The Afro-Hispanic speech of the municipio of Cuajinicuilapa, Guerrero

Material Information

The Afro-Hispanic speech of the municipio of Cuajinicuilapa, Guerrero
Althoff, Francis Daniel, 1955-
Publication Date:


Subjects / Keywords:
African Americans ( jstor )
Creoles ( jstor )
Language ( jstor )
Linguistics ( jstor )
Pidgins ( jstor )
Slavery ( jstor )
Spoken communication ( jstor )
Verbs ( jstor )
Vowels ( jstor )
Words ( jstor )

Record Information

Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
029218026 ( ALEPH )
41352994 ( OCLC )

Full Text







Copyright 1998


Francis Daniel Althoff, Jr.

Cora Mae Knight Andrew


In meora

This study owes its existence primarily to John M.

Lipski. Indefatigable researcher and unexcelled scholar, he

has shepherded this work from its sketchiest beginnings to

its final form with infinite patience and a quiet confidence

that it would finally come to fruition. The esteem in which

I hold him can only be hinted at in these few words. I

extend to him my deepest thanks for his guidance and the

many courtesies and graces that go far beyond the call of

mere professional obligation.

My sincere appreciation to the other members of my

supervisory committee: to David A. Pharies for skillfully

managing the details of my program in Gainesville and for

his keen editorial eye; to Haig Der-Houssikian for his

encouragement, enthusiasm and insistence on clarity; to

Florencia Cortds-Conde for taking on yet another task for no

reason other than her own generosity. Geraldine Cleary

Nichols, chair of the Romance Languages department, has been

gracious in facilitating the completion of my program; I

extend my thanks to her for smoothing my path.

Many thanks are due also to Margarita Hidalgo for

initiating the chain of contacts in Mexico, beginning at the

Instituto Nacional Indigenista (INI) in Mexico City with

Jesus Rubiell (now deceased) and Bdrbara Cifuentes. My

appreciation and thanks to Guerrero state coordinator of the

INI in Tixtla, Josd Manuel Figueroa; to Magali Meza Herrera,

director of the Direcci6n General de Culturas Populares in

Chilpancingo; and Francisca Aparicio Prudente and Maria

Cristina Diaz Pdrez who were finishing their fieldwork in

San Nicolds. Jaime Rique Alarc6n, director of the secondary

school in San Nicolds, offered me hospitality and invaluable

assistance in making contacts within the community. The

participants were indispensable to this study, and I thank

them deeply for willingly sharing some of their lives with

this unlikely g .

The Tinker Foundation and the University of Florida

Center for Latin American Studies are gratefully

acknowledged for a grant that made travel and initial

fieldwork possible.

I would also like to thank the many friends and

colleagues from the University of Florida who have enriched

my life, especially Mark and Silvia Cox, Lynn Scott, Michael

Rao, Lynn Knox Salzburg, Antonio Gil and Brian and Brenda

Rae Polding.

Finally, the love and support of my family have been

unwavering. My parents, Dan and Sybil Althoff, my aunt and

uncle, Bea and Joe Carpentieri, my sister Carolyn Althoff

Buttner, and my "brother of choice" Stephen Ferrell have

contributed in so many essential ways to this project that a

complete accounting would be impossible. Although they know

how I feel, I want to express publicly my love and gratitude

to them.


ACKNOWLEDGMENTS ........................................... iv

ABSTRACT .................................................. xi


1 INTRODUCTION ............................................ 1

Mexican Spanish ...................................... 1
Gulf Coast and Southern Pacific Coast Spanish ........ 4
Calls for Research .......................................... 5
Research Purposes and Questions ...................... 7
Organization of the Study ............................ 9
Notes ................................................ 12

2 THEORETICAL CONSIDERATIONS ........ ....................... 13

Languages in Contact ................................... 15
Pidgin Language .................. ..................... 16
Portuguese Contacts in Africa ................... 18
Literary Evidence of Africanized Portuguese ..... 19
Literary Evidence of Africanized Spanish ........ 24
Creole Genesis in Africa .............................. 25
Processes of Creole Development and Evolution ........ 26
Monogenetic and Polygenetic Theories of Creoles ...... 31
For a Monogenetic Theory of Afro-Hispanic
Language ...................................... 33
Questioning a Monogenetic Theory of Afro-
Hispanic Language .............................. 42
Ibero-Romance Creoles ...........................48
Features of Afro-Hispanic Caribbean Speech............50
Summary .. .. .............................. .......... 54
Notes.................... .. .............. ..... 56


Sor Juana............................................... 61
Parochial villancicos of Puebla........................ 63
18th Century Songs and Dances.......................... 64
Modern Descriptions of Costa Chica Spanish ........... 66
Summary................................................ 73
Notes................................................ 75


4 AFRICANS IN MEXICO ..................................... 76

The Slave Trade to Mexico ............................. 79
First Period of African Slavery in Mexico ....... 82
Second Period of African Slavery in Mexico ...... 83
Final Period of African Slavery in Mexico ....... 86
Decline and End of African Slavery in Mexico .... 95
Ethnic Origins of Slaves in Mexico ................... 97
The Conditions of Slavery ............................. 99
Conditions of Slavery in Africa ................. 101
Conditions of Slavery Aboard Ship ............... 102
Conditions of Slavery in the Factorias .......... 105
Conditions of Slavery in Mexico ................. 106
The Sociolinguistic Conditions of Slavery ............ i1
A Case of Slave Resistance in Mexico ................. 116
Two Coasts of Slavery ................................. 118
Summary ............................................... 121
Notes ................................................. 125


History of the Costa Chica ........................... 130
Origins of the Afromestizo ........................... 133
The "Lords" of Cuajinicuilapa ........................ 136
Afromestizo Independence in the Costa Chica .......... 138
Group Identity Among Afromestizos .................... 142
Cuajinicuilapa in 1949 ............................... 144
Cuaji in the Nineties ................................ 146
Notes................................................ 152

6 SITE SELECTION AND METHODOLOGY ....................... 154

San NicolAs Tolentino ................................ 154
Data Gathering ....................................... 161
Linguistic Interviews ................................ 163
Data Excluded from Analysis .......................... 165
Summary ............................................... 166
Notes ................................................. 167

7 ANALYSIS OF THE SPEECH DATA .......................... 169

Morphology ........................................... 170
Direct Address and the Use of Voseo............. 170
Variability of Grammatical Gender ............... 173
Anomalous Noun Forms ............................ 175
Verb Mood and Tense ............................. 177
Subject-Verb Agreement .......................... 183
Productive Derivational Processes ............... 185
Archaisms ....................................... 186
Syntax ................................................ 189
Pronouns ......................................... 189


Clitics ......................................... 192
Preposition Loss ................................ 193
Lack of Subordination ............................ 196
Intransitive Verbs Used Transitively ............ 198
Verbs ser and estar ............................. 199
Verbs conocer and s ......................... 200
Phonetics and Phonology ............................... 201
Vowels .......................................... 201
Vowel raising ............................... 201
Vowel lowering .............................. 204
Other variations ........................... 205
Diphthongization ............................ 205
Monophthongization .......................... 206
Suppression of hiatus ...................... 207
Hypercorrection of hiatus .................. 210
Consonants ....................................... 211
Voiced stops /b/, /d/, /g/ ................. 211
Consonant cluster simplification ........... 213
Behavior of liquids /l/ and /r/ ............ 214
Nasals and nasalization .................... 218
The behavior of standard /f/ ............... 220
Archaic and epenthetic [h] ................. 222
Behavior of /s/ ............................. 224
Other Phenomena .................................. 226
Paragoge .................................... 226
Aphaeresis .................................. 227
Prothesis ................................... 229
Metathesis .................................. 230
Apocope ...................................... 231
Haplology ................................... 232
Local Lexicon .......................................... 233
Words of Indigenous Origin ...................... 233
Words of Likely Indigenous Origin ............... 235
Words of African Origin ........................ 237
Words of Unknown Origin ......................... 237
Summary .............................................. 238
Notes ................................................ 247

8 CONCLUSIONS ............................................ 251

Examining the Assumptions ............................. 252
The Research Questions ............................... 258
Research Question One ............................ 259
Research Question Two ............................ 261
Research Question Three ......................... 261
Research Question Four .......................... 262
Suggestions for Further Research ..................... 263
Conclusion ............................................. 268
Notes .................................................. 271


A INFORMANT DATA .......................................


C TRANSCRIPTION CONVENTIONS ............................

D INFORMANT 1 ..........................................






















............. oo eooo .. ... o... oo ..o.o

.......oo.... o .. o........ oo ..oo o ...

............. oo ... ....o .....o ..ooe o ....

.o .... oo .o ...............oo ..o ........

........... eo ...oeoo .. ..o o .............


AND 19..o .. o ........o ........ o .......

eo ................................ o .oo

REFERENCES ..............................................

BIOGRAPHICAL SKETCH .....................................























Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Francis Daniel Althoff, Jr.

May 1998

Chairman: John M. Lipski
Major Department: Romance Languages and Literatures

The people of the 'county' of Cuajinicuilapa

in the Mexican state of Guerrero are among the few groups in

Mexico that are of identifiably African origin. The Spanish

they speak has striking similarities to that of other

African-derived groups living in and along the coast of the

Caribbean and the Gulf of Mexico. The linguistic

correspondences between Afro-Mexicans of the Pacific coast

and Afro-Hispanic groups living in the circum-Caribbean has

led some researchers, notably the linguist German de Granda,

to hypothesize that an original language, derived from a

Portuguese-based pidgin brought to the Americas by west

African slaves, underlies the Spanish spoken by Afro-

Hispanics in this hemisphere. The present study adduces

linguistic, ethnographic and historical evidence in order to

test Granda's hypothesis.

The study examines the process of pidgin and creole

language formation and notes those features that have been

identified with the Afro-Iberian speech, beginning with

early attestations from the Portuguese theatre and

continuing with later manifestations of "black" language in

Spanish and Mexican literature. A critical examination of

the primary historical record used to support the presence

of a pidginized or creolized speech reveals that such a

language was found only among a comparatively small number

of Africans and thus probably did not constitute a

widespread substratum for the later acquisition of local

varieties of Spanish. The psychological and physical

traumas of slavery also impeded the acquisition of European


The history of slavery in Mexico and the socio-

linguistic conditions there indicate that Africans were

subjected to early pressure to assimilate linguistically

and, further, that there are few records of Africanized

Spanish in Mexico.

The data recorded on-site show some morphophonological

patterns of likely African origin have been retained in the

local speech. The study nevertheless concludes that the

evidence is insufficient to posit an earlier underlying

Afro-Portuguese basilectal speech: the speech of the

1fliclpi has been defined by interaction with indigenous

people and with south Mexican Spanish.



Lengua y sociedad, linguistica e historia
son entidades por completo inseparables:
son las dos caras indivisibles de esa moneda
de oro puro que es la vieja y noble filologia.
--Juan M. Lope Blanch

Mexican Spanish

The study of the Spanish spoken in Mexico has produced

a body of dialectological information that until recent

years was fairly thin (Lope Blanch 1991), but has burgeoned

noticeably in the past quarter century. An inspection of

the pertinent literature reveals in the first place that

Mexican Spanish is hardly a monolithic entity. Although the

concept of Mexican Spanish may first bring to mind

(especially in the United States) the characteristic "sing-

song" intonations of northern Mexico, a clear-eyed and

attentive examination of the linguistic realities of a

country as diverse as Mexico will quickly dispel the notion

of a homogeneous national speech: indeed, one might almost

refer to the "Spanishes" spoken in Mexico. This is not to

say that there is no standard variety of Spanish in Mexico

or that there are no common characteristics of Spanish in

Mexico (Lipski 1994a:279). In fact, the norma culta

'educated standard' of Mexico City and other urban centers

has been described, after Castilian Spanish, as "probably


the most prestigious variety of the Spanish-speaking world,

and certainly the most reputable in the Latin American

subcontinent" (Hidalgo 1996:68).

While studies of the norma culta and urban speech in

general have yielded important contributions to the

knowledge of the Mexican linguistic landscape, they have

historically also dominated Mexican dialectology. The

varieties of Spanish spoken in the more remote and

marginalized areas have only rarely received the same

careful attention as urban dialects. To illustrate, of the

346 entries related to Mexican Spanish in a recent

bibliographical survey (Fulk 1993a, 1993b), we find only 48

entries (14%) dedicated to studies of regional varieties,

and of those 48 entries, more than half (25) concern the

speech of the Yucatan peninsula.' In that same

bibliographical survey, the states of Chihuahua and Tabasco

show the most number of studies (after the YucatAn's 25)

with only four each, followed by three entries relating to


The dearth of studies bearing on other regional speech

varieties does not indicate, of course, that they are less

worthy of description and study. Rather, the greater number

of Yucatecan studies reflects the extraordinarily rich

linguistic atmosphere of that peninsula and, in particular,

the fascination it holds for one author. The Atlas

Linguistico de Mdxico (Lope Blanch, Alcald Alba et al. 1990)

is an impressive effort to redress some of the imbalances in

Mexican dialect studies. Conceived of in 1965, the atlas

project was developed by Lope Blanch and a core group of

five researchers. After completing pilot studies, teams of

two to three fieldworkers visited 193 localities across

Mexico, including each of the state capitals, over the

course of 15 years. The number of sites surveyed within a

region depended essentially on its population: simply put,

for sparsely inhabited areas (e.g. the northern deserts,

parts of Quintana Roo), there are fewer points included than

for more populous zones. At each location, the teams

interviewed seven informants; typically, four of them were

tape-recorded and three were asked to respond to a

questionnaire instrument (Lope Blanch 1991:158-63). The

results of this project are still being published, but have

already begun to elucidate the linguistic geography of

Mexico in a way heretofore impossible. Indeed, previous

delineations of Mexican dialect zones have almost always

relied on the suggestions first made by the renowned

Dominican philologist Henriquez Urefia in 1921 (q.v.), who

proposed six broad regions:

(1) North (consisting of Baja California, Sonora,

Sinaloa, Chihuahua, Durango, Coahuila, Nuevo Le6n, most of


(2) Central (Mdxico state, Puebla, Tlaxcala, Hidalgo,

Querdtaro, Guanajuato, San Luis Potosi, Aguascalientes,


Zacatecas, Michoac~n, the Veracruzan highlands, the national

capital [Mdxico, D. F.] and the related subgroup of Jalisco,

Colima and Nayarit);

(3) Gulf Coast (Tabasco, Campeche, the lowlands of

Veracruz and a part of Tamaulipas);

(4) South (Morelos, Guerrero and Oaxaca);

(5) YucatAn (YucatAn and Quintana Roo); and

(6) Chiapas (linguistically a part of Central American

[Guatemalan] Spanish).

Although this classification has been useful as a

heuristic, Henriquez Urefia made his distinctions based on

incomplete and somewhat asystematic data as well as on his

own intuition (Lope Blanch 1991:156), and it has become

clear that some possible relationships bear closer


Gulf Coast and Southern Pacific Coast Spanish

The classification of the speech of Guerrero, Oaxaca

and Morelos as a unit in Henriquez Urefia's vision misses

some essential connections with the speech of the Gulf

coast. Specifically, the speech of the states of Veracruz,

Tabasco and Campeche share phonetic features with varieties

of southern Pacific coast Spanish in Oaxaca and Guerrero.

These similarities (e.g. weakening of syllable-final /s/,

velarization of word-final /n/, some neutralization of /I/

and /r/) are noted in Lipski (1994a); Lope Blanch (1989a)

also notes some phonological and lexical similarities

between the two coasts. The Isthmus of Tehuantepec--where

the Gulf and Pacific coasts narrow to form a "waist" only

approximately 200 kilometers (120 miles) wide--may provide a

geographic and historical link between the two lowland


It is the potential links between the Gulf coast

variety and southern Pacific coast speech that partially

motivate this study. On the Gulf coast, the port of

Veracruz was the point of entry for thousands--i.e. the

overwhelming majority--of African slaves who were forced to

labor in New Spain.2 As the trade in humans came to its

eventual end, the town of Campeche, also located on the Gulf

coast, was the last official slaving port. These two ports,

Veracruz to the north and Campeche to the south, form the

lips of a topographic "funnel" directly into the Isthmus of

Tehuantepec, which might well have served as an escape route

for fleeing slaves to make their way to the Pacific coast.

In such a case, which is purely speculative, they would

undoubtedly have brought with them their own distinctive

patterns of speech which have historically been associated

especially with the Caribbean islands of Cuba, Puerto Rico

and the Dominican Republic, and to only a somewhat lesser

degree with the coasts of Venezuela and Colombia.

Calls for Research

The Spanish spoken by Africans and their descendants in

the Americas has been an important focus of dialectological

research in recent years. Calls for research in this field

date at least as far back as the early 1970s (cf. Granda

1978) and have been frequently reiterated (e.g. Lipski

1989). One of the dangers of investigating "black Spanish"

results from its popularity, especially in the 19th and 20th

centuries, as a literary device.3 Lipski has noted that

the immense notoriety and popularity of this literary

language "has deflected attention away from the existence of

real cases of black Spanish, not as used by writers and

actors, but as spoken in isolated linguistic enclaves

throughout the Americas and in Africa" (1985a:63).

Indeed, among the least known of Mexico's many

linguistic realities is the speech of its African-derived

populations. The African presence in Mexico as a whole has

historically been minimized, if not ignored or even denied.

The process of m in Mexico--generally understood as

the intermarriage of European and Amerind peoples, destined

to result in a single national population and identity--has

not been conducive to the recognition and preservation of an

African legacy. This third racial heritage in Mexico has

nevertheless managed to survive in certain areas of the

country, notably along the Gulf of Mexico from Veracruz to

Tabasco and Campeche; and along the Pacific, following the

coast of Guerrero state from Acapulco, tracing a southerly

route just into the neighboring state of Oaxaca, an area

known in its entirety as the Costa Chica. This area,


specifically the 'county' of Cuajinicuilapa on the

Costa Chica, is one of the few remaining enclaves of

identifiably African origin in Mexico that still has the

potential to shed some light on Spanish as spoken by

Africans and later by their mixed-blood descendants who

continued to live, until relatively recent times, in a state

of near isolation.

Research Purposes and Questions

One of the reasons underlying this study of the Spanish

of the of Cuajinicuilapa is to seek to fill some

descriptive lacunae. We have seen above that most

descriptions and analyses of Mexican Spanish have

concentrated almost exclusively on urban varieties with some

few forays into more isolated, rural zones. Moreover, the

number of the Afro-Mexican texts extant is extremely small:

the substantial transcriptions appended will augment the

linguistic research materials available for further

investigation. An additional benefit of the appendixes is

that they provide supplementary ethnographic and

anthropological testimony regarding many aspects of life in

the z, although they themselves do not refer

explicitly to the speech of the village under study. The

data contained in these texts--which are artifacts of the

linguistic elicitation techniques employed--may be useful to

folklorists, ethnographers and anthropologists in assessing

cultural aspects of mesiKaj and assimilation.

Apart from the interest inherent in the study of any

speech variety, the language of the m of

Cuajinicuilapa is important for hypotheses concerning the

origins of a supposedly pan-Afro-Hispanic pidgin or creole

language. It has been argued, primarily by Granda (1978),

that Africans, upon being enslaved by Europeans, learned a

pidgin Portuguese which they brought with them to the

Americas. This African-influenced Portuguese language would

have served as the substrate upon which acquisition of local

varieties of Spanish would then have taken place. It is

hypothesized that if groups of Africans had lived separately

from Spanish American colonial societies for a substantial

amount of time (such as in the slave refugee communities

known as palJnque), then it is likely that their speech

would exhibit signs of a divergent, creole-like trajectory

of development, if not evolve into a full-blown creole

language. In Mexico, such an enclave of Africans was

reported in the Costa Chica in the early years of the 16th

century near Cuajinicuilapa. The second major purpose of

this study, therefore, is to assess the speech of this

micipio to determine if there are any elements present
that would tend to support or disconfirm the hypotheses that

(1) A Portuguese-based pidgin or creole language was

originally spoken by Africans in Mexico;

(2) A Portuguese-based pidgin or creole language was

originally spoken by Africans in the area currently

encompassed by the m of Cuajinicuilapa;

(3) A creole or creole-like4 language emerged among

the descendants of Africans (i.e. Afromestizos) in the

muicipio; and

(4) The local creole or creole-like language later

underwent a process of decreolization which resulted in a

variety of Spanish essentially indistinguishable from other

dialects of rural Mexican speech.

Organization of the Study

In order to test these hypotheses, an essential

theoretical background is provided in Chapter 2. Beginning

with definitions and some problems related to terminology,

an overview of the consequences of language contact between

Portuguese and Africans (such as the language of Africans in

poetry and the theatre) is presented. The "monogenetic

hypothesis" for Afro-Hispanic language in the Americas as

proposed and developed by Granda (1978) is the framework

used for analyzing the corpus at hand. This hypothesis,

along with some related critical observations, is expounded.

There is also an examination and description of the

characteristics of the literary texts that have been used to

situate Africanized Spanish in the circum-Caribbean in

support of the monogenetic hypothesis.

The third chapter examines the presence and

characteristics of Africanized Spanish in Mexico as recorded

in texts from the 17th and 18th centuries. The speech of

the Costa Chica and the municip.i of Cuajinicuilapa has also

been described in some writings beginning in the 1930s.

The fourth chapter presents the history of the slave

trade to Mexico and the conditions facing African slaves

that would certainly have had an impact on their ability to

acquire some variety of a European language. Additionally,

the sociolinguistic circumstances surrounding slavery in

Mexico are examined in detail, as well as the case of the

most famous 2alengue in Mexico, located in modern Veracruz

state. The case of the 2 of Yanga has relevance to

the development of Afro-Hispanic language in Mexico.

The fifth chapter presents the history of the

of Cuajinicuilapa, drawing primarily upon the work of

Aguirre Beltrdn (1958) who, especially in the 1940s, carried

out the pioneering work on the African in Mexico. This

chapter highlights the historic isolation and distinctive

culture of the Afromestizos in this area dating from the

mid-1500s, emphasizing its general suitability for updated

field studies. The nearly fifty intervening years, however,

have modified the original rural characteristics of

Cuajinicuilapa, and a nearby, more suitable alternative site

for investigation was located.

The of Cuajinicuilapa contains within it

several small, independent communities. Chapter 6 describes

the selection of the town of San Nicol~s Tolentino for

fieldwork for this study, as the locus of Costa Chica

Afromestizo culture and research had clearly shifted there

since 1949. The essential cultural equivalence of San

Nicol~s, Cuajinicuilapa and other smaller, neighboring towns

is also noted in Aguirre Beltr~n (1958), thus eliminating

any significant doubts about the comparable earlier forms of

speech. Technical details concerning the recording of data

and the protocols guiding the elicitation of the corpus are

also discussed.

The linguistic data obtained from the participants are

presented for analysis and discussion in Chapter 7. The

analysis is based upon a comparison of characteristics that

have been identified as having an African origin, those

having a strong association with Africanized populations in

the Spanish-speaking Americas and those merely

characteristic of rural Spanish in the Americas. The data

are assessed with regard to morphology, syntax, phonetics

and phonology, other locally significant morphophonological

phenomena and the etymology of some of the distinctive

regional vocabulary.

The summarizing analysis of the entire study is found

in Chapter 8. A discussion of the creole hypothesis and the

specific research questions presented in this introductory

chapter leads to concluding observations and suggestions for

further research on Afro-Mexican Spanish.

1. The Mayan presence in the Yucatan and its unique contact
situation with Spanish certainly justify serious linguistic
interest. But it should be also noted that 13 (52%) of
those 25 studies of Yucatecan Spanish are the product of a
single author, Juan M. Lope Blanch, dean of Mexican
2. This study prefers to use adjectives derived from the
names of continents rather than color to refer to people of
different ethnic or racial origins. Thus, "African" is used
in preference to "black," "European" in preference to
"white," etc. The peoples indigenous to the Americas are
generally referred to as such. Current practice among the
indigenous peoples of the United States also generally
allows the use of "Indian" as an inoffensive adjective and
noun; the use of that term in this study is, however,
sparing. Likewise, the adjectives "native" and "aboriginal"
are used only rarely. Where direct quotations and English
translations have been provided, the original vocabulary has
been retained.
3. The authenticity of 16th and 17th century literary
"black" Spanish is not questioned to the same degree.
Although there are certain phonetic conventions, stock
phrases and stereotyped characterizations, theatrical
representations of black speech of the Golden Age are not
generally regarded as baseless inventions. See Chapter 3 for
a discussion of this topic.

4. Terminology is discussed in the following chapter.


Any study, such as this, treating a speech variety that

is explicitly purported to have specific geographic, ethnic

or racial origins or associations must begin by defining the

term. To that end, "Afro-Hispanic" as used in the title of

this study is considered to be Spanish as spoken originally

by groups of Africans and later by their descendants living

in community, whether in early 16th century Spain or, as in

the present case, late 20th century Mexico. A wide spectrum

of manifestations is subsumed under this definition. One of

these must logically have been the speech of individual

Africans who would have learned Spanish with varying degrees

of success. Although the individual language outputs would

have shown considerable variation based on natural

differences, the universality of this experience for African

slaves arriving in Spanish America would permit classifying

their interlanguage as a group variety. Another

manifestation of "Afro-Hispanic" language is the Africanized

(or bozal) Spanish variety richly attested in theatre and

poetry. Yet another would be a pan-Latin American standard

Castilian (cf. Hidalgo 1990), absolutely indistinguishable

from that spoken by any other ethnic or racial group per se.


Some mode of speech regarded as a "black," "African" or

"Africanized" variety of Spanish was certainly depicted by a

number of observers and writers of past centuries.

Sustained scholarly interest in this Afro-Hispanic language

is of fairly recent date, coming to acquire a particular

momentum especially in the second half of the 20th

century.' Although there is no present-day Spanish-

speaking society in which there exists a dialect analogous

to African-American Vernacular English (Lipski 1994a:128-

29), this does not erase or minimize the fact that Africans

were portrayed in literature as speaking a distinctive form

of Spanish (e.g. Lipski 1986b, 1992, 1995a), particularly

during the two centuries following colonization.

The present state of "black" Spanish, quite apart from

its intrinsically interesting origins, occupies a

controversial place in synchronic dialectological studies.

There are those who maintain that Spanish speech features

alleged to be "black" (i.e. of an African substrate origin)

may be adequately explained within the matrix of extant

variations and theories of dialect diffusion. Other

scholars assert that the confluence of certain phonological,

syntactic and lexical structures coupled with demographic,

ethnographic and other extralinguistic data points

inescapably to an Africanized speech variety that must, at

the very least, be considered a proto-Afro-Hispanic pidgin.

The present chapter will elaborate a framework of analysis

based upon theories of language contact bearing upon the

creation and evolution of pidginized and creolized varieties

of Spanish in the Americas in order to assess more

accurately the place of Costa Chica dialect within the

larger complex of Afro-Hispanic speech.

Languages in Contact

Curiosity about what happens when at least two

mutually unintelligible languages come into contact within a

delimited social and geographic space has given rise,

predominantly within the past 50 years, to the formal study

of pidgin and creole languages (creolistics).2 Prior to

that point, these modalities did not count, from a Western

perspective, as "serious" languages worthy of study in their

own right: they were sometimes humorously, sometimes

deprecatingly regarded as corrupt or aberrant forms of

correct, civilized European speech.3 In hindsight, it is

clear that these judgements reflected more the Europeans'

attitudes toward the speakers of those languages rather than

a fair assessment of "broken English" or "bastard

Portuguese" as languages per se (cf. Holm 1988:1). As

linguistic analysis grew more sophisticated and less overtly

ethnocentric, the recognition of pidgins and creole

languages as intact systems and not "broken" European speech

marked the true emergence of creolistics as a scientifically

based undertaking.


Although simple curiosity cannot be entirely discounted

as a motivating factor, commercial and military rivalries

were the main engines driving the establishment of European

outposts and colonies in Africa, Asia and the Americas. Led

by the Portuguese in the mid-15th century, there began years

of intensive language contact between European explorers,

traders, colonizers and the local populations. In the

context of the present study, the linguistic ferment arising

as a result of contact between Portuguese and the various

languages of the African Atlantic coast is of particular


Pidgin Language

In a language contact situation, it is natural for

people to attempt verbal communication even where there is

no mutually intelligible variety of language available.

Although systems of gestures and signs have been

successfully used for communication,' the more common and

preferred option is verbalization. Very short-term,

pragmatic linguistic transactions between individuals (such

as between a taxi driver and a foreign tourist) do not

constitute the necessary or sufficient conditions for the

genesis of a pidgin; technically these exchanges are

examples of jargon, or a prpigin (Rickford and McWhorter

1996:241), a form of speech which entirely lacks fixed norms

(Holm 1988:5).5 For a pidgin language to emerge, there

must be two or more language groups in contact who use a

mode of speech that is not the fully realized, adult-level

language of either group (Bakker 1995:26; see also note 2).

Characterizations of a pidgin language vary among

scholars, as do the definitions of such basic linguistic

concepts as "word," "sentence," "dialect" and even
"language" itself, but there is common agreement among

specialists that there j such a group of languages, and

that they exist or have had a previous existence recorded in

all parts of the world (Romaine 1994:163; Smith 1995; see

also Holm 1989 and Reinecke 1937). Although the exact

definitions for a pidgin language vary somewhat from

linguist to linguist, it is clear that pidgins share certain

similarities: (1) they arise in situations of extended or

repeated contact between groups of speakers of two (or more)

languages, whether through trade, enslavement or migration

(Rickford and McWhorter 1996:238); (2) they are (with rare

exceptions) not spoken natively by any group; (3) the

lexicon is generally derived from the group with more power

(the superstrate or lexifier language); (4) they possess

greatly reduced syntactic, morphological and lexical

structures in comparison with the lexifier language; and (5)

they must be learned.6

Although pidgins serve well enough for limited,

typically commercial purposes, they do not possess the range

of expressive resources necessary for them to be considered

adequate for other functions. There do arise contact

situations in which communicative requirements expand from

the commercial into other domains, and the original pidgin

is pressed into service to meet those new and quite

different demands. In such instances, it is possible for

stable pidgins and expanded pidgins to appear (Muhlhausler

1986). An expanded pidgin is considered to be a pidgin that

has become the primary language of its speakers before

"nativization" has taken place (Rickford and McWhorter

1996:246). Africanized Spanish, particularly in the

Caribbean, is believed by several leading scholars to have

its origins in an expanded Portuguese-lexified pidgin spoken

on the west coast of Africa.

Portuguese Contacts in Africa

The history of Portuguese seafaring in the 1400s is

thrilling not only for its sheer audacity but for its wide-

ranging linguistic implications. Naro (1978:317) recounts

that until the last quarter of the 15th century Portuguese

was an unknown language on the West African coast, but about

a century later, and certainly by the first quarter of the

17th century, it had become quite widespread there.

Although the first contacts between Europeans and sub-

Saharan Africans were without doubt conducted on the African

continent, Naro believes that there was no on-site


The possibility that the contact of ocean traders and
Africans would give rise to a "trade language" was
simply not fulfilled. If any direct linguistic contact
was established between African and European, this


occurred only in Europe, when the Africans were taught
some version of Portuguese, questioned about their
country, and trained to serve as guides or
interpreters. In view of the principal motivation for
the system of verbal communication that developed in
this way I propose to call this system the
reconnaissance language. (1978:319-20; original

Whether this "reconnaissance language" was taught to

Africans brought to Europe or whether it emerged initu in

Africa is of some consequence for theories of pidgin and

creole development.7 At this point, however, it is

sufficient to state that Africans became a notable presence

on the Iberian peninsula soon after regular commercial

activity began between Portugal and West Africa in the late


Literary Evidence of Africanized Portuguese

As noted above, there is good reason to accept the

existence of a so-called "black" Spanish, as there are

numerous instances of this speech recorded in the literary

works of Spain, beginning in the 16th century. Yet it is

not in Spain, but rather in Portugal, where the first

attestations of an Africanized European language are


The Portuguese literary predecessors of "black" speech

are critical to the later appearance of Afro-Spanish

language both in verse and in the theatre. As demonstrated

in depth in Chapter 3, the Portuguese were the first

Europeans to organize the trade in African slaves, building

upon the advantage of their pioneering explorations of the


west African coast. The first sub-Saharan Africans arrived

in Portugal as early as 1441 and later arrived in such

numbers that a significant portion of Lisbon's population

was African from the late 15th century through the mid-16th

(Saunders 1982). It would be unusual, then, not to find

Africans appearing in some capacity in the theatre of the

period; most frequently they were presented as comic


Although Africans first appear in Portuguese literature

in 1516 in the anthology entitled Cancioneiro Geral,' one

of the most useful sources of early Afro-Iberian speech is

undoubtedly the dramatic works of Gil Vicente (ca. 1470 -

ca. 1536). His literary production included many appearances

by Africans, as well as other marginalized groups such as

Moors, Gypsies and Jews. The most in-depth linguistic

analysis of Vicente's dramatic language is found in Teyssier

(1959). Although his study concentrates on the phonetic and

phonological presentation of Africanized Portuguese,

Teyssier's work also reveals some of the syntactic and

morphological characteristics that are later employed by his

Portuguese imitators in the Vicentine school0 as well as

in Spanish theatre and poetry. Although Teyssier does not

discuss syntax, an examination of the excerpts he has

reproduced shows, for example, reduced pronominal morphology

(mi, a ni in lieu of standard eu 'I'); occasional loss of
the copula ser in conjunction with the occasional use of the

bozal copula s 'is/are/be'; lack of agreement between
adjective and noun; frequent use of the infinitive and 3 sg.

verb forms instead of standard conjugated forms (Teyssier

Beginning with the first depiction of an African woman

in his Fr~goa d'Amor (1524), the Portuguese spoken by

Africans in Gil Vicente's theatre exhibits the following
phonological features as identified by Teyssier (1959):"
(1) Loss of final consonants /r/, /s/ and /1/:

renr> purugu[nltaa 'to ask'; qr > q 'to
want'; Tordesillas > TQ eilla 'id.'; tmos > temo 'we
have'; otal > Purutugd;

(2) Paragoge (especially in words with post-tonic,
word-final /r/): cover > chovere 'to rain'; molher >

muyere 'woman'; a > amoro 'love'; senhor > senhoro 'sir';
(3) Consonant-cluster simplification: triste > trisse
'sad'; vestido > besiro dresseded]; outra > ota

(4) Substitution of /d/ by [r]: dnheir > rinheiro
'money'; tudo > t 'everything';
(5) Neutralization of /r/ and /r/: Multiple vibrant
/r/ is simply replaced by alveolar /r/: terra > tera

'land'; ferro > Iero 'iron';"2
(6) Word-internal substitution of /1/ by [r]: ele >

ere 'he'; sjmnet > saramonete 'red mullet';

(7) Widespread vocalic variation and instability:

Derguntando > bruguntando 'asking'; folgar > furugd 'to

enjoy'; fermso > farmosa 'beautiful'; me fard > ma faraa

'he will make me'; minha > minho 'my/mine';

(8) General reduction of diphthongs (monophthong-

ization): pDja > pg.a 'then'; deixa > dexa 'he allows';

S> tredora 'traitor'; muito > muto 'very; much';

dou > r 'I give';

(9) Widespread aphaeresis: enojada > n 'angry';

escrivAo > crivam 'scribe'; escravo > cravo 'slave';

e > condiro 'hidden'.

The attempts of African characters to produce

Portuguese speech were certainly amenable to comic treatment

given the sensibilities of the era, but at the same time

Teyssier insists that this language of the stage does in


veut rdproduire, ou tout au moins dvoquer dans ses
grandes lignes, le parler rdel des esclaves nbgres
que l'on trouvait en grand nombre au Portugal &
l'4poque de Gil Vicente.

'intend to reproduce, or at the very least broadly
evoke the real speech of black slaves that were found
in large numbers in the Portugal of Gil Vicente's era.'

Although Teyssier describes the idea of linking these

forms to African substrates as "tempting," he simultaneously

warns against such "dangerous" undertakings, noting the

large variety of African languages that could have served

for any number of substrates. Moreover, he adds that there

are no words or forms of African origin in Vicente's work,

such as those that have entered Brazilian Portuguese.

Teyssier does recognize, however, the numerous points of

overlap between the characteristics of the literary lingua

de preto 'black language' and those of modern Portuguese-

lexified creoles ("les parlers n4gro-portugais") of west

Africa. Among other elements, Teyssier points specifically

to the lack of final /r/ in infinitives, confusion between

/l/ and /r/ and the reduction of diphthongs (1959:248-49).

Naro (1978) uses precisely these and other literary

data to attempt to demonstrate that an Africanized or

pidginized Portuguese was spoken first in Europe, and that

it was learned and used by Africans who had been brought

there originally and specifically for that purpose (318).

On the basis of historical and textual evidence, he shows

that the first example of pidginized Portuguese spoken by an

African character ("A min rrey de negro estar Serra Lyoa"

from 1455) antedates the first hint of any use of Portuguese

in Africa by more than 20 years (1978:320). Although Naro

is careful to avoid making facile comparisons with the

phonotactics of west African languages it nevertheless would

be unrealistic to assume that the characteristics listed

above could be ascribed wholly to a simplified "foreigner

talk" phonology as produced by European teachers.

("Foreigner talk" has been described as "a mixture of

cultural conventions and genuine natural intuitions on


language simplification" [Holm 1988:106] produced by native

speakers for the benefit of non-native speakers and language

learners.) Teyssier himself recognizes this by allowing for

such items as the confusion of liquids /1/, /r/ and the

tendency to reduce diphthongs as being of African influence.

Literary Evidence of Africanized Spanish

As the trade between Africa and the Iberian peninsula

increased, the number of Africans brought into Spain as

slaves from Portugal increased and reached its highest level

in the 16th century, particularly in Andalusia (Pike 1967).

In tandem with the increased presence and visibility of the

African minority population, the popularity of Vicente's and

his successors' work in Portugal gave rise to numerous

imitations in Spain, and "black" speech figures in some of

the work of Spain's greatest literary lights such as Lope de

Vega and Luis de G6ngora (Lipski 1986b:8-9). A great deal

of study has been devoted to "black" or bozal Spanish as a

linguistic phenomenon, revealing that the characteristics

already present in the Africanized speech of the Portuguese

theatre were also repeated in theatrical and poetic works of

Spain: indeed, Naro insists that the "reconnaissance

language" has identical Portuguese and Spanish versions

(1978:327). It should be noted, however, that the original,

most pidgin-like features of early Golden Age bozal Spanish

eventually evolve out of literary Africanized Spanish,

leaving primarily phonological divergences as the

stereotypical depiction of how Africans "should" speak on


Creole Genesis in Africa

As is well known, the Portuguese language played a

hugely important role in west Africa, particularly as the

language of the first 100 years of the frightful large-scale

slave trade (see Chapter 4). As the Portuguese strengthened

and expanded their commercial contacts with coastal African

kingdoms, they established a series of firia (Spanish

factorf) or slave depots on neighboring islands (e.g. Sdo

Tomd, Annobon, Cape Verde) as well as on the African

mainland (e.g. Loanda). As part of these settlements, the

Portuguese placed single European male colonists or l a

to direct the trade on-site. These l, far removed

from home and family, predictably intermarried with African

women. Their offspring, the filhos da terra or the so-

called 'children of the land', would have been the first

propagators of a true Portuguese-based creole. Granda

(1978:229) believes that the sociocultural conditions in the

Portuguese-influenced areas of west Africa were perfect for

the genesis of creole language (cf. Alvarez Nazario


The disagreements between linguists regarding

definitions have already been duly acknowledged above.

Despite all caveats, it is nevertheless fairly

uncontroversial to state that a creole language is a

pidgin that has acquired native speakers."5 In the case of

the Portuguese living in west Africa, the offspring of the

lancados and African women would clearly have been those

first native speakers. It is virtually impossible that those

first native speakers, children of the dominant Europeans,

would have been permitted to fill the cargo holds of the

slave ships headed for the distant markets of Lisbon,

Seville and the New World. They would have remained

resident in the Portuguese African territories and would

have formed the nucleus of a local mestizo elite.

Processes of Creole Development and Evolution

It is clear that Portuguese-based creoles developed in

west Africa and that they are still spoken today, e.g. S~o

Tomense (Ferraz 1979) and Cape Verde Crioulo (Lopes da Silva

1984) among several others (see Smith 1995:348-49 and Holm

1989 for a more complete listing). How creoles developed

not only in west Africa, but in other areas of the world has

been and still is the subject of lively theoretical debates

with implications for general linguistics and particularly

language acquisition (e.g. Andersen 1983, Wekker 1996).

The theoretical issues involved in pidgin and creole

language studies move quickly from competing definitions

into more substantive questions of development.6 The

common underlying assumption is that a creole language will

have developed from an "impoverished" or reduced base

language, whether this is defined as a pidgin or a jargon.

The most frequent scenario of creole development begins with

a jargon. This jargon then acquires socially accepted

syntactic and lexical norms as it becomes a stable pidgin.

The stable pidgin becomes an expanded pidgin before it

coalesces ultimately into a creole language (MUhlhdusler


The variety of definitions and processes of development

can lead to widely divergent research agendas. For example,

under one possible path of development, the types of creole

considered to be the purest for studying linguistic

universals are those arising from "situations in which the

human linguistic capacity is stretched to the uttermost"

(Bickerton 1981:4). As a result, in this particular school

of research (Bickerton's language bioprogram hypothesis), a

creole is restricted to mean gn1y languages which:

(1) Arose out of a pidgin which had not existed for
more than a generation.

(2) Arose in a population where not more than 20
percent were native speakers of the dominant language
and where the remaining 80 percent was composed of
diverse language groups. (1981:4)

Under these two conditions, numerous creoles (e.g. Tok

Pisin, Rdunion Creole and perhaps the creole of the

Portuguese enclave of Macao) are excluded from analysis

under the bioprogram hypothesis, as Bickerton acknowledges

(1981:4). Consequently, his study is limited in scope to

Hawaiian Pidgin and Hawaiian Creole English. While such

restrictions provide for much cleaner data and a more

rigorous methodology for testing his hypothesis, they

thereby exclude an enormous range of languages that by any

other definition(s) are clearly creoles. The unintended

effect is that under this program, creoles without a clear

date of origin (i.e. almost all others) are again pushed to

the periphery of serious linguistic inquiry (see McMahon

1994:270-80 for a discussion of the pros and cons of this

research hypothesis). It is very probable that the

Portuguese creoles of west Africa do not meet Bickerton's

criteria for testing the bioprogram hypothesis."

The terminological and related theoretical issues do

not stop, however, with definitions and disputes regarding

which language is "more creole" than another. In addition,

there exists the further possibility that a language may

fall into a classification known as a crejlj (also termed

semi- or quasi-creole). A creoloid language is one that

does not meet the criteria for being considered an ideal

type of creole, i.e. one that has either developed outside

those pathways considered typical of creole development or

has emerged under very fluid social environments. Three

such circumstances have been described as follows:

(1) Whereas true creoles develop where there is a
radical break in language transmission, many
languages appear to have developed with only a
partial break.

(2) A number of languages with no known pidgin
ancestor nevertheless exhibit many of the alleged
typological properties of creole languages.


(3) Next to mixing between fully developed linguistic
systems one also finds mixing between full systems and
developing systems, such as pidgins. The results of
this second type of mixture tend to be structurally
similar to established creoles. (Muhlhdusler 1986:10)

More succinctly, Holm summarizes the outcome of these

circumstances by noting that a creoloid language is one that

superficially resembles creole languages in some way (e.g.

by being morphologically simpler than a possible source

language), but which appears never to have undergone

creolization (1988:10).

The complexities surrounding the origin and development

of pidgins and creoles do not, unfortunately, stop even at

that point. A creole language, however defined or

originally formed, may continue its evolution along a post-

creole continuum (DeCamp 1971). The post-creole continuum,

as its name suggests, is a system composed of gradients.

The variety of the creole that is maximally distant from the

lexifier language is called the basilect; the variety most

approximating the lexifier language, the acrolect. Between

these two polar varieties exists a series of intermediate

types or mesolects. In a famous analysis of Jamaican

Creole, DeCamp was able to chart the Jamaican post-creole

continuum as an implicational scale, in which the presence

of a feature (basilectal or acrolectal) would imply the

presence of other features. Although suggestive and

groundbreaking at the time, there were some aspects of the

data analysis that proved troubling, and implicational


scaling for creoles has been largely abandoned. The concept

itself has remained attractive, however, and the term post

creole has survived as a useful term for describing a creole

language under pressure from the local standard of the

lexifier language to restructure itself towards the standard

as part of the process known as decreolization (cf. Megenney


Adding two final notes of caution, it should not be

assumed that there is a direct process that proceeds

inexorably from jargon, to pidgin, to stable pidgin, to

expanded pidgin, to creole, to post-creole, ending in an

inevitable merger with the lexifier or superstrate language.

First, it must be remembered that "[c]reolization can take

place at any point during the pidgin's life cycle, ranging

from a jargon to an expanded pidgin" (Romaine 1994:169; cf.

M~hlhdusler 1986:8).-" Second, a language may also

undergo a process of recreolization in which (typically

younger) speakers will deliberately choose to employ a more

basilectal level of speech, a process noted for West Indian

creole English in the United Kingdom (Romaine 1988:188-203)

and Tok Pisin in Papua New Guinea (Romaine 1994:171). As we

see from this brief discussion alone, there is no more a

predestined course of development or guaranteed evolutionary

outcome for pidgins and creoles than for any other natural


Monogenetic and Polygenetic Theories of Creoles

Although there have been increasing levels of

theoretical sophistication in creole studies in general,

most work relating to the origin, transmission and survival

of Africanized Iberian language(s), whether termed

"pidginized," "creolized" or "creoloid," has tended to

follow older traditions of inquiry. These earlier proposals

have the disadvantage of not being as nuanced as some of the

proposals discussed above; in comparison to both the rigor

and subtleties of, for example, Bickerton's (1981, 1984)

Language Bioprogram Hypothesis, they appear fairly simple.

Some of these proposals, nevertheless, have devoted

adherents to this day, as will be shown.

These older perspectives of analysis may be divided

into theories of monogenesis and polygenesis. The

monogenesis hypothesis in its strongest form proposes that,

because of their similarities in structure, many (if not

most) pidgins and creoles throughout the Pacific and

Atlantic developed as a result of contact with an earlier

nautical pidgin Portuguese, which in turn had its likely

origins in the Mediterranean trade language known as Sabir;

the overt differences in lexicon may be attributed to a

process of relexification (Whinnom 1965), whereby the

original lexical items are replaced by others from a new

superstrate language. In such cases, a new dominant

language would have been one of the results of change in

colonial administration: e.g. the original Spanish

superstrate of Jamaica was supplanted by English. The

weaker version of the hypothesis concentrates more modestly

on the similarities between the Atlantic creoles and

pidgins, whether derived lexically from English, French,

Dutch, Portuguese or Spanish. In particular, the concept of

relexification of an Afro-Portuguese pidgin to Africanized

Spanish is very attractive and plausible in light of the

high degree of mutual intelligibility of 15th and 16th

century Portuguese and Spanish (cf. McWhorter 1995:232). In

their important work, both Teyssier (1959) and Naro (1978)

(building on Teyssier's observations) make this point

explicitly. Confirming what had been simply a piece of

received wisdom, more recently Jensen (1989:851) has

demonstrated experimentally that native speakers of South

American Spanish and monolingual Brazilian Portuguese

speakers are able to understand each other's recorded spoken

language with a 50% to 60% degree of accuracy without any

prior training. This finding is important in that modern

American Spanish and Brazilian Portuguese are far more

dissimilar than were their respective 15th and 16th century

Iberian antecedents.

The proposals in favor of polygenesis, on the other

hand, dispute the idea that pidgins and creoles originated

from a single source language: "By definition all theories

besides monogenesis imply polygenesis, and many believe that


pidginization and creolization occurred in different places

at different times but under parallel circumstances that

produced parallel results" (Holm 1988:52). In short, there

is no single, well-articulated hypothesis regarding

polygenesis: the non-monogenetic views may include theories

of origin incorporating language universals; imperfect

second-language learning; and "baby talk"/"foreigner talk"

hypotheses of pidgin-creole genesis.2"

For A Monogenetic Theory of Afro-Hispanic Language

Most of the work done in the field of Afro-Hispanic

pidgin and creole languages has rested on the weaker version

of the monogenetic theory as described above; the non-

monogenetic analyses have come more as reactions to these

rather vigorously argued monogenetic proposals than as

independently motivated appraisals of those languages'

origins. Foremost among the proponents for monogenesis in

Spanish America is Granda (1978; see also 1977, 1988).

Beginning with his writings of the late 1960s and early

1970s,21 Granda has consistently championed the monogenetic

theory, specifically for areas of the Caribbean basin. He

maintains that the recorded presence of creolized (creoloid,

semi- or quasi-creole) varieties of Spanish and the large

size of the African slave population originally brought into

the region argue for the earlier existence of an Afro-

Portuguese pidgin transplanted to the Americas. One of the

most intriguing pieces of evidence supporting this assertion

was uncovered by Granda (1978:350-61) in an early 17th

century text written by a Jesuit missionary in the Caribbean

port of Cartagena (modern Colombia). In a 1627 description

of his activities as missionary to and among the African

slaves arriving aboard the slave ships, the priest Alonso de

Sandoval (1576-1651) writes tantalizingly that the creoles

and natives of Sdo Tomd communicated with other African

nations by means of

un gdnero de lenguaje muy corrupto y revesado de la
portuguesa, que llaman lengua de San Thomd, al modo que
ahora nosotros entendemos y hablamos con todo gdnero de
negros y naciones con nuestra lengua espahola corrupta,
como comdnmente la hablan todos los negros.

'a type of very corrupted and contorted Portuguese
language that they call language of Sdo Tomd, in the
same way that we now understand and speak with all
types of blacks and nations in our corrupted Spanish
language as all the blacks commonly speak it.'
(Sandoval 1956/1627:94)

This passage does indeed seem to confirm that a

Portuguese pidgin or creole was spoken at that time in and

around Sdo Tomd as Granda (1978:356) indicates and, of

course, S&o Tom4 creole exists to this day. This same

citation alludes fairly clearly to a similar type of Spanish

being spoken in the coastal city of Cartagena itself, but

Granda goes on to draw some highly enthusiastic exegetical


En efecto, la redacci6n del p~rrafo del Padre
Sandoval no deja la menor duda de que se trata de un
habla 'criolla' implantada entre los negros esclavos de
la Amdrica espahola con cardcter de generalidad y, lo
que es mds importante, constituida del mismo modo .
que el "habla de San Thomd."

'Indeed, Father Sandoval's paragraph does not leave
the slightest doubt that we are dealing with a "creole"
speech implanted generally among the black slaves of
Spanish America and, what is more important, formed in
the same way as the "speech of Sio Tome."'

With such a reliable eyewitness account to bolster his

position, Granda speaks specifically in favor of the process

of relexification and not one, but in effect, two


[E]Il espahol de las zonas negras de la Amdrica
espahola se habria configurado en su estado actual
a travds de un proceso de relexificacidn y
reestructuraci6n hacia la lengua dominante a
partir de un b criollo, de matriz
portuguesa pero de lxico rdpidamente castellanizado.
Es a travds de este estadio criollo intermedio
como hay que concebir en general la influencia
africana en el espahol de Amdrica, y no directamente
desde los diferentes c6digos lingfisticos propios de
las diversas agrupaciones del continente negro (yoruba,
bantd, ewe-fon, etc.). [S]i consideramos el
criollo de las Areas hispanoamericanas de poblaci6n
negra como el sustrato sobre el que se ha desarrollado
el espahol actual de estas zonas, las lenguas
aborigenes africanas representarfan un
subsustrato, diferenciado segdn sus condicionamientos

'[T]he Spanish of the black areas of Spanish America
would have achieved its present state through a process
of relexification and restructuring towards the
dominant language starting from a creole .
originating from a Portuguese matrix, but with a
rapidly Hispanized lexicon. It is via this
intermediate creole stage that one must generally
imagine the African influence on American Spanish, and
not directly from the different linguistic codes of the
diverse groups from the black continent (Yoruba, Bantu,
Ewe-Fon, etc.). [I]f we consider the creole of
the black-populated Spanish American areas as the
substrate upon which the current Spanish of these zones
developed, the African aboriginal languages would
represent a sub-substrate, differentiated
according to their historical conditioning.' (Granda
1978:208-09; original emphasis)

According to this view, it is the relentless efficacy

of this same process of relexification22 and restructuring

that accounts for the scarcity of the linguistic evidence

attesting to an African language substrate in most of the

areas where the African-derived population is located: the

operation of this process, in effect, eliminates most of the

evidence of its having operated on the local language at

all. Granda suggests that the best evidence of an African

"sub-substrate" in most of the areas with an African

presence would be found in the lexicon:

[L]os escasos elementos que, procedentes de las
lenguas aborigenes africanas, se han incorporado al
espahol actual de las dreas de poblaci6n negra en
Hispanoamdrica deben ser rastreados en el nivel ldxico,
mds superficial y mejor adecuado a una tan dificil y
problem~tica transmisi6n, mientras que los
presuntamente adscribibles a los niveles fondtico y
morfosint~ctico de influencia son, de modo general
aunque con posibles excepciones, mds correctamente
analizables dentro de evoluciones internas del sistema
castellano. ...

'I[T]he few elements proceeding from African aboriginal
languages that have become incorporated into the
current Spanish of black-populated areas of Spanish
America must be traced at the lexical level, which is
more at the surface and better suited to such a
difficult and problematic transmission; whereas those
presumably ascribable to the phonetic and morpho-
syntactic levels of influence are generally, but with
possible exceptions, more correctly analyzable within
the internal evolution of the Castilian system. .... I
(Granda 1978:209-10)

Indeed, the presence of African lexical items in American

Spanish has been a subject of inquiry for many years (e.g.

Ortiz 1924), and the Spanish vocabulary is the area in which

African contributions have been the most plentiful (e.g.

Alvarez Nazario 1974, Del Castillo Mathieu 1982, Megenney

1983a) and least controversial (Lipski 1985a:99).

In this paradigm there are also certain special

sociohistorical circumstances under which closer links with

elements of a pidginized Portuguese (i.e. other than

vocabulary) may be sought with a greater likelihood of

success. Some of these special social conditions include

more recent dates of African immigration or slavery, and the

existence of religious rituals with an accompanying

"special" or "secret" language of an African or creolized

base (Granda 1978:210-11; cf. also Ortiz L6pez 1996; Lipski


Moreover, the suggestion is clear that there are

specific places where such a language would be more likely

to have survived the pressures of standardization:

Al igual que ha ocurrido en cuanto a otros rasgos
culturales, los ndcleos de cimarrones, por su actitud
psicol6gica de repulsa a la transculturaci6n y su
aislamiento y hostilidad respecto a la sociedad
hispanoamericana que los rodeaba han actuado como
"areas relegadas" tambidn en sentido lingUistico y,
en esta calidad, han mantenido situaciones arcaicas
en cuanto al habla, superadas y rebasadas mucho
tiempo antes por los grupos de afroamericanos
sometidos con intensidad a la acci6n transformadora de
la norma ling~istica general en Hispanoamdrica.
Tambidn en cuanto a la lengua los palenques o cumbes
se han comportado como recept~culos de preservaci6n
de rasgos "en conserva."

'In the same way as it has happened with regard to
other cultural features, the maroon settlements,
through their psychological attitude of rejection
towards transculturation and their isolation and
hostility to the Spanish American society which
surrounded them, have acted as "set-aside areas" also
in a linguistic sense and, as such, they have

maintained archaic situations regarding their speech,
situations overcome and surpassed much earlier by
Afroamerican groups intensely subjected to the
transformative action of the general linguistic norm in
Spanish America. Also with regard to language, the
palenques or cumbes [isolated and fortified settlements
of escaped slaves] have acted as jars for the
preservation of "pickled" features.' (Granda 1978:378)

Although Africans were sold as slaves in all parts of

Spanish America (Curtin 1969, Mellafe 1964, Rout 1976,

Zelinsky 1949), the areas that saw the importation of the

greatest number of Africans, that had the most factories

(transshipment depots for slaves) and where slavery endured

the longest time are all in or around the Caribbean. As a

result, linguistic research on the possible remnants of an

Afro-Portuguese pidgin or creole in Spanish has naturally

concentrated on this region, particularly the Greater

Antilles (Puerto Rico, Cuba and Hispaniola [the Dominican

Republic]) and those mainland countries along the Caribbean

basin (Venezuela, Colombia and Panama).23

Following the lines of research described by Granda

(1978), a great deal of work has been focused on the

Caribbean area where African immigration remained strong

well into the 19th century. William Megenney (1982, 1983a,

1983b, 1984, 1985a, 1985b, 1985c, 1986, 1989, 1990a, 1990b,

1990c, 1990d) is one of those linguists who has clearly

adopted the hypothesis that an Afro-Portuguese pidgin

underlies the vernacular Spanish of the circum-Caribbean.

For example, he states (1990a:19) that a pidgin language

formed on Sdo Tomd was carried by Africans to the New World

where it served as the basis for Caribbean creole Spanish.

Traces of this Caribbean creole Spanish, he maintains, are

found in modern Papiamentu, Palenquero, the speech of the

Choc6 in northern Colombia, and the "bozal speech" of Cuba.

Consistent with Granda's position, Megenney posits a pidgin-

creole language as a "filter" through which African speech

passed on its way to vernacular Spanish (1985a:230). In the

case of the Dominican Republic, he even asserts that the

European population, outnumbered by Africans, adopted a

common creole language "basado en el lenguaje africanoide

que hablaban los esclavos y ampliado por el castellano de

los colonos" 'based on the Africanized language that the

slaves spoke and expanded by the Castilian of the colonists'


The evidence he marshals in favor of this position is

not found generally in the syntax of circum-Caribbean

varieties of Spanish, but rather in surviving Portuguese

vocabulary and usages, lexical items derived from African

sources and phonetic characteristics. Megenney has done

extensive work on the speech of the Dominican Republic, the

speech of coastal Venezuela and Palenquero. It is his work

on the non-creolized varieties of Spanish that are of

interest, as the status of Palenquero as a creole with Afro-

Portuguese origins is not at issue but rather the existence

of an earlier pan-Caribbean Afro-Portuguese pidgin that has

left detectable traces in modern vernaculars.

Megenney has found in the case of the Dominican

Republic evidence of African-derived vocabulary (1982,

1990a), and relates such phenomena as vocalization of final

infinitival /-r/, the neutralization of liquids /l/ and /r/

and the variability of vowel realizations to an African

base. Indeed, as noted above, he attributes the entire

spectrum of Dominican pronunciation to an Africanized base

augmented by colonial Spanish. Another author (Lorenzino

1993) working on the same topic finds support for a process

of semicreolization or partial creolization (cf. Holm 1988),

based on some limited morphological simplification and

higher-than-standard use of subject pronominals resulting

from elided /s/ in verbs.

In the case of Barlovento, Venezuela (1985a, 1990b:153-

59, 169; 1990c:246-56) his work presents many of the same

correspondences (e.g. a high degree of nasalized vowels;

loss and aspiration of implosive and final /s/;

neutralization of liquids /1/ and /r/; alternation of /r/,

/1/ and /d/; velarization of final /n/) which he interprets

as possible evidence of an Africanized substrate. Other

work (1984, 1985b) treats evidence of Portuguese in two

undisputed creoles (Papiamentu and Palenquero) and in the

so-called Afro-Cuban creole, whose existence and status has

been and remains controversial (see Martinez Gordo 1982).

Also following the monogenetic hypothesis championed by

Granda is the German linguist Matthias Perl. Where Megenney

has concentrated his attention chiefly on the Dominican

Republic, Venezuela, and Colombian Palenquero, Perl (1984,

1985, 1988, 1989a, 1989b, 1989c) has done extensive

investigation precisely on the above-mentioned Afro-Cuban

creole (habla bozal). Besides Perl, several other authors

(Otheguy 1973; Granda 1978; Megenney 1984, 1985b; Fernandez

Marrero 1989) accept the premise that this Afro-Cuban speech

derives from an original Afro-Portuguese pidgin or creole.

Based primarily on morphological and phonological evidence

(the use of ta [from estd 'is'] followed by infinitive; the

3 sg. pronoun k [alternating with nA] undifferentiated for

gender; neutralization of liquids /l/ and /r/ [Otheguy

1973]; lack of common prepositions _q 'at/to' and dg

'of/from' and relative pronoun que 'that/who/which'; reduced

verbal morphology; loss of definite article; obligatory use

of subject pronoun; reduplication, etc. [Perl 1984, 1985,

1988 1989b]), supplemented by historical texts (cf.

Fernandez Marrero 1989), Perl (1989a, 1989c) finds links

with Portuguese-based creoles in the Caribbean and in

Africa, bolstering Granda's monogeneticist position,

although he himself does not describe Afro-Cuban as a full-

fledged creole.

Although Armin Schwegler is best known for his in-depth

(e.g. 1996a) research into the Palenquero creole language in

Colombia, he has also recently related the phenomenon of

double negation (e.g. no sd no in contrast to standard

Spanish "simple negation" n 'I don't know') found in

Palenquero (1991) to the same feature in Dominican Spanish

(1996b). This syntactic development has been recorded in

the speech of Afro-Latin Americans in Brazil, Colombia and

Cuba but is unknown in European varieties of Spanish and

Portuguese and is unattested in the extant bozal texts.

Schwegler claims a special status for double negation in

contrast to other traits advanced in favor of monogenesis

porque los oponentes de la teoria afroportuguesa
ya no podrdn, como siempre han podido hacer, invalidar
la evidencia al invocar la posibilidad (remota o no) de
influencias africanas substratisticas directas, cambios
internos, u origenes dialectales peninsulares. ...

.. because the opponents of the Afro-Portuguese
theory can no longer, as they have always been able to
do, invalidate the evidence by invoking the possibility
(remote or not) of direct African substratal
influences, internal changes or [Iberian) peninsular
dialectal origins. .' (1996b:290)

Indeed, this particular phenomenon does not appear to have

any Ibero-Romance language parallel outside the African-

settled areas of this hemisphere and may be the strongest

syntactic evidence to date for the existence of an Afro-

Portuguese pidgin or basilect.

Questioning a Monogenetic Theory of Afro-Hispanic Language

At this point we must indicate that not all researchers

share Granda's sanguine views of a transplanted Afro-

Portuguese pidgin. For example, Alleyne (1971), Laurence

(1974), L6pez Morales (1980), and more recently McWhorter

(1995) point out the varying circumstances and periods of

the slave trade in the Spanish Caribbean--especially the

disparate origins and languages of the enslaved Africans--

that may account for the substantial lack of Spanish-based

creoles in the circum-Caribbean. Goodman (1987b) takes

sharp issue with the assumption of a widespread Portuguese

pidgin along the African coast, finding (in contrast to Naro

1978) little evidence of its being used among the captives

destined for slavery. He believes, rather, that the

prisoners probably used an African language to communicate

among themselves:

[I]t has often been suggested that the slaves, speaking
many diverse languages, learned Pidgin Portuguese in
the various trade forts and slave barracks along the
coast (where they were often held many months and even
occasionally as long as a year or more prior to
shipment) as a means of communicating not merely with
their European captors, but even among themselves. Yet
the slave populations of these locales were far more
homogeneous linguistically than those of the various
colonies to which they were ultimately sent. The
predominant local language, Kimbundu and Kikongo in the
Bantu area, Ewe-Fon on the Slave Coast (modern Dahomey
[now Benin] and Togo), Akan on the Gold Coast (modern
Ghana), Wolof in Senegal, and so forth, would, in all
likelihood, have served as the captives' lingua franca,
to the extent that one was needed. Even those who did
not speak it upon arrival at the coast were more likely
to learn it in captivity than any European language,
and in the interim they could surely have found people
to interpret for them. (Goodman 1987b:364)

Moreover, a closer reading of the very text cited by

Granda as early evidence for a widespread Afro-Portuguese

pidgin reveals that the communicative situation was not as

clear as Granda has depicted it (see also Ortiz L6pez

1996:42-43). For example, Sandoval refers specifically to

the need for interpreters in ministering to the temporal and

spiritual needs of his African parishioners:

Habida la noticia del capitulo pasado, sirve de poco si
no hay intdrprete o lengua de la naci6n del enfermo o
adulto sano a quien por su medio se ha de catequizar,
bautizar o confesar, si ellos no entienden la nuestra o
nosotros no sabemos la suya. Y la dificultad estA en
que de ordinario sus amos no tienen intdrpretes ni se
les da nada por buscarlos; y nosotros [sic] parece
moralmente imposible que aprendamos todas estas lenguas
por ser tanta su multitud y no haber alguna general.

'In light of the previous chapter, it does little good
if there is no interpreter or native language of the
ill or healthy adult by means of which he is to be
instructed, baptized or confessed if they do not
understand our language or we do not know theirs. And
the difficulty is that their owners ordinarily do not
have interpreters nor do they care at all to look for
any; and it seems morally impossible for us to learn
all these languages on account of their great multitude
and because there is no general language. .. .
(Sandoval 1956/1627:335)

According to Father Sandoval's account, many Africans

were quite multilingual, but in other African languages.

From the context of the excerpt which follows, Sandoval

makes a clear distinction between the (semi-)acculturated,
"Latinized" Africans (ladinos) speaking various African

languages as well as Spanish, and those arriving directly

from Africa (bozales) whose linguistic repertoire was

strictly African:

Y porque asi como las lenguas e intdrpretes ladinos
suelen hablar varias lenguas, asi los negros bozales
tambidn las suelen hablar y entender, para que asi se
les bautice, confiese y remedie con cualquiera de las
que entendieren, sin que haya de ser fuerza busc~rseles
las suyas genuinas y naturales. ..

'And just as the lading interpreters usually speak
several languages, so the bozal blacks usually speak
and understand them as well, so in that way they may be
baptized, confessed and saved using any of the ones
they may understand, without it being necessary to seek

out for them their own genuine and native languages.
S.' (Sandoval 1956/1627:338)

From these references it is not difficult to wonder why, if

indeed so many slaves arrived in Cartagena speaking some

pidginized version of Portuguese, there would be such a need

for African interpreters. Sandoval, who was by all accounts

a learned, pious, dedicated and compassionate man, would

have had little difficulty in making himself understood in

light of the above-cited mutual intelligibility of

Portuguese and Spanish, considering especially the extreme

similarity between sacramental and theological terms in the

two languages. His motivation to be understood would have

been extraordinary since we see from the citations above his

sincere concern for the salvation of the Africans' souls:

Sandoval genuinely mourned the death of slaves who died

outside the Church and who therefore lacked the certainty of

entering paradise.2"

Although the recent observations of Schwegler (1996b)

present a challenge to the analyses and critiques of

"opponents" of monogenesis, virtually all other data from

circum-Caribbean dialects of Spanish and historical texts

have indeed been analyzed by various scholars who find solid

reasons to question their inclusion as positive evidence for

monogenesis. For example, those who favor the hypothesis of

an Afro-Portuguese basis for Caribbean vernacular Spanish

often cite Cuban commentaries and texts of the late 18th and

early 19th centuries (e.g. Fernandez Marrero 1989; Perl

1984, 1989a respectively). Work by the Cuban author Lydia

Cabrera, who became renowned for her transcriptions of

stories told to her by elderly Afro-Cubans in their own

lengua bozal has also frequently been presented as reliable

evidence that remnants of an Africanized Spanish were heard

in Cuba through the 1950s. Some linguists--notably two from

modern Cuba--have been skeptical of such claims: Valdds

Bernal (1978, 1987) and Alpizar Castillo (1989). Valdds

Bernal makes several pertinent observations. He notes that

it is difficult to determine which phonetic influences on

Cuban Spanish are specifically African and not simply

coincidental with European Spanish; even more difficult to

identify are sub-Saharan influences on the morphology and

syntax of Cuban Spanish in light of the numerous west

African nationalities who entered Cuba, no single one of

which ever clearly predominated over the others (1978:102).

Although there has been an undeniable influence on the

lexicon from African languages, Valdds Bernal also calls

into question the reliability of the transcriptions of texts

(i.e. librea 'notebooks') in which "African" ritual

languages were recorded, as well as comments by 19th century

writers (e.g. Esteban Pichardo) and Lydia Cabrera. With

regard to Lydia Cabrera, he notes that songs and stories in

bozal language were handed down from generation to
generation, and that this does not mean that this speech

remains alive--a belief that he attributes specifically to

foreign scholars (1978:89).

Also questioning the interpretation of early texts,

Alpizar Castillo finds that an early catechism intended for

the instruction of newly arrived African slaves in Cuba (cf.

Fernandez Marrero 1989), supposedly written in bozal speech,

shows very few differences from Spanish: there are more

aspirations than standard Spanish; some archaisms; some

apocopated forms found predictably in informal, rapid speech

(e.g. pA < R 'for'; na < nada 'nothing'); disregard for

or misuse of verb tenses; the non-standard use of some

Spanish vocabulary. In short, he observes that the most

that could be said about bozal speech is that there were

regionalisms and archaisms mixed with errors of syntax

typical of those learning the language from Europeans who,

in turn, did not speak it well themselves (Alpizar Castillo


A broad survey of Afro-Caribbean Spanish is provided by

Lipski (1993). Identified specifically by Schwegler

(1996b:290) as one of the "opponents" of the Afro-Portuguese

monogenetic theory, Lipski argues that "(t]he presence of a

stable creole, if one existed at all, is overshadowed by a

wide range of pidginized varieties, none of which embodies

the creole structures which support the 'relexification' of

a previously-acquired Portuguese" (Lipski 1993:2).

Similarly, Ortiz L6pez (1996) finds evidence in present-day

Cuba that would allow for a partially creolized variety, but

not a full-fledged creole.

Ibero-Romance Creoles

In spite of any shortcomings in the monogenetic theory

as elaborated by Granda, the research carried out in that

framework on Caribbean varieties of Spanish has yielded some

interesting lexical and phonological data, but there is no

evidence to date that points unequivocally to the prior

widespread existence and use of a Portuguese-based pidgin or

creole among Africans shipped to this hemisphere. This is

not to say, however, that creole languages with a strong

African content never developed from an Iberian base in the

circum-Caribbean. Granda's hypotheses seem to find support

especially in the case of Palenquero, a creole language of

northern Colombia which has been and continues to be the

object of intense linguistic scrutiny (e.g. Montes Giraldo

1962, Bickerton and Escalante 1970, Friedemann and Patifio

Rosselli 1983, Megenney 1986, Patifio Rosselli 1989, Granda

1989, Schwegler 1996). This language, spoken by

approximately only 4,000 people and which is at risk of

extinction (Lipski and Schwegler 1993), presents many of the

characteristics considered typical of creoles such as pre-

verbal tense, mood and aspect (TMA) markers, limited

pronominal allomorphy, invariable verb forms and shows,

moreover, a considerable Bantu influence. Palenquero has

its origins in the speech of maroons and their descendants

who settled in a remote mountainous area; these

circumstances are consistent with the special

sociohistorical conditions mentioned by Granda (cited


Also following special circumstances of development is

Papiamentu (more frequently spelled Papiamento). Spoken on

Curagao, Aruba and Bonaire by some 200,000 people,

Papiamentu developed from either Portuguese or Spanish

(which Iberian language is the lexifier still occasions

vigorous debates) but has shown recent trends favoring

Hispanization due to the islands' proximity to Venezuela.

Papiamentu also shows some lexical influence from Dutch (the

superstrate language since 1634) and English. The abrupt

change of colonial overlord followed by arrival of

Portuguese-speaking Sephardic Jews (see Goodman 1987b) and

the concomitant loss of contact with the original

metropolitan Spanish favored the consolidation of a local

speech variety into a creole. Papiamentu, in contrast to

other creole languages, enjoys high local prestige. There

is also some official recognition of Papiamentu; it is used

in a number of newspapers; it has a well-established

literary tradition, is a language of broadcasting and it has

recently come to play a role in education. Dutch, however,

continues as the dominant language of instruction and of

government (Kouwenberg and Muysken 1995; Appel and

Veerhoeven 1995; for origins and grammatical information see

also Lenz 1928, Goilo 1953, 1972; Wijk 1958, Ferrol 1982,

Goodman 1987b). Palenquero and Papiamentu are the only two

surviving creole languages in the hemisphere that may have

developed from Spanish; in each case they also have a clear

present link with neighboring varieties of Spanish.

Features of Afro-Hispanic Caribbean Speech

Because the scourge of slavery was for so long so

heavily concentrated in the circum-Caribbean, phonological

traits associated with that region have frequently also been

associated with the speech of Africans and their descendants

in general. These phonological characteristics, which have

been said to comprise the elements of a "radical" (as

opposed to a "conservative") pronunciation (Guitart 1978),

include such traits as the aspiration and/or deletion of

word- and syllable-final /s/, velarization of word-final

/n/, neutralization of the phonemic distinction between the

liquids /l/ and /r/; the realization of /d/ as [r] also is

cited as a particularly Afro-Hispanic attribute (cf. Lipski


Slavery in Spanish America was, of course, not confined

just to Cuba, Puerto Rico, Santo Domingo and the coastal

zones of Panama, Colombia and Venezuela (Mellafe 1964,

Curtin 1969). During the course of the trade, African

slaves were transported to virtually every area of the

continent, including Ecuador and Peru on the south Pacific

coast, and Uruguay and Argentina on the south Atlantic


coast. It has been tempting to see an African influence in

the phonology of Caribbean varieties of Spanish, but many of

the same features (e.g. aspirated or deleted syllable- and

word-final /s/, velarized final /n/) appear in regional

dialects that have had virtually no connection with, or

influence from, African contacts, such as Chile and Paraguay

(cf. Canfield 1981:17). The phonological and syntactic

similarities between the dialects of southern Spain and the

Canary Islands and "radical" dialects of Spanish America

(particularly the Caribbean) have complicated all attempts

to disentangle "African" from "Andalusian" language traits.

In light of the fact that many of the port cities of

America, both on the Atlantic and Pacific, were also those

that served as slaving depots and exhibit "radical"

features, the conflation of African and Andalusian features

is a nearly intractable problem. Given a proper combination

of demographics (e.g. Boyd-Bowman 1964) and Old World

Spanish regional dialects (e.g. Cataldn 1958), the behavior

of "Africanized," "Caribbean" or "radical" Spanish may be

plausibly and completely accounted for without invoking an

Afro-Portuguese component (Lipski 1994a:133). This fact

places additional importance on those few areas of African

settlement that are fairly distant from the historic slaving

ports, such as the rnmiciQ of Cuajinicuilapa: it is only

in such areas that there is a reasonable possibility, as

Granda has noted, in discovering authentic African or Afro-

Portuguese linguistic elements.

As noted above, apart from Papiamentu and Palenquero no

clear Afro-Iberian creole languages emerged in the Americas.

Nevertheless, there are particular syntactic and

morphological configurations in addition to the phonological

characteristics of "radical," "Caribbean," or bozal Spanish

that have been used to support the putative earlier

existence of a generalized Afro-Iberian creole or pidgin

brought from the coasts of Africa. The following syntactic

and morphological features have been typically identified

with Caribbean bozal (creoloid, semi-creole) speech (Lipski


(1) Non-inverted interrogative sentences (e.g. t

quieres? instead of standard ZOud quieres ti.i? 'What do you


(2) Obligatory use of subject pronouns that in

standard Spanish are optional and typically suppressed;

(3) "Personal" infinitives in such constructions as

para td hacer eso in contrast to standard para aue hagas eso

'for you to do that', suggestive of the Portuguese personal


(4) Lack of gender and number agreement in nouns and


(5) Loss of the common prepositions 'at/to' and d


(6) Occasional copula deletion;

(7) Loss of articles;

(8) Postposed demonstratives (W I instead of esa

piedra or la piedra esa 'that stone');

(9) Lack of the obligatory subordinating conjunction

a 'that';

(10) Use of the pronoun vos, particularly in areas

where the standard pronoun of informal/intimate address is

td 'you';

(11) Use of the generalized preposition na 'in/to

[the]' of almost certain Portuguese origin (but questioned

in Alleyne 1971:170);

(12) A third person subject pronoun Q or el unmarked

for gender;

(13) The use of tenet 'to have' as the existential

instead of standard haber;

(14) Frequent preposing of adverb ms 'more' in

negative expressions such as mAs nunca, ms nada in lieu of

standard nunca mds 'never again', nada mAs 'nothing else'.

If we include the previously cited characteristics of

early attestations of bozal Portuguese and Spanish

(reproduced here for reference: Loss of word-final /r/, /1/

and /s/; paragoge [senhor > senhoro 'sir']; consonant

cluster simplification; /d/ > [r]; neutralization of

distinctions /r/ and /1/, /F/ and /r/; monophthongization;

aphaeresis [escravQ > cravo 'slave']; generalized vocalic

instability; and frequently there is also included the

phenomenon of "intrusive nasalization" in forms such as

munglave < g 'very serious' and Minguel < Miguel

'Michael' [Megenney 1985c:191]), we have formed a matrix of

characteristics which can be used as a basis for judging, to

the degree it may be possible at this point, the likely

presence and use of an Africanized Iberian speech in Mexico

both past and present.

We have clearly seen that the contacts established by

the Portuguese in Africa led to the emergence of creole

languages in that continent (leaving entirely aside the

importance of Portuguese in Asian creoles). The existence

of S&o Tomense, Principense, Cape Verde Crioulo and other

creoles are the living evidence of the longstanding

linguistic interaction between two (and more) very different

groups of people. The growth and use of Portuguese-lexified

creoles among people living in Africa, however, has never

been at issue. It is, rather, whether the clusters of

historical, demographic, geographic, cultural and linguistic

data in toto support the hypothesis that a Portuguese pidgin

or creole was transplanted from Africa to Spanish America

and, more specifically, to Mexico.

The theories of language contact surveyed in this

chapter predict that in periods of extended or repeated

contact between two or more groups speaking mutually

unintelligible languages, the group of lesser power

(speaking a substrate language) will typically adopt some

version of the more powerful group's language (the

superstrate). How the less-powerful group acquires the

pidgin is a matter of ongoing controversy with implications

for theories of language acquisition. As contacts between

the groups move beyond the strictly commercial, the pidgin

will acquire certain minimal norms (i.e. will become a

stable pidgin) and from that point will tend to expand

further to fill the new range of communicative functions.

Although the original reduced or pidgin Portuguese (Naro's

reconnaissance language) may have been taught to Africans in

Portugal who had been brought there to learn to interpret,

the emergence of a true creole language from a pidgin occurs

at the birth of a generation of native speakers, which

clearly occurred in west Africa. Granda's (1978) work

hypothesizes that Africans destined for slavery brought some

version of this reconnaissance language with them to the New

World and that this speech achieved enough stability to

serve as the substrate for the acquisition of local

vernacular Spanish in the circum-Caribbean. It is clear

from the historical record that Africans arriving in

Portugal and Spain seldom achieved mastery of their captors'

language; the early texts from those countries show similar

divergences from the standard forms. The existence of

creole languages (Papiamentu and Palenquero) beside Afro-

Caribbean varieties of Spanish showing intriguing

similarities between them (see list above) is a strong

motive for investigating the likelihood of a pan-Caribbean,

proto-Afro-Hispanic language. What is often overlooked is

the fact that Mexico borders not only on the Gulf, but also

on the Caribbean and was a major destination for west

Africans for more than three centuries. The following

chapter examines historical texts from Mexico in order to

determine to what degree Africanized speech is represented

in Spain's largest and most important New World colony.


1. These investigations into Afro-Hispanic speech are
coincidentally contemporaneous with studies relating to the
variety of English associated with African descendants in
the United States (e.g. Wolfram 1969, Dillard 1972, Labov

2. The minimum number of languages in contact required for
the emergence of a pidgin language is contested. Classical
definitions (e.g. Whinnom 1965) stipulate there must be at
least three; others allow the broader definition adopted
here: "A pidgin language is defined as a reduced language
with input from speakers of at least two linguistic
communities whose languages are mutually unintelligible"
(Clements 1992:77; see also Fox 1983:105; Bakker 1995:29).

3. What non-European cultures think or have thought of
pidgin and creole languages is far beyond the scope of this
study, but Romaine (1988:109-13) offers the opinions of some
native creole speakers. Among Western linguists, Hugo
Schuchardt (1842-1927) was among the very first Europeans to
study these languages in any depth and is therefore
considered the father of modern pidgin and creole studies.

4. The most famous examples of communicating via signs and
gestures are the "sign language" used among the indigenous
people of North America's Central Plains and the signing
systems used by Deaf communities (Bakker 1995:29).

5. Bickerton (1994:122) maintains that pidgins, together
with the speech of children under two years of age and the
speech of adults who have been deprived of language in their
formative years, do not constitute "normal human language,"
but rather a "protolanguage," which can also be produced by
trained apes. A jargon or pre-pidgin would, by Bickerton's
criteria, also be considered "protolanguage."

6. Holm (1988:4-5) offers a lengthy definition and makes
special reference to deliberate simplification employed by
superstrate speakers in the genesis of a pidgin. Bakker
(1995:26) is more succinct. This summary of similarities is
derived from both authors' descriptions.

7. Naro's (1978) conclusions have recently been challenged.
For example, Clements (1992) maintains that Naro's
"reconnaissance language" was not a new pidgin variety but
"merely an instance of foreigner talk" (75). The island of
Arguin as a site where pidgin Portuguese might have emerged
has also been proposed as an alternative to a European
origin. A sometimes bitter exchange on these topics between
his critics and Naro can be found in Goodman (1987a), Naro
(1988), Goodman (1988), Clements (1992), Naro (1993) and
Clements (1993).

8. There is earlier (11th century) evidence, however, of
Africans speaking a pidginized Arabic in Mauritania (Holm

9. Although the compilation Cancioneiro Geral is dated
1516, the first poem in Africanized Portuguese ("A min rrey
de negro estar Serra Lyoa") is believed to date from 1455
(Teyssier 1959:228). The dating of this poem also figures
in the controversy mentioned in note 7 above.

10. See also Megenney (1990d) for an examination of creole
elements in 16th century Portuguese texts by Henrique da
Mota, Antonio Ribeiro as well as by Gil Vicente.

11. Only those phonological features that have particular
relevance to and resonance in later manifestations of Afro-
Spanish are listed here. See Teyssier (1959:243-50) for his
more complete description. Megenney (1990d:362, 371) also
describes many of these same phenomena in the texts of other
authors of the Portuguese Renaissance.

12. As in Spanish, the Portuguese multiple vibrant is
depicted word-initially in standard spelling by a single
letter r, and so any differences in pronunciation in this
position would not typically have any ready orthographic
representation. There are exceptions, however, as we see in

the mid-15th century poem "A min rrey de negro estar Serra

13. Translations are my own throughout and are provided as
a courtesy to interested readers.

14. The bibliography relating to Dozal language of the
Spanish Golden Age (roughly 1550-1650) is extensive. Some
of the principal articles treating this variety are Chasca
(1946); Castellano (1961); Weber de Kurlat (1962); Jason
(1967); Granda (1978:216-33). Lipski (1986a) is a detailed
comparative analysis of the features associated with bozal
Spanish over the course of 300 years; Lipski (1986b) is an
excellent succinct examination of the poetic and theatrical
manifestations of bozal Spanish specifically during the
Spanish Golden Age (siglo de oro). It is Lipski (1986b) who
notes that bozal Spanish did not always blindly follow the
original Portuguese literary models, but followed its own
course of evolution, eventually discarding the more pidgin-
like constructions of its Lusitanian antecedents.

15. Even here, in keeping with the fractious nature of
linguistic definitions, there is at least one objection to
the use of the term "native speakers;" some prefer to refer
to the creole as being the "principal language" of its
speakers (MHhlhausler 1986:9).

16. The literature treating theoretical issues of pidgin
and creole development is extensive and widely available.
Some of the most thorough recent accounts are M~hlhdusler
(1988), Holm (1988), Romaine (1988), Arends, Muysken and
Smith (1995). Additional helpful readings are found in
L6pez Morales (1989), Fasold (1990), McMahon (1994), Romaine
(1994) and Rickford and McWhorter (1996).

17. It is difficult to make a completely clear distinction
between the grammatical characteristics of an expanded
pidgin and a creole. The primary difference, of course, is
that an expanded pidgin has no native speakers and is the
object of second language learning; the creole is typically
the product of first language acquisition. Bakker
nevertheless notes that "Apparently, the criterion whether
or not a language is learned as a mother tongue is not
sufficient to allow us to distinguish pidgins and creoles
structurally" (1995:27).

18. Briefly, the language bioprogram hypothesis argues that
children who have been exposed to inadequate first language
input (e.g. a pidgin) for acquisition will produce
grammatical rules for which they have not had models
provided, thus giving rise to a creole. This language
bioprogram is a species-specific, genetic endowment

compatible with--although not identical to--Chomsky's
Universal Grammar (Bickerton 1981:298; cf. also Chomsky
1988). This hypothesis has been and remains controversial;
its content and aims are outside the scope of the present
work. See Holm (1988:61-65) for a useful overview of the

19. The leap from pidgin to creole in the span of only one
generation has been proposed, for example, in the
development of Hawaiian Creole English (Bickerton 1980,

20. A summary and discussion of language universalist
approaches is found in Muysken and Veenstra (1995); a
discussion of imperfect second-language learning and the
"baby talk/foreigner talk" issue is available in Besten,
Muysken and Smith (1995). See also Ferguson (1975) for his
seminal experimental exploration into the nature and origin
of foreigner talk. Naro (1978) is an eloquent exposition of
the foreigner talk origins of a pidgin; on this topic see
also Clements (1992).

21. Granda (1978) is a compilation of his most important
articles, including his earliest on this topic.

22. The process of relexification does not always result in
the creation of a creole language. Media Lengua in Ecuador
is an example of a speech variety that has undergone massive
relexification from Spanish but retains its Quechua
grammatical structures. Languages such as Media Lengua and
Michif (Cree grammar, French lexicon) are not considered
creoles proper, but "mixed languages" (Bakker and Muysken
1995; Muysken 1997).

23. As a consequence of the intense interest in the circum-
Caribbean varieties of Spanish as they relate to a proposed
Afro-Portuguese pidgin or creole, a substantial bibliography
has arisen, only a small portion of which is cited here.

For Puerto Rico: Alvarez Nazario (1974) treats the
African presence on the island and their linguistic
contributions in brilliant detail; Maule6n Benitez (1974) is
a primarily phonological examination of the speech of an
Afro-Puerto Rican village supplemented by some lexical
information; see also Granda (1978).

For Cuba: Ortiz (1924); Otheguy (1973); Martinez Gordo
(1982); Fernandez Marrero (1989); Perl (1984, 1985, 1988,
1989a, 1989b, 1989c); Ortiz L6pez (1996).

For the Dominican Republic: GonzAlez and Benavides

(1982); Terrell (1982); Megenney (1982, 1990a); Lorenzino

For Venezuela: Megenney (1985a, 1990b, 1990c); Alvarez
(1991, 1992).

For Colombia (excluding treatments of Palenquero):
Fl6rez (1951), Granda (1977, 1978); Del Castillo Mathieu

For Panama: Lipski (1990b).

24. Granda acknowledges that Sandoval knew Portuguese
"perfectamente bien" 'perfectly well' (1988:42). At the
same time he also cites a passage from Sandoval referring to
the need and use of African interpreters (1988:46); there
is, however, no mention of this fact in Granda (1978).

25. The salient phonological aspects of a "radical" Spanish
dialect are not listed explicitly by Guitart (1978); these
have been adduced by other linguists and include features
commonly associated with Afro-Hispanic speech. Rather, the
article cited provides a convenient rubric for
classification that has been widely adopted.


One of the greatest difficulties confronting the

researcher in Mexico is the general paucity of historical

texts purporting to represent Afro-Hispanic language in the

subcontinent (Zimmermann 1995). There are, nevertheless,

some sources, fairly well distributed diachronically, that

do portray the Spanish as spoken by Africans in Mexico.

Although there is a rich corpus of Iberian bozal texts, the

written sources of "black" speech in the New World are

fairly scant (cf. Lipski 1993). The only available

historical texts of Africanized Spanish in Mexico come from

poetic and musical compositions.

Mordant poet, nun, proto-feminist and the "tenth Muse"

of Spanish letters, Sor ('Sister') Juana Inds de la Cruz

(1651-1695) occupies a unique place in Mexican humanism and

intellectual activity; her life and writings have inspired a

prodigious biographical and critical bibliography. Her

poetic production includes villancicos 'religious songs'

celebrating various Church feast days. Among these are

vinQi2o purportedly as sung by Africans, such as the
following excerpt from the feast day of San Pedro Nolasco,

1677. Rather than its being a laudatory, jubilant or


reverential song, the content below shows a sharp criticism

of the Church (Mansour 1973:61), couched in

characteristically bozal language:

La otra noche con mi conga
turo sin durmf pensaba,
que no quiele gente plieta,
como eya so gente branca.
Sola saca la Pahola;
ipues, Dioso, mila la trampa,
que aunque neglo gente somo,
aunque nos dici cabaya!

The other night with my Congo woman
sleepless I was thinking
that she [the Church] doesn't want black people,
only white people like her.
She only saves the Spaniards,
but God, look at this deception,
that although black, we are people,
though she calls us horses!' (Mansour 1973:61)

This text contains many examples of the well-known

interchangeability of the liquids /1/ and /r/ (/r/ > [1] in

giee, p mila, neglo; /l/ > [r] in branch); there is

also substitution of /d/ > [r] in turo; loss of infinitive-

final /r/ (durm); loss of final /s/ (sozo < somos); certain

vocalic shifts, most particularly raising (turo, durmi <

t dormir; dici < dice); erroneously marked gender and

number (l, cabaya < los esnafioles, caballos);

paragoge (Dioso < Dios) and aphaeresis (Pahola < espaholes).

There are two possible Portuguese forms in so 'only' and

rnca 'white' (cf. Spanish s61o and blanca), and the

raising of unstressed /o/ to [u] may be easily accepted as a

Portuguese influence as well. Syntactically, there are no


discrepancies in this text from standard Spanish, confirming

Lipski's (1986b) observation that early, pidgin-like

features evolve out of late Golden Age literature. In

short, this is a classic bozal text from the late Spanish

Golden Age, which even contains the formulaic lament of the

enslaved African, "aunque neglo gente somo."

Paradoxically, it is the very classic nature of this

text, representative of a particular type of poetry, that

weakens claims to authenticity.' Although Sor Juana was

resident in Mexico City and would have known of Africans in

the city, she was also entirely conversant with contemporary

trends in Peninsular literature including the figure of the

bozal and his characteristic speech as found in the work of

writers such as Quevedo and G6ngora; moreover, she had been

in Puerto Rico and knew bozal speech first-hand. Sor

Juana's examples are important, nevertheless, for situating

an Africanized Spanish in the capital of the Viceroyalty.2

Parochial villancicos of Puebla

Also within the same tradition of religious songs is a

collection of villancicos brought to light by Megenney

(1985c; but see also Stevenson 1974, 1975). Discovered in

the cathedral of Puebla, the possible dates of their

transcription range over more than a century, from 1606 to

1715 (Megenney 1985c:162). These villancicos present

characteristics identical to those written by Sor Juana, and

among them is one with a distinctly Portuguese base; this

last fact does lend support to the hypothesis of a

pidginized Afro-Portuguese spoken among Africans in Mexico.

The other five villancicos have rather more boZal elements

than Sor Juana's, reflecting an early date of probable

composition, independent of their transcription date. The

more bozal features in the remaining five ji c that

Megenney identifies as having a Spanish base include such

forms as noso (Portuguese nosso, Spanish nuestro 'our') and

siolo, siQ (Spanish sehor, Portuguese senhor 'sir'), as

well as the very bozal copula g& 'is/are/be', (see e.g.

"Guineo a 5" in Megenney 1985c:175-76; "Negrilla," 188-89).

The discovery of these texts is fairly compelling evidence

of an early Afro-Hispanic speech, heavily influenced by

Portuguese, if not a pidgin or creole per se.3

18th Century Songs and Dances

As the slave trade declined rapidly in Mexico during

the 18th century (see Chapter 4) there was nevertheless

continuing contact with Cuba via the Gulf coast port of

Veracruz. Cuba had a vigorous slave trade during the late

18th century which peaked by the mid-19th century as its

sugar cane industry was developed. Afro-Cuban and Afro-

Cuban-influenced songs and dances achieved wide currency in

many parts of Mexico, including the so-called Facico songs

(from the Africanized pronunciation of the name Franci).

These Facicos, which were apparently widely known and

circulated in Mexico as late as the 1950s, often included

lyrics that have clear Afro-Portuguese influences: "Maria

Facica querd casd / con home banco de Portug&" 'Maria

Francisca wants to marry / a white man from Portugal'

(Mendoza 1956:1103). Various dances such as the sacamandi,

chucumbe and m were known as "African" and were also

frequently denounced to the Inquisition and condemned by the

Church as being too suggestive. In all cases these dances

appear to have been learned in Cuba and brought to Mexico.

The lyrics of "black" songs in the early 18th century

maintain a definite bozal flavor, (i.e. forms such as Dioso,

sa, the prepositional particle na, etc. appear in the

lexicon) but this "bozal effect" is most frequently achieved

by means of the formula of interchanging liquids /l/ and

/r/, and by including onomatopoeic "Africanisms" such as cu-

cu-cd and Zanguangu in refrains (see especially "Ha

negliyo" in Mendoza 1956:1102-03). The masked Danza de los

Negros 'Dance of the Negroes' among the indigenous Mixe
people in Oaxaca state attests to the African presence on

the Pacific coast, and it has been suggested that it may

actually have originated in the Costa Chica of Guerrero

(Mendoza 1956:1106). This assertion appears dubious,

however, as current cultural expressions among the

Afromestizos of the Costa Chica, curiously enough, do not

include any of the dances or songs mentioned by Mendoza as

typical of the Gulf coast or of Oaxaca state, nor have they

been recorded by other researchers (Aguirre Beltr~n 1958,

Manzano Ahorve 1991). The local dances of the Costa Chica

include the artesa, danced by a couple upon a tarima

(hollowed log shaped like a canoe); and the chilena, which

originated from the cueca sung and performed by the crew of

a Chilean squadron visiting Acapulco in the early 19th

century (L6pez Barroso 1967:76-77; Luna Mayani 1976:195-96).

The corridos 'folk ballads' for which the residents of

Cuajinicuilapa and environs are most renowned do not include

any of the Portuguese-like forms attested in the 18th

century "black" songs (see Gutidrrez Avila 1985).

Modern Descriptions of Costa Chica Spanish

The earliest modern reference to the Spanish spoken in

the Costa Chica comes from an account provided by a census

taker in 1870:

Los habitantes que forman la municipalidad son
descendientes de raza africana, por lo que su idioma es
el espahol antiguo, y no poseen mds industria que las
siembras de algod6n y maiz, pero en corta cantidad, asi
como la cria de vacas y bestias caballares.

'The inhabitants who make up the municipality are
descendants of the African race, for which reason their
language is old Spanish, and they have no industries
other than growing cotton and corn, but in small
quantities, as well as cattle and horse raising' (J. M.
L6pez, cited in Aguirre Beltr~n 1958:63).4

Unfortunately, what L6pez meant by "old Spanish" spoken

by the descendants of Africans is open to speculation. In

previous work, it has been suggested that L6pez "heard their

language as something 'other' than his own and that .

the African ancestry of the people had something to do with

the way they spoke Spanish" (Althoff 1994:247). We have

seen in the previous chapter that there are many features

that, taken together, have been used as keys to the

identification of Africanized Spanish. What features L6pez

may have had in mind as he commented on their language is

beyond learning, but his is a prescientific recognition of

the distinct nature of a regional speech that may have

originated among maroons living in the area since the late

1500s (see Chapter 5).

In this century Heredia (1935) presents a modest

description of the Spanish spoken in the Costa Chica town of

Ometepec, an important regional center. Chiefly consisting

of a list of regional lexical items, Heredia's article notes

the common raising of /e/ to [i] (e.g. ensehar > inseiar 'to

teach') and the near-universal aspiration of syllable- and

word-final /s/. She points out commonalities between the

vocabulary of Ometepec, located very near the Pacific coast,

and that of the YucatAn, Campeche and Tabasco on the Gulf of

Mexico. Although many of these words are identical in form

on both coasts, there are regional differences in meaning.

The similarity in vocabulary, however, is suggestive of some

possible connection between the two varieties across the

Isthmus of Tehuantepec (these implications are discussed in

Chapter 8).

The most complete description and analysis of Costa

Chica Spanish is found in Aguirre Beltr~n (1958). As part

of an ethnographic treatment of the town of Cuajinicuilapa,

Aguirre Beltrdn has described a number of characteristics

that coincide with many of the features described above as

typical of bozal Spanish and Portuguese. Basing his report

on fieldwork carried out in late 1948 and early 1949, he

records (1958:203-13):

(1) Usual aspiration and even loss of syllable- and

word-final /s/;

(2) Paragoge, particularly in forming the diminutive

of proper nouns (Ram6n > Monche; SebastiAn > T );5

Aguirre Beltr~n does not record any examples of the type

sefor > sehoro as seen in earlier Afro-Mexican texts;

(3) The wide use of the second person singular

pronoun vos (transcribed DQd) with accompanying distinctive

verb morphology;

(4) Word-final /r/ was often eliminated (dot6, cal6,

Rr< doctor 'doctor', calor 'heat', Rrir 'to give


(5) Loss of word-final and intervocalic /d/ (verdA <

verdad 'truth', toabia < todavia 'still');

(6) Evidence, again from the diminutives of proper

nouns, that both intervocalic /d/ and /r/ may be substituted

by [1] (Isidoro > Lita, > Lila);

(7) Widespread variation in standard vowels,

typically in atonic syllables (olesa < pii 'police',

mormuyo < muai1l.Q 'murmur');

(8) Aphaeresis (tubo < estuvo '3 sg. was', horito <

sehorito locally 'patriarch of extended family');
(9) Apocope (pA < Dar 'for', g& <.casa 'house', DU

< p '3 sg. is able');

(10) Prothesis (atraisionar, abntz < traicionar

'to betray', levantar 'to raise').

Among his other findings are a rare, but locally

general, behavior of /f/ (analyzed in Chapter 7) and a

distinctive regional lexicon already noted in part by

Heredia (1935). Aguirre Beltr~n is very aware, however, of

the potentially strong African contribution to the language

of the culehos 'residents of Cuajinicuilapa', and notes

that the differences in speech there are significant and

numerous enough to "surprise" anyone who converses with them

(1958:201). He points out two factors involved in the

likely development of the local variety: the isolation in

which the inhabitants lived until earlier this century and

"la variabilidad de lenguas habladas por quienes fijaron su

residencia definitiva en la planicie costanera" 'the

variability [sic] of languages spoken by those who settled

permanently on the coastal plain' (1958:202). Following

Granda's (1978:372-73) recommendation, it has been noted

above that the element of isolation in particular makes this
. 'county' an attractive site for investigating

potential African or Afro-Portuguese linguistic links. The

already-attested traits of the local speech which have been

correlated with other Africanized varieties of Iberian

languages, serve to confirm the suitability of the .

of Cuajinicuilapa for further investigation (see also

Chapters 5 and 6).

L6pez Barroso (1967) is a compendium of information

relating to the judicial district of Abasolo which includes

the of Cuajinicuilapa. Although there is no

specific information on syntax or phonology, among its

contents is a regional word list that contains some items of

interest. Consistent with an Afro-Portuguese or bozal

origin are the words nenco (cf. Spanish negro) 'black', paye

(padre) 'father' and &udo (nudo) 'knot' (1967:189).

Although these forms are not recorded specifically in the
. of Cuajinicuilapa, their attestation within this

Costa Chica district, headquartered in the regional center

of Ometepec, would reasonably allow their use as collateral

data. The phenomenon of intrusive nasalization appears in

nenco, similar to attested nengro, nengrg < n in bozal

speech (Lipski 1992). The palatalization of standard /n/ to

[fi] (nudo > &Ud) has also been identified with Afro-

Hispanic phonotactics, although this has also been found in

other non-Africanized dialects (Lipski 1994a:126). The

appearance of R so similar in form to Portuguese VA&

'father', is intriguing, but there is no corresponding

feminine form included in that short lexicon (but see

below). These forms, however, contribute to a picture of

the regional speech that may plausibly include Africanized


The most important collection of regional vocabulary

and expressions is undoubtedly Aparicio Prudente et al.

(1993b). Although this work is strictly descriptive, the

forms found therein greatly expand the lexical data

available for study. Consisting of approximately 1700

entries described as being in daily use, it is divided into

sections describing illnesses, flora and fauna, regional

cooking and foods, agricultural implements, weights and

measures, names, surnames and hypocoristics, toponyms and

miscellaneous expressions and sayings. Among the

expressions recorded are words referring to the language of

both regional and non-regional speech groups. The editors

note that the coastal residents recognize each other through

the distinctiveness of their speech (Aparicio Prudente et

al. 1993b:xi, 15) and refer to the standard speech as

"hablar fisico" 'physical talk', a style of speaking "sin

contraer las palabras, aspirar ni suprimir las letras"

'without contracting words, aspirating or suppressing

letters'. The residents of the rnmiciQ of Cuajinicuilapa

themselves say they speak "mocho" 'choppy' or "payo"

'countrified'. The data contained here show the

complementary forms m and Day_ (1993b:15, 17), which are

forms used to call out to a girl and boy, respectively, when

one does not know the child's name. Confirmed again is the


phenomenon of intrusive nasalization (muncho < mucho 'much';

nunca < nuca 'nape of the neck' [1993b:17-18]). Finding its

first mention in the literature for this region is the

frequent tendency for /o/ to be raised to [u], both in tonic

and atonic positions. The editors, employing a digraph ou

to represent this raised vowel, record creou, fi and

gourda < creo 'I believe', frIQ 'cold' and gord 'fat'


Academic work published in the 1980s and early 1990s

relating to the of Cuajinicuilapa has otherwise

focused on literature of the oral and folkloric traditions

(Neff 1988, Gutidrrez Avila 1988b, 1993; Gutidrrez Avila and

del Rio Azurmendi 1988) as well as on the nature and

function of the corrido 'folk ballad' in Costa Chica society

(Gutidrrez Avila 1985, 1988a, Moedano 1988). While all of

these studies have intrinsic ethnographical and

anthropological interest, as a rule they shed no new light

either on the historical origins of the African-descended

population or on their speech. Anthropology, in fact, has

been the main discipline driving research in and about the

Costa Chica (see Chapter 6); the appearance of Washington

(1992) follows in this same social sciences tradition.

The stories and legends in Aparicio Prudente et al.

(1993a) make it exceptional in providing transcriptions that

largely follow the conventions of this study. Based on

taped interviews with local residents, these narratives not

only offer insights into local lore but into the structure

of their language as well. The most striking linguistic

feature of this book, however, is the glossary of local

terms and usages, but on a lesser scale than that found in

Aparicio Prudente et al. (1993b).

Linguistic studies of a more formal, academic nature

apart from the information provided in Aguirre Beltr~n

(1958) have been all but nonexistent. The most recent work

bearing specifically on the language of the ..i of

Cuajinicuilapa is Althoff (1994). This article offers some

history on the origins of the residents, recapitulating some

of the information described in Aguirre Beltr~n (1958), and

provides some new data on the local speech. In the context

of increasing urbanization, mass media and education, this

article predicts the eventual disappearance of some of the

more distinctive features of the local phonology but the

survival of the unique regional lexicon (Althoff 1994:253-


Beginning with the classically bozal~ yjvil igQ of the

Golden Age, the speech of Africans and their descendants has

been depicted on Mexican soil. Because the African was so

often depicted in the theatre as a humorous figure speaking

outlandishly, the reliability of these texts as a record of

Africanized Spanish has often been called into question.

Although there were undoubtedly exaggerations and some

distortions of actual features, this study aspires to the

position described by Lipski (1992:285), neither to accept

the texts uncritically nor reject them out of hand: the

working (and necessary) assumption here is that Spanish

bozal speech, although prone to be exaggerated by dramatists

and poets (e.g. in the wildly indiscriminate interchanging

of liquids /l/ and /r/), was more or less accurately

recorded in their compositions. More recent descriptions

and observations of local language, although often

impressionistic, may be regarded as good-faith attempts to

record accurately those linguistic features regarded as

especially typical of the region.

In assessing Afro-Hispanic texts in Mexico, we have

seen definite parallels between the characteristics

associated with bozal Spanish as well as some associated

with (Afro-)Caribbean Spanish. In particular, the evidence

described by Aguirre Beltr~n (1958), augmented by later

descriptive works and compendia, and the known history of

the j .Q and the Costa Chica (see Chapter 5), suggest

the possibility that an Africanized variety of Spanish, if

not a full creole, could have been spoken along the south

Pacific coast of Mexico.

The following chapter examines details of the slave

trade to Mexico that seized Africans and brought them to

this continent. More importantly, it examines the

conditions under which Africans were forced to live at

different points in their enforced immigration and the

potential effects those living conditions might have had on

language acquisition.

1. The accuracy of slave speech in the United States as
depicted in the songs of Stephen Foster or the novels of
Mark Twain (Samuel Clemens) may be interrogated in the same
way: there is certainly some basis in linguistic reality,
but there is also undoubtedly some degree of stereotyping
and exaggeration.

2. African languages were also heard in colonial Mexico
City, although it was usually dangerous for anyone, European
or African, to speak them: Europeans might fall under
suspicion of being crypto-Jews and Africans of fomenting a
slave rebellion. Alberro indicates that "Angola" language
was sometimes used by Portuguese merchants living in Mexico
City who had learned it in Africa or from their slave
nannies (1979:146). Portuguese merchants were frequently
suspected of secretly being Jews, who were distinctly
personae non gratae in Spain and especially in Spain's

3. The texts found in Stevenson (1974, 1975) reflect very
much the characteristics described in Megenney (1985c).

4. Aguirre Beltr~n (1958:63, 239) cites a text I have been
unable to locate but whose publication data are reproduced
here verbatim:

Arce, Francisco 0. 1870. Memoria presentada ante la
H. Legislatura del Estado de Guerrero, por el Gobernador del
mismo: J. L. L6pez. Estadistica del Distrito de Ometepec,
fechado en Ometepec, Enero de 1870. Ciudad Guerrero.

5. Although Aguirre Beltrdn devotes a great deal of space
to discussing local diminutives and hypocoristics, Lipski
(1995c) has shown, using the theory of prosodic morphology,
that these forms are produced according to unvarying
patterns across all dialects of Spanish.


Throughout its history as an independent nation, and

especially since the Revolution of 1910, Mexico has proudly

considered itself to be a country that is mestizo to its

very core, a raza c6smica 'cosmic race' born of the union of

the Spaniard and the indigene. Moreover, this conception of

Mexican national parentage is supported and reinforced by

the existence and activities of official bodies such as the

Instituto Nacional Indigenista and the Consejo Nacional para

la Cultura y las Artes, among others. In the popular

consciousness as well as in the official textbooks and

histories, the contributions of native peoples and Europeans

are those that constitute the essence of the process of

miz ~j.e resulting ultimately in la mexicanidad, a

distinctive national, cultural and ethnic identity. And

while it would be impossible to underestimate the importance

of either the indigenous or European lineage to this

process, it is apparent that the reality of the Mexican

ethnic heritage is somewhat more textured than is generally

understood. A truly complete account of Mexico's history

must include mention and discussion of the African presence

on the subcontinent.

While people of African descent in many Spanish-

speaking American countries, such as Cuba, Puerto Rico, the

Dominican Republic, Panama, Colombia and Venezuela, comprise

a large and visible segment of the population, their seeming

absence in contemporary Mexico may suggest, even among some

Mexicans, that there was never any African presence in the

country (Palmer 1976:vii).

This misapprehension is rather easily explained by

taking into account a couple of factors. First, Africans

and their descendants in Mexico have always been a very

small minority in an overwhelmingly Amerind population. The

exact number of Africans introduced into New Spain (Mexico)

as slaves has never been established with complete certainty

due to some incomplete records and what is presumed to have

been a very high degree of slave smuggling (Palmer 1970:16-

17; 1976:2). The most commonly accepted estimate is that

200,000 slaves were imported during the course of almost

exactly three centuries beginning in 1519 (Aguirre Beltr~n

1944:431; Curtin 1969:46). Although New Spain was at one

time (1570-1650) the largest holder of African slaves in the

New World (Palmer 1970:4), in comparison to the indigenous

population, both the Africans and the Spaniards remained

distinctly minority populations: taken as a single sum,

these two groups never constituted any more than one to two

percent of Mexico's entire population (Aguirre Beltrdn

1972:198). In the 16th century, 71 percent of the immigrant

population was African, 29 percent European. This

proportion remained more or less constant in the 17th

century, but by the 18th century had become 65 percent

African and 35 percent European. As a percentage of the

entire Mexican population, these figures increase if we

consider the combined African and mixed African-indigenous

(i.e. Afromestizo) population, but still indicate a minority

status. The numbers in this combined African-Afromestizo

category range from approximately seven-tenths of one

percent of New Spain's population in 1570 to 10.2 percent in

1810. The figures for the combined indigenous and European

groups are 99.07 percent and 89.4 percent, respectively

(Aguirre Beltr~n 1972:234).

The other factor accounting for Afro-Mexico's relative

invisibility is cultural rather than demographic. While it

is true that the relative numbers of Africans and their

descendants in Mexico were never large, their presence,

contributions and importance were additionally minimized--

indeed, made nearly to vanish--by "la magnitud que entre

nosotros siempre alcanz6 la naturaleza mistica de lo"ndio"

'the magnitude that the mystical nature of everything Indian

always achieved among us' (Aguirre Beltr~n 1958:11; original

emphasis). In spite of the initial interest generated

primarily by Aguirre Beltr~n's historical and ethnographic

work of fifty years ago, scholarly production addressing

Africans and African-descended culture in Mexico remained

relatively scarce until the mid-1980's when a renewed

interest in Mexico's African roots became evident.

The Slave Trade to Mexico

of all the colonial powers later to employ involuntary

servitude in their overseas territories, Spain was the first

to institute the "peculiar institution" in the New World

(Palmer 1976:1), and the first half-dozen Africans to arrive

in Mexico came with the original conguistadores (Aguirre

BeltrAn 1972:19; Meyer and Sherman 1995:215). The

definitive arrival of the Spanish in 1519 marked not only

the beginning of the apocalyptic end of the Mdxico-

TenochtitlAn empire, but also the beginning of the African

presence in Mexico.

The Spanish slave commerce in Africans certainly did

not begin with explorations and expeditions to the New

World, although it did indeed reach its highest levels in

the new hemisphere.' Slaves, both from north and sub-

Saharan Africa, were a common sight on the Iberian

peninsula, particularly in Seville (Pike 1967) and Lisbon

(Saunders 1982). These two cities came to be the most

important centers for the slave trade in Europe. In Spain,

MAlaga, CAdiz and Valencia also reported substantial numbers

of slaves (Palmer 1970:6; Rout 1976:16). The sub-Saharan

African slaves permitted to be sold in Spain were required

to have received at least rudimentary instruction in

Christianity as well as baptism. These partially

Europeanized Africans, or laino, were likewise the only

ones originally permitted into the New World, but as the

native populations in the Indies (and later in Mexico)

succumbed to the devastating results of epidemics and

mistreatment, the demand for labor in the overseas Spanish

colonies would grow extremely acute.

The Spanish slave traffic to the Americas began even

before Columbus's fourth and final voyage (1502-1504) to the

New World. The first regulatory measures by the Spanish

Crown regarding slavery in the New World were instituted in

1501 with instructions to Nicolds de Ovando, governor of

Hispaniola, forbidding slaves who were Jews, Moors or new

converts to Christianity2 to land on the island. These

orders called instead for the importation of negros ladinos,

the Christianized Africans, who were considered to be free

from exposure to pernicious and heretical ideas (Aguirre

Beltrdn 1972:16).

As noted above, it was clear, however, that the need

for labor would eventually outstrip the available number of

Christianized African slaves who had spent the requisite

amount of time on the Iberian peninsula for some minimal

catechization and acculturation. The year 1518, a scant 17

years after Ovando received his orders, was of special

significance in the history of slavery in Spanish America

for two reasons: first, the Spanish Crown gave its

permission for non-Europeanized, non-Christianized'


Africans, the so-called negros bozales 'wild Negroes', to be

conveyed directly from Africa to Hispaniola; and second, the

Crown approved an organized, large-scale trade in slaves.

In this year, a cjdula 'decree' incorporating both of these

points was granted by the Emperor Carlos V to one of his

court favorites, Laurent de Gouvenot. The cddula authorized

4,000 slaves to be sent to the Indies over the course of

nine years. Gouvenot in turn sold the 4,000 li nias (one

licencia being required to import one slave) to three

Genoese merchants living in Seville. These merchants agreed

to deliver 1,000 slaves to Cuba and the remaining 3,000 to

Hispaniola. They further managed to reserve to themselves

the right to dispatch an indeterminate number of African

slaves to the Yucatan and other parts of the Mexican

mainland, whether discovered or yet to be discovered

(Aguirre Beltr~n 1972:17-18). Thus began the organized--and

eventually enormous--slave trade to Spanish America which in

turn led to a massive infusion of African blood into the

racial mix of the Spanish colonies. The cddula granted to

Gouvenot in 1518 can also be considered the first instance

of an asiento, the official monopoly granted to an

individual slave trader or slaving company to supply the

almost insatiable demand for labor (Palmer 1976:9). The

reader should keep in mind that the Spanish, while

enthusiastic customers who provided a strong and ready

demand, never really controlled the supply of slaves.

Beginning with Gouvenot's Genoese subcontractors, the

traffic in slaves was, with some exceptions, always managed

and directed by non-Spaniards.

First Period of African Slavery in Mexico

There were three main periods of forced African

migration into Mexico (Palmer 1976:3). The first period,

which may fairly be called the periodo de las licencias

'period of licenses' (Aguirre Beltrdn 1972), began in 1519

with the arrival of the Spanish, and ended in 1580 after a

disastrous epidemic. Between 1519 and 1580 the slave trade

was carried on under the authority of numerous a and

lncia granted to individuals of various nationalities.

In many cases the individuals authorized to import slaves

were simply "los numerosos empleados que pasaban a la Nueva

Espaha a ocuparse de su administraci6n" 'the numerous

employees who were crossing to New Spain to take charge of

its administration' (Aguirre Beltr~n 1972:20). The largest

single number of jfl ja granted by the Crown to an

individual during this time was 23,000 to Hernando de Ochoa

Ochandiano in 1552. The contract was to have been a

monopoly of seven years, but it appears this particular

contract never actually took effect (Aguirre Beltr~n

1972:23). The size of the contract does serve, however, to

indicate how hugely demand had increased from the original

Gouvenot aiento for 4,000 slaves in 1518.

During the 61 years of the first period of African

slavery in Mexico, literally thousands of licncia were

issued to various courtiers, officials, explorers, settlers,
conquistadores (including Herndn Cortds in 1542) and

entrepreneurs. Indeed, licencias were available to nearly

anyone who could afford them: the Crown had discovered that

either issuing licencias with a tax or simply selling them

on the open market was a very good means of raising revenue

to pay for His Majesty's wars (Aguirre Beltrdn 1972:23).

Second Period of African Slavery in Mexico

The second period, from 1580 to 1650, witnessed the

height of slavery in Mexico, both in terms of numbers of

slaves in the country and in its economic significance

(Palmer 1976:3). This is also the period corresponding

broadly to the years of the unified Crowns of Portugal and

Spain. King Felipe II, son of the Emperor Carlos V of Spain

and Princess Isabel of Portugal (hence grandson and

legitimate heir of Manuel I of Portugal), ascended the

Lusitanian throne in 1581 following the death of the

childless king Sebastian in 1578 (Cantarino 1995:143).'

The political unity of the two kingdoms further consolidated

the Portuguese hold on the Spanish slave trade. Indeed, the

slave trade had to that point been so organized as to make

the Portuguese indispensable to the traffic (Aguirre Beltrdn

1972:33). By virtue of the 1479 Treaty of Alcagovas, Spain

had already ceded to Portugal the exclusive rights to trade

and explore on mainland Africa (Rout 1976:15). This

arrangement was rather strictly observed by the Spanish, as

evidenced by the fact that Spain's only sub-Saharan African

possession, Spanish Guinea (since 1968, the Republic of

Equatorial Guinea) was only officially acquired by them in

1778, with effective settlement coming much later (Lipski


Towards the end of the 16th century, the Portuguese saw

their effective slaving monopoly officially sanctioned and

strengthened by the Crown of Spain. In 1595, Felipe II

ordered a special jun 'commission' composed of members of

the Real Hacienda 'Royal Treasury' and the Consejo de Indias

'Council of the Indies' to discuss a general contract for

sending slaves to Spain's American possessions. From that

point onward, this junta would re-form each time it was

necessary to establish a new asiento. It was in this same

year of 1595 that an asiento was granted to--or more

accurately, bought at auction by--the Portuguese Pero G6mez

Reynel, which specified that 3,500 slaves be delivered alive

(of the 4,250 to be embarked in Africa) to the port of

Cartagena on the coast of present-day Colombia. Two

thousand of these were to be distributed according to the

Crown's orders; the remaining 1,500 were permitted to be

sold as the a 'holder of the asiento' wished

(Aguirre Beltrdn 1972:38). Due to some unresolvable

conflicts with viceregal customs officials, the asiento with

G6mez Reynel ended in 1599, some five years short of its

originally stipulated length (Aguirre Beltr~n 1972:40-41).

The difficulties G6mez Reynel had in the administration

of his asiento were not sufficient to discourage other

bidders from seeking the monopoly. In 1601 another aiento

was awarded to Juan Rodriguez Coutinho (also known as Jodo

Roiz Coutinho), who, conveniently enough, was also the

governor of the Portuguese colony of Luanda on the Angolan

coast. (Luanda was one of the main African fti, or

warehousing and transshipment points, in the Portuguese-

controlled slave trade.) This aiento was plagued by

lawsuits occasioned under the previous anUsa ; it ended

finally with the death of Juan Rodriguez Coutinho himself in

1603 (Aguirre Beltr~n 1972:42-43).

The late governor's brother, Gonzalo V~ez Coutinho,

then picked up the remaining years of the monopoly under an

asiento dated 1605. The Coutinhos had obligated themselves

to provide 38,250 slaves to the service of Spain's New World

colonies over the course of nine years. This contract, as

the previous one, provided for a yearly number of 4,250

slaves; again, the annual number of slaves to be delivered

alive was only 3,500--a figure that allowed for an

appallingly high mortality rate of nearly 18 percent.

Aguirre Beltr~n notes that the majority of these people were

destined for arrival in Mexico, as New Spain was the only


market capable of absorbing such a large number of slaves at

that time (1972:43-44).

With the exception of a five year period (1609-1614)

following the Coutinho asientos, a period which was

characterized by contraband slave imports and the use of

some individual licencias, the Portuguese were in complete

possession of the slaving monopoly from 1595 until 1639.5

Aguirre Beltr~n indicates that the period from 1580 to 1640

also marked the height of the Portuguese slaving trade to

Spanish America (1972:48). The last asiento granted to the

Portuguese came to its conclusion at a time when Portugal

was beginning to reassert its autonomy from Spanish

political domination by armed resistance. During the period

of Portuguese hegemony over the Spanish slave trade, it is

estimated that 35,089 Africans arrived in New Spain (Aguirre

Beltr~n 1972:217-218).

Final Period of African Slavery in Mexico

The third and final period of slavery in this

chronology began in 1650 and lasted until 1827 (Palmer

1976:3). In contrast to the relatively uncomplicated

arrangements in the slave trade that had prevailed until

1650, the situation became increasingly fluid (Mellafe

1964:37). In light of her decline as the most powerful

economic and military power in Europe, Spain was forced to

deal more and more with Protestant slaving nations with

which she had less than amicable relations.

After the last Portuguese as came to its

conclusion in 1639, there was a period of approximately 20

years in the trade with no clear direction or ae sa. In

1662 a new contract was finally granted to three Genoese

merchants, Domingo Grillo, Ambrosio Lomelln and his brother

Agustin Lomelin, who were known collectively as the Grillos.

Although contractually they were permitted to seek slaves

from any nation other than Portugal willing to supply them

(including even the hated English), these merchants obtained

most of their slaves from the Dutch factor on Curagao

(Aguirre Beltr~n 1972:56). Portuguese sources had been

closed to the Grillos as a result of Lisbon's rebellion

against Spain, and although commercial and diplomatic

relations were soon restored between the two governments,

the Portuguese never again played a major role in

introducing African slaves to Spanish America (Aguirre

Beltrdn 1972:56-57).

The Grillos successfully fulfilled the terms of their

asiento, which were to bring 24,500 Africans to the ports of

Cartagena, Portobelo (Panama) and Veracruz over the course

of seven years.'

The next a, granted in 1674, was held by the

Spaniards Antonio Garcia and Sebastian de Siliceo. They

undertook to bring 20,000 Africans to America, at the rate

of 4,000 slaves for each of five years. Although it was

nominally controlled by Spaniards, it has been suggested

that this asiento served merely as a facade for the banking

and commercial interests represented by the Dutchman

Balthazar Coymans (Aguirre Beltrdn 1972:59-60). If this is

true, then the interests represented by Coymans were

apparently not concerned enough to rescue the a from

bankruptcy: in 1676 Garcia and Siliceo failed to meet their

financial obligations to the government, and the Spanish

Crown ordered the a into the hands of the corporate

trading enterprise called the Consulado y Comercio de

Sevilla 'Chamber of Commerce of Seville'. Garcia and
Siliceo had been able to hold the contract only thirteen

months (Aguirre Beltrdn 1972:60).

The Co!nsad met with only partial success in taking

over the previous asiento. In the face of large shipping

losses, the Cnsuldo itself was forced to seek permission

to place part of its obligation into yet another party's

hands (Aguirre Beltrdn 1972:60). Juan Barroso del Pozo, a

Genoese merchant living in Seville, then entered the slave

trade under the terms of the asiento granted to the

C. Fulfilling these terms, however, was extremely

difficult as this contract forbade that any slaves be

acquired from the Dutch-held island of Curagao. Barroso del

Pozo was in a quandary; he was unable to find any other

legally permitted suppliers able or willing to assist him

(Aguirre Beltrdn 1972:61).

Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ERH6MBVUC_43WU61 INGEST_TIME 2014-07-17T21:08:12Z PACKAGE AA00022865_00001