrowed some interlan-
guage or substrate features from contact with the previous varieties); and
the standard American English first spoken by educated freedmen and

now spoken by millions of African Americans – usually with AAE phono-
logical features to signal a positive awareness of ethnic identity. This is not
conjecture: Kautzsch and Schneider (2000) document idiolects in South
Carolina that they identify as creole, semi-creolized, and non-creole. The
accuracy of their identification is borne out by the linguistic features in
these idiolects. As Schneider (1989) suggests, pockets of greater restruc-
turing developed in localities with demographics more similar to those
in South Carolina’s low country that led to the emergence of Gullah,
which underwent relatively complete creolization, making AAE the
product of a few presumably independent creolization processes in localities with
an exceptionally dense black population . . . with a certain degree of leveling, mix-
ture and perhaps loss or even spreading of such forms in the post-emancipation
period. (Schneider 1989:278)
2.1.3 The emergence of Gullah

As in other colonies, the sociolinguistic situation of Africans in coastal
South Carolina was determined by the crop they were brought there to
raise. In this case it was the cultivation of rice that created a setting for
the creolization of English – relatively complete if compared to the partial
restructuring of AAE, but still far from the fundamental restructuring of
English in Suriname.
In 1670 Charles II granted the Lords Proprietors of Carolina on the
mainland a patent that included the Bahama Islands, creating a single
colony that endured as such for the next half century. In the same year
Charleston, in what is now South Carolina, began to be settled by some
800 British settlers and 300 slaves. According to Wood (1974:24–25),
almost half the whites and more than half the blacks came from Barba-
dos. Other settlers included French Huguenots and religious dissenters
from the British Isles and from other colonies in British North America
(Baptists, Quakers, and Catholics), as well as settlers from Bermuda, the
Bahamas, the Leeward Islands, and Jamaica (ibid; Joyner 1984:13). Dur-
ing the first years of settlement the colonists traded with the Indians for
deerskins and also raised livestock, sending meat to Barbados in exchange
for slaves and sugar. By the 1690s it became clear that the coastal low
country was suitable for raising rice, for which there was a growing mar-
ket in southern Europe and the Caribbean. Although the British settlers
had no experience in cultivating rice, their slaves did since it had long
been grown in various parts of West Africa. The intensive labor needed
for its cultivation led to a great increase in the importation of slaves: by
1708 the colony’s population of 8,000 was equally divided between whites
Table 4. Estimated population of South Carolina, 1685–1775
1685 1700 1715 1730 1745 1760 1775
Total 11,900 14,100 19,200 33,400 62,400 97,500 179,400

Red 84% 53% 27% 5% 2% 1% 0.3%
White 12% 27% 29% 29% 33% 40% 40%
Black 4% 20% 45% 65% 65% 60% 60%
(Based on Wood 1989:38, cited by Winford 1997.)
and blacks, but by 1740 there were 40,000 slaves as opposed to only
20,000 whites (Wood 1974:143–152). The proportion of inhabitants of
American Indian, European, and African ancestry underwent the shifts
documented in table 4.
These figures suggest that during the first generation of colonization,
blacks made up a small minority who either learned English as a sec-
ond language through relatively close contact with native speakers under
frontier conditions, or else arrived from Barbados or other parts of the
British West Indies with some knowledge of Creole English. In the next
generations, however, the black population (coming increasingly straight
from Africa with little or no knowledge of English or Creole) became
the majority, leaving white native speakers a minority of about one-third
to two-fifths of the population. It must be remembered that the above
figures are for all of South Carolina not just the coastal low country,
where the concentration of blacks was much higher. Segregation was in-
stituted in South Carolina in 1720 after the colony was separated from
the Bahamas and North Carolina, and a 1726 survey of two low-country
parishes indicated that two-thirds of the slaves lived on plantations, de-
fined as landholdings with 25–100 hands (Wood 1974:160). As noted
previously, by the nineteenth century blacks made up some 80 percent
of the the population of the tidewater area of South Carolina – and over
95 percent in its rural districts (Reinecke 1937:491). In coastal Georgia
(settled after 1733) the proportion was not quite so great, but Gullah
came to predominate there as well.
The social and linguistic relations in Gullah country differed consider-
ably from those in Virginia and elsewhere in the American South. Quite
apart from the continuing existence of creolized English there today,
the above historical, social and demographic data support the develop-
ment of a creole continuum in this area during the eighteenth century –
one that developed in a society with a relatively high percentage of native
speakers of English, keeping the speech of most creole speakers quite
mesolectal in comparison to Caribbean varieties.
2.1.4 AAE: the early nineteenth century

In 1790 the United States’ 757,000 blacks (only 60,000 of whom were
free) made up over 19 percent of the country’s population. Large-scale
slavery was restricted mainly to the tobacco-growing area of the Upper
South (Virginia, Maryland, and North Carolina) and central Kentucky
and Tennessee, as well as the coastal areas of South Carolina and Georgia,
where rice and indigo could be grown (Bailey 2001:58). However, in
1793 the invention of the cotton gin made short-staple cotton a viable
cash crop, leading to the spread of new cotton plantations westward into
the interior South. During the first half of the nineteenth century, up to a
million slaves were moved to Alabama, Mississippi, and newly acquired
areas as far west as Louisiana and Texas (2001:59). By 1830 parts of the
interior South, such as the Mississippi Delta, the Alabama Black Belt, and
the lower Brazos Valley in Texas, had ratios of blacks as high as those in
Gullah country (2001:60). Mufwene (2000b:247) claims that the cotton
plantations were on average smaller than the tobacco plantations in the
Upper South and rice plantations in coastal South Carolina and Georgia,
but Bailey (2001:60) claims they were often just as large. While many of
the slaves brought to these states came from the Upper South and spoke
early AAE, many also came from coastal South Carolina and Georgia
and spoke Gullah. Still others were brought directly from Africa and, to
a lesser extent, the Caribbean until the importation of slaves was made
illegal in 1808 – although it continued until the Civil War (1861–65).
The latter slaves, speaking restructured or non-native varieties of
English learned from the first two groups, seem likely to have contributed
to a heterogeneous mixture of AAE varieties, so it is not surprising that
later creolists such as Stewart (1968) and Dillard (1972) had the impres-
sion that there existed a widespread “Plantation Creole” continuum at
this time: evidence of creole features was indeed widespread outside the
original Gullah area. Stewart (1968:52) claimed that decreolization be-
gan well before the Civil War; however, Mufwene (2000b:247) concludes
that this was the period when AAE and Gullah were stabilizing, referring
to similarities in how they were represented in fiction then and how they
are spoken today. Pointing to the end of the illegal importation of slaves,
he surmises that
The increasing number of native speakers and the attrition of non-native speakers
enabled the vernaculars that had been developing till then among descendants of
Africans to stabilize, or at least not to undergo any form of restructuring motivated
by language contact. (Mufwene 2000b:247)
Bailey, however, interprets the same social and linguistic evidence as sup-
porting restructuring:
Given the fact that a substantial portion of slaves coming into most of these areas
[of the interior South] spoke some form of English, we would hardly expect the
development of a full-scale creole language, but we might well expect a variety of
English showing the clear imprint of the creole and African languages that formed
part of the matrix from which it derived. According to Bailey/Thomas (1998a),
that is precisely what we find in the early varieties of AAVE in the interior South.
(Bailey 2001:60)
2.1.5 AAE: diaspora varieties and the ex-slave narratives

One of the basic problems in reconstructing the history of the devel-
opment of AAE has been the fact that it was rarely written, and what
written fragments that do exist in fiction or travelers’ accounts are of
limited reliability, most having been written by non-native speakers of
AAE. Moreover, it is not possible to identify exactly which of the many
varieties spoken by African Americans is being portrayed (native, non-
native, creole, etc.) unless something is known about the origin and back-
ground of the individual speaker (Holm 1988–89:498). One attempt to
avoid this dilemma has been to examine the modern speech of the descen-
dants of groups of African Americans who left the United States at earlier
points in its history, although the information this provides is usually in-
direct and sometimes misleading. Such groups include those who went
to the Bahamas in the 1780s (Holm 1983); Sierra Leone in the 1790s
(Hancock 1971; Huber 2001); Nova Scotia in the nineteenth century
(Poplack 2000); the Samaná peninsula of the Dominican Republic in the
1820s (ibid.); and Liberia, also in the early nineteenth century (Singler
1989). However, there are pitfalls in this method, as I have discovered:
for example, Holm (1988–89:501) concluded that
Decreolization [of AAE in the United States] would appear to have begun during
this same time span [1780–1820] since the descendants of the emigrants to the
Bahamas speak creolized English while those in Samaná do not.
However, a closer examination of the historical records revealed that

the creole nature of the folk speech on these southern Bahamian islands should
not be interpreted as evidence that AAVE had been fully creolized on the mainland
before 1780 (and later decreolized) since the language that was brought there was
in all likelihood eighteenth-century Gullah rather than eighteenth-century AAVE.
This new interpretation undermines the most compelling evidence we knew of to
support the hypothesis that AAVE was ever fully creolized, and it provides further
indication that AAVE was – from its beginning – the product of semi-creolization.
(Hackert and Holm forthcoming)
Another source of information about earlier forms of AAE is the body

of transcribed interviews with elderly former slaves conducted during
the 1930s and later, known as the Works Project Administration Ex-Slave
Narratives (e.g. Schneider 1989). The problem of disentangling the lin-
guistic system of the transcriber from that of the speaker has been partly
resolved with the Ex-Slave Recordings, mechanically recorded interviews
with African Americans born in the American South between 1844 and
1861 (Bailey, Maynor, and Cukor-Avila 1991). Although there is some
controversy over how much speakers’ grammars might change over their
lifetimes, these recordings have been of great value in casting light on the
structure of AAE since the middle of the nineteenth century. A recent
comparative quantitative study of these sources (Kautzsch 2002:257)
concludes that properly chosen interviews “can be regarded as very close
approximations to speech” – much more so than letters, since literacy is
so closely tied to the learning of standard English.
2.1.6 AAE: the late nineteenth century

It is not necessary for AAE ever to have been fully creolized for it to
have undergone decreolization. However, if it was never a full creole,
“decreolization” may be a misleading term. Mufwene (2000b) prefers
“debasilectalization,” and Holm (2000:32) refers to this reapproachment
with non-restructured varieties as “secondary leveling.” It has been as-
sumed that African Americans had more freedom of movement and better
access to education after emancipation, and that this accelerated the reap-
proachment of AAE to the non-restructured lects with which they were in
contact. Evidence for a straightforward movement in this direction over
time, however, is mixed. Poplack et al. (2000:100) admit that the rate of
plural -s marking is lower in early AAE than in contemporary AAE (but
not that this is due to decreolization; see section 4.4.1), while Kautzsch
(2002:260) finds no evidence of decreolization in any of his quantitative
studies of earlier AAE.
Bailey (2001:64ff.) points out that the area where cotton was grown
continued to expand as rapidly after the Civil War as before it, with share-
cropping or farm tenancy replacing slavery as the mechanism for meeting
the intensive labor demands of cotton. The main difference was that after
the war, poor whites were also ensnared in this system and kept there
by the lack of alternative sources of credit. Bailey makes his point with
two maps of a Georgia plantation: in 1860 laborers are confined to the
slave quarters, but in 1881 the shacks of both black and white tenant
farmers are scattered over the plantation. He concludes that
The spread of tenancy to Whites and the spatial reorganization of the plantation
created new contexts for Black–White speech relationships throughout the South,
contexts which allowed for interaction among African Americans and Whites that
was probably more widespread than before the Civil War and took place among
people who were closer to being socioeconomic equals. . . . Hence we might expect
their vernaculars to to have much in common and . . . they do share many features
that emerged during the late 19th and early 20th centuries. (Bailey 2001:66)
However, the late nineteenth century also saw the rise of white sup-
remacists who enacted Jim Crow laws throughout the South to segregate
and disfavor African Americans socially, economically, and educationally.
This situation triggered emigration.
2.1.7 AAE: the twentieth century

Almost a million African Americans left the South during the great migra-
tion, which began even before the opening of new economic opportunities
in the North that came with the First World War and lasted until well after
the Second World War. Whatley (1981:93) points out that the increasing
migration of blacks to the North as well as the West Coast is reflected
in the decreasing proportion left in the South: 85 percent of the national
total in 1920, 77 percent in 1940; and 60 percent in 1960. By 1970, over
a third of all African Americans lived in just seven major urban areas:
New York, Philadelphia, Baltimore, Washington D.C., Detroit, Chicago,
and Los Angeles (Bailey 2001:66). These areas became more and more
segregated: the African American population of Detroit, for example, in-
creased from 37 percent in the 1960s to 83 percent in 2000 (Wolfram
forthcoming). The restricted social environment of blacks in the ghet-
tos of Northern cities fostered the continuation of features brought from
the South, but this speech took on a distinctively urban character. Labov
(1972b:xiii) described AAE as having a “relatively uniform grammar”
from one region to another, as one might expect from a variety that had
recently migrated. The question is whether AAE and other American
varieties of English have been growing apart. Mufwene speculates that
Perhaps the late nineteenth century was also the time when African and Euro-
pean Americans were particularly eager to identify some linguistic peculiarities
as ethnic markers and thus made divergent selections of features from the pool
of variants that they had shared until then. Such selections would have enhanced
ethnic-dialectal differences that may have been minor at the dawn of the twentieth
century. (Mufwene 2000b:251)
Poplack (2000:1) interprets this as endorsing her own position that
the many grammatical distinctions between contemporary varieties of AAVE and

American and British English are relatively recent developments, possibly initi-
ated during the post-Civil War period, as suggested by Mufwene [2000b] . . . in
a social context highly propitious to racial segregation and divergence.
Rickford (1999:xix) notes that “Accelerated AAVE use – especially of

salient features like be – and scathing criticism of those who ‘talk white’
are part of a symbolic statement by today’s young people of awareness of
and pride in African American identity.”
Kautzsch (2002) interprets his quantitative study of earlier AAE nega-
tion patterns as evidence that
In the first half of the nineteenth century AAE has a largely nonstandard English
basis. During the second half of the nineteenth century it gradually – though
slowly and partially – moves closer to standard English. Finally, a renewed de-
viation away from standard English seems to have started in the early twenti-
eth century, which is supported in some cases by present-day data (cf. ain’t for
didn’t; negative concord). This return to nonstandard patterns might result from
increased segregation of African Americans from the early twentieth century on-
wards, which at the same time favors the emergence of a black linguistic identity.
(2002:88)
Concerning the relationship between Southern varieties of AAE and
Southern White Vernacular English (SWVE), Cuckor-Avila (2001:116–
119) found that there were linguistic correlates to the reduced contact
between speakers of the two varieties after the Second World War due to
the decline of share-cropping, northern migration of blacks, and other
factors. Some of the earlier shared features have survived only in AAE
(e.g. copula absence after you, lack of third person singular present -s
marking of verbs), and others have evolved only in AAE (e.g. be + verb +
-ing). However, AAE has retained still other features that had not been
shared with SWVE (zero third personal singular copula, remote time been,
ain’t for didn’t). It is the last group which makes the position of Poplack
(2000) untenable.
However, it is still possible that historically (or at least currently) the
grammars of AAE and other varieties are diverging rather than con-
verging, as was assumed by those who saw AAE mainly as the result
of the decreolization of a full-fledged creole (e.g. Stewart 1968). In 1986
Labov and Harris argued that increasing racial segregation had produced
a “BEV [Black English Vernacular] that is more remote from other di-
alects than . . . reported before” (1986:4). The work of Bailey and Maynor
(1987, 1989) (cf. Bailey [2001] above) in Texas appeared to provide inde-
pendent confirmation of this hypothesis, suggesting that it might be a gen-
eral pattern in the United States. A symposium, published as Fasold et al.
(1987), raised the points “that Labov and his colleagues had failed to pro-
vide comparison points in either real or apparent time, and that Bailey
and Maynor needed an intermediate age group to minimize the the pos-
sibility of age-grading” (Rickford 1999:262). However, Rickford’s own
recent work in Palo Alto, California (1999:274) confirms this growing
divergence, particularly in the AAE use of invariant be and copula ab-

sence (see section 3.1.1). He notes that the first feature is “now very
salient as a distinctively black form . . . be is an invariant lexical item,
which can be consciously adopted and rapidly spread like slang terms and
other lexical items” (1999:276). Finally, he notes that “Black teenagers
are less assimilationist than their parents and especially their grandpar-
ents, and more assertive about their rights to talk and act their ‘natural
way’ ” (1999:274).
Whatever the current state of the relationship between AAE and other
American varieties of English, this brief historical survey demonstrates
that, given what we now know about what is required for full creoliza-
tion to take place, that is not what happened to AAE in most parts of
the American South. On the other hand, it has become clear that the
features which modern AAE shares with Caribbean varieties of Creole
English but not with Southern White Vernacular English cannot all be
explained through the kind of borrowing associated with language con-
tact, i.e. between a less-restructured AAE and more-restructured varieties
of English brought in from Africa, from the Caribbean, or from Gullah
country. While contact between creolized and non-creolized varieties of
English can explain the borrowing of lexical items (including ones with
syntactic implications, such as the complementizer say, discussed in sec-
tion 5.1.2) and even the borrowing of morphosyntactic features, it cannot
explain systemic differences like the patterning of zero copulas (section
3.1.4).
The only remaining explanation for the origin of such features is that
AAE is the result of English having undergone a less-extensive degree of
restructuring than that which led to Gullah or the English-based creoles of
the Caribbean and Suriname. Such limited restructuring would still have
allowed for the features described in chapters 3–5, i.e. both the reduction
of inflectional morphology and the retention of certain substrate features
such as a correlation between the form of the copula and its following
syntactic environment.
2.2 Afrikaans: the sociolinguistic setting of its development
2.2.1 Afrikaans: the seventeenth century

In 1652 the Dutch East India Company established the first European
settlement on Africa’s Cape of Good Hope in order to supply their ships
with provisions for the long voyage from Holland to the Dutch East
Indies. For the first six years the principal languages in contact were
Dutch (mainly the dialects of south Holland) and Khoi, spoken by the
original inhabitants of the Cape with whom the Dutch bartered for cattle.
By 1656 it was reported that the Khoi “were beginning to speak Dutch
fairly well” (cited by Scholtz 1970:85); they soon began working for the
settlers as nursemaids and herdsmen. In 1658 the colony received its
first shipload of slaves from West Africa and Angola; at this point “there
were 166 whites against 194 slaves and freedmen, together with some
Hottentot hangers-on” (Reinecke 1937:597). Thus the slaves found two
kinds of Dutch in use: that of the settlers (among themselves); and that of
the Khoi (used with the settlers). The latter probably played an important
role in shaping the pidgin, which did not have SVO word order as did
the Niger-Congo languages of the slaves, but rather SOV, the underlying
word order of both Dutch and Khoi (den Besten 1989:220).
In 1685 a colonial official reported that the settlers’ children were pick-
ing up the broken Dutch of the non-whites (Valkhoff 1972:40). In 1677
190 slaves were brought to the Cape from the Asian colonies that the
Dutch had taken from the Portuguese in India, Ceylon, and the East
Indies (Reinecke 1937:568); they seem likely to have made up a substan-
tial portion of the colony’s 350 slaves in 1691 (1937:597). As in Batavia,
the new Dutch capital of the East Indies, their lingua franca was creole
Portuguese, which was spoken throughout Portugal’s former possessions.
Hesseling (1897) argued that the use of Malyo-Portuguese at the Cape
Colony was both early and widespread, and that this was a much more
important factor in the partial creolization of Afrikaans than the influence
of Khoi, although this interpretation has been seriously challenged.
It has long been asserted that contact with other European languages
played a role in the relatively rapid evolution of Dutch into Afrikaans. In
the 1660s the Europeans were mostly Dutch (numbering 104), but there
were also a considerable number of Germans (40) and individuals of other
nationalities (23) (Reinecke 1937:576). However, most of the Germans
were men who had come to work for the Dutch East India Company;
many knew the Low German dialects closely related to Dutch, and most
soon learned Dutch and became assimilated (Scholtz 1970:84). In 1688
some 225 French-speaking Huguenots sought refuge in the colony; they
made up a full sixth of the white population at the time, but the Dutch
governor ensured their assimilation by requiring their farms to be inter-
spersed among those of the Dutch rather than in a single settlement. The
resulting social contact led to intermarriage and their acceptance as “true
Afrikaners,” so that their language disappeared within some forty years
(ibid.).
It seems clear that the crucial period in the formation of Afrikaans was
the second half of the seventeenth century, by the end of which “gender
distinction had disappeared completely; wij ‘we’ was replaced by ons ‘us’;
Windhoek MOZAM-
B O T S W A N A
Rehoboth BIQUE
K a l a h a r i D e s e r t LIMPOPO
NAMIBIA
Pretoria(A) GA
N
Namaland GAUTENG LA
A
NORTH Johannesburg
M
SWAZILAND
U
(E)
MP
WEST .
a a lR
V
Griqua- FREE STATE
land KWAZULU-
.
Little Ora n ge R Bloemfontein
NATAL
Namaqualand Or LESOTHO
NORTHERN CAPE a n ge R . Durban
(E)
EASTERN
CAPE
WESTERN East London (E)
CAPE
Cape Town
(A=E) Port Elizabeth
(E) (A) Largely Afrikaans-speaking
0 300 km
(E) Largely English-speaking
0 200 miles
Map 2 Southern Africa. Originally published in Holm (1989) Pidgins

and Creoles Volume 2 (Cambridge University Press).
the present and perfect had coalesced [although Raidt (1983) places this
later]; and some verbs had already lost the infinitive suffix -en” (Hesseling
1897, translation of 1979:7). Although further development took place
over the next century, during which Cape Dutch became what can now
be recognized as Afrikaans (Raidt 1983:28), a major factor in the crucial
early period was clearly the restructured Dutch of the Khoi and the slaves.
Since the geography of the Cape was unfavorable for the kind of plan-
tation agriculture that arose in the New World, there were no large slave
holders except for the Dutch East India Company itself. While the slaves
from Mozambique, Madagascar, and Asia tended to live in the Cape
Town area, the Khoi often lived out in the country among the Dutch
as domestics and farm hands (van Rensburg, personal communication).
Their partial assimilation was not only cultural and linguistic but also
genetic. Few Dutch women accompanied the first settlers to the Cape,
and in the first twenty years of the colony, some 75 percent of the chil-
dren born to female slaves were fathered by Dutch colonists (Valkhoff
1966:206), as were many children born to Khoi women, so that the Cape
Coloured population began to grow early on. In the seventeenth century

such children received their freedom on reaching adulthood, provided
they spoke Dutch and were confirmed members of the church (Raidt
1983:11). During this early period freed slaves of mixed ancestry were
not viewed as a separate ethnic group; they had all the civil rights of whites,
with whom they intermarried (ibid.), although this situation changed in
the eighteenth century. Early “Hottentot Dutch” can be viewed either
as a pidgin (den Besten 1989) or as the early stage of a continuum of
interlanguages that the Khoi evolved in their acquisition of Dutch (van
Rensburg 1989).
2.2.2 Afrikaans: the eighteenth century

During the late seventeenth and early eighteenth centuries many Khoi
retreated into the interior to avoid European domination, often taking
with them a knowledge of Dutch. The detribalization of those who re-
mained at the Cape was hastened by a severe smallpox epidemic in 1713,
which decimated the Khoi. Those who fled took with them what was
left of their social system, and those who remained merged into the
colony’s population of free blacks and Coloureds. By the end of the eigh-
teenth century most of their descendants were monolingual speakers of
Dutch (Combrink 1978:77). From the middle of the century onwards
the European settlers began moving east and then north to areas popu-
lated by Bantu-speaking peoples, but it is important to bear in mind that
within the Cape Colony itself the slaves never greatly exceeded the white
population. In 1708 there were 1,723 free burghers, who owned 1,298
slaves; in 1798 there were 21,746 Europeans, with some 25,000 slaves
(Reinecke 1937:597). Table 5 summarizes the demographic figures given
above.
Although these figures are incomplete, they allow some inferences re-
garding language contact in so far as group origin indicates likelihood
of a first or second language knowledge of Dutch, or, in 1798, the fore-
runner of Afrikaans. (It should be noted that the estimated figure for the
Khoikhoi is for the southwestern Cape from the Oliphants to the Breede
River; however, the relevant figure is the unknown number of Khoikhoi
who were in regular contact with the other groups.) These figures are help-
ful in assessing the social and linguistic developments from 1652 until the
end of the eighteenth century, when Afrikaans had clearly emerged as a
system quite distinct from Dutch, which most Afrikaners could no longer
speak (Raidt 1983:18). Unlike Negerhollands, the fully creolized Dutch
that developed in the Danish West Indies (Holm 1988–89:325–328),
Afrikaans developed in a colony that had remained under Dutch control
Table 5. Estimated population of the Cape Colony, 1658–1798
1658 1660s 1677 1688 1691 1708 1798
White 166 1,723 21,746

Dutch 104
German 40
French 225
Other 23 1,125
Slaves
African and freed 194
Asian 190
Total 350 1,298 25,000
Khoikhoi 50,000?∗ ? ? ? ? ?
(Based on Reinecke 1937:568–697; ∗ Elphick and Malherbe 1989:3.)
up to that point. The maintenance of standard Dutch as the language of

government, education, and religion undoubtedly played an important
role in preventing Cape whites from adopting most of the creole features
used by non-whites, as happened in the West Indies. However, there is
unambiguous evidence that some such borrowing did in fact take place
in the case of Afrikaans, although to a more limited extent (see below).
It is also of primary significance that whites continued to make up a full
half of the speakers of Afrikaans, compared to the 6 percent white mi-
nority of the speakers of Negerhollands. White Afrikaans speakers were
able to impose a much more European character on their local culture
and language, much like the large minority or majority of whites in the
American South.
2.2.3 Afrikaans: the nineteenth century

The French occupation of the Netherlands in 1795 led Britain to seize
the Cape Colony, which it gained permanently in 1814. The arrival of
5,000 British settlers in 1820 brought the first large body of Europeans
not to be assimilated into Afrikaans culture. As in Ceylon, Guiana, and
other Dutch colonies that the British took over at this time, English came
to replace Dutch in government and education, although the Afrikaners
firmly maintained it in their religion with the 1637 Statenbijbel in standard
Dutch. Afrikaners had begun moving north into the interior in search of
better pastures from the middle of the eighteenth century, but when the
British abolished slavery in 1833, some 12,000 Afrikaners left the Cape
for what became Natal in the east, despite Zulu resistance. However,
Britain annexed Natal in 1843 and many of the Afrikaner farmers or Boers
migrated to the interior, establishing the republics of the Orange Free

State and the Transvaal in the 1850s, where standard Dutch was made
the official language despite the difficulty it presented many Afrikaans
speakers. The British, Dutch, and even many of its own speakers looked
down upon the Boers’ “corrupt kitchen Dutch” and considered it inca-
pable of serving any serious purpose. However, in the second half of the
nineteenth century some Afrikaners, feeling increasingly threatened by
the political and cultural dominance of the English, began a movement
to promote the use of their language in writing (Deumert 1999). Citing
the example of the Negerhollands Bible for the Danish West Indies, it
was proposed that the Statenbijbel should be translated into Afrikaans to
facilitate its use among the Coloureds and poor whites who had difficulty
reading Dutch. Despite long opposition, this was finally accomplished in
1933, and the translation also came into general use among whites. Sup-
port for the new literary language remained divided until the Boer War
(1899–1902), after which the victorious British forced the Boer republics
to join the Union of South Africa.
2.2.4 Afrikaans: the twentieth century

In 1910 “Dutch” was given equal status with English within the Union
of South Africa, but in 1925 it was made explicit that “Dutch” included
the newly standardized Afrikaans language, i.e. Afrikaans Nederlands
or ‘African Dutch.’ The three main regional varieties of the language
are (West) Cape Afrikaans, Orange River Afrikaans, and East Cape
Afrikaans. East Cape formed the basis of the standard, but also has
nonstandard sociolects. West Cape (Kaaps) is spoken in the area that
made up the Cape Colony in 1750 and has many features of the nonstan-
dard speech of the Coloureds, who today make up 54 percent of Cape
Town’s population (Makhudu 1984:134). Orange River Afrikaans, spo-
ken by by the descendants of the Khoi who retreated to the Orange River
area in the eighteenth century, is “widely recognized as a pidgin or creole”
variety in South Africa (van Rensburg 1989:136), although it appears to
be a post-creole. It is spoken by several groups, including the Griqua, a
group of mixed Khoi, slave, and white ancestry that migrated from the
Cape Colony north to the Orange River area in the eighteenth century
(Rademeyer 1938; van Rensburg 1989). It is also spoken by the Rehoboth
Basters, the descendants of Dutch farmers who married Khoi women on
the northern fringes of the Cape Colony in the eighteenth century. In the
middle of the nineteenth century they were driven north by encroach-
ing white farmers and attacking indigenous peoples; in 1870 they settled
in Rehoboth in what was to become the German protectorate of South
West Africa in 1884, thus bringing Afrikaans to the area that recently be-
came independent Namibia (van der Merwe 1989, 1990; Luijks 2000).
Finally, nonstandard Afrikaans is also spoken by many of the 127,000
Cape Malays, descendants of the slaves brought from Asia. It was their
Islamic clergy who first put Afrikaans into writing in the nineteenth cen-
tury, using Arabic script. Kotzé (1984, 1989) considers the Cape Malays
“a small but linguistically significant core group, since it can be shown
that the nonstandard characteristics that are viewed as typical of Cape
Afrikaans are quantitatively strongest in the speech of this community”
(1984:42).
2.3 BVP: the sociolinguistic setting of its development
2.3.1 BVP: the sixteenth century

Because African slaves were brought to Brazil in such great numbers
that they and their descendants came to make up the majority of the
population in certain parts of the country – and of the country as a whole
at certain points in its history – it would be reasonable to expect a more
fully restructured variety of Portuguese to have developed there, much
as creoles developed under apparently similar sociolinguistic conditions
elsewhere. As Guy (1981) put it:
From the social historical standpoint, our question probably would not be “Was
Portuguese creolized in Brazil?”, but rather “How could it possibly have avoided
creolization?” (1981:309)
However, there is a paucity of evidence to show that a widespread, sta-

ble creole ever became firmly established in Brazil. This may be due to
the way in which differing sociolinguistic conditions there affected lan-
guage transmission. Yet certain features of the nonstandard variety indi-
cate the influence of Amerindian, African, and creole languages. Brazil
eventually became the world’s greatest importer of slaves, receiving 38
percent of all Africans brought to the New World, as compared to the 4.5
percent who went to British North America (Curtin 1975:41). But the
point at which these Africans arrived, the sociolinguistic situation they
found, and the portion of the population they made up are factors of
crucial importance regarding the degree of restructuring that Portuguese
underwent in Brazil.
In one important respect Brazil was unlike the first places where
Portuguese-based creoles became established: the Cape Verde Islands
and São Tome, which the Portuguese colonized in 1462 and 1485, re-
spectively. These islands off the coast of western Africa had no inhabitants
when the Portuguese came upon them, so there was no local language
to learn when settlers arrived from Portugal and slaves were brought
from the African mainland, putting their languages in contact and lead-
ing to the development of new creoles within a generation. Brazil, by
contrast, was already inhabited. When the Portuguese began to explore
its vast coast after 1500, they found it to be inhabited by Amerindians
speaking closely related varieties of Tupi. As the Portuguese established
settlements from the 1530s onwards, contact among the various Indian
subgroups increased and there evolved a common Tupi vocabulary fit-
ting into a shared syntactic framework which was relatively free of com-
plicated morphology. This koine, which the Portuguese also learned for
contact with the Indians, came to be called the Lı́ngua Geral, or general
language for communication throughout the colony. Sampaio (1928:3)
claimed that during the first two centuries of colonization, Lı́ngua Geral
was the principal language of three-quarters of Brazil’s population,
albeit with growing bilingualism in Portuguese. Even the Brazilian-born
Portuguese settlers, often raised by Tupi-speaking nurses, used the lan-
guage with ease and seemed to have a strong emotional attachment to it
(Reinecke 1937:692). There was a great deal of intermarriage between
the Portuguese and Tupis, and in some parts of Brazil people of such
mixed ancestry came to predominate; their mother tongue was Lı́ngua
Geral. One such group was the bandeirantes of São Paulo, who carried
the flag of Portugal ever further inland in their search for slaves and gold,
taking their language with them.
Thus it is not clear what language(s) African slaves encountered on ar-
riving in Brazil during the colony’s first two hundred years. They may well
have had to learn Lı́ngua Geral more often than Portuguese, as suggested
by Reinecke (1937:549), so that during the linguistically crucial first gen-
erations of the colony there was little opportunity for a fully restructured
variety of Portuguese to become established among African slaves unless
they had brought such a language with them from Africa. If Africans first
learned Lı́ngua Geral as a second language and their descendants learned
it as their mother tongue, their later shift to a Tupi-influenced but unre-
structured variety of Portuguese along with the rest of the population dur-
ing the eighteenth century would have established such Portuguese as a
second language to be learned by slaves newly arrived from Africa. Under
these circumstances there would have been no need for the establishment
of a creole. In any case, the first generations of Africans arriving in Brazil
did not encounter the same kind of linguistic situation as those coming
to the off-shore islands of Africa did, or indeed as did those arriving on
Caribbean islands whose native populations had all but disappeared.
Brazil also differed from most of the areas where creole languages de-
veloped in that Africans or their descendants made up only a quarter of
Dominica
Martinique
Caribbean Sea
St Lucia
ao
Aruba raç ire St Vincent Barbados
Cu Bona
Grenada
Tobago
Trinidad
A T L A N T I C
Caracas
VENEZUELA O C E A N
Georgetown
Paramaribo
British
G Dutch Cayenne
Bogotá U I French
SURA N A
INA
COLOMBIA M
RIO
AMAPÁ
BRANCO
P A R Á
AMAZONAS MARANHÃO RIO
CEARÁ GRANDE
DO NORTE
U N I T E D S T A T E S PARAIBA
PIAUÍ PER
N AM
BUCO
ALAGOAS
GUAPORÉ
O F B R A Z I L SERGIPE
BAHIA
PERU MATO GROSSO GOIÁS
Salvador
La Paz Brasilia
BOLIVIA
Helvécia
MINAS GERAIS
ESPIRITO
SANTO
A N
SÃO PAULO
PA RIO DE JANEIRO
R
O C E
A São
G Paulo Rio de Janeiro
PARANÁ
U
E
AY
Asunción
I L
I C
SANTA CATARINA
ATLANTIC
P A C I F
C H
ARGENTINA RIO GRANDE O C E A N

DO SUL
0 250 500 750 km
0 250 500 miles

URUGUAY
Map 3 South America, 1830–1956. Originally published in Darby and

Fuller (1978) The New Cambridge Modern History Atlas (Cambridge
University Press).
the population from 1600 to 1650; it was not until the 1770s that they
constituted over 50 percent of the population (Marques 1976:359, 435),
reaching 65 percent in 1818 (Reinecke 1937:556). Yet whites, the group
most likely to speak unrestructured Portuguese, never made up more than
a third of the population until the second half of the nineteenth century
(see table 6), when slavery ended and the linguistic dice had already been
Table 6. Estimated population of Brazil, 1538–1890
1538–1600 1601–1700 1701–1800 1801–1850 1851–1890
African-born 20% 30% 20% 12% 2%

Creole Africans – 20 21 19 13
Integrated 50 10 8 4 2
Amerindians
Mixed – 10 19 34 42
European-born 30 25 22 14 17
Creole whites – 5 10 17 24
(Based on Mussa 1991:163, cited in Mello 1997:85.)
cast. Thus there were proportionately fewer whites in Brazil than in most
of the American South, although they always exceeded 20 percent, the
maximum proportion of native speakers associated with full creolization
(see introduction to this chapter).
Yet national population figures for large countries can obscure the local
conditions of particular speech communities. For example, in what be-
came the United States whites soon far outnumbered all other groups, but
blacks still predominated in certain areas such as coastal South Carolina
and Georgia, where a creolized variety of English developed and sur-
vived (section 2.1.3). The same seems likely to have been the case in the
sugar-growing areas of Brazil, where labor-intensive production required
plantations with a majority of African slaves since Indians proved unsuit-
able for this work. Lucchesi (2000:42ff.) claims that “the expansion of
the Portuguese language in Brazil until the beginning of the 18th century
occurred parallel to the expansion of the sugar-plantation society in the
Northeast” [my translation].
Marques (1976:362) points out that each sugar mill in Brazil required
a minimum of 80 slaves, besides the hundreds needed to work the fields.
The number of such mills increased rapidly, from 1 in 1533, to 60 in
1570, 130 in 1585, and 346 in 1629. According to Silva Neto (1963:79),
of the 14,000 Africans in Brazil in 1597, almost all were in the sugar-
growing settlements of Bahia and Pernambuco. At this time there were
10,000 Africans in Pernambuco, making up fully half of its population,
which also included some 8,000 whites and 2,000 Indians. Moreover,
these Africans constituted a considerable proportion of the small total
population of Brazil at this linguistically crucial early stage (numbering
only a few tens of thousands), making it more susceptible to influence
from the fully creolized Portuguese brought to Brazil from Africa. Evi-
dence that a more highly restructured variety of Portuguese had been the
language of coastal Brazilian sugar plantations rather than Lı́ngua Geral

can be deduced from the fact that many of the earliest Portuguese sugar
planters and their slaves came to Brazil by way of São Tomé (Ivens Ferraz
1979:19). The Portuguese had settled this island in the decades following
their discovery of it in 1470, intermarrying with the slaves they brought
from the mainland. A creolized variety of Portuguese evolved along with
the cultivation of sugar on large plantations. The prosperity that this
brought to São Tomé during the first half of the sixteenth century waned
during the second half as slave rebellions and maroon attacks eventually
destroyed the island’s economy. The Portuguese began abandoning São
Tomé in large numbers, many going to Brazil (ibid.). It is unlikely that
they would have left behind the greatest financial asset needed to estab-
lish sugar plantations in Brazil, namely their creole-speaking slaves from
São Tomé. Moreover, there are abundant phonological, syntactic, and
lexical features linking São Tomé Creole Portuguese and BVP (Holm
1992b).
2.3.2 BVP: the seventeenth century

Further evidence of the use of restructured Portuguese on Brazilian sugar
plantations comes from New Holland, the empire that the Dutch tried
to carve out of northeastern Brazil from the modern states of Sergipe to
Maranhão from 1630 to 1654. After the Dutch seized this area from
Portugal in their struggle for independence from Spain (united with
Portugal under a single monarch from 1580 to 1640), they found allies
in many of the marranos (crypto-Jews) who had settled in Brazil to avoid
the Inquisition in Spain and Portugal. The more tolerant Dutch al-
lowed them to revert openly to Judaism, and they were joined by other
Sephardic Jews whose families had earlier fled the Iberian Peninsula to
take refuge in Holland (Goodman 1987). The Dutch and the Sephardic
Jews seem likely to have communicated in Spanish or Portuguese (or re-
structured varieties of these) with one another and with their Brazilian
slaves, mistresses, and wives. Few Dutch women came to Brazil, which is
why the Netherlanders were unable to establish their language or culture
there during the period they held this colony – a full generation (Boxer
1965:227). When the Portuguese regained the area in 1654, the Dutch
and most of their Jewish collaborators were forced to leave Brazil. Many
resettled in the Carribean regions, particularly in Dutch holdings in the
Guianas and on islands such as Curaçao (Goodman 1987).
Modern Papiamentu, the creole language of the Netherlands Leeward
Antilles (Curaçao, Aruba, and Bonaire), reveals unmistakable Portuguese
influence in its most basic vocabulary (Schuchardt 1882:895; Jeuda
1991). In fact Hancock (1969b:26) identifies Papiamentu as an off-

shoot of a Brazilian variety of pidgin or creole Portuguese. However,
the Portuguese element in Papiamentu has been obscured by the creole’s
three centuries of close contact with the Spanish spoken by the inhabi-
tants of nearby Venezuela.
There is also a strong Portuguese element in the lexicon of Saramac-
can, an English-based creole spoken in the interior of Suriname (formerly
Dutch Guiana). This has taditionally been attributed to the influence of
the Portuguese-speaking Jewish plantation owners who fled from Brazil
and made up 75 percent of the entire European population of Suriname
at the end of the seventeenth century (Rens 1953:22). However, there
has been considerable debate as to how Portuguese actually came to in-
fluence Papiamentu and Saramaccan. It has been suggested (e.g. Lenz
1928; Voorhoeve 1973) that this influence was via a pidginized variety of
Portuguese used in the slave trade between Africa and the New World,
which was partly learned by slaves and carried over into the languages
which evolved in the Caribbean area. However Goodman (1987) pre-
sented considerable evidence supporting the view that the Portuguese
element in Papiamentu, Saramaccan, and certain other creoles in the
Caribbean area (e.g. the French-based creole of Guyane) was introduced
by refugees from Dutch Brazil. If this is indeed the case, then Portuguese,
whatever its degree of restructuring, must have been the language of
Brazil’s coastal sugar plantations since at least the early 1600s. Under
these circumstances, Lı́ngua Geral could not have acted as a buffer to
prevent the restructuring of Portuguese there.
It has recently been debated that “no slaves could have been taken
from Brazil to Surinam” (Ladhams 1999:232) and thus “the origin of
the Portuguese element in the Surinam Creoles should not be sought
in Brazil” (Arends 1999:204). However, Goodman (1987:368) asserted
that “According to van Dantzig (1968:77) they took slaves along as
well” and pointed out that the 1,200 refugees from Brazil who landed
in Guadeloupe and Martinique in 1654 included 300 slaves (1987:391).
Smith (1999) reasons that the presence of pidgin Portuguese features
found in the Portuguese-derived vocabulary of Saramaccan but not in
Sranan supports their origin in Dutch Brazil. To avoid a lengthy digres-
sion here, the interested reader is referred to these articles, but it should
be pointed out that the transfer of restructured Portuguese from Per-
nambuco to the Caribbean area does not require the massive transfer of
slaves: it is well known that in the seventeenth and eighteenth centuries
whites were often proficient in creole languages and appear to have been
the agent of their transfer in some cases, such as that of Creole French to
Louisiana; cf. Hull (1983).
Furthermore, there is linguistic evidence that the Portuguese taken

to the Caribbean area arrived via Brazil rather than Africa. Reinecke
(1937:467) noted that the Saramaccan word plakkeh (now spelled pulakë)
“a kind of eel” comes from the north Brazilian form poraquê (elsewhere
in Brazil piraquê) “electric eel,” ultimately from Tupi pura’ke (da Cunha
et al. 1982). Since the Saramaccans were not in direct contact with Tupi
speakers, this word must have come into their language via Brazilian
Portuguese. There is further evidence in the Sranan word kabúgru
(whence Dutch karboeger) “a mixed Amerindian/negro group living in
Western Suriname” from the Brazilian Portuguese word caboclo (Smith
2002:151). The latter, meaning “indio; mestiço de branco com indio” is
derived from Tupi kariuoka “homen branco” (da Cunha et al. 1982).
There is also linguistic evidence that the Portuguese brought to the
Caribbean area had already undergone some restructuring. Wullschlägel
(1856:328) notes a proverb in Suriname that he tentatively identifies as
“Neger-portugiesisch?”: “Praga beroegoe no mata caballo.” This is ap-
parently the Brazilian proverb, “Praga de burrico não mata cavalo,” or
“The braying of an ass doesn’t kill a horse,” i.e. “The insults of the weak
don’t affect the strong” (Diva Penha Lopes, personal communication).
The lack of de “of ” in the NP “praga beroegoe” indicates a posses-
sive construction also found in BVP and São Tomé CP (4.3.3). It should
be noted that beroegoe also shows the effect of vowel harmony (Holm
2000a:151); this and the zero articles in both the Surinamese and mod-
ern BVP proverb indicate restructuring (Inverno and Swolkien 2001).
Even though the English-based Surinamese creoles could have influ-
enced these last two features, they have a different possessive construction
[possessor + possessed], whereas the construction in the proverb is praga
beroegoe, or [possessed + possessor]. Thus this proverb is perhaps the
best indirect evidence that the language of Brazil’s coastal sugar-growing
areas in the sixteenth and seventeenth centuries was Portuguese (rather
than Lı́ngua Geral) and that this Portuguese had undergone some degree
of restructuring.
Besides Lı́ngua Geral, another factor complicates the reconstruction of
language transmission in Brazil as this might relate to restructuring. This
factor is the retention of African languages over many generations and
among large numbers of people. Such retention was largely absent from
other New World societies in which creoles developed, where Africans
were often put into linguistically heterogeneous groups to prevent com-
munication in a language that their overseers could not understand in
order to foil possible plots to revolt. However, in Brazil linguistic ho-
mogeneity seems to have been valued since it enabled older generations
of slaves to teach newcomers more easily (Le Page and Tabouret-Keller
1985:33). Until 1600, most slaves came from West Africa; from then
until about 1660, Bantu-speaking Angolan and Congolese slaves pre-
dominated, and afterwards they came from both parts of Africa (Marques
1976:361). Influence from both sources is confirmed in the etymological
study of African words surviving in Bahian BVP by Castro and Castro
(1980:46): of 1,955 words, 967 (49.5 percent) were of Bantu origin,
while 988 (50.5 percent) were of West African origin. Mendonça
(1933:28) claimed that the West African or Sudanic people predomi-
nated in Bahia, while Bantu speakers predominated to the north and
south, but this may be an oversimplification.
In the northeastern part of Brazil, African languages seem likely to have
been used as lingua francas in the quilombos or maroon settlements es-
tablished by escaped slaves (Reinecke 1937:557). In 1579, a Jesuit priest
wrote that the
foremost enemies of the colonizers are revolted Negroes from Guiné in some
mountain areas, from where they raid and give much trouble, and the time may
come when they will dare to attack and destroy farms as their relatives do on the
island of São Thomé. (quoted by Kent 1979:174)
That time came when the Portuguese were distracted by the attacking
Dutch in 1630 and the great “Negro Republic” of Palmares was estab-
lished in Pernambuco. Its fortified villages had thousands of inhabitants
with well-organized governments; Palmares endured until the end of the
seventeenth century (ibid.). Freitas (1978:48, my translation) notes that
Os palmarinos falavam uma lı́ngua toda sua, que compreendia formas da lı́ngua
portuguesa, das lı́nguas africanas e secundariamente, das lı́nguas indı́genas.
Quando queriam se entender com eles, moradores e autoridades recorriam a
“lı́nguas” ou seja intérpretes.
[The inhabitants of Palmares spoke a language that was all their own and included
elements from Portuguese, African languages and, to a lesser degree, Indian lan-
guages. When they wanted to communicate with them, the local Europeans and
their authorities took recourse to “linguists” or interpreters.]
Clear evidence that the Portuguese of such maroons was restructured can
be found in the text of a quilombo song:
Folga nego, branco não vem cá
Se vié, o diabo há de levá . . . .
Pau há de levá.
[Relax, black [friend], Whites won’t come here
If they come, the devil take them . . .
They’ll get a good beating.]
(Freitas 1978:43; translated by Heliana Mello)
2.3.3 BVP: the eighteenth and nineteenth centuries

At the beginning of the eighteenth century it was estimated that in Bahia
the proportion of blacks to whites [was] twenty to one . . . in the city of São
Salvador alone twenty-five thousand Negroes were catechized and instructed in
the Angolan tongue. (Nash 1926:127)
There is further evidence that Portuguese – or a restructured variety of

it – was the language of the coastal sugar plantations rather than Lı́ngua
Geral. The latter clearly predominated in São Paulo and other areas where
Indians and mestiços worked on fazendas raising cattle. The shift from
Lı́ngua Geral to Portuguese in the interior seems to have been triggered
by the gold rush in Minas Gerais during the first half of the eighteenth
century. Although the gold had been discovered by bandeirantes who spoke
Lı́ngua Geral, Portuguese soon become the common language in the min-
ing region (Carvalho 1977:27). This was due not only to the great influx
of men from Portugal but also to the great influx of African slaves (to do
the actual mining) from the coastal sugar-growing areas where such slaves
were plentiful and “the inhabitants were proficient in Portuguese; we be-
lieve that they had never actually abandoned the language since their com-
mercial activity kept them in constant contact with Portugal” (Carvalho
1977:27). However, an African-based lingua franca was also used during
the gold rush in Minas Gerais, apparently among newly arrived slaves
from Africa. From a detailed description of this Lingoa Minna written
in 1731 and published two centuries later (Peixoto 1945), Hazel Carter
(personal communication) determined that the language was based on
Fon, an eastern variety of Ewe. As the mines became exhausted after the
1750s, there was a general movement of population away from Minas
Gerais, and this probably played a key role in spreading a newly leveled
variety of Portuguese – the forerunner of BVP – throughout the settled
parts of Brazil, at the expense of Lı́ngua Geral. As Portuguese gradually
came to predominate, it is likely that successive generations of bilinguals
had decreasing competence in Lı́ngua Geral and increasing competence
in Portuguese.
It would be surprising indeed if there had not been considerable in-
terpenetration of the two languages on all linguistic levels under such
conditions. Describing the linguistic situation in the Amazon region in
the nineteenth century as this language shift moved deeper inland, Hartt
(1872:72) noted that
many Portuguese idioms have crept into the Tupı́; but on the other hand, the
Portuguese, as spoken in the Amazonas, besides containing a large admixture of
Tupı́ words, is corrupted by many Tupı́ idioms.
It would seem probable, therefore, that the Lı́ngua Geral variety of Tupi
left a strong mark on the BVP of rural peasants (Silveira Bueno 1963).
The process involved was language shift, which can involve considerable
restructuring (Thomason and Kaufman 1988:110–146). Different lin-
guists’ attribution of BVP features to the influence of both Tupi and West
African languages presents no real dilemma: such totally unrelated lan-
guages may well coincidentally share structural similarities (e.g. syllabic
structure rules, progressive nasalization, a lack of many kinds of inflec-
tions) which simply converged to reinforce one another in shaping BVP.
African languages continued to be spoken by large numbers of people
in Brazil during the eighteenth and nineteenth centuries. Yoruba became
established as a lingua franca in the state of Bahia and as late as 1900 an
observer noted that
The Nagô [i.e. Yoruba] language is in fact much spoken at Bahia by almost all the
old Africans of different nationalities and by a large number of [black] creoles and
mulattoes. When in this state it is said that a person speaks the Lı́ngua da Costa,
invariably the Nagô is meant. (Nina Rodrigues, quoted by Reinecke 1937:553)
Although it is no longer spoken natively, Yoruba is still used as a litur-

gical language in Bahia (Holm and Oyedeji 1984). I recorded it there
in 1983 during a religious ceremony, and its identity was later con-
firmed by Oyedeji, a native speaker from Nigeria who listened to the
tapes. Remnants of Bantu languages have also been identified (Fry, Vogt,
and Gnerre 1981). Such long-term survival of African languages as lin-
gua francas among different African ethnolinguistic groups and their de-
scendants seems likely to have worked against the retention of an early
Portuguese creole for such communication since it was not needed for
this purpose.
It is important to stress that European Portuguese continued to reassert
itself in Brazil. Silva Neto (1963:68ff.) notes that when the Portuguese
court fled to Rio de Janeiro during the Napoleonic wars, “tudo se foi
re-europeizando” [“everything was re-Europeanized”], from houses to
clothes. In 1808 Rio had 50,000 inhabitants and the number of whites
was much lower than blacks; by 1835 the population was 110,000
with many more whites, given the arrival of 24,000 Portuguese and a
large number of other Europeans (ibid.). To this day, Rio’s inhabitants
are known for their palatalization of plural -s in the European Portuguese
manner, a prestige pronunciation that occurs only sporadically outside
Brazil’s former captial. Political independence from Portugal in 1822 saw
the creation of an empire that lasted until the abolition of slavery in 1888.
However, the establishment of independent Brazilian cultural institu-
tions did not mean that the grammar of European Portuguese lost any
status in education. Silva Neto (1963:90) quotes a priest in the north-

eastern province of Maranhão writing in 1819:
At present the language of this country is Portuguese; educated people speak it
very well; however, among the rustics there is a certain dialect, which I believe is
the result of the mixture of the languages of the diverse nations that have inhabited
Maranhão.
A sociolinguistic factor relevant to the survival of a restructured ver-

nacular was race relations. Although observers from societies with more
rigid racial caste systems sometimes believe Brazil to be a country without
racism, black Brazilians are usually quick to disabuse one of this notion.
Although Brazilians of all races have been equal before the law since abo-
lition, without any form of officially sanctioned segregation, blacks still
tend to be poor and powerless, while the rich and powerful still tend
to be white or light skinned. An important historical factor determining
the present racial structure of Brazilian society was the official sixteenth-
century colonial policy of encouraging Portuguese men to marry native
women in their colonies in order to establish a local part-Portuguese com-
munity with cultural and political loyalty to Portugal. In Portuguese India
the founding governor, Afonso de Albuquerque, carried out this policy
by granting such couples a state-subsidized dowry (Marques 1976:249).
As a country with a small population – just over a million during this
period (Boxer 1969:49) – it was the only way Portugal could maintain its
far-flung trading empire, extending from Brazil to what is today Indone-
sia. Portuguese women simply remained at home in Portugal during this
early period. Marques (1976) notes that
on the whole the [early] social patterns of Brazil copied that of the Atlantic
islands, particularly Cape Verde and São Tomé. . . . cross-breeding derived
from the needs of nature, not from race equality. The whites were always con-
sidered superior to the others and held most offices of leadership, although
tolerance and respect toward both mulattoes and mamelucos [Portuguese-Tupi
mestiços] reached a higher level in Brazil than probably anywhere else (1976:255,
360)
2.3.4 BVP: the twentieth century

The greater frequency with which Portuguese men fathered children by
their slave women probably led to the far higher rate of manumission in
Brazil than in the Caribbean. Racial mixing worked against the mainte-
nance of the rigid caste system that helped to preserve creolized language
varieties elsewhere. Later, as Afro-Brazilians had fewer barriers to face in
rising socially, they also had more incentive to learn Standard Brazilian
Portuguese as a mark of their standing; such circumstances seem likely

to have accelerated the loss of non-Portuguese features in their speech.
As education started to become more widely available, Sousa de Silveira
noted in 1921 that “even the Negroes speak better today than they used
to” (reprinted in Pimentel Pinto 1981:27).
The language-related problems facing modern Brazil are still daunting:
out of a total population of 165 million, 17 percent, or some 28 million
Brazilians, are illiterate (Famighetti 1997:746) and likely to speak only
nonstandard BVP. Azevedo (1989:869) notes that
Vernacular speakers must learn to read and write in a dialect they neither speak
nor fully understand, a circumstance that may have a bearing on the high dropout
rate in elementary schools: as recently as 1983 only 70 out of every thousand
students reportedly finished the 8th grade.
Today BVP is the language usually spoken by lower class Brazilians

with little education. It differs considerably from Standard Brazilian Por-
tuguese (SBP), the literary language usually spoken by educated middle-
and upper-class Brazilians, particularly in formal circumstances. It should
be noted that BVP is not a variety spoken only by black Brazilians. The
most divergent varieties of BVP and SBP are the extremities of a contin-
uum of lects that relate more to social class and education than to race.
Although blacks are certainly overrepresented in the lower class and un-
derrepresented in the upper class, the structure of Brazilian society is
such that all sociolects have speakers of all races, just as other aspects
of Brazilian culture are shared by all ethnic groups. Social variables for
language usage are further complicated by the fact that most middle- and
upper-class Brazilians are in fact bidialectal:
Educated speakers are often ambivalent about the vernacular, caught as they are
between a prescriptive norm based largely on written models . . . and a linguistic
reality that departs considerably from that ideal model. In a society where literacy
and grammaticality of language use are held as the hallmark of education, such
ambivalence is inevitable when one is aware that the occurrence of stigmatized
features in one’s own speech is too spontaneous and frequent to be dismissed as
resulting from occasional slips of the tongue. It is apparent that rules considered
typical of the vernacular are present in the native linguistic repertory of educated
speakers, who acquire the standard largely through normative coaching, which
includes not only formal school instruction but also pressure from family and
peers (Azevedo l989:862)
Thus educated Brazilians are comparable to educated African Ameri-

cans, who use the standard in writing and speaking in formal situations
but often use the nonstandard in other social situations to signal intimacy
or solidarity (1989:868). It is often not clear which variety is a speaker’s
first or dominant language. There are rural usages that are unlikely to
be found in the speech of educated Brazilians, although many such us-
ages have become part of urban sociolects through immigration from the
countryside.
2.3.5 Helvécia Portuguese

The most extreme case of nonstandard rural usage that has been docu-
mented is the restructured Portuguese spoken in the village of Helvécia
in the state of Bahia (HP), recorded by Silveira Ferreira in 1961 but pub-
lished some years later (1985). HP is described by Baxter (1987, 1992,
1997), and Lucchesi (2000). The paucity of the information available
about the social history and linguistic structure of HP makes it difficult
to extrapolate about the general sociolinguistic conditions that shaped
this variety. Discovered during the course of a dialect survey of the state
of Bahia, HP was described in a short article (Silveira Ferreira 1985) that
received little attention until recently.
Helvécia (from the Latin name for Switzerland) is an isolated, one-
street village in the southern part of the state of Bahia, which Silveira
Ferreira reached by jeep, traveling for seven hours over a rough road
from Nanuque. She stayed there only one night, but the young people
she first met (who spoke the usual variety of regional BVP)
diziam que naquela cidade havia muita gente que falava diferente, “engraçado,”
principalmente os mais velhos, e acrescentavam ainda que muitas vezes era difı́cil,
para eles mismos, filhos da terra, entenderem. [“They said that in that town there
were a lot of people who spoke differently, ‘funny,’ mainly the oldest people, and
they added that many times it was difficult even for them – natives of the town –
to understand them.”] (Silveira Ferreira 1985:23)
Helvécia was founded in the eighteenth century by Swiss (speaking both

German and French) and Germans, who bought slaves in the state of
Bahia in order to raise coffee. It was first referred to in 1818 under the
name Colônia Leopoldina, and described forty years later (Tölsner 1858,
cited in Neeser 1951) as a prosperous community of 40 plantations with
200 whites (mostly German and Swiss, with some French and Brazilians)
and 2,000 blacks. Lucchesi (2000) notes that Helvécia had a relatively
high proportion of African-born slaves (some 40 percent) at this time.
Silveira Ferreira concluded that
With the decline of the coffee industry, the whites gradually abandoned the area,
leaving as their legacy the name of the town, mixed descendants, and some family
names. The blacks, however, remained isolated, continuing to speak a creole that
must have had general currency among them since vestiges of it persisted until
1961. (1985:22)
Lucchesi (2000:84) describes HP as the result of semi-creolization. From

what is known about the community’s history, its main distinguishing fea-
ture is that it was founded by non-native speakers of Portuguese. German
seems likely to have served as a lingua franca among the first gener-
ations of Europeans, who probably learned much of their Portuguese
from their slaves in order to communicate with them. If this was the case,
the Europeans’ Portuguese could not have served – as seems likely else-
where in Brazil – as a model for decreolization. Since the slaves were
originally purchased in Bahia in the eighteenth century, they must have
already known some Portuguese – apparently in a restructured or sec-
ond language form, to judge from the variety that survived there until
the twentieth century. It is possible that this variety was also influenced
by the second language Portuguese of the first generation of Europeans,
although their children must have become proficient in the Portuguese
of the slaves. The isolation of the community was certainly not complete,
since some Brazilians joined it, slaves from Africa were brought in (at least
until abolition in 1888), and trade with the outside brought it prosperity.
Contact with the less-restructured BVP of the region seems likely to have
increased over the years, bringing the speech of younger people more in
line with that of other communities in the area. Although the sociolin-
guistic history of Helvécia Portuguese is not typical of BVP varieties in a
number of respects, and its existence cannot be viewed as unambiguous
evidence that fully creolized Portuguese was ever widely spoken in Brazil,
it does demonstrate that Portuguese was restructured there to varying
degrees.
2.4 NSCS: the sociolinguistic setting of its development

Today the Caribbean basin constitutes a generally definable dialect area
of American Spanish, although there is some disagreement on exactly
how far inland this dialect is spoken on the mainland. It includes the
varieties found in the Greater Antilles (Cuba, the Dominican Republic,
and Puerto Rico), pockets in Trinidad and Guyana, the Caribbean coasts
of Venezuela and Colombia, and some coastal areas of Central America
and Mexico, as well as offshoots in the Miami and New York City areas
of the United States.
2.4.1 NSCS: the sixteenth to eighteenth centuries

The Spanish colonized many parts of this area for over a century be-
fore the establishment of British, French, and Dutch colonies in the sev-
enteenth century. The indigenous Arawakan languages of the Greater
U . S . A .
AT L A N T I C
OCEAN
Miami
Gulf of Mexico
Mexico
Puerto Rico
Cuba
Mexico
Jamaica Haiti
Belize
Dominican Republic
Coastal area of Guerrero Honduras & Caribbean Sea
and Oaxaca Nicaragua
Guatemala Aruba, Curaçao, Bonaire
El Salvador
Cartagena Trinidad
Small groups of Afro-Hispanics without Panama
numerical importance.These groups Caracas
probably never spoke an Afro-Hispanic Costa Rica Barlovento
variety
Parts of the State of
Bogotá state of Zulia Miranda
The lowlands of Colombia Venezuela
(Chocó, Cauca Valley, Nariño) Colombia
Popayán
Ecuadorian coast (especially Patía (Colombia)
the province of Esmeraldas)
Chota Valley (Ecuador)
Quito
Ecuador
Guayaquil
P A C I F I C
Brazil
Peruvian coast (especially
between Trujillo and Chiclayo)
O C E A N Per u
Lima
Peruvian coast (especially

B ol i v i a
between Chincha and Ica)
Map 4 ‘Localización aproximada de áreas afro-hispanas en América’.

Originally published in Perl/Schwegler (eds) (1998) América Negra:
panorámica actual de los estudios lingüı́sticos sobre variedades hispanas, por-
tuguesas y criollas (Iberoamericana). Reproduced by kind permission of
Iberoamericana.
Antilles died out during the course of the sixteenth century as their speak-
ers were virtually exterminated by forced labor and disease or absorbed
into the Hispanicized population. During the same period the large-scale
importation of African slaves began, but in most parts of the Caribbean
basin the populations of the Spanish colonies did not become overwhelm-
ingly black by the middle of the eighteenth century as did those of the
British, French, and Dutch, who in the main concentrated more on labor-
intensive sugar plantations. Until the nineteenth century the Spanish is-
lands generally had small farms raising various crops (e.g. tobacco, coffee,
cattle) on which a small number of slaves worked alongside poor whites

and a steadily increasing number of free mulattos.
It has been claimed that “Whites in Iberian America felt less hostility
toward and a closer affinity with people of color, especially if they were
light-skinned, than did the French or Anglo-Saxons” (Cohen 1980:108).
Conditions on the Spanish islands were thus more conducive to the
Africans’ linguistic and cultural assimilation than they were on the over-
whelmingly black sugar islands with the rigid racial caste system of plan-
tation slavery (Mintz 1971). Spanish slave laws were also more liberal;
under the Cuban system called coartación, slaves could purchase their
own freedom by paying instalments. Through this and more prevalent
manumission, over 40 percent of Cuba’s non-white population was free
by 1774 (Laurence 1974:492) – although these figures were to change in
the next century. At this time Cuba’s population was 56 percent white,
18 percent free colored, and 26 percent slave (1974:495). During this
period slaves accounted for only 10 percent of Puerto Rico’s population
(Reinecke 1937:266), although at the the beginning of colonization in
the sixteenth century whites were outnumbered four to one by Indians
and African slaves (Reinecke et al. 1975:144). De Granda (1970b:76)
interprets the linguistic data in Alvarez Nazario (1961) as evidence of
a Puerto Rican creole that existed until the nineteenth century. In the
Dominican Republic, where over 80 percent of the population has some
African ancestry (Reinecke et al. 1975:146), similar social conditions
seem to have led to fairly thorough linguistic assimilation. However,
on the Caribbean coast of Colombia, which also has a sizeable popu-
lation of African descent, there is a direct reference to an early pidgin in
Sandoval (1627), a Jesuit who lived there from 1605 to 1652 (de Granda
1970a). He refers to a language used between Spaniards and African
slaves of various ethnolinguistic groups, which was “corrupt Spanish,
as the Negroes commonly speak it . . . influenced by the Portuguese
they call the language of São Tomé” (Sandoval 1627: p. 94 of 1956
edn). This and the later emergence of Palenquero Creole Spanish in
northern Colombia provide sufficient evidence that a Spanish-based pid-
gin based on Afro-Portuguese did in fact exist in the Caribbean, but it
does not confirm the speculation of Bickerton and Escalante (1970:262)
that there existed “a Spanish-based creole spoken in many parts of the
Caribbean during the sixteenth and seventeenth centuries” since we do
not know (apart from Curaçao, where there is an unresolvable ques-
tion regarding the continuity of Spanish after the Dutch takeover) the
fate of pidgins and jargons that may have been spoken elsewhere in the
region.
2.4.2 NSCS: Habla bozal in the nineteenth century

However, there is fairly clear evidence that a pidgin developed on Cuba
during the nineteenth century, although it is less clear that it ever became a
true creole. Large-scale sugar plantations were rapidly developed in Cuba
towards the end of the eighteenth century when the Spanish attempted
to gain the markets for sugar lost by the French in Saint Domingue after
independent Haiti plunged into chaos. This caused the importation of
slaves into Cuba to increase steadily: the proportion of the population
that they made up grew from 26 percent in 1774, to 36 percent in 1817,
and 43 percent in 1841. Cuba’s bozales or African-born slaves developed
a language which Pichardo (1862:vii) described thus:
Este lenguaje es común e idéntico en los Negros, sean de la Nación que fue-
sen, y se conservan eternamente, a ménos que hayan venido mui niños: es un
Castellano desfigurado, chapurrado, sin concordancia, número, declinación, ni
conjugación . . . en fin, una jerga más confusa miéntras más reciente la inmi-
gración.
[This speech is uniform among the Negroes, no matter from which nation they
come, and is retained throughout life, unless they came very young. It is distorted
and mutilated Castilian, without concord, number, declension, or conjugation . . .
in sum, a jargon the more confused the more recent the arrival.]
Based on transcriptions of the speech of aged Cubans of African de-

scent recorded earlier in the twentieth century and published later
(Cabrera 1969), Otheguy (1973) concluded that the variety was a creole
(1973:335). Although he was undoubtedly right in considering it to be
“related to other Caribbean Creoles . . . through a common West African
origin” (1973:332), this may have been via the typological similarity of
the languages which influenced both rather than via a creole family tree,
since there is evidence that bozal Spanish was not a creole. Its use was
confined to slaves born in Africa, so it was not a mother tongue; Pichardo
(1862:iii) states quite specifically that “Los Negros Criollos hablan como
los blancos del paı́s de su nacimiento o vecinidad” [“the Negroes born
in Cuba talk like the local whites”]. However, certain features of bozal
Spanish survived among rural black Cubans (Reinecke 1937:271), sug-
gesting that there was partial nativization resulting in partial restructuring
(Figueroa 1999; Ortiz 1998).
2.4.3 NSCS: the twentieth century

More recently, Green (1997:190ff.) analyzed the partially restructured
variety of Dominican Spanish referred to as Pororó in Villa Mella and also
Table 7. Estimated proportion of whites in the Greater

Antilles in the late eighteenth century
Colony Year Percent Source
Cuba 1774 56% (Laurence 1974:495)

Puerto Rico 1776 45% (Alvarez Nazario 1974:76)
Santo Domingo 1794 34% (Lorenzino 1991:13)
spoken in Cambita, two small rural communities in the San Cristóbal dis-
trict near Santo Domingo. A considerable proportion of their population
is descended from African slaves who worked on local sugar plantations
during the colonial period. Today the survival of the Afro-Hispanic roots
of their culture is evident not only in the local speech variety but also
in their cofradı́as, or brotherhoods, showing syncretism with African re-
ligious fraternities (1997:6). Green based her analysis of the grammar
of this variety of NSCS on her recordings of thirty-three speakers living
in these communities, ten of whom spoke basilectal varieties (1997:4).
It should be said that John Lipski (personal communication, October 9,
2002) has suggested that the speech of two of these thirty-three speak-
ers may be abnormal because they may have “a congenital and probably
genetically-based problem.” This seems a somewhat impressionistic di-
agnosis, but I have not quoted the speech of these informants without
noting that they, rather than others, used a particular structure.
Finally, important new light on the relationship between fully and par-
tially restructured Spanish in the Caribbean comes from a recent study
(Schwegler and Morton 2002) documenting morphosyntactic features
hitherto unknown to linguists in the regional vernacular Spanish of bilin-
gual speakers of Palenquero Creole Spanish in Colombia.
In assessing the historical likelihood of partial restructuring in the re-
gion as a whole, it has to be borne in mind that sociolinguistic conditions
on each Spanish island differed to some degree. However, each colony
was part of the same political, economic, and cultural system, and the
similarity of the distinctive features of the Spanish spoken in each country
today reflects the strength of their ties to one another during the Spanish
colonial period. Although colonial census figures for the Greater Antilles
are disparate, taken in different years with demographic groupings that
are not always comparable, we can get an overall impression of the so-
ciolinguistic setting for the development of NSCS if we compare the
proportion of whites on each island at the end of the eighteenth century.
These figures will be discussed in the concluding section, 2.6.
2.5 VLRF: the sociolinguistic setting of its development
2.5.1 VLRF: the seventeenth and eighteenth centuries

Réunion, a small volcanic island lying some 400 miles east of Madagascar
in the Indian Ocean, was uninhabited when the Portuguese first came
upon it in 1511. It was settled in 1663 by some Frenchmen and
Malagasies from Madagascar. The island, called Bourbon until 1848,
became a permanant colony as an outpost of the French East India
Company in 1665. In 1671 the population consisted of 36 whites, 37
Malagasies, and 3 mixed children born on the island (Papen 1978:9–10).
In 1674 the colonists wrote to the company and “asked for wives,
most of them having had to marry Negresses, their slaves” (quoted by
Cohen 1980:49). Some young Indo-Portuguese women were brought
from Daman and Surate (Chapuis forthcoming) and from Pondicherry,
the last a French enclave in southeast India earlier held by Portugal
(Stein 1984:90), and by 1686 the island’s population of 269 included
144 children born to French fathers and mothers who were either
Indo-Portuguese, French, or Malagasy (Baker and Corne 1986:167).
Chaudenson (1974:455) notes that of the 36 Frenchmen with families
in 1686, 12 had Indo-Portuguese wives, 14 had Malagasy wives, and 10
had French wives. They and their children made up 75 percent of the
island’s population, the rest being Malagasy and Indian slaves.
The vernacular French of Réunion appears to have taken something
approaching its present form during this early period. In court testi-
mony taken some time between 1715 (Chaudenson 1974:1147) and 1723
(Baker and Corne 1982:6), a sentence appears whose features are those
of modern VLRF (Hull 1979:212): “Moin la parti maron parce qu’Alexis
l’homme de jardin l’était qui fait à moin trop l’amour,” “I ran away be-
cause Alexis, the gardener, made amorous advances” (my emphasis). In
modern VLRF la marks the perfect (or “passé perfectif ” according to
Watbled 2002) and lete ki marks the durative past (or “passé imperfectif”
[ibid.]; see section 3.5.1). Because of the latter construction, Baker and
Corne (1982:111) referred to the vernacular of the early colonists as lete
ki French. According to their interpretation, this variety was influenced
by the French spoken as a second language by the non-French wives and
slaves, and it became the first language of the children (particularly of
mixed marriages) who were born on the island. During this early period
of settlement, free people made up the majority of Réunion’s population,
and they usually had access to at least one parent who was a native speaker
of French. Although some restructuring took place under these condi-
tions, it resulted more from the borrowing of isolated non-French features
St Denis IN D I A N
0 10 km Ste Marie
0 5 miles N1 Bellepierre Ste Suzanne
La Montagne
Le Port OC E A N
La Possession
W St Andre
St Paul Dos-d'Ane in
Cirque dw
IN D I A N de 2277 a
Mallate rd St Benoit
Grand
Ilet Salazie
S
2992
id
O C E A N L
e
2896 3000
Trois Bassins e e Ste Rose
w
a Cilaos
N1
r Plaine
St Leu d des Palmistes
Plaine des 1606
LEGEND (a) Makes La Plaine
Entre des Caffres Vo l c a n i c
Les Deux S 2631
Areas above 400 meters Avirons Etang
Sale i
Tampon d Area
Predominantly (East) Indian Pierrefonds e
St Louis
N3
Predominantly White
Plaine des
Predominantly Mixed and Black Petite Ile Gregues
St Pierre
Scattered groups of (East) Indians N2
predominant St Philippe
Main roads St Joseph
LEGEND (b) St Denis

1 0 10 km
4 2
Basilectal varieties of VLRF 3 0 5 miles
5
6
7
Acrolectal varieties of VLRF Le Brulé
LePort
8
9
Dos-d'Ane
11
Grand Ilet
14 12
IN D I A N 18 17
Grand Place
13 Saint Benoit
10
19 16
Lanouvelle
20
Trois Bassins
O C E A N
23 21 22
Lachaloupe
Cilaos
26
25
Tevelave
28 29 27
24
AF RIC A Plaine des Cafres Bois
31 Riv. Blanc
St Louis 30 Entre Deux
33 32
IND I A N
OC E A N
34
35
38 36
37
ar
asc
St Pierre
d ag
38
Ma
40
Réunion St Philippe
Map 5 (a) ‘Carte de L’ı̂le de la Réunion’. Réunion Ethnic groups. Orig-

inally appeared in the Chaudenson (1998) Créolisation du français et fran-
cisation du créole: Le cas de Saint-Barthélemy et de la Réunion (previously
unpublished paper presented at the Conference on Degrees of
rather than the complete reanalysis of the base language’s structure that
characterizes full-fledged pidginization and creolization.
The cultivation of coffee on Réunion started in 1715, beginning a
plantation system that had far-reaching effects on the island’s demog-
raphy. Over the next half-century the island’s population grew from
2,000 in 1717 (of whom 45 percent were whites according to Chau-
denson 1974:458–459) to 27,700 in 1767, of whom only 5,300 were
free as opposed to 22,400 slaves (Papen 1978:13), brought in mainly
from Madagascar and East Africa, but also from India and even West
Africa. Baker and Corne (1982:111, 1986:166) hypothesize that as the
proportion of the slave population rose from 50 percent to 80 percent of
the total, the new slaves began a period of pidginization and creolization,
leading to the emergence of the predecessor of the modern basilectal
Créole des Bas (see below) and a continuum of varieties ranging from
the regional varieties then spoken in France to the partially restructured
vernacular lects of the island, which had already stabilized before this
period began but which may have taken on additional creole features.
However, the sociolinguistic situation on Réunion during the eighteenth
century differed in one important respect from that of Mauritius during
the same period, or that of Haiti in the late seventeenth century. In both
of the latter colonies, which developed fully creolized varieties of French,
there was a rapid influx of a large, non-French-speaking slave population
with little access to the French used by native speakers. On Réunion, on
the other hand, the first half century of settlement had led to a variety
of French that had undergone only limited restructuring, which was
accessible to the quarter or more of the population that consisted of
slaves. This group had sufficient contact not only with the early colonists
but also with the later slaves to act as an effective agent of language
transmission, so that the restructuring of the local French spoken on
Réunion (i.e. lete ki French in Baker and Corne’s terminology) was not
as complete as the pidginization and creolization of French on Mauritius
and Haiti.
<
Map 5 (cont.)
Restructuring in Creole Languages, University of Regensburg, June 24–
27 1998). Reproduced by kind permission of Robert Chaudenson.
(b) ‘Réseau ALRé’. Réunion Lects. Originally appeared in the Chau-
denson (1998) Créolisation du français et francisation du créole: Le cas de
Saint-Barthélemy et de la Réunion (previously unpublished paper pre-
sented at the Conference on Degrees of Restructuring in Creole Lan-
guages, University of Regensburg, June 24–27 1998). Reproduced by
kind permission of Robert Chaudenson.
2.5.2 VLRF: the nineteenth and twentieth centuries

The French Republic’s emancipation of slaves in 1794 led to an open
rebellion among the colonists on Réunion, but this was soon diverted
by the attacks of the British. Napoleon re-established slavery in 1803
and renamed the island Bonaparte. By 1810 the British had captured
Réunion, Mauritius, and the Seychelles; in 1814 only Réunion was re-
turned to the French, who promptly renamed it Bourbon. Freed from the
administrative and economic domination of Mauritius (Papen 1978:16),
Réunion began a rapid expansion of the cultivation of sugar cane, which
soon replaced coffee as the island’s main crop. When slavery was perma-
nently abolished by the French Republic in 1848 (which renamed the
island Réunion), its population of 110,000 included some 60,000 slaves.
Most of the freed blacks refused to continue working for their former
masters and established themselves on the island’s high central plateaus,
also inhabited by the descendants of Maroons who had been fleeing there
to escape slavery on the lowland plantations since the beginning of the
eighteenth century.
The economic situation of whites holding small farms had been deteri-
orating since the late eighteenth century. Land was becoming increasingly
scarce because of the continuing immigration of French peasants and a
high birth rate. Holdings became even smaller since land was inherited
not by the eldest son but rather by all children, among whom it was di-
vided equally. Many whites became impoverished and took refuge in the
higher elevations of the volcanic mountain slopes, where they lived like
the Maroons who had fled there before them by hunting, fishing, and
poaching. The absorption of small holdings into large sugar plantations
aggravated this situation, and by 1836 two-thirds of the island’s white
population was described as “indigent” (Papen 1978:13). The colonial
authorities viewed with increasing alarm the deforestation carried out by
the Petits Créoles, the poor whites who were joined by freed slaves in the
highlands. Chapuis (forthcoming, citing Grondin 1998) notes that later
the whites came to be referred to as Petits Blancs to set them apart from
the non-white populations of this area. In 1869 it had been observed
that “They have tried, through the centuries – after an initial racial mix-
ing that is beyond any doubt – to remain as white as possible, despite
the fact that some Petits Blancs . . . are practically black” (quoted in
Bourquin 1994:114, trans. by Chapuis). The modern local connotations
of the word Créole in reference to people (which is certainly related to
the use of the same term in reference to their language, regardless of
what this term might mean elsewhere) is made clear in the following
passage:
ici . . . Créole est devenu synonyme de Réunionnais, avec nuances: la majorité

des Créoles ne sont pas exactement blancs mais tous les natifs de la Réunion ne
sont pas admis à la dignité de Créole; ainsi l’enfant du “zoreil” de passage ou
celui du Chinois . . .
[Here . . . Creole has become the synonym of Réunionese, with nuances: the
majority of Creoles are not exactly white, but not all the natives of Réunion can
be called Creoles; e.g. the child of a passing metropolitan Frenchman, or that of
a Chinese . . .] (Vaxelaire 1992:147, my translation)
This complex social identity surely bears a relationship to the group’s

speech, with its mixture of features associated both with Europeans and
non-Europeans.
Baker and Corne (1982:111) see the Créole des Hauts or highland
vernacular as descended from metropolitan and the earlier local lete ki
French, but influenced by the Créole des Bas or lowland vernacular that
developed during the eighteenth century among slaves and others, and
was then learned as a second language by the indentured servants from
southern India who were brought in after the abolition of slavery in 1848
to supply the need for labor on the sugar plantations. Today the poor
white population (30 percent of the total of 692,000 in 1997) typically in-
habits the mountainous interior, while the coastal lowlands are occupied
by people whose ancestry is Indian (25 percent) or African and mixed
(45 percent). The latter groups tend to speak more basilectal lowlands
VLRF, while the whites tend to speak a less basilectal archaic highlands
variety (Baker and Corne 1982:12, 110), although there is a consider-
able overlap of features. Chaudenson (2000) confirms this opposition,
noting that speaking the lowlands basilect is called kozé an nwar, “talk-
ing black” (2000:365), although he notes that non-whites living among
whites talk like whites and vice versa. He refers to the highlands variety
as français créolisé, or ‘Creole-influenced French’, and the lowlands vari-
ety as “un créole moins basilectal que dans d’autres aires,” ‘a creole that
is less basilectal than those found elsewhere’ (2000:373). The lowlands
basilect is marked by the unrounding of the French rounded vowels and
the depalatalization of French palatal fricatives, as well as certain mor-
phosyntactic constructions (see section 3.5.1). These markers are well
known and there is no problem in mutual comprehension (2000:368).
However, Chapuis (forthcoming) asserts that
. . . a clear distinction between the lect of . . . a Petit Blanc and a Kaf or dark-skinned
speaker (a mixture of African and Malagasy origin) cannot be established linguis-
tically. . . . What has, however, been considered a watershed between acrolectal
and basilectal speech varieties is the use of the long form of the verb for the past
progressive by the Petits Blancs and the use of the té i form by the rest of the
population. Not that the same speaker cannot use both in consecutive sentences
or even the same sentence (Cellier 1985a).
This opposition of lects was valid until the 1960s; there is now a
complex interrelationship of these varieties with more acrolectal urban
speech – a continuum with two distinct basilectal extremities. Although
the most basilectal lects of Réunion French are clearly more restructured
than the other four languages examined in this volume, one of the lin-
guists who knows the structure of the VLRF best concludes that “. . . dans
l’évolution du français au créole [réunionnais] la créolisation a été in-
complète . . .” [In the evolution of French to Réunion Creole, creolization
has been incomplete] (Cellier 1985a:135). Yet Watbled (forthcoming) is
also right to assert of VLRF (“créole”): “En dépit des origines françaises,
les structures créoles obéissent à des principes grammaticaux différents”
[‘In spite of their French origins, Réunionnais structures follow different
grammatical principles.’]
2.6 Common sociolinguistic factors

We now return to the discussion of the role of sociolinguistic factors in
determining the structure of these new language varieties resulting from
contact. If we focus on the proportion of whites in the five settings out-
lined above, an important factor in the partial restructuring of languages
emerges. Since race was usually an indicator of whether an individual
spoke the European lingua franca as a first or second language during the
earliest period, it can be inferred that in all of these situations the maxi-
mum percentage of native speakers associated with full creolization – i.e.
20 percent according to Bickerton (1981:4) – was considerably exceeded.
If we compare the above figures to the proportion of Europeans in the late
eighteenth century in some speech communities where full creoles devel-
oped (Holm 1988–89), we find a considerable discrepancy (table 8).
In the case of Réunion, the figures on the ethnicity of the population
in the late eighteenth century would seem to be at variance with the
above pattern: in 1767 over 80 percent of the population were slaves and
only 20 percent were free. This apparent discrepancy stems from the un-
usual circumstances of the early development of the VLRF: as pointed
out in section 2.5.1, the first half century of settlement had led to an
only slightly restructured variety of French that was quite accessible to
the slave population. In 1717 Réunion’s population was comparable to
those in table 8 where partially restructured languages emerged, being
45 percent white and 55 percent non-white. The non-whites had had
sufficient contact with the French settlers during the earlier period to
Table 8. Estimated proportion of whites in various societies in the

late eighteenth century
Colony Developing language Percent
Virgin Islands Negerhollands Creole Dutch c. 6%

Jamaica Jamaican Creole English c. 8%
(rural) Curaçao Papiamentu Creole Spanish/Portuguese c. 7%
Virginia African American English c. 59%
Brazil Brazilian Vernacular Portuguese c. 32%
Cape Colony Afrikaans c. 47%
Cuba c. 56%
Puerto Rico Nonstandard Caribbean Spanish c. 45%
Santo Domingo c. 34%
learn their language with only minor restructuring and then act as an
effective agent in transmitting this local French to the new slaves who
arrived in the great influx after 1717, with whom they also had suffi-
cient contact to allow for something approaching normal language learn-
ing. Thus, the partially restructured French that had developed up to
1717 remained the island’s language, although it surely acquired more
non-French features over the next half century, particularly on the low-
land plantations.
To summarize, partially rather than fully restructured languages de-
veloped in those societies with a higher proportion of native speakers of
the European lingua franca. This important point will be discussed more
fully in the concluding chapter.
Before proceeding with a structural comparison of the five languages
examined here, it should be pointed out that no claim is being made
that these newer language varieties are any more closely related to one
another genetically than the older languages out of which they grew, but it
is being claimed that they underwent a similar kind of partial restructuring
triggered by parallel social and linguistic factors.
To keep the comparison of their structure manageable, we will deal first
with the morphosyntax of the verb phrase (chapter 3), then with the mor-
phosyntax of the noun phrase (4), and finally with the syntax of clauses
(5), presenting the facts about each language variety in turn in each area
of grammar so they can be more readily compared. In our conclusion
(6), we will propose a theory to account for the the sociohistorical and
linguistic facts that we have surveyed.
3 The verb phrase
Introduction
The verb phrase has been of central importance in contact linguistics.
While it is true that no particular set of syntactic features can iden-
tify a language as having undergone restructuring without reference to
its sociolinguistic history, it is also true that the structure of the verb
phrase has been of primary importance in distinguishing creole vari-
eties (e.g. Jamaican) from non-creole varieties (e.g. Caymanian English)
of the same lexical base. In the Caribbean, the non-creoles have their
European system of tense marking (e.g. auxiliary verbs and verbal in-
flections) more or less intact, whereas the creoles have a radically differ-
ent way of dealing with tense and aspect. With few exceptions, basilec-
tal Atlantic creole verbs have no inflections; instead, they are preceded
by particles indicating tense (the time of an action’s occurrence) or as-
pect (referring to its duration, recurrence, completion, etc.). These often
have the outer form of auxiliary verbs from the lexical source language
(which occupy a similar position and usually serve a similar function),
but semantically and syntactically they are much more like the prever-
bal tense and aspect markers in many of the creoles’ African substrate
languages.
The data below show that verbs in the partially restructured languages
have few if any inflections – far fewer than their European source lan-
guages. In some cases their auxiliary verbs can take on the semantic or
even syntactic uses of preverbal markers in fully creolized languages of
the same lexical base, giving rise to constructions quite unlike those in
their source languages. Further non-European constructions found in
these varieties include exotica like discontinuous double negation and
non-verbal predicates without copulas.
72
The verb phrase 73
3.1 The AAE verb phrase
3.1.1 AAE verbal morphology

In AAE the simple present tense is usually indicated by the verb stem
without any -s inflection in the third person singular, e.g.
(1) AAE: Where Miss Annie . . . live now. (Schneider 1989:65)
Working-class white Americans usually confine such deletion to a single
lexical item (“he don’t”) and only in the negative (although “he do”
is found in the dialects of southwestern England according to DARE).
In contrast, Wolfram (1969:36) found that lower working-class African
American males in Detroit deleted the third person present ending 74
percent of the time.
The -s inflection can also occur with other persons in AAE:
(2) AAE: I members de first shoes I ever had. (ibid.; Labov et al.
1968).
The omission of the third person singular -s is highly stigmatized and
considered a marker of social class.
In AAE a verb with past reference does not need to be marked for
the past tense; speakers may alternate between inflected and uninflected
forms:
(3) AAE: They taught me mighty good, they teach me good.
(Holm 1991:235)
Schneider (1989:81) found that of some 8,000 verbs in a past-tense con-
text in the ex-slave narratives (recordings made between 1935 and 1974
of AAE speakers born before 1861; see section 2.1.5), 75 percent were
morphologically marked, indicating that the category of past tense is part
of the grammar of AAE, although actual marking of verbs is optional.
3.1.2 AAE auxiliaries/preverbal markers

Mufwene (1983) shows that the semantics of time reference in the AAE
verbal system bears out a kinship to the English-based creoles of the
Caribbean. However, the use of been as a creole-like preverbal marker of
anterior tense or remote past is relatively rare in AAE, even in the ex-slave
recordings, e.g.
(4) AAE: I got on a cowboy shirt now that I brought from Texas.
Been have it all my days. (Holm 1991:235)
Decreolizing varieties often replace anterior been with did, had, or was;
these are frequently less deviant from standard usage and thus less stig-
matized. AAE been plus verb can be made to conform more closely to
standard English morphology as an auxiliary in a progressive construc-
tion, despite the semantic mismatch:
(5) AAE: I hear jus’ as good now as I ever been hearing. (ibid.)
Some older AAE speakers occasionly use unstressed did to mark the
past:
(6) AAE: Let me see how that did come up. (ibid.)
Another remnant of preverbal been may be the AAE use of had in con-
structions that do not conform syntactically or semantically to non-AAE
usage:
(7) AAE: Today I had went to work. (Cukor-Avila 2001:105)
In Gullah and the Caribbean creoles the completive aspect marker done is
followed by the uninflected form of the verb, but in the ex-slave narratives
it is followed by the past participle:
(8) AAE: Bout eight o’clock he done been all around.
(Cukor-Avila 2001:238)
In decreolizing Caribbean varieties, progressive aspect is indicated not
by the basilectal preverbal marker de but by the English verbal suffix
-ing (without be as an auxiliary verb); this is the construction found
throughout the ex-slave recordings, e.g.
(9) AAE: They all going home now. (Cukor-Avila 2001:236)
Although -ing is clearly an inflectional morpheme in English, its status
in decreolizing varieties is less unambiguous.
In a number of English-based creoles unstressed does marks habitual
aspect, a dialectal usage also found in England and Ireland. A parallel
construction is found in the ex-slave recordings:
(10) AAE: An’ I does enjoy certain of his show. (Cukor-Avila
2001:237)
In the Bahamas, this habitual preverbal marker does has the reduced
forms is and ’s:
(11) Bahamian CE: They is be in the ocean. (Holm with Shilling
1982:111)
The verb phrase 75
Rickford (1980) suggests that the complete loss of these reduced forms
left be itself with habitual force in some varieties:
(12) Bahamian CE: Sometimes you be lucky. (Holm 1988–89:160)
(13) Bahamian CE: They just be playing. (ibid.)
Invariant be has also taken on the force of a habitual marker in modern
AAE:
(14) AAE: Those boys be messing with me. (Cukor-Avila
2001:105–107)
This usage is not usually found in the speech of whites, and has become
much more widespread in AAE as a marker of ethnic identity since the
middle of the twentieth century (ibid.). The habitual meaning of this
construction can be emphasized with steady:
(15) AAE: Them brothers be rappin steady. (Baugh 1983:86)
Another apparent AAE innovation is the combination be done for the
future perfect (Cukor-Avila 2001:104–107):
(16) AAE: We be done washed all the cars by the time JoJo gets
back with the cigarettes. (Baugh 1983:78)
3.1.3 AAE negation

Some nonstandard features of AAE negation, such as most uses of ain’t,
are also found in other nonstandard varieties of British and American
English, and their origin is not connected to contact with creolized va-
rieties of English. However, AAE ain’t can also be used to negate verbs
understood to refer to past action, such as the following:
(17) AAE: He ain’t do it. ‘He didn’t do it.’ (Rickford 1999:8)
Kautzsch (2002:45) notes that AAE “preverbal ain’t appears to have lost
its potential to occur in the present tense and has been restricted to past
tense contexts.” This use of ain’t is not found in the speech of whites
in the American South who use ain’t in other contexts (Cukor-Avila
2001:105–107). Taken together, the AAE uses of ain’t seem related to
the use of ain’t in decreolizing varieties to replace the preverbal negator
no found in more basilectal varieties:
(18) Bahamian CE: Stone at sea bottom no know sun hot. (Holm
with Shilling 1982:143)
(19) Bahamian CE: Bookie ain’ know who do it yet. (Holm with
Shilling 1982:3)
This seems likely to have converged with the use of ain’t in other non-
standard varieties corresponding to standard haven’t in the present per-
fect tense (with partially parallel past reference). Similarly, AAE multi-
ple negation or negative concord (negating not only the auxiliary verb
but also all the indefinite pronouns in the sentence) is also found in
white speech (Cukor-Avila 2001:105–107). However, AAE and creole
English can extend negation to noun phrases as definite as proper
nouns:
(20) AAE: We don’ want no six-month investigation! (AAE speaker,
Euronews, 11/7/02)
(21) Bahamian CE: They can’t sell that in no Haiti (Holm with
Shilling 1982:143)
For emphasis, AAE can invert the negative auxiliary and the indefinite
subject:
(22) AAE: Don’t nobody like him. ‘Nobody likes him.’ (Sells,
Rickford, and Wasow 1996)
Negative concord can also be transferred across clauses:
(23) AAE: It ain’t no cat can’t get in no coop. ‘There isn’t any cat
that can get into any coop.’ (Labov 1972a:130)
Howe and Walker (2000:110) point out that such clause-external con-
cord (like negative inversion) is also found in non-AAE varieties of non-
standard English but is apparently not documented in any English-based
creole. Sentence (23) is particularly difficult for monodialectal speakers
of standard English to parse because it also contains a zero subject rela-
tive pronoun (section 5.1.2) and AAE it’s ‘there is’, also found in SWVE.
This construction seems likely to be linked to Bahamian CE it have idem.
via the use of it’s for both ‘it is’ and ‘it has’ (Holm 2000a:200). There are
parallels in Bantu languages as well as creoles based on French, Span-
ish, and Dutch; among partially restructured varieties, BVP uses tem
‘[it] has’ (versus EP há) and NSCS uses tiene ‘[it] has’ (versus S hay)
(ibid.).
3.1.4 AAE non-verbal predicates

Non-verbal AAE predicates have received particular attention in the lit-
erature. Labov (1969) did a quantified study of the absence of forms of
The verb phrase 77
be in certain phonological and syntactic environments, which he related

to social variables. Holm (1984) related the AVE patterns to those in
Atlantic creoles and the African languages that influenced them to trace
the role of restructuring in AAE’s genesis and development.
Expressed forms of the copula are normally required in standard
English and all British dialects. Poplack (2000:20) concedes that “zero
copula is perhaps the only variant studied in this volume which cannot
be identified as a legacy of English.” Walker (2000:67) implies he has
counter evidence regarding this point: “Regardless of the lack of histor-
ical examples, zero copula does exist in other nonstandard varieties of
English, in locales such as Alabama (Feagin 1979), Mississippi (Wolfram
1974), and Yorkshire (Tagliamonte, p.c.).” Aside from dealing with the
likelihood that zero copulas came into the non-AAE varieties of Alabama
and Mississippi through contact with AAE, Walker needs to provide more
precise information than reference to a personal communication, since
many varieties of informal English can omit copulas when they are in-
verted auxiliaries (e.g. “You going?”).
There is massive variation of the AAE zero form of the copula with the
expressed forms; sometimes this variation occurs almost within the same
sentence:
(24) AAE: They all dead. All of them’s dead.
Where they at . . . where they is.
The Yankee be to the landing, they drunk.
(Holm 1991:239)
The AAE pattern has parallels in Gullah, Jamaican, and ultimately
Yoruba (ibid.), providing evidence that AAE resulted from the partial
restructuring of English under the influence of similar creole and African
languages.
3.2 The Afrikaans verb phrase
3.2.1 Afrikaans verbal morphology

Unlike Dutch, Afrikaans verbs have no inflections to indicate person or
number, as can be seen in this comparison of present tense forms:
(25) A: om te help ‘to help’ cf. D: helpen idem
ek help ons help ik help wij helpen ‘I/we help’
jy help julle help jij helpt jullie helpen ‘you help’
hy help hulle help hij helpt zij helpen ‘he helps/they
help’
The relative simplicity of Afrikaans morphology is especially noticeable

in its regularization of suppletive forms, such as in the verb meaning
‘to be’:
(26) A: om te wees ‘to be’ cf. D: zijn idem
ek is ons is ik ben wij zijn ‘I am/we are’
jy is julle is jij bent jullie zijn ‘you are’
hy is hulle is hij is zij zijn ‘He is/they are’
There has been a tendency towards morphological simplification in such
West Germanic languages as English and Dutch for at least the past half
millennium, and there are regional varieties of both languages in which
verbal morphology is even simpler than in the standard. However, the
loss of inflections is also characteristic of language contact, and there
can be little doubt that this tendency in Dutch was accelerated when
the language came into contact with other languages at the Cape.
Evidence for this can be found in the fact that while the base form of the
Afrikaans verb is usually derived from the ik form of the Dutch verb (e.g.
help above), this is not always the case. For example, some Afrikaans
verbs are based on the Dutch infinitive, e.g. A gaan ‘go’ (cf. D gaan
‘to go’ vs. D ik ga ‘I go’). Such a mixture of loss and retention of Dutch
inflections in Afrikaans is strongly suggestive of contact with languages
lacking such inflection (e.g. Khoi or Creole Portuguese); as in full cre-
oles, the most frequent superstrate forms were taken to be the base forms
regardless of the presence or absence of what had been inflections in the
source language.
The Dutch simple past tense, with its inflections and many irregular
forms, is – with very few exceptions – not found in Afrikaans, which uses
for the past a construction derived from the Dutch present perfect, i.e.
the auxiliary het ‘have’ plus a past participle beginning with the prefix
ge- (as in Dutch) but followed by a stem with almost none of the Dutch
irregularities, e.g. “Ek het geskryf ” ‘I wrote; I have written.’ Lockwood
(1965:210) notes that in Afrikaans
The past part[iciple] is formed from the present stem, so that ablaut no longer
plays any role. Thus the most typical feature of the Germanic verbal system,
the division into strong and weak classes, has vanished from Afrikaans. This has
occurred in no other Germanic language – Pidgin English excepted.
Afrikaans can indicate past action not by a morphological change in the

verb but by a temporal adverb (especially toe ‘then, at that time’) used
with the present tense form of the verb (except for the few verbs with
irregular past tense forms, e.g. was):
The verb phrase 79
(27) A: Ek het op die systraat gestaan. Toe kom die motorkar

vinnig om die hoek en ry vas teen die lamppaal. Daar was ‘n
harde geraas . . . ‘I was standing on the pavement. Then the
motor-car came rapidly round the corner and rode slap into
the lamppost. There was a loud noise.’ (Burgers
1963:118–119)
Such a use of the present-tense form of the verb for the past is unknown in
Dutch, but pidgins characteristically indicate tense with adverbials rather
than inflections.
3.2.2 Afrikaans auxiliaries/preverbal markers

Finally, a more highly restructured nonstandard variety, Orange River
Afrikaans (ORA, which has other non-European features also found in
the eighteenth-century Afrikaans of non-Europeans), sporadically uses
ga or ge as a preverbal marker of past tense, comparable in form and
meaning to both the substrate Nama marker gye or go and the Dutch
past participle prefix ge- (Roberge 1994:73–74).
Another creole-like preverbal marker in ORA is gedaan, indicating
completive aspect:
(28) ORA: Jij mijn Cameraat gedaan vast maken
you my comrade PERF-ANT fast make
‘You (have) tied up my comrade.’ (den Besten 1989:225,
quoting Franken 1952:50)
The ORA preverbal marker lê indicates durative aspect (Rademeyer
1938:78–79); it also occurs in the non-standard Kaaps variety of
Afrikaans (KA), e.g.
(29) KA: Hy lê wag daar. ‘He always waits there.’ (Makhudu
1984:88–80)
3.2.3 Afrikaans negation

Afrikaans has a postverbal negator (as does Dutch); if any other element
follows the verb, there is a second negator at the end of the sentence (not
found in Dutch):
(30) A: Sy eet nie. ‘She isn’t eating.’
she eat NEG (Combrink 1978:79)
(31) A: Sy eet nie pap nie. ‘She does not eat porridge.’
she eat NEG porridge NEG (ibid.)
Combrink (1978:84) notes that the second element of the double nega-
tion found in Afrikaans seems most likely to have been influenced by
the postverbal (and therefore sentence-final) negator in Khoi languages
like Nama. A study of written documents from the period 1830–44
(Nienaber 1965) revealed that among Coloured speakers of Afrikaans
whose ancestors had spoken Khoi, 40 percent of all cases of negation
had the second element, but among white speakers only 13 percent did.
Later, when Afrikaans was being standardized, its structural differences
from Dutch were emphasized to increase its Abstand, or distancing from
the other standard, and double negation became fully accepted in white
Afrikaans.
3.3 The BVP verb phrase
3.3.1 BVP verbal morphology

There is a drastic reduction of verbal inflection in BVP, which cannnot
be due solely to a phonological tendency towards the loss of final -s. The
intimate second person forms used in Portugal (e.g. the singular tu partes
and the obsolescent plural vós partis for ‘you leave’) have been replaced
by polite forms taking third person endings (singular você parte and plural
vocês partem) in most parts of Brazil, except in prayer. The feature dis-
tinguishing singular from plural in the third person is nasalization, which
often does not occur. The most coherent explanation for this variation is
partial restructuring.
Without the intimate second person forms, the six distinct verbal in-
flections for person in the present tense in European Portuguese (EP) are
reduced to four in Standard Brazilian Portuguese (SBP) – confining the
discussion to one conjugation and noting that there are parallel contrasts
in the other two:
(32) SBP: eu parto ‘I leave’

você/ele parte ‘you leave/he leaves’
nós partimos ‘we leave’
vocês/eles partem ‘you/we leave’
However, in BVP denasalization of the third person plural ending can

yield a three-way contrast, with distinctive endings only in the first person
singular and plural. This might be attributed to a phonological rule for
denasalization. In the data I collected there were a number of sentences
like the following:
The verb phrase 81
(33) BVP: Os alunos . . . que não conhece . . .

‘The students . . . who don’t know . . .’ [cf. SBP
conhecem /konyesẽ/]
However, the loss of contrasting inflections is unlikely to be due to a
purely phonological rule because in some lects of BVP the first person
plural ending is also nondistinctive (Marroquim 1934:115–116), an un-
ambiguous indication of restructuring:
(34) BVP: eu parto ‘I leave’
você/ele parte ‘you leave/he leaves’
nós parte ‘we leave’
vocês/eles parte ‘you/they leave’
A further confirmation of the morphological rather than phonological
nature of the rules that are needed to account for the loss of endings for
person on BVP verbs is found in the rural variety of Ceará, where forms
following even the first person singular pronoun can take what in SBP
is the third person singular ending, e.g. eu da ‘I give’ (cf. SBP eu dou)
or eu sabe ‘I know’ (cf. SBP eu sei) according to Jeroslow (1974:142,
171). Because of the reduced verbal paradigm, BVP makes greater use of
subject pronouns than EP, which is a Pro-drop language that uses them
for emphasis only. In BVP subject pronouns are required for all persons
except the first singular, since it usually maintains its own distinctive
verbal inflection (Mello 1997).
It should be noted that most Atlantic creole verbs, which take no in-
flections, appear to be derived from the imperative form of the verb in
the lexical donor language, rather than the infinitive (or possibly the third
person singular form, although this is could not be the case in the English-
based creoles). This can be seen more clearly in the case of many irregu-
lar verbs; for example, Guiné-Bissau CP bay and Papiamentu bai, both
meaning ‘go’, are from P vai ‘go!’ (or possibly [ele] vai ‘[he] goes’). It is
significant that this is also the form found in BVP, e.g. “Nós vai lá’ ‘We
go there’ (cf. SBP “Nós vamos lá”) (Rodrigues 1974:208).
This two-way inflectional contrast can also be found in the preterit:
(35) SBP BVP

eu parti parti ‘I left’
você/ele partiu partiu ‘you/he left’
nós partimos partiu ‘we left’
vocês/eles partiram partiu ‘you/they left’
In the imperfect tense, a single form (partia) can be used for all persons,
replacing the three-way contrast in SBP. In BVP other inflected tense
forms are rarely used, and the subjunctive mood tends to be replaced by
the indicative (Azevedo 1989:866–867).
The tense system of Helvécia Portuguese appears to have been quite
divergent from that of even rural varieties of BVP. For example, one
informant in Silveira Ferreira (1985) used é (a SBP present-tense form
for ‘is’) with what could only be past-time reference.
3.3.2 BVP auxiliaries/preverbal markers

The Portuguese-based creoles use preverbal markers to signal categories
of tense and aspect; there are some partially parallel markers in rural va-
rieties of BVP such as foi ‘PAST’ (literally ‘was’) and vivia ‘HABITUAL
PAST’ (literally ‘lived’):
(36) BVP: Eli foi dis . . . ‘He PAST said. . . .’ (McKinney 1982:6)
(37) BVP: Eli vivia trabayava ‘He HABITUAL-PAST worked’
(McKinney 1982:7)
3.3.3 BVP negation

Negation of the BVP verb can be handled three ways (Schwegler 1996b),
as in the following sentences, each of which means ‘He doesn’t know’:
(38) l. Before the VP: BVP: Ele não sabe.
2. Before and after the VP: Ele não sabe não.
3. After the VP: Ele sabe não.
In pattern 2, utterance-final não alternates with nada ‘nothing’:
(39) BVP: El’ não falô issu nada. ‘He didn’t say that.’
Older Portuguese had only pattern l, while modern EP has 1 and 2,
the latter for emphasis only. Schwegler concludes that in BVP there is a
change in progress towards pattern 3. One possible external factor may
have been contact with creoles having patterns 2 and 3. São Tomense
Creole Portuguese has pattern 2:
(40) São Tomense CP: I’nẽ na ka ‘tlaba na’i fa
they not ASPECT work here not
‘They do not work here.’ (Ferraz 1976:36)
while the closely related creole of nearby Prı́ncipe has pattern 3 with
only utterance-final fa. Discontinuous double negation is also found in
some African languages such as Ewe, which surrounds the verb with the
disjunctive negators me . . . o, the first element of which can sometimes
be omitted (Boretzky 1983:102).
The verb phrase 83
3.3.4 BVP non-verbal predicates

Some rural varieties of BVP have non-verbal predicates consisting of noun
phrases, adjective phrases, and locative phrases, all without copulas:
(41) BVP: Eu mininu. ‘I [was] a child’
Ela loka pur eli. ‘She [is] crazy about him’
i eli ali ‘and he [was] there’ (McKinney
1975:15)
These constructions are parallel to “zero copula” structures in AAE,
(section 3.1.4), the Atlantic creoles, and their substrate African languages
(Holm 1988–89:175–178).
3.4 The NSCS verb phrase
3.4.1 NSCS verbal morphology

While scholars of NSCS disinclined to accept any external influence on
its structure have long resorted to postulating the internal development
of phonological rules to account for the loss of inflectional morphology
in the VP, the similar but more pervasive restructuring of some varieties
of rural BVP (see section 3.3 above) suggests that phonology alone does
not provide an adequate explanation for the reduced verbal inflections
in either BVP or NSCS, forcing the issue of morphological restructuring
resulting from contact:
(42) Spanish NSCS Portuguese BVP
I buy compro compro compro compro
you buy compras compra compras compra
s/he buys compra compra compra compra
we buy compramos compramo compramos compra⇐
you buy compran compra compram compra
they buy compran compra compram compra
Traditionally the loss of inflectional distinctions in NSCS was explained
as purely phonological: syllable-final /s/ weakens to /h/ or disappears al-
together, as in the Spanish of Andalusia (which, of course, was in contact
with Arabic for many centuries); and syllable-final /n/ weakens to vowel
nasalization or disappears altogether. However, there appear to be no
regional varieties of European Portuguese that have undergone parallel
phonological changes in any comparable way (pace Naro and Scherre
2000), and the BVP loss of an entire syllable of inflection (P compramos
becoming BVP compra) can only be accounted for by the same kind of
morphological restructuring which, on a more extensive scale, produced

the Portuguese-based creoles. Because of the sociolinguistic parallels in
the history of NSCS and BVP, it is reasonable to deduce by analogy
that it was contact-influenced morphological simplification rather than
only internally motivated phonological rules that led to the reduced in-
flectional distinctions of not only BVP but also NSCS. In fact, the same
phonological rules may well have been initiated by externally motivated
morphosyntactic rules, as discussed in section 3.3.1 above.
3.4.2 NSCS auxiliaries/preverbal markers

NSCS is less restructured than BVP and there are no unambiguous cases
of syntactically induced loss of inflections in most modern varieties of
NSCS as there are in BVP. However, some nineteenth-century texts from
Puerto Rico and Cuba reveal exactly this kind of morphological change.
Note that in the first clause of (43) below, the verb has the third per-
son singular ending quiere despite its first person subject, yo ‘I.’ In the
second clause, the form ta (from the Spanish auxiliary está, as in está
cantando ‘s/he is singing’) seems to function as a preverbal marker of
durative aspect, as in creoles based on Spanish and Portuguese around
the world.
(43) NSCS: y mientre ma te quiere yo . . . tú no ta queré a mı́
‘and the more I love you . . . you don’t love me’
cf. S: y mientras mas te quiero . . . tú no me quieres
(Álvarez Nazario [1974]:190)
This variety of Afro-Caribbean Spanish was like most Spanish and
Portuguese-based creoles in that it used the third person singular form
of the present tense as the general form of the verb. Green (1997:136ff.)
found some sporadic contemporary examples of this in the unusually
basilectal variety of Dominican Spanish she studied (see section 2.4.3):
(44) NSCS: . . . nosotroh iban . . . ‘we went’ (Green 1997:138)
(cf. S: nosotros ı́bamos)
Spanish is a Pro-drop language in which verbal inflections indicate per-
son, usually making subject pronouns unnecessary except for emphasis.
However, because NSCS has a reduced verbal paradigm like BVP, it re-
quires greater use of subject pronouns than other varieties of Spanish to
make clear which person is the subject.
3.4.3 NSCS negation

Like AAE, NSCS has extended the rules of multiple negation or nega-
tive concord from those of the standard (negating the verb and all the
The verb phrase 85
indefinite pronouns in the sentence) to negating nouns as definite as a

proper noun:
(45) NSCS: José no fue a ningún Boston. ‘José didn’t go to

Boston.’ (Puerto Rican informant [February 1988], cf. S: José
no fue a Boston.
The verb in NSCS can take double negation (Schwegler 1996b), as in

(46):
(46) NSCS: Pero yo no me acueldo na deso no.

‘But I don’t remember anything about that.’ (Ortiz 1996:200)
This construction occurs in the vernacular Spanish spoken not only in

the Greater Antilles but also in coastal Colombia and Venezuela – but
not in European dialects of Spanish. There is a parallel construction
in Palenquero, the creolized Spanish of northern Colombia (Schwegler
1991; Dieck 2000):
(47) Palenquero CS: Nu abla ma nu.

NEG speak more NEG
‘Don’t say any more.’ (Bickerton and
Escalante 1970:259)
3.4.4 NSCS non-verbal predicates

Non-verbal predicates are rare in NSCS, but do occur sporadically. Green
(1997:183ff.) found zero copulas in the unusually basilectal variety of
Dominican Spanish that she studied:
(48) NSCS: Yo Bido Dobe. (Green 1997:185)

1s [am] Bido Doble
‘I am a Bido Doble.’ [family names]
Zero copulas are also documented for the NSCS of Venezuela (Alvarez
1990:125).
3.5 The VLRF verb phrase
3.5.1 VLRF verbal morphology

The verbal systems of the various vernacular lects of Réunion French
form a continuum, with uninflected single-form verbs with preverbal
auxiliaries (or possibly tense-mood-aspect markers more similar to those

in the verbal systems of the fully creolized languages) usually occur-
ring more frequently at the basilectal end of the continuum. Appar-
ently inflected verbs and auxiliaries analogous to forms in the verbal
system of European varieties of French (possibly reanalyzed as single
morphemes) are more frequent at the acrolectal end of the VLRF con-
tinuum – although a single individual may use both of the alternative
forms. Thus Chaudenson (2000:362) contrasts basilectal VLRF moin te
ki dans ‘I was dancing’ with acrolectal mi dansé idem. (cf. F je dansais
[dãse]).
Some of the apparently conjugated VLRF verb forms can be seen in the
following table from Baker and Corne 1982:27 (using their orthographic
system, but omitting some allomorphs):
(49) Selected VLRF verb forms
eat serve put say come know be
Present tense mâz serv me(t) di viê kone le
Past tense mâze serve mete dize vne konese lete
Future (negative) mâzra servira metra dira viê(n)ra konetra s(o)ra
Past participle mâze servi met/mi di vni koni (e)te
Infinitive mâze servir met(r) dir vnir konet(r) et(r)
Despite the phonemic representation, the inflected French verb forms
from which these VLRF forms derive are usually unambiguous (e.g. le
from il est ‘he is’). However, the status of the above VLRF “inflections”
as distinct morphemes is dubious, given sentences like VLRF i fémalra
pa (cf. F Ça ne fera pas mal) ‘That won’t hurt’ (Cellier 1985a:86). In
fact, VLRF rapa seems to function as an unbound postverbal marker of
(negative) future tense (Chapuis forthcoming).
Corne asserts that
The morphosyntax of R[éunion] C[reole] derives, as will be readily apparent,

from (mainly seventeenth-century varieties of) French . . . the semantics of the
R[éunion] C[reole] verbal system are also basically “French” (Baker and Corne
1982:13, 29)
However, Ramassamy’s 1985 dissertation was not yet available when

Corne made this assertion. Her study, based on the speech of mono-
lingual speakers of the most basilectal variety of VLRF, the Créole des
Bas, demonstrates the profound (if subtle) influence on the VLRF of
Malagasy, its earliest and most important substrate language. Chapuis
(forthcoming) notes how this study reveals that the VLRF (like Mala-
gasy, but unlike French) distinguish processive predicates (i.e. with a
The verb phrase 87
verb that is semantically a process, taking a subject that is the agent of

the action), by preceding such verbs with the agentive subject marker
(ASM) i:
(50) VLRF: Zan i vyen manzé aswar.

Jean ASM come eat tonight
‘Jean is coming to dinner tonight.’ (Ramassamy
1985:146)
Such processive predicates are opposed to non-processive predicates (se-
mantically not a process but a quality, situation, etc., taking a non-
agentive subject), which are preceded by the copula-like particle (CLP)
le (cf. F il est ‘he is’) with various forms marked for past (te) or future
tense (sra).
(51) VLRF: Mon rob lé rouz.

Is dress CLP red
‘My dress is red.’ (Ramassamy 1985:154)
The agentive subject marker i also has various forms conveying different
tenses and aspects, but these are distinct from the corresponding forms
of le. Although the distinction between the above two predicates is easy
to understand, corresponding to the distinction between a (non-stative)
verb and a copula in French and other Western Indo-European languages,
other such distinctions are more subtle:
(52) VLRF: La kaz i ékléré.
DET house ASM light up
‘The house was lighting up.’ (Ramassamy 1985:68)
(53) VLRF: La kaz lé ékléré.
DET house CLP light up
‘The house was lit up.’ (Ramassamy 1985:69)
Chapuis (forthcoming) points out that VLRF verbs are inherently stative,
and the agentive subject marker i destativizes them. For other interpre-
tations of VLRF i, see Béraha (2002), Caid (2002), Cellier (1985a), and
Watbled (forthcoming).
3.5.2 VLRF auxiliaries/preverbal markers

There is no doubt that the VLRF are characterized by a considerable
amount of restructuring. Forms of auxiliaries (e.g. avuar ‘have’ and etr
‘be’ above) combine with the verb to form tenses such as the following
(Baker and Corne 1982:26):
(54) Present m i mâz ‘I eat’

Past muê lete ki mâze ‘I ate’
m i mâze ””
Future: definite m i sava mâze ‘I’m going to eat’
indefinite m ava mâze ‘I’ll eat’
negative m i mâzra pa ‘I won’t eat’

Future in past muê te i sa mâze
Conditional muê nore mâze ‘I was going to eat’

Perfect muê la mâze
‘I have eaten’
muê la fin mâze

Pluperfect muê lave mâze
‘I had eaten’
muê te fin mâze
Future Perfect mi sra fin mâze ‘I’ll have eaten’
It should be noted that the English glosses in (54) were provided not
by Baker and Corne (1982:26) but rather by the present author with the
help of J.-Ph. Watbled, who analyzes the VLRF verbal system somewhat
differently (Watbled 2002). This system includes periphrastic construc-
tions expressing progressive aspect in various tenses (muê l’apre mâze ‘I
am eating’), a recent past (mi i sort mâze ‘I have just eaten’), and an
imminent future (muê le pur mâze ‘I am about to eat’).
But it is not just the pronominal form, the agentive subject marker i,
and the lack of agreement for person and number that make the above
structures seem un-French. There are two alternative perfect markers,
one which is from French (la < il a; lave < il avait) and the other (fin <
F finir ‘to finish’) which is French in form only, that are used semantically
and syntactically as in fully creolized varieties based on French. Moreover,
some of the semantic oppositions (e.g. definite versus indefinite future)
appear not to be French at all. (However, Watbled notes in a personal
communication that he believes that the VLRF do not, in fact, make this
last distinction, which is, to be sure, found in Mauritian Creole French.)
3.5.3 VLRF negation

To negate a verb, standard French places discontinuous negators (e.g.
ne . . . pas) around the first verbal element in the verb phrase, which can
be either the main verb (55) or an auxiliary (56):
(55) F: Je ne sais pas. ‘I don’t know.’
Is NEG know NEG
The verb phrase 89
(56) F: Ils n’ ont pas dormi. ‘They haven’t slept.’

3p NEG have NEG slept
VLRF maintain the pattern of colloquial metropolitan French, in which

the first negator can be omitted:
(57) VLRF: Mi koné plu. ‘I don’t know any more.’

I know NEG (Cellier 1985a:135)
(58) VLRF: Zot la pa vu. ‘They haven’t seen.’
They PERF NEG seen (ibid.)
Fully creolized varieties like Mauritian CF have developed an alternative

system of verbal negation, in which the single negator pa always precedes
the entire verb phrase. Comparing the three systems, it is striking how
close the syntax of VLRF remains to that of standard French:
(59) Standard French VLRF Mauritian CF

Je ne mange pas. Mi mâz pa. Mo pa manzé.
Je n’ai pas mangé. Muê la pa mâze. Mo pa ti manzé.
‘I don’t eat.’
‘I haven’t eaten.’
(Watbled, personal communication)
Cellier (1985a:136) notes that this similarity of the surface structure of

VLRF to that of French suggests the “. . . simplification du français plus
que . . . la création d’un nouveau système linguistique” (‘simplification
of French rather than the creation of a new linguistic system’).
3.5.4 VLRF non-verbal predicates

If forms of etr ‘be’ are considered auxiliary verbs in the structures above
(e.g. muê le pur mâze ‘I am about to eat’), then VLRF have verbal predi-
cates with a copula followed by an adjective:
(60) VLRF: muê le malad. ‘I am sick.’

muê lete malad. ‘I was sick.’ (Baker and Corne
1982:26)
This VLRF structure contrasts with the zero copula in the non-verbal
predicates of fully creolized varieties:
(61) Mauritian CF: Mo morisien. ‘I am Mauritian’ (Cellier

1985a:136)
However, if the above forms derived from F être simply indicate stativity
(or non-processiveness) and tense, then malad in sentence (60) above
could be considered a verb meaning ‘to be sick,’ specified by a prever-
bal marker. This interpretation is supported by the fact that Malagasy,
the principal substrate language, has no copula, although it does have a
topicalizer that can occur between the subject and a nominal predicate
(Keenan 1978:301). Zero copulas can also occur in VLRF non-verbal
predicates that consist of an NP:
(62) VLRF: Sa ê bug fite. ‘That is one cunning fellow.’ (Baker
and Corne 1982:30)
Cellier (1985a:85) notes that VLRF sa (cf. F ça ‘that’) varies with VLFR
se (cf. F c’est ‘that is’) as a “présentatif,” or highlighter, before NPs, as
above. However, there are non-verbal predicates not only after sa but also
after NPs:
(63) VLRF: Lé chipèk, kom inn papang. (Cellier 1985a:73)
“Grasshoppers” are the same as a “buzzard” [referring
to an unpleasant woman].
3.6 A comparison of the Verb Phrase

While no claim is being made that the newer language varieties examined
above are any more closely related genetically or typologically than the
older languages out of which they developed, it is being claimed that they
underwent a similar kind of partial restructuring triggered by partly simi-
lar social and linguistic factors. For this reason, the structural differences
between the verb phrases of the newer and older languages are compared
with one another below in an effort to identify any common patterns of
restructuring.
3.6.1 A comparison of verbal morphology

The five varieties examined above all have highly reduced inflectional
morphology to indicate personal endings on verbs in comparison to
the European languages from which they evolved: AAE, Afrikaans, and
VLRF need no such endings at all (although AAE can have -s for all
persons) and rural varieties of BVP can have only one distinctive ending
for the first person singular, whereas NSCS has distinctive endings in
both the singular and plural of the first person. Indication that this mor-
phological simplication was due to language contact as well as a general
tendency in the European language towards the loss of inflection can be
The verb phrase 91
seen in the fact that sometimes inflected forms in the source language
were selected as base forms in the newer variety (e.g. AAE is, Afrikaans
gaan, BVP vai, NSCS ta or VLRF le).
3.6.2 A comparison of auxiliaries/preverbal markers

Although past-tense marking is categorical in BVP, NSCS, and VLRF, it
is optional in both AAE and Afrikaans. This optionality would appear to
be due to the influence of earlier varieties that were more highly restruc-
tured in which tense was indicated not by inflections but by adverbials or
preverbal markers. Remnants of the latter can be detected in modern con-
servative varieties (AAE been, ORA ge, rural BVP foi, basilectal NSCS
ta, and VLRF lete) along with preverbal aspect markers that appear to
be the remnants of an earlier, more creole-like, verbal system.
3.6.3 A comparison of negation

Superstrate influence can be seen in multiple negation in AAE (as well
as BVP and NSCS), but other aspects of negation suggest the influence
of language contact. The coincidental similarity of different substrates
seems to have led to discontinous double negation in Afrikaans and in
BVP and NSCS, as in some of the creole languages which influenced
them.
3.6.4 A comparison of non-verbal predicates

Although expressed copulas have apparently always been categorical in
Afrikaans, such copulas vary with non-verbal predicates in AAE, some
varieties of BVP and NSCS, as well as VLRF. Non-verbal predicates seem
likely to be due to the influence of partially overlapping sets of substrate
languages, some of which are not shared with Afrikaans.
4 The noun phrase
Introduction
None of the inflectional morphology of the noun phrase in the European
source languages is preserved – at least functioning as such – in fully
creolized languages. This inflectional morphology can be relatively com-
plex. In the English NP, nouns are marked for plural number; while they
take no inflectional marking for grammatical gender, they do take a pos-
sessive inflection. Dutch nouns, in addition to taking plural inflections,
co-occur with articles marked for gender as well as number, and these
and other modifiers agree with the head; however, Dutch possessive in-
flections have largely given way to periphrastic constructions. In Spanish
and Portuguese modifiers such as determiners and adjectives must be
marked by inflections to agree with the head of the NP in gender (usually
-o for masculine and -a for feminine) and number (zero for singular, -s
for plural). While number agreement is audible in spoken Spanish and
Portuguese, it is often not audible in spoken French, in which the plural-
izing inflection (usually -s) is often silent. Still the covert system of plural
marking of French, in which nouns co-occur with articles and other mod-
ifiers marked for plurality, is nearly as salient as the overt system of plural
marking in Spanish and Portuguese. Furthermore, personal pronouns in
all five European languages must agree with the noun to which they refer
in number and gender – the latter not only in the third person singular,
but also in all persons of the plural in Spanish. These pronouns are also
usually marked for case to indicate their function as subjects, objects, and
possessives.
All of these inflections are considerably reduced in the partially re-
structured varieties, except for number marking in standard Afrikaans
and gender marking in BVP and NSCS. Moreover, new constructions
with free morphemes have arisen to indicate possession and grammatical
categories unknown in the European source languages, such as associative
plurals.
92
The noun phrase 93
4.1 The AAE noun phrase

AAE nouns are variably marked for number and possession, although in
some varieties the possessive marker is categorically absent. AAE personal
pronouns usually mark case.
4.1.1 Number in the AAE noun phrase

As in many West African languages, Caribbean English Creole nouns
are not inflected for number, although when relevant plurality can be
indicated by juxtaposing a noun with a morpheme that is homophonous
with the pronoun meaning ‘they’:
(64) Yoruba: àwo.n o.kùnrin [literally ‘they men’] i.e. ‘the men’
(Rowlands 1969:195–197)
(65) Jamaican CE: dem bwai / de bwai-dem ‘the boys’
This plural marking usually implies definiteness and is confined to ani-
mate nouns. Of course the form is related to the British and American
dialectal demonstrative “them boys,” but creoles frequently derived their
definite articles not from those of their lexical source languages (which
tend not to receive emphasis) but rather from the latter’s demonstratives
(Holm 1988–89:191). Some parallel constructions can be found in the
language of the ex-slaves:
(66) AAE: them wagon (Holm 1991:240)
However, the English -s inflection also occurs frequently:
(67) AAE: two looms (ibid.)
In quantitative studies of contemporary AAE, the -s pluralizer is nearly
always present (Wolfram 1969:143). Rickford (1999:7) notes that its ab-
sence is “much less frequent” than the absence of the homophonous
verbal or possessive inflection. However, in earlier varieties of AAE such
as the ex-slave narratives, there is so much variation that it is not always
clear that the -s morpheme is anything more than a stylistic variant:
(68) AAE: had hounds . . . them hound . . . . . six mens . . .
six mans . . . six men (Holm 1991:240)
Poplack et al. (2000:100) claim that this grammatical approachment of
AAE to standard English is not evidence of decreolization since the va-
rieties of English out of which early AAE grew (which they claim were
not influenced by African or creole languages) also had zero marking for
some plurals (e.g. “two bushel ”). However, they are unable to offer any
evidence that the rate of zero plural marking in these varieties was in any
way comparable to that in early AAE. Thus the lower rate of plural -s
absence in contemporary AAE is indeed evidence of a leveling of features
between AAE and other American varieties, particularly in light of the
finding of Rickford (1999:273) that his recent studies of this feature in
Palo Alto show “no appreciable change” in the rates found by Wolfram
(1969) or Labov et al. (1968) that might support a growing divergence
between AAE and other varieties.
It has been suggested that such nonstandard forms as mans and chil-
drens resulted from decreolization, i.e. the acquisition of the English
pluralizing morpheme and its use with what were considered monomor-
phemic lexical items. Schneider (1989:161) rejects this hypothesis, citing
such British dialect forms as foots and feets as likelier sources; however,
the burden of proof would seem to be upon him to demonstrate that there
were British dialect models for all or even most such AAE forms, and that
their use was widespread among southern whites.
In addition to the simple plural, AAE and other varieties of English
have an associative plural after names of persons:
(69) AAE: Felicia an’ them done gone. (Mufwene 1998:73)
The construction, often pronounced /nεm/, here means ‘Felicia’s friends
or family or associates.’ Mufwene (ibid.) notes that AAE shares this
construction “with English creoles, rather than with other varieties of
English,” although he adds in a footnote that it is also used by whites
in the American South. It does indeed seem to be related to a syntacti-
cally and semantically parallel construction in a number of African and
Atlantic creole languages of various lexical bases:
(70) Yoruba: àwo.n Táı́wò [literally ‘they Taiwo’], i.e. ‘Taiwo and
his family, schoolmates or friends’ (Rowlands 1969:196)
(71) Miskito Coast CE: di sukya dem [literally ‘the medicine-
man they’], i.e. ‘the medicine man and his lot’ (Holm
1988–89:193)
Actually the AAE construction is also found in informal use in many
parts of the United States (DARE) and England (R. Hudson, personal
communication), although it may well have originated in Africa:
(72) Nonstandard E: Mary an’ them came over yesterday.
The tendency of emphasized and conjoined pronouns to take the object
case in informal English (whatever their function in the sentence) proba-
bly facilitated the borrowing of the AAE associative plural structure into
other varieties of English.
The noun phrase 95
4.1.2 Gender in AAE

For all practical purposes, standard English lacks grammatical gender;
even British bureaucrats are now nervous about calling a ship she, let
alone a country (Michael Pye, personal communication). For this reason,
gender agreement between elements in the verb phrase is as irrelevant
to AAE as it is to standard English. Dillard (1972:56) mentions AAE-
speaking children occasionally producing sentences like “He a nice little
girl,” but this is not a feature of adult AAE, although it can be found in
Gullah and other varieties of creole English.
4.1.3 Possession in the AAE noun phrase

The English-based creoles indicate possession by juxtaposition rather
than inflection:
(73) Miskito Coast CE: di uman biebi ‘the woman’s baby’ (Holm
1978:286)
This is also found in AAE:
(74) AAE: the white folk kitchen (Holm 1991:241)
However, Schneider (1989:162) found the possessive morpheme present

in over 90 percent of the 377 cases in which it was possible in the ex-
slave narratives. He notes that while there is variable use of the possessive
inflection in all northern urban varieties of AAE, there are southern vari-
eties in which the morpheme’s absence is categorical, suggesting that the
suffix has been gaining ground as a part of decreolization (1989:164).
Cukor-Avila (2001:106–107) lists the absence of both plural and posses-
sive -s as features found only in AAE, not found in the vernacular speech
of whites in the American South.
4.1.4 Pronouns in AAE

In the original pronominal system of most of the Caribbean creoles, it
would appear that no distinction was made for gender or case, and the
same form also served as a possessive determiner. Possible remnants of
such a system can be found in the language of the ex-slaves, although
these usages are less frequent in current AAE:
(75) AAE: Well the master had promise’ to, to give we all forty dollars
a month in pay. (Holm 1991:241–242).
(76) AAE: We had we own lawyers. (ibid.)
The use of standard English object pronouns in subject position is also

encountered in the ex-slave narratives:
(77) AAE: When us all leaves dis old world. (Schneider 1989:66)
Schneider concludes that “there is some evidence that this grammati-
cal variable was more subject to some degree of creolization than oth-
ers” (1989:177). AAE’s second person plural pronoun, you all /yɔ/, is
likely to be an innovation. The pronoun of the same person and number
is derived from various African languages in Caribbean Creole English
(Holm 1988:203–204), probably motivated by the need for a form of you
that is unambiguously plural (cf. similar dialectal forms, e.g. yous, you
guys, you chaps, etc.). In fact Eastern Caribbean CE has all you; the
AAE form, generalized throughout Southern American English, may be
a calque on Twi mó nyina, literally ‘you all.’ In AAE (but seldom in the
speech of Southern whites) there is a corresponding possessive:
(78) AAE: It’s y’all ball. (Rickford 1999:7)
Like other varieties of English including CE, AAE has pleonastic subject
pronouns:
(79) AAE: That teacher, she yell at the kids. (Fasold and Wolfram
1970:81)
These pronouns seem likely to facilitate the parsing of AAE sentences
containing relative clauses without subject relative pronouns (see section
5.1.2). Their existence in CE may also have facilitated the adoption of
the postnominal dem pluralizer (cf. Cassidy and Le Page 1980:147).
Finally, Wolfram (forthcoming) notes that “The regularization of mine
to mines in ‘The book is mines’ is quite robust in most varieties of AAVE,
though it appears more typical of preadolescent speakers than older
speakers.”
4.2 The Afrikaans noun phrase

In comparison to the seventeeth-century Dutch dialects from which it
developed, Afrikaans reveals the effects of language contact through the
loss of inflectional morphology and the replacement of bound by un-
bound morphemes in the noun phrase. Although the distinction between
common gender (e.g. D de man ‘the man’) and neuter (D het huis
‘the house’) was lost in Afrikaans (die man, die huis), the distinction
between singular and plural survived, but a number of morphosyntactic
rules were lost or reinterpreted.
The noun phrase 97
4.2.1 Number in the Afrikaans noun phrase

Afrikaans generally marks nouns with a plural inflection, usually either -s
after certain sounds or -e elsewhere. However, like many earlier Dutch
dialects, Afrikaans uses -s to mark the plural on many nouns whose etyma
form the plural with -en (often pronounced as a schwa) in standard
Dutch, which also uses -s to mark other plurals. Like Dutch and other
Germanic languages, Afrikaans also has irregular plurals with a vowel
change:
(80) A: stad ‘city’ versus stede ‘cities’ (Donaldson 1993:69–75)
The final consonant clusters of earlier Dutch nouns often underwent
simplification:
(81) D: kast > A: kas ‘cupboard’ (ibid.)
However, the Afrikaans plural ending -e usually preserves the consonant
lost in the singular:
(82) D: kasten > A: kaste ‘cupboards’ (ibid.)
However, this is not always the case:
(83) D: kost ‘food’ > A: kos ‘food’, kosse ‘foods’ (ibid.)
This is comparable to AAE des’ ‘desk,’ which has also undergone conso-
nant cluster simplification (Rickford 1999:4), taking the plural form for
words ending in a sibilant: desses. In Afrikaans this phenomenon has led
to non-historical hypercorrect forms in nonstandard varieties:
(84) D: bus, bussen > A: bus, buste ‘bus, busses’ (Donaldson 1993:69–75)
(cf. D: buste ‘bust, bossom’)
The Afrikaans plural markers -s and -e can sometimes both be used on

the same noun:
(85) A: beddens ‘beds’ (cf. D: bedden, plural of bed) (ibid.)
The plural endings can also alternate in Afrikaans with different mean-
ings, e.g. man ‘man’ becomes manne in the sense of ‘units (e.g. in war,
sports)’ but mans in the sense of gender, e.g. mans en dames (cf. D
mannen for both) (Burgers 1963:54).
Moreover, a free morpheme unknown in standard Dutch is used
in Afrikaans to form an associative plural (section 4.1.1): hulle ‘they,
them’ (from dialectal Dutch hun-lui, literally ‘them-people,’ according
to Lockwood 1965:210). In informal standard Afrikaans hulle can be
used only after [+human] nouns like pa ‘dad’ and ma ‘mom’, e.g. pa-
hulle ‘Dad and the others’ (Makhudu 1984:64). However, in Kaaps, the
nonstandard Afrikaans of Coloureds, hulle can also be used with nouns
that are [−human] but [+animate], e.g. die bome- hulle ‘the trees’ (ibid.).
The word goed ‘goods’ can also be used as an associative plural marker
in nonstandard and Orange River Afrikaans, e.g. Ma goed ‘Mom and the
others.’ Webb (1989) notes that in Khoi the reduced form of the mascu-
line plural pronoun gu ‘they’ can be used after a noun to give it plural
meaning, e.g. khoe ‘man’ becomes khoegu ‘the men.’
4.2.2 Gender in the Afrikaans noun phrase

As mentioned above, the Dutch distinction between common gender
(e.g. D de man ‘the man’) and neuter (D het huis ‘the house’) was lost
in Afrikaans (die man, die huis). Because of this, the rules governing
the marking of these distinctions in Dutch became opaque in Afrikaans.
In Dutch, for example, attributive adjectives usually take an -e ending:
(86) D: de jonge man ‘the young man’
There is an exception to this rule before neuter singular nouns with cer-
tain determiners:
(87) D: een goed boek ‘a good book.’
This morphosyntactic rule became a phonological rule in Afrikaans: poly-
syllabic adjectives take an -e (aangename weer ‘pleasant weather’) but
monosyllabic adjectives (except those ending in d, f, g, s, and certain
others) do not (A ‘n sterk man ‘a strong man’). The problem is that the
phonological rule in Afrikaans has scores if not hundreds of exceptions;
Donaldson admits, “It is impossible to formulate water-tight rules for the
inflection of adjectives” (1993:170; see also Lass 1990).
4.2.3 Possession in the Afrikaans noun phrase

Possession is no longer normally indicated with inflectional morphemes
on common nouns in modern Dutch (cf. Vondels werken ‘Vondel’s
works’), bur rather with the preposition van ‘of ’ (de werken van Vondel).
The latter is a frequent structure in standard Afrikaans as well, which
also has another method of indicating possession with the particle se
after nouns:
(88) A: pa se hoed ‘father’s hat.’
This is apparently an extension of the function of colloquial Dutch z’n (cf.
D zijn ‘his’); z’n can be used only after masculine nouns (e.g. vader z’n
hoed idem), although there is no such gender restriction on the Afrikaans
The noun phrase 99
particle. Kotzé (personal communication) notes that there is a variant of

the Afrikaans structure with se which is not restricted to animates:
(89) A: Die stoel wat by die tafel staan se poot.
the chair which by the table stands POSS leg
‘the leg of the chair standing by the table’
Interestingly, the Afrikaans possessive marker se, like the possessive ’s
in (colloquial) English, marks the end of an NP and is not a nominal
inflection, as in my aunt who lives in New Jersey’s son, or (possibly more
salonfähig) the King of England’s crown.
4.2.4 Pronouns in Afrikaans

Like the European Portuguese clitic object pronouns that became free
morphemes in BVP (section 4.3.4), the unemphatic personal pronouns
of spoken and written Dutch were not preserved in Afrikaans, which has
only the forms corresponding to the Dutch emphatic pronouns – with the
sole exception of the set phrase “jy weet” ‘you know,’ in which the vowel of
the pronoun tends to be reduced to a schwa, like that of the corresponding
Dutch unemphatic pronoun – and, coincidentally, that of the English filler
phrase, “y’ know” (Donaldson 1993:126). Otherwise singular personal
pronouns in Afrikaans are quite similar to those of Dutch, marking not
only case distinctions but also gender distinctions in the third person
(rare in fully restructured Atlantic creoles). However, the plural personal
pronouns make fewer of the Dutch case distinctions for subjects, objects,
and possessives:
(90) Subject Object Possessive

Dutch: 1. wij ‘we’ ons ‘us’ ons/onze ‘our’
2. jullie ‘you’ jullie ‘you’ jullie ‘your’
3. zij ‘they’ hen/hun ‘them’ hun ‘their’
Afrikaans: 1. ons ‘we’ ons ‘us’ ons ‘our’
2. julle ‘you’ julle ‘you’ julle ‘your’
3. hulle ‘they’ hulle ‘them’ hulle ‘their’
The Afrikaans plural possessive determiners above can form possessive

pronouns when followed by s’n:
(91) A: ons s’n ‘ours’
Afrikaans can also form possessives through the use of se after the relative
pronoun wat (section 5.2.2) and the relative and interrogative pronoun
wie ‘who’:
(92) A: Wie se boek is dit? ‘Whose book is this?’ (Burgers 1963:101)

When no noun is being modified, this construction also serves as a pos-
sessive pronoun:
(93) A: Wie s’n is dit? ‘Whose is this?’ (ibid.)
Compare standard Dutch “Wiens boek is dit?” or “Van wie is dit boek?”
with colloquial Dutch “Wie z’n boek is dit?” (Donaldson 1984:72). The
Afrikaans possessive relative pronoun construction (“Die man wie se
boek ek geleen het . . .” ‘The man whose book I borrowed . . .’) is also
from spoken Dutch (“De man die z’n boek ik geleend heb . . .”) rather
than written Dutch (“De man wiens boek ik geleend heb . . .”).
However, there is also convincing evidence that coloquial Dutch was
not the only source of the se possessive construction. The converging in-
fluence of parallel particles in Khoi and Creole Portuguese is more clearly
seen in the use of se and s’n after personal pronouns – constructions un-
known in Dutch. In standard Afrikaans possessive pronouns are formed
with s’n after the pronouns u ‘you,’ ons ‘we,’ julle ‘you (plural),’ and
hulle ‘they’:
(94) A: Die boek is ons s’n. ‘The book is ours.’
Moreover, nonstandard varieties of Afrikaans use this particle to form
possessive determiners in an equally un-Dutch construction: Makhudu
(1984) notes the frequent occurence in Coloured Afrikaans of julle se
‘your (plural)’ (but not jy se ‘your [singular]’), while Griqua, or Orange
River Afrikaans, has parallel forms:
(95) ORA: hy se huis ‘his house’ (den Besten 1978:31)
In the Khoi language Nama, the genitive particle di also occurs after
pronouns as well as nouns:
(96) Nama: //êib di ómi [literally ‘he ’s house’], i.e. ‘his house’ (ibid.)
(The symbol // represents a voicless lateral click.) The Malayo-Portuguese
genitive marker sa is also used this way:
(97) Papia Kristang CP: eli-sa mai ‘his mother’ (Hancock l969:41,
1975:229)
Furthermore, a parallel marker is used the same way in Bazaar Malay
(Hancock 1975:229). Although the use of se after pronouns to form
possessive determiners in nonstandard Afrikaans could be viewed as an
extension of the rule for its use after nouns in standard Afrikaans (which
The noun phrase 101
could then be traced to colloquial Dutch), the case for its survival from the
ancestral languages of the Griqua and Cape Malay in their restructured
varieties of Afrikaans seems more convincing. Of course these sources
are by no means mutually exclusive, and the most likely scenario is that
their influence converged in the development of the modern Afrikaans
constructions with se (den Besten 1978:38).
In Afrikaans the reflexive pronouns are identical to the object pronouns:
(98) A: Die professor trek hom in sy studeerkamer terug ‘The
professor retreats into his study.’ (Brachin 1985:141)
However, -self can be added for emphasis (Donaldson 1993:290). In
modern Dutch, the third person singular and plural reflexive pronoun is
zich (Donaldson 1984:172), a borrowing from German that did not be-
come prevalent in Dutch until the eighteenth century (Brachin 1985:13),
well after the beginning of Afrikaans. Thus Brachin (1985:141) com-
ments on “the archaic nature of Afrikaans.”
4.3 The BVP noun phrase

The BVP noun phrase preserves the categorical gender agreement found
in SBP, but tends to mark plural number only on the first item in
an NP. Gender and number agreement are categorically absent in the
Portuguese-based creoles and variably absent in Helvécia Portuguese,
although the relevant data from the latter is very limited. BVP object pro-
nouns differ from those of SBP in their case marking and word order, in
which they resemble more the pronouns of the Atlantic creoles.
4.3.1 Number in the BVP noun phrase

SBP requires that all determiners, nouns, and adjectives in a noun phrase
be marked for plural number:
(99) SBP: os livros velhos ‘the old books’
However, BVP often indicates plurality by adding -s to only the first
element (usually a determiner), leaving the plural -s inflection optional
on following nouns and adjectives:
(100) BVP: um dos mais velho orixás ‘one of the most ancient
deities’ (Holm 1987:417)
SBP: um dos mais velhos orixás
However, it is not always the first element that is marked:
(101) BVP: o meus irmão ‘[the] my siblings’ (ibid.)

SBP: os meus irmãos
(102) BVP: todo os mais velho ‘all the most ancient [ones]’ (ibid.)
SBP: todos os mais velhos.
Regarding the tendency of BVP noun phrases to mark only the initial
element for number – Guy (1989) found that over 95 percent had such
marking – this pattern may, as he suggested, represent a survival of the
system of marking plurality at the beginning of noun phrases in many
Niger-Congo languages. An early variety of restructured Portuguese in
Brazil may have had an optional system of marking plurality comparable
to that of São Tomé CP, in which nẽ, the word for ‘they,’ is used before
nouns, e.g. nẽ mwala ‘the women.’ This is comparable to the parallel use
for the word for ‘they’ in various African and creole languages discussed in
section 4.1.1. Through decreolization, this plural marker could have been
replaced by a plural form of the definite article, os (which also functions as
the object pronoun ‘them’). This is suggested by the attestation of earlier
BVP forms such as osêle ‘they,’ apparently a combination of a pluralizing
os plus ele ‘he,’ instead of SBP eles (Mendonça 1933:67). Support for
this interpretation can be found in the parallel use of Cape Verdean CP
ũʃʃ (cf. the P plural indefinite article uns ‘some’) as a pluralizer:
(103) Cape Verdean CP: ũʃʃ rapaz´ı̃ ‘(some) boys’ (Almada 1961:92)
It is clear that the variable rule for -s is both phonological and mor-
phosyntactic. As a phonological rule it operates on (synchronically) single
morphemes:
(104) BVP: somo ‘[we] are’
SBP: somos
Guy (1989) found clearly monomorphemic instances such as BVP onibu
‘bus’ (SBP onibus) (personal communication). He goes on to point out
that
at the same time a variable syntactic rule of NP plural marking is required to
account for phrases such as as vez, os espanhol, as nação, because if they
resulted from simple S-deletion, they should be as veze, os espanhoi, as
naçõe
due to certain irregularities in the formation of some SBP plurals. Viewed

historically, the variable marking for number in modern BVP makes sense
only if the variety evolved from an uninflected variety which began bor-
rowing inflections from SBP at a stage when the latter’s system of number
agreement within noun phrases and between subjects and verbs was still
The noun phrase 103
opaque to speakers of BVP. The inflections were probably first applied

randomly (as in decreolizing English the boy go / the boys goes / the boy
goes / the boys go) in free variation. The syntactic rules of the more fre-
quent inflections (-s and nasalization) alternating with their absence in
turn led to BVP phonological rules for the same alternation that could
then be applied to single morphemes. Naro and Scherre (2000) have
claimed that the BVP phonological rules for the variability of -s and
the nasalization of vowels were inherited from European Portuguese, but
such variation is very unusual in Portugal.
It is not clear from the available data whether number agreement in HP
is absent or resembles that of BVP. Regarding NP number agreement in
the colloquial speech of educated Brazilians, Azevedo (1989:867) notes
that it is frequently that of BVP, e.g. “Prova uns pãozinho” ‘Try some
rolls’ (SBP pãezinhos).
4.3.2 Gender in the BVP noun phrase

Gender marking appears to be categorical in both SBP and BVP. How-
ever, it is not a part of the grammar of the Atlantic creoles or most of
their Niger-Congo substrate languages (Holm 1988–89:195). There is
evidence that in HP there is no gender agreement between nouns (or
pronouns) and adjectives:
(105) HP: ’Ela E ’mu˜itu sa’idu ‘she is very meddlesome.’ Silveira

Ferreira (1985:30–31)
SBP: ela é muito saida
Silveira Ferreira (ibid.) notes that there is an absence of gender agree-

ment between nouns and articles in HP as well, and that nouns do not
necessarily take the gender marking of SBP:
(106) HP: ‘ũa a’bota ‘an abortion’ (ibid.)

SBP: um abôrto
The variability of gender agreement in HP is the topic of a dissertation

by Lucchesi (2000), who notes that such variation is disappearing among
younger speakers. He points out that its origin can only be explained by
what he calls the “irregular transmission” of rules of morphology, since
there are no known instances of the phonological variation of /o/ and /a/,
marking masculine and feminine forms, respectively, in the history of the
Portuguese language.
Bonvini (2000:402) notes that in the ritual “Lı́ngua dos Pretos Velhos”
gender agreement within the NP is variable:
(107) LVP: mia povu ‘my people’ (cf. SBP: meu povo) (ibid.)
(108) LVP: tera foi kiadu ‘the earth was created’ (cf. SBP: a terra foi
criada) (ibid.)
Finally, Amaral (1928 [1976]:73) notes that in the language of ex-slaves,
the pronoun ele (SBP ‘he’) could also be used in reference to females,
like the third person singular pronoun in basilectal Atlantic creoles and
in most of the Niger-Congo languages that formed their substrate.
4.3.3 Possession in the BVP noun phrase

Possession is normally indicated by the preposition de in BVP, as in
standard Portuguese:
(109) SBP: A casa de Maria ‘Maria’s house’
However, in some rural varieties of BVP the preposition can be omitted:
(110) BVP: kaza Maria [literally ‘house Maria’] (Jeroslow 1975)
There is a parallel possessive construction in São Tomé Creole Por-
tuguese, in which the preposition di can also be omitted:
(111) São Tomé CP: donu di losa
or donu losa ‘[the] owner [of the] farm’
(Ivens Ferraz 1979:69)
4.3.4 Pronouns in BVP

One of the most striking features of BVP is the use of personal pronoun
forms that can be used only for emphatic subjects in EP as direct objects
in BVP, replacing the clitics of the standard:
(112) BVP: Ela chamou eu. ‘She called me’ (Azevedo l989:863)
SBP: Ela chamou-me.
This usage is also frequently found in the casual speech of educated
Brazilians; Azevedo recorded a linguist saying the following:
(113) BVP: . . . impediu eles de passar ‘prevented them from
passing’ (Azevedo 1989:864)
SBP: . . . impediu-os de passar
In many Atlantic creoles there is no distinctive case marking for subject
and object pronouns. Moreover, these creoles always preserve their basic
SVO word order with object pronouns, unlike Romance languages, in
The noun phrase 105
which direct and indirect object pronouns usually occur before the verb.
BVP preserves this word order as well:
(114) BVP: Esses porco aı́, nós ganhemo eles. ‘Those pigs, we got
them as a gift’ (ibid.)
SBP: Esses porcos, nós os ganhamos
Turning from personal to reflexive pronouns, the latter are an integral

part of SBP grammar:
(115) SBP: João cortou-se com faca. (Mello 1997:153)

‘John cut himself with a knife.’
However, these do not commonly occur in BVP; instead, the following

constructions occur:
(116) BVP: a. João cortou com faca.

b. João cortou ele com faca.
c. João cortou ele mesmo com faca. (ibid.)
The first BVP structure without any pronoun is found throughout the
Atlantic creoles and many of their substrate languages, in which any tran-
sitive verb can have not only an active meaning but also a passive one if
its subject is a plausible object:
(117) Papiamentu CS: E yama Maria. ‘She is called Mary.’

(Holm 1988:83)
cf. S: Ella se llama Marı́a.
(118) São Tomé CP: E ple’de.. ‘He got lost.’ (Ivens Ferraz 1979:72)
cf. EP: Ele se perdeu.
(119) Bambara: Tò dun Mali la. ‘Millet porridge is eaten
in Mali.’
millet-porridge eat Mali in
(Holm 1988:84)
4.4 The NSCS noun phrase

Plural marking in the noun phrase is much reduced in NSCS when com-
pared to the standard, but in most modern varieties this appears to be
through phonological rather than morphological rules – although ear-
lier contact could well have played a role in the development of these
phonological rules.
4.4.1 Number in the NSCS noun phrase

The tendency of NSCS to delete word-final -s, which marks the plural
forms of nouns, as well as their modifiers such as articles and adjectives
in the standard, appears to be the result of a phonotactic tendency in
NSCS to simplify a CVC syllabic structure to CV. At least this has been
the prevailing interpretation in Spanish dialect studies. This tendency
does not always do away with plural marking, which can be indicated by
a weakening of /s/ to /h/, especially in feminine forms:
(120) NSCS: lah niña ‘the girls’ (Lorenzino in Holm, Lorenzino and
Mello 2000:203)
S: las niñas
However, the total loss of /s/ throughout the NP is more frequent when
the noun is masculine because the NSCS plural masculine article lo [cf.
S los] still contrasts with the singular masculine form el:
(121) NSCS: lo hombre ‘the men’ [cf. S: los hombres] (ibid.)
A distinct plural form is also retained in masculine nouns ending in a con-
sonant, such as dictador ‘dictator,’ which in standard Spanish require
an e before the plural -s:
(122) NSCS: lo dictadore esto ‘those dictators’
S: los dictadores estos (ibid.)
In a more highly restructured variety of NSCS spoken in the Chocó region
of Colombia, the occurrence of the plural form lo dictador without the
-e of the allomorph of the Spanish plural morpheme (Ruı́z Garcı́a 2001)
is as clear evidence of morphological restructuring (as opposed to mere
phonological rules) as BVP plurals forms like as vez instead of as
veze (cf. SBP as vezes ‘the times’) (Guy, personal communication;
see section 4.3.1).
Studies by Poplack (1978), Terrel (1979), and Nuñez Cedeño (1980)
analyze such plural marking in the noun phrase in the regional Spanish
of Puerto Rico, Cuba, and the Dominican Republic, respectively, which
in each case is more pervasive the lower the socioeconomic class of the
speaker.
More recently, Green (1997:190ff.) analyzed the use of -se as a plural
marker in an unusually basilectal variety of Dominican Spanish (see sec-
tion 2.4.3). The weakening and loss of final /s/ in plural nouns ending in a
stressed syllable appears to have motivated an analogical reinterpretation
of -se as a plural marker in these vernacular Dominican varieties:
(123) NSCS: lo cafése ‘cafés’ (cf. S: los cafés)
The noun phrase 107
by analogy with NSCS “lo mese ” (cf. S “los meses”) (Lorenzino in

Holm, Lorenzino, and Mello 1999:46).
Although the extent of the use of pluralizing -se is unclear in other va-
rieties of Antillean Spanish in Cuba and Puerto Rico, Green (1997:190–
196) indicates that in the Dominican varieties she studied, its use has
spread to nouns not stressed on the last syllable:
(124) NSCS: mucháchose ‘boys’; látase ‘tin cans’ (ibid.)
4.4.2 Gender in the NSCS noun phrase

Nonstandard Spanish throughout the Caribbean retains gender agree-
ment in the noun phrase, which is not found in creoles based on Spanish
and other Romance languages. These creoles usually derive nouns and
adjectives from the masculine form in the superstrate:
(125) Palenquero CS: cabeza malo ‘bad head’ (cf. S: cabeza mala)
(Lorenzino 1993:118).
(126) Papiamentu CS: baka gordo ‘fat cow’ (cf. S: vaca gorda) (ibid.)
However, Schwegler and Morton (2003) note the sporadic lack of gender
agreement in the Spanish spoken by bilingual Palenqueros:
(127) PS: cosa importantı́simo ‘important things’ (ibid.)

cf. S: cosas importantı́simas
4.4.3 Possession in the NSCS noun phrase

Possession in NSCS is formed with the preposition de as in standard
Spanish, and no other structure (such as that in BVP discussed in section
4.3.3) occurs.
4.4.4 Pronouns in NSCS

NSCS personal pronouns show no reduction of case or gender mark-
ing. However, in the NSCS of bilingual speakers of Palenquero (section
1.5), Schwegler and Morton note that reflexive pronouns can be omitted
(128d), as they are in Palenquero and other Atlantic creoles based on
Romance languages (cf. section 4.3.3), or take the subject form of per-
sonal pronoun plus mimo (cf. S mismo as in the emphatic construction
yo mismo ‘I myself.’ Of the structures below, they note that (128a) and
(128b) are common, while (128c) and (128d) are sporadic:
(128) PS: a. Yo me lavo. ‘I wash myself ’ (ibid.)

b. Yo me lavo yo mimo.
c. Yo lavo yo mimo.
d. Yo lavo. (2003: 118)
In the Dominican Republic, the structure in (126d) was used by one of
the informants whose speech was rejected as abnormal by Lipski (Green
1997:207). However, Lipski (1998) notes the absence of reflexive pro-
nouns in earlier bozal Spanish.
4.5 The VLRF noun phrase
4.5.1 Number in the VLRF noun phrase

In VLRF nouns and their modifiers have “no rules of gender or num-
ber agreement” (Cellier 1985a:36). For example, the following could be
interpreted as either singular or plural:
(129) VLRF: marmay té i sort lékol (ibid.)
child PAST ASM leave school
‘The child was/the children were leaving the school.’
Above VLRF marmay comes from a collective noun in French which can
only be singular (la marmaille ‘troupe de petits enfants’; cf. marmot ‘petit
garçon’ [familiar]), much like E cousinage. The F word marmaille cannot
refer to an individual, but VLRF marmay can:
(130) VLRF: . . . lo marmay evidaman li koné pa tousalà . . .

(C. Warnecke, personal communication)
‘. . . the child obviously doesn’t know all that . . .’
However, in VLRF if an indication of plurality is essential, there is a

free morpheme bann (cf. F bande ‘band, gang’) which can be placed
before the noun as a pluralizer, as in the fully creolized varieties of French
in the Indian Ocean:
(131) VLRF: bann marmay té i sort lékol. (ibid.)

‘The children were leaving the school.’
This unbound pluralizer can also form an associative plural, as in Bann
Payèt ‘the Payet family’ (Honoré 2002:37).
4.5.2 Gender in the VLRF noun phrase

Although articles in VLRF appear to match those of standard French
regarding gender (e.g. la kaz ‘the house’; cf. the feminine article in F
The noun phrase 109
la case ‘the hut’), there is no gender agreement between VLRF nouns

and their other modifiers such as possessive determiner and adjectives,
as there is in French:
(132) VLRF: mon kaz lé gran (Cellier 1985a:19)
my house is big
(Cf. the feminine forms in the equivalent F “ma maison est grande
[grãd]” versus the masculine forms mon ‘my’ and grand [grã] ‘big’.)
Cellier (1985a:19) notes that “this calls into question the existence of
lexical gender despite the apparent analogy with French [regarding ar-
ticles]; thus la is only a frozen noun indicator and kaz does not behave
like a feminine noun because there is no agreement between the adjective
and the noun.” While the VLRF certainly do not have the same kind of
gender agreement as French, the VLRF may indeed have true articles:
Chapuis (forthcoming) notes that Malagasy has definite articles that are
distinct from demonstrative adjectives (Keenan 1978:297) and suggests
that this helped to ensure the survival of definite articles in VLRF. Chau-
denson (1974:366) notes the use of some VLRF adjectives based on the
F feminine form (e.g. e grãd légliz ‘a large church’ (cf. F une grande
église), but these may be acrolectal forms influenced by the standard lan-
guage; they appear to be sporadic and most usual in set phrases such as
toponyms (e.g. grãd ravin; cf. F Grande Ravine).
From this perspective, the VLRF articles that seem to indicate number
and gender as in French may simply preserve the form of the French
article, as do the initial syllables of certain words in fully creolized varieties
of French:
(133) Seychellois CF: Fer letour lakaz. ‘Take a tour
of the house.’
cf. F: Faites le tour de la maison. (D’Offay and
Lionnet 1982:238)
The existence of analogous forms in VLRF such as lakaz ‘house’ (Chau-
denson 1974:349), however rare, suggests remnants of earlier stages of
more intense restructuring (e.g. the mid-eighteenth century). Nonethe-
less, the synchronic co-occurrence of VLRF in ‘a’ and lo ‘the’ (cf. F un ‘a
[masculine]’ and le ‘the [masculine singular]’ on the one hand, coupled
with the synchronic co-occurrence of VLFR inn ‘a’ and la ‘the’ (cf. F une
‘a’ [feminine] and la ‘the’ [feminine singular] on the other hand, make the
question of the existence of grammatical gender and number more com-
plicated to determine in VLRF than in fully restructured creoles like Sey-
chellois. Still, Ramassamy (1985:249) observes that although the VLRF
possessive pronouns (lé-myenn ‘mine,’ lé-tyenn ‘yours’ etc.) are derived
from the feminine plural French forms (les miennes, les tiennes), the
VLRF forms do not carry the gender and number references of their et-
yma, making clear that the very concepts of grammatical gender and num-
ber are at least much more marginal in VLRF than they are in French.
4.5.3 Possession in the VLRF noun phrase

Possession, usually indicated by the prepositional de in French, is indi-
cated by a noun complement in VLRF, i.e. without a preposition:
(134) VLRF: la kaz son momon (Ramassamy 1985:227)
the house [of] her mother (cf. F la maison de sa maman)
This construction is also found in fully creolized varieties:
(135) Mauritian CF: lakaz lapay mo tohtoh (Baker 1972:83)
house straw my uncle, i.e. ‘my uncle’s straw
house’
This construction is extended to more abstract connotations of de, such
as association, juxtaposition, etc.:
(136) VLRF: lankazman ti-fiy (Baker 1972:373)
‘the girl’s engagement’ (cf. F l’engagement de la fille)
(137) VLRF: bo la mer (Baker 1972:392)
‘by the sea’ (cf. F au bord de la mer)
In fact, Cellier (1985a:106) states that the French preposition de does
not exist in VLRF except in fossilized forms like “piedmang” (<pied de
mangue ‘mango tree’), and even this has a basilectal form without the d
of de: “piemang” (Watbled, personal communication).
4.5.4 Pronouns in VLRF

VLRF have the following personal pronouns (Ramassamy 1985:231–
239):
(138) VLRF: Singular Plural
1st person mwen ‘I’ nou ‘we’
2nd person ou/toué ‘you’ zot ‘you’
3rd person li ‘s/he, it’ zot ‘they’
bann-la ∼ banna ‘they’
Most are derived from the French disjunctive or emphatic pronouns,
while zot appears to come from F vous autres ‘you [others]’ or les or
eux autres ‘the others,’ and can refer to either ‘you [plural]’ or ‘they.’
The noun phrase 111
An alternative term for ‘they’ is bann-la (with the variant banna) from
F bande ‘band, gang’ (cf. the pluralizer bann, section 4.5.1). There is
a parallel here to those Atlantic creoles and their substrate languages in
which the pronoun meaning ‘they’ is juxtaposed to a noun to emphasize
plurality (see section 4.1.1). The VLRF pronoun bann-la was used by
slaveholders to refer to their slaves and can have a pejorative connotation
(Ramassamy 1985:237). Note that unlike their French counterparts (e.g.
il, elle; ils, elles), none of the above VLRF personal pronouns is marked
for gender. However, Cellier notes that VLRF has borrowed èl (cf. F elle
‘she’): “èl is a gallicism that usually (but not always) has a [+human] ref-
erent; this loan can be basilectal; it is marked for natural gender and does
not imply the introduction of grammatical gender” (Cellier 1985b:335,
my translation).
When personal pronouns are used as subjects before the agentive sub-
ject marker i, there is an obligatory contraction of the two morphemes as
follows (Cellier 1985b:234):
(139) VLRF: Singular Plural
1st person mi ni
2nd person wi/ti zot i
3rd person li zot i
Before the future marker va there is a parallel contraction to ma, wa, ta,
li va; na, zot va, but this is optional and the full forms (e.g. mwen va)
are found as well (ibid.).
The full pronouns can function not only as subjects but also as objects.
They simply follow prepositions (èk li ‘with him/her’), but as direct (140)
and indirect objects (141), all except bann-la are preceded by a-:
(140) VLRF: boug la lap rogad a-nou.

‘That man is looking at us.’ (Ramassamy 1985:235)
(141) VLRF: mwen la donn a-li en liv.
‘I gave him/her a book.’ (ibid.)
Note, however, that a- can also precede non-object pronouns:
(142) VLRF: manz a-ou. ‘[you singular] eat!’ (Ramassamy

1985:330)
(143) VLRF: manz a-zot. ‘[you plural] eat!’ (ibid.)
This a- seems to reinforce the emphatic nature of the pronoun, and thus
can be optional in preverbal position; it is obligatory in the post-verbal
position (Watbled, personal communication).
(144) VLRF: a-ou èk a-li va genyé. ‘You and he will win.’
(Ramassamy 1985:385)
(145) VLRF: ou èk li va genyé. idem. (ibid.)
While this a- is compatible (at least in form) with the French preposition
à ‘to,’ it should be noted that the stressed pronouns in a number of
varieties of restructured Portuguese are also marked by a- (< P a mim
‘to me’) (Holm 1988:203), and these forms may have entered VLRF via
Indo-Portuguese, reinforced by the initial a- in free-standing object and
possessive (but not subject) personal pronouns in Malagasy (Chaudenson
1974:954). Finally, it should be noted that the VLRF pronouns often
do not follow the case marking and word order of the non-emphatic
French pronouns. This makes them morphosyntactically more similar to
pronouns in fully creolized varieties of French (although none of these
takes an initial a-), which is compatible with the origin of the VLRF
pronouns in the French disjunctive pronouns.
VLRF does not have the reflexive verbs found in French (Cellier
1985a:68, 73). When the subject and object refer to the same person,
VLRF can express the latter simply by using the normal object pronoun
(146) or le ko:r (147); cf. F le corps ‘the body,’ used this way in archaic
and regional French, with parallels in many languages of the world includ-
ing creoles (Holm 1988:205), and Malagasy tena ‘body, self ’ (Chapuis
forthcoming).
(146) VLRF: li balãs ali. (cf. F il se balance)
‘He balances himself.’ (Chaudenson 1974:345)
(147) VLRF: zòt i pé pan zòt ko:r. (cf. F ils peuvent se pendre)
‘They can [go] hang themselves.’ (ibid.)
Reflexive verbs in French can also correspond to intransitive verbs in
VLRF:
(148) VLRF: mi apèl Paul. (cf. F je m’appelle Paul)
‘My name is Paul.’ (Cellier 1985a:175)
Note the similarity of VLRF apèl in (148) above to the corresponding
verb in a fully creolized variety of French:
(149) Haitian CF: Li rélé Mari. ‘She is called Mary.’ (Holm 1988:83)
4.6 A comparison of the noun phrase

The differences between the noun phrases of the partially restructured
languages examined here and their source languages are compared with
one another below to identify any common patterns of restructuring.
The noun phrase 113
4.6.1 A comparison of number in the noun phrase

In comparison with their source languages, number marking in the noun
phrase is considerably reduced in all the varieties under discussion except
standard Afrikaans. Plural marking mainly involves nouns in AAE and
Afrikaans, but in BVP and NSCS there is variable number agreement be-
tween head nouns and their modifiers. Phonological rules could account
for most of this variation in NSCS, but in BVP the variation clearly orig-
inated in morphosyntactic restructuring. In VLRF nouns take no plural
inflections whatsoever; they co-occur with article-like morphemes that
can indicate number, but there is no number agreement between nouns
and adjectives or other modifiers.
In AAE, Afrikaans, and VLRF there is a free morpheme of the same
form as the pronoun meaning ‘they’ which can serve as a plural or an
associative plural marker. A parallel construction is found in a number of
Atlantic creoles and their African substrate languages. The vestigial BVP
form osêle ‘they’ suggests that a parallel form may also have existed in
BVP.
4.6.2 A comparison of gender

Like English, AAE has no rules of gender agreement in the noun phrase.
Afrikaans nouns have not retained the covert gender marking of their
Dutch etyma, and the Dutch morphological rule for gender agreement
between adjectives and nouns has been reinterpreted as an (often un-
workable) phonological rule in Afrikaans. However, gender agreement in
the noun phrase is categorical in BVP (except for the Helvécia variety
and the Lı́ngua dos Pretos Velhos) and NSCS.
It is interesting that the BVP and NSCS mark number variably but
gender categorically, while in Afrikaans number marking is almost cat-
egorically present (except in nonstandard varieties) but gender mark-
ing is categorically absent (except with personal pronouns). Lucchesi
(2000:329ff.) tries to account for this in BVP by pointing out that the
phonological variability of /o/ and /a/ in the marking of gender is less plau-
sible than the phonological variation of /s/ and the nasalization of vowels
in marking number in subject–verb agreement, but it is equally implau-
sible that either of these rules was inherited from EP (sections 4.3.1 and
4.3.2). The explanation may lie in the evolutionary tendencies of the older
language varieties out of which the newer ones developed. While number
marking is still categorical in standard Dutch, Portuguese, and Spanish,
the marking of gender distinctions between masculine and feminine was
already breaking down in seventeenth-century Dutch (Brachin 1985:67).
This historical fact seems likely to have contributed to the complete loss
of gender marking on modern Afrikaans nouns, although contact with

African and Asian languages lacking this distinction is surely relevant as
well (note that modern Dutch nouns retain a distinction between com-
mon gender and neuter). Similarly, the optional marking of number in
many of the languages with which Spanish and Portuguese were in con-
tact in the New World (and with which Dutch was in contact in the Cape
Colony) seems to have worked against the categorical marking of number
in BVP and NSCS – and in nonstandard varieties of Afrikaans.
VLRF lacks number and gender agreement between nouns and ad-
jectives altogether. This seems likely to be related to substrate influence,
and it can be seen as evidence that VLRF underwent considerably greater
restructuring than BVP or NSCS. However, in VLRF, the synchronic
co-occurrence of morphemes derived from the singular versus plural and
masculine versus feminine forms of French articles and nouns makes the
question of the existence of grammatical gender and number more com-
plicated to determine than in the other restructured varieties discussed
here.
4.6.3 A comparison of possession

In the seventeenth century the inflectional marking of possession with -s
was being lost in Dutch but not in English; however, this construction
did not survive in Afrikaans or some varieties of AAE, probably because
there was no parallel inflection in the other languages in contact. Juxta-
position [possessor + possessed] appears to have prevailed in early AAE,
being both a universal and a substrate strategy for indicating possession
and associated relationships. However, Afrikaans adopted the possessive
marker of colloquial Dutch [possessor HIS possessed], which converged
with constructions in the languages with which it was in contact. BVP
and NSCS took over the standard Portuguese and Spanish construction
[possessed OF possessor], but the sporadic absence of the preposition in
some rural varieties of BVP suggests that in earlier stages the preposition
was variably absent, as it is in some modern varieties of creole Portuguese.
VLRF use a parallel construction without the French preposition de.
4.6.4 A comparison of pronouns

In all the restructured varieties except NSCS the personal pronouns of
the source languages underwent some loss of case distinctions, appar-
ently because of the influence of substrate languages or the restructur-
ing associated with adult second language acquisition. Reduction of case
marking was also influenced by the fact that Afrikaans, BVP, and VLRF
The noun phrase 115
all selected the emphatic or disjunctive personal pronouns of their source

languages (a characteristic strategy in language contact); these are less
often marked for case and more often follow SVO word order. Due to
decreolization the case discrepancies between AAE and standard English
pronouns have largely disappeared, but the partly similar discrepancies
between Afrikaans and Dutch have become fixed with the standardiza-
tion of Afrikaans. BVP appears to have reformulated a morphosyntactic
rule favoring emphatic (i.e. subject) pronouns as free morphemes fol-
lowing verbs over the object-case-marked clitics of standard European
Portuguese.
Reflexive pronouns are omitted in VLRF and in most varieties of BVP
(and an unusually basilectal variety of NSCS, Palenquero Spanish) as
they are in the Atlantic creoles, corresponding to constructions in sub-
strate languages. The difference in reflexive pronouns between Dutch and
Afrikaans appears to have nothing to do with language contact, resulting
instead from an innovation in Dutch after Afrikaans had already come
into existence.
5 The structure of clauses
Introduction
With the exception of Afrikaans, the partially restructured languages
examined here have the subject-verb-object word order for declarative
sentences that is found in both their Western European superstrate and
Niger-Congo substrate. In fact, this is also the basic word order in all
of the Atlantic creoles. Brazilian Vernacular Portuguese favors the SVO
order even for object pronouns, where European Portuguese can have
SOV order. Afrikaans is unlike the other varieties in that it has basic SOV
word order, but it is like them in so far as it follows the word order of
its superstrate, Dutch, which coincides with the SOV order of some of
Afrikaans’ substrate languages.
While the European lexical source languages can require the inversion
of the subject and the verb (or auxiliary) to transform a statement into
a question, this is not a part of the syntax of Niger-Congo languages or
full-fledged creoles. Instances of creole-like non-inversion of subject and
verb (or auxiliary) that would be unacceptable in the source language can
occur freely in all of the partially restructured varieties under discussion
except Afrikaans, which adheres strictly to Dutch word order.
5.1 The structure of AAE clauses
5.1.1 AAE word order

African American English has the usual English subject-auxiliary inver-
sion (or lack of it) in questions that can be answered “yes” or “no,” e.g.
“Can I go?” (Burling 1973:68). However, unlike standard English, AAE
has optional inversion with question words in the main clause, i.e. both
of the following occur:
(150) AAE: Where can I go?

Where I can go? (ibid.)
116
The structure of clauses 117
In embedded questions, which have no subject-auxiliary inversion in stan-

dard English, inversion is again optional in AAE:
(151) AAE: I wonder {where can I go}.
I wonder {where I can go}. (ibid.)
In embedded yes/no questions, AAE may have no connecting if or
whether but does have inversion:
(152) I wonder {can I go}. (ibid.)
In this respect AAE is unlike English-based creoles, which have no such
inversion at all and therefore happen to match standard English word
order in embedded questions:
(153) Jamaican CE: Dem aks mi {if a want i}.
‘They asked me {if I wanted it}.’ (Hancock
1979b:14)
A case might be made for AAE being more similar to Irish English, in
which direct questions are also embedded:
(154) Irish English: I don’t know {is that right or not}.
(Barry 1982:108)
While Irish English might well have served as a model for AAE at an
earlier period (Rickford 1986), the AAE pattern of subject-auxiliary in-
version could also be the result of partial restructuring or decreolization.
Bahamian English, which seems to be either more restructured or at
an earlier stage of decreolization than AAE (or both), has no subject-
auxiliary inversion in the basilect but frequent inversion in the upper
mesolect, even in embedded questions. Thus one finds (155) varying
with (156):
(155) Bahamian CE: I can go? (Holm 2000a:236)
(156) Bahamian CE: Can I go? (ibid.)
The same occurs in embedded questions:
(157) Bahamian CE: I don’t know {where I can go}. ∼ I don’t
know {where can I go}. (ibid.)
5.1.2 Dependent clauses in AAE

The structure of many AAE relative clauses is parallel to their equivalents
in standard English, but this is not always the case. In AAE the zero form
of the relative pronoun can be used not only for the object of the verb as
in standard English (e.g. “The man { he is hiring} is my uncle”) but

also for the subject:
(158) AAE: He got a gun { sound like a bee}. (Dillard 1972:68)
This structure is also found in Creole English (CE), e.g.
(159) Jamaican CE: De man { owe me money} gone a Cuba.
(Cassidy 1961:57)
According to Cassidy, Jamaican CE, “like the Niger-Congo languages . . .
gets along with paratactic constructions” (ibid.). However, Yoruba, which
is a Niger-Conger language, does have a relativizer (tı́ ‘who, which’) but
in “spoken Yoruba tı́ is often omitted,” e.g.
(160) Yoruba: as.o. { mo rà lánǎ} n’ı̀yı́ (Rowlands 1969:90)
cloth I bought yesterday this-is
‘This is the cloth I bought yesterday.’
However, “It cannot be omitted where its omission would produce am-
biguity,” e.g.
(161) Yoruba: màlúù {tı́ kò nı́ ı̀rù}
cow which no has tail
‘a cow which has no tail’ (ibid.)
In fact, the Yoruba relativizer tı́ sometimes seems more like a European
subordinator than a relative pronoun, in that it can introduce a clause
which requires its own subject pronoun even though the relativizer itself
would have this function in a European language:
(162) Yoruba: èmi {tı́ mo fún e. nı́ gbogbo owó yı̌}
1s REL 1s give 2s OBJ all money DEM
‘I who [I] gave you all this money.’ (Rowlands 1969:88)
Bickerton (1981:63) speculates that creoles may have been “born with-
out surface relativizers” and gives examples of zero subject relative pro-
nouns in Guyanese CE, Seychellois CF, and Annobón CP (1981:62–63).
AAE’s zero subject relative pronouns have been used to support its
creole history, but Tottie and Harvie (2000) provide convincing evidence
that the varieties of English that British settlers brought to the New World
also contained this construction. Not only was it the predominant form in
Middle English, as in Chaucer’s “I saugh a beest was lyk an hound”
(Tottie and Harvie 2000:202), but it is also found throughout British
regional varieties (Orton et al. 1978, map S5). This construction appar-
ently converged with the zero subject relativizers in African and creole
languages to favor the selection of this form in AAE. Although there is a
lack of published research in this area, Tottie and Harvie point out that
a study of one individual’s AAE (McKay 1969) reveals that “zero is the
most frequently used relativizer, with 54% of all cases, more than half of
which are subjects; that comes second with 38%, and what accounts for
9%; except for quotations from the Bible, who and which do not occur”
(Tottie and Harvie 2000:200). While zero subject relativizers account for
41 percent of the tokens in the ex-slave recordings, they account for only
2 percent and 5 percent in the modern spoken English of Americans and
Britons, respectively (2000:224).
However, it is ironic that Tottie and Harvie (2000:223) find the En-
glishness of AAE confirmed by the fact that “the Gullah relativizers wuh
and weh are totally lacking in our data” since these can be traced to
England’s Northcountry dialect (Holm with Shilling 1982:218). As a
matter of fact, Kautzsch (2002:213) notes that in his early AAE data
there are indeed a few sporadic occurrences of “relative clauses intro-
duced by non-spatial where, which makes it hard to categorically deny
any creole influence on AAE relative constructions.”
Dillard notes that some speakers of AAE seem to hypercorrect clauses
without relative pronouns, supplying not only an object relative pronoun
but also the clause’s original non-relative object pronoun:
(163) AAE: Dem little bitty hat {what dey wearin’ dem now}.
(Dillard 1972:68)
In AAE ambiguity in sentences without a subject relative pronoun can
often be cleared by a pleonastic subject pronoun (section 4.1.4) marking
the verb of the main clause:
(164) AAE: The boy { won} he did a three.
‘The boy who won did a three.’ (Smith 1973:94)
Regarding subordinate clauses, most in AAE are identical in structure to
those in standard English, with the notable exception of the use of say to
introduce a quotation:
(165) AAE: They told me {say they couldn’t get it}. (Rickford
1977:212)
This construction is also found in a number of English-based creoles:
(166) Krio CE: A yεri {se Olu de fes di buk kam}.
‘I heard that Olu is bringing the book along.’ (Yillah
forthcoming.)
(167) Gullah CE: dε lɔ {sε wi tu ol}.
‘They admit that we’re too old.’ (Turner 1949:211)
Turner (1949:201) pointed out the formal and syntactic similarity of
Gullah sε and Twi sε ‘that, saying’ and English say. Cassidy (1961:63)
noted that the pronunciation of Jamaican se is /sε/ when it means ‘that’

rather than /sey/, leading him to support the connection with Akan se.
Boretzky (1983:177) finds the lexical borrowing of se into the creoles an
inadequate explanation in light of the fact that the Surinamese creoles
have completely different forms, i.e. Sranan tak(i) and Saramaccan táa,
leading him to believe that the substrate influence on this construction
lay in the grammar rather than the lexicon.
(168) Sranan CE: M sab {tak a tru}. ‘I know that it’s true’
(Voorhoeve 1962:26)
The existence of parallel structures in creoles not based on English (Holm
2000a:208–209) supports Boretzky’s conclusion.
5.2 The structure of Afrikaans clauses
5.2.1 Afrikaans word order

As noted above, Afrikaans differs from the other varieties in this study,
following the SOV word order of its superstrate Dutch and some of
its substrate languages such as Khoi and Indo-Portuguese (den Besten
1986:187). Like Dutch, Afrikaans is considered an SOV language be-
cause this is the word order found in its dependent clauses, whether
these are subordinate (169) or relative (170):
(169) A: Jy weet {dat ek dit môre doen}.
2s know that 1s 3s tomorrow do
‘You know I’m doing it tomorrow.’ (Donaldson 1993:365)
(170) A: Dit is die mense {wat langsaan bly}.
3s be ART people REL next door live
‘These are the people who live next door.’ (ibid.)
SVO can be found in independent clauses with one verb: the verb occupies
the second position, which (as in German) usually follows the subject
(171), although time adverbials can also occupy the first position, leaving
the verb second and the subject third (172):
(171) A: Hy is siek vandag.
3s be sick today
‘He is sick today.’ (Donaldson 1993:362)
(172) A: Vandag is hy siek.
today be3s sick
‘Today he’s sick.’ (ibid.)
However, the fact that the main verb occurs at the end of the main clause
whenever there is an auxiliary verb (which then occupies the second
position) is taken as evidence that Afrikaans, like Dutch, has an underly-

ing SOV order:
(173) A: Ek sal dit môre doen.

1s will 3s tomorrow do
‘I will do it tomorrow.’ (Donaldson 1993:363)
This word order sets Afrikaans apart from the Atlantic creoles, which all
have strict SVO word order – including Negerhollands CD. However, it
confirms Afrikaans’ status as one of the partially restructured languages,
which follow the main word order of their superstrates except for some
creole-like variation in questions.
Note that in yes/no questions, Afrikaans follows the (inflected) verb-
subject word order of Dutch (and English in sentences with auxiliaries
and the equivalent of be):
V S
(174) A: Voel jy naar?
feel 2s nauseous
‘Do you feel nauseous?’ (Donaldson 1993:370)
(175) A: Sal jy dit asseblief vir my doen?
will 2s that please for me do
‘Will you please do that for me?’ (ibid.)
However, if there is a question word, Afrikaans follows (inflected) verb-

second word order (again like Dutch or English auxiliaries).
QW V S
(176) A: Waar bly jy?
where live 2s
‘Where do you live?’ (Donaldson 1993:323)
(177) A: Waar is die wildtuin?
where be the game park
‘Where is the game park?’ (Donaldson 1993:327)
Note, however, that in indirect questions Afrikaans has two possible word
orders. Following the word order of embedded questions in Dutch, the
verb can be final (the order normally found in written Afrikaans):
QW S V
(178) A: Hy sal seker weet {waar die wildtuin is}.
‘He’ll know for sure where the game park is.’ (ibid.)
Or the embedded question can follow the inverted V-S order of the direct
question, a syntactic innovation also found in AAE (see section 5.1.1):
QW V S
(179) A: Hy sal seker weet {waar is die wildtuin}. (ibid.)
5.2.2 Dependent clauses in Afrikaans

As noted above, Afrikaans follows Dutch word order in both main clauses
(verb or auxiliary second) and dependent clauses (verb and auxiliaries
last). Relative clauses in Afrikaans differ from those of Dutch, which
has various relative pronouns depending on gender and number, in that
Afrikaans makes almost exclusive use of a single relative pronoun wat
(which can never be omitted) for all antecedents:
(180) A: Ek wil die boek hê {wat jy in jou hand het}.

‘I want the book which you have in your hand.’
(Donaldson 1993:146)
This wat combines with the possessive marker se (section 4.2.3) to form
the relative ‘whose’ or ‘of which’: die tafel wat se poot af is ‘the table, the leg
of which is off’ (den Besten 1996:12) – although [+human] antecedents
can take the relative wie se. The only other exception to the exclusive
use of wat is when the relative is the object of a preposition, in which
case the Dutch forms are used when there is pied-piping (181a); if the
preposition is stranded (181b), then wat must be used (181b):
(181a) A: die probleem waarvan jy praat (den Besten 1996:12)

‘the problem whereof you speak’
(181b) A: die problem wat jy van praat (ibid.)
The origin of the form of the Afrikaans relative is Dutch wat, which can
occur as a relative pronoun after an indefinite (e.g. alles wat je zegt ‘every-
thing you say’). This form (and the parallel Negerhollands CD relativizer
wa) may have been selected over competing forms for grammaticaliza-
tion as the relativizer due to the influence of nonstandard or archaic
Dutch usage, or simply because it is more salient. Den Besten analyzes
it as a “gespecialiseerd voegwoord” [specialized conjunction] (1996:13)
that is a “marker for noninterrogative WH-movement in finite clauses”
(1981:141).
The indeclinability of wat can sometimes lead to ambiguity:
(182) A: Die Engelse soldate {wat dié Boere verslaan het}

‘The English soldiers whom those Boers defeated . . .’ OR
‘The English soldiers who defeated those Boers . . .’
(Donaldson 1984:64)
However, the latter meaning can be made unambiguous by using the

personal direct object marker vir:
(183) A: Die Engelse soldate {wat vir dié Boere verslaan het} . . . (ibid.)
This use of vir is an Afrikaans innovation unknown in Dutch. Brachin
(1985:140) attributes its origin to Malay, but it could also come from the
object-marker ku in Malayo-Portuguese or per in Indo-Portuguese:
(184) Malayo-Portuguese: Eli-sa mai mandá ku éli bai skóla.
‘His mother sent OBJ him to school.’
(Holm 1988–89:293).
In Afrikaans subordinate clauses, the subordinator dat is omitted as of-
ten as the English subordinator that, unlike the equivalent Dutch form,
which is never omitted:
(185) A: Ek weet {dat jy dit gedoen het}. ‘I know that you did it.’
(186) A: Ek weet {jy het dit gedoen.} ‘I know you did it.’
(Donaldson 1993:146)
(Note that without dat, the subordinate clause takes main clause word
order.) Regarding the possible omission of the Afrikaans subordinator
dat, Brachin (1985:141) again attributes this to Malay.
5.3 The structure of BVP clauses
5.3.1 BVP word order

While Portuguese is basically an SVO language, it is like other Romance
languages in that object pronouns can precede the verb, particularly in
European Portuguese and standard Brazilian Portuguese:
(187) EP, SBP: Não os vi. ‘I didn’t see them.’ (Cunha 1982:279)
NEG 3p saw-1s
In these varieties of Portuguese, object pronouns can also occur after the
verb as unemphasized clitics:
(188) EP, SBP: Ela chamou-me. ‘She called me.’ (Azevedo 1989: 863)
3s called 1s
However, in BVP these clitics are usually replaced with emphatic subject
pronouns, requiring SVO word order:
(189) BVP: Ela chamou eu. ‘She called me.’ (ibid.)
3s called 1s
Moreover, “subject-verb inversion in question constructions is not com-

monly found in BVP” (Mello 1997:167):
(190) BVP: Onde você mora? ‘Where do you live?’
where 2s live
This structure is paralleled in the Atlantic creoles and many of their
African substrate languages (Holm 1988–89:212–214), but not in the
Portuguese of Europe. There subject-verb inversion is required
(191) EP: Onde mora você?
where live you
unless the question word is emphasized:
(192) EP: Onde é que você mora?
where is-it that you live
Lemle (1976:77) notes that the following question structures can be ob-
served in urban Rio de Janeiro for ‘Where did you fall?’:
(193) MOST FORMAL: a Onde você caiu?
b Onde caiu você?
CAREFUL COLLOQUIAL: c Você caiu onde?
d Onde é que você caiu?
e Onde foi que você caiu?
f Onde é que foi que você caiu?
MOST INFORMAL: g Onde que foi que você caiu?
h Onde que você caiu?
The last structure, (193h) without é, is unknown in European Portuguese
but etymologically derivable from the structure illustrated in (190) above.
Here BVP que is parallel to a creole highlighter used after question words:
k’ in Capeverdean CP and ki in Guiné-Bissau CP (Baptista, Mello, and
Suzuki forthcoming). However, the most surprising result of Lemle’s
survey is that one of the most formal constructions (193a Onde você
caiu?) – presumably considered standard Brazilian Portuguese – is not
possible in European Portuguese.
5.3.2 Dependent clauses in BVP

In BVP almost all relative clauses are introduced by the relative pronoun
que, unlike the standard, which has a number of different relativizers
such as quem ‘whom’ (as the object of a preposition), cujo ‘whose,’
etc. In some rural or isolated varieties of BVP, such as that of Ceará,
the relativizer que (which takes the form ki) can be omitted, whether
it is functioning as the subject of the relative clause (195) or as the

object (196).
(194) BVP: moʃtré u pãu {ki eu kumia}
‘I showed the bread that I was eating.’ (Jeroslow
1974:193)
(195) BVP: u fradi morava nu sobradu { era
‘The priest lived on-the second-floor, [which] was
múitu áutu}
very high.’ (1974:194)
(196) BVP: i tudus kuienu us pedasu { kirı́a kumé}
‘And all choosing the pieces [that] they-wanted to-eat.’
(1974:195)
While the zero form of the relativizer is apparently unusual in other vari-
eties of BVP, it is also found in Helvécia Portuguese (Baxter and Lucchesi
1997:78).
To make clear the syntactic function of BVP que within the relative
clause, a resumptive pronoun can be used, e.g. BVP can preserve the
direct object in its original position even though it is also represented
by que – a structure unknown in standard Portuguese (but cf. a parallel
structure in AAE sentence [163] above):
(197) BVP: Esse rapaz, {que eu conheci ele} . . .
‘This guy that I met [him]’
(Amaral 1928 [1976] 78)
The resumptive pronoun can also be the object of a preposition, which
remains in its original position:
(198) BVP: O aluno {que eu conheço o pai dele.}
‘the student that I know the father of him’
‘The student whose father I know . . .’
However, the latter construction (198) is also found in colloquial Euro-
pean Portuguese. To work out its origin, it would be relevant to know if it
is attested in Europe before 1500; if not, it may have been imported from
Brazil. Standard Portuguese uses the relative determiner cujo to indicate
possession:
(199) SBP: O aluno {cujo pai eu conheço} . . .
‘The student whose father I know . . .’
Like colloquial European Portuguese, BVP can delete the resumptive
pronoun in sentences like (198) above, a construction which is preferred
because it avoids a stigmatized structure. While this deletion can some-

times result in a standard construction, as in the case of resumptive object
pronouns (197), it can also result in nonstandard structures, as when the
prepositional phrase in (198) is deleted:
(200) BVP: O aluno {que eu conheço o pai }
the student that I know the father
(201) BVP: aukãsó a kaza {ki tijã duxmidu}
‘He-reached the house [in] which he had slept.’
(Jeroslow 1974:193)
Such deletion of a stigmatized construction resulting in a structure that
is still nonstandard is characteristic of decreolization.
Although relative clauses with deleted resumptive pronouns are con-
sidered less socially marked, they are frequently ambiguous, as seen in
the possible interpretations of the following:
(202) BVP: a menina que falei . . .
SBP: a menina {com quem falei} . . . ‘The girl I talked to’
a menina {de quem falei} . . . ‘The girl I talked about’
a menina {por quem falei} . . . ‘The girl I spoke for’
As Tarallo (1986) notes, relative clauses beginning with a preposition plus
quem are only found in written Portuguese in Brazil:
(203) BVP: Esse fulano aı́, {com quem eu nunca tive aula}
‘This guy with whom I never had a class’
In speech que is the relativizer, and the prepositional phrase either occurs
in its original form and position (204) or not at all (205):
(204) BVP: Esse fulano aı́, {que eu nunca tive aula com ele}
(205) BVP: Esse fulano aı́, que eu nunca tive aula.
Baxter (1987) notes that relative clauses structurally parallel to the BVP
clauses above can be found in the Portuguese-based creoles of Africa,
such as São Tomé CP:
(206) São Tomé CP: omi {ku zõ sa ka fla n- e}
man who John PROG talk about him
‘the man that John is talking about’
To this could be added parallel constructions in other fully creolized
varieties of Portuguese and Spanish:
(207) Angolar CP: OmE {ki m ba kw E} . . .
the man that I went with [him] (Lorenzino
forthcoming)
(208) Cape Verde CP: kel ome {ke n fala k’ el} . . .

That man that I spoke with [him] (Baptista,
Mello, and Suzuki forthcoming)
(209) Guiné-Bissau CP: N mora na kasa {ku bu mora-ba n el}
I live in house that you live ANT in [it]
‘I live in the house that you used to live in.’
(ibid.)
(210) Papiamentu CS: E homber {ku m’a papia kun’ e}
The man that I PAST speak with [him]
a papia malu.
PAST speak bad
‘The man that I spoke with spoke badly.’
(Michel forthcoming)
Given the significant number of speakers of such creolized varieties of

Portuguese in Brazil during the first centuries of settlement from Eu-
rope and Africa while the predecessor of modern BVP was taking form,
their influence in making this structure a part of the vernacular seems
very likely. It should be noted that the above data provide clear counter
evidence to the position of Lucchesi (2000:34): “. . . não consideramos
provável a transferência de estructuras das lı́nguas do substrato para o
português brasileiro” [‘we consider it improbable that any structures were
transfered from substrate languages to Brazilian Portuguese’].
Regarding subordinate clauses, in some rural or isolated varieties of
BVP, such as that of Ceará, the complementizer que (which takes the
form ki) can be omitted:
(211) BVP: eli sabi {ki nãu se}

‘He knows that I do not know.’ (Jeroslow 1974:199)
(212) BVP: eu se { eu koj̃esu a mãga} (ibid.)
‘I know [that] I am familiar with the range.’
While the zero form of the subordinator ‘that’ is apparently unusual in

other varieties of BVP, it is also found in Helvécia Portuguese (Baxter
and Lucchesi 1997:78) as well as Cape Verdean and Guiné-Bissau Creole
Portuguese (Baptista, Mello, and Suzuki forthcoming).
5.4 The structure of NSCS clauses
5.4.1 NSCS word order

NSCS has nothing like the BVP use of subject pronouns as direct objects
after verbs, although some basilectal varieties achieve SVO word order by
replacing direct object pronouns before the verb with quasi-disjunctive

object pronouns marked by a after the verb.
NSCS is unlike most other dialects of Spanish (but like the Atlantic
creoles and their substrate languages) in that it usually requires expressed
subject pronouns (see section 4.3.4) and permits these pronoun subjects
to take SVO word order after non-subject question words, which in stan-
dard Spanish require subject–verb inversion:
(213) NSCS: Qué tú dices?
What you say, i.e. ‘What do you say?’
cf. S: Qué dices (tú)? (Holm 1989:308)
In the Atlantic creoles (but not in NSCS) SVO can be found in questions
even when the subject is a noun:
(214) Papiamentu CS: Unda e buki ta? ‘Where is the book?’
where DET book is (Maurer 1988:44)
5.4.2 Dependent clauses in NSCS

The form que predominates over other relative pronouns in NSCS, in-
cluding zero: even nineteenth-century Afro-Puerto Rican texts indicate
its fairly consistent presence:
(215) NSCS: ¿Cómo ba queré señorita que son tan bonita, uno
hombre ası́ tan feo?
‘How could you, who are so beautiful, love a man as
ugly as I am?’ (Alvarez Nazario 1961 [1974]: 385)
Zero relative pronouns, which do not occur in standard Spanish, did
occur in the unusually basilectal variety NSCS studied by Green (1997)
in the speech of the informant that Lipski thought to be mentally deficient
(section 2.4.3).
(216) NSCS: No, ahı́ hay una ta mara.
‘No, there is one that is bad there.’ (Green 1997:166)
This structure is parallel to that in AAE sentence (164) above, with a zero
relative pronoun acting as the subject of a relative clause.
The NSCS subordinator que can also take a zero form, which never
occurs in standard Spanish; in the bozal Spanish of Cuba, the comple-
mentizer que was often omitted:
(217) NSCS: Dice jagüey tá chiquito. ‘He says that the liana
is small.’
cf. S: Dice que el jagüey está chiquito (Cabrera 1969,
cited in de Granda 1978:486)
However, the presence of the complementizer is no less frequent. Simi-

larly, there is a variable absence of the complementizer que introducing
subordinate clauses in the basilectal variety of Dominican Spanish stud-
ied by Green (1997):
(218) NSCS: . . . dice se pusien acechalo ‘. . . they say that
they started to spy on him’
cf. S: . . . dicen que se pusieron a acecharlo. (Green 1997:166).
Another speaker of the same lect used para as a complementizer to
mark an infinite, like English to, which has no equivalent in standard
Spanish:
(219) NSCS: Yo no temo para hablá, conversá con nalie. (ibid. 168)
I NEG fear to speak converse with no one.
‘I do not fear speaking, conversing with anyone.’
Kihm (1994) notes that (unlike Spanish and Portuguese) Guiné-Bissau

CP uses pa (cf. P para ‘for’) to mark infinitive-like nominal uses of
verbs:
(220) Guiné-Bissau CP: Pa lei i yera un tarbaju difı́sil.
‘To read was a hard job.’
(Cited in Holm 1988:170)
In many Atlantic creoles based on Spanish and Portuguese, pa can mark
a tensed clause with a subject and verb, e.g.
(221) Papiamentu CS: Mi ke pa bo bai. (Goilo 1972)
1s want COMP you go
‘I want you to go.’
This construction is also found in the NSCS of bilingual Palenquero
speakers:
(222) PS: La abuelita mı́a no querı́a pa mı́ ir al monte.
‘My grandmother didn’t want me to go to the fields.’
cf. S: La abuelita no querı́a que yo fuera al monte.’ (Schwegler
and Morton 2002:150)
5.5 The structure of VLRF clauses
5.5.1 VLRF word order

Unlike French, in which subject-verb inversion can occur in questions,
in VLRF it does not occur:
QW S V
(223) VLRF: ousa vi rès? ‘Where do you live?’
QW V S
cf. F: Où habitez-vous? (Cellier 1985a:20)
However, it should be noted that strategies for avoiding subject-verb in-

version were well established in French before the eighteenth century,
when VLRF emerged (A. Kihm, personal communication):
QW S V
(224) F: Où est-ce que vous habitez?
In yes/no questions, the interrogative nature of the VLRF structure is

signaled solely by rising intonation (Ramassamy 1985:323):
S VP
(225) VLRF: Zan la vni zordi? ‘Did John come today?’ (ibid.)
John PAST come today
While this is also possible in standard French, the latter has other syntactic
strategies for marking such questions.
It should be noted that Malagasy, which played a significant role as a
substrate for the VLRF, has VOS sentence structure (Chapuis forthcom-
ing). However, a frequent Malagasy existential construction followed by
a relative clause lent itself to reinterpretation as SV(O) word order in
VLRF, which – like Malagasy – has optional relative pronouns (see sec-
tion 5.5.2):
(226) Malagasy: m- isy zaza izay m- itomany

PRES exist child (who) PRES cry
‘There is a child who cries.’ ‘Some child is crying.’
(ibid.)
(227) VLRF: Nana zanfan i plèr.
Exist child ASM cry
‘There is a child who is crying.’ ‘Some child is crying.’
(ibid.)
In fact, VLRF existential nana (cf. F il y en a ‘there are some’) became

reinterpreted as ‘to have’ in sentences like the following:
(228) VLRF: Larivyèr sinni nana bokou dlo zordi

river St.-Denis exist much water today
‘There is a lot of water in the St.-Denis River today.’
‘The St.-Denis River has a lot of water today.’ (ibid.)
As Chapuis (forthcoming) notes, “Indeed, here, from a Popular French

point of view, rivyèr sinni appears to be the subject: ‘The Saint-Denis
River,’ while, from the basilectal point of view, it constitutes an adverbial
phrase: ‘in the Saint-Denis River,’ since no preposition is used with place
names in the VLRF.”
Another feature of VLRF word order appears also to have developed
due to substrate influence. In French, the direct object must precede the
indirect (indicated by the preposition à):
S V DO IO
(229) F: Jean donne l’argent à Paul
‘John gives the money to Paul’; ‘John gives Paul the money’
(Chapuis forthcoming)
In the VLRF, however, no preposition is used and either object can come
first when neither is a pronoun:
IO DO
(230) VLRF: Zan la di momon tout mon ségré
John PAST tell mother all 1s secrets
‘John told Mother all my secrets.’ (Ramassamy
1985:271)
DO IO
(231) VLRF: Zan la di tout mon ségré momon. (ibid.)
Malagasy has the same lack of constraints on the order of direct and
indirect objects (Chapuis forthcoming citing Keenan 1975:251, 267–
268).
5.5.2 Dependent clauses in VLRF

Ramassamy (1985:229) notes that in VLRF relative clauses immediately
follow the noun they modify without a relativizer:
(232) VLRF: mang vert {mwen la manzé} la té maf.
mango green 1s PAST eat DET was over-ripe
‘The green mango that I ate was over-ripe.’ (ibid.)
While relative pronouns are obligatory in French, the VLRF pattern fol-
lows that of Malagasy, in which the relative pronoun izay can also be
omitted (Chapuis forthcoming citing Keenan 1978:296). Ramassamy
further notes (1985:230) that the end of the NP can be indicated by
la (as above), which demarcates the syntactic unit and facilitates the
parsing of the sentence. Similarly, Malagasy “frames” noun phrases with
an initial and final demonstrative (Chapuis forthcoming citing Keenan

1978:299). Cellier (1985a:150) notes that ke (cf. F qui) can function as
an overt relativizer, which Watbled (personal communication) describes
as acrolectal:
(233) VLRF: lété kèt soz {ke lété dur}. ‘It was something that was
hard.’(ibid.)
(234) VLRF: na dmoun i gardé le feu asoir (ibid.)
EXIST people ASM watch ART fire this-evening
‘There are people who watched the fire this evening.’
(ibid.)
cf. F: il y a des gens qui gardaient le feu ce soir.
Ramassamy (1985:378) does note that relativizers like ousa ‘where’ can
occur but are not obligatory. The fact that such relativizers occur op-
tionally as objects of prepositions (which are themselves optional) again
suggests influence from French:
(235) VLRF: mwen la pran le ti-santyé {(par-ousa) nou la pasé lot
zour}.
‘I took the little path by which we went the other day.’
(ibid.)
Again unlike French, VLRF can omit ke (cf. F que) functioning as
a subordinator. Ramassamy (1985:364) treats the juxtaposition of the
subordinate clause without a subordinator as the normal VLRF structure
after certain verbs, such as those of saying or perception:
(236) VLRF: li krwa { ou ri [sic: kri?] a-li}.
‘He thinks [that] you are calling him.’ (ibid.)
However, Cellier (1985a:150) notes that this kind of sentence can take
the subordinator ke (or its variant k), again possibly under the influence
of standard French:
(237) VLRF: mi kroi {k i apelé “tandif.”}
‘I thought that it was-called “tandif”.’ (ibid.)
From the perspective of French, the subordinator is also missing in
clauses following what seems to be a preposition, pou (cf. F pour ‘for’):
(238) VLRF: Mwen la di a-zot sa {pou li konèt}. (Ramassamy
1985:349)
‘I told you that so he would know.’
cf. F: Je vous ai dit ça pour qu’il le sache.
Ramassamy (1985:350) calls this a “pseudo-infinitif” since there are no

other grounds to justify positing the existence of infinitives in VLRF.
This construction, found throughout the Atlantic creoles – see sentence
(221) above – lends itself to translation with an infinitive in a number of
European languages (‘. . . for him to know’), but it is actually a tensed
clause and pou is best treated as a subordinator meaning ‘so that’ (Holm
1988:168–70). Watbled (personal communication) considers sentence
(238) as containing an infinitive clause; he points out that a corresponding
finite construction, more like the French equivalent of (238) above, also
exists:
(239) VLRF: Mwen la di a-zot sa {pou k(e) li koné.}
5.6 A comparison of clause structure

To reiterate, no claim is being made here that the newer language varieties
discussed above are any more closely related genetically or typologically
than the older languages out of which they developed. However, it is
being claimed that they underwent a similar kind of partial restructuring
triggered by partly similar social and linguistic factors. For this reason, the
differences in the structure of clauses of the newer and older languages
are compared with one another below to identify any common patterns
of restructuring that might characterize this kind of partial restructuring.
5.6.1 A comparison of word order

The five language varieties examined here have the word order for declar-
ative sentences that predominates in both their superstrate and substrate
languages. For all but Afrikaans this is SVO, which is also the basic word
order of all of the Atlantic creoles; for Afrikaans, that word order is SOV.
Although the inversion of the subject and the verb (or auxiliary) to trans-
form a statement into a question can or must occur in all the superstrate
languages, this is not a part of the syntax of full-fledged creoles or their
substrate languages. Non-inversion can or must occur in all the partially
restructured varieties except Afrikaans, which adheres strictly to Dutch
word order.
5.6.2 A comparison of dependent clauses

AAE relative clauses differ from those of standard English mainly in
that subject relative pronouns can be omitted; this structure is found in
some English-based creoles as well as earlier and nonstandard varieties of
English, which appear to have had a converging influence on AAE. On the

other hand, AAE’s use of the subordinator say to introduce a quotation
is clearly of creole and African origin. Dependent clauses in Afrikaans
closely follow the structure found in Dutch, except that Afrikaans has
a single flectionless relative pronoun wat (except for complex relatives)
and has a zero subordinator as an allomorph of dat.
BVP relative clauses differ from those of standard Portuguese in that
BVP has only one form of the relative pronoun, que, corresponding to
the single simplex form of the Afrikaans relativizer, wat. The structure of
BVP relative clauses follows patterns found in Portuguese-based creoles
and nonstandard European Portuguese. The avoidance of stigmatized
resumptive pronouns suggests the camouflaging strategies characteris-
tic of decreolization. NSCS and BVP differ from standard Spanish and
Portuguese in that the complementizer que can be omitted both as a rel-
ativizer and as a subordinator, but this occurs only in unusually basilectal
or archaic varieties. Zero seems to be the default form of both kinds of
VLRF complementizers, although ke or k also occurs, apparently under
the influence of French.
In summary, clauses in the partially restructured varieties examined
here have many of the structural features of their European source
languages, although they also reflect morphosyntactic simplification
and have some features found in fully creolized languages and their
substrates.
6 Conclusions
Introduction
This concluding chapter will attempt to relate the social information in
chapter 2 to the linguistic information in chapters 3–5 in order to provide
a basis for a theory that can account for the facts known about the five
languages examined here: what led to their partial restructuring and how
this process affected their structure.
6.1 Social factors in partial restructuring

Chapter 2 pointed to a single, overriding social factor in the develop-
ment of African American English, Afrikaans, Brazilian Vernacular Por-
tuguese, Nonstandard Caribbean Spanish, and the Vernacular Lects of
Réunion French as varieties distinct from both unrestructured overseas
varieties of their source languages (e.g. the English of Ontario, the ex-
tinct Dutch of New York and New Jersey, the Portuguese of Madeira,
the Spanish of Chile, or the French of Quebec) and completely restruc-
tured creole languages (e.g. Guyanese Creole English, the extinct Creole
Dutch of the Virgin Islands, Guiné-Bissau Creole Portuguese, Palen-
quero Creole Spanish, or Mauritian Creole French). That social factor is
the demographic balance, during the first century of a new language’s de-
velopment, of native speakers versus non-native speakers of the European
source language.
Parkvall (2000) is certainly correct in his conclusion that this demo-
graphic ratio is not the only social factor that determines the degree of
language restructuring: there are other relevant factors, such as an incom-
ing population already having some fluency in a common restructured
language brought in from elsewhere, such as the English-Creole-speaking
slaves imported into the southern American colonies during the seven-
teenth century (introduction to chapter 2 and section 2.1.1). Moreover,
time is certainly a relevant factor: the importance of a society’s final demo-
graphic balance can be overridden by the earlier emergence of a common
135
partially restructured language spoken by a population that is fully ac-

cessible to a later incoming population, as was the case in Barbados and
Réunion (sections 2.1.1 and 2.5.1–2.6, respectively).
However, we can conclude from both Parkvall (2000) and chapter 2
of the present study that the ratio between native and non-native speak-
ers of the source language during the first century of a new language’s
development is indeed the most important social factor in determining
the structure of that language. Where native speakers made up a strong
majority in the new society, unrestructured overseas varieties developed.
Where non-native speakers made up a strong majority, fully restructured
creole languages developed. Neither process was particularly pretty: it is
well known that many plantations that depended on slave labor were in
fact death camps that simply consumed the men and women brought
there to work. However, the ethnic changes necessary for unrestructured
overseas language varieties to flourish (except for those on previously un-
inhabited islands like Madeira, of course) were not much more attractive:
massive immigration of European colonists, who controlled the wealth,
government, and cultural institutions of the colonies; coupled with the
extermination, absorption, or retreat of speakers of indigenous languages.
The partial restructuring of languages occurred in new societies where
neither group – neither native nor non-native speakers, which in the be-
ginning meant neither Europeans nor non-Europeans – were numerous
enough completely to overwhelm the other group culturally.
The fact that partially rather than fully restructured languages devel-
oped in societies with a higher proportion of native speakers of the Euro-
pean lingua franca is logical: there were simply more native speakers to
provide non-native speakers with samples of the language from which the
latter could derive the rules needed to speak it. Despite social stratifica-
tion, learners still had better access to the target language than they did
in those plantation societies where fully creolized languages developed.
This led to two defining characteristics of the resulting languages: first,
the non-native version of the European language was never as completely
restructured as a fully creolized language; second, as the partially restruc-
tured language acquired native speakers – often among the descendants
of Europeans as well as non-Europeans – it developed into an identifying
community language that could draw on features not only from the non-
native lingua franca but also from native-speaker varieties of the European
language.
The five partially restructured languages examined here continued to
be in contact not only with native-speaker varieties of the European
language (either acquired by local whites via normal transmission or
brought in by new arrivals from the colonizing country) but also with
Conclusions 137
the fully restructured pidgins and creoles spoken by newly arrived slaves:
by Caribbean and African slaves brought to the American South; by
Asians and Africans brought to the Cape Colony; by African slaves arriv-
ing via Cape Verde and São Tomé to Brazil and the Spanish Caribbean;
and by Africans, Malagasies, and Indians brought to Réunion. Further-
more, these partially restructured languages were in varying degrees of
contact with the pidgins and subsequent creoles that developed in nearby
areas where sociolinguistic conditions were favorable to fuller restructur-
ing, producing the forerunners of Gullah in the American South, Orange
River Afrikaans in the Cape Colony, Helvécia Portuguese in Brazil, Habla
Bozal and Pororó in the Spanish Caribbean, and the Créole des Bas on
Réunion.
Thus, even after the local pidgin or jargon of the earliest contact period
ceased to be used, non-European features could still be borrowed into the
version of the local language used by the monolingual descendants of the
non-European groups. And, because humans signal their social identity
and solidarity with others through their choice of linguistic variables (Le
Page and Tabouret-Keller 1985), and because the social identities of the
descendants of both the European and the non-European groups evolved
over time, as did their relationships to one another, the variety spoken by
each group tended to grow less dissimilar to that of the other through the
two-way borrowing of features on all linguistic levels.
6.2 Linguistic factors in partial restructuring

Unlike the social factors discussed above, the linguistic facts surveyed
in the three preceding chapters are so many and various that they lend
themselves less readily to generalizations. Table 9 is presented in order to
allow a better overview of the morphosyntactic changes that characterize
partial restructuring.
Space constraints have made necessary a good deal of encoding in
this table; the following explanation is intended to help crack the code.
First, the features are described by number without abbreviations be-
low, followed by examples in AAE wherever possible (since readers know
the grammar of modern standard English, which is often close to AAE’s
source language), followed by the number of the sentence in the preced-
ing chapters, where further discussion of the morphosyntactic feature in
question can be found.
1. The absence of an inflection indicating the third person singular form
of the present tense:
AAE: Where Miss Annie . . . live now. (1)
Table 9. Key morphosyntactic features in partial restructuring
AAE Sb AFR Sb BVP Sb NSCS Sb VLRF Sb E D P S F
Verb phrase
Verbal morphology∗
1. Zero 3s PRES infl. + + + + 0 + 0 + + + 00 000
2. Zero 1p PRES infl. + + + + + + 0 + + + 00 000
3. Zero PAST infl. + + + + 0 + 0 + 0 00 000
Aux./ preverbal marker
4. Semantic influence + + + + + 00 000
Negation
5. Negative concord + 0 + + + 0 0 +++
6. Discontinuous double 0 + + + + + + 0 00 +0+
7. Non-verbal predicates + + 0 + + + + + + 00 000
Noun phrase
Number
8. Zero plural infl.∗ + + 0 + + + + + + + 00 000
9. Unbound pluralizer + + 0 + + + 0 + + + 00 000
10. Associative plural + + + + 0 + 0 + 0 + 00 000
Gender
11. No agreement in NP + + + + 0 + 0 + + + +0 000
Possession
12. [possessor 0 possessed] + + 0 + 0 + 0 + 0 + 00 000
13. [possessed 0 possessor] 0 + 0 + + + 0 + + + 00 000
Pronouns
14. Reduced case marking + + + + + + 0 + + + 00 000
15. Zero reflexive pronoun 0 + 0 + + + + + + + 00 000
Clauses
Word order
16. QW S-V/Aux (direct)∗∗ + + 0 + + + + + + + 00 000
Dependent clauses
17. Zero subject REL + + 0 + + + + + + + 00 000
18. Zero subordinator “that” + + + + + + + + + + +0 000
Total number of +’s 15 9 13 9 14
Sb = relevant substrate language(s) + = attested presence of feature

infl. = inflection 0 = attested absence of feature
see also List of Abbreviations, p. xix = not applicable or unknown
∗
refers to spoken, non-suppletive forms only
∗∗
with S immediately following QW (non-echo question)
2. The absence of an inflection indicating the first person plural of the

present tense:
BVP: nós parte ‘we leave’ (cf. SBP nós partimos ibid.) (34)
3. The absence of an inflection indicating past tense:
AAE: They taught me mighty good, they teach me good. (3)
It has to be stipulated that the above three features (as well as feature 8,
the absence of a plural inflection on nouns etc.) refer only to spoken,
non-suppletive forms in order to exclude irregular standard English
Conclusions 139
past forms like put or silent standard French inflections like (il ) vient
‘he comes.’ Of course, it is arbitrary to assign the inflection on E we
go or F il parle ‘he speaks’ either + or 0.
4. Semantic influence of a (creole) preverbal marker on a (source
language) auxiliary verb:
AAE: I got on a cowboy shirt now. . . . Been have it all my days. (4)
5. The presence of a negator before a verb requires the negative form of
indefinite determiners, pronouns, etc.
AAE: We don’ want no six-month investigation! (20)
Assigning the presence of negative concord a plus in this chart seems to
imply that it results from restructuring, which is not at all clear. It is a
standard feature in the Romance languages and in earlier, nonstandard,
varieties of English and Dutch (Brachin 1985:22). It is found in Atlantic
creoles of all lexical bases (Holm 2000:195–196), but its presence in
substrate languages is unknown.
Apropos of double negatives, part of the problem in decoding Table 9
has to do with negative designations for certain features, e.g. 1 (Zero 3s
PRES infl.) or 11 (No agreement in NP). This encoding was necessary to
avoid the opacity of having the same symbol (e.g. +) randomly encoding
both restructuring and the lack of it. As the table stands, + encodes
restructuring except in a few cases where the European superstrates
(E D P S F) differ among themselves structurally (e.g. feature 11: all
except English have gender agreement within the NP). In sum, no nega-
tive (or zero feature, such as the absence of an inflection) equals a negative
(0, e.g. the absence of the Zero 3s PRES infl. in standard English), while
its presence equals a positive (+, e.g. the presence of the Zero 3s PRES
infl. in AAE).
6. The presence of a negator both before and after the verb:
BVP: Ele não sabe não. literally ‘He doesn’t know no.’ (38)
7. The absence of the equivalent of a form of ‘be’:
AAE: They all dead. (24)
8. The absence of a plural inflection on nouns (or other elements of the
noun phrase):
AAE: them hound (68)
9. A separate word (often meaning ‘they’) to indicate plurality:
AAE: them hound (68)
10. The use of a pluralizing word to indicate a person’s usual associates:
AAE: Felicia an’ them done gone. (69)
11. The absence of inflections indicating gender agreement between a
noun and its modifiers:
VLRF: mon kaz lé gran ‘my house is big’ versus F ma maison est
grande (132)
12. Nouns have no inflection identifying them as the possessor of the

following noun:
AAE: the white folk kitchen (74)
13. Nouns do not follow a preposition indicating that they possess the
preceding noun:
BVP: kaza Maria ‘Maria’s house’ (cf. SBP a casa de Maria)
(110)
14. The forms of personal pronouns do not necessarily indicate their
grammatical function in the sentence as in the source language:
AAE: We had we own lawyers. (76)
15. Reflexive pronouns required in the source language can be omitted:
BVP: João cortou com faca. ‘John cut [himself] with a knife.’ cf.
SBP: João courtou-se com a faca. (115)
16. In direct questions with a question word, the subject can precede the
verb or auxiliary:
AAE: Where I can go? (150)
It should be explained that the subject must immediately follow the
question word to exclude constructions like Where [is it that] I can
go? Furthermore, it is a direct question and not the repetition of an
indirect question for confirmation: Where I can go?
17. A relative pronoun functioning as the subject of the clause can be
omitted:
AAE: He got a gun { sound like a bee}. (158)
18. The equivalent of ‘that’ introducing a subordinate clause can be
omitted:
NSCS: Dice jagüey tá chiquito. ‘He says that the liana is small.’
(217)
This survey of the salient morphosyntactic features that distinguish
these five partially restructured varieties from the standard variety of their
source language (which does not imply that the latter was necessarily the
most relevant source of the former) focuses on those features found in a
number of the restructured varieties rather than in just one, such as the
unbound possessive marker se in Afrikaans. This is because the point of
table 9 is to provide an overview of the general structural tendencies of
languages that have undergone partial restructuring.
Many of these tendencies can be characterized as structural reduc-
tion: reduced morphological marking for person or tense on verbs; for
number on nouns and other elements in the NP; or for case on personal
pronouns. Sometimes this reduction means the loss of syntactic complex-
ities (such as subject-verb inversion in questions) or the loss of function
words (the reflexive pronoun, the preposition equivalent to ‘of ’ indicating
Conclusions 141
possession, the subordinator equivalent to ‘that’). However, the loss of

these particular features rather than others does not seem to be random:
the losses that took place tend to make the partially restructured varieties
more like their substrate languages.
Of course isolating Niger-Congo languages formed an important part
of the substrate of most of the partially restructured languages examined
here, so it is not surprising that these tend to be more isolating, too.
However, it is also a universal in second language acquisition that adults
tend, when possible, to isolate grammatical information in unbound mor-
phemes rather than inflections. Yet there is no need to choose one of these
linguistic processes over the other in accounting for the structure of the
new varieties: each one obviously reinforced the other. Furthermore, the
fully restructured creoles of the same lexical base with which the partially
restructured varieties may have been in contact bear the mark of their sub-
strates even more strongly, but again there is no need to choose among
these three forces pulling in the same direction. In the final analysis, there
is no question that the partially restructured varieties bear the stamp of
their substrate, since they have innovative structures that represent not
a reduction of the structure of their superstrate languages but rather an
addition to it from their substrate: features 4 (the semantic influence of
preverbal markers on auxiliaries), 6 (discontinuous double negation), 7
(non-verbal predicates), 9 (unbound pluralizers), and 10 (associative plu-
rals). And there are many more such features that occur in only one or
two partially restructured varieties: the AAE complementizer say (5.1.2),
the VLRF agentive subject marker i (3.5.1), the resumptive pronouns in
BVP relative clauses (5.3.2), etc. Regarding the last feature, it is prob-
ably not entirely coincidental that this feature in BVP sentence (197) is
also found in AAE sentence (163): it is the same strategy for repairing
ambiguity in a variety that has undergone restructuring with all the poten-
tial for impairment of communication that this may have brought about.
Other such parallels include the option for subject-verb (or auxiliary) in-
version found in embedded questions in both AAE (151) and Afrikaans
(179). The cunning that languages appear to have to repair and renew
themselves makes August Schleicher’s nineteenth-century conceit – that
languages are living organisms in a Darwinian world – seem less quaint;
of course, such cunning is that of a language’s speakers, who are indeed
such organisms.
Finally, table 9 confirms subjective impressions about the degree of
restructuring which each of these varieties has undergone (cf. section
1.7). Arranging them according to their total number of positive features
(indicating restructuring) yields the following hierarchy:
15 African American English

14 Vernacular Lects of Réunion French
13 Brazilian Vernacular Portuguese
9 Afrikaans
9 Nonstandard Caribbean Spanish
Of course, the accuracy of such a quantitative indication of their degree
of restructuring depends on the representativeness of the features chosen
for table 9. Still, this hierarchy is confirmed by a comparison of the two
varieties with the most closely related source languages, BVP and NSCS.
Features indicating restructuring that are not found in NSCS are present
in BVP: e.g. features 2, 9, 13, and 14.
6.3 Linguistic processes in partial restructuring

I proposed an earlier version of the theory below (Holm 2000b) to ac-
count for what is known about partial restructuring, based on the results
of the research of my students, myself, and others. This model includes
the findings of Mello (1997), who observes that a number of linguis-
tic processes must have combined to trigger the partial restructuring of
Brazilian Vernacular Portuguese; and Green (1997), who in her study of
Nonstandard Caribbean Spanish, saw that, in addition to contact with
more fully restructured varieties, what sets this kind of partial restructur-
ing apart from other kinds of language shift (cf. Thomason and Kaufman
1988) is that it occurs among a shifting population speaking a number
of different first languages. Both of these factors place the kind of partial
restructuring discussed here closer to full creolization than language shift
by linguistically homogeneous groups, as Thomason acknowledges:
in ordinary shift situations there is no negotiation among shifting speakers,
because they all share the same L1; and if they learn the T[arget] L[anguage] im-
perfectly, they are likely to make similar or identical “errors.” These “errors” may
include (among other things) marked features of their original L1. (Thomason
2001:255)
However, the kind of partial restructuring discussed here is more like full
creolization in that it presupposes shift by a linguistically heterogeneous
population. This implies negotiation in the acceptance of substrate fea-
tures, so that highly marked features lacking transparency are less likely
to be accepted into the emerging vernacular. However, marked features
that are transparent – e.g. AAE say introducing quotations, as in sentence
(165) – are acceptable. More importantly, the existence of a number of
different first languages among the shifting population is crucial in that
it precipitates “. . . a forced shift because of an urgent need for a contact
medium in a new multilingual contact situation” since bilingualism in
Conclusions 143
two of the substrate languages cannot provide a solution to the crisis in

communication (Thomason and Kaufman 1988:307).
The theoretical model for partial restructuring presented here presup-
poses a population with different first languages shifting to a typologically
distinct target language (itself an amalgam of varieties in contact, includ-
ing fully restructured ones) under social circumstances that partially re-
strict access to the target language as normally used among a minority or
weak majority population of native speakers. It predicts that some or all
of the following linguistic processes will shape the resulting restructuring:
1. language drift, following internal tendencies within the target lan-
guage, particularly phonotactic, morphological, or syntactic simplifi-
cation;
2. primary leveling, preserving lexical or structural features that are
archaic, regional, or rare in the target language, sometimes extending
them to new contexts;
3. imperfect language shift by the entire population, perpetuating
structural features from ancestral languages and interlanguages in the
speech of monolingual descendants;
4. language borrowing, incorporating structural features from fully
pidginized or creolized varieties of the target language spoken by new-
comers or found locally but confined to areas where sociolinguistic
conditions were favorable to full restructuring;
5. secondary leveling, or the possible loss of features not found in the
target language (from any of the above processes) if there is contin-
ued contact with the target language and it is perceived to have more
prestige.
These processes result in a new variety with a substantial amount of the
target language’s structure intact, but also with a significant number of
substrate or interlanguage structural features, i.e. a partially restructured
vernacular.
The order in which these processes are listed above follows the order
in which they occur. The first two refer to the Western European lexi-
cal source languages taken overseas: all had been undergoing drift (1) in
Europe; when their different regional and social varieties came into closer
contact in an overseas colony, they underwent the first stage of dialect
leveling (2). The overridingly important process in partial restructuring
is language shift (3) of the dominated non-European population to the
language of the dominating Europeans. After this there is language bor-
rowing; it is convenient to distinguish the new variety’s borrowing of what
are largely substrate features (4) from its borrowing of features of the lex-
ical source language (5), parallel to the decreolization of fully creolized
varieties.
6.4 Comparative studies in restructuring

The aim of this book has been to demonstrate the usefulness of the con-
cept of the partial restructuring of languages. That this concept has eluded
linguists up to now is not surprising: partially restructured languages had
simply fallen between the cracks of theory, being neither unrestructured
overseas varieties nor fully restructured creoles. Each was compared only
to varieties of its lexical source language, so it was not possible for any pat-
tern to emerge. It is the comparison of such varieties not only with their
source languages but also with one another – focusing on their sociolin-
guistic histories as well as their synchronic structure – that makes it clear
that despite their dissimilar vocabularies they are, in a very important
sense, the same kind of language, resulting from the same sociolinguistic
processes.
Recent research comparing different partially restructured languages
(e.g. Holm, Lorenzino, and Mello 1999) clearly demonstrates that in-
sights gained from the study of one can cast light on others. While schol-
ars of NSCS disinclined to accept any external influence on its structure
have long resorted to postulating internally motivated phonological rules
to account for the loss of inflectional morphology on verbs and within
the noun phrase, the similar but more pervasive restructuring of BVP
indicates that phonological rules alone are an inadequate explanation,
forcing the issue of morphological restructuring resulting from contact
(see section 3.4.1). Because of the sociolinguistic parallels in the history
of NSCS and BVP, it can be argued by analogy that what led to the
reduced inflectional morphology of not only BVP but also NSCS was
contact-induced morphological simplification rather than just phonolog-
ical rules, which themselves seem more likely to have been motivated by
such contact than random language-internal forces. Moreover, the differ-
ence in the degree of restructuring evident in NSCS and BVP, objectively
confirmed in section 6.2, points to a simple but important observation:
that restructuring can indeed take place to differing degrees. This issue
is now settled.
Another insight from comparative study can be found in Afrikaans,
whose sociolinguistic history has been shown to have defining parallels
with the other four language varieties examined in this study. There is a
salient contrast between the lack of conjugational inflections in Afrikaans
and their presence in its source language, Dutch (section 3.2.1). At first
glance, this too might seem attributable to phonological rules such as
the simplification of final consonant clusters. In fact, such a phonological
rule has long been evoked to account for the loss of certain inflections
Conclusions 145
in another speech variety with important parallels in its sociolinguistic

history, AAE. The phonological constraints on the deletion of the mor-
pheme {-s} (marking the plural or possessive of nouns or the third person
singular present tense of verbs or even the contracted form of is) or of
{-d} (marking the past and past participle form of verbs) do not in them-
selves provide an adequate explanation for this deletion, as Labov (1969)
implied. Indeed, the flectionless simplicity of AAE habitual be is the mir-
ror image of Afrikaans is, and the only credible explanation is that both
varieties underwent morphological restructuring. Since there is no con-
vincing evidence that AAE itself was ever fully creolized (and considerable
demographic evidence to the contrary), we can safely conclude that AAE
underwent partial restructuring as it is defined above.
Such comparative work demonstrates the usefulness of this theoretical
model in accounting for the known facts of the historical development and
synchronic structure of the five partially restructured languages examined
in this study: identifying common sociolinguistic patterns in their histo-
ries should lead us to examine their linguistic structures aided by insights
gained from the study of other varieties that fall into the same category.
The concept of a partially restructured language is a useful generalization
about a group of languages that facilitates predictions about all members
of this category. The discovery of such falsifiable generalizations is what
science is supposed to be about.
As noted in the introduction to this book, there has been a recent move-
ment to place pidgin and creole linguistics within the broader scope of
contact linguistics (Thomason 1997), including language varieties result-
ing not only from the kind of restructuring associated with pidginization
and creolization (to whatever degree) but also those resulting from such
processes as intertwining (Bakker and Muysken 1994), koineization, and
indigenization (Siegel 1997). Such studies promise to increase our under-
standing of the range of possible outcomes of language contact, and this
understanding will surely shed new light on many kinds of restructuring.
For example, Siegel (1997) examined immigrant koines (e.g. overseas
Hindi), indigenized varieties (such as Singapore English), and even rena-
tivized Hebrew, concluding that the adoption of features in the leveling
that took place in all of these was affected by certain factors: frequency,
regularity, salience, and transparency. This suggests a solution for the
long search for principles that guide the selection of substrate features
into pidgins, creoles, and partially restructured varieties during their gen-
esis, and their adoption of other features during their development: the
likelihood of a lexical or structural feature being selected is greater if it is
more frequent, more regular, more salient, or more transparent. Siegel’s
observation of this tendency in contact between different varieties of the

same language is supported by Boretzky’s (1986) observation of the same
tendency in contact between different languages.
Contact linguistics should continue to apply the insights gained from
the study of one kind of restructuring to the study of other kinds of re-
structuring. Certainly it is time to re-examine what we understand not
only about Romani but also Romanian, Old French, and Middle English,
as well as all the other languages whose histories are known to have in-
volved substantial contact.
References
Adam, L. 1883. Les idiomes négro-aryen et maléo-aryen: essai d’hybridologie linguis-

tique. Paris: Maisonneuve.
Almada, M. D. de Oliveira. 1961. Cabo Verde: contribuição para o estudo do dialecto
falado no seu arquipélago. Estudos de ciências polı́ticas e sociais, No. 55.
Lisbon: Junta de Investigações do Ultramar.
Alvarez, Alexandra. 1990. Venezuelan Spanish: the search for a creole continuum.
Ph.D. dissertation, Georgetown University.
Alvarez Nazario, Manuel. 1961. El elemento afronegroide en el español de Puerto
Rico. San Juan de Puerto Rico: Instituto de Cultura Puertorriqueña (2nd
edn., 1974).
Amaral, Amadeu. 1928. O Dialeto Caipira. (3rd edn. 1976, São Paulo:
HUCITEC, Secretaria da Cultura, Ciência e Tecnologia).
Arends, Jacques, ed. 1995. The Early Stages of Creolization. Amsterdam Philadel-
phia: John Benjamins.
1999. The origin of the Portuguese element in the Surinam Creoles. In
M. Huber and M. Parkvall, eds., pp. 195–208.
Arends, J., P. Muysken, and N. Smith, eds. 1994. Pidgins and Creoles: an intro-
duction. Amsterdam/ Philadelphia: John Benjamins.
Azevedo, Milton M. 1989. Vernacular features in educated speech in Brazilian
Portuguese. Hispania 72, pp. 862–872.
Bailey, Guy. 2001. The relationship between African American and White vernac-
ulars in the American South: a sociocultural history and some phonological
evidence. In S. Lanehart, ed., pp. 53–92.
Bailey, Guy, and Natalie Maynor. 1987. Decreolization? Language in Society 16,
pp. 449–473.
1989. The divergence controversy. American Speech 64, pp. 12–39.
Bailey, Guy, Natalie Maynor, and Patricia Cukor-Avila, eds., 1991. The Emergence
of Black English: texts and commentary. Amsterdam: John Benjamins.
Baker, Philip. 1972. Kreol: a description of Mauritian Creole. London: Hurst.
1982. The Contribution of Non-Francophone Immigrants to the Lexicon of Mauri-
tian Creole. Ph.D. diss., University of London.
1996. Pidginization, creolization, and français approximatif. Review of Des
ı̂les, des hommes, des langues, by Robert Chaudenson. Journal of Pidgin and
Creole Languages 11:1, pp. 95–120.
Baker, Philip, and Chris Corne. 1982. Isle de France Creole: affinities and origins.
Ann Arbor: Karoma.
147
148 References
Bakker, Peter, and Pieter Muysken. 1994. Mixed languages and language inter-
twining. In Arends et al., eds., pp. 41–84.
Baptista, Marlyse, Heliana R. de Mello, and Miki Suzuki. Forthcoming.
Capeverdean and Guiné-Bissau Creole Portuguese. In J. Holm and P.
Patrick, eds.
Barry, M. V. 1982. The English language in Ireland. In R. Bailey and M. Görlach,
eds., English as a World Language. Ann Arbor: University of Michigan Press,
pp. 84–133.
Baugh, John. 1983. Black Street Speech: its history, structures, and survival. Austin:
University of Texas Press.
Baxter, Alan N. 1987. Creole-like traces in rural Brazilian Portuguese dialects.
Department of Spanish, La Trobe University, Bundoora, Australia. 47
pp. MS.
1992. A contribuição das comunidades afro-brasileiras isoladas para o debate
sobre a crioulização prévia: um exemplo do estado da Bahia. In E. d’Andrade
and A. Kihm, eds., pp. 7–35.
1997. Creole-like features in the verb system of an Afro-Brazilian variety of
Portuguese. In A. Spears and D. Winford, eds., pp. 265–288.
Baxter, Alan N., and Dante Lucchesi. 1997. A relevância dos precessos de
pidginização e crioulização na formação da lı́ngua portuguesa no Brasil.
Estudos lingüı́sticos e literários 19, pp. 65–84.
1999. Un paso más hacia la definición del pasado criollo del dialecto afro-
brasileño de Helvécia (Bahia). In K. Zimmermann, ed., pp. 119–141.
Béraha, Jeanette. 2002. La particle verbal i en réunionnais. Paper presented
at the Colloque International des Études Créoles, Réunion.
Bhattacharjya, Dwijen. 2000. The Genesis and Development of Nagamese: its social
history and linguistic structure. Ph.D. dissertation, City University of New
York.
Bickerton, Derek. 1973. On the nature of a creole continuum. Language 49,
pp. 640–649.
1981. Roots of Language. Ann Arbor: Karoma.
1984. The language bioprogram hypothesis. Behavioral and Brain Sciences 7,
pp. 173–221.
Bickerton, Derek, and Aquiles Escalante. 1970. Palenquero: a Spanish-based
creole of northern Colombia. Lingua 24, pp. 254–267.
Bloomfield, Leonard. 1933. Language. London: Allen and Unwin.
Bollée, Annegret. 1977. Zur Entstehung der französischen Kreolen-dialekte im Indi-
schen Ozean: Kreolisierung ohne Pidginisierung. Geneva: Droz.
Bonvini, Emilio. 2000. La langue des pretos velhos (vieux noirs) au Brésil: un
créole à base portugaise d’origine africaine? Bulletin de la Société de Linguis-
tique de Paris XCV, pp. 389–416.
Boretzky, Norbert. 1983. Kreolsprachen, Substrate und Sprachwandel. Wiesbaden:
Harrassowitz.
1986. Verbkategorien im Fantse und im Jamaican Creole. Ms.
Boshoff, S. P. E. 1921. Volk en taal in Suid-Afrika. Pretoria: De Bussy.
1959. Die wonder van die ontstaan van Afrikaans. In M. S. du Buisson, ed.,
Die wonder van Afrikaans. Johannesburg: Voortrekkerpers, pp. 25–35.
References 149
Bosman, Daniel B. 1923. Oor die ontstaan van Afrikaans. Amsterdam: Swets and
Zeitlinger.
1947. Die ontstaan van Afrikaans. In C. M. van den Heever and P. de Villiers
Pienaar, eds., Kultuurge skiedenis van die Afrikaner. Cape Town, Bloem-
fontein, and Port Elizabeth: Nasionale Pers 2, pp. 1–22.
Bourquin, A. 1994. Étude d’une catégorie sociale: les Petits Blancs de la Réunion.
Thèse de doctorat. Université J. Moulin, Lyon.
Boxer, C. R. 1963. Race Relations in the Portuguese Colonial Empire, 1415–1825.
Oxford: Clarendon Press.
1965. The Dutch Seaborne Empire. New York: Knopf.
1969. The Portuguese Seaborne Empire, 1415–1825. New York: Knopf.
Brachin, Pierre. 1985. The Dutch Language: a survey. Leiden: Martinus Nijhoff.
Burgers, M. P. O. 1963. Teach Yourself Afrikaans. London: The English Univer-
sities Press.
Burling, R. 1973. English in Black and White. New York: Holt, Rinehart, and
Winston.
Byrne, F., and J. Holm, eds. 1993. Atlantic Meets Pacific. Amsterdam/Philadelphia:
John Benjamins.
Cabrera, Lydia. 1969 [1954]. El Monte. Miami: Ediciones C.R.
Caid, Leila. 2002. Le morphème verbal i en créole réunionnais. Paper pre-
sented at the Colloque International des Études Créoles, Réunion.
Calvet, Louis-Jean and Robert Chaudenson. 1998. Saint-Barthélemy. Une énigme
linguistique. Paris: Didier Erudition.
Carden, G., M. Goodman, R. Posner, and W. A. Stewart. 1991. A 1671 French
creole text from Martinique. Paper presented to the Society for Pidgin and
Creole Linguistics, Chicago.
Careno, Mary F. do. 1997. Vale do Ribeira: a voz e a vez das comunidades negras.
São Paulo: Editora Arte e Ciência.
Carlin, Eithne B., and Jacques Arends, eds. 2002. Atlas of the Languages of Suri-
name. Koninklijk Instituut voor Taal-, Land- en Volkenkunde, Caribbean
Series 22. Leiden: KITLV Press.
Carvalho, Felix de. 1977. Falares crioulos de Brasil: um tema em debate. Rio
de Janeiro: Dpto. de Letras, Pontı́fica Universidade Católica, dissertação de
mestrado.
Cassidy, Frederic G. 1961. Jamaica Talk: three hundred years of the English language
in Jamaica. London: Macmillan.
Cassidy, Frederic G., and Joan H. Hall, eds. 1985 . . . Dictionary of American
Regional English. Cambridge, Mass: Belknap Press.
Cassidy, Frederic G., and R. B. Le Page. 1980. Dictionary of Jamaican English.
Cambridge: Cambridge University Press.
Castro, Y. de Pessoa de, and G. A. de Souza Castro. 1980. Culturas africanas nas
Américas: um esboço de pesquisa conjunta na localização dos empréstimos.
Afro-Asia 13, pp. 27–50.
Cellier, Pierre. 1985a. Comparaison syntaxique du créole réunionnais et du français
(Réflexions pré-pédagogiques). Université de la Réunion.
1985b. Description syntaxique du créole réunionnais: essai de standardisation.Thèse
de doctorat d’Etat. Aix: Université de Provence.
150 References
Chapuis, Daniel. forthcoming. Aspects of Restructuring in Vernacular Lects of

Réunion French. Ph.D. dissertation, City University of New York.
Chaudenson, R. 1974. Le lexique du parler créole de la Réunion. Paris: Champion.
2 vols.
1992. Des ı̂les, des hommes, des langues. L’Harmattan: Paris.
1995. Les Créoles. Que sais-je? Paris: Presses Universitaires de France.
2000. Créolisation du français et francisation du créole: Les cas de Saint-
Barthélemy et de la Réunion. In I. Neumann-Holzschuh and E. W. Schnei-
der, eds., pp. 361–381.
Coelho, F. Adolpho. 1880–1886. Os dialectos românicos ou neolatinos na Africa,
Asia, e América. Boletim da Sociedade de Geographia de Lisboa. Reprinted
1967 in Estudos Lingüı́sticos Crioulos, J. Morais Barbosa. Lisbon: Acadêmia
Internacional de Cultura Portuguesa.
Cohen, W. B. 1980. The French Encounter with Africans: white response to blacks,
1530–1880. Bloomington London: Indiana University Press.
Combrink, J. 1978. Afrikaans: its origin and development. In L. W. Latham and
K. P. Prinsloo, eds., Language and Communication Studies in South Africa.
Oxford: Oxford University Press, pp. 69–95.
Couto, Hildo H. do. 1996. Introdução ao estudo das lı́nguas crioulas. Brası́lia:
Editora da Universidade de Brası́lia.
1997. Falar capelinhense: um dialeto conservador do interior de Minas Gerais.
Colóquio internacional: substandard e mudança no português do Brasil. Berlin:
Ibero-Amerikanisches Institut.
Craig, Beth. 1991. American Indian English. English World-Wide 12:1, pp. 25–
62.
Crystal, David. 1986. A Dictionary of Linguistics and Phonetics. Oxford: Blackwell.
Cukor-Avila, Patricia. 2001. Co-existing grammars: The relationship between the
evolution of African American and White Vernacular English in the South.
In Sonja L. Lanehart, ed. 2001, pp. 93–127.
Cunha, Celso Ferreira da. 1981. Lı́ngua, Nação, Alienação. Rio de Janeiro: Editora
Nova Fronteira (Coleção Logos).
1982. Gramática da Lı́ngua Portuguesa. Rio de Janeiro: FENAME (Fundação
Nacional de Material Escolar, Ministério da Educação e Cultura).
Curtin, Philip D. 1975. The Atlantic slave trade. In A. M. Pescatello, ed., The
African in Latin America. New York: Knopf, pp. 35–48.
da Cunha, A. G. et al., 1982. Dicionário etimológico Nova Fronteira da lı́ngua por-
tuguesa. Rio de Janeiro: Editora Nova Fronteira.
d’Andrade, Ernesto, and Alain Kihm, eds. 1992. Actas do colóquio sobre crioulos
de base lexical portuguesa. Lisbon: Colibrı́.
DeCamp, David. 1961. Social and geographical factors in Jamaican dialects. In
R. B. Le Page, ed., Creole language studies II. Proceedings of the Conference
on Creole Language Studies (University of the West Indies, Mona, 1959).
London: Macmillan.
1971. Towards a generative analysis of a post-creole speech continuum. In
Hymes, ed., 1971, pp. 349–370.
DeGraff, Michel, ed., 1999. Language Creation and Language Change. Cambridge,
Mass: MIT Press.
References 151
2001. Against creole genesis as “abnormal transmission.” Paper presented to

the Linguistic Society of America, Washington, DC.
de Granda, Germán. 1968. La tipologı́a “criolla” de dos hablas del área lingüı́stica
hispánica. Thesaurus 23, pp. 193–205.
1970a. Un temprano testimonio sobre las hablas “criollas” en África y América.
Thesaurus 25, pp. 1–11.
1970b. On the study of the creole dialects in Spanish-speaking areas. Orbis 19,
pp. 72–81.
1975. Planteamientos y necesidades actuales de los estudios lingüı́sticos
afrohispano-americanos. Anuario de Letras 12, pp. 53–82.
1976. Algunos rasgos morfosintácticos de posible origen criollo en el habla
de áreas hispanoamericanas de población negra. Anuario de Letras 14,
pp. 5–22.
1978. Estudios lingüı́sticos hispánicos, afrohispánicos y criollos. Madrid: Editorial
Gredos.
1987. Situación actual de los estudios lingüı́sticos afrohispanoamericanos. The-
saurus 31, pp. 481–501.
1998. Estudios sobre el español caribeño (1985–1995): resumen y evaluación
de resultados. In M. Perl and A. Schwegler, eds., pp. 62–69.
de Kleine, Christa. 1997. The verb phrase in Afrikaans: Evidence of creolization?
In D. Winford and A. K. Spears, eds., pp. 289–307.
Deltel, J. R. 1969. Le Créole de la Réunion. Annales de Nice 7, pp. 29–34.
den Besten, Hans. 1978. Cases of possible syntactic interference in the develop-
ment of Afrikaans. Amsterdam Creole Studies 2, pp. 5–56.
1981. Marking WH-Movement in Afrikaans. In P. Muysken, ed., pp. 141–
179.
1985. Die doppelte Negation im Afrikaans und ihre Herkunft. In N. Boretzky
et al. eds., Akten des 1. Essener Kolloquiums über Kreolsprachen und Sprach-
kontakte vom 26.1.1985 an der Universität Essen. Bochum: Studienverlag
Dr. N. Brockmeyer, pp. 9–42.
1986. Double negation and the genesis of Afrikaans. In P. Muysken and
N. Smith, eds., Substrata versus Universals in Creole Genesis. Amsterdam:
John Benjamins, pp. 185–230.
1987. Review article of Raidt 1983. Journal of Pidgin and Creole Languages 2,
pp. 67–92.
1989. From Khoekhoe foreigner talk via Hottentot Dutch to Afrikaans: the
creation of a novel grammar. In Pütz and Dirven, eds., pp. 207–250.
1993. The Khoikhoi basis for the Afrikaans associative construction. Paper
presented to the Society for Pidgin and Creole Linguistics, Amsterdam.
1996. Afrikaans relatief “wat” en de West-Germanse relativisatiesystemen. In
H. den Besten, J. Goossens, F. Ponelis, and P. van Reenen, eds., Afrikaans
en variëteiten van het Nederlands. Amsterdam: P. J. Meertens-Instituut. Taal
en Tongval, Thema nummer 9, pp. 12–34.
Denis, Béatrice. 1998. St. Thomas French: a semi-creole? 62-p. ms., Department
of Linguistics, City University of New York.
Deumert, Andrea. 1999. Variation and Standardisation: the case of Afrikaans
(1880–1922). Ph.D. dissertation, University of Cape Town.
152 References
Dieck, Marianne. 2000. La negación en Palenquero: análisis sincrónico, estudio

comparativo y consecuencias teóricas. Frankfurt: Vervuert/Madrid: Iberoamer-
icana.
Dillard, J. L. 1972. Black English: its history and usage in the United States. New
York: Random House.
D’Offay, Danielle, and Guy Lionnet. 1982. Diksyonner kreol-franse/Dictionnaire
créole seychellois-français. Hamburg: Buske.
Donaldson, Bruce C. 1984. Dutch Reference Grammar. Leiden: Martinus Nijhoff.
1993. A Grammar of Afrikaans. Berlin, New York: Mouton de Gruyter.
du Toit, P. J. 1905. Afrikaansche studies. (dissertation, Ghent). Ghent: A. Siffer.
Elphick, R., and V. C. Malherbe. 1989. The Khoisan to 1828. In R. Elphick
and H. Giliomee, eds., The Shaping of South African Society, 1652–1840.
Middletown, Conn.: Wesleyan University Press, pp. 3–65.
Famighetti, Robert, ed., 1997. The World Almanac and Book of Facts 1998.
Mahwah, NJ: World Almanac Books.
Fasold, Ralph W., and Walt Wolfram. 1970. Some linguistic features of Negro
dialect. In R. Fasold and R. Shuy, eds., pp. 87–104.
Fasold, Ralph W., and Roger W. Shuy, eds., 1970. Teaching Standard English in
the Inner City. Washington, D.C.: Center for Applied Linguistics.
Fasold Ralph, W., William Labov, Fay Bayel Vaughn-Cooke, Guy Bailey, Walt
Wolfram, Arthur K Spears, and John Rickford. 1987. Are black and white
vernaculars diverging? American Speech 62:1, pp. 3–80.
Feagin, C. 1979. Variation and Change in Alabama English: a sociolinguistic study
of the White community. Washington: Georgetown University Press.
Ferguson, C. A., and S. B. Heath, eds., 1981. Language in the USA. Cambridge:
Cambridge University Press.
Ferraz, Luiz. 1976. The origin and development of four creoles in the Gulf of
Guinea. African Studies 35:1, pp. 33–38.
Figueroa A., V. Jesús. 1998. Rasgos semi-criollos en el español no estándar de la
región suroriental cubana. In K. Zimmermann, ed., pp. 411–440.
Folb, Edith A. 1980. Running Down Some Lines: the language and culture of
Black teenagers. Cambridge, Mass.: Harvard University Press (see glossary,
pp. 227–260).
Franken, J. L. M. 1952. Taalhistoriese bydraes. Amsterdam, Capetown: Balkema.
Freitas, Décio. 1978. Palmares: A Guerra dos Escravos. Rio de Janeiro: Graal.
Fry, P., C. Vogt, and M. Gnerre. 1981. Mafamba e caxapura: na encruzilhada
da identidade. Revista de Ciências Sociais 24:3, pp. 373–389.
Gilbert, Glenn G. 1985. Hugo Schuchardt and the Atlantic creoles: a newly
discovered manuscript “On the Negro English of West Africa.” American
Speech 60:1, pp. 31–63.
Gilbert, Glenn. ed. 1987. Pidgin and Creole Languages: essays in memory of John
E. Reinecke. Honolulu: University Press of Hawaii.
ed. 2002. Pidgin and Creole Linguistics in the Twenty-first Century. New York:
Peter Lang.
Goilo, E. R. 1972. Papiamentu Textbook. Aruba: de Wit.
Goodman, Morris. 1987. The Portuguese element in the American creoles. In
G. Gilbert, ed., pp. 361–405.
References 153
Grondin, R.-C. 1998. La construction historique des catégories sociales. L’exemple des
Petits Blancs des Hauts. Mémoire de DEA. Université de la Réunion.
Green, Katherine. 1997. Non-standard Dominican Spanish: evidence of partial re-
structuring. Ph.D. dissertation, City University of New York. Ann Arbor:
University Microfilms International.
Guy, Gregory R. 1981. Linguistic Variation in Brazilian Portuguese. Ph.D. disser-
tation, University of Pennsylvania.
1989. On the nature and origins of Popular Brazilian Portuguese. In Estudios
sobre español de América y lingüı́stica afroamericana. Bogotá: Instituto Caro y
Cuervo, publication no. 83, pp. 227–245.
Hackert, Stephanie, and John Holm. Forthcoming. Southern Bahamian: trans-
ported AAVE or transported Gullah? In John Lipski ed., African-American
English and its Congeners. Amsterdam Philadelphia: John Benjamins.
Hahn, Theophilus. 1882. On the Science of Language and its Study, with Special
Reference to South Africa. Cape Town: Herrmann Michaelis.
Hancock, Ian F. 1969. The Malacca Creoles and their language. Afrasia 3,
pp. 38–45.
1971. A Study of the Sources and Development of the Lexicon of Sierra Leone Creole.
Ph.D. thesis, School of Oriental and African Studies, University of London.
1975. Malacca Creole Portuguese: Asian, African, or European? Anthropologi-
cal Linguistics 17(5): 211–236.
1979. The relationship of Black Vernacular English to the Atlantic creoles.
Working paper of the African and Afro-American Studies and Research
Center, University of Texas at Austin.
1987. A preliminary classification of the Anglophone Atlantic creoles, with
syntactic data from 33 representative dialects. In G. Gilbert, ed., pp. 264–
334.
Hartt, Charles F. 1872. Notes on the Lingoa Geral or modern Tupi of the
Amazonas. Transactions of the American Philological Association 3, pp. 58–76.
Healy, Maureen. 1993. The parallel continuum model for Suriname: a prelimi-
nary study. In F. Byrne and J. Holm, eds., pp. 279–289.
Henriquez Ureña, Pedro. 1987 [1940]. El español en Santo Domingo. Santo
Domingo: Editora Taller.
Hesseling, D. C. 1897. Het Hollandsch in Zuid Afrika. De Gids 61, pp. 138–62;
reprinted as Dutch in South Africa: In T. L. Markey and Paul T. Roberge,
eds. and trans., On the Origin and Formation of Creoles: a miscellany of articles,
Ann Arbor: Karoma, 1979.
Hickey, Raymond. 1997. Arguments for creolisation in Irish English. In R. Hickey
and S. Puppel, eds., Language History and Linguistic Modeling: a festschrift for
Jacek Fisiak on his 60th Birthday. Berlin: Mouton de Gruyter, pp. 969–1038.
Hickey, Raymond, ed. forthcoming. Legacies of Colonial English. Cambridge:
Highfield, Arnold R. 1979. The French Dialect of St. Thomas, U.S. Virgin Islands.
Ann Arbor: Karoma.
Holm, John. l978. The Creole English of Nicaragua’s Miskito Coast: its sociolinguistic
history and a comparative study of its lexicon and syntax. Ph.D. thesis, University
College, London. Ann Arbor: University Microfilms International.
154 References
1980. African features in White Bahamian English. English World-Wide 1:1,

pp. 45–65.
ed. 1983. Central American English. Heidelberg: Groos/Amsterdam: Benjamins.
1983. On the relationship of Gullah and Bahamian. American Speech 58:4,
pp. 303–318.
1984. Variability of the copula in Black English and its creole kin. American
Speech 59:4, pp. 291–309.
1987. Creole influence on Popular Brazilian Portuguese. In G. Gilbert, ed.
pp. 406–430.
1988–89. Pidgins and Creoles. 2 vols. Cambridge: Cambridge University Press.
1989. Do creoles have inflectional morphology? Paper presented at the Work-
shop on Creole Morphology, University of Amsterdam.
1991. The Atlantic creoles and the language of the ex-slave recordings. In
G. Bailey et al., eds., pp. 231–247.
1992a. A theoretical model for semi-creolization. Paper presented to the Soci-
ety for Caribbean Linguistics, Barbados.
1992b. Popular Brazilian Portuguese: a semi-creole. In E. d’Andrade and
A. Kihm, eds., pp. 37–66.
1994. The history of the English Language in the West Indies. In The Cambridge
History of the English Language, vol. V, R. Burchfield, ed., English in Britain
and Overseas: Origins and Development, pp. 328–381.
2000a. An Introduction to Pidgins and Creoles. Cambridge: Cambridge Univer-
sity Press.
2000b. Semi-creolization: problems in the development of theory. In I.
Neumann-Holzschuh and E. W. Schneider, eds., pp. 19–40.
2001. The semi-creole identity of Afrikaans lects: parallel cases of partial re-
structuring. Journal of Germanic Linguistics 13:4, pp. 353–379.
2002. The study of semi-creoles in the 21st century. In G. Gilbert, ed. Pidgin
and Creole Linguistics in the Twenty-first Century. Peter Lang: New York,
pp. 173–197.
Holm, J., G. A. Lorenzino, and H. R. de Mello. 1999. Diferentes grados
de reestructuración en dos lenguas vernáculas: el español caribeño y el
portugués brasileño. In Luı́s Ortiz, ed., El Caribe hispánico: perspectivas
lingüı́sticas actuales. Frankfurt: Vervuert, pp. 43–60.
Holm, John, and A. Goke Oyedeji. 1984. The Yoruba Language in the New
World. Oso 3:1, pp. 83–90.
Holm, John, and Peter Patrick, eds. forthcoming. Comparative Creole Syntax.
London: Battlebridge Press.
Holm, John, with Alison Shilling. 1982. Dictionary of Bahamian English. Cold
Spring, NY: Lexik House.
Honoré, Daniel. 2002. Dictionnaire d’Expressions Créoles–Semi-lo-mo. Réunion:
Editions Udir.
Howe, Darin M., and James A. Walker. 2000. Negation and the creole-origins
hypothesis: evidence from Early African American English. In S. Poplack,
ed., pp. 109–140.
Huber, Magnus. 2001. The Nova Scotia–Sierra Leone connection: new evidence
on an early variety of African American Vernacular English in the diaspora.
References 155
Paper presented at the meeting of the Society for Pidgin and Creole Linguis-
tics, Coimbra.
Huber, Magnus, and Mikael Parkvall, eds. 1999. Spreading the Word: the issue
of diffusion among the Atlantic Creoles. London: University of Westminster
Press.
Hull, Alexander. 1979. On the origin and chronology of the French-based creoles.
In I. Hancock et al., eds., Readings in Creole Studies. Ghent: Story-Scientia,
pp. 201–216.
1983. Créole louisianais et créole haı̈tien: ressemblances et différences.
Paper presented at the 4e Colloque International du Comité International
des Etudes Créoles, Lafayette.
Hymes, Dell, ed. 1971. Pidginization and Creolization of Languages. Cambridge:
Inverno, Liliana and Dominika Swolkien. 2001. O artigo definido zero em dois
contextos especı́ficos no Português do Brasil e no Crioulo de Cabo Verde.
Paper presented at the meeting of Associação: Crioulos de Base Lexical
Portuguesa e Espanhola, Coimbra.
Ivens Ferraz, L. 1979. The Creole of São Tomé. Johannesburg: Witwatersrand Uni-
versity Press.
Jeroslow, Helen [McKinney]. 1974. Rural cearense Portuguese: a study of one variety
of non-standard Brazilian Portuguese. Dissertation: Cornell University, Ithaca,
NY.
Jeuda, David M. 1991. The Portuguese element in the Papiamentu lexicon. Paper
presented to the Society for Pidgin and Creole Linguistics, Chicago.
Joyner, C. 1984. Down by the Riverside: a South Carolina slave community. Urbana,
Chicago: University of Illinois Press.
Kautzsch, Alexander. 2002. The Historical Evolution of Earlier African American
English: an empirical comparison of early sources. Berlin, New York: Mouton
de Gruyter.
Kautzsch, Alexander, and Edgar W. Schneider. 2000. Differential creolization:
some evidence from Earlier African American Vernacular English in South
Carolina. In I. Neumann-Holzschuh and E. W. Schneider, eds., pp. 247–
274.
Kaye, Alan S. 1990. Observations on pidginistics and creolistics. Semiotica 78:3/4,
pp. 285–348.
Keenan, E. L. 1975. Remarkable subjects. In C. N. Li, ed., Subject and Topic.
New York: Academic Press, pp. 249–301.
1978. The syntax of subject-final languages. In W. P. Lehman, ed., Syntactic
Typology: studies in the phenomenology of language. Austin: University of Texas
Press, pp. 266–327.
Kent, R. K. 1979. Palmares: an African state in Brazil. In R. Price, ed., Maroon
Societies: rebel slave communities in the Americas. Baltimore: Johns Hopkins
University Press, pp. 170–190.
Kloeke, G. G. 1950. Herkomst en groei van het Afrikaans. Leiden: Universitaire
Pers.
Kotzé, E. F. 1984. Afrikaans in the Malay Quarter: a diachronic perspective.
Tydskrif vir Geesteswetenskappe. 24:1, pp. 41–73.
156 References
1989. How creoloid can you be? Aspects of Malay Afrikaans. In M. Pütz and
R. Dirven, eds., pp. 251–264.
Kruisinga, E. 1906. De oorsprong van het Afrikaans. Taal en Letteren 16,
pp. 417–439.
Labov, William. 1969. Contraction, deletion and inherent variability of the
English copula. Language 45:4, pp. 715–751.
1972a. Negative attraction and negative concord. In Language in the Inner City:
Studies in the Black English Vernacular. Philadelphia: University of Pennsyl-
vania Press, pp. 130–196.
1972b. Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press.
1982. Objectivity and commitment in linguistic science: the case of the Black
English trial in Ann Arbor. Language in Society 11, pp. 165–201.
Labov, William, Paul Cohen, and John Lewis. 1968. A Study of the Non-standard
English of Negro and Puerto Rican Speakers in New York City. Report on Co-
operative Research Project 3288. New York: Columbia University.
Labov, William, and Wendell A. Harris. 1986. De facto segregation of Black and
White vernaculars. In David Sankoff, ed., Diversity and Diachrony. Amster-
dam, Philadelphia: John Benjamins.
Ladhams, John. 1999. The Pernambuco connection? An examination of the na-
ture and origin of the Portuguese elements in the Surinam Creoles. In M.
Huber and M. Parkvall, eds., pp. 209–240.
Lanehart, Sonja L. 2001. Sociocultural and Historical Contexts of African American
English. Amsterdam, Philadelphia: John Benjamins.
Laurence, Kemlin M. 1974. Is Caribbean Spanish a case of decreolization? Orbis
23:2, pp. 484–499.
Lass, R. 1990. How to do things with junk: exaptation in language evolution.
Journal of Linguistics 26, pp. 79–102.
Lemle, Miriam. 1976. Heterogenidade dialetal: um apelo a pesquisas. In
E. M. P. Lobato, ed., Lingüı́stica e ensino do vernáculo. Rio de Janeiro: Tempo
Brasileiro.
Lenz, Rodolofo. 1928. El papiamento, la lengua criolla de Curazao (la gramática
más sencilla). Santiago de Chile: Balcells.
Le Page, R. B. and A. Tabouret-Keller. 1985. Acts of Identity: creole-based
approaches to language and ethnicity. Cambridge: Cambridge University
Press.
Lipski, John M. 1993. On the non-creole basis for Afro-Caribbean Spanish. Al-
buquerque: Research Paper Series no. 24. Albuquerque: The University of
New Mexico.
1994. A new perspective on Afro-Dominican Spanish: the Haitian contri-
bution. Albuquerque: Latin American Institute of the University of New
Mexico.
1998. Perspectivas sobre el español bozal. In M. Perl and A. Schwegler, eds.,
pp. 293–328.
2000. Bozal Spanish: Restructuring or creolization? In I. Neumann-Holzschuh
and E. W. Schneider, eds., pp. 437–468.
Lockwood, W. B. 1965. An Informal History of the German Language, with chapters
on Dutch and Afrikaans, Frisian and Yiddish. Cambridge: Heffer.
References 157
Lorenzino, Gerardo. 1991. On the semi-creole status of Dominican Spanish. City

University of New York, Dept. of Linguistics, 41-p. MS.
1993. Algunos rasgos semicriollos en el español popular dominicano. Anuario
de Lingüı́stica Hispánica 9, pp. 111–126.
1998. El español caribeño: antecedentes sociohistóricos y lingüı́sticos. In
M. Perl and A. Schwegler, eds., pp. 26–39.
Forthcoming. Angolar. In J. Holm and P. Patrick, eds.
Lucchesi, Dante. 2000. A variação na concordância de gênero em uma comunidade
de fala afro-brasileira: Novos elementos sobre a formação do português popular do
Brasil. Ph.D. diss., Universidade Federal do Rio de Janeiro.
Luijks, Carla. 2000. The Realisation of Syntactic Principles in Non-standard
Afrikaans: the Correspondence of Jan Jonker Afrikaner (1820–1889). Ph.D.
thesis, University of Cape Town.
Makhudu, Dennis. 1984. Is Afrikaans a creole language? M.A. thesis, Depart-
ment of Linguistics, Southern Illinois University at Carbondale.
Markey, Thomas. 1982. Afrikaans: creole or non-creole? Zeitschrift für Dialektolo-
gie und Linguistik 49:2, pp. 169–207.
Marques, A. H. de Oliveira. 1976. History of Portugal. New York: Columbia Uni-
versity Press.
Marroquim, Mário. 1934. A lı́ngua do Nordeste (Alagôas e Pernambuco). São Paulo:
Companhia Editora Nacional.
Maurer, Philippe. 1988. Les modifications temporelles et modales du verbe dans le pa-
piamento de Curaçao (Antilles Néerlandaises). Hamburg: Helmut Buske Verlag.
1995. L’angolar: un créole afro-portugais parlé à São Tomé. Hamburg: Helmut
Buske Verlag.
McKay, J. R. 1969. A Partial Analysis of a Variety of Nonstandard Negro English.
Ph.D. dissertation, University of California, Berkeley.
McKinney, Helen [Jeroslow]. 1975. Creole characteristics in rural Brazilian
Portuguese. Paper presented at the International Congress on Pidgins and
Creoles, University of Hawaii.
1982. Serial verbs in a non-standard Brazilian dialect: possible creolization.
Paper presented at the Conference of the American Association of Teachers
of Spanish and Portuguese, New Orleans.
McWhorter, John. 1998. Identifying the creole prototype: vindicating a typolog-
ical class. Language 74:4, pp. 788–818.
2000. Defining “creole” as a synchronic term. In I. Neumann-Holzschuh and
E. W. Schneider, eds., pp. 85–123.
ed., 2000. Language Change and Language Contact in Pidgins and Creoles.
Amsterdam, Philadelphia: John Benjamins.
2002. The rest of the story: restoring pidginization to creole genesis theory.
Journal of Pidgin and Creole Languages 17:1, pp. 1–48.
Megenney, William W. 1976. El elemento sub-sahárico en el léxico costeño de
Colombia. Revista Española de Lingüı́stica 6, pp. 405–451.
1985. Africa en Venezuela: su herencia lingüı́stica y su cultura literaria. Revista
Montalbán 15, pp. 1–54.
1990. Africa en Santo Domingo: su herencia lingüı́stica. Santo Domingo: Editorial
Tiempo.
158 References
1999. Cuba y Brasil: etnohistória del empleo religioso del lenguaje afroamericano.
Miami FL: Ediciones Universal.
Mello, Heliana Ribeiro de. 1997. Brazilian Vernacular Portuguese: evidence of partial
restructuring. Ph.D. dissertation, City University of New York. Ann Arbor:
University Microfilms International.
2001. Situando o Português Brasileiro através da perspectiva da análise crı́tica
do discurso. Paper presented at the meeting of Associação: Crioulos de Base
Lexical Portuguesa e Espanhola, Coimbra.
Melo, Gladstone Chaves de. 1946. A Lı́ngua do Brasil. Rio de Janeiro: Livraria
AGIR Editora. Repr. 1975. Rio de Janeiro: Editora da Fundação Getúlio
Vargas.
Mendonça, Renato. 1933. A influência africana no português do Brasil. Repr. 1973.
Rio de Janeiro: Civilização Brasileira/MEC.
Mesthrie, Rajend, and Paul T. Roberge, eds., 2001–02. Focus on Afrikaans
Sociohistorical Linguistics. Journal of Germanic Linguistics 13:4–14:1.
Michel, Abigail. Forthcoming. Papiamentu. In J. Holm and P. Patrick, eds.
Mintz, S. W. 1971. The socio-historical background to pidginization and cre-
olization. In D. Hymes, ed., pp. 481–498.
Moñino, Yves and Armin Schwegler, eds., 2002. Palenque, Cartagena y Afro-
Caribe: histórica y langua. Tübingen: Max Niemeyer.
Montgomery, M., and G. Bailey, eds. 1986. Language Variety in the South: per-
spectives in black and white. University, AL: University of Alabama Press.
Mufwene, Salikoko. 1983. Some observations on the verb in Black English
Vernacular. African and Afro-American Studies and Research Center Papers,
series 2. University of Texas at Austin.
1987. Review of Montgomery and Bailey, eds., 1986. Journal of Pidgin and
Creole Languages 2:1, pp. 93–110.
1991. Review of Holm 1988–89. Language 67, pp. 380–387.
1994. On decreolization: the case of Gullah. In M. Morgan, ed., Language
and the Social Construction of Identity in Creole Situations. Center for African
Studies, UCLA, Los Angeles.
1996. The founder principle in creole genesis. Diachronica 13, pp. 83–134.
1997. Jargons, pidgins, creoles, and koines: what are they? In A. Spears and
D. Winford, eds., pp. 35–70.
1998. The structure of the noun phrase in AAVE. In S. Mufwene et al. eds.,
pp. 69–81.
2000a. Population contacts and the evolution of English. The European English
Messenger 9, pp. 9–15.
2000b. Some sociohistorical inferences about the development of African
American English. In S. Poplack, ed., pp. 233–263.
2000c. Creolization is a social, not a structural, process. In I. Neumann-
Holzschuh and E. W. Schneider, eds., pp. 65–84.
2001. African-American English. In John Algeo, ed., Cambridge History of the
English Language. Vol. 6: English in North America.
Mufwene, Salikoko S., John R. Rickford, Guy Bailey, and John Baugh, eds.,
1998. African-American English: structure, history and use. London, New York:
Routledge.
References 159
Mühlhäusler, Peter. 1982. Tok Pisin in Papua New Guinea. In R. Bailey and
M. Görlach, eds., English as a World Language. Ann Arbor: University of
Michigan Press, pp. 439–466.
Mussa, A. B. N. 1991. O papel das lı́nguas africanas na história do português
do Brasil. Unpublished master’s thesis, Universidade Federal do Rio de
Janeiro.
Muysken, Pieter, ed. 1981. Generative Studies on Creole Languages. Dordrecht:
Foris.
Naro, Anthony J., and Miriam Lemle. 1976. Syntactic diffusion. In S. B. Steever,
C. A. Walker, and S. S. Mufwene, eds., Papers from the Parasession on Di-
achronic Syntax, Chicago: Chicago Linguistic Society, pp. 221–240.
Naro, Anthony J., and M. Marta P. Scherre. 2000. Variable Concord in
Portuguese: the situation in Brazil and Portugal. In J. McWhorter, ed.,
pp. 235–255.
Nash, Roy. 1926. The Conquest of Brazil. New York: Harcourt, Brace.
Neeser, Hermann. 1951. A Colônia Leopoldina. Bahia: Centro de Estudos Ba-
ianos, Publicação no. 5.
Neumann-Holzschuh, Ingrid, and Edgar W. Schneider, eds. 2000. Degrees of
Restructuring in Creole Languages. Amsterdam, Philadelphia: John Benjamins.
Nienaber, P. J. 1965. Iets naders oor die ontkenning in Afrikaans. In P. J. Nien-
aber, ed., Taalkundige opstelle. Cape Town: Balkema, pp. 22–38.
Nuñez Cedeño, Rafael. 1980. La fonologı́a moderna y el español de Santo Domingo.
Santo Domingo: Taller.
Oldendorp, C. G. A. 1777. Geschichte der Mission der evangelischen Brüder auf den
caraibischen Inseln S. Thomas, S. Croix und S. Jan. Barby: C. F. Laur.
Ortiz, Fernando. 1924. Glosario de afronegrismos. Havana: El Siglo XX.
Ortiz López, Luis A. 1996. El contacto lingüı́stico afro-hispánico en la génesis y
evolución del español de Cuba. Ph.D. dissertation, University of New Mexico,
Albuquerque.
1998. Huellas etno-sociolingüı́sticas bozales y afrocubanas. Frankfurt-am-Main:
Vervuert/Madrid: Iberoamericana.
ed. 1999. El Caribe hispánico: perspectivas lingüı́sticas actuales. Homenaje a
Manuel Álvarez Nazario. Frankfurt-am-Main: Vervuert/Madrid: Iberoamer-
icana.
Orton, H., S. Sanderson, and J. Widdowson. 1978. The Linguistic Atlas of England.
London: Croom Helm.
Otheguy, Ricardo. 1973. Spanish Caribbean: a creole perspective. In C. Bailey
and R. Shuy, eds., New Ways of Analyzing Variation in English. Washington:
Georgetown University Press, pp. 323–339.
Papen, Robert A. 1978. The French-based Creoles of the Indian Ocean: an analysis
and comparison. Ph.D. dissertation, University of California at San Diego.
Parkvall, Mikael. 2000. Reassessing the role of demographics in language re-
structuring. In I. Neumann-Holzschuh and E. W. Schneider, eds., pp. 185–
213.
2002. Cutting off the branch. In G. Gilbert, ed., pp. 355–367.
Peixoto, Antônio da Costa. 1945. Obra nova de Lı́ngua Geral de Mina. Lisbon:
Agência Geral das Colônias.
160 References
Perl, Matthias. 1985. El fenómeno de descriollización del “habla bozal” y el

lenguaje coloquial de la variante cubana del español. Anuario de Lingüı́stica
Hispánica 1, pp. 191–202.
1988. Rasgos poscriollos léxicos en el lenguaje coloquial cubano. Thesaurus 43,
47–64.
1989. El “habla bozal”: ¿una lengua criolla de base española? Anuario de
Lingüı́stica Hispánica 5, pp. 205–220.
Perl, Matthias, and Armin Schwegler, eds. 1998. América negra: panorámica ac-
tual de los estudios lingüı́sticos sobre variedades hı́spanas, portuguesas y criollas.
Frankfurt: Vervuert/ Madrid: Iberoamericana.
Pichardo, Esteban. 1862. Diccionario provincial casi razonado de voces cubanas.
Havana: Imprenta la Antilla.
Pimentel Pinto, Edith, ed. 1981. O português do Brasil: textos crı́ticos e teóricos. Rio
de Janeiro: Livros Técnicos e Cientı́ficos.
Ponelis, F. A. 1993. The Development of Afrikaans. (Duisburger Arbeiten zur
Sprach- und Kulturwissenschaft, 18). Frankfurt-am-Main: Peter Lang.
Poplack, Shana. 1978. On Deletion and Disambiguation in Puerto Rican Spanish: a
study of verbal /n#/. New York: Centro de Estudios Puertorriqueños.
2000. Introduction. In S. Poplack, ed., pp. 1–32.
Poplack, Shana, ed. 2000. The English History of African American English. Oxford:
Blackwell.
Poplack, Shana, and David Sankoff. 1987. The Philadelphia story in the Spanish
Caribbean. American Speech 62:4, pp. 291–314.
Poplack, Shana, Sali Tagliamonte, and Ejike Eze. 2000. Reconstructing the
source of early African American English plural marking: a comparative study
of English and Creole. In S. Poplack, ed., pp. 73–105.
Pütz, Martin, and René Dirven, eds. 1989. Wheels within Wheels: Papers of the
Duisburg Symposium on Pidgin and Creole Languages. Frankfurt-am-Main:
Peter Lang.
Rademeyer, J. H. 1938. Kleurling-Afrikaans: Die taal van die Griekwas en Rehoboth-
Basters. Amsterdam: Swets & Zeitlinger.
Raidt, Edith. 1974. Nederlandse en Kaapse spreektaal in die 17de en 18de eeu.
In F. Odendal, ed., Taalkunde–’n lewe. Studies opgedra aan W. Kempen. Cape
Town: Tafelberg, pp. 90–104.
1983. Einführung in Geschichte und Struktur des Afrikaans. Darmstadt: Wis-
senschaftliche Buchgesellschaft.
1991. Afrikaans en sy europese verlede. Cape Town: Nasou.
Raimundo, Jacques. 1933. O elemento afro-negro na lı́ngua portuguesa. Rio de
Janeiro: Renascença Editora.
Ramassamy, Ginette. 1985. Syntaxe du créole réunionnais: analyse de corpus
d’unilingues créolophones. Thése pour le Doctorat de 3ème cycle, Université
René Descartes, Paris V.
Rawley, J. A. 1981. The Transatlantic Slave Trade: a history. New York, London:
Norton.
Reinecke, John. 1937. Marginal Languages: a sociological survey of the creole lan-
guages and trade jargons. Ph.D. dissertation, Yale University. Ann Arbor: Uni-
versity Microfilms International.
References 161
Reinecke, J. E., S. M. Tsuzaki, D. DeCamp, I. F. Hancock, and R. E. Wood, eds.,

1975. A Bibliography of Pidgin and Creole Languages. Honolulu: University
Press of Hawaii.
Rens, L. L. E. 1953. The Historical and Social Background of Surinam’s Negro
English. Academisch proefschrift. Amsterdam: North Holland.
Révah, Israel S. 1963. La question des substrats et superstrats dans le domaine lin-
guistique brésilien: les parler populaires brésiliens doivent-ils être considerés
comme des parlers “créoles” ou “semi-créoles”? Romania 84, pp. 433–450.
Rickford, John R. 1977. The question of prior creolization in Black English. In A.
Valdman, ed., Pidgin and Creole Linguistics. Bloomington: Indiana University
Press, pp. 190–221.
1986. Social contact and linguistic diffusion: Hiberno English and New World
Black English. Language 62:2, pp. 245–289.
1987. Dimensions of a Creole Continuum. Stanford: Stanford University Press.
1997. Prior creolization of African-American Vernacular English? Sociohis-
torical and textual evidence from the 17th and 18th centuries. Journal of
Sociolinguistics 1:3, pp. 315–336.
1999. African American Vernacular English: Features, Evolution, Educational
Implications. Oxford: Blackwell.
Roberge, Paul T. 1994. The formation of Afrikaans. Stellenbosch Papers in
Linguistic PLUS, 23. Department of General Linguistics, University of
Stellenbosch.
Roberts, Ian. 1999. Verb movement and markedness. In M. DeGraff, ed.,
pp. 287–327.
Rodrigues, Ada Natal. 1974. O Dialeto Caipira na Região de Piracicaba. São Paulo:
Editora Ática.
Rowlands, E. C. 1969. Yoruba. London: The English Universities Press.
Ruı́z Garcı́a, Martha E. 2001. El español popular de Chocó, Colombia: evidencia
de una reestructuración parcial. Ph.D. dissertation. Ann Arbor: University
Microfilm International.
Sampaio, Theodoro. 1928. O Tupı́ na geografia nacional. Baı́a: Secção da Escola
de Aprendizes.
Sandoval, Padre A. de. 1627. De instauranda Aethiopum salute. Seville: F. de Lira.
Repr. 1956, Bogotá: Empresa Nacional de Publicaciones.
Schneider, Edgar W. l981. Morphologische und syntaktische Variablen im amerikani-
schen Early Black English. Frankfurt am Main: Peter Lang.
1989. American Earlier Black English: morphological and syntactic variables.
Tuscaloosa: University of Alabama Press.
1990. The cline of creoleness in English-oriented creoles and semi-creoles in
the Caribbean. English World-Wide 11:1, pp. 79–113.
Scholtz, J. du Plessis. 1963. Taalhistoriese opstelle. Pretoria: J. L. van Schaik.
1970. Internal History of Afrikaans. Standard Encyclopedia of Southern Africa.
Cape Town: Nasou, pp. 80–113.
1980. Wording en ontwikkeling van Afrikaans. Cape Town: Tafelberg.
Schuchardt, Hugo. 1882. Kreolische Studien I: Ueber das Negerportugiesische
von S. Thomé (Westafrika). Sitzungsberichte der kaiserlichen Akademie der
Wissenschaften zu Wien 101:2, pp. 889–917.
162 References
1885. Review of Focard’s Du patois créole de l’ı̂le de Réunion. Literaturblatt für

germanische und romanische Philologie 6, pp. 513–515. Translated into English
as Further remarks on Réunionnais 1979. In T. L. Markey, ed. and trans.,
The Ethnography of Variation: selected writings on pidgins and creoles. Ann Arbor:
Karoma, pp. 15–17.
1889. Beiträge zur Kenntnis des kreolischen Romanisch V. Allgemeineres
über das Indoportugiesische (Asioportugiesische). Zeitschrift für romanische
Philologie 13, pp. 476–516.
Schwegler, Armin. 1991. Negation in Palenquero: synchrony. Journal of Pidgin
and Creole Languages 6, pp. 165–214.
1993. Subject pronouns and person/number in Palenquero. In F. Byrne and
J. Holm, eds., pp. 145–161.
1996a. Chi ma nkongo: lengua y rito ancestrales en al Palenque de San Basilio
(Colombia). Frankfurt: Vervuert.
1996b. La doble negación dominicana y la génesis del español caribeño. His-
panic Linguistics 8:2, pp. 247–315.
Schwegler, Armin, and Thomas Morton. 2003. Vernacular Spanish in a micro-
cosm: Kateyano in El Palenque de San Basilio (Colombia). Revista Interna-
cional de Lingüı́stica Iberoamericana 1:1, pp. 97–159.
Sells, Peter, John R. Rickford, and Thomas A. Wasow. 1996. Negative inversion
in African American Vernacular English. Natural Language and Linguistic
Theory 14, pp. 591–627.
Siegel, Jeff. 1997. Mixing, leveling, and pidgin/creole development. In A. Spears
and D. Winford, eds., pp. 111–149.
Silva Neto, Serafim da. 1950a. Falares Crioulos. Brası́lia 5, pp. 1–28.
1950b. [2nd edn. 1963]. Introdução ao estudo da lı́ngua portuguesa no Brasil. Rio
de Janeiro: Departamento de Imprensa Nacional.
Silveira Bueno, F. 1963. Les langues indigènes du Brésil et leur influence sur le
portugais. Orbis 12, pp. 226–240.
Silveira Ferreira, Carlota da. 1985. Remanescentes de um falar crioulo brasiliero.
Revista Lusitana 5 (Lisbon), pp. 21–34.
Singler, John V. 1986. Remarks in response to Derek Bickerton’s “Creoles and
universal grammar: the unmarked case?” Journal of Pidgin and Creole Lan-
guages 1:1, pp. 141–145.
1989. Plural marking in Liberian Settler English. American Speech 64:1,
pp. 40–64.
1990. On the use of sociohistorical criteria in the comparison of creoles. Lin-
guistics 28:4, pp. 645–660.
1995. The demographics of creole genesis in the Caribbean: a comparison of
Martinique and Haiti. In J. Arends, ed., pp. 203–232.
Slomanson, Peter. 1993. On a problem in the grammar of Orange River
Afrikaans. Paper presented to the Society for Pidgin and Creole Linguistics,
Amsterdam.
Smith, Johannes J. 1927. The evolution and recognition of the Afrikaans language.
Official Year Book of the Union 8 (1925), pp. 14–23.
1952. Theories about the origin of Afrikaans. Johannesburg: Witwatersrand
University Press.
References 163
Smith, Norval. 1999. Pernambuco to Surinam 1654–1655? The Jewish slave

controversy. In M. Huber and M. Parkvall, eds., pp. 251–298.
2002. The history of the Surinamese creoles II. In E. B. Carlin and J. Arends,
eds., pp. 131–151.
Smith, Norval, and Tonjes Veenstra, eds., 2001. Creolization and Contact.
Amsterdam/ Philadelphia: John Benjamins.
Smith, Riley B. 1973. Interrelatedness of certain deviant grammatical structures
in Negro Nonstandard Dialects. In R. H. Bentley and S. D. Crawford, eds.,
Black Language Reader. Glenview, Ill. and Brighton, England: Scott, Fores-
man and Company, pp. 90–96.
Spears, Arthur K., and Donald Winford, eds. 1997. Pidgins and Creoles: structure
and status. Amsterdam Philadelphia: John Benjamins.
Stein, P. 1984. Kreolisch und Französisch. Tübingen: Niemeyer.
Stewart, William A. 1962. Creole languages in the Caribbean. In F. A. Rice,
ed., Study of the Role of Second Languages in Asia, Africa and Latin America.
Washington, D.C.: Center for Applied Linguistics, pp. 34–53.
1965. Urban Negro speech: sociolinguistic factors affecting English teaching.
In R.W. Shuy, ed., Social Dialects and Language Learning. Champaign, Ill.:
National Council of Teachers of English, pp. 10–18.
1967. Sociolinguistic factors in the history of American Negro dialects. Florida
FL Reporter 5. Repr. in Wolfram and Clarke, eds., 1971, pp. 74–89.
1968. Continuity and change in American Negro dialects. Florida FL Reporter
6. Repr. in Wolfram and Clarke, eds., 1971, pp. 51–73.
Tagliavini, Carlo. 1931. Creole, Lingue. Enciclopedia Italiana di scienze, lettere ed
arti. Milan: Istituto Giovanni Treccani. Vol. 11, pp. 833–835.
Tarallo, Fernando. 1986. On the alleged creole origin of Brazilian Portuguese:
untargeted syntactic change. Paper presented at the Creole Workshop, LSA
Institute, City University of New York. In Ian Roberts and Mary A. Kato,
eds., 1993. Português brasileiro: uma viagem diacrônica. Campinas: Editora da
UNICAMP, pp. 35–68.
Terrel, T. 1979. Final /s/ in Cuban Spanish. Hispania 62, pp. 599–612.
Thomason, Sarah G. 1997. A typology of contact languages. In A. Spears and
D. Winford, eds., pp. 71–88.
2001. Contact-induced language change and Pidgin/Creole genesis. In N.
Smith and T. Veenstra, eds., pp. 249–262.
2002. Column: Creoles and genetic relationship. Journal of Pidgin and Creole
Languages 17:1, pp. 101–109.
Thomason, S. G., and T. Kaufman. 1988. Language Contact, Creolization and
Genetic Linguistics. Berkeley: University of California Press.
Tölsner, Carl A. 1858. Die Colonia Leopoldina in Brasilien. Göttingen: Dieterich-
sche Universitäts-Buchdruckerei.
Tottie, Gunnel, and Dawn Harvie. 2000. It’s all relative: relativization strate-
gies in Early African American English. In S. Poplack, ed., pp. 198–
230.
Turner, L. D. 1949. Africanisms in the Gullah dialect. Repr. 1974. Ann Arbor:
University of Michigan Press.
Valdman, Albert. 1978. Le créole: structure, statut et origine. Paris: Klincksieck.
164 References
Valkhoff, Marius. 1964. Notes socio-linguistiques sur le parler créole de la

Réunion. Revue Philologique 17, pp. 723–735.
1966. Studies in Portuguese and Creole with special reference to South Africa.
Johannesburg: Witwatersrand University Press.
1972. New light on Afrikaans and Malayo-Portuguese. Louvain: Peeters.
van Dantzig, A. 1968. Het Nederlandse aandeel in de slavenhandel. Bussem: Van
Dishoek.
van der Merwe, H. J. J. M. 1963. Die ontstaan van Afrikaans. In H. J. J. M.
van der Merwe, ed., Inleiding tot die Taalkunde. Pretoria: J. L. van Schaik,
pp. 67–88.
1968. Die ontstaan van Afrikaans. In H. J. J. M. van der Merwe, ed., Afrikaans–
sy Aard en Ontwikkeling. Pretoria: J. L. van Schaik, pp. 15–66.
van der Merwe, Elizabeth G. 1989. Die ontwikkeling, stand en toekoms van
Afrikaans in S.W.A./Namibië. Ph.D. diss., Potchefstroomse Universiteit.
1990. Oorlamafrikaans. Paper presented to the Linguistic Society of Southern
Africa, University of Stellenbosch.
Van Name, Addison. 1869–70. Contributions to creole grammar. Transactions of
the American Philological Association 1, pp. 123–167.
van Rensburg, M. C. J. 1984. Oranjerivier-Afrikaans. In M. C. J. van Rensburg,
ed., Die Afrikaans van die Griekwas van die tagtigerjare. Bloemfontein: Uni-
versity of the Orange Free State, 1:2, pp. 513–518.
1989. Orange River Afrikaans: a stage in the pidgin/creole cycle. In M. Pütz
and R. Dirven, eds., pp. 135–151.
Vasconcellos, José Leite de. 1901. Esquisse d’une dialectologie portugaise. Paris:
Aillaud.
Vaxelaire, Daniel. 1992. Vingt-et-un jours d’histoire: Ile de La Réunion. Saint-Denis,
Réunion: Azalées Editions.
Viljoen, W. J. 1896. Beiträge zur Geschichte der cap-holländischen Sprache. Strass-
burg: K. Trübner.
Vintilă-Rădulescu, I. 1976. Le créole français. The Hague: Mouton.
Voorhoeve, Jan. 1962. Sranan Syntax. Amsterdam: North Holland.
1973. Historical and linguistic evidence in favor of a relexification theory in
the formation of creoles. Language in Society 2, pp. 133–145.
Wagner, Max L. 1949. Lingua e dialetti dell’America spagnola. Florence: Le
Lingue Estere.
Walker, James A. 2000. Rephrasing the copula: contraction and zero in Early
African American English. In S. Poplack, ed., pp. 35–72.
Warantz, Elissa. 1983. The Bay Islands English of Honduras. In Holm, ed.,
pp. 71–94.
Washabaugh, William. 1983. Creoles of the off-shore islands: Providencia, San
Andrés, and the Caymans. In Holm, ed., pp. 157–179.
Watbled, Jean-Philippe. 2002. Le système verbal du créole réunionnais. Paper
presented at the Colloque International des Études Créoles, Réunion.
Forthcoming. Grammaire créole et grammaire français. In Gudrun Ledegen,
ed., Anciens et nouveaux plurilinguismes: Les actes de la 6e table ronde du Moufia.
Paris: L’Harmattan.
References 165
Webb, V. N. 1989. The presumed pidgin origins of Orange River Afrikaans.

Paper presented at the 18th conference on New Ways of Analyzing Variation
in English, Duke University.
Wentworth, Harold. 1944. American Dialect Dictionary. New York: Crowell.
Whatley, Elizabeth. 1981. Language among Black Americans. In C. A. Ferguson
and S. B. Heath, eds., pp. 92–110.
Williams, Jeffery P. 1985. Preliminaries to the study of the dialects of White West
Indian English. Nieuwe West-Indische Gids 59, pp. 27–44.
Winford, Donald. 1997. On the origins of African American Vernacular English –
A creolist perspective, Part I: The sociolinguistic background. Diachronica
14:2, pp. 305–344.
1998. On the origins of African American Vernacular English – A creolist
perspective, Part II: Linguistic Features. Diachronica 15:1, pp. 99–154.
2000. “Intermediate” creoles and degrees of change in creole formation:
the case of Bajan. In I. Neumann-Holzschuh and E. W. Schneider, eds.,
pp. 215–246.
Wolfram, Walt. 1969. A Sociolinguistic Description of Detroit Negro Speech. Wash-
ington, D.C.: Center for Applied Linguistics.
1974. The relationship of white southern speech to vernacular black English.
Language 30, pp. 498–527.
Forthcoming. Urban AAVE grammar. In Edgar W. Schneider, ed., English in
the Americas and the Caribbean. Varieties of English Handbook, Vol. 1. Berlin
etc.: Mouton de Gruyter.
Wolfram, W., and N. H. Clarke, eds. 1971. Black–White Speech relationships. Wash-
ington D.C.: Center for Applied Linguistics.
Wood, Peter H. 1974. Black Majority: Negroes in colonial South Carolina from 1670
through the Stono Rebellion. New York: Knopf.
1989. The changing population of the colonial South: an overview by race and
religion, 1685–1790. In P. H. Wood et al., eds., Powhatan’s Mantle: Indians
in the colonial Southeast. Lincoln: University of Nebraska Press, pp. 35–103.
Wullschlägel, H. R. 1856. Deutsch-Negerenglisches Wörterbuch. Repr. 1965.
Amsterdam: S. Emmering.
Yillah, Sorie. Forthcoming. Krio. In J. Holm and P. Patrick, eds.
Ziegler, Douglas V. 1976. A Preliminary Study of Afro-Cuban Creole. M.A.
thesis, University of Rochester.
1977. Some Notes on Puerto Rican Creole Spanish. MS.
Zimmermann, Klaus, ed. 1999. Lenguas criollas de base lexical española y
portuguesa. Frankfurt-am-Main: Vervuert/Madrid: Iberoamericana.
Index
a- (VLRF pronoun prefix), 111–112 Afrikaans, 2, 7, 8, 9, 13, 23

abbreviations, xix–xx, 138 auxiliaries, 79
abolition of slavery “Coloured” Afrikaans, see Kaaps
in Brazil, 56 dependent clauses, 122
in British Empire, 45 East Cape versus West Cape, 46
in Réunion, 68 Flytaal, 14
in USA, 36, 38 gender in NP, 98
Abstand, 80 Griqua Afrikaans, 14, 46, 100, 101
acrolect, 7, 69, 86 historical development, 41–47, 71, 137
Adam, L., 19 Kaaps, 14, 23, 46, 79, 98
African American English, 8, Malay Afrikaans, 14, 47, 101
28 negation, 79–80
attitudes towards, 1 number in NP, 97–98
auxiliary verbs, 73–75 Orange River Afrikaans, 14, 23, 46–47,
“creolist” theory, 10, 12 79, 98, 100, 137
decreolization of, 6 possession, 98–99
dependent clauses, 120 pronouns, 99–101
diaspora varieties, 37 standardization of, 46
ex-slave narratives, 37 study of, 13–15
historical development, 29–41, 71, verbal morphology, 77–79
135, 137 word order, 120–122
negation, 75–76 Afro-Iberian linguistics, 16, 18
non-verbal predicates, 76–77 Afro-Portuguese, 62
number in NP, 93–94 Afro-Seminole, 33
possession, 95 age-grading, 40, 95, 96
pronouns agentive subject marker, 86–87, 88, 111,
study of, 4, 10–13 141
varieties, 33–34 agreement
verbal morphology, 73 gender, 101, 103–104, 108–110, 139
word order, 116–117 number, 103, 106–107, 108
African languages in Brazil, 53–56 subject-verb, 73, 88
Bantu versus West African, 54 agriculture, see plantation slavery
Ewe-Fon, 55 ain’t (AAE ‘didn’t’), 40, 75–76
Yoruba, 56 Alvarez Nazario, M., 18
Africans (see also West Africans) ambiguity, 126, 141
in American South, 36, 47 American Indian English, 3–4, 21
in Brazil, 47, 48–49, 50, 55 American Indians
in Cape Colony, 45 in Brazil, 50
in Réunion, 67, 137 in South Carolina, 35
in South Carolina, 35, 50 in Virginia, 32
in Spanish Caribbean, 61–62 American South, 33, 39, 41, 137
in Virginia, 32 analyticity, 5
166
Index 167
Andalusian Spanish, 17, 83 negation, 82

Angola, 42, 54 non-verbal predicates, 83
Angolar CP, 126 number in NP, 101
animate, 93, 98, 99 possession, 104
anterior tense, 73 pronouns, 104–105
Arawakan languages, 60 study of, 15–17
archaic features, 10, 101 verbal morphology, 80–82
articles, see determiners British Caribbean, 28, 29, 32, 35
aspect, 72, 86 British colonists, 45
assimilation, 41, 42, 43, 61, 62 British dialects, 1, 10, 77, 119, 139
associative plural, 92, 94, 97, 98, 108, 141 British North America, 33, 47
Atlantic creoles, xiv, 5, 81, 103, 107, 111,
116, 121, 124, 139 Cabrera, L., 18
attrition, xiv Cajun French, 23
Australian Aboriginal English, 4 camouflaging, 134
auxiliary verbs, 72, 120 Cape Colony, 41–46, 137
comparison, 91 population, 45
influence of preverbal markers on, 72, Cape Coloured, 43, 44, 46
73–75, 79, 82, 84, 85, 87–88 Cape Dutch, see Afrikaans
Cape Malays, 42, 43, 47
Bahamian (C)E, xii, 10, 37, 117 Cape of Good Hope, 41–46
link to Gullah, 13, 34, 35, 37 Cape Town, 43
of whites, 11 Cape Verde CP, 26, 27, 47, 57, 102,
Bahia, 50, 55, 56, 59 127, 137
Bajan see Barbados CE Caribbean Spanish, see Non-standard
Baker, P., 20 Caribbean Spanish
Bambara, 105 Carolinas, 30, 34
bandeirantes, 48, 55 case marking, 92, 95, 99, 101
Bantu languages, 44 caste dialects, 7, 21
Barbados, 29–32, 34, 35 Caymanian English, 9, 11, 72
CE, 26, 27, 136 Central American (C)E, 11
basilect, 3, 5, 7, 9, 23, 30, 64, 69, 72, 86 Central American Spanish, 60
Baxter, A., 16 Ceylon, see Sri Lanka
Bay Island English, 11 Chapuis, D., 19, 21
be (AAE habitual), 40, 41, 75 Chaudenson, R., xv, 20
be done (AAE future perfect), 75 Chocó, 106
been (AAE remote past), 40, 73–74 civil rights movement, 1
Bermuda, 26, 34 Civil War (in USA), 36, 39
Bickerton, D., 9, 24, 27 clause structure
bidialectalism, 58 in AAE, 116–120
bioprogram, 25, 28 in Afrikaans, 120–123
Black English, 1 in BVP, 123–127
Boer War, 46 in NSCS, 127–129
Bollée, A., 20 in VLRF, 129–133
Boomfield, L., 7, 9 coartación, 62
borrowing, 137, 143 code switching, xiv
Bourbonnais, 20, 65 Coelho, A., 15
bozal Spanish, see Habla Bozal cognition, xv
Brazilian Vernacular Portuguese, 2, 8, Colombian Spanish, 17, 18, 60, 62, 85,
18 106
auxiliaries, 82 Colônia Leopoldina, 59
dependent clauses, 124 Coloured, see Cape Coloured
gender in NP, 103–104 comparative method, xiv
historical development, 26, 47–60, 71, comparative studies in restructuring,
137 144–146
comparison Creole Spanish (see also Palenquero,

of auxiliaries/preverbal markers, 91 Papiamentu)
of dependent clauses, 133–134 in Cuba? 63
of gender, 113–114 pan-Caribbean? 18
of historical development, 70–71 creolistics, xi
of negation, 91 creolisant, 6, 8
of number marking in NP, 112–113 creolization, xi, xiv, 3
of partially restructured vernaculars, full versus partial, 4–5, 12, 24, 134, 135,
21–23 141, 145
of possession, 114 crops, see plantation slavery
of pronouns, 114 Cuban Spanish, 8, 17, 18, 60, 62, 63, 71,
of verbal morphology, 90–91 84, 106
of word order, 133 Curaçao, 51, 62, 71
completive aspect, 74, 79
complexification, 141 de (CE progressive marker), 74
consonant cluster simplification, 97 “debasilectalization”, 38
contact linguistics, 3, 145, 146 DeCamp, D., 7
continuum, 7, 20, 85 decreolization, 2–7, 9, 10, 11, 16, 18,
convergence/divergence debate, 11, 40 20, 38, 60, 74, 75, 93, 102, 103,
convergence of influence, 5, 118 126, 143
copula, 41, 87 versus partial restructuring, 4, 11
absence of (see also non-verbal definite articles, 93
predicates), 23, 40, 41, 74, 90, 139, definite future, 88
141 definiteness, 93
comparison, 91 DeGraff, M., xiv
in AAE, 76–77 de Granda, G., 17, 18
in BVP, 83 degrees of restructuring, see gradience
in NSCS, 85 dem, see pluralizer
in VLRF, 89–90 demographic ratio, see native-speaker
copula-like particle, 87 ratio
Corne, C., 20 demonstratives, 93
Couto, H., 16 denasalization, 80
Craig, B., 21 den Besten, H., 14
creole, xiii depalatalization, 69
influence, 9, 11, 21 dependent clauses
prototype, xv, 28 comparison, 133–134
versus non-creole features, 22–23 in AAE, 120
Créole des Bas, see VLRF in Afrikaans, 122
Créole des Hauts, see VLRF in BVP, 124
Creole English in NSCS, 128
in Caribbean, 2, 7, 8, 41 in VLRF, 129–133
in North America, 30, 32, 33, 34–35, 41 derivational affixes, xv
in Suriname, see Surinamese creoles destativizer, 87
Creole French determiners
in Caribbean, xvi, 6 articles, 109
in Louisiana, 26, 52 absent, 53
in Indian Ocean, see Mauritian, gender agreement, 108–110
Seychelles CF number agreement, 101
Creole Portuguese detribalization, 44
in Africa, 47 diaspora varieties of AAE, 37
in Asia, 42 Dillard, J., 2, 36
in Brazil, 47, 50, 55 direct object
in South Africa, 13–14, 42 marker, 123
Créoles (i.e. people) word order, 123, 131
in Réunion, 68 does (CE habitual marker), 74
Index 169
Dominican Spanish, 17, 18, 19, 26, 60, German, 14, 59

62, 63, 71, 84, 85, 106, 108 dialects, 42
done (CE, AAE completive marker), 74 gradience, xv, 4, 5, 8–10, 12, 22, 25, 26,
drift, xv, 5, 15, 78, 90, 113, 143 141, 144
durative aspect, 79, 84 Green, K., 17, 19, 21, 142
Dutch, 7, 41–46, 116, 120 Guadeloupe, xvi, 26
archaic, 122 Guiné-Bissau CP, 81, 127, 129
colloquial, 98 Gullah, xii, 2, 26, 28, 29, 33, 36, 37, 41,
dialects, 97, 122, 139 77, 95, 119
Dutch Brazil, 51, 52 historical development, 26, 35
Dutch East India Company, 41, 42, 43 Guy, G., 16
Guyanais CF, 26
East Africans Guyanese CE, 28
in Réunion, 67 Guyanese Spanish, 60
East Indies see Indonesia
Ebonics, 1 habitual aspect, 75
education, 3, 21, 58 Habla Bozal, 17, 18, 108
emancipation see abolition of slavery historical development, 63, 137
embedded questions, 117 had in AAE, 74
emphasis, 81, 82, 94, 124 Haiti, 25, 26, 27, 67
ethnic linguistic marking, 32, 34, 39, 40 Halbkreolisch, 6
Eurocentrism, 15 Hancock, I., 10
European Portuguese, 56, 80, 83, 123, 125 Helvécia Portuguese, 16, 59–60, 82, 101,
European Spanish, see Peninsular Spanish 103, 125, 127, 137
Europeans (see also whites) Hesseling, D., 7
in Brazil, 49–50 Hiberno English, see Irish English
in Cape Colony, 45 highlighter, 90, 124
in South Carolina, 35 Hispanicization, 61
in Spanish Caribbean, 64 historical fragments of language, 2, 37
in Virginia, 32, 71 historical linguistics, xiv
Ewe-Fon, 55, 82 Hottentot, see Khoi
existential constructions, 76, 130 Huguenots, 34, 42
ex-slave narratives, 11, 37, 73, 93, hypercorrection, 97, 119
119
i (VLRF verb marker), 69, 86–87, 111,
foreigner’s English, 2, 33 141
fossilized forms, 78, 86, 91, 110 ideology, 1–3
founder effect, 14, 27, 32 Ile de France CF, 19
français avancée, 20 imperative, 81
français créolisé, 69 identity and language, 136, 137
freed slaves, 34, 44, 62 indentured servants
French, 42, 59 from British Isles, 29, 30
French East India Company, 65 from India, 69, 137
frequency, 145 India, 42, 69
full creolization, see creolization indigenization, 4, 145
indigenous languages, replacement of,
Geechee, xii 136
gender, 25, 42, 92, 95, 107, 108–110, indirect object, 131
139 Indo-European languages, xiv
common, 96 Indonesia, 42
comparison, 113–114 Indo-Portuguese, 6, 9, 19, 65, 112, 120,
lexical, 109 123
natural versus grammatical, 111 infinitive, 78, 81
genetic relation, xiv–xv, 71 marker, 129
Georgia, 32, 33, 35 pseudo-infinitif, 133
inflections, xv lexicon, 22
fossilized, 78, 86, 91 Liberia, 37
loss of, xv, 3, 4, 5, 22, 41, 43, 72, 78, Lı́ngua dos Pretos Velhos, 16, 103
90, 140, 144 Lı́ngua Geral, 48, 51, 52, 55–56
verbal, 11, 72, 86 linguistic processes in partial restructuring,
influence of full creoles, 137 142–143
-ing (AAE progressive marker), 74 linguistic traits of partial restructuring, 22,
innovation and partial restructuring, 121, 142
141 Lipski, J., 18
interior South, 36, 37 Lorenzino, G., 19
interlanguage, 14, 16, 33, 44 Louisiana CF, 26, 52
intermarriage, 42, 44, 48, 57
“intermediate” varieties, 30 Madagascar, 43, 65, 67
intertwining, xiii, 4, 145 main clauses, 120
intransitives from reflexives, 112 Makhudu, D., 14
invariant be, see be Malagasy, 65, 86, 90, 109, 112, 130–131,
inversion of S/V (or Aux), 116–117, 121, 137
124, 128, 129–130, 133, 140, 141 Malay
Irish English, xiii, 4, 30, 117 Bazaar, 100
isolating languages, xv, 5–10, 141 Malayo-Portuguese, 13–14, 42, 78, 100,
it’s (AAE existential), 76 123
manumission, 57
Jamaican CE, 25, 26, 30, 34, 71, 72, 77, mamelucos, 57
118, 120 Maranhão, 51, 57
Jeroslow, H. (see also McKinney, H.), 15, marked features, 142
16 Markey, T., xiv
Jews maroons, 28, 51, 54–68
in Brazil, 51 Martinique, xvi, 25, 26, 27
in Curaçao, 51–52 Maryland, 33
in Suriname, 52–53 Mauritian CF, 19, 20, 26, 28, 67, 88
juxtaposition, see [possessor + possessed] McKinney, H. (see also Jeroslow, H.),
15
Kaaps, see Afrikaans McWhorter, J., xiii, xv
Kaf, 69 Megenney, W., 18
Khoi, 13, 14, 15, 41–46, 78, 80, 100, Mello, H., 3, 15, 16, 21, 142
120 mesolect, 7, 117
koineization, 4, 48, 145 Mexican Spanish, 60
Kotzé, E., 14 Miami Spanish, 17, 60
Krio, 119 migration of speakers, 3, 28, 30, 39, 40,
135–136
la (VLRF perfect marker), 65 Minas Gerais, 55
Labov, W., 10, 11 mines (AAE possessive pronoun), 96
language borrowing, see borrowing Miskito Coast CE, 10, 94
language change, xiv, xvi morphological versus phonological rules,
language contact, xvi, 78, 90, 91, 144, 145 15, 80, 81, 83, 98, 102, 103, 105,
language drift, see drift 106
language shift, see shift Mozambique, 43
language transmission, see transmission Mufwene, S., xii, xiv, xv, 9, 12
of language Mühlhäusler, P., 6
Leeward Islands, 30, 34
lete ki (VLRF durative past marker), 65, Nama, 79, 80, 100
67, 69 Namibia, 47
leveling, 34, 143, 145 nasalization (see also denasalization), 56,
primary, 143 80
secondary, 38, 143 Natal, 45
Index 171
native-speaker ratio, 3, 12, 21, 24–29, 70, in BVP, 101–105

135–136 in NSCS, 105–108
in American South, 33, 145 in VLRF, 108–112
in Brazil, 50 Nova Scotia, 37
in Cape Colony, 45 number marking in NP, 92
in Réunion, 67, 70–71 comparison, 112–113
in South Carolina, 35 in AAE, 93–94
in Spanish Caribbean, 64 in Afrikaans, 97–98
in Virginia, 32 in BVP, 103
negation in NSCS, 106–107
comparison, 91 in VLRF, 108
discontinuous double, 72, 80, 82, 85,
88, 91, 139, 141 objects: order of direct/indirect, 131
in AAE, 75–76 Orange Free State, 46
in Afrikaans, 79–80 Orange River Afrikaans, see Afrikaans
in BVP, 82 orthography, xx
in NSCS, 84–85
in VLRF, 88–89 palatalization (see also depalatalization),
negative concord, 76, 84, 91, 139 56
clause-external, 76 Palenquero CS, 17, 18, 19, 62, 85, 107
negative inversion, 76 Palenquero Spanish, 18, 64, 107, 108,
postverbal, 79, 80, 82 129
Negerhollands, 6, 26, 44, 45, 46, 71, 121, Palmares, 54
122 Papen, R., 20
Negro Non-Standard English, 1 Papia, 16
New York City Spanish, 17, 60 Papia Kristang CP, 100
Niger-Congo languages, xiv, 5, 42, 102, Papiamentu, 6, 17, 26, 27, 71, 81, 105,
103, 116, 141 107, 127, 129
no (CE verbal negator), 75 Portuguese element in, 51–52
Nonstandard Caribbean Spanish, 2, 8 parameter settings, xv
auxiliaries, 84 Parkvall, M., xiii, 23, 24–27
dependent clauses, 128 partial restructuring, xiii, 2
gender in NP, 107 and substrate influence, 141
historical development, 60–64, 71, 137 theoretical model, 5, 7–10, 142–143
negation, 84–85 versus decreolization, 4, 23
non-verbal predicates, 85 past tense
number in NP, 106–107 in AAE, 73, 138
possession, 107 in Afrikaans, 78–79
pronouns, 107–108 Peninsular Spanish, 85
study of, 17–19 periphrasis, 88
verbal morphology, 83–84 Pernambuco, 50, 52, 54
word order, 127–128 petits blancs (see also poor whites), 68, 69
non-stative (see also stative), 87 phonological versus morphological rules,
non-verbal predicates (see also copula 15, 80, 81, 83, 98, 102, 103, 105,
absence), 23, 40, 41, 72, 74, 90, 106
141 internal motivation of phonological
comparison, 91 rules, 144
in AAE, 76–77 phonotactic rules, 106
in BVP, 83 Pidgin Dutch, 42, 43, 44
in NSCS, 85 Pidgin English, 2, 34–35
in VLRF, 89–90 Pidgin French, 28
North Carolina, 33, 35 Pidgin Portuguese, 52
noun phrase Pidgin Spanish, 63
in AAE, 93–96 pidginization, 3
in Afrikaans, 96–101 pidginization index, 10, 27
pied-piping, 122 preverbal markers, 11

plantation agriculture, 43 influence on auxiliary verbs, 72, 73–75,
coffee, 59, 61, 67 79, 82, 84, 85, 87–88, 139, 141
cotton, 36, 38 primary leveling, 143
rice, 34, 36 prior creolization, 16
sugar, 30, 50, 61, 68 processive, 86–87
tobacco, 29, 30, 32, 36, 61 Pro-drop, 81, 84, 128
“Plantation Creole”, 36 progressive aspect, 74
pleonastic subject pronoun, 96 pronouns
pluralizer = ‘they’, 93, 108, 111, 139, case marking, 92, 95, 99, 101, 104, 112,
141 140
in associative plural, 94, 97, 139 clitic, 99, 104, 115, 123
plural marking in NP, 92 comparison, 114
irregular, 97 conjoined, 94
Ponelis, F., 14 disjunctive, 110, 112, 115, 128
poor whites emphatic, 94, 104, 110, 111, 115
in Cape Colony, 46 gender marking, 95, 99, 104, 111
in Réunion, 68–69 in AAE, 95–96
Poplack, S., 12 in Afrikaans, 99–101
Pororó, 63 in BVP, 104–105
Portuguese, 47 (see also Brazilian in NSCS, 107–108
Vernacular Portuguese, European in VLRF, 110–112
Portuguese, Helvécia Portuguese, interrogative, 100
Standard Brazilian Portuguese) object, 101, 112
influence on Guyanais CF, 52 pleonastic subject, 96, 119
influence on Papiamentu, 51–52 possessive, 99–101, 122
influence on Saramaccan, 52–53 reflexive, 101, 105, 107, 112, 115, 140
[possessed (OF) possessor], 53, 98, 104, relative, 76, 96, 122, 125
110, 140 resumptive, 125, 134, 141
possession subject, 81, 84, 128
comparison, 114 word order, 104, 112, 123
in AAE, 95 prototype, creole, xv, 28
in Afrikaans, 98–99 Puerto Rican Spanish, 17, 18, 60, 62, 71,
in BVP, 104 84, 85, 106
in NSCS, 107
in VLRF, 110 quantitative studies, 93
possessive questions
determiner, 95, 99–101 direct, 121
marker on nouns, 92, 98 embedded, 121–122
pronoun, 99–101 word order, 116, 121, 128, 130
[possessor + possessed], 53, 140 question words, 124
in CE, AAE, 95 quilombos, see maroons
[possessor HIS possessed], 98–99,
100–101 race relations
in colloquial Dutch, 98 in Brazil, 57
post-creole, 1–9, 20, 23, 46 in South Africa, 43, 44
post-verbal negation, see negation in Spanish Caribbean, 61–62
post-verbal marker, 86 in USA, 32, 39, 41
power relationships, 3 racial segregation, 35, 39, 40, 57
prepositional phrase deletion, 126 Raidt, E., 14
prepositions Rademeyer, J., 14
absence before place names, 131 “radical” creole, 25
absence before indirect object, 131 rapidity of creole genesis, xvi
preposition stranding, 122 regional speech forms, 10, 11
present for past, 73, 78–79, 82, 91 regularity, 145
Index 173
Rehoboth Basters, 46 secondary leveling, 38, 143

Reinecke, J., xi, 8, 13, 19 segregation, see racial segregation
relative semi-creolization, xi–xiii, 8, 9, 12, 15, 19,
clauses, 117–119, 120, 122, 124–127, 37, 60
131, 133 Seychelles CF, 19, 20, 109
determiners, 125 share-cropping, 38, 40
pronouns, 122, 124, 131, 132 shift, 4, 16, 56
complex/simplex, 134 of heterogeneous language community,
resumptive, 125, 134 142, 143
subject, absent, 76, 96, 117–119, 125, of homogeneous language community,
128, 130, 133, 134, 140 30
religious rites, language of, 16, 18, 56 Siegel, J., 4
remote past, 73 Sierra Leone, 37
renativization, 145 Silva Neto, S., 9, 15
repidginization, 9 slaves, 36
restructuring index, 25, 26 from Africa, 32, 33, 36, 43, 45, 67
Réunionnais, see Vernacular Lects of from Caribbean, 32, 33, 36
Réunion French from Asia, 42, 43, 45, 67
Rickford, J., 10, 12, 24, 27 status, 30
Rio de Janeiro, 56 social factors in partial restructuring, 21,
Roberge, P., 13, 15 24–71, 135–137
Romance languages, 104, 139 sociolinguistics, xiv, 106
sociolinguistic history, 3, 24
-s (E verbal inflection: third person of AAE, 29–41
singular), 40 of Afrikaans, 41–47
in AAE, 73, 137 of BVP, 47–60
(E, D noun inflection: plural), 97, 139 of Gullah, 34–35
in AAE, 93–94 of Helvécia Portuguese, 59–60
in British dialects, 93, 94 of NSCS, 60–64
(NP inflection: plural) of VLRF, 65–70
in French, 108–112 South, see American South
in Portuguese, 56, 101 South Africa, 13, 46
in Spanish, 106–107 South African philological school, 14, 15
St. Barts or St. Barthélemy, 20, 23 South Carolina, 33, 35
St. Kitts, xvi, 26, 27, 30 Southern White Vernacular English, 23,
St. Thomas, 25, 26 40, 41, 75, 76, 95, 96
salience, 145 SOV word order, 42, 116, 120, 133
Samaná, 37 speech relationships
São Paulo, 48 black-white, 39, 40, 41
São Tomé CP, 17, 26, 27, 47, 57 Spanish, 52
influence on BVP, 51, 82, 102, 104, Spanish Caribbean, 84–85, 137
105, 126 demographics, 64
influence on Palenquero CS, 62 language contact in, 60–64
Saramaccan, 7, 120 “spontaneous development” theory, 14
Portuguese element in, 52–53 Sranan CE, 26, 27, 53, 120
say (CE, AAE quotative), 41, 120, 134, Sri Lanka, 42, 45
141, 142 Standard Brazilian Portuguese, 57–59,
Schneider, E., 10, 12 80–81
Scholtz, J., 14 standardization, 13, 21, 46
Schuchardt, H., 6, 19 stative (see also non-stative), 87
Schwegler, A., 18 Stewart, W., 7, 8, 36
se (Afrikaans possessive), 98–99, 122, stigmatized forms, 73, 74
140 strong restructuring, 23
-se (NSCS plural marker), 106–107 structural distance, 7–10, 12, 25
second language acquisition, 5, 14, 18, 35 structural reduction, 140
subject/verb (or Aux) inversion, Venezuelan Spanish, 17, 18, 60, 85

see inversion verbal morphology
subjunctive mood, 82 comparison of, 90–91
subordinate clauses, 120, 123, 127, 129 in AAE, 73
subordinator, 118 in Afrikaans, 77–79
subordinator, absent, 123, 127, 128, in BVP, 80–82
132, 134, 140, 141 in NSCS, 83–84
substrate, xiv, 4, 5, 139 in VLRF, 20, 85–87
influence of, 15, 16, 18, 33, 41, 72, 86, verb phrase
134, 141, 143 in AAE, 73–77
superstrate, xiv, xv, 5, 6 in Afrikaans, 77–80
Surinamese creoles (see also Saramaccan, in BVP, 80–83
Sranan), 7, 8, 28, 30, 34, 41, 52, in NSCS, 83–85
120 in VLRF, 20, 85–90
SVO word order, 42, 104, 116–119, 120, verb-last word order, 122
121, 123, 130, 133 verb-second word order, 120, 121, 122
syllable structure, 56 Vernacular Lects of Réunion French,
syntax-internal model, xv, xvi 2, 10
synthetic languages, xv, 5 auxiliaries, 87–88
systematicity within language, 5 Créole des Bas, 20, 67, 69, 86, 137
Créole des Hauts, 20, 69
Tagliavini, C., 6 dependent clauses, 131–133
target language, access to, 136 historical development, 65–70, 136,
te (VLRF verb marker), 69, 87 137
tense, 72, 86 negation, 88–89
theoretical model for partial restructuring, non-verbal predicates, 89–90
142–143 number marking in NP, 108–112
Thomason, S. G., xiv, 3–4, 9, 24, 142 possession, 110
time factor in restructuring, 24, 27–29, study of, 19–21
135 time factor in restructuring, 28
Tok Masta, 6 urban Creole, 20
tone, xv, 25 verbal morphology, 85–87
topicalizer, 90 word order, 129–131
transmission of languages, xv, 3, 21, Virginia, 29, 30, 32, 33, 35, 36
67, 71 Virgin Islands, 71
normal, xiv, 71 VSO word order, 130
irregular, 103
transparency, 142, 145 weak restructuring, 23
Transvaal, 46 West Africans
Trinidadian Spanish, 60 in Cape Colony, 42
Tupi, 15, 48, 53, 55–56 in Barbados, 29
typology, xiii, xiv–xv in Guyana, 28
in Mauritius, 20, 28
unbound pluralizer, see pluralizer in Réunion, 67
unrestructured overseas varieties, 3, 33, in South Carolina, 35
34, 135, 136 in Virginia, 29, 32
universals, xv, 5 whites
unrounding of vowels, 69 as creole speakers, 52
Upper South, 36 in Brazil, 49–50
urban ghettos, 39 in Cape Colony, 45
in South Carolina, 35
Valkhoff, M., 14, 15, 19 in Spanish Caribbean, 64
Van Name, A., xvi, 17, 24 in Virginia, 32, 71
van Rensburg, M., 14 white supremacists, 39
Vasconcellos, J., 7 WH-movement, 122
Index 175
Winford, D., 4, 12 you all (AAE, SWVE pronoun), 96

withdrawal of superstrate, 28 Yoruba, 77
word order, 101, 104, 116–117, associative plural, 94
119 in Brazil, 56
comparison, 133 pluralizer, 93
in AAE, 116–117 relative clauses, 118
in Afrikaans, 120–122
in BVP, 123–124 zero copula, see copula absence
in NSCS, 127–128 zoreil, 69
in VLRF, 129–131 Zulu, 45

John Holm-Languages in Contact The Partial Restructuring of Vernaculars (2004) PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

John Holm-Languages in Contact The Partial Restructuring of Vernaculars (2004) PDF

Uploaded by

Copyright:

Available Formats

This page intentionally left blank

There is widespread agreement that certain non-creole vernaculars are

John Holm is Chair, English Linguistics and Director, Graduate Pro-

Cambridge University Press

© John Holm 2004

This book is in copyright. Subject to statutory exception and to the provision of

First published in print format 2003

ISBN-13 978-0-511-06755-6 eBook (EBL)

ISBN-13 978-0-521-43051-7 hardback

Cambridge University Press has no responsibility for the persistence or accuracy of

List of maps page ix

1. The study of partially restructured vernaculars 1

2. Social factors in partial restructuring 24

3. The verb phrase 72

4. The noun phrase 92

5. The structure of clauses 116

1 United States: ‘Coloured Population, 1900.’ Originally

1 Degree of restructuring and social factors page 26

The naming of books is a tricky business, and it has consequences. In the

linguists’ self-licence to go around the world baptizing some vernaculars “cre-

use, many African Americans and Brazilians find semi-creole or semi-crioulo

means in linguistics. Genetically dissimilar languages can, by coincidence,

(e.g. definite/indefinite opposition) and those which are “ornamental”

My greatest debt is to my former doctoral students at the City University

a paper at a conference on Réunion (2002); from the University of Puerto

A, AFK (standard) Afrikaans

underwent decreolization, or the loss of many of its distinctive creole fea-

with non-European languages, today have some 200 million speakers,

such processes as intertwining (Bakker and Muysken 1994), koineization,

1.1 Partial restructuring versus decreolization

have had more difficulty developing an adequate theoretical model for

Ueberall wo eine kreolische Mundart gesprochen wird, liegt Mischung mit

The first situation produces a variety similar to what Mühlhäusler

This view was not elaborated into a full-blown theory of decreolization

the non-standard speech of present-day American Negroes still seems to exhibit

DeCamp (1961, 1971) developed the idea of a continuum of lects for

The first recognition of a whole category of such languages can be found

As recently as l962, Stewart considered Suriname to have the only real

between the Surinamese and West Indian varieties of creolized English is

The term is also used in this sense by Thomason and Kaufman in

Bickerton (1984:176–178) proposed what he called a pidginization in-

The theoretical importance of gradience in creolization was signaled

1.2 The study of African American English (AAE)

I was further led to question some of the basic assumptions of decre-

In 1986 there began a debate as to whether AAE and white varieties of

A more radical view (which seems inherently unlikely, given what is

of varieties such as Gullah and AAE, which stand at differing distances

While Schneider considered the term semi-creole unhelpful because of the

More recent work on AAE has focused increasingly on those sociolin-

distinctions between contemporary varieties of AAVE and American and

1.3 The study of Afrikaans

characterization set off a debate that continued for a century, at times

What might be called the “Amsterdam school” of Afrikaans scholars

1.4 The study of Brazilian Vernacular Portuguese (BVP)

her to suspect prior creolization (McKinney 1975). Guy (1981) exam-

both BVP and NSCS, as is the recently founded Associação: Crioulos de

1.5 The study of Nonstandard Caribbean Spanish (NSCS)

conferences on Palenquero and NSCS (Moñino et al. 2002) as well as the

1.6 The study of Vernacular Lects of Réunionnais

dialectal mosaic” (1967:126). Deltel (1969) proposed that there was a

kind of restructuring found in normal language transmission. This

1.7 The comparison of partially restructured vernaculars

Regarding the linguistic make-up of each variety, the following factors